Workshop proceeding - final.pdf - Faculty of Information and ...
Workshop proceeding - final.pdf - Faculty of Information and ...
Workshop proceeding - final.pdf - Faculty of Information and ...
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
The 3 rd VACPS Research <strong>Workshop</strong><br />
Proceedings<br />
Victorian Association <strong>of</strong><br />
Chinese PhD Students <strong>and</strong> Scholars<br />
– VACPS –<br />
EN Building, Hawthorn Campus,<br />
Swinburne University <strong>of</strong> Technology<br />
24 Oct 2009
版 权 声 明<br />
经 作 者 同 意 , 本 论 文 集 收 录 了 第 三 届 维 多 利 亚 州 中 国 博 士 生 及 其 青 年 学 者 优 秀 论 文 报<br />
告 会 所 接 收 的 学 术 论 文 。 该 论 文 , 及 其 摘 要 、 图 、 表 和 公 式 等 , 其 版 权 属 于 作 者 所<br />
有 。 墨 尔 本 中 国 博 士 沙 龙 等 主 办 方 对 此 版 权 不 提 出 要 求 。<br />
未 经 作 者 本 人 许 可 , 任 何 人 不 得 局 部 或 者 全 部 变 更 、 转 载 、 复 制 、 改 动 、 散 布 、 展<br />
示 、 发 行 该 论 文 , 否 则 主 办 方 有 权 利 依 法 追 究 法 律 责 任 。 墨 尔 本 中 国 博 士 沙 龙 对 该 版<br />
权 声 明 具 有 最 终 解 释 权 。<br />
特 此 声 明 。<br />
墨 尔 本 中 国 博 士 沙 龙<br />
Disclaimer<br />
This <strong>proceeding</strong> is only for the archive purpose, <strong>and</strong> information included is authorized by the<br />
authors. Rights <strong>and</strong> intellectual properties related to the manuscripts, abstracts, figures, tables,<br />
equations <strong>and</strong> other materials are held by their respective authors. The Third VACPYS<br />
<strong>Workshop</strong> committee claims no copyright to the content <strong>of</strong> the manuscripts <strong>and</strong> abstracts.<br />
No part <strong>of</strong> this <strong>proceeding</strong> or the related files may be reproduced or transmitted in any form,<br />
by any means, without the prior written permission from the respective authors.<br />
VACPS reserves all rights <strong>of</strong> interpretation <strong>of</strong> this disclaimer.
Organizing Committee<br />
Victorian Association <strong>of</strong> Chinese PhD Students <strong>and</strong> Scholars 2009<br />
Mr. Erjiang (Frank) Fu, RMIT University; President <strong>of</strong> VACPS 2009<br />
Mr. Duo (Winston) Wu, Monash University; Vice President <strong>of</strong> VACPS 2009<br />
Mr. Peng (Peter) Hao, The University <strong>of</strong> Melbourne; Secretary General <strong>of</strong> VACPS 2009<br />
Mr. Xiao Liu, Swinburne University <strong>of</strong> Technology; Head <strong>of</strong> Organization <strong>and</strong> Membership<br />
Team, VACPS 2009<br />
Mr. Haidong (Rick) Zheng, RMIT University; Head <strong>of</strong> Media Team, VACPS 2009<br />
Dr. Gui (John) Qin, RMIT University; Head <strong>of</strong> Business Team, VACPS 2009<br />
Ms. Ping Yu, RMIT University; Secretary <strong>and</strong> Treasurer Team, VACPS 2009<br />
Ms. Xiaohuan (Iris) Yan, Monash University; Secretary <strong>and</strong> Treasurer Team, VACPS 2009<br />
Mr. Wei Tao, Swinburne University <strong>of</strong> Technology; Organization <strong>and</strong> Membership Team,<br />
VACPS 2009<br />
Mr. Yu Zong, Victoria University; Organization <strong>and</strong> Membership Team, VACPS 2009<br />
Mr. Jian Zhong, RMIT University; Media Team, VACPS 2009<br />
Ms. Yini Wang, Deakin University; Media Team, VACPS 2009<br />
Ms. Nan Fu, Monash University; Academy Team, VACPS 2009<br />
Ms. Ying Lu, Monash University; Business Team, VACPS 2009<br />
i
Editorial Board<br />
Editor-in-Chief:<br />
Nan Fu<br />
Academy Team, VACPS 2009<br />
Symposium Chair, Bioscience <strong>and</strong> Biotechnology, The 3 rd VACPS <strong>Workshop</strong> 2009<br />
Editorial Board:<br />
Erjiang Fu<br />
President, VACPS 2009<br />
Duo Wu<br />
Vice President, VACPS 2009<br />
Symposium Co-Chair, Bioscience <strong>and</strong> Biotechnology, The 3 rd VACPS <strong>Workshop</strong> 2009<br />
Peng Hao<br />
Secretary General, VACPS 2009<br />
Liang Chen<br />
Symposium Chair, Electrical <strong>and</strong> Electronic Engineering, The 3 rd VACPS <strong>Workshop</strong> 2009<br />
Xiaohuan Yan<br />
Symposium Co-Chair, Electrical <strong>and</strong> Electronic Engineering, The 3 rd VACPS <strong>Workshop</strong> 2009<br />
Xiaohui Zhao<br />
Symposium Chair, <strong>Information</strong> <strong>and</strong> Communication Technologies, The 3 rd VACPS <strong>Workshop</strong><br />
2009<br />
Xiao Liu<br />
Symposium Co-Chair, <strong>Information</strong> <strong>and</strong> Communication Technologies, The 3 rd VACPS<br />
<strong>Workshop</strong> 2009<br />
Haidong Zheng<br />
Symposium Chair, Material Science, The 3 rd VACPS <strong>Workshop</strong> 2009<br />
Wei Tao<br />
Symposium Co-Chair, Material Science, The 3 rd VACPS <strong>Workshop</strong> 2009<br />
Ying Lu<br />
Symposium Chair, Business <strong>and</strong> Management, The 3 rd VACPS <strong>Workshop</strong> 2009<br />
Gui Qin<br />
Symposium Co-Chair, Business <strong>and</strong> Management, The 3 rd VACPS <strong>Workshop</strong> 2009<br />
ii
Preface<br />
On the occasion <strong>of</strong> the publication <strong>of</strong> the Proceeding <strong>of</strong> the 3 rd Victorian Chinese PhD<br />
students <strong>and</strong> Young Scholars Research <strong>Workshop</strong>, I would like to extend my congratulations<br />
on the successful organization <strong>of</strong> the workshop. I would also express my congratulations on<br />
the research achievements presented on the workshop. Personally, I was very impressed by<br />
the quality <strong>and</strong> scope <strong>of</strong> the presentations on the workshop.<br />
As we are living the <strong>Information</strong> Age, communications <strong>and</strong> networking are part <strong>of</strong><br />
important aspects <strong>of</strong> the pr<strong>of</strong>essional career development for each <strong>of</strong> PhD students <strong>and</strong> young<br />
scholars no matter where you are. In that sense, Victorian Association <strong>of</strong> Chinese PhD<br />
Students <strong>and</strong> Young Scholars (VACPS) has played an important role in Australia, particular<br />
in Victoria. I sincerely wish that VACPSYS is more successful in creating a bridge not only<br />
among Chinese students but also between Australia <strong>and</strong> China in the exchange <strong>of</strong> science,<br />
technology <strong>and</strong> culture in future.<br />
Pr<strong>of</strong>essor Min Gu<br />
Fellow, the Australian Academy <strong>of</strong> Science<br />
Fellow, the Australian Academy <strong>of</strong> Technological Sciences <strong>and</strong> Engineering<br />
Pro Vice-Chancellor, Swinburne University <strong>of</strong> Technology (International Research<br />
Collaboration)<br />
It is our great honour to support the 3rd Victorian Chinese PhD students <strong>and</strong> Young<br />
Scholars Research <strong>Workshop</strong>. It was a great success for this annual event held at Swinburne<br />
University <strong>of</strong> Technology this year. Organising such a workshop requires a significant<br />
dedication, I would congratulate all who were involved.<br />
The series <strong>of</strong> the Victorian Chinese PhD student s <strong>and</strong> Young Scholars Research<br />
<strong>Workshop</strong>s provide an excellent platform for presenting <strong>and</strong> sharing ongoing research work<br />
among students across a variety <strong>of</strong> disciplines. It also plays an important role for networking.<br />
I wish <strong>and</strong> have no doubt that this event will continue to be fruitful.<br />
Pr<strong>of</strong>essor Yun Yang<br />
Associate Dean (Research), <strong>Faculty</strong> <strong>of</strong> <strong>Information</strong> & Communication Technologies <strong>of</strong><br />
Swinburne University <strong>of</strong> Technology<br />
iii
Organizing Committee<br />
Editorial Board<br />
Preface<br />
Table <strong>of</strong> Contents<br />
Table <strong>of</strong> Contents<br />
Scientific Advisory Board………………………………………………………….…. 1<br />
Acknowledgements……………………………………………………………………. 2<br />
List <strong>of</strong> Awardees………………………………………………………………………... 3<br />
Oral Presentation Sessions<br />
Invited Presentations…………………………………………………………………... 4<br />
Career planning <strong>and</strong> Job Interview in Australian ICT Companies………………. 5<br />
Novel medical therapies to treat ectopic pregnancies……………………………. 6<br />
The role <strong>of</strong> mitochondria in atrial fibrillation with Rheumatic heart valvular<br />
disease………………………………………………………………………………… 7<br />
The characteristics <strong>of</strong> the InGaAs/GaAs quantum wells <strong>and</strong> its application in<br />
VECSEL lasers……………………………………………………………………….. 8<br />
Electrical <strong>and</strong> Electronic Engineering………………………………………………. 9<br />
Track-Before-Detect Procedures for Low PRF Surveillance Rader…………….. 10<br />
Fast realization <strong>of</strong> Automatic optical <strong>and</strong> infrared image registration…………... 27<br />
Impulsive Interference Detection Method Based on Morlet Wavelet <strong>and</strong><br />
Maximum Likelihood Estimation……………………………………………………. 33<br />
Demonstration <strong>and</strong> Performance Analysis <strong>of</strong> an Uplink based on Digitized RFover-Fiber<br />
Signal Transport…………………………………………………………. 34<br />
Fiber Nonlinearity Compensation for CO-OFDM Systems with Periodic<br />
Dispersion Maps……………………………………………………………………… 35<br />
Bioscience <strong>and</strong> Biotechnology………………………………………………………. 36<br />
Effect <strong>of</strong> mini-tyrosl-tRNA synthetase / mini-tryptophanyl-tRNA synthetase on<br />
ischemic angiogenesis in rats with acute myocardial infarction………………… 37<br />
Distribution <strong>of</strong> three forms alpha-melanocyte stimulating hormone in the brain<br />
<strong>of</strong> male <strong>and</strong> female sheep by using HPLC <strong>and</strong> Mass Spectrometry Analysis… 49<br />
Investigation <strong>of</strong> the mechanism <strong>of</strong> the cardioprotective effect <strong>of</strong> flavonols…….. 61<br />
Efficient Ethanol production in glucose/xylose co-fermentations by a novel coculture<br />
scheme <strong>of</strong> Zymomonas mobilis <strong>and</strong> Pichia stipitis………………………. 62<br />
Dual mode roll-up effect in a multicomponent near adiabatic adsorption<br />
process………………………………………………………………………………... 63<br />
Effect <strong>of</strong> glucose <strong>and</strong> sodium chloride on the stability <strong>of</strong> aged glucose oxidase<br />
from Aspergillus niger………………………………………………………………... 64<br />
i<br />
ii<br />
iii<br />
iv<br />
iv
Paper as a low-cost base material for diagnostic <strong>and</strong> environmental sensing<br />
applications…………………………………………………………………………… 65<br />
<strong>Information</strong> <strong>and</strong> Communication Technologies…………………………………... 66<br />
Data Management in Cloud Scientific Workflow Systems……………………….. 67<br />
HC_AB: A New Heuristic Clustering Algorithm based on Approximate<br />
Backbone……………………………………………………………………………… 73<br />
Web Page Prediction Based on Conditional R<strong>and</strong>om Fields……………………. 74<br />
Provenancing Qualifications Using MEASUR within Higher Education<br />
Institutions: An Australian Case…………………………………………………….. 75<br />
A Multi-modal Gesture Recognition System in a Human-Robot Interaction<br />
Scenario……………………………………………………………………………….. 76<br />
Distributed Agent based Interoperable Virtual EMR System for Healthcare<br />
System Integration…………………………………………………………………… 77<br />
Searching for Fair Joint Gains in Agent-based Negotiation……………………... 78<br />
Material Science………………………………………………………………………… 79<br />
Mesoporous Silica-Templated Assembly <strong>of</strong> Luminescent Polyester Particles… 80<br />
Grain refinement <strong>of</strong> pure magnesium by back pressure equal channel angular<br />
pressing at room temperature………………………………………………………. 91<br />
Dielectrophoretic Separation <strong>of</strong> Carbon Nanotubes <strong>and</strong> Polystyrene<br />
Microparticles…………………………………………………………………………. 96<br />
Use <strong>of</strong> a Radially Polarized Beam for Ultra-low Energy Threshold for Cancer<br />
Photothermal Therapy with Gold Nanorods……………………………………….. 112<br />
Electrospinning Thermoplastic Polyurethane-contained Collagen Nan<strong>of</strong>ibers<br />
for Tissue Engineering Applications………………………………………………... 116<br />
Uniform Coating <strong>of</strong> WO x on TiO 2 Nanotubes for Enhanced Electrochromic<br />
Performance………………………………………………………………………….. 129<br />
Nanoparticle Trapping by Surface Plasmon Resonance Assisted Thermal<br />
Forces…………………………………………………………………………………. 130<br />
Business <strong>and</strong> Management…………………………………………………………… 131<br />
Union as a Social Regulator <strong>of</strong> Market Risk? --Empirical Evidence from<br />
Pr<strong>of</strong>essional Union Leader Program in B City…………………………………….. 132<br />
Developing Dynamic Capability through Partnership: The Role <strong>of</strong> Capabilities. 141<br />
Early Childhood Education Matters: Functions, Situations, <strong>and</strong> Perceptions…. 148<br />
Five Dimensions <strong>of</strong> Entrepreneurship: A Study <strong>of</strong> First <strong>and</strong> Second<br />
Generations Chinese Entrepreneurs in Melbourne………………………………. 153<br />
An Exploration <strong>of</strong> Country <strong>of</strong> Origin Effect on Union Policies in Chinese<br />
Multinational Enterprises…………………………………………………………….. 154<br />
Poster Session………………………………………………………………………….. 155<br />
v
Scientific Advisory Board<br />
(listed alphabetically by last name)<br />
Dr. Peggy Chan, Department <strong>of</strong> Chemical Engineering, Monash University<br />
Dr. Jinjun Chen, Centre for Complex S<strong>of</strong>tware Systems <strong>and</strong> Services, Swinburne University<br />
<strong>of</strong> Technology<br />
Pr<strong>of</strong>. Xiao Dong Chen, Department <strong>of</strong> Chemical Engineering, Monash Univeristy<br />
Dr. Zhaolin Chen, Department <strong>of</strong> Electrical <strong>and</strong> Computer Systems Engineering, Monash<br />
University<br />
Dr. Shanfei Feng, Department <strong>of</strong> Marketing, Monash University<br />
Dr. Wenyu Guo, Centre for Power Transformer Monitoring, Diagnostics <strong>and</strong> Life<br />
Management, Monash University<br />
Pr<strong>of</strong>. Yuemei Guo, <strong>Faculty</strong> <strong>of</strong> Business <strong>and</strong> Economics, Monash University<br />
Dr. Baohua Jia, Centre for Micro-Photonics, Swinburne University <strong>of</strong> Technology<br />
Dr. Cha<strong>of</strong>eng Li, Department <strong>of</strong> Electrical Engineering, The University <strong>of</strong> Melbourne<br />
Dr. Xiangping Li, Centre for Micro-Photonics, Swinburne University <strong>of</strong> Technology<br />
Mr. Ke Liu, Department <strong>of</strong> Electrical <strong>and</strong> Computer Systems Engineering, Monash<br />
University<br />
Dr. Hongming Ma, <strong>Faculty</strong> <strong>of</strong> Education, Monash University<br />
Pr<strong>of</strong>. Paul Peiris, School <strong>of</strong> Natural Sciences, University <strong>of</strong> Western Sydney<br />
Dr. Ramanie Samaratunge, Department <strong>of</strong> Management, Monash University<br />
Dr. Dongxia Xu, Department <strong>of</strong> Electrical Engineering, The University <strong>of</strong> Melbourne<br />
Dr. Gu<strong>and</strong>ong Xu, Centre for Applied Informatics, Victoria University<br />
Mr. Qi Yang, Department <strong>of</strong> Electrical Engineering, The University <strong>of</strong> Melbourne<br />
Pr<strong>of</strong>. Yun Yang, Centre for Complex S<strong>of</strong>tware Systems <strong>and</strong> Services, Swinburne University<br />
<strong>of</strong> Technology<br />
Dr. Abu Zadek, School <strong>of</strong> Electrical <strong>and</strong> Computer Engineering, RMIT University<br />
Ms. Yuan Zhou, Department <strong>of</strong> Electrical Engineering, The University <strong>of</strong> Melbourne<br />
1
Acknowledgements<br />
We gratefully acknowledge the support from the following organizations:<br />
中 国 驻 墨 尔 本 总 领 事 馆 教 育 组<br />
Education Office, The Consulate General <strong>of</strong> The P. R. China in Melbourne<br />
斯 文 本 科 技 大 学 中 国 学 生 学 者 联 合 会<br />
Chinese Students <strong>and</strong> Scholars Association, Swinburne University <strong>of</strong> Technology<br />
斯 文 本 科 技 大 学 信 息 与 通 信 学 院 微 光 电 子 研 究 中 心<br />
<strong>Faculty</strong> <strong>of</strong> ICT & Centre <strong>of</strong> Micro-Photonics, Swinburne University <strong>of</strong> Technology<br />
We gratefully appreciate the following sponsors:<br />
2
List <strong>of</strong> Awardees<br />
Best Presentations<br />
Bioscience <strong>and</strong> Biotechnology<br />
Dr. Rui Zeng, “Effect <strong>of</strong> Mini-tyrosl-tRNA Synthetase / Mini-tryptophanyltRNA<br />
Synthetase on Ischemic Angiogenesis in Rats with Acute Myocardial Infarction”, Sichuan<br />
University & Monash University<br />
Business <strong>and</strong> Management<br />
Henri Lee, “Five Dimensions <strong>of</strong> Entrepreneurship: A Study <strong>of</strong> First <strong>and</strong> Second<br />
Generations Chinese Entrepreneurs in Melbourne”, Swinburne University <strong>of</strong> Technology<br />
Electrical <strong>and</strong> Electronic Engineering<br />
Xiaobo Deng, “Track-Before-Detect Procedures for Low PRF Surveillance Rader”,<br />
University <strong>of</strong> Electronic Science <strong>and</strong> Technology <strong>of</strong> China & The University <strong>of</strong><br />
Melbourne<br />
<strong>Information</strong> <strong>and</strong> Communication Technologies<br />
Minyi Li, “Searching for Fair Joint Gains in Agent-based Negotiation”, Swinburne<br />
University <strong>of</strong> Technology<br />
Material Science<br />
Chen Zheng, “Dielectrophoretic Separation <strong>of</strong> Carbon Nanotubes <strong>and</strong> Polystyrene<br />
Microparticles”, RMIT University<br />
Best Posters<br />
Wei Tao, Bao Hongchun, Dru Morrish <strong>and</strong> Min Gu, “Supercontinuum generation for fiberoptic<br />
nonlinear microscopy”, Swinburne University <strong>of</strong> Technology<br />
Gang Li, Ranjeet Singh, Dan Li, Chunxia Zhao, Liying Liu, <strong>and</strong> Paul A. Webley, “Synthesis<br />
<strong>of</strong> biomorphic zeolite honeycomb monoliths with 16000 cells per square inch”, Monash<br />
University<br />
3
Career planning <strong>and</strong> Job Interview in Australian ICT Companies<br />
* Invited speaker<br />
1. Telstra<br />
Victor Li 1*<br />
In this talk, Victor Li will share his experiences on career planning <strong>and</strong> some practical<br />
tips on job interview <strong>and</strong> writing resumes.<br />
BIO: Victor Li got his bachelor <strong>and</strong> master degree in electronic engineering in Beijing<br />
Institute <strong>of</strong> Technology, <strong>and</strong> PhD in Computer <strong>and</strong> S<strong>of</strong>tware Engineering in the University <strong>of</strong><br />
Melbourne. He is currently the Group Manager for Enterprise Solutions in Telstra Network<br />
<strong>and</strong> Technology. He is the Chief Architect in Telstra network operations, a well known expert<br />
in streaming media <strong>and</strong> SOA. He has been working in ICT area for over 15 years. He used to<br />
work in Institute <strong>of</strong> Computing Technology, Chinese Academy <strong>of</strong> Science, Motorola<br />
S<strong>of</strong>tware Lab Centre <strong>and</strong> Falcom in Germany. The annual budgets for his projects are over<br />
$100M in the past three years. The company founded by him in China was successfully<br />
floated on New York Stock Exchange in 2007.<br />
5
Novel medical therapies to treat ectopic pregnancies<br />
UW Nilsson 1 a , YE Gao 1,4 a * , TG Johns 2 , E Dimitriadis 3 , E Menkhorst 3 , Joanne Mockler 1 ,<br />
BRG Williams 2 , S Tong 1<br />
* Invited speaker<br />
a These authors contributed equally to this work.<br />
1. Centre for Women’s Health Research, Monash Institute <strong>of</strong> Medical Research, Victoria<br />
2. Centre for Cancer Research, Monash Institute <strong>of</strong> Medical Research, Victoria<br />
3. Prince Henry’s Institute, Victoria, Australia<br />
4. Department <strong>of</strong> Obstetrics <strong>and</strong> Gynecology, the Second Affiliated Hospital, School <strong>of</strong><br />
Medicine, Xi'an Jiaotong University, Xi’an, China<br />
Introduction: Ectopic pregnancies are those implanting outside the uterus <strong>and</strong> they can<br />
be life-threatening emergencies. Most are treated surgically. Methotrexate (MTX) is<br />
occasionally used, buts has limited efficacy. Hence, it is only used for small ectopic<br />
pregnancies. Epidermal growth factor receptor (EGFR) inhibitors are used clinically to treat<br />
cancers. They come in tablet form <strong>and</strong> have few side-effects. Given that the placenta is<br />
dependent on EGFR signalling, EGFR inhibition may be a novel therapeutic approach to cure<br />
un-ruptured ectopic pregnancies <strong>of</strong> any size.<br />
We set out to develop a novel medication based treatment to cure ectopic pregnancy by<br />
combining MTX <strong>and</strong> EGFR tyrosine kinase inhibitor.<br />
Methods/Results: In vitro testing: Firstly, we confirmed the expression <strong>of</strong> EGFR in<br />
JEG-3 cells <strong>and</strong> human 1st trimester placental tissue by immunostaining. Next, we added<br />
various treatment to placenta-derived cells, BeWo (± syncytialisation) cells, JEG-3 cells <strong>and</strong><br />
primary 1st trimester extravillous cytotrophoblast cells, <strong>and</strong> assayed viability 72 hours later<br />
using the CellTiter-Blue® Cell Viability Assay (Promega). We also dynamic monitored MTX<br />
interaction with JEG-3 cells using xCELLigence real-time cell analyzing system. EGFR<br />
inhibitor alone did not cause trophoblast cell death compared to control. MTX alone<br />
decreased cell viability in a dose-responsive manner. Interestingly, combination <strong>of</strong> MTX <strong>and</strong><br />
EGFR inhibitor are supra-additive in decreasing placental tissue viability in vitro.<br />
In vivo testing: JEG-3 xenografted in SCID mice (n=3-5 per group) was used as an in<br />
vivo model to test effects <strong>of</strong> MTX <strong>and</strong> EGFR inhibitor. Compared with vehicle-treated mice,<br />
MTX or EGFR inhibitor can inhibit JEG-3 xenograft growth, <strong>and</strong> are supra-additive when<br />
combined.<br />
Conclusion: Combination methotrexate (MTX) <strong>and</strong> EGFR inhibition can potently<br />
regress placental tissue. It may be a novel therapeutic approach to medically cure ectopic<br />
pregnancy, potentially replacing surgery with tablets.<br />
BIO: Yan’e Gao is an Associate Pr<strong>of</strong>essor in the Department <strong>of</strong> Obstetrics & Gynecology, 2 nd<br />
Affiliated hospital, Xi’an Jiaotong University School <strong>of</strong> Medicine. She is a specialist in the<br />
maternity clinical practice <strong>and</strong> has worked as an attending physician for more than 15 years.<br />
She got her PhD in Anytomy, Histology & Embryology in Xi’an Jiantong University School<br />
<strong>of</strong> Medicine in 2003, <strong>and</strong> she is now an active researcher in the area <strong>of</strong> ectopic pregnancies<br />
<strong>and</strong> uterine cell in vitro culture.<br />
6
The role <strong>of</strong> mitochondria in atrial fibrillation with Rheumatic heart<br />
valvular disease<br />
Runwei Ma 1,2 *<br />
* Invited speaker<br />
1. Department <strong>of</strong> Cardiovascular Surgery, The First Hospital Affiliated Kunming Medical<br />
College, Yunnan 650032, P.R.China<br />
2. Department <strong>of</strong> Cardiothoracic Surgery, Monash Hospital, Clayton, Vic 3168, Australia<br />
Rheumatic heart disease (RHD) is a common disease in developing country. Most <strong>of</strong> the<br />
disease is concomitant with Atrial Fibrillation (AF). The mechanism <strong>of</strong> AF in RHD is<br />
different with other AF. We know that the function <strong>of</strong> mitochondria is associated with<br />
cardiomyocyte necrosis <strong>and</strong> apoptosis which will remodel the structure <strong>of</strong> atrium, especial<br />
fibrosis. Tissue fibrosis results from an accumulation <strong>of</strong> fibrillar collagen deposits, occurring<br />
most commonly as a reparative process to replace degenerating myocardial parenchyma with<br />
concomitant reactive fibrosis, which causes interstitial expansion. Fibrosis is a hallmark <strong>of</strong><br />
arrhythmogenic structural remodeling. In addition, there is great difference in heart function<br />
<strong>of</strong> patients with same detriment <strong>of</strong> valves, as well as AF. We postulate that there is a<br />
difference in mitochondria’s function which is influenced by gene mutation or difference <strong>of</strong><br />
mitochondria which caused by inheritance or disease. We plan to use DNA chip technique to<br />
study the difference <strong>of</strong> mitochondrial gene in RHD patients with <strong>and</strong> without AF.<br />
BIO: Runwei Ma is the head <strong>of</strong> the Department <strong>of</strong> Cardiovascular Surgery in the First<br />
Affiliated Hospital <strong>of</strong> Kunming Medical College, <strong>and</strong> is responsible for the Cardiovascular<br />
Surgeon. He has successfully applied a number <strong>of</strong> novel techniques in surgical operations. Dr<br />
Runwei got his MD degree in Kunming Medical College in 2007 <strong>and</strong> now he is a visiting<br />
scholar in the Medical Center Monash University. His research focuses on the improvement<br />
<strong>of</strong> cardiovascular conditions using both traditional drug treatment <strong>and</strong> genetic therapy.<br />
7
The characteristics <strong>of</strong> the InGaAs/GaAs quantum wells <strong>and</strong> its application<br />
in VECSEL lasers<br />
* Invited speaker<br />
1. Beijing University <strong>of</strong> Technology, China<br />
Yanrong Song 1*<br />
We introduce a novel kind <strong>of</strong> laser-vertical external cavity surface emitting laser which<br />
combinds the advantages <strong>of</strong> solid state lasers <strong>and</strong> semiconductor lasers. The design <strong>of</strong> the gain chip,<br />
microavity <strong>and</strong> the whole setup are given <strong>and</strong> a comprehensive modeling <strong>of</strong> vertical-external-cavity<br />
surface-emitting laser, reflective spectrum, edge-emitting photoluminescence spectrum, surfaceemitting<br />
photoluminescence spectrum, material gain spectrum <strong>and</strong> the output characteristics are<br />
calculated. The experiments about CW laser, double frequency output, mode locking laser are<br />
demonstrated.<br />
BIO: Pr<strong>of</strong>. Yanrong Song works in College <strong>of</strong> Applied Sciences, Beijing University <strong>of</strong><br />
Technology, China, <strong>and</strong> now works as a visiting scholar in Centre for Micro-Photonics,<br />
Swinburne University <strong>of</strong> Technology. She got her MS. Degree <strong>and</strong> Ph.D degree in Optics in<br />
Shanxi University. She received the Bachelor degree in Semiconductor Physics <strong>and</strong> devices<br />
from Tianjin University.<br />
Her insterested research areas include in ultrafast processing <strong>and</strong> mode-locked lasers,<br />
Biology sensor with nano-material, vertical external cavity <strong>of</strong> surface emitting semiconductor<br />
lasers, ultra-fast Ti: Sipphire lasers <strong>and</strong> amplifiers.<br />
8
Track-Before-Detect Procedures for Low PRF Surveillance Rader<br />
Xiaobo Deng 1 * , Yiming Pi 1 , Mark Morel<strong>and</strong>e 2 , Bill Moran 2<br />
* Presenter<br />
1. University <strong>of</strong> Electronic Science <strong>and</strong> Technology <strong>of</strong> China, Chengdu, 610054, China<br />
2. The University <strong>of</strong> Melbourne, Parkville, VIC 3010, Australia<br />
In this paper we present a dynamic programming (DP) based tracking-beforedetect (TBD)<br />
procedures with reference to a low pulse repetition frequency (PRF) surveillance radar<br />
framework. In order to avoid nonlinear transformation <strong>and</strong> meanwhile preserve the<br />
ambiguous Doppler information which is eliminated in most <strong>of</strong> the literature, we model the<br />
target dynamics in the range-Doppler domain. The target state evolutions in the ambiguous<br />
range-Doppler domain are considered as a hybrid system with the ambiguous number deemed<br />
as mode variable. Ambiguity number (or mode) transitions are governed by guard conditions<br />
that are state-dependent. We present a DP based method for joint maximum a posteriori<br />
(MAP) estimation <strong>of</strong> target’s trajectory in the ambiguous range-Doppler maps <strong>and</strong> the<br />
corresponding ambiguity sequence, both assumptions <strong>of</strong> known <strong>and</strong> unknown nuisance<br />
parameters are considered. The detection <strong>and</strong> tracking performance <strong>of</strong> the proposed procedure<br />
is investigated under several system settings. The effect <strong>of</strong> the prior uncertainty <strong>of</strong> the<br />
nuisance parameters (target power <strong>and</strong> noise variance) on the performance is also studied.<br />
1. Introduction<br />
Early detection <strong>and</strong> trajectory estimation <strong>of</strong> moving targets from remote surveillance radars is a<br />
very challenging problem. The detection <strong>and</strong> tracking strategy should be power efficient to deal with<br />
the targets in low signal-to-noise ratio (SNR), while the complexity should not hamper the process <strong>of</strong><br />
early decisions. The stringent detection specification is well met by track-before-detect (TBD)<br />
approaches. TBD procedures allow simultaneous detection <strong>and</strong> tracking using unthresholded data <strong>and</strong>,<br />
shows superior detection performance over the conventional methods.<br />
Early studies on TBD methods mainly focused on its application in optical <strong>and</strong> infrared sensors as<br />
in [1-5]. Extensions <strong>of</strong> TBD methods to pulse-Doppler radar system have been considered in [6-8].<br />
Unlike in the infrared <strong>and</strong> optical scenarios, TBD approaches operate on a sequence <strong>of</strong> twodimensional<br />
grey-scale images, in radar scenario, TBD approaches might apply to a sequence <strong>of</strong> threedimensional<br />
data sets, defined by azimuth, range <strong>and</strong> Doppler (radial velocity), making the complexity<br />
requirement much more stringent. Another difference is that, in radar applications, the inherent<br />
properties <strong>of</strong> the azimuth- range-Doppler data space can be fully exploited to improve detection <strong>and</strong><br />
tracking performance. For example, in [6], a method for using TBD procedure in pulse-Doppler radars<br />
is described, wherein the author exploits the Doppler information to define which range cell a target<br />
would have been in the previous frame, <strong>and</strong> thus greatly simplify the search process involved in the<br />
Viterbi algorithm. However, there is a drawback with this method if the observed Doppler is<br />
ambiguous, which is the case when the radar under our consideration works with a low pulse<br />
repetition frequency (PRF). In [7], the authors maximize the data set over the observed apparent<br />
Doppler shifts, <strong>and</strong> then employ a Viterbi-like tracking algorithm in the range-azimuth domain. In [8],<br />
the authors have also suggested taking advantage from the structure <strong>of</strong> the range-Doppler domain to<br />
efficiently perform Viterbi-like tracking <strong>of</strong> multi-target. However, since the observed Doppler is<br />
ambiguous, it’s also eliminated as in [7].<br />
In order to exploit the Doppler information more effectively, we treat it in a more subtle way. The<br />
observed Doppler is a modulo-PRF version <strong>of</strong> the true Doppler, <strong>and</strong> thus we represent the Doppler<br />
information using observed apparent Doppler <strong>and</strong> the corresponding ambiguity number. The<br />
ambiguity number is deemed as a mode variable. In order to avoid the nonlinear transformation<br />
between the target’s state <strong>and</strong> measurement (defined in the ambiguous range-Doppler domain), the<br />
target states are defined in the ambiguous range-Doppler domain. The evolutions <strong>of</strong> target state in<br />
successive range-Doppler maps are analyzed <strong>and</strong> modeled as a hybrid (or multiple-model) system [9,<br />
10
10]: The transitions <strong>of</strong> target locations (defined by range <strong>and</strong> apparent Doppler) are governed by the<br />
ambiguity number, i.e., mode variable, while the ambiguity number transitions are state-dependent.<br />
Exploiting the admissible transitions <strong>of</strong> target locations <strong>and</strong> ambiguity number, we have proposed a<br />
dynamic programming (DP) based method for joint maximum a posteriori (MAP) estimation <strong>of</strong><br />
target’s trajectory in the ambiguous range-Doppler maps <strong>and</strong> the corresponding ambiguity number<br />
sequence. The proposed method is investigated under both known <strong>and</strong> unknown nuisance parameters<br />
(target power <strong>and</strong> noise variance). The detection <strong>and</strong> tracking performances <strong>of</strong> the proposed algorithm<br />
are studied with respect to different system settings. Moreover, the effect <strong>of</strong> the prior uncertainty <strong>of</strong><br />
the nuisance parameters on the detection <strong>and</strong> tracking performance is also investigated.<br />
2. Signal processing model for low PRF surveillance radars<br />
The physical scenario considered here is shown in Fig.1. A low PRF surveillance radar is used for<br />
monitoring a given area with an electronically scanned antenna. The surveillance area is divided into<br />
small angular regions, visited by the antenna at a constant revisit intervalT R<br />
. For many long range<br />
targets, passing from one elevation beam to another during several scans happens infrequently due to<br />
the geometric constraints; however, crossing azimuth beams happens much more frequently.<br />
Therefore, each angular region is composed <strong>of</strong> a single elevation <strong>and</strong> N a azimuthal beam-pointing<br />
positions. The azimuthal beamwidth is Δ<br />
a<br />
. The pulse repetition time is T p<br />
<strong>and</strong> the number <strong>of</strong> train<br />
pulses coherently processed is Nd<br />
. At each visit, say, the kth<br />
, the radar provides a multi-dimensional<br />
data map (with four dimensions <strong>of</strong> range, azimuth, elevation <strong>and</strong> Doppler), each pixel is the reflected<br />
power out <strong>of</strong> a signal processing unit, e.g., discrete Fourier transform(DFT), <strong>of</strong> a surveillance radar.<br />
Fig.1. Radar surveillance zone illustration<br />
The conventional signal processing method for surveillance radar first thresholds the data maps to<br />
extract target plots (range, azimuth, elevation <strong>and</strong> Doppler locations), <strong>and</strong> then performs tracking<br />
based on these plots. For targets <strong>of</strong> high SNR, this method works well, however, for targets <strong>of</strong> relative<br />
low SNR, such as stealthy military aircraft <strong>and</strong> cruise missiles, thresholding throws away potentially<br />
useful information, causing missing detections. Besides, since we are concerned with low PRF radar,<br />
the Doppler plots are ambiguous; it’s hard to resolve the Doppler ambiguity under low SNR in the<br />
framework <strong>of</strong> conventional methods. The TBD algorithms use several successive unthresholded data<br />
maps to jointly determine the presence <strong>of</strong> a target <strong>and</strong> its trajectory (if present). TBD methods exploit<br />
all the useful information among several scans <strong>and</strong> are especially preferred when encountering low<br />
SNR targets. Besides, Doppler ambiguity is also implicitly resolved in the framework <strong>of</strong> TBD methods.<br />
To simplify the exposition, we assume that the azimuth <strong>and</strong> elevation information are obtained<br />
through antenna processing. The only relevant coordinates are range <strong>and</strong> Doppler. In an actual<br />
implementation, one can choose to process azimuth-elevation <strong>and</strong> range-Doppler maps separately.<br />
11
3. Dynamic model for approaching targets<br />
As concerns the choice <strong>of</strong> target dynamic model, it’s tied to the adapted tracking coordinate system.<br />
The most natural coordinate is the Cartesian coordinate. In a Cartesian system, the target state is<br />
described by ( x, xyy , , ) T x y<br />
x& y&<br />
is the velocity vector.<br />
Φ = & & , where ( , ) k<br />
k<br />
k<br />
is the target’s position; ( , ) k<br />
2 2<br />
Velocity vector can be described by speed <strong>and</strong> heading pair (, v α , where v = x& + y& ,<br />
<strong>and</strong> arctan ( x / y)<br />
α = & & .<br />
However, in the situation under our consideration, the measurements are the reflected power on the<br />
ambiguous range-Doppler (i.e., range-apparent Doppler) grid. Thus, the choice <strong>of</strong> the Cartesian<br />
coordinate requires a nonlinear transformation which results in increased complexity. It’s <strong>of</strong> interest to<br />
directly model the target dynamics in the ambiguous range-Doppler domain. As illustrated in Fig. 2,<br />
an approaching target travels at a velocity v <strong>and</strong> headingα . Heading is defined as the angle from the<br />
velocity direction to the north direction. β denotes the difference between the inverse bearing <strong>and</strong><br />
target heading,<br />
β = α − ( θ + π)<br />
(1)<br />
0<br />
) k<br />
Fig.2. Cartesian coordinate with radar located at the original <strong>and</strong> approaching target<br />
The range trajectory for the approaching target is<br />
2 2 2<br />
rt () = r + vt −2rvtcosβ<br />
0 0<br />
2<br />
dr() t 1 dr () t 2<br />
r0 |<br />
t=<br />
0<br />
t |<br />
2 t= 0<br />
t (2)<br />
≈ + ⋅ + ⋅<br />
dt 2 dt<br />
2<br />
1( vsin β ) 2<br />
= r0<br />
− ( vcos β ) t+<br />
t<br />
2 r0<br />
2<br />
Let λ = c/ Ft<br />
be the radar wavelength; fdc<br />
= 2vcos β / λ , the Doppler centroid; fdr<br />
=− 2( vsin β ) /( λr0<br />
) ,<br />
the Doppler rate. (2) can be rewritten as<br />
λ λ 2<br />
rt () = r0<br />
− fdct− fdrt<br />
(3)<br />
2 4<br />
The Doppler trajectory can be written as<br />
2 dr( t)<br />
fd()<br />
t =− = f dc<br />
+ f dr<br />
t (4)<br />
λ dt<br />
The discrete-time target dynamic model can be obtained by sampling the above continuous-time<br />
model. Sampling the above continuous-time models at the interval <strong>of</strong>T R<br />
, we get<br />
fdk , + 1<br />
= fd()| t<br />
t= ( k+<br />
1) T<br />
= f ( 1)<br />
R dc+ fdr k+<br />
TR<br />
(5)<br />
= f + f T<br />
dk , dr R<br />
12
λ<br />
λ<br />
2<br />
rk+ 1<br />
= r()| t<br />
t= ( k+<br />
1) T<br />
= r0<br />
− f ( 1) ( )<br />
R<br />
dc<br />
k + TR − fdc<br />
kTR<br />
+ TR<br />
2 4<br />
λ λ 2<br />
= rk − fd,<br />
kTR − fdrTR<br />
2 4<br />
(6)<br />
Expressing the true Doppler fdk<br />
,<br />
in terms <strong>of</strong> observed apparent Doppler f<br />
da,<br />
k<br />
( 0≤ f da , k<br />
< PRF<br />
), <strong>and</strong><br />
ambiguity number m k<br />
,<br />
fdk ,<br />
= fdak ,<br />
+ mkPRF<br />
(7)<br />
The apparent Doppler fda,<br />
k<br />
is a modulo-PRF version <strong>of</strong> the true Doppler,<br />
fda, k<br />
= mod ( fd , k<br />
, PRF)<br />
(8)<br />
where mod ( A,<br />
B)<br />
denotes the modulo- B version <strong>of</strong> A . Defining the target state vector<br />
as s ( ,<br />
,<br />
) T<br />
k<br />
= x<br />
k<br />
fdr k<br />
, where x ( ,<br />
,<br />
) T<br />
k<br />
= rk fda k<br />
is the target location in ambiguous range-Doppler maps, the<br />
evolution <strong>of</strong> target state can be written as a hybrid system<br />
s = k 1<br />
f ( s k, mk) + h( s + k, mk)<br />
w (9)<br />
k<br />
where<br />
⎛ λ<br />
λ 2 ⎞<br />
⎜rk − ( fda, k<br />
+ mkPRF)<br />
TR − fdr,<br />
kTR<br />
2 4<br />
⎟<br />
⎜<br />
⎟<br />
f( sk, mk) = ⎜ mod ( fda, k<br />
+ fdr,<br />
kTR,PRF) ⎟<br />
(10)<br />
⎜<br />
⎟<br />
⎜<br />
fdr,<br />
k<br />
⎟<br />
⎝<br />
⎠<br />
⎛λT<br />
2 /4⎞<br />
⎜ ⎟<br />
h( sk, mk)<br />
= ⎜⎜ TR<br />
⎟<br />
(11)<br />
1 ⎟<br />
⎝ ⎠<br />
wk<br />
is the Doppler rate increment noise, it’s assumed to be zero mean white Gaussian noise with<br />
variance Q = var( w k<br />
) . The ambiguity number, mk<br />
∈{1, 2,..., M } is considered as the mode variable. In<br />
the following, we use the ambiguity number <strong>and</strong> mode interchangeably.<br />
3.2 Transition probabilities <strong>of</strong> ambiguity number<br />
The target location transitions among successive ambiguous rang-Doppler maps depends on the<br />
ambiguity number, while the ambiguity number transitions occur when the apparent Doppler falls<br />
within a subset, known as guard condition, <strong>of</strong> the state space. For example, as illustrated in Fig. 3, a<br />
target decelerates at a constant rate. When the observed Doppler has reached a certain region, say,<br />
below 5 at k = 3 , the ambiguity number starts to change from m<br />
3<br />
= 7 to m<br />
4<br />
= 6 . This implies that the<br />
evolution <strong>of</strong> the ambiguity number depends on the state, specifically, the apparent Doppler. It makes<br />
sense to define the state-dependent mode transition probabilities [9, 10]<br />
( x ) = pm ( = j| m = i, x ) (12)<br />
π<br />
ij k k + 1<br />
k k<br />
In the following, we consider the mathematic expression <strong>of</strong> the state dependent ambiguity<br />
number transition probability π<br />
ij<br />
( x<br />
k<br />
). To start with, we assume only transitions to adjacent ambiguity<br />
numbers are feasible. This assumption limits the magnitude <strong>of</strong> the Doppler rate <strong>and</strong> its increment noise<br />
variance, achieving more realistic target behavior. Specifically, for a decelerating target shown in<br />
Fig.3, only transition from mode i to mode j = i − 1 is feasible, <strong>and</strong> the transition occurs only when the<br />
observed Doppler has fallen below a certain value. We express the guard conditions as linear<br />
inequalities. Thus, a mode transition from mode i to mode j = i − 1 occurs when the following guard<br />
condition is met.<br />
Cx<br />
ij k<br />
≤ δij<br />
(13)<br />
C<br />
ij<br />
= (0,1) , <strong>and</strong>δ ij<br />
is the guard threshold.<br />
We can assume δ ij<br />
as deterministic, then it’s much easier to determine π<br />
ij<br />
( x<br />
k<br />
)<br />
13
⎧π0<br />
if j = i <strong>and</strong> Cx<br />
ij k<br />
> δij<br />
⎪<br />
π<br />
ij<br />
( xk ) = ⎨1 − π0<br />
if j = i−1 <strong>and</strong> Cijx k<br />
≤δij<br />
(14)<br />
⎪<br />
⎩0 else<br />
whereπ 0<br />
is the probability <strong>of</strong> staying in the same mode, the guard thresholdδ ij<br />
is determined by the<br />
level <strong>of</strong> Doppler rate <strong>and</strong> its incremental noise rate increment.<br />
Due to the uncertainties in target dynamics, δ<br />
ij<br />
is well modeled as a r<strong>and</strong>om variable. For<br />
simplicity, we assume δ ij<br />
is a Gaussian distributed variable with mean δ ij<br />
<strong>and</strong> variance Σ ij<br />
, denoted<br />
byδ N ( δ ; δ , Σ ), then the conditional mode transition probability from mode i to mode j = i − 1 is<br />
ij ij ij ij<br />
given by<br />
π ( x ) = p δ ≥ C x | m = i,<br />
x<br />
( )<br />
ij k ij ij k k k<br />
(15)<br />
Sinceδ ij<br />
is Gaussian distributed, then (15) is given by<br />
( δ )<br />
π ( x ) = 1 −Φ C x − ; Σ (16)<br />
ij k ij k ij ij<br />
where Φ( δ ij<br />
; Σ ij ) denotes a Gaussian cumulative probability function (CDF) with mean δ ij<br />
<strong>and</strong><br />
variance<br />
Σ ij<br />
.The probability that target stays the same ambiguity number is given by<br />
π ( x ) = 1 −π<br />
( x )<br />
(17)<br />
ii k ij k<br />
Fig.3. Illustration <strong>of</strong> state-dependent ambiguity number transition<br />
4. Measurement Statistical Model<br />
Let x()<br />
t be the signal transmitted by the radar. For simplicity, the transmitted signal is a<br />
rectangular pulse train x() t = a()exp( t j2 π Ft t ), where F t<br />
is the carrier frequency <strong>and</strong> at () is the pulse<br />
envelope <strong>of</strong> the form<br />
⎧⎪ 1:0 ≤mod( t,<br />
T p ) ≤τ<br />
(18)<br />
at () = ⎨<br />
⎪⎩ 0:others<br />
whereτ is the pulse width.<br />
The received baseb<strong>and</strong> signal during the kth<br />
visit, k = 1,..., K the can be written as<br />
H : y () t = g () t<br />
0<br />
k<br />
k<br />
Nd<br />
⎛<br />
2r<br />
k ⎞<br />
H1<br />
: yk() t = ∑Aa k ⎜t−nTp−( k−1) TR− exp ( j2π<br />
f t<br />
, ) + g ()<br />
dk k<br />
t<br />
n=<br />
0<br />
c<br />
⎟<br />
⎝<br />
⎠<br />
(19)<br />
14
wherein A k<br />
is the received signal amplitude, <strong>and</strong> it’s assumed to be unchanged during each pulse train.<br />
r <strong>and</strong> f d,<br />
k<br />
are the unknown target range, <strong>and</strong> Doppler at kth visit, respectively. c is the speed <strong>of</strong> light,<br />
k<br />
gk<br />
() t is the complex white Gaussian noise with variance σ 2 g<br />
. After range sampling, the slow-time<br />
discrete observation on a given range cell, say the ith<br />
, i = 1,..., Nr<br />
, can be written as<br />
H<br />
0<br />
:<br />
y ( n) = g ( nT )<br />
ki , ki , p<br />
( j2π<br />
f nT<br />
d,<br />
k p)<br />
H : y ( n) = A exp + g ( nT )<br />
1 ki , k ki , p<br />
(20)<br />
for n= 1,..., N<br />
d<br />
. The dimension <strong>of</strong> range cell is given by Δ<br />
r<br />
= cτ /2.<br />
For extracting moving targets from the undesired background clutter, Doppler processing is<br />
desirable. For each range cell i , i = 1,..., Nr<br />
at the kth<br />
scan, a DFT is applied to the slow-time<br />
observation. We first consider the case when only noise present. The N<br />
d<br />
-point DFT <strong>of</strong> yki<br />
,<br />
( n)<br />
is given<br />
by<br />
N d −1<br />
⎛ 2π<br />
nl ⎞<br />
Yki ,<br />
(| l H0 ) = ∑ gki ,<br />
( nTp)exp⎜−j<br />
⎟<br />
l=<br />
0<br />
⎝ N<br />
, l= 1,..., N (21)<br />
d<br />
d ⎠<br />
Yki<br />
,<br />
( l| H0)<br />
is Gaussian distributed since it’s linear combination <strong>of</strong> the complex Gaussian<br />
2 2<br />
variable gki<br />
,<br />
( nT<br />
p)<br />
. The mean <strong>of</strong> Yki<br />
,<br />
( l| H 0<br />
) is zero <strong>and</strong> the variance isσ<br />
= N d<br />
σ g<br />
.<br />
When a target is present, we ignore leakage effect <strong>and</strong> assume the apparent Doppler lies in<br />
frequency bin , which means<br />
ki ,<br />
( |<br />
1)<br />
l 0<br />
( )<br />
f = mod f , PRF = l PRF/ N<br />
da, k d , k 0 d<br />
The expression forY<br />
ki ,<br />
( l | H1)<br />
is given by<br />
(22)<br />
⎧⎪ NdAk + Yk, i( l| H0)<br />
l = l0<br />
Yki<br />
,<br />
( l| H1)<br />
= ⎨<br />
(23)<br />
⎪⎩ Yki<br />
,<br />
( l| H0)<br />
l ≠ l0<br />
2<br />
Y l H is also a Gaussian r<strong>and</strong>om variable with mean N A <strong>and</strong> covariance σ = . Through DFT,<br />
the Doppler domain is divided into Nd<br />
cells, the dimension <strong>of</strong> which is Δ<br />
d<br />
= PRF/ N d<br />
. The reflected<br />
power on the ambiguous range-Doppler grid is given by<br />
z = | Y ( l)| 2 ,( i,<br />
l)<br />
∈S , k = 1,..., K (24)<br />
k<br />
{ k,<br />
i<br />
}<br />
where S = { 0..., N} × { 0..., N }<br />
r<br />
d<br />
denotes the set <strong>of</strong> all resolution cells in the ambiguous range-Doppler grid.<br />
The total number <strong>of</strong> resolution cells is N= Nr× Nd. The center <strong>of</strong> each cell ( il) , is defined to be<br />
(,)<br />
at ( iΔr, lΔd)<br />
. For each pixel <strong>of</strong> zk<br />
, z il<br />
k<br />
, being the power <strong>of</strong> a complex Gaussian variable, follows either<br />
a exponential distribution (if only noise exists) or a noncentral chi-square distribution with two<br />
degrees <strong>of</strong> freedom (if a targets exists). Assume the pixels <strong>of</strong> z k<br />
are independent, the likelihood<br />
function <strong>of</strong> z k<br />
is given by<br />
(,) il<br />
1 ⎛ z ⎞<br />
k<br />
p( zk<br />
| bk, H0 ) = ∏ exp<br />
2 ⎜−<br />
2 ⎟ (25)<br />
(,) il∈S<br />
σ ⎝ σ ⎠<br />
if there is no target exists or<br />
( i0, l0)<br />
( i0, l0)<br />
1 ⎛ u<br />
2<br />
k<br />
+ z ⎞ ⎛ uz ⎞<br />
k<br />
k k<br />
p( zk | xk, bk)<br />
== exp I<br />
2 ⎜− 2 ⎟× ⎜ ⎟<br />
0<br />
×<br />
2<br />
σ ⎝ σ ⎠ ⎜ σ ⎟<br />
⎝ ⎠<br />
(26)<br />
(,) il<br />
1 ⎛ z ⎞<br />
2 exp<br />
k<br />
∏ ⎜−<br />
2 ⎟<br />
(,) il∈S<br />
,(,) il ≠( i0, l0)<br />
σ ⎝ σ ⎠<br />
if there is a target located in cell ( i0, l0)<br />
<strong>of</strong> the ambiguous range-Doppler, where I()<br />
0<br />
⋅ is the zero-order<br />
2<br />
modified Bessel function; uk = ( NdAk)<br />
is the expected target power after Doppler processing.<br />
2<br />
2<br />
bk<br />
= σ under H0<br />
<strong>and</strong> bk<br />
= ( σ , uk)<br />
under H<br />
1<br />
, denotes the nuisance parameters.<br />
k<br />
The set <strong>of</strong> all data maps up to the k th visit might be denoted by Z = ( z ,..., 1<br />
z<br />
k ) , the likelihood<br />
k<br />
function <strong>of</strong> under H takes the form<br />
Z<br />
1<br />
d<br />
k<br />
2<br />
N d<br />
σ g<br />
15
k<br />
k k k<br />
( ) p(<br />
p Ζ | X , B = ∏ zk | x<br />
k,<br />
b<br />
k)<br />
(27)<br />
k = 1<br />
k<br />
k<br />
given the target trajectory X = ( x1, x2,..., xk<br />
) <strong>and</strong> nuisance parameter B = ( b1<br />
,..., b k<br />
).<br />
5. TBD Procedures<br />
In our situation, given the set <strong>of</strong> unthresholded ambiguous range-Doppler data maps up to<br />
k<br />
the k th visit, Z = ( z ,..., 1<br />
z<br />
k ) , the TBD procedures aim to compute optimal (in MAP sense) target<br />
k<br />
trajectory X = ( x1, x2,..., xk<br />
) in ambiguous range-Doppler map, <strong>and</strong> its ambiguity number<br />
k<br />
k k<br />
sequence Μ = ( m1<br />
,..., m k<br />
) . The close-form solution <strong>of</strong> X , Μ is hardly feasible. Moreover, the<br />
k<br />
existence <strong>of</strong> nuisance parameters, B in the measurements also increases the complexity <strong>of</strong> solving<br />
this problem. An exhaustive search over the ambiguous range-Doppler maps, mode parameter <strong>and</strong><br />
unknown nuisance parameter space results in unbearable computational burden. Luckily, advantage<br />
k<br />
k<br />
can be taken from the Markovian property <strong>of</strong> the target trajectory X <strong>and</strong> mode sequence Μ , the<br />
k<br />
k<br />
search for the MAP estimates <strong>of</strong> X <strong>and</strong> Μ can be solved using DP methodology. In the following<br />
k<br />
k<br />
section, we first derive the joint MAP estimates <strong>of</strong> X <strong>and</strong> Μ assuming the known nuisance<br />
parameters <strong>and</strong> then extend the approach to the case <strong>of</strong> unknown nuisance parameters.<br />
5.1 Known Nuisance Parameters<br />
k<br />
k<br />
When the nuisance parameter is known, the joint MAP estimates <strong>of</strong> X <strong>and</strong> Μ is given by<br />
ˆ k<br />
( , ˆ k k k k<br />
X Μ ) = arg max p( X , Μ | Z ) (28)<br />
k X , Μ<br />
k<br />
k k k<br />
Direct maximization <strong>of</strong> p( X , Μ | Z ) is extremely difficult since a close-form solution may not exist.<br />
Brute-force maximization results in an exhaustive search over the ambiguous range-Doppler maps <strong>and</strong><br />
mode parameters, which represents a huge computation burden. Making use <strong>of</strong> the Markovian<br />
k<br />
k<br />
property <strong>of</strong> X <strong>and</strong> Μ , as shown in Section 3, the search for target trajectory <strong>and</strong> mode sequence that<br />
k k k<br />
maximize the posteriori function p( X , Μ | Z ) can be solved using DP methodology.<br />
The maximization in (28) is split into two parts<br />
ˆ k ˆ k k k k<br />
( X , Μ ) = argmax<br />
⎡<br />
max p( , | )<br />
⎤<br />
, k 1 k 1<br />
k m ⎢<br />
X M Z<br />
x − −<br />
k⎣X , M<br />
⎥<br />
(29)<br />
⎦<br />
The inner maximization is denoted by<br />
( , ) max ( k , k | k<br />
I x m = p X M Z ) (30)<br />
k k k<br />
,<br />
k−1 k−1<br />
X M<br />
Ik( xk, mk)<br />
can be interpreted as the merit <strong>of</strong> the c<strong>and</strong>idate trajectory ending in any state xk<br />
<strong>and</strong> mode mk<br />
at the kth visit. The desired iterative relation is obtained using Bayes’ theorem to<br />
express Ik+ 1( x<br />
k+ 1,<br />
mk+<br />
1) in terms <strong>of</strong> Ik( x<br />
k, mk)<br />
. From Bayes’ theorem it follows that<br />
k+ 1 k+ 1 k+<br />
1<br />
p( X , Μ | Z )<br />
p( zk+ 1| xk+ 1) p( xk+ 1| xk, mk) p( mk+<br />
1| mk, xk)<br />
(31)<br />
k k k<br />
= p( X , Μ | Z )<br />
k<br />
p( zk+<br />
1<br />
| Z )<br />
Using (31), Ik+ 1( x<br />
k+ 1,<br />
mk+<br />
1)<br />
can be written in the form<br />
Ik+ 1( xk+ 1, mk+<br />
1)<br />
k+ 1 k+ 1 k+<br />
1<br />
= max p( X , Μ | Z )<br />
(32)<br />
X<br />
k<br />
, Mk<br />
⎧p( zk+ 1| xk+ 1) p( xk+ 1| xk, mk) p( mk+<br />
1| mk, xk)<br />
⎫<br />
= max ⎨ Ik( xk, mk)<br />
k,<br />
m<br />
k<br />
⎬<br />
x k⎩<br />
p( zk+<br />
1<br />
| Z )<br />
⎭<br />
k<br />
Equation (32) constitutes the desired iterative relation. The function p ( zk<br />
+ 1<br />
| Z ) is used for<br />
normalization purposes only <strong>and</strong> can be dropped without affecting the maximizer. The probability<br />
density p ( z )<br />
k+ 1<br />
| x<br />
k+<br />
1<br />
is obtained from(26). Working with the natural logarithm <strong>of</strong> (32), we transform<br />
the merit function into an additive rather than a multiplicative function.<br />
16
I<br />
( x , m )<br />
k+ 1 k+ 1 k+<br />
1<br />
(,) il<br />
( i0, l<br />
( 0)<br />
i0, l0)<br />
1 ⎛ 1 z ⎞ 2 u<br />
1 1 1<br />
k 1z<br />
k+ uk+ + z<br />
⎛ ⎛ ⎞⎞<br />
k+<br />
+ k+<br />
1<br />
= ln − ln + ln I<br />
2 ∑ ⎜ −<br />
2 2 ⎟−<br />
⎜ ⎜ ⎟⎟<br />
2<br />
0 2<br />
σ (,) il∈S<br />
,(,) il ≠( i0, l0)<br />
⎝ σ σ ⎠ σ ⎜ ⎜ σ ⎟ ⎟<br />
⎝ ⎝ ⎠⎠<br />
+ max ln ( | , ) + ln ( | , ) + ( , )<br />
xk,<br />
mk<br />
{ px x m pm m x I x m }<br />
k+ 1 k k k+<br />
1 k k k k k<br />
where ( i0, l0)<br />
is the target location in ambiguous range-Doppler map when in state x k + 1. Note that adding<br />
the same bias to the merit function <strong>of</strong> all states in any given frame will not affect the maximization<br />
result. Thus, we are justified in dropping the first two terms on the right side <strong>of</strong> (33), which are<br />
independent <strong>of</strong> ( x , m ).<br />
k+ 1 k+<br />
1<br />
I<br />
( x , m )<br />
k+ 1 k+ 1 k+<br />
1<br />
( i0, l<br />
( 0)<br />
i0, l0)<br />
u<br />
2 u<br />
1 1<br />
k 1z<br />
k+ + z<br />
⎧<br />
⎫<br />
k+<br />
⎪<br />
⎛ ⎞<br />
+ k+<br />
1 ⎪<br />
=− +ln I<br />
2 ⎨ ⎜ ⎟<br />
0<br />
2 ⎬<br />
σ ⎪ ⎜ σ ⎟<br />
⎩ ⎝ ⎠⎪⎭<br />
+ max ln ( | , ) + ln ( | , ) + ( , )<br />
xk,<br />
mk<br />
{ p x x m p m m x I x m }<br />
k+ 1 k k k+<br />
1 k k k k k<br />
where we are using the same symbol I for the redefined merit function.<br />
Although the maximization in (34) is said to take place over all locations <strong>and</strong> ambiguity numbers<br />
in the previous visit, in practices, only a small set <strong>of</strong> locations <strong>and</strong> ambiguity numbers which are<br />
physically admissible need to be considered. Fig. 4 shows admissible search regions among successive<br />
frames. The search region in range dimension depends on ambiguity number. For a cell (, i j)<br />
at<br />
k + 1st frame, given the ambiguity number mk<br />
, say, m<br />
k<br />
= 2 , the corresponding search region in range<br />
dimension at k th frame is well defined by RT(<br />
mk<br />
= 2) . When conditioned on a different ambiguity<br />
number, say, m = 6 , the search region is redirected to a further region in range, R ( m = 6)<br />
.<br />
k<br />
T<br />
k<br />
(33)<br />
(34)<br />
Fig.4. Illustration <strong>of</strong> admissible search regions among successive visits<br />
17
As illustrated in Fig.4, the structure information <strong>of</strong> the ambiguous range-Doppler maps can be well<br />
exploited to limit the search region to a small set <strong>of</strong> admissible locations <strong>and</strong> ambiguity numbers.<br />
Limiting the search region can reduce the computational burden.<br />
Once the merit function <strong>of</strong> last frame has been calculated, the outer maximization <strong>of</strong> (29) is<br />
performed to yield the last location <strong>and</strong> ambiguity number estimates<br />
( xˆ<br />
, mˆ<br />
) = argmax I ( x , m )<br />
(35)<br />
k k k k k<br />
xk,<br />
mk<br />
The previous location <strong>and</strong> ambiguity number estimates are found by tracing backwards from ( xˆ<br />
, ˆ<br />
k<br />
mk)<br />
.<br />
A <strong>final</strong> remark is that the merit function in (34) consists <strong>of</strong> two parts, the measurement<br />
information part <strong>and</strong> the transition probability information parts (including mode conditioned state<br />
transition <strong>and</strong> state dependent mode transition). When the noise σ<br />
2 is known to be very high, the<br />
measurement information part represented by the first two terms is trivial. As a consequence, the<br />
transition probability information parts are much more relied to determine the target trajectory. When<br />
there is no prior information about the target dynamics, all transition probabilities from feasible<br />
previous state <strong>and</strong> mode to the current state <strong>and</strong> mode can be assumed equal. The first two terms in the<br />
maximization can be treated as constants <strong>and</strong> dropped.<br />
5.2 Unknown Nuisance Parameters<br />
k<br />
When the nuisance parameter B is unknown, we assume it’s an unknown deterministic sequence.<br />
k<br />
k<br />
In fact, the target power sequence U = ( u1,..., u k<br />
) in B could also be considered as a Markov sequence,<br />
but for TBD application, this information is weak comparing to the information on the Markovian<br />
k<br />
k<br />
k<br />
property <strong>of</strong> X <strong>and</strong> Μ . For the deterministic parameter B without a priori information, its MAP<br />
k<br />
k<br />
estimate is equivalent to its ML estimate. The MAP estimates <strong>of</strong> X <strong>and</strong> Μ , <strong>and</strong> ML estimates <strong>of</strong><br />
k<br />
k k k k<br />
B can be obtained by maximizing the conditional joint PDF p( Z , X , Μ | B ).<br />
ˆ k<br />
( , ˆ k<br />
, ˆ k k k k k<br />
X Μ B ) = arg max p( X , Μ , Z | B ) (36)<br />
,<br />
k k k X Μ , B<br />
The above maximization is split into two parts<br />
ˆ k ˆ k ˆk k k k k<br />
( X , Μ , B ) = argmax<br />
⎡<br />
max p( , , | )<br />
⎤<br />
, k 1<br />
,<br />
k 1<br />
k ,<br />
k<br />
k m ⎢<br />
X Μ Z B<br />
x − −<br />
⎣X Μ B<br />
⎥<br />
(37)<br />
⎦<br />
The inner maximization is denoted by<br />
( , ) max ( k , k , k | k<br />
Γ x m = p X M Z B ) (38)<br />
k k k<br />
,<br />
X<br />
k−1 Mk−<br />
1 , Bk<br />
The desired iterative relation is obtained using Bayes’ theorem to express Γ<br />
k<br />
k<br />
<strong>of</strong> Γ ( x , m ). Since X is independent <strong>of</strong> B , from Bayes’ theorem it follows that<br />
k k k<br />
p<br />
k+ 1 k+ 1 k+ 1 k+<br />
1<br />
( X , Μ , Z | B )<br />
k+<br />
1 k k k k+1<br />
=( p zk+ 1, xk+ 1, mk+<br />
1| xk, mk, B ) p( Z , X , Μ | B )<br />
zk +1<br />
xk + 1<br />
b<br />
k + 1<br />
, while k + 1<br />
( x , m<br />
k+ 1 k+ 1 k+<br />
1<br />
) in terms<br />
Since the depends only on <strong>and</strong><br />
x depends on xk<br />
<strong>and</strong> mk<br />
; m<br />
k + 1<br />
depends<br />
on x <strong>and</strong> m , the first term on the right side <strong>of</strong> (39) is broken down into<br />
k k<br />
k+<br />
1<br />
( zk+ 1, xk+ 1, k+<br />
1| xk, k, B )<br />
p m m<br />
(40)<br />
= p( zk+ 1| xk+ 1, bk+ 1) p( xk+ 1| xk, mk) p( mk+<br />
1| xk, mk)<br />
Using (39) <strong>and</strong> (40), we can write Γk+ 1( x<br />
k+ 1,<br />
mk+<br />
1)<br />
in the form<br />
Γk+ 1( xk+ 1, mk+<br />
1)<br />
k+ 1 k+ 1 k+ 1 k+<br />
1<br />
= max p( X , Μ , Z | B )<br />
,<br />
Xk , Μk Bk+<br />
1<br />
= max p( zk+ 1| xk+ 1, bk+ 1) max { p( xk+ 1| xk, mk) p( mk+<br />
1| mk, xk) Γk( xk, mk)<br />
}<br />
bk+<br />
1<br />
xk,<br />
mk<br />
which is the required iterative equation. We work with the natural logarithm <strong>of</strong> (41) to transform the<br />
merit function into an additive function<br />
Γk+ 1( xk+ 1, mk+<br />
1)<br />
= max ln p( z | x , b ) +<br />
(42)<br />
bk<br />
+ 1<br />
xk,<br />
mk<br />
k+ 1 k+ 1 k+<br />
1<br />
{ p x x m + p m m x +Γ x m }<br />
max ln ( | , ) ln ( | , ) ( , )<br />
k+ 1 k k k+<br />
1 k k k k k<br />
(39)<br />
(41)<br />
18
where we are using the same symbol Γ for the redefined merit function <strong>of</strong> (42). Note that similar to<br />
(34), the merit function in (42) also consists <strong>of</strong> two parts, the measurement information part<br />
(represented by the first term on the right side) <strong>and</strong> the transition information part (represented by the<br />
last three terms on the right side). Equation (42) differs (34) only in the measurement information part,<br />
where the former has to perform maximization over the space <strong>of</strong> nuisance parameters b<br />
k + 1<br />
. We now<br />
consider the ML estimate <strong>of</strong> the nuisance parameters. The logarithm <strong>of</strong> the likelihood function in (42)<br />
can be written as<br />
2 sk+ 1+<br />
uk+<br />
1<br />
0 0<br />
( ) { ( )}<br />
( i<br />
ln , l ) 2<br />
p zk+ 1| xk+ 1, bk+ 1<br />
=−Nlnσ<br />
− + ln I<br />
2<br />
0<br />
2 uk+ 1zk+<br />
1<br />
/ σ<br />
(43)<br />
σ<br />
(,) il<br />
where s = ∑ S +<br />
; ( i0, l0)<br />
denotes the corresponding cell when a target locates at x k +1<br />
.<br />
k+ 1<br />
z<br />
(,) il∈<br />
k 1<br />
We evaluate the partial derivatives <strong>of</strong> the logarithm likelihood function as<br />
( i0, l0) 2<br />
( )<br />
I ( )<br />
( 0, 0)<br />
1<br />
2<br />
1 1<br />
/<br />
i l<br />
z<br />
uk+ zk+<br />
σ<br />
k+ 1<br />
xk+ 1 k+ 1 1<br />
k+<br />
1<br />
=− +<br />
2<br />
2<br />
∂u ( i0, l0) 2<br />
k+ 1<br />
σ I ( )<br />
1<br />
0<br />
2 u<br />
1 1<br />
/<br />
k<br />
k<br />
zk<br />
σ σ u<br />
+<br />
+ +<br />
( i0, l0) 2<br />
I (<br />
( )<br />
0, 0)<br />
1<br />
2 u<br />
1 1<br />
/ i l<br />
k+ zk+<br />
σ<br />
k+ 1<br />
xk+ 1 k+ 1<br />
2 z<br />
1 1<br />
k<br />
u<br />
k+ k+<br />
k<br />
= − −<br />
2<br />
2<br />
2 2<br />
( i0, l0) 2 2<br />
2<br />
∂σ<br />
( σ ) σ I0( 2 uk+ 1zk+<br />
1<br />
/ σ ) ( σ )<br />
∂ln p | , b 1<br />
z<br />
( z<br />
)<br />
∂ ln p | , b s + u N<br />
+ 1 + 1<br />
where I() = I() ′ is the first-order modified Bessel Function. Equating (44) to zero, we obtain<br />
1 0<br />
( i0, l0) ( uˆk+ zk+<br />
2<br />
σˆ<br />
)<br />
( i0, l0) ( uˆk+ zk+<br />
2<br />
σˆ<br />
)<br />
I1 2<br />
1 1<br />
/<br />
uˆ<br />
k + 1<br />
( i0, l0)<br />
I 1<br />
0<br />
2<br />
1 1<br />
/ zk<br />
+<br />
= (46)<br />
Substituting the result to (45), <strong>and</strong> equating it to zero, we have<br />
2<br />
uˆk<br />
+ 1= s k + 1− Nˆσ<br />
(47)<br />
By solving (46) <strong>and</strong> (47) jointly, we can find the MAP estimate <strong>of</strong> the unknown<br />
2<br />
parameter bˆ 2<br />
ˆ ˆ<br />
k+ 1<br />
( σ , uk+<br />
1)<br />
. Notice that both ˆ σ <strong>and</strong> u ˆk + 1<br />
enter the function <strong>of</strong> (46) in a nontrivial way,<br />
therefore, solving the function will be an iterative numerical process, which is computationally<br />
expensive.<br />
Now we consider a simple way <strong>of</strong> estimating (<br />
2<br />
, ). The target signature only affects the cell it<br />
σ u k + 1<br />
locates, as a result, if a target is present <strong>and</strong> locates in cell ( i0, l0)<br />
at k + 1st frame, the pixel ( i0 , l0<br />
z )<br />
k + 1<br />
is a<br />
(,) il<br />
noncentral chi-square variable, while the other pixels z<br />
k + 1<br />
, (, il)<br />
∈ S , (, il) ≠ ( i0, l0)<br />
, are exponentially<br />
distributed variables. Both the noncentral chi-square <strong>and</strong> exponential variables share the same<br />
2<br />
parameterσ , therefore, we can estimateσ 2 (mean <strong>of</strong> the exponential distribution) by averaging the<br />
exponentially distributed pixels<br />
( i0, l0)<br />
2 1<br />
( , ) 1 1<br />
ˆ σ il sk − zk<br />
= ∑ z<br />
+ +<br />
(,) il ,(,) il ( i 1<br />
0, l0)<br />
k +<br />
=<br />
∈ ≠<br />
N −1 S<br />
(48)<br />
N −1<br />
Substituting (48) into (47) gives<br />
( i0, l0)<br />
Nzk+<br />
1<br />
− sk+1<br />
uˆ<br />
k + 1<br />
=<br />
(49)<br />
N −1<br />
Substituting (48) <strong>and</strong> (49) into (42) <strong>and</strong> dropping the terms independent <strong>of</strong> ( xk+ 1,<br />
mk+<br />
1)<br />
, we end up with<br />
a modified merit function<br />
Γ ( x , m )<br />
k+ 1 k+ 1 k+<br />
1<br />
( i0, l0)<br />
( i0, l0) ( i0, l0)<br />
(2N−1)<br />
z 2 ( N 1)( N z<br />
1<br />
k 1<br />
sk)<br />
z<br />
k+<br />
−s<br />
⎧<br />
⎫<br />
k ⎪<br />
⎛ −<br />
+<br />
− ⎞<br />
k+<br />
1 ⎪<br />
=− +ln I<br />
( i 0<br />
0, l0) ⎨ ⎜<br />
⎟<br />
( i0, l0)<br />
⎬<br />
sk −z ⎜<br />
k+ 1<br />
sk −z<br />
⎟<br />
k+<br />
1<br />
⎩⎪<br />
⎝<br />
⎠⎭⎪<br />
+ max ln ( | , ) + ln ( | , ) +Γ ( , )<br />
xk,<br />
mk<br />
{ p x x m p m m x x m }<br />
k+ 1 k k k+<br />
1 k k k k k<br />
where we are using the same symbol Γ for the redefined merit function <strong>of</strong> (50).<br />
Once the merit function <strong>of</strong> last frame has been calculated, the last state <strong>and</strong> mode estimates are<br />
obtained using<br />
(44)<br />
(45)<br />
(50)<br />
19
( xˆ<br />
, m ˆ ) = argmax Γ ( x , m )<br />
(51)<br />
k k k k k<br />
xk,<br />
mk<br />
The previous states <strong>and</strong> modes are found by tracing backwards from ( xˆ<br />
, mˆ<br />
).<br />
k<br />
k<br />
6. Simulation <strong>and</strong> Results<br />
In our simulation, an approaching target, referring back to Fig. 2, travels at initial range r<br />
0<br />
= 300 km,<br />
with velocity v = 2.8 Mach, <strong>and</strong> headingα = 225 deg. Since the target is not travelling along the radial<br />
direction, the Doppler is inevitably time-varying. We set the Doppler rate increment variance<br />
2 2<br />
Q = (6Hz/s ) .Fig. 5 <strong>and</strong> Fig.6 show the trajectories <strong>of</strong> target range <strong>and</strong> Doppler, respectively. The<br />
target Doppler information is represented by the observed apparent Doppler <strong>and</strong> corresponding<br />
ambiguity number.<br />
Fig. 5. target range trajectory<br />
Fig. 6. target Doppler trajectory, represented by apparent Doppler trajectory <strong>and</strong> ambiguity number<br />
trajectory<br />
The ambiguity number transits for i to j = i − 1 when the guard condition is satisfied. To account for<br />
the r<strong>and</strong>om variations in Doppler increment, δ ij<br />
is modeled as a Gaussian distributed variable with<br />
20
mean δ<br />
ij<br />
= 50 Hz <strong>and</strong> variance Σ<br />
ij<br />
= 16 , denoted byδij<br />
N ( δij<br />
;50,16), <strong>and</strong> then, as discussed in Section<br />
3, the state-dependent ambiguity number transition probability is<br />
⎧Φ( Cx<br />
ij k<br />
−δij ; Σ<br />
ij ) if j = i<br />
⎪<br />
πij ( xk ) = ⎨1 −Φ( Cijx k<br />
−δij; Σ<br />
ij ) if j = i−1<br />
(52)<br />
⎪<br />
0 else<br />
⎪⎩<br />
A maximum target velocity v<br />
max<br />
= 3Mach is assumed in our simulation.<br />
We consider here a low PRF surveillance radar using a carrier frequency <strong>of</strong> 1.2 GHz <strong>and</strong><br />
transmitting trains <strong>of</strong> N<br />
d<br />
= 64 rectangular pulses with pulse width τ = 1.2 μs<br />
<strong>and</strong> PRF=860 Hz. Radar<br />
revisit interval is T<br />
R<br />
= 4 s . For the reader’s sake, the main surveillance radar parameters are<br />
summarized in Table I .After discretization <strong>and</strong> Doppler processing, the received data are transformed<br />
into the ambiguous range-Doppler maps. For low PRF surveillance radars, the measured range is<br />
unambiguous while the measured Doppler is ambiguous. Since we are primarily concerned with early<br />
warning <strong>of</strong> approaching targets, the range cells under consideration are limited to the<br />
furthest N<br />
r<br />
= 256 cells. Thus, after each visit, we obtained a 256× 64 ambiguous range-Doppler map. In<br />
the simulation, the unknown target’s power u k<br />
at each visit is r<strong>and</strong>omly drawn from an exponential<br />
distribution with average return power u . Note that we ignore the power variation caused by the<br />
changing range since the range variation is small in comparison to the distance between the target <strong>and</strong><br />
radar. The SNR is defined in range-Doppler domain. The SNR <strong>of</strong> the target at kth visit is expressed<br />
2<br />
as (SNR) = 10log( u / σ ) .<br />
k<br />
k<br />
Table I Radar System Parameters<br />
It’s clear that SNR experienced by the target varies at each visit due to the variation <strong>of</strong> target power. In<br />
the following, in order to characterize the detection performance <strong>of</strong> the proposed TBD algorithms, we<br />
define an average SNR as<br />
2<br />
SNR = 10log( u / σ ) (53)<br />
Notice that, since we assume the number <strong>of</strong> coherently processed pulses N<br />
d<br />
= 64 , the average received<br />
SNR per pulse is obtained by subtracting 10log 64 18dB from the reported values in range-Doppler<br />
domain.<br />
Fig.7 shows a mesh plot <strong>of</strong> the reflected power in ambiguous range-Doppler maps. The target is<br />
located at cell (231, 14). When the SNR is high (13dB) <strong>and</strong> it’s easy to discern the target as a high<br />
peak (Fig.7(a)). When the SNR is relative low (10dB), it’s not easy to distinguish the target from the<br />
noise (Fig.7(b)). TBD procedures provide a way to improve the detestability <strong>of</strong> low SNR target by<br />
jointly processing several range-Doppler maps. In the following simulation, we set the number <strong>of</strong><br />
jointly processed ambiguous range-Doppler maps to K = 6 .<br />
21
(a)<br />
(b)<br />
Fig. 7. mesh plot <strong>of</strong> the reflected power in ambiguous range-Doppler maps: (a) SNR=13dB, (b)<br />
SNR=10dB<br />
Fig. 8 shows the <strong>final</strong> merit function for a fixed mode m<br />
K<br />
= 5 . When no target exists, it’s obvious that<br />
the merit function tends to noisy (Fig.8(a)). When a target with an average SNR=10dB is present (the<br />
nuisance parameters are assumed known), it’s clear that the noisy merit function has a distinct peak at<br />
a particular location in the ambiguous range-Doppler map (Fig.8(b)). The corresponding target <strong>final</strong><br />
state can be extracted by thresholding the resultant merit function.<br />
6.1 Selection <strong>of</strong> Threshold<br />
In order to calculate the threshold, we first have to define the probability <strong>of</strong> false alarm P fa<br />
. In our<br />
situation, Pfa<br />
is defined as the probability <strong>of</strong> detecting a false track in the absence <strong>of</strong> target (i.e., under<br />
the hypothesis H<br />
0<br />
). This is equal to the probability <strong>of</strong> the maximum noise merit function exceeding<br />
the threshold VT<br />
P<br />
fa<br />
= Pr ( max Γ<br />
K<br />
(<br />
K<br />
, mK ) ><br />
T<br />
xK,<br />
mK<br />
x V ) (54)<br />
where x K<br />
∈{noise states} .<br />
Given the desired P fa<br />
, the threshold is determined as the (1−<br />
Pfa<br />
)-quantile <strong>of</strong> the test statistic, i.e.,<br />
max<br />
x ,<br />
( , ) given in<br />
K m<br />
Γ<br />
K K<br />
x<br />
K<br />
mK<br />
(54). Unfortunately, the statistical distribution <strong>of</strong> the test statistic is rather<br />
complicated [11], sometimes even infeasible. This mainly because 1) the max operator in the DP<br />
recursion brings in nontrivial nonlinearities <strong>and</strong> non-Gaussianity; 2) the samples <strong>of</strong> the test statistic are<br />
22
actually correlated since the backward search process in the DP recursion may end with a common<br />
previous location xk −1<br />
for two different locations at frame k . The threshold is, hence, evaluated via<br />
computer simulation.<br />
(a)<br />
(b)<br />
Fig. 8. mesh plot <strong>of</strong> the merit function ( K = 6) for a fixed mode m<br />
K<br />
= 5 : (a) no target exists (b) a target<br />
with average SNR=10dB exists<br />
6.2 Performances Evaluation<br />
We now move to the analysis <strong>of</strong> the detection <strong>and</strong> tracking performance <strong>of</strong> the proposed algorithm.<br />
As in [7], we define the following performance measures for evaluating the proposed TBD algorithms.<br />
Definition 1 The probability <strong>of</strong> target detection P d<br />
is defined as the probability <strong>of</strong> the maximum <strong>final</strong><br />
merit function exceeds the threshold VT<br />
<strong>and</strong> meanwhile, the corresponding retrieved cell, say, ( iˆ<br />
, ˆ<br />
K<br />
l<br />
K<br />
),<br />
is within ε r<br />
cells in range <strong>and</strong> ε d<br />
cells in Doppler from the actual location, ( iK<br />
, l<br />
K<br />
), <strong>and</strong> the ambiguity<br />
number estimate m ˆ K<br />
is equal to the true one. Hence<br />
Pd = Pr<br />
⎛( max<br />
K( K, mK)<br />
⎞<br />
⎜ Γ x > V<br />
T<br />
,<br />
) ∪E ⎟<br />
⎝ xK<br />
m<br />
d<br />
(55)<br />
K<br />
⎠<br />
where the event E d<br />
is expressed as<br />
abs( iˆ<br />
K<br />
−iK)<br />
≤ ε<br />
r, abs( lˆ<br />
K<br />
−lK)<br />
≤ ε<br />
d<br />
, mˆ K<br />
= m K<br />
. (56)<br />
23
Definition 2 The probability <strong>of</strong> target detection <strong>and</strong> reliable tracking P dT ,<br />
is defined as the probability<br />
<strong>of</strong> making a correct target detection <strong>and</strong> <strong>of</strong> recovering a target trajectory such that each retrieved<br />
state, ( iˆ<br />
ˆ<br />
k, lk)<br />
, k = 1,... K , is within ε<br />
r<br />
cells in range <strong>and</strong> ε d<br />
cells in Doppler from the actual location, <strong>and</strong><br />
the estimated ambiguity number sequence ˆ K<br />
Μ = ( mˆ ˆ<br />
1,..., m K<br />
) is equal to the true one. Hence<br />
PdT ,<br />
= Pr<br />
⎛(<br />
max<br />
K( K, mK)<br />
V<br />
⎞<br />
⎜ Γ x ><br />
T<br />
∪<br />
dT , ⎟<br />
⎝ xK,<br />
m<br />
) E (57)<br />
K<br />
⎠<br />
where the event<br />
E dT ,<br />
is expressed as<br />
abs( iˆ<br />
k<br />
−ik)<br />
≤ ε<br />
r, abs( lˆ<br />
k<br />
−lk)<br />
≤ εd, k = 1,... K , Μˆ K = Μ<br />
K<br />
(58)<br />
It’s evident that the overall system performance depends on the choice <strong>of</strong> the tolerance factors: ε r<br />
<strong>and</strong> ε d<br />
. The smaller the tolerance factors, the larger the SNR is required in order to maintain a certain<br />
value <strong>of</strong> detection probability P d<br />
<strong>and</strong> P under a preset probability <strong>of</strong> false alarm rate. In the<br />
dT ,<br />
following, the detection <strong>and</strong> tracking performance with respect to average SNR <strong>and</strong> tolerance factors<br />
are investigated.<br />
We first consider the case that the nuisance parameters are known in advance; the corresponding<br />
−6<br />
implementation <strong>of</strong> the DP recursive is given in (33). In Fig.9, we set P fa<br />
= 10 , <strong>and</strong> plot P<br />
d<br />
<strong>and</strong> P versus average SNR under different settings <strong>of</strong> tolerance factor. The tolerance<br />
dT ,<br />
factor ( ε<br />
r, ε<br />
d)<br />
from up to down is (4,2) , (2,1) <strong>and</strong> (0,0) . ( ε<br />
r, ε<br />
d) = (0,0) means correct detection <strong>and</strong><br />
estimation. We note that as the tolerance factor becomes smaller (i.e., the estimation accuracy<br />
becomes higher), both P d<br />
<strong>and</strong> P dT ,<br />
decreases considerably. We can also notice that compared to P dT ,<br />
,<br />
P is less sensitive to the tolerance factor. Satisfactory detection <strong>and</strong> tracking performance can be<br />
d<br />
achieved when the tolerance factor is (4,2) .<br />
Fig. 9. P <strong>and</strong><br />
d<br />
P dT ,<br />
versus the average SNR in range-Doppler domain, with known nuisance parameters,<br />
( ε , ε ) = (4, 2), (2, 1), (0, 0) from up to down<br />
r<br />
d<br />
We now investigate the situation when the nuisance parameters are unknown; the DP recursive with<br />
−6<br />
unknown nuisance parameters is given in (50). In Fig. 10, we set P fa<br />
= 10 , <strong>and</strong> report P d<br />
24
<strong>and</strong> P dT ,<br />
versus the average SNR under several values <strong>of</strong> tolerance factor. For the sake <strong>of</strong> comparison,<br />
the tolerance factor ( ε<br />
r, ε<br />
d)<br />
from up to down is also set to (4,2) , (2,1) <strong>and</strong> (0,0) . Similar result to Fig. 10<br />
is obtained: both P d<br />
<strong>and</strong> P decreases considerably with the decreasing <strong>of</strong> tolerance factors. By<br />
dT ,<br />
inspection, we note that P d<br />
<strong>and</strong> P dT ,<br />
suffer an approximately 1dB loss due to the prior uncertainty as to<br />
the nuisance parameters.<br />
Fig. 10. P <strong>and</strong><br />
d<br />
P versus the average SNR in range-Doppler domain, with unknown nuisance<br />
dT ,<br />
parameters,<br />
( ε , ε ) = (4, 2), (2, 1), (0, 0) from up to down<br />
r<br />
d<br />
7. Conclusions<br />
In this paper we have investigated the TBD procedures with reference to a low PRF surveillance<br />
radar framework. We give physical <strong>and</strong> statistical models for the received power measurements on the<br />
ambiguous range-Doppler maps. In order to avoid the nonlinear transformation between the target’s<br />
state coordinates <strong>and</strong> measurement coordinates (defined by range <strong>and</strong> apparent Doppler); we model<br />
the target’s dynamics directly in the measurement coordinate. Exploiting the target dynamics in the<br />
ambiguous range-Doppler maps as a hybrid system, we have proposed DP based methods for joint<br />
MAP estimation <strong>of</strong> target’s trajectory in the ambiguous range-Doppler maps <strong>and</strong> the corresponding<br />
ambiguity number sequence. The detection <strong>and</strong> tracking performance <strong>of</strong> the proposed procedure is<br />
studied using two measures: 1) detection probability P d<br />
, 2) detection <strong>and</strong> reliable tracking<br />
probability P dT ,<br />
. Simulation results show satisfactory detection <strong>and</strong> tracking performance is achieved<br />
when the tolerance factor is set to within four cells in range <strong>and</strong> two cells in Doppler from the true<br />
location. The impact <strong>of</strong> the uncertainties in the nuisance parameters (target power <strong>and</strong> noise variance)<br />
has also been studied. Simulation shows that both P d<br />
<strong>and</strong> P dT ,<br />
suffer an approximately 1dB loss due to<br />
the prior uncertainty <strong>of</strong> the nuisance parameters.<br />
25
Acknowledgments<br />
Xiaobo Deng would like to acknowledge his useful discussion with Dr. Xuezhi Wang<br />
(Department <strong>of</strong> EEE, The University <strong>of</strong> Melbourne), <strong>and</strong> thank the China Scholarship Council for<br />
supporting his visit at the Department <strong>of</strong> EEE, The University <strong>of</strong> Melbourne.<br />
References:<br />
[1] Reed, I., Gagliardi, R. <strong>and</strong> Stotts, L.: ‘A recursive moving-target-indication algorithm for optical<br />
image sequences’, IEEE Trans. Aerosp. Electron. Syst., 1990, 26, (1), pp. 434-440<br />
[2] Barniv, Y.: ‘Dynamic programming solution for detecting dim moving targets’. IEEE Trans.<br />
Aerosp. Electron. Syst., 1985, 21, (1), pp. 141-155<br />
[3] Arnold, J., Shaw, S., <strong>and</strong> Pasternack, H.: ‘Efficient target tracking using dynamic programming.<br />
IEEE Trans. Aerosp. Electron. Syst., 1993, 29, (1), pp. 44-56<br />
[4] Xiong. Y, Peng, J. X., Ding, M. Y., <strong>and</strong> Xue, D. H.: ‘An extended track-before-detect algorithm<br />
for infrared target detection’, IEEE Trans. Aerosp. Electron. Syst., 1997, 33, (3), pp. 1087-1092<br />
[5] Tonissen, S. M., <strong>and</strong> Evans, R. J.: ‘Performance <strong>of</strong> dynamic programming techniques for trackbefore-detect’,<br />
IEEE Trans. Aerosp. Electron. Syst., 1996, 32, (4), pp. 1440-1451<br />
[6] Wallace, W. R.: ‘The use <strong>of</strong> track-before-detect in pulse-Doppler radar’. Proc. Int. Conf. Radar<br />
2002, pp. 315-319<br />
[7] Buzzi, S., Lops, M., <strong>and</strong> Venturino, L.: ‘Track-before-detect procedures for early detection <strong>of</strong><br />
moving target from airborne radars’, IEEE Trans. Aerosp. Electron. Syst., 2005, 41, (3), pp. 937-<br />
954<br />
[8] Buzzi, S., Lops, M., <strong>and</strong> Ferri, M.: ‘Track-before-detect procedures in a Multi-target environment’,<br />
IEEE Trans. Aerosp. Electron. Syst., 2008, 44, (3), pp. 1135-1148<br />
[9] Seah, C. <strong>and</strong> Hwang, I.: ‘State estimation for stochastic linear hybrid systems with continuousstate-dependent<br />
transitions: an IMM approach’, IEEE Trans. Aerosp. Electron. Syst., 2009, 45, (1),<br />
pp. 376-392<br />
[10] Blom, H., <strong>and</strong> Bloem, E.: ‘Exact Bayesian <strong>and</strong> particle filtering <strong>of</strong> stochastic hybrid systems’,<br />
IEEE Trans. Aerosp. Electron. Syst., 2007, 43, (1), pp. 50-70<br />
[11] Johnston, L. A., <strong>and</strong> Krishnamurthy, V.: ‘Performance analysis <strong>of</strong> a dynamic programming track<br />
before detect algorithm’, IEEE Trans. Aerosp. Electron. Syst., 2002, 38, (1), pp. 228-242<br />
26
Fast realization <strong>of</strong> Automatic optical <strong>and</strong> infrared image registration<br />
Weiping Yang 1 2 * , Xuezhi Wang 2 , Bill Moran 2<br />
* Presenter<br />
1. National University <strong>of</strong> Defence Technology, Hunan, 410073, China<br />
2. The University <strong>of</strong> Melbourne, Parkville, VIC 3010, Australia<br />
Because <strong>of</strong> the requirement <strong>of</strong> practical applications, it is necessary for image<br />
registration between different source images. However, the image characteristics from<br />
different imaging sensors are different, so the wanted results can not be obtained if the<br />
intensive images are directly used. In order to solve this problem, edge detection <strong>and</strong><br />
extraction methods should be applied, then the correlation tactics is used. But because the<br />
st<strong>and</strong>ard cross correlation method is time expensive, fast realization method should be<br />
adopted. Here we use the changeable resolution method to do this, <strong>and</strong> get good results with<br />
only less than one fortieth time consuming <strong>of</strong> the st<strong>and</strong>ard algorithm.<br />
Keywords automatic image registration, edge detection, changeable resolution, normalized<br />
cross correlation<br />
1. Introduction<br />
Water resource is the key factor <strong>of</strong> human being social development, <strong>and</strong> paid more attention by<br />
the world. The issue <strong>of</strong> how to use the limited freshwater effectively is one <strong>of</strong> the subjects in all over<br />
the world. In order to study automatic water saving irrigation technique, we need to know the water<br />
status <strong>of</strong> the plants. To say further, if we want to know the plant water status, we need to know the<br />
canopy temperature at first. Therefore, researchers have done many fruitful results [1-5] . Using<br />
nondestructive optical image sensing technique to obtain the canopy temperature is a challenging<br />
project. The first step is to automatically register the infrared image with the optical image, then<br />
extract the temperature <strong>of</strong> the interested regions, <strong>and</strong> <strong>final</strong>ly can get the temperature estimation <strong>of</strong> the<br />
canopy.<br />
As we all know, researchers have presented many algorithms on image registration <strong>and</strong> image<br />
match, for example, cross correlation [6] , mutual information [7] , correlation ratio [8] , <strong>and</strong> SIFT based<br />
methods [9] . However, for image registration between different source image pairs, there are some<br />
practical problems if we directly use the intensive images to do the image registration processing, in<br />
reference [10] we have pointed out, here we only discuss its fast realization. It is because <strong>of</strong> the<br />
existing differences between optical <strong>and</strong> infrared images that we should edge the images at first, then<br />
using certain image registration method to process the edge images. Because <strong>of</strong> the time consuming,<br />
st<strong>and</strong>ard cross correlation( normalized cross correlation) method is improper, the changeable<br />
resolution method based normalized cross correlation may be the best choice. Experiment tests show<br />
that the changeable resolution normalized cross correlation method has not only the same good<br />
registration results, but also fast running speed with less than one fortieth time consuming <strong>of</strong> the<br />
st<strong>and</strong>ard cross correlation method.<br />
The paper is organized as follows: In Section 2 the algorithm is described <strong>and</strong> the flow diagram is<br />
given. Section 3 is the implementation <strong>of</strong> the above algorithm. The experiment results are presented in<br />
Section 4, then followed by the discussion <strong>and</strong> conclusion.<br />
2. Changeable resolution algorithm based on normalized cross correlation<br />
When we want to register the IR image to the reference optical image, we should obtain its<br />
location <strong>and</strong> rotation angle. In general, the IR image is wholly overlapped with the reference optical<br />
image. Moreover, we assume that the space resolution <strong>of</strong> the two images is the same. And then the<br />
changeable resolution algorithm based on normalized cross correlation can be shown as Fig 1.<br />
Because it needs to calculate the correlation coefficient pixel by pixel, normalized cross<br />
correlation (NCC) algorithm is time expensive. In order to solve the problem <strong>of</strong> the slowly running<br />
27
speed, several methods can be utilized. Here we use the changeable resolution algorithm to accelerate<br />
it. The key step <strong>of</strong> the changeable resolution algorithm is to reduce the resolution at first, then roughly<br />
do image registration in the lower resolution images, <strong>and</strong> then do it accurately near the possible<br />
positions in the full resolution images, <strong>final</strong>ly we can get the registration results.<br />
Optical image<br />
IR image<br />
Edge extraction<br />
Edge extraction<br />
Resolution Reducing<br />
Resolution Reducing<br />
NCC<br />
Roughly registration<br />
NCC at certain positions<br />
Selecting optimal position<br />
Rotation angle estimation<br />
Image registration parameters<br />
Fig 1 Flow diagram <strong>of</strong> changeable resolution algorithm based on NCC<br />
Suppose Im O <strong>and</strong> Im IR (image size is M×N) st<strong>and</strong>s for the optical <strong>and</strong> IR image respectively, f is<br />
the image zoom factor, Im Ol <strong>and</strong> Im IRl (image size is Mh×Nh) st<strong>and</strong>s for the lower resolution one,<br />
accordingly, Im Oe , Im IRe , Im Ole , Im IRle are the edged images <strong>of</strong> the corresponding images, then the<br />
generalize cross correlation can be expressed as<br />
ρ ( u,<br />
v)<br />
=<br />
l<br />
ρ(<br />
k,<br />
l)<br />
=<br />
Mh<br />
Nh<br />
∑∑<br />
i= 0 j=<br />
0<br />
M<br />
N<br />
∑∑<br />
i= 0 j=<br />
0<br />
(Im<br />
(Im<br />
Oleu, v<br />
( i,<br />
j)<br />
− ImIRle<br />
( i,<br />
j))<br />
σ<br />
lu,<br />
v<br />
k , l<br />
σ<br />
IR<br />
IRl<br />
Oek, l<br />
( i,<br />
j)<br />
− ImI<br />
Re<br />
( i,<br />
j))<br />
σ<br />
σ<br />
Where ρ<br />
l<br />
( u,<br />
v)<br />
is the cross correlation coefficient calculated from the lower resolution image pair<br />
Im Ole <strong>and</strong> Im IRle , (u,v) is the coordinate index <strong>of</strong> the optical image Im Ole , Im Oleu,v<br />
is an image located in<br />
(u,v)th <strong>of</strong> the image Im Ole <strong>and</strong> its size is the same as image Im IRle , σ lu,v <strong>and</strong> σ IRl are the st<strong>and</strong>ard<br />
deviation <strong>of</strong> the corresponding images respectively; ρ ( k,<br />
l)<br />
is the cross correlation coefficient<br />
calculated from the raw resolution image pair Im Oe <strong>and</strong> Im IRe , (k,l) is the coordinate index <strong>of</strong> the<br />
(1)<br />
(2)<br />
28
optical image Im Oe ,<br />
Im Oek,l<br />
is an image located in (k,l)th <strong>of</strong> the image Im Oe <strong>and</strong> its size is the same as<br />
image Im IRe , σ k,l <strong>and</strong> σ IR are the st<strong>and</strong>ard deviation <strong>of</strong> the corresponding images respectively.<br />
3. Algorithm implementation <strong>of</strong> the changeable resolution algorithm<br />
(a) Raw optical image<br />
(b) Raw IR image<br />
Fig 2 Raw images<br />
In order to implement the algorithm, the first step is edge extraction. It is important to the results<br />
<strong>of</strong> image registration. There are many methods on edge detection <strong>and</strong> extraction, such as Prewitt,<br />
Sobel, Robert operators <strong>and</strong> Canny algorithm, <strong>and</strong> so on. As the above mentioned, the infrared image,<br />
different with the reference optical image(Fig 2), has only a lower resolution <strong>and</strong> less details, we need<br />
to reserve its edge information while edge processing. Here the edge operation is based on the Sobel<br />
operator, <strong>and</strong> it can be described as follow,<br />
⎡1<br />
0 −1⎤<br />
⎡ 1 2 1 ⎤<br />
S =<br />
⎢ ⎥<br />
⎢<br />
2 0 − 2 , , (3a)<br />
1<br />
⎥ S<br />
⎢<br />
⎥<br />
2<br />
=<br />
⎢<br />
0 0 0<br />
⎥<br />
⎢⎣<br />
1 0 −1⎥⎦<br />
⎢⎣<br />
−1<br />
− 2 −1⎥⎦<br />
⎡2<br />
1 0 ⎤ ⎡ 0 1 2⎤<br />
S =<br />
⎢ ⎥<br />
⎢<br />
1 0 −1<br />
,<br />
3<br />
⎥<br />
S<br />
⎢ ⎥<br />
(3b)<br />
4<br />
=<br />
⎢<br />
−1<br />
0 1<br />
⎥<br />
⎢⎣<br />
0 −1<br />
− 2⎥⎦<br />
⎢⎣<br />
− 2 −1<br />
0⎥⎦<br />
Using these four operators to convolute with the raw images, we can get<br />
2<br />
2<br />
∑∑<br />
E ( i,<br />
j)<br />
= Im( i + m −1,<br />
j + n −1)*<br />
S ( m,<br />
n)<br />
. k=1,2,3,4<br />
k<br />
m= 0 n=<br />
0<br />
E( i,<br />
j)<br />
= max E ( i,<br />
j)<br />
k<br />
k<br />
k<br />
(5)<br />
Where E( i,<br />
j)<br />
is the edge <strong>of</strong> the point (i,j) <strong>of</strong> image Im.<br />
After edge detection, we can get the edged image Im Oe , Im IRe , Im Ole , Im IRle , then using NCC<br />
algorithm to do the rough registration. On account <strong>of</strong> possible disturbing, we choose some points with<br />
the former higher correlation coefficients, marked as {( u<br />
k<br />
, vk<br />
)}.<br />
When we get the c<strong>and</strong>idate points{ ( u k<br />
, v , next step is to do the accurate image registration near<br />
k<br />
)}<br />
these specified points <strong>of</strong> the full resolution images. Here we adopt a weighted coefficient c to ensure<br />
that the <strong>final</strong> point with the maximum correlation coefficient is the optimal one.<br />
ρ<br />
k<br />
= c ρl<br />
( uk<br />
, vk<br />
) + (1 − c)<br />
ρ(<br />
u′<br />
k<br />
, v′<br />
k<br />
)<br />
(6)<br />
ρ max ρ<br />
* =<br />
k<br />
k<br />
k<br />
(4)<br />
(7)<br />
29
Where u ′ , v′<br />
) is the point position in the full resolution image Im Oe , ρ u ′ , v′<br />
) is the correlation<br />
(<br />
k k<br />
(<br />
k k<br />
coefficient <strong>of</strong> accurate image registration, ρl<br />
( uk<br />
, vk<br />
) is the correlation coefficient <strong>of</strong> rough image<br />
registration. Through the step, we can get the registration position u ′ , ′ ) .<br />
( * v *<br />
k k<br />
The <strong>final</strong> step is the estimation <strong>of</strong> rotation angle. At present the permitted rotation angle θ range is<br />
from -10°to 10°. We divide the range into 200 angles. To each angle, we calculate its correlation<br />
coefficient, then choose the angle θ * with the maximum correlation value.<br />
4. Experiment results<br />
The above algorithm was tested using many image pairs. All <strong>of</strong> the results are successful with<br />
tolerable errors(Fig 3). In order to test the running speed, contrasting experiment was done between<br />
the changeable resolution algorithm <strong>and</strong> st<strong>and</strong>ard NCC in the same running environment <strong>and</strong><br />
conditions, the results can be shown in table 1. The experiments show that the changeable resolution<br />
algorithm has the same good registration results with a fast speed, only spending less than one fortieth<br />
running time compared with the st<strong>and</strong>ard algorithm. Because we don’t know the true registration<br />
positions, we compare them with the results <strong>of</strong> manual registration.<br />
Fig 3 Raw image pairs <strong>and</strong> their registration results<br />
Table 1 experiment results <strong>of</strong> changeable resolution algorithm based on NCC<br />
Number <strong>of</strong> Success Running time per pair Registration error<br />
Image pairs number (average)<br />
(max, min, average)<br />
St<strong>and</strong>ard NCC 20 20 610s (6,0,2.6)<br />
Changeable<br />
resolution<br />
algorithm<br />
20 20 15s (6,0,2.6)<br />
All <strong>of</strong> the tests are done in the condition <strong>of</strong> f = 2 <strong>and</strong> using Sobel operator to get the image edges.<br />
The contrasting experiments with the NMI (normalized mutual information) <strong>and</strong> CR(correlation ratio)<br />
algorithms have also been done( Fig 4, Table 2). From these figures <strong>and</strong> tables, we can see the<br />
changeable resolution algorithm based on normalized cross correlation is the best choice.<br />
30
Table 2 experiment results <strong>of</strong> several different algorithms<br />
Number <strong>of</strong> Image Success* Registration error<br />
pairs<br />
number (max, min, average)<br />
Changeable resolution 10 10 (6,0,2.6)<br />
algorithm<br />
NMI algorithm 10 8 (9,0,2.7)<br />
CR algorithm 10 0 Null<br />
Remark: *here a successful registration is whose position error is less than 10 pixels.<br />
(a) Image pair 1(Optical(left), IR(right))<br />
(299,120,-0.8)<br />
(b) registration result 1 <strong>of</strong> Changeable Res Algo<br />
(c) registration result 1 <strong>of</strong> NMI alg (304,125,-0.5) (d) registration result 1 <strong>of</strong> CR Alg (318,257,0.5)<br />
(e) Image pair 2(Optical(left), IR(right)) (f) registration result 2 <strong>of</strong> Changeable Res Algo (364,153,1.5)<br />
(g) registration result 2 <strong>of</strong> NMI alg (363,157,1.0) (h) registration result 2 <strong>of</strong> CR Alg (283,93,3.9)<br />
Fig 4 Raw image pairs <strong>and</strong> their registration results using different algorithms<br />
5. Discussion<br />
From the experiment results we can find that NMI <strong>and</strong> CR algorithms are unsuitable to the<br />
different source image registration because these two methods are dependent on the intensity<br />
distributions <strong>of</strong> the images, while the changeable resolution algorithm is related to the edge <strong>of</strong> the<br />
images, that is to say, the latter algorithm is dependent on the image features <strong>and</strong> independent with the<br />
image intensities. Though they have good results in many applications, the image registration methods<br />
31
ase on SIFT are also unsuitable to our application because common key points can hardly be found.<br />
The important reason is that the key points are wholly dependent on the image intensities.<br />
With regard to the running speed, though the images are only zoomed out at 2 times, a higher<br />
running efficiency is obtained. The changeable resolution algorithm only takes less than one fortieth<br />
running time <strong>of</strong> the st<strong>and</strong>ard pixel by pixel cross correlation method, but they have same registration<br />
results, so the changeable resolution algorithm is a fast <strong>and</strong> effective realization <strong>of</strong> the NCC method.<br />
6. Conclusion<br />
The paper presented a fast realization algorithm <strong>of</strong> automatic optical <strong>and</strong> infrared image registration.<br />
It reduces the computation considerably by utilizing changeable resolution algorithm <strong>and</strong> has obtained<br />
inspiring results at a fast running speed. It can also be used in other applications. As we mentioned<br />
above, the image registration is only the first work to do, the aim is to estimate the canopy temperature.<br />
But delightful thing is that it is our ongoing research, <strong>and</strong> some great progresses have been made.<br />
Moreover, how to get more accurate registration <strong>and</strong> how to make algorithm suit to more variational<br />
background are also our ongoing research.<br />
7. Acknowledgement<br />
We would like to thank Mr Ashley Wheaton <strong>and</strong> Dr Nicola Cooley for their providing images <strong>and</strong><br />
some useful discussions.<br />
References:<br />
[1] H.G. Jones. Use <strong>of</strong> thermography for quantitative studies <strong>of</strong> spatial <strong>and</strong> temporal variation <strong>of</strong><br />
stomatal conductance over leaf surfaces. Journal <strong>of</strong> Plant, Cell <strong>and</strong> Environment (1999) 22, 1043–<br />
1055.<br />
[2] H.G. Jones. Use <strong>of</strong> infrared thermometry for estimation <strong>of</strong> stomatal conductance as a possible aid<br />
to irrigation scheduling. Journal <strong>of</strong> Agricultural <strong>and</strong> Forest Meteorology 95 (1999) 139-149.<br />
[3] L. Guilioni , H.G. Jones, I. Leinonen, J.P. Lhomme. On the relationships between stomatal<br />
resistance <strong>and</strong> leaf temperatures in thermography. Journal <strong>of</strong> Agricultural <strong>and</strong> Forest Meteorology<br />
148 (2008): 1908 – 1912.<br />
[4] O. Grant, L. Tronina1, H. G. Jones <strong>and</strong> M. M. Chaves (2007) Exploring thermal imaging variables<br />
for the detection <strong>of</strong> stress responses in grapevine under different irrigation regimes. Journal <strong>of</strong><br />
Experimental Botany, Vol. 58, No. 4, pp. 815–825, 2007.<br />
[5] Wheaton, A.D., Cooley, N., Dunn, G., Goodwin, I., Needs, S., (2007) Evaluation <strong>of</strong> infrared<br />
thermography to determine the crop water status <strong>of</strong> Cabernet Sauvignon grapevines. Poster paper.<br />
13th Australian Wine Industry Technical Conference, Adelaide, 28 July – 2 August.<br />
[6] Du-Ming Tsai, Chien-Ta Lin. Fast normalized cross correlation for defect detection. Journal <strong>of</strong><br />
Pattern Recognition Letters 24 (2003) 2625–2631.<br />
[7] A. Roche, G. Mal<strong>and</strong>ain, etc. The correlation ratio as a new similarity measure for multimodal<br />
image registration. Medical Image Computing <strong>and</strong> Computer Assisted Intervention - MICCAI'98,<br />
1496: 1115-1124,1998.<br />
[8] S. Klein. Evaluation <strong>of</strong> Optimization Methods for Nonrigid Medical Image Registration Using<br />
Mutual <strong>Information</strong> <strong>and</strong> B-Splines. IEEE Transactions on Image Processing, 16(12): 2879-2890,<br />
2007.<br />
[9] Jian Chen, Jie Tian. Real-time multi-modal rigid registration based on a novel symmetric-SIFT<br />
descriptor. Progress in Natural Science 19 (2009): 643–651.<br />
[10] Yang Weiping, Wang Xuezhi, etc, Automatic optical <strong>and</strong> IR image fusion for plant water stress<br />
analysis, Proc <strong>of</strong> 12 th International Conference on <strong>Information</strong> Fusion, 1053-1059, 2009.<br />
32
Impulsive Interference Detection Method Based on Morlet Wavelet <strong>and</strong><br />
Maximum Likelihood Estimation<br />
Yuan Y. He 1 2 * , Chang J. Yu 2 , <strong>and</strong> Tai F. Quan 2<br />
* Presenter<br />
1. Harbin Institute <strong>of</strong> Technology, Harbin, 150001, China<br />
2. The University <strong>of</strong> Melbourne, Parkville, VIC 3010, Australia<br />
The performance <strong>of</strong> the frequency monitor system (FMS) <strong>of</strong> high frequency (HF) radar is<br />
degraded by signal corruption due to impulsive interferences such as lightning <strong>and</strong> meteor<br />
echoes. These interferers raise the FMS spectrum noise level, as a result, FMS can’t pick out<br />
the real minimum disturbance frequency for HF radar. It is desirable to extract the impulsive<br />
interference components from the FMS data without removing any other information.<br />
However, Current wavelet detecting impulsive interference techniques, not matching<br />
impulsive interference very well, always lead to non-impulsive interference components<br />
degraded <strong>and</strong> lost. In this paper, we construct a more suitable impulsive interference detection<br />
method, by using the Morlet wavelet, matching well the shape <strong>of</strong> the impulsive interference<br />
signal, decomposition algorithm <strong>and</strong> maximum likelihood estimation thresholding rule,<br />
specifically designed to detect impulsive interference, so as to less disruption to the nonimpulsive<br />
interference information when extracting the impulsive interference. The<br />
effectiveness <strong>of</strong> the proposed technique has been proved by both simulated <strong>and</strong> practical<br />
experiments.<br />
33
Demonstration <strong>and</strong> Performance Analysis <strong>of</strong> an Uplink based on Digitized<br />
RF-over-Fiber Signal Transport<br />
Yizhuo Yang 1 * , Christina Lim 1 , Prasanna Gamage 1 <strong>and</strong> Ampalavanapillai Nirmalathas 1<br />
* Presenter<br />
1. The University <strong>of</strong> Melbourne, Parkville, VIC 3010, Australia<br />
In this paper, we demonstrate <strong>and</strong> analytically analyze a digitized RF-over-fiber transport<br />
scheme for uplink transmission in a Radio-over-Fiber system. We achieve significant<br />
improvement in both signal-to-noise ratio (SNR) <strong>and</strong> dynamic range. The experimental <strong>and</strong><br />
analytical results show that digitized RF link is less susceptible to nonlinearity <strong>and</strong> fiber<br />
attenuation which makes this scheme well suited for long-reach optical backhaul for wireless<br />
networks.<br />
34
Fiber Nonlinearity Compensation for CO-OFDM Systems with Periodic<br />
Dispersion Maps<br />
Liang B. Y. Du 1 * <strong>and</strong> Arthur J. Lowery 1<br />
* Presenter<br />
1. Electrical <strong>and</strong> Computer Systems Engineering, Monash University, Clayton, VIC 3800,<br />
Australia<br />
We show that the nonlinear limit <strong>of</strong> CO-OFDM systems at 100+ Gbps in periodic<br />
dispersion maps can be increased by 4 dB using nonlinear precompensation. We also show<br />
compensation is beneficial in WDM systems.<br />
35
Effect <strong>of</strong> mini-tyrosl-tRNA synthetase / mini-tryptophanyl-tRNA synthetase<br />
on ischemic angiogenesis in rats with acute myocardial infarction<br />
Rui Zeng 1 * , Yu-cheng Chen 1 , Zhi Zeng 1 , Wei-qiang Liu 1 , Xiao-xia Liu 2 , Rui Liu 3 , Ou<br />
Qiang 3 , Xian Li 3 , Smith AI 4<br />
* Presenter<br />
1. Department <strong>of</strong> Cardiology, West China Hospital, School <strong>of</strong> Clinic Medicine, Sichuan<br />
University, Chengdu ,China .610041<br />
2. Department <strong>of</strong> Epidemiology <strong>and</strong> Health Statistics,School <strong>of</strong> Public Health, Sichuan<br />
University,Chengdu ,China .610041<br />
3. Laboratory <strong>of</strong> Peptides Related with Human Diseases,The National Laboratory <strong>of</strong><br />
Biomedicine, Sichuan University, Chengdu,China .610041<br />
4. Department <strong>of</strong> Biochemistry <strong>and</strong> Molecular biology, Monash university, Melbourne,<br />
Australia.3800<br />
Background: To clarify the mechanism <strong>of</strong> the angiogenesis effect <strong>of</strong> mini-TyrRS/mini-<br />
TrpRS in rodent primates with acute myocardial infarction. Methods: Left coronary artery<br />
ligation was used to establish the model <strong>of</strong> acute myocardial infarction in rats (Sprague-<br />
Dauley male rats,250-300g, 2-3 months old), mini-TyrRS/mini-TrpRS were administered by<br />
coronary artery polyvinyl catheter injection to rats. Rats were r<strong>and</strong>omly divided into four<br />
experimental groups: sham operated group; coronary artery ligation (CAL); CAL+mini-<br />
TyrRS(20μl, twice daily, 600μg. Kg -1 .day -1 );. CAL+mini-TrpRS(20μl, twice daily, 600μg.<br />
Kg -1 .day -1 ). The experiment was carried out at 4 time points as the 3rd, 7th, 14th, <strong>and</strong> 28th<br />
day after ligation. To determine whether mini-TyrRS/mini-TrpRS affected the angiogenesis<br />
activity <strong>of</strong> rats with myocardial infarction, we measured the myocardial infarction size by<br />
TTC staining, <strong>and</strong> the density <strong>of</strong> capillary using immunohistochemistry staining to investigate<br />
the expression <strong>of</strong> VIII factor. Results: Compared with CAL group<br />
(9.3%,15.0%,30.9%,39.3%),the myocardial infarction size <strong>of</strong> mini-TyrRS group at 3rd, 7th,<br />
14th, <strong>and</strong> 28th day were respectively 8.0%,10.9%,22.4%,28.1%, as for mini-TrpRS group,<br />
were 10.7%,18.8%,37.5%,51.2%, but only in day 14th <strong>and</strong> 28th has significant difference,<br />
p
A rodent model <strong>of</strong> acute myocardial infarction(MI) was first developed in the rat[7].More recently<br />
a murine equivalent has been described[8],providing the means to exploit the increasing availability <strong>of</strong><br />
many useful transgenic <strong>and</strong> knockout mouse strains. Complete occlusion <strong>of</strong> the left anterior<br />
descending(LAD)coronary artery induces an acute MI. The resulting ischaemia in the left ventricular<br />
wall has been visualized with Evans blue/TTC perfusion assays[9-12],which allows quantification <strong>of</strong><br />
the infarct area. Further work by Guo et al.[13] has demonstrated ischaemic preconditioning after<br />
short-term LAD occlusion in mouse, thus validating the physiologic relevance <strong>of</strong> this infarct size.<br />
Aminoacyl-tRNA synthetases catalyze the first step <strong>of</strong> protein synthesis that consists <strong>of</strong> the<br />
aminoacylation <strong>of</strong> tRNAs. But they have a broad repertoire <strong>of</strong> functions beyond protein synthesis,<br />
including transcriptional <strong>and</strong> translational regulation as well as cell signaling [14]. Recently, it has<br />
been demonstrated that two <strong>of</strong> the tRNA synthetases, human tyrosyl-tRNA synthetase (TyrRS) <strong>and</strong><br />
human tryptophanyl-tRNA synthetase (TrpRS), have novel cytokine functions[15]. This<br />
demonstration established a link between protein synthesis <strong>and</strong> signal transduction. At the same<br />
time,mammalian TyrRS <strong>and</strong> TrpRS have also been shown to regulate angiogenesis [16-21].<br />
Under apoptotic conditions in culture, full lengthTyrRS is secreted, <strong>and</strong> two distinct cytokines can<br />
then be generated by an extracellular protease such as leukocyte elastase[15].The NH2-terminal<br />
catalytic fragment, mini-TyrRS, binds strongly to the CXC-chemokine receptor CXCR1 <strong>and</strong>, like IL-8,<br />
functions as a chemoattractant for polymorphonuclear leukocytes (PMNs), to promote<br />
angiogenesis[15], whereas the full-length enzyme lacks cytokine activity.The catalytic core domain <strong>of</strong><br />
TrpRS is a close homologue <strong>of</strong> the catalytic domain <strong>of</strong> TyrRS [22-24]. In normal cells, human TrpRS<br />
exists as two forms. The major form is the full-length protein, <strong>and</strong> the other is a truncated TrpRS<br />
(mini-TrpRS) in which most <strong>of</strong> the extra NH2-terminal domain is deleted because <strong>of</strong> alternative<br />
splicing <strong>of</strong> the pre-mRNA[25-26], with Met-48 being deduced as the NH 2 -terminal residue <strong>of</strong> mini-<br />
TrpRS [25]. PMN elastase digestion <strong>of</strong> recombinant full-length TrpRS produced two fragments<br />
designated T 1 -TrpRS <strong>and</strong> T 2 -TrpRS, respectively.These fragments were similar in size to mini-TrpRS.<br />
In addition, mini-TrpRS <strong>and</strong> T2-TrpRS blocked vascular endothelial growth factor-stimulated<br />
angiogenesis in both chick cell adhesion molecule <strong>and</strong> mouse matrigel assays in vivo [18-19]. The<br />
full-length enzyme lacks cytokine activities. The construction <strong>and</strong> relationship <strong>of</strong> TyrRS, TrpRS, <strong>and</strong><br />
their variants is shown in Fig. 1.<br />
Fig 1. Schematic representation <strong>of</strong> human TyrRS, TrpRS <strong>and</strong> their variant constructs used in this study.<br />
Shaded regions <strong>of</strong> full-length TyrRS <strong>and</strong> TrpRS represent COOH- <strong>and</strong> NH 2 -terminal appended<br />
domains, respectively. Numbers on the left <strong>and</strong> right correspond to the NH 2 - <strong>and</strong> COOH-terminal<br />
residues relative to the human full-length enzymes, respectively.<br />
Despite these intriguing in vitro actions, no studies have examined the mini-TyrRS <strong>and</strong> mini-<br />
TrpRS in physiological or pathological settings in vivo, especially in myocardial ischemia condition.<br />
Therefore, the purpose <strong>of</strong> this study was to determine whether exogenous mini-TyrRS augments<br />
angiogenesis while mini-TrpRS inhibited, <strong>and</strong> to preliminary comprehend their angiogenesis<br />
mechanism. All <strong>of</strong> this may be helpful for healing ischemia diseases, such as CAHD.<br />
2. Methods<br />
38
2.1 Materials<br />
Sprague-Dauley(SD) male rats,(250-300g, 2-3 month old) were provided from experimental<br />
animal center <strong>of</strong> Sichuan university.(Sichuan, China). Human recombinant mini-TyrRS <strong>and</strong> mini-<br />
TrpRS were from aTyr Pharma (La Jolla, CA).Ⅷ factor related antigen <strong>and</strong> antibody was from<br />
Boster(Wuhan, china). Immunohistochemisty staining kit was purchased from Zhongshan<br />
Goldenbridge (Beijing, China). Revert Aid First Str<strong>and</strong> cDNA Synthesis Kit was purchased from<br />
MBI Company (Lithuania).<br />
2.2.Ethics<br />
All animal procedures were conducted with prior institutional ethical approval under the<br />
requirements <strong>of</strong> the Chinese Prevention <strong>of</strong> Cruelty to Animals Act <strong>and</strong> the Code <strong>of</strong> Practice for the<br />
Care <strong>and</strong> Use <strong>of</strong> Animals for Scientific Purposes. Prior clearance was obtained from the Animal<br />
Experimentation Ethics Committees <strong>of</strong> West china Medical Centre <strong>and</strong> Institutes <strong>of</strong> Animal Science.<br />
The animals <strong>of</strong> this study were inspected by members <strong>of</strong> the West china Medical Centre Animal Ethics<br />
Committee.<br />
2.3.Left anterior descending (LAD) artery ligation<br />
Animals were anaesthetized with 100g/L chloral hydrate (0.2ml per 100g body weight injected<br />
intraperitoneally).The analgesic buprenorphine was given preoperatively(10-20 µg/kg body<br />
weight).Occlusion <strong>of</strong> the left anterior descending coronary artery(LAD) was performed as previously<br />
reported[11],under sterilec conditions, with minor alterations. Briefly,the anaesthetized animal was<br />
incubated endotracheally in a supine position, <strong>and</strong> ventilated with a Harvard Mouse Mini-<br />
Vent(Harvard apparatus, Marchhugstetten,Germany),which supplied 0.2-0.25ml room air 120 times<br />
per minute. The animal was moved onto its right side, <strong>and</strong> a left thoracotomy in the third intercostal<br />
space was provided access to the beating heart. After removing the pericardium,the LAD was<br />
visualized with a stereomicroscope (Leica MZ6,Heerbrugg,Switzerl<strong>and</strong>),<strong>and</strong> occluded with 8/0<br />
prolene suture. The suture position <strong>of</strong> the LAD coronary artery was 0.3mm distal to the<br />
atrioventricular junction. Occlusion was confirmed by observation <strong>of</strong> left ventricular pallor<br />
immediately post ligation <strong>and</strong> an electrocardiogram was used to observe changes such as widening <strong>of</strong><br />
QRS <strong>and</strong> ST-T segment elevation. The chest was closed, the lungs re-inflate <strong>and</strong> the animal moved to<br />
a prone position until spontaneous breathing occurred. Animal were monitored closely for signs <strong>of</strong><br />
infection at the surgical site; none were observed in any animals.<br />
2.4.Experimental groups<br />
A total <strong>of</strong> 80 rats(200-250g) were divided into four different groups(n=20 per each group, <strong>and</strong><br />
n=5 for each time point).⑴sham group: animals underwent a thoracotomy with removal <strong>of</strong> the<br />
pericardium, but no coronary artery ligation (CAL). No suture was placed in the sham animals’heart,<br />
in order to avoid unintended vessel damage or occlusion; ⑵ CAL group, but no mini-TyrRS/mini-<br />
TrpRS-siRNA injection;⑶: CAL+mini-TyrRS(20μl, twice daily, 600μg. Kg -1 .day -1 ) ⑷: CAL+mini-<br />
TrpRS(20μl, twice daily, 600μg. Kg -1 .day -1 ). Mini-TyrRS or mini-TrpRS were administrated by<br />
coronary artery polyvinyl catheter injection to rats.<br />
2.5.Histologic expression<br />
All rats were killed after ligation 3rd, 7th, 14th, <strong>and</strong> 28th day ,the heart was removed <strong>and</strong> fixed in<br />
fresh 4% paraformaldehyde, pH 7.4.The tissue was processed <strong>and</strong> embedded in paraffin using routine<br />
histological procedures. Five micrometre transverse step sections were collected every 200µm<br />
through the entire ventricle(approximately 10-12 sections per animal),<strong>and</strong> stained with Haematoxylin<br />
<strong>and</strong> Eosin(HE).the cells were observed with an inverted phase-contrast microscope (Olympus, Japan)<br />
<strong>and</strong> photographed. The total tube area were analyzed in 3 different fields at ×400 magnification<br />
2.6.Measurement <strong>of</strong> irreversible ischemic injury<br />
After ligation 3rd, 7th, 14th, <strong>and</strong> 28th day, 1% <strong>of</strong> Evans blue solution (5 mL),was infused into the<br />
abdominal vena cava to delineate the ischemic area at risk <strong>of</strong> the left ventricle. The heart was excised<br />
<strong>and</strong> cross-sectioned from the apex to the atrioventricular groove into four specimens <strong>of</strong> 0.8mm in<br />
thickness with the use <strong>of</strong> a stereoscope. The two middle heart slices were incubated in 2,3,5-<br />
39
triphenyltetrazolium chloride (TTC) solution (1%) for 30 min in phosphate buffer at 37℃. Sections<br />
were fixed overnight in 4% paraformaldehyde for contrast enhancement between stained <strong>and</strong><br />
unstained tissue. TTC stained the viable tissue with red, while the necrotic tissue remained discolored.<br />
The sections were then placed between two cover slips <strong>and</strong> digitally photographed using a Nikon<br />
coolpix S10 camera, <strong>and</strong> quantified with the weight respectively. The area <strong>of</strong> irreversible injury (TTC<br />
negative) is presented as a percentage <strong>of</strong> the area(the irreversible injury area/the total weight <strong>of</strong><br />
ventricles).<br />
2.7.Measured the density <strong>of</strong> capillary<br />
The streptavidin peroxidase (SP) immunohistochemical method was used to detect the expression<br />
<strong>of</strong> factor VIII in myocardial infarction margin areas. The dilution <strong>of</strong> factor VIII rat monoclonal<br />
antibody (Santa Cruz, CA, USA) was 1:100. The procedure was performed according to the<br />
manufacturer’s instructions.The positive cells were identified, counted <strong>and</strong> analyzed under the<br />
inverted phase-contrast microscope (Olympus, Japan) in 3 different fields(0.1mm 2 ) at 400<br />
magnification using Image-proplus 6.0 s<strong>of</strong>tware.Results were normalized by arbitrarily setting the<br />
total tube area <strong>of</strong> control to 100%.<br />
2.8. Real time fluorescent quantitation PCR (RT-PCR)<br />
After ligation, myocardial tissues were harvested at the times indicated, washed twice with icecold<br />
phosphate-buffered saline (PBS) <strong>and</strong> collected by centrifugation. Total RNA was isolated using<br />
the Trizol reagent (MRC, USA) according to the manufacturer's instructions. Total RNA (5 µL) was<br />
converted to complementary DNA (cDNA) using Revert Aid First Str<strong>and</strong> cDNA Synthesis Kit. A<br />
5µL aliquot <strong>of</strong> the resulting cDNA was used as template for PCR amplification with the following<br />
primers: mini-TrpRS, P1 (forward, 5’- CCC TGC TGC ACT CCA CCT T-3’), P2 (reverse, 5’- ACG<br />
CAT GCT TAT TGA CCT TG-3’); mini- TyrRS, P3 (forward, 5’- CAT CTG ATG AAT CCT ATG<br />
GTT -3’), P4 (reverse, 5’- GGA TCA CAA ACT CGG ACT TA-3’); β-actin, P5 (forward, 5’-GCC<br />
AAC ACA GTG CTG TCT -3’), P6 (reverse, 5’-AGG AGC AAT GAT CTT GAT CTT -3’). The<br />
amplifications were performed by an initial denaturation (94 °C for 2 min), followed by 45 cycles <strong>of</strong><br />
denaturation, annealing <strong>and</strong> extension (94 °C for 20 s, 54 °C for 20 s, 72 °C for 30 s), <strong>and</strong> a <strong>final</strong><br />
extension (72 °C for 5 min). The transcript <strong>of</strong> β-actin was also amplified by RT-PCR from the same<br />
cDNA template <strong>and</strong> was used as an internal control. All the primers were designed <strong>and</strong> synthesized by<br />
Genepharma (Shanghai, China). The identity <strong>of</strong> each PCR product was confirmed by DNA sequencing.<br />
The pictures were scanned <strong>and</strong> analyzed by the Gel Doc 1000 gel imaging system.<br />
2.9.Statistical analysis<br />
Results are expressed as mean ± st<strong>and</strong>ard deviation. Comparison <strong>of</strong> means was performed by<br />
means <strong>of</strong> the analysis <strong>of</strong> variance procedure (Student–Newman–keuls test, SPSS 13.0 for Windows).<br />
P
Fig. 2. The representative photomicrographs <strong>of</strong> ne<strong>of</strong>ormative blood capillaries in infarcted LV area,<br />
stained with hematoxylin <strong>and</strong> eosin. Black arrowheads point to newly formed blood vessels <strong>and</strong><br />
ne<strong>of</strong>ormative blood capillary at different time point. Sham, sham operation; CAL, coronary artery<br />
ligation; CAL + mini-TyrRS, mini-TyrRS (20μl, twice daily, 600μg. Kg -1 .day -1 ); CAL + mini-TrpRS,<br />
mini-TrpRS (20μl, twice daily, 600μg. Kg -1 .day -1 ).<br />
3.2.Measurement <strong>of</strong> irreversible ischemic injury(TTC staining)<br />
No myocardial infarction was found in the sham group, , but obvious infarction areas were seen in<br />
other groups. Compared with CAL group (9.3%,15.0%,30.9%,39.3%),the myocardial infarction size<br />
<strong>of</strong> mini-TyrRS group at 3rd, 7th, 14th, <strong>and</strong> 28th day were respectively 8.0%,10.9%,22.4%,28.1%, as<br />
for mini-TrpRS group, were 10.7%,18.8%,37.5%,51.2%, but only in day 14th <strong>and</strong> 28th has significant<br />
difference, p
Fig 3. Mini-TyrRS-siRNA / mini-TrpRS-siRNA transfected TTC stain after myocardial infarction.<br />
Black arrowheads point to myocardial infarction zone. Sham, sham operation; CAL, coronary artery<br />
ligation; CAL + mini-TyrRS, mini-TyrRS (20μl, twice daily, 600μg. Kg -1 .day -1 );CAL + mini-TrpRS,<br />
mini-TrpRS (20μl, twice daily, 600μg. Kg -1 .day -1 ). No myocardial infarction was found in the sham<br />
group, but more obvious infarction areas were seen in other groups. Mini-TrpRS group had the most<br />
myocardial infarction areas, while in mini-TyrRS group were present the least .<br />
Fig 4 Measurement <strong>of</strong> irreversible ischemic injury area <strong>of</strong> mini-TyrRS /mini-TrpRS injected.<br />
Compared with CAL, the myocardial infarction area was decreased in mini-TyrRS group, while<br />
increased in mini-TrpRS group, but only in day 14 <strong>and</strong> 28 considered statistically significant<br />
difference .Data are shown as the means±S.D. (n = 5 for each time point, each group). #P < 0.05, vs.<br />
CAL group. No any infarction size were seen in sham.<br />
3.3.Measured the density <strong>of</strong> capillary(SP method)<br />
The streptavidin peroxidase (SP) immunohistochemical method was used to detect the expression<br />
<strong>of</strong> factor VIII in myocardial infarction areas. The density <strong>of</strong> capillary <strong>of</strong> mini-TyrRS group at 3rd, 7th,<br />
14th, <strong>and</strong> 28th day were respectively 5.5 mount/0.1mm 2 ,8.6 mount/0.1mm 2 , 11.7mount/0.1mm 2 ,13.4<br />
42
mount/0.1mm 2 , as for mini-TrpRS group, were 1.2 mount/0.1mm 2 ,1.6 mount/0.1mm 2 , 1.9<br />
mount/0.1mm 2 ,1.8 mount/0.1mm 2 , has significant difference compared with CAL group (3.0<br />
mount/0.1mm 2 ,5.6 mount/0.1mm 2 , 6.7mount/0.1mm 2 ,5.5 mount/0.1mm 2 ), p
Fig.6 Angiogenesis (capillary density, measured at day 3, 7, 14 <strong>and</strong> 28 after ligation) was inhibited by<br />
mini-TrpRS <strong>and</strong> augmented by mini-TyrRS, while sham had no angiogenesis effect compared with<br />
CAL. Data are shown as the means±S.D. (n = 5 for each time point, each group). #P < 0.05, ##P <<br />
0.01 vs. CAL group.<br />
Fig.7 Mini-TyrRS mRNA expression in infarcted left ventricular tissue in CAL rats at different time<br />
points. (A) mRNA <strong>of</strong> mini-TyrRS at 3 days;(B) mRNA <strong>of</strong> mini-TyrRS at 7 days; (C) mRNA <strong>of</strong> mini-<br />
TyrRS at 14 days; (D) mRNA <strong>of</strong> mini-TyrRS at 28 days. Sham, sham-operation; CAL, coronary artery<br />
ligation; CAL + mini-TyrRS (20μl, twice daily, 600μg. Kg -1 .day -1 ). Lane 1, Sham; Lane 2, CAL;<br />
Lane 3, CAL + mini-TyrRS. Data are shown as the means±S.D. (n = 5 for each time point, each<br />
group). * P < 0.05,**P < 0.01 vs. sham operation group; #P < 0.05, ##P < 0.01 vs. CAL group.<br />
44
3.4.The mRNA expression <strong>of</strong> mini-TyrRS / mini-TrpRS at infarction marginal zone<br />
RT-PCR was used to detect the mRNA expression <strong>of</strong> mini-TyrRS/mini-TrpRS at infarction zone.<br />
The results indicated that: Compared with sham group, the expression level <strong>of</strong> CAL was increase <strong>and</strong><br />
had significantly different in mini-TyrRS group(P < 0.05 or P < 0.01), after administration <strong>of</strong> mini-<br />
TyrRS, there was a significant increase in expression level compared with the CAL group <strong>and</strong> sham<br />
group.(P < 0.05 or P < 0.01).But the expression level between sham-operated <strong>and</strong> CAL groups was<br />
not significantly different in mini-TrpRS group(p>0.05), after administration <strong>of</strong> mini-TrpRS, there was<br />
a significant increase in expression level compared with the CAL group <strong>and</strong> sham group.(P < 0.05 or<br />
P < 0.01).see Fig.7-8.<br />
Fig.8 Mini-TrpRS mRNA expression in infarcted left ventricular tissue in CAL rats at different time<br />
points. (A) mRNA <strong>of</strong> mini- TrpRS at 3 days;(B) mRNA <strong>of</strong> mini- TrpRS at 7 days; (C) mRNA <strong>of</strong> mini-<br />
TrpRS at 14 days; (D) mRNA <strong>of</strong> mini- TrpRS at 28 days. Sham, sham-operation; CAL, coronary<br />
artery ligation; CAL + mini- TrpRS (20μl, twice daily, 600μg. Kg -1 .day -1 ). Lane 1, Sham; Lane 2,<br />
CAL; Lane 3, CAL + mini- TrpRS. Data are shown as the means±S.D. (n = 5 for each time point). * P<br />
< 0.05,**P < 0.01 vs. sham operation group; #P < 0.05, ##P < 0.01 vs. CAL group.<br />
4. Discussion<br />
Preclinical models <strong>of</strong> myocardial ischaemia have been reported in several large animals species,<br />
including dogs[27-28]<strong>and</strong> goats [29].The model that most closely resembles the response seen in<br />
humans is the pig ameroid model[30],which has been used in a variety <strong>of</strong> therapeutic studies[31-<br />
39].However, the expense <strong>and</strong> practical dem<strong>and</strong>s <strong>of</strong> porcine surgical facilities severely limit the extent<br />
<strong>of</strong> such studies,precluding the use <strong>of</strong> this model for large scale screening studies <strong>of</strong> novel therapeutic<br />
approaches.<br />
In normal cells, human TrpRS exists as a full length form <strong>and</strong> as a truncated form designated mini-<br />
TrpRS, which is produced by alternative splicing[40]. Expression <strong>of</strong> mini-TrpRS is highly stimulated<br />
in human cells by the addition <strong>of</strong> IFN-c [41-43]. Although both human full-length TrpRS <strong>and</strong> mini-<br />
TrpRS are enzymatically active in aminoacylation, they differ in angiostatic activity[23; 44]. The same<br />
45
phenomenon is also present between full-length TyrRS <strong>and</strong> mini-TyrRS. Researches have found a<br />
kind <strong>of</strong> angiogenesis regulatory factor, CXC chemotactic factor, but their function was dependent on<br />
whether or not they have the Glu-Leu-Arg (ELR) motif [45]. The CXC chemotactic factors which<br />
have the ELR motif could promote angiogenesis, while CXC chemotactic factors which do not have<br />
the ELR motif would be the antagonistic factors <strong>of</strong> angiogenisis. The situation is analogous to human<br />
mini-TyrRS <strong>and</strong> mini-TrpRS. Mini-TyrRS has an ELR motif <strong>and</strong> promotes angiogenesis, but mini-<br />
TrpRS does not have the ELR motif. Therefore, their function was totally opposite[46].<br />
We were aware that angiogenic <strong>and</strong> angiostatic factors may work together to regulate angiogenesis.<br />
To test our hypothesis that mini-TyrRS (containing a natural ELR sequence) is an angiogenic factor,<br />
<strong>and</strong> that mini-TrpRS(containing a natural ELQ sequence) is an angiostatic factor, we evaluated<br />
whether mini-TyrRS/mini-TrpRS induced myocardial cells to format new capillaries using rats<br />
myocardial infarction model. We found out that irreversible ischemic injury areas was decreased <strong>and</strong><br />
new capillaries formation was increased after mini-TyrRS administration by using TTC staining(Fig.3-<br />
4) <strong>and</strong> the SP immunohistochemical method(Fig.5-6), but in mini-TrpRS group, myocardial infarction<br />
zone was increased <strong>and</strong> new capillaries formation was decreased after administration. In contrast,<br />
angiogenesis was not observed with full-length TyrRS <strong>and</strong> TrpRS. Interestingly, angiogenesis is<br />
stimulated by either mini-TyrRS or is inhibited by mini-TrpRS in rats myocardial infarction models,<br />
raising the possibility that mini-TyrRS/mini-TrpRS stimulates a common downstream signaling event.<br />
Thus, naturally occurring fragments <strong>of</strong> the two proteins involved in translation, TyrRS <strong>and</strong> TrpRS,<br />
have opposing activity on angiogenesis in the rat model. The opposing activities <strong>of</strong> the two tRNA<br />
synthetases suggest tight regulation <strong>of</strong> the balance between pro- <strong>and</strong> antiangiogenic stimuli, but the<br />
exact mechanism <strong>and</strong> relationship between them has not been clearly demonstrated. More research<br />
should be done to further explore the mechanism.<br />
References:<br />
1.Losordo DW, Vale PR, Symes JF, Dunnington CH, Esak<strong>of</strong> DD, Maysky M, Ashare,AB Lathi K,<br />
Isner JM. Gene therapy for myocardial angiogenesis ;initial clinical results with direct myocardial<br />
injection <strong>of</strong> phVEGF165 as sole therapy for myocardial ischemia. Circulation 1998;98:2800-4.<br />
2.Laham RJ, Sellke FW, Edelman JD, Pearlman JD, Ware A, Brown DL, Gold JP , Simons M. Local<br />
perivascular delivery <strong>of</strong> basic fibroblast growth factor in patients undergoing coronary bypass<br />
surgery :results <strong>of</strong> a phase I r<strong>and</strong>omized, double-blind, placebo-controlled trial. Circulation<br />
1999;100:1865-71.<br />
3.Rosengart TK, Lee LY, Patel SR, Kligfield PD, Okin PM, Hackett NR, Isom OW , Crystal RG. Sixmonth<br />
assessment <strong>of</strong> a phase I trial <strong>of</strong> angiogenic gene therapy for the treatment <strong>of</strong> coronary artery<br />
disease using direct intramyocardial administration <strong>of</strong> an adenovirus vector expressing the<br />
VEGF121 cDNA. Ann Surg 1999;230,466-72.<br />
4.Rosengart TK, Lee LY, Patel SR, Sanborn TA, Parikh M, Bergman GW, Hachamovitch R, Szulc M,<br />
Kligfield PD, Okin PM, Hahn RT, Devereux RB, Post MR, Hackett NR, Foster T, Grasso TM,<br />
Lesser ML, Isom OW, Crystal RG. Phase I assessment <strong>of</strong> direct intramyocardial administration <strong>of</strong><br />
an adenovirus vector expressing the VEGF121 cDNA to individuals with clinically significant<br />
severe coronary artery disease. Circulation 1999;100:468-74.<br />
5.Symes JF, Losordo DW, Vale PR, Lathi KG, Esak<strong>of</strong> DD, Maysky M, Isner JM. Gene therapy with<br />
vascular endothelial growth factor for inoperable coronary artery disease. Ann Thoracic Surg<br />
1999;68:830-7.<br />
6.Vale PR, Losordo DW, Milliken CE, McDonald MC, Gravelin LM, Curry CM, Esak<strong>of</strong> DD, Maysky<br />
M, Symes JF ,Isner JM. R<strong>and</strong>omized, singe-blind, placebo-controlled pilot study <strong>of</strong> catheter-based<br />
myocardiol gene transfer for therapeutic angiogenesis using left ventricular electromechanical<br />
mapping in patients with chronic myocardial ischemia. Circulation 2001;103:2138-43.<br />
7.Pfeffer MA, Pfeffer JM, Fishbein MC, Fletcher PJ, Spadoro J, Kloner RA, Braunwald E. Myocardial<br />
infarct size <strong>and</strong> ventricular function in rats. Circulation Res 1979 ;44:503-12.<br />
8.Michael LH, Entman ML, Hartley CJ, Youker KA, Zhu J, Hall SR, Hawkins HK, Berens K,<br />
Ballantyne CM. Myocardial ischemia <strong>and</strong> reperfusion : a murine model. Am J Physiol<br />
1995;269:2147-54.<br />
46
9.Du XJ, Gao X, Ramsey D. Surgical methods <strong>of</strong> inducing transverse aortic stenosis <strong>and</strong> myocardial<br />
infarction in the mouse. Asia Pacific Heart 1998;7:187-92.<br />
10.Lutgens E, Daemen MJAP, de Muinck ED, Debet J, Leenders P, Smits JFM. Chronic myocardial<br />
infarction in the mouse :cardiac structural <strong>and</strong> functional changes. Cardiovascular Res<br />
1999;41:586-93.<br />
11.Michael LH, Ballantyne CM, Zachariah JP, Gould KE, Pocius JS, Taffet GE, Hartley CJ, Pham TT,<br />
Daniel SL, Funk E, Entman ML. Myocardial infarction <strong>and</strong> remodeling in mice. Effect <strong>of</strong><br />
reperfusion. Am J Physiol 1999;277:660-8.<br />
12.Scherrer-Crosbie M, Steudel W, Ullrich R, Hunziker PR, Liel-Cohen N, Newell J, Zar<strong>of</strong>f J, Zapol<br />
WM, Picard MH. Echocardiographic determination <strong>of</strong> risk area size in a murine model <strong>of</strong><br />
myocardial ischemia. Am J Physiol 1999;277:986-92<br />
13.Guo Y, Wu WJ, Qin Y ,Tang XL, Yang Z, Bolli R.Demonstration <strong>of</strong> an early <strong>and</strong> a late phase <strong>of</strong><br />
ischemic preconditioning in mice. Am J Physiol 1998;275:1375-87.<br />
14. Martinis SA, Plateau P, Cavarelli J, Florentz C .AminoacyltRNA synthetases: a family <strong>of</strong><br />
exp<strong>and</strong>ing functions. EMBO J 1999;18:4591–6.<br />
15.Wakasugi K, Schimmel P.Two distinct cytokines released from a human aminoacyl-tRNA<br />
synthetase. Science 1999;284:147–51.<br />
16.Wakasugi K, Schimmel P.Highly differentiated motifs responsible for two cytokine activities <strong>of</strong> a<br />
split human tRNA synthetase. J Biol Chem 1999;274:23155–9.<br />
17.Wakasugi K, Slike BM, Hood J, Ewalt KL, Cheresh DA, Schimmel P. Introduction <strong>of</strong> angiogenesis<br />
by a fragment <strong>of</strong> human tyrosyl-tRNA synthetase. J Biol Chem 2002;277:20124–6.<br />
18.Wakasugi K, Slike BM, Hood J, Otani A, Ewalt KL, Friedl<strong>and</strong>er M, Cheresh DA, Schimmel P.A<br />
human aminoacyl-tRNA synthetase as a regulator <strong>of</strong> angiogenesis. Proc Natl Acad Sci<br />
2002;99:173–7.<br />
19.Otani A, Slike BM, Dorrell M, Hood J, Kinder K, Ewalt KL, Cheresh D, Schimmel P, Friedl<strong>and</strong>er<br />
M.A fragment <strong>of</strong> human TrpRS as a potent antagonist <strong>of</strong> ocular angiogenesis.Proc Natl Acad Sci<br />
USA 2002;99:178–83.<br />
20.Ewalt KL, Schimmel P.Activation <strong>of</strong> angiogenic signaling pathways by two human tRNA<br />
synthetases. Biochemistry 2002;41:13344–9.<br />
21.Ibba M.Aminoacyl-tRNA synthesis. Annu Rev Biochem 2000;69:617–50.<br />
22.Ribas de Pouplana L, Frugier M, Quinn CL, Schimmel P.Evidence that two present-day<br />
components needed for the genetic code appeared after nucleated cells separated from eubacteria.<br />
Proc Natl Acad Sci USA 1996;93:166–70.<br />
23.Lee SW, Cho BH, Park SG, Kim S.Aminoacyl-tRNA synthetase complexes: beyond translation. J<br />
Cell Sci 2004;117:3725–34.<br />
24.Fleckner J, Martensen PM, Tolstrup AB, Kjeldgaard NO ,Justesen, J.Differential regulation <strong>of</strong> the<br />
human interferon inducible tryptophanyl-tRNA synthetase by various cytokines in cell lines.<br />
Cytokine 1995;7:70–7.<br />
25.Kise Y, Lee SW, Park SG, Fukai S, Sengoku T, Ishii R, Yokoyama S, Kim S, Nureki O.A short<br />
peptide insertion crucial for angiostatic activity <strong>of</strong> human tryptophanyl-tRNA synthetase. Nat<br />
Struct Mol Biol 2004;11:149–56.<br />
26.Jorgensen R, Sogaard TM, Rossing AB, Martensen PM, Justesen J.Identification <strong>and</strong><br />
characterization <strong>of</strong> human mitochondrial tryptophanyl-tRNA synthetase. J Biol Chem<br />
2000;275:16820–6.<br />
27.Schaper W, De Bradb<strong>and</strong>er M, Lewi P. DNA synthesis <strong>and</strong> mitoses in coronary collateral vessels<br />
<strong>of</strong> the dog. Circulation Res 1970;28:671-9.<br />
28.Shou M, Thirumurti V,Sharnini Rajanayagam MA, Lazarous DF, Hodge E, Stiber JA, Pettiford M,<br />
Elliot E, Shah SM, Unger EF. Effect <strong>of</strong> basic fibroblast growth factor on myocardial angiogenesis<br />
in dogs with mature collateral vessels.J Am College Cardiogl 1997;29:1102-6.<br />
29.Mannion JD, Blood V, Bailey W ,Bauer TL, Magno MG, Dimeo F,Epple A, Spinale FG.The effect<br />
<strong>of</strong> basic fibroblast growth factor on the blood flow <strong>and</strong> morphologic features <strong>of</strong> a latissimus dorsi<br />
cardiomyoplasty. J Thoracic Cardiovascular Surg 1996;111:19-28.<br />
30.Ware JA, Simons M. Angiogenesis <strong>and</strong> cardiovascular disease. Oxford university press, 1999,New<br />
York.<br />
47
31.Giordano FJ, Ping P, McKirnan D,Nozaki S, DeMaria AN, Dillmann WH, Mathieu CO, Hammond<br />
K. Intracoronary gene transfer <strong>of</strong> fibroblast growth factor-5 increases blood flow <strong>and</strong> contractile<br />
function in an ischemic region <strong>of</strong> the heart. Nature Med 1996;2:534-9.<br />
32.Hariawala MD, Horowitz JR, Esak<strong>of</strong> D, Sheriff DD, Walter DH, Keyt B, Isner J, Symes JF.VEGF<br />
improves myocardial blood flow but produces EDRF-Mediated hypotension in porcine hearts. J<br />
Surg Res 1996;63:77-82.<br />
33.Lopez JJ, Laham RJ, Stamler A, Pearlman JD, Bunting S, Kaplan A, Carrozza JP, Sellke<br />
FW.VEGF administration in chronic myocardial ischemia in pigs. Cardiovascular Res<br />
1998;40:272-81.<br />
34.Mack CA, Patel SR, Schwarz EA, Zanzonico P, Hahn RT, Ilercil A, Devereux RB, Goldsmith SJ,<br />
Christian TF, Sanborn TA, Kovesdi I, Hackett N, Isom OW, Crystal RG, Rosengart TK. Biologic<br />
bypass with the use <strong>of</strong> adenovirus-mediated gene transfer <strong>of</strong> the complementary deoxyribonucleic<br />
acid for vascular endothelial growth factor 121 improves myocardial perfusion <strong>and</strong> function in the<br />
ischemic porcine heart. J Thoracic Cardiovascular Surg 1998;115:168-77.<br />
35.Tio RA, Tkebuchava T, Scheuermann TH, Lebherz C, Magner M, Kearny M, Esak<strong>of</strong> DD, Isner JM,<br />
Symes JF. Intramyocardial gene therapy with naked DNA encoding vascular endothelial growth<br />
factor improves collateral flow to ischemic myocardium. Huamn Gene Ther 1999;10:2953-60.<br />
36.Patel SR, Lee LY, Mack CA,Polce DR, El-Sawy T, Hackett NR, Ilercil A, Jones EC, Hahn RT,<br />
Isom OW, Rosengart TK, Crystal RG.Safety <strong>of</strong> direct myocardial administration <strong>of</strong> an adenovirus<br />
vector encoding vascular endothelial growth factor 121 .Human Gene Ther 1999;10:1331-48.<br />
37.Laham RJ, Rezaee M, Post M, Novicki D, Sellke FW, Pearlman JD, Simons M, Hung<br />
D.Intrapericardial delivery <strong>of</strong> fibroblast growth factor-2 induces neovascularization in a porcine<br />
model <strong>of</strong> chronic myocardial ischaemia. J Pharmacol Exp Therapeutics 2000;292:795-802.<br />
38.Sato K, Laham RJ, Pearlman JD, Novicki D, Sellke FW, Simons M, Post MJ. Efficacy <strong>of</strong><br />
intracoronary versus intravenous FGF-2 in a pig model <strong>of</strong> chronic myocardial ischemia. Ann<br />
Thoracic Surg 2000;70:2113-8.<br />
39.Sato K, Wu T, Laham RJ, Johnson RB, Douglas P, Li J, Sellke FW, Bunting S, Simons M, Post MJ.<br />
Efficacy <strong>of</strong> intracoronary or intravenous VEGF in a pig model <strong>of</strong> chronic myocardial ischemia. J<br />
Am College Cardiol 2001;37:616-23.<br />
40.Yeh CH, Peng HC, Huang TF.Accutin, a new disintegrin,inhibits angiogenesis in vitro <strong>and</strong> in vivo<br />
by acting as integrin alphavbeta-antagonist <strong>and</strong> inducing apoptosisl.J Blood 1998;92:3268–76<br />
41.Tolstrup AB, Bejder A, Fleckner J, Justesen J.Transcriptional regulation <strong>of</strong> the interferon-c<br />
inducible tryptophanyltRNA synthetase includes alternative splicing. J Biol Chem 1995;270:397–<br />
403.<br />
42.Fleckner J, Rasmussen HH, Justesen J.Human interferon gamma potently induces the synthesis <strong>of</strong><br />
a 552 kDa protein highly homologous to rabbit peptide chain release factor <strong>and</strong> bovine tryptophan.<br />
Proc Natl Acad Sci 1991;88:11520–4.<br />
43.Shaw AC, Rossel LM, Roepstorff P, Justesen J, Christiansen G, Birkelund S. Mapping <strong>and</strong><br />
identification <strong>of</strong> interferon gamma-regulated HeLa cell proteins separated by immobilized pH<br />
gradient two dimensional gel electrophoresis. Electrophoresis 1999;20:984–93.<br />
44.Tzima E, Reader JS, Irani-Tehrani M, Ewalt KL, Schwartz MA, Schimmel P. VE-cadherin links<br />
tRNA synthetase cytokine to anti-angiogenic function. J Biol Chem 2005;280:2405–8<br />
45. Clark LI, Dewald B, Geiser T ,Moser B, Baggiolini M.Platelet factor 4 binds to interleukin 8<br />
receptors <strong>and</strong> activates neutrophils when its N terminus is modified with Glu-Leu-Arg. Proc Natl<br />
Acad Sci 1993;90:3574–7.<br />
46.Jeong EJ, Hwang GS, Kim KH, Kim MJ, Kim S, Kim KS.Structural analysis <strong>of</strong> multi-functional<br />
peptide motifs present in human bifunctional tRNA synthetase: identification <strong>of</strong> RNA binding<br />
residues <strong>and</strong> functional implications for t<strong>and</strong>em repeats. Biochemistry 2000;39:15775–82.<br />
48
Distribution <strong>of</strong> three forms alpha-melanocyte stimulating hormone in the<br />
brain <strong>of</strong> male <strong>and</strong> female sheep by using HPLC <strong>and</strong> Mass Spectrometry<br />
Analysis<br />
Rui Zeng 1,2 * , A.Ian.Smith 2 , Iain.J.Clarke 3<br />
* Presenter<br />
1. Department <strong>of</strong> Cardiovascular diseases, West China Hospital, School <strong>of</strong> Clinic Medicine,<br />
Sichuan University, Chengdu 610041, China<br />
2. Department <strong>of</strong> Biochemistry & Molecular Biology, <strong>Faculty</strong> <strong>of</strong> Medicine, Nursing <strong>and</strong><br />
Health Sciences, Monash University, Vic 3800,Australia<br />
3. Department <strong>of</strong> Physiology, <strong>Faculty</strong> <strong>of</strong> Medicine, Nursing <strong>and</strong> Health Sciences, Monash<br />
University, Vic 3800, Australia<br />
To compare des-acetyl α-MSH(des-α-MSH), acetyl α-MSH(act-α-MSH) <strong>and</strong> di-acetyl α-<br />
MSH(di-act-α-MSH) between the male <strong>and</strong> female sheep in ten special brain tissues. All <strong>of</strong><br />
this will let us ascertain whether different gonadal steroids causes any change in acetylated<br />
peptide. Mass spectrometry (MS) was used for molecular weight determination, HPLC/EIA<br />
was also used for quantitation. Before MS, Immunomagnetic protein isolation using<br />
Dynabeads Protein A also provides a fast <strong>and</strong> reliable method for capturing Ig for small scale<br />
purification. Mass spectrometry analysis indicated that: In female sheep, all three forms <strong>of</strong> α-<br />
MSH were detected in hypothalamus, neurointermediate lobe, cortex, pre-optic area <strong>and</strong><br />
median eminence. In arcuate nucleus, des- <strong>and</strong> act-α-MSH were present, <strong>and</strong> only acetylated<br />
forms(including mono-act <strong>and</strong> di-act) were detected in anterior pituitary. For male sheep, only<br />
in neurointermediate lobe <strong>and</strong> pre-optic area were present all three forms <strong>of</strong> α-MSH. In<br />
hypothalamus, median eminence, cortex, anterior pituitary, des- <strong>and</strong> act-α-MSH were detected.<br />
We can only find des-α-MSH in arcuate nucleus. Cerebellum, spinal cord <strong>and</strong> brain stem have<br />
no any forms <strong>of</strong> α-MSH both in male <strong>and</strong> female sheep. HPLC/EIA results showed that in<br />
hypothalamus, neurointermediate lobe, pre-optic area <strong>and</strong> median eminence, there is no any<br />
difference in des-α-MSH between male <strong>and</strong> female sheep, but female sheep has more act-α-<br />
MSH <strong>and</strong> di-act-α-MSH compared with male sheep.P < 0.05. In arcuate nucleus, there is no<br />
any difference in des-α-MSH between male <strong>and</strong> female sheep, but female sheep has more actα-MSH<br />
only compared with male sheep.P < 0.05. In anterior pituitary, there is more des-α-<br />
MSH in male sheep, but female sheep has more act-α-MSH <strong>and</strong> di-act-α-MSH compared with<br />
male sheep. P < 0.05. Finally, in cerebral cortex, female sheep has more des-α-MSH, act-α-<br />
MSH <strong>and</strong> di-act-α-MSH compared with male sheep.P < 0.05. These data suggest that N-<br />
acetylation <strong>of</strong> products formed by the processing <strong>of</strong> α-MSH can markedly adapt to their<br />
biological properties. The results also illustrated that the distribution <strong>of</strong> α-MSH may be<br />
affected by the sex steroid hormone.<br />
Keywords<br />
α-MSH, POMC, sheep, HPLC, Mass Spectrometry<br />
Introduction<br />
Proopiomelanocortin (POMC) is a precursor protein that contains the sequence for several<br />
bioactive peptides including adrenocorticotropin (ACTH), β-endorphin(β-EP), <strong>and</strong> melanocytestimulating<br />
hormone (α-, β- <strong>and</strong> γ- MSH). The functional peptides have been detected in a variety <strong>of</strong><br />
tissues.<br />
Post-translational processing <strong>of</strong> prohormones is an important biochemical mechanism employed<br />
by many secretory cells to generate a diversity <strong>of</strong> biological peptides from the same gene product, a<br />
process that occurs in a tissue-specific manner (1). Similarly to many prohormones, POMC is<br />
49
subjected to sequence-specific cleavages to generate bioactive products by the action <strong>of</strong> the<br />
prohormone convertases (PCs) 1 <strong>and</strong> 2 (2). Thus far, the processing <strong>of</strong> POMC has been described in<br />
the pituitary <strong>and</strong> in the arcuate nucleus (ARC) <strong>of</strong> the hypothalamus. In the ARC, similar to the pars<br />
intermedia <strong>of</strong> the pituitary (3-6), POMC is initially cleaved by PC1 to generate proadrenocorticotrophin<br />
(pro-ACTH) <strong>and</strong> β-lipotrophin. Pro-ACTH is further cleaved by PC1 to generate<br />
a 16 kDa N-terminal peptide <strong>and</strong> ACTH. ACTH is further cleave to generate ACTH (1-17) <strong>and</strong><br />
corticotrophin like intermediate lobe peptide (CLIP). Then, carboxypeptidase E enzyme removes C-<br />
terminal basic amino acids from ACTH (1-17) , <strong>and</strong> the peptidyl α-amidating mono-oxigenase (PAM)<br />
enzyme amidates the peptide to generate des-acetyl-α-melanocyte stimulating hormone (des-α-MSH).<br />
Finally, an N-acetyl-tranferase enzyme converts des-α-MSH to acetyl-α-MSH (act-α-MSH) <strong>and</strong><br />
double- acetyl-α-MSH (di-act-α-MSH)(7-9). POMC (10,11) is also generated in the commissural part<br />
<strong>of</strong> the nucleus <strong>of</strong> the solitary tract (NTS). The amount <strong>of</strong> act-α-MSH in the rodent brain is<br />
controversial, because some studies report undetectable levels <strong>of</strong> the N-acetylated form in the<br />
hypothalamus (12-14). On the other h<strong>and</strong>, it is well established that synthetic act-α-MSH is far more<br />
potent than des-α-MSH in reducing food intake when administered centrally in rats (15,16). However,<br />
the fate <strong>of</strong> its processing <strong>and</strong> mechanism underlying the different potencies <strong>of</strong> the two forms <strong>of</strong> α-<br />
MSH is poorly understood.<br />
The amino acid sequence is the foundation <strong>of</strong> a protein, <strong>and</strong> any degradation or modification <strong>of</strong> a<br />
protein’s primary structure may affect its biological properties <strong>and</strong> functions. Numerous conventional<br />
analytical techniques have been applied to the determination <strong>of</strong> the primary structure <strong>of</strong> therapeutic<br />
proteins.To investigate post-translational processing, we needs to know each product including<br />
acetylated products. Usually methods is high-performance liquid chromatography (HPLC) was used<br />
for protein separation <strong>and</strong> purity analysis first, <strong>and</strong> then enzyme immunoassay(EIA) for quantitation,<br />
but the disadvantage <strong>of</strong> this method was that it would cost at least two days to acquire the results.<br />
Another choice was mass spectrometry (MS), using by molecular weight determination, for its fast<br />
(only several seconds you can obtain all the results), accurate <strong>and</strong> reliable, except that it could not be<br />
used for quantitation. The amino acid sequence information <strong>and</strong> molecular weight for the three α-MSH<br />
were seen in Table.1.<br />
Table 1 Amino acid sequence information <strong>and</strong> molecular weight <strong>of</strong> three α-MSH<br />
Peptide<br />
MW<br />
Amino acids<br />
Des - Ac - α-MSH (sheep) 1622.4 13<br />
Sequence(One-Letter Code): SYSMEHFRWGKPV-NH2<br />
Sequence(Three-Letter Code): H - Ser - Tyr - Ser - Met - Glu - His - Phe - Arg -<br />
Trp - Gly - Lys - Pro - Val - NH2<br />
Ac -α-MSH (sheep) 1664.9 13<br />
Sequence(One-Letter Code): Ac-SYSMEHFRWGKPV-NH2<br />
Sequence(Three-Letter Code): Ac - Ser - Tyr - Ser - Met - Glu - His - Phe - Arg -<br />
Trp - Gly - Lys - Pro - Val - NH2<br />
Di - Ac-α-MSH (sheep) 1706.8 13<br />
Sequence(One-Letter Code): Di-ac-SYSMEHFRWGKPV-NH2<br />
Sequence(Three-Letter Code): Di-ac - Ser - Tyr - Ser - Met - Glu - His - Phe -<br />
Arg - Trp - Gly - Lys - Pro - Val - NH2<br />
The purpose <strong>of</strong> this study was therefore threefold: (i)to determine levels <strong>of</strong> three forms α-MSH in<br />
10 different brain tissues extracts;(ii)to compare the different type <strong>of</strong> α-MSH between the male <strong>and</strong><br />
female sheep;(iii) to compare the results determined by HPLC/EIA <strong>and</strong> mass spectrometry analysis.<br />
All <strong>of</strong> this will let us ascertain whether different gonadal steroids causes any change in acetylated<br />
peptide.<br />
50
Materials <strong>and</strong> Methods<br />
Ethics<br />
All animal procedures were conducted with prior institutional ethical approval under the requirements<br />
<strong>of</strong> the Australian Prevention <strong>of</strong> Cruelty to Animals Act 1986 <strong>and</strong> the Code <strong>of</strong> Practice for the Care <strong>and</strong><br />
Use <strong>of</strong> Animals for Scientific Purposes. Prior clearance was obtained from the Animal<br />
Experimentation Ethics Committees <strong>of</strong> Monash Medical Centre <strong>and</strong> Victorian Institutes <strong>of</strong> Animal<br />
Science. The animals <strong>of</strong> this study were inspected by members <strong>of</strong> the Monash Medical Centre Animal<br />
Ethics Committee.<br />
Animals<br />
Adult Corriedale male sheep(55kg,n=3) <strong>and</strong> female sheep(53kg, n=3) were maintained on pasture or<br />
in feedlots under natural conditions. For intensive experimentation, the animals were housed in<br />
individual pens with natural lighting <strong>and</strong> temperature <strong>and</strong> were fed ad libitum lucerne chaff with free<br />
access to water. Before experiment, animals were conditioned to pen-housing <strong>and</strong> h<strong>and</strong>ling for 1 week.<br />
Animals in this experiment were euthanised by overdose <strong>of</strong> 20 ml sodium pentobarbital (Lethabarb;<br />
Virbac, Peakhurst, NSW, Australia) iv <strong>and</strong> the brains were removed, then Hypothalamus , Anterior<br />
pituitary gl<strong>and</strong> (A.P), Cerebellum (CEREB), Arcuate nucleus (ARC), Neurointermediate lobe (NIL),<br />
Median eminence (M.E), Cerebral cortex, Spinal cord, Brain stem(B/S), <strong>and</strong> Pre-optic area(POA)<br />
were dissected. All tissues were frozen on dry ice <strong>and</strong> stored at -80°C.<br />
Peptide Extraction<br />
Tissues were immediately immersed in 2N acetic acid (each 600 μl for 300mg sample) <strong>and</strong> boiled for<br />
10 min. The samples were then homogenized by using electric homogenisers. Vortex, st<strong>and</strong> on ice<br />
while homogenizing the rest <strong>of</strong> the samples, <strong>and</strong> once completed leave on ice for at least half an hour.<br />
Finally, centrifuged at 6,000rpm at 4°C for 30 min. Supernatants were collected, <strong>and</strong> re-constituted in<br />
distilled water.<br />
Experiment 1: Mass spectrometry analysis <strong>of</strong> α-MSH in ten brain tissues<br />
Co-immunoprecipitation (Peptide purification)<br />
Polyclonal α-MSH-antibody (1:1000; Bachem, Bubendorf, Switzerl<strong>and</strong>) was cross-linked to<br />
Dynabeads protein A (Dynal Biotech, Olso, Norway) according to the manufacturer’s protocol.<br />
Tissues lysates were precleared with IgG Dynabeads-protein A for 10 min <strong>and</strong> then incubated with α-<br />
MSH-Dynabeads overnight at 4 °C, then α-MSH-immunoprecipitated complexes were washed five<br />
times with immunoprecipitation buffer (10mM Tris/HCl, pH 7.8,1mM EDTA,150mM NaCl,1mM<br />
NaF,0.5%NonidetP-40,0.5% glucopyranoside, 1μg/ml aprotinin, <strong>and</strong> 0.5 mM phenylmethylsul-fonyl<br />
fluoride).Proteins were eluted by 0.1M citrate buffer (elution buffer, pH2-3 ) <strong>and</strong> then processed by<br />
Mass Spectrometry(ABI 4700 Proteomics Analyzer ,Applied Biosystems, USA).<br />
Peptide samples for mass spectrometry analysis<br />
Peptide samples were purified <strong>and</strong> concentrated using a Zip Tip (Millipore) which has C18 resin fixed<br />
at its end. The resin was rinsed according to the manufacturer’s instructions with 10 μl 0.1% TFA <strong>and</strong><br />
50% ACN. Peptides were eluted in 10 μl 50% ACN in 0.1 % TFA. A 0.5 μl volume <strong>of</strong> the<br />
concentrated peptide-containing sample was mixed with a saturated solution <strong>of</strong> α-cyano-4-<br />
hydroxycinnamic acid (0.5 μl). Each sample (0.5 μl) was spotted on the mass spectrometer sample<br />
plate . The mass spectrometer determines the mass <strong>of</strong> the peptides <strong>and</strong> the sequence (by collisioninduced<br />
dissociation). From the masses <strong>of</strong> the peptide fragments, sequence data were determined by<br />
comparison with known sequences.<br />
Experiment 2: HPLC/Enzyme immunoassay <strong>of</strong> α-MSH in ten brain tissues<br />
α-MSH Separation<br />
α-MSH peptides were separated by HPLC using a Symmetry C18 (5μm) column (4.6 × 150mm)<br />
(Waters, Milford, USA). The chromatographic run was performed with an aqueous phase containing<br />
0.1% trifluoroacetic acid (TFA) <strong>and</strong> an organic phase containing 0.085% TFA in acetonitrile(ACN):<br />
methanol (80:20) gradient at a flow rate <strong>of</strong> 1.0 ml/min. The gradient used for separation was run 52<br />
mins; the total run time was 72 mins .α-MSH peptides eluted between 25-30% acetonitrile: methanol.<br />
51
Fractions (1 ml) were dried <strong>and</strong> rehydrated with 50μl <strong>of</strong> EIA buffer (1.42g/l Na 2 HPO 4 , 0.2g/l KH 2 PO 4 ,<br />
5g/l BSA, 1ml/l Tween 20 in water, pH 7.4) <strong>and</strong> assayed to identify the different forms <strong>of</strong> α-MSH<br />
using Enzyme immunoassay (EIA).<br />
Enzyme immunoassay<br />
The EIA plates were first coated with Sheep anti-αMSH (1:10000; Bachem, Bubendorf, Switzerl<strong>and</strong>)<br />
in EIA coating buffer (4.3g NaHCO 3 , 5.3g Na 2 CO 3 , in 1L water, pH 9.4) over night at 4°C. The plate<br />
was then washed in EIA wash buffer (1.42g Na 2 HPO 4 ,0.2g KH 2 PO 4 , 1ml Tween 20 in 1L water,<br />
pH7.4) <strong>and</strong> blocked (8g NaCl, 1.42g KH 2 PO 4 , 0.2g/KCl, 5.0g BSA in 1L water, pH 7.4) for 1h at<br />
room temperature. After 3 washes, St<strong>and</strong>ards (acetylated αMSH , 0.5 – 1024pg/50μl; Bachem,<br />
Torrance, CA)<strong>and</strong> samples were added at 50μl/well for 90mins at room temperature. After removing<br />
the supernatant, 50μl <strong>of</strong> biotinylated αMSH (1:100000, Bachem, Torrance, CA) was added to all wells<br />
except blank <strong>and</strong> non specific binding wells for 90mins. The plate was washed 3 times followed by<br />
50μl <strong>of</strong> streptavidin-HRP (BD Pharmingen, Franklin Lakes, NJ) for 1h. Finally, after 3 washes, 50μl<br />
tetramethylbenzidine (TMB) (Pierce, Rockford, IL) was added to each well until a strong blue reaction<br />
developed in the total binding wells (30 mins). The reaction was terminated by the addition <strong>of</strong> 50μl 2N<br />
H 2 SO 4 <strong>and</strong> the plates were read at 450nm. The sensitivity <strong>of</strong> this assay was 4pg/100μl.<br />
Statistical analysis<br />
Results are expressed as mean ± st<strong>and</strong>ard deviation. Comparison <strong>of</strong> means was performed by means <strong>of</strong><br />
the analysis <strong>of</strong> variance procedure (Student–Newman–keuls test, SPSS 13.0 for Windows). P
Fig.1 (A) Mass Spectrometry results for male <strong>and</strong> female sheep in hypothalamus. All three forms <strong>of</strong> α-<br />
MSH were detected in female sheep, but only des- <strong>and</strong> act-α-MSH in male sheep. 1622 for des-α-<br />
MSH ; 1664 for act-α-MSH; 1706 for di-act-α-MSH .(B) HPLC/EIA results indicated that there is no<br />
any difference in des-α-MSH between male <strong>and</strong> female sheep, but female sheep has more act-α-MSH<br />
<strong>and</strong> di-act-α-MSH compared with male sheep.* P < 0.05, vs. male sheep group.<br />
53
Fig.2 (A) Mass Spectrometry results for male <strong>and</strong> female sheep in arcuate nucleus. Des- <strong>and</strong> act-α-<br />
MSH were detected in female sheep, but only des-α-MSH in male sheep. 1622 for des-α-MSH; 1664<br />
for act-α-MSH. (B) HPLC/EIA results indicated that there is no any difference in des-α-MSH between<br />
male <strong>and</strong> female sheep, but female sheep has more act-α-MSH compared with male sheep.* P < 0.05,<br />
vs. male sheep group.<br />
Fig.3 (A) Mass Spectrometry results for male <strong>and</strong> female sheep in anterior pituitary. Only acetylated<br />
forms(including mono-act <strong>and</strong> di-act) were detected in female sheep <strong>and</strong> des- <strong>and</strong> act-α-MSH in male<br />
sheep. 1622 for des-α-MSH ; 1664 for act-α-MSH; 1706 for di-act-α-MSH . (B) HPLC/EIA results<br />
54
indicated that there is more des-α-MSH in male sheep, but female sheep has more act-α-MSH <strong>and</strong> diact-α-MSH<br />
compared with male sheep.* P < 0.05, vs. male sheep group.<br />
Fig.4 (A) Mass Spectrometry results for male <strong>and</strong> female sheep in neurointermediate lobe. All three<br />
forms <strong>of</strong> α-MSH were detected in both female <strong>and</strong> male sheep. 1622 for des-α-MSH ; 1664 for act-α-<br />
MSH; 1706 for di-act-α-MSH. (B) HPLC/EIA results indicated that there is no any difference in desα-MSH<br />
between male <strong>and</strong> female sheep, but female sheep has more act-α-MSH <strong>and</strong> di-act-α-MSH<br />
compared with male sheep.* P < 0.05, vs. male sheep group.<br />
55
Fig.5 (A) Mass Spectrometry results for male <strong>and</strong> female sheep in cerebral cortex. All three forms <strong>of</strong><br />
α-MSH were detected in female sheep, but only des- <strong>and</strong> act-α-MSH in male sheep. 1622 for des-α-<br />
MSH ; 1664 for act-α-MSH; 1706 for di-act-α-MSH . (B) HPLC/EIA results indicated that female<br />
sheep has more des-α-MSH, act-α-MSH <strong>and</strong> di-act-α-MSH compared with male sheep.* P < 0.05, vs.<br />
male sheep group.<br />
56
Fig.6 (A) Mass Spectrometry results for male <strong>and</strong> female sheep in pre-optic area. All three forms <strong>of</strong> α-<br />
MSH were detected in both female <strong>and</strong> male sheep. 1622 for des-α-MSH ; 1664 for act-α-MSH; 1706<br />
for di-act-α-MSH. (B) HPLC/EIA results indicated that there is no any difference in des-α-MSH<br />
between male <strong>and</strong> female sheep, but female sheep has more act-α-MSH <strong>and</strong> di-act-α-MSH compared<br />
with male sheep.* P < 0.05, vs. male sheep group.<br />
Fig.7 (A) Mass Spectrometry results for male <strong>and</strong> female sheep in median eminence. All three forms<br />
<strong>of</strong> α-MSH were detected in female sheep, but only des- <strong>and</strong> act-α-MSH in male sheep. 1622 for des-α-<br />
MSH ; 1664 for act-α-MSH; 1706 for di-act-α-MSH. (B) HPLC/EIA results indicated that there is no<br />
any difference in des-α-MSH between male <strong>and</strong> female sheep, but female sheep has more act-α-MSH<br />
<strong>and</strong> di-act-α-MSH compared with male sheep.* P < 0.05, vs. male sheep group.<br />
57
Table 2 The distribution <strong>of</strong> three forms α-MSH in brain tissues by Mass spectrometry analysis<br />
Male<br />
Female<br />
des- mono-act di-act des- mono-act di-act<br />
Hypothalamus + + — + + +<br />
Anterior pituitary + + — — + +<br />
Arcuate nucleus + — — + + —<br />
Neurointermediate lobe + + + + + +<br />
Median eminence + + — + + +<br />
Cerebral cortex + + — + + +<br />
Pre-optic area + + + + + +<br />
Spinal cord — — — — — —<br />
Brain stem — — — — — —<br />
Cerebellum — — — — — —<br />
Discussion<br />
Some researches indicated that no authentic α-MSH was detected in extracts <strong>of</strong> rat brain regions<br />
or human pituitary. In these tissues the only α-MSH variant detected was characterized as des-acetyl<br />
α-MSH. Conversely, acetylated α-MSH was found to be the predominant form in rat pituitary(12). It is<br />
the same result <strong>of</strong> female sheep in our research, Only acetylated forms(including mono-act <strong>and</strong> di-act)<br />
were detected, but except acetylated α-MSH ,we can still find des-acetyl α-MSH in male sheep.<br />
The role <strong>of</strong> the N-acetyl group has been extensively studied in conjunction with<br />
melanocortinergic regulation <strong>of</strong> food intake. Peripheral injection <strong>of</strong> α-MSH in viable yellow obese<br />
mice resulted in significantly lower food intake <strong>and</strong> body weight gain than in mice injected with the<br />
same doses <strong>of</strong> desacetyl-α-MSH (17). Tsujii <strong>and</strong> Bray (16) injected acetyl-α-MSH <strong>and</strong> des-acetyl-α-<br />
MSH intracerebroventricularly (i.c.v.) into food-deprived rats <strong>and</strong> found that acetyl-α-MSH produced<br />
a highly significant reduction <strong>of</strong> food intake at doses <strong>of</strong> 100 <strong>and</strong> 250 pmol whereas des-acetyl-α-MSH<br />
had no effect at any dose between 100 <strong>and</strong> 2500 pmol. Similarly, after i.c.v. administration to fasted<br />
rats, doses <strong>of</strong> 1, 3, or 6 nmol acetyl-α-MSH resulted in decreased food intake (18). Des-acetyl-α-MSH<br />
did not produce a significant effect at these doses but reduced food intake at a dose <strong>of</strong> 25 nmol. The<br />
most recent demonstration <strong>of</strong> differential effects <strong>of</strong> MSH acetylation on the regulation <strong>of</strong> food intake<br />
has been provided by Guo et al.2004(19), who again demonstrated that i.c.v. injection <strong>of</strong> 3 nmol<br />
acetyl-α-MSH, but not des-acetyl-α-MSH, resulted in inhibition <strong>of</strong> food intake in rats.<br />
Differential effects <strong>of</strong> acetylated <strong>and</strong> non-acetylated α-MSH have also been noted in the<br />
adrenal cortex, adipose tissue, <strong>and</strong> bone. In addition to ACTH or melanocortin-2 receptors<br />
(MC2R),MC3R <strong>and</strong> MC5R are expressed in the adrenal cortex (20,21).Both acetyl-α-MSH<br />
<strong>and</strong> des-acetyl-a-MSH stimulate aldosterone, corticosterone, <strong>and</strong> cortisol secretion from<br />
human adrenocortical cells in vitro (22). Des-acetyl-α-MSH is much more potent than acetylα-MSH,<br />
but both are far less effective than ACTH (1–24). ACTH (1–24) elicited significant<br />
secretion <strong>of</strong> all three steroids at doses <strong>of</strong> 1.0 pmol whereas the threshold dose was 1.0 nmol<br />
for des-acetyl-α-MSH <strong>and</strong> 100 nmol for acetyl-α-MSH. The relative potencies <strong>of</strong> ACTH <strong>and</strong><br />
α-MSH agree with the 104-fold difference in potency between the two peptides in rat adrenal<br />
(23). However, others have failed to find any evidence for steroidogenic activity or MC2R<br />
binding <strong>of</strong> α-MSH in rat or mouse adrenal (24,25). Des-acetyl-, di-acetyl- <strong>and</strong> mono-acetyl-α-<br />
MSH have been demonstrated to have lipolytic activity in rabbits both in vivo <strong>and</strong> in vitro,<br />
but none <strong>of</strong> the forms is effective in rats (24). The threshold dose <strong>of</strong> des-acetyl-α-MSH to<br />
increase plasma free fatty acid (FFA) concentration in rabbits was 1 mg/kg, but for the two<br />
acetylated forms was 0.33 mg/kg. As determined by peak levels <strong>of</strong> FFA, the relative potencies<br />
<strong>of</strong> diacetyl-/monoacetyl-/desacetyl-α-MSH were 8.3/5.3/1. The effects <strong>of</strong> the three peptides<br />
on lipolytic activity in rabbit adipose tissue slices followed similar patterns except that the<br />
threshold doses <strong>and</strong> potencies <strong>of</strong> acetyl-α-MSH <strong>and</strong> di-acetyl-α-MSH were equivalent. MC2R,<br />
MC4R, <strong>and</strong> MC5R have now all been identified in adipose tissue by quantitative reverse<br />
58
transcription-polymerase chain reaction analysis (RT-PCR) (26,27). . These data suggest that<br />
N-acetylation <strong>of</strong> products formed by the processing <strong>of</strong> α-MSH can markedly alter their<br />
biological properties.<br />
In our research, we detected the different N-acetylation forms <strong>of</strong> α-MSH in ten brain<br />
tissues <strong>of</strong> female <strong>and</strong> male sheep. The results illustrated that the distribution <strong>of</strong> α-MSH may<br />
be affected by the sex steroid hormone, <strong>and</strong> further researches should be done in the future.<br />
Acknowledgments<br />
This work were finished at Department <strong>of</strong> Biochemistry & Molecular Biology <strong>and</strong> Department <strong>of</strong><br />
Physiology, <strong>Faculty</strong> <strong>of</strong> Medicine, Nursing <strong>and</strong> Health Sciences, Monash University .The authors are<br />
grateful to Dr. David Steer(Bio) <strong>and</strong> Josie Lawrence(Bio) for their valuable discussion on MS analysis;<br />
Iresha Hanchapola (Bio) ,Melanie Clarke(Physi) <strong>and</strong> Alex<strong>and</strong>ra Rao (Physi) for their help <strong>of</strong><br />
HPLC/EIA.<br />
References:<br />
1. Nillni EA ,Sevarino KA. The biology <strong>of</strong> pro-thyrotropin-releasing hormonederived peptides.<br />
Endocr Rev 1999; 20: 599-648.<br />
2. Benjannet S, Rondeau N, Day R, Chretien M, Seidah NG. PC1 <strong>and</strong> PC2 are proprotein convertases<br />
capable <strong>of</strong> cleaving proopiomelanocortin at distinct pairs <strong>of</strong> basic residues. Proc Natl Acad Sci<br />
USA 1991; 88: 3564-3568<br />
3. Barnea A, Cho G, Porter JC. A reduction in the concentration <strong>of</strong> immunoreactive corticotropin,<br />
melanotropin <strong>and</strong> lipotropin in the brain <strong>of</strong> the aging rat. Brain Res 1982;232: 345-353.<br />
4. Emeson RB , Eipper BA. Characterization <strong>of</strong> pro-ACTH/endorphin-derived peptides in rat<br />
hypothalamus. J Neurosci 1986; 6: 837-849.<br />
5. Gramsch C, Kleber G, Hollt V, Pasi A, Mehraein P, Herz A. Proopiocortin fragments in human <strong>and</strong><br />
rat brain: beta-endorphin <strong>and</strong> alpha-MSH are the predominant peptides. Brain Res 1980;192: 109-<br />
119.<br />
6. Orwoll E, Kendall JW, Lamorena L, McGilvra R. Adrenocorticotropin <strong>and</strong> melanocyte-stimulating<br />
hormone in the brain. Endocrinology 1979;104: 1845-1852.<br />
7. Cone RD, Lu D, Koppula S, Vage DI, Klungl<strong>and</strong> H, Boston B, Chen W, Orth DN, Pouton C,<br />
Kesterson RA. The melanocortin receptors: agonists, antagonists, <strong>and</strong> the hormonal control <strong>of</strong><br />
pigmentation. Recent Prog Horm Res 1996;51: 287-317.<br />
8. Pritchard LE, Turnbull AV, White A. Pro-opiomelanocortin processing in the hypothalamus: impact<br />
on melanocortin signalling <strong>and</strong> obesity. J Endocrinol 2002; 172: 411-421.<br />
9. Wilkinson CW. Roles <strong>of</strong> acetylation <strong>and</strong> other post-translational modifications in melanocortin<br />
function <strong>and</strong> interactions with endorphins. Peptides 2006; 27: 453-471.<br />
10. Bronstein DM, Schafer MK, Watson SJ, Akil H. Evidence that beta endorphin is synthesized in<br />
cells in the nucleus tractus solitarius: detection <strong>of</strong> POMC mRNA. Brain Res 1992; 587: 269-275.<br />
11. Palkovits M, Mezey E, Eskay RL. Pro-opiomelanocortin-derived peptides (ACTH/betaendorphin/alpha-MSH)<br />
in brainstem baroreceptor areas <strong>of</strong> the rat. Brain Res 1987; 436: 323-338.<br />
12. Evans CJ, Lorenz R, Weber E, Barchas J D. Variants <strong>of</strong> alpha-melanocyte stimulating hormone in<br />
rat brain <strong>and</strong> pituitary evidence that acetylated alpha-MSH exists only in the intermediate lobe <strong>of</strong><br />
pituitary. Biochem Biophys Res Commun 1982; 106: 910–919.<br />
13. Emeson RB, Eipper BA. Characterization <strong>of</strong> pro-ACTH/endorphin-derived peptides in rat<br />
hypothalamus. J Neurosci 1986; 6: 837–849.<br />
14. Jegou S, Tranch<strong>and</strong>-Bunel D, Delbende C, Blasquez C. Vaudry H. Characterization <strong>of</strong> alpha-<br />
MSH-related peptides released from rat hypothalamic neurons in vitro. Brain Res Mol Brain Res<br />
1989; 5, 219–226.<br />
15. Abbott CR., Rossi M, Kim M, AlAhmed SH, Taylor GM, Ghatei MA, Smith DM, Bloom SR.<br />
Investigation <strong>of</strong> the melanocyte stimulating hormones on food intake. Lack Of evidence to support<br />
a role for the melanocortin-3-receptor. Brain Res 2000; 869: 203–210.<br />
16. Tsujii S, Bray GA. Acetylation alters the feeding response to MSH <strong>and</strong> b-endorphin. Brain Res<br />
Bull 1989; 23:165–169.<br />
59
17. Shimizu H, Shargill NS, Bray GA, Yen TT, Gesellchen PD. Effects <strong>of</strong> MSH on food intake, body<br />
weight <strong>and</strong> coat color <strong>of</strong> the yellow obese mouse. Life Sci 1989;45:543–52.<br />
18. Abbott CR, Rossi M, Kim M-S, AlAhmed SH, Taylor GM, Ghatei MA, Smith DM, Bloom SR.<br />
Investigation <strong>of</strong> the melanocyte stimulating hormones on food intake. Lack <strong>of</strong> evidence to support<br />
a role for the melanocortin-3-receptor. Brain Res 2000;869:203–210.<br />
19. Guo L, Mu¨ nzberg H, Stuart RC, Nillni EA, Bjørbæk C. N-acetylation <strong>of</strong> hypothalamic α-<br />
melanocyte-stimulating hormone <strong>and</strong> regulation by leptin. Proc Natl Acad Sci USA 2004;<br />
101:11797-11802.<br />
20. Dhillo WS, Small CJ, Gardiner JV, Bewick GA, Whitworth EJ, Jethwa PH, Seal LJ, Ghatei MA,<br />
Hinson JP, Bloom SR. Agouti-related protein has an inhibitory paracrine role in the rat adrenal<br />
gl<strong>and</strong>. Biochem Biophys Res Commun 2003;301:102–107.<br />
21. Liakos P, Chambaz EM, Feige JJ, Defaye G. Expression <strong>of</strong> ACTH receptors (MC2-R <strong>and</strong> MC5-R)<br />
in the glomerulosa <strong>and</strong> the fasciculata-reticularis zones <strong>of</strong> bovine adrenal cortex. Endocr Res<br />
1998;24:427–432.<br />
22.Henville KL, Hinson JP, Vinson GP, Laird SM. Actions <strong>of</strong> desacetyl-α-melanocyte-stimulating<br />
hormone on human adrenocortical cells. J Endocrinol 1989;121:579–583.<br />
23. Baumann JB, Eberle AN, Christen E, Ruch W, Girard J. Steroidogenic activity <strong>of</strong> highly potent<br />
melanotropic peptides in the adrenal cortex <strong>of</strong> the rat. Acta Endocrinol 1986;113:396–402.<br />
24. Rudman D, Hollins BM, Kutner MH, M<strong>of</strong>fitt SD, Lynn MJ. Three types <strong>of</strong> α-melanocytestimulating<br />
hormone: bioactivities <strong>and</strong> half-lives. Am J Physiol 1983; 245:47–54.<br />
25. Schio¨th HB, Chhajlani V, Muceniece R, Klusa V, Wikberg JES. Major pharmacological<br />
distinction <strong>of</strong> the ACTH receptor from other melanocortin receptors. Life Sci 1996; 59:797–801.<br />
26. Boston BA. The role <strong>of</strong> melanocortins in adipocyte function. Ann NY Acad Sci 1999; 885:75–84.<br />
27. Hoggard N, Hunter L, Duncan JS, Rayner DV. Regulation <strong>of</strong> adipose tissue leptin secretion by α-<br />
melanocyte stimulating hormone <strong>and</strong> agouti-related protein: further evidence <strong>of</strong> an interaction<br />
between leptin <strong>and</strong> the melanocortin signalling system. J Mol Endocrinol 2004;32:145–153.<br />
60
Investigation <strong>of</strong> the mechanism <strong>of</strong> the cardioprotective effect <strong>of</strong> flavonols<br />
Chengxue Qin, 1,2 * Richard A. Hughes, 1 Spencer J. Williams, 2 Owen L. Woodman 3<br />
* Presenter<br />
1. The Department <strong>of</strong> Pharmacology, University <strong>of</strong> Melbourne, Parkville, 3010<br />
2. School <strong>of</strong> Chemistry <strong>and</strong> Bio21 Molecular Science <strong>and</strong> Biotechnology Institute, University<br />
<strong>of</strong> Melbourne, Parkville, 3010<br />
3. School <strong>of</strong> Medical Sciences, RMIT University, Bundoora, 3083<br />
We have reported that 3’, 4’-dihydroxyflavonol (DiOHF) reduces infarct size after<br />
myocardial ischaemia/reperfusion in sheep [1] <strong>and</strong> ischaemic stroke in rats [2]. It is however<br />
not known whether the cardioprotective action <strong>of</strong> this flavonol relies on its anti-calcium or<br />
antioxidant activities.<br />
In this study, we investigated the mechanism <strong>of</strong> the beneficial action <strong>of</strong> flavonols using<br />
our recently discovered single-action anti-calcium compound (4’-OH-3’-OCH 3 flavonol) [3]<br />
<strong>and</strong> the complementary antioxidant compound (DiOHF-6-succinamic acid) [4] in rat isolated<br />
hearts. Hearts were perfused with physiological solution using a Langendorff apparatus <strong>and</strong><br />
subjected to global, no-flow 20-minute ischaemia followed by 30-minute reperfusion. Hearts<br />
(n = 6~8 each group) were r<strong>and</strong>omly treated with vehicle (0.05% DMSO), DiOHF, 4’-OH-3’-<br />
OCH 3 flavonol or DiOHF-6-succinamic acid. Treatments (10 -5 M) were included in the<br />
perfusate for 10 min before ischaemia <strong>and</strong> during reperfusion. In vehicle treated hearts,<br />
LV+dP/dt was significantly reduced at the end <strong>of</strong> reperfusion (60±8%) compared to the preischaemic<br />
value. Furthermore, total lactate dehydrogenase (LDH) release was significantly<br />
elevated during reperfusion in vehicle treated hearts (895±121 U/L) compared to shams<br />
(70±10 U/L), indicating significant myocardial damage (p
Efficient Ethanol production in glucose/xylose co-fermentations by a novel<br />
co-culture scheme <strong>of</strong> Zymomonas mobilis <strong>and</strong> Pichia stipitis<br />
Nan Fu 1 * , Paul Peiris 1 , Julie Markham 1 , John Bavor 1<br />
* Presenter<br />
1. School <strong>of</strong> Natural Sciences, University <strong>of</strong> Western Sydney, Locked Bag 1797, Penrith<br />
South DC, NSW 1797, Australia<br />
The efficiency <strong>of</strong> bioethanol production from lignocellulose would be significantly<br />
improved if the glucose <strong>and</strong> xylose in the hydrolysate could be simultaneously converted to<br />
ethanol. This study designed a novel fermentation scheme, using a glucose-fermenting strain,<br />
Zymomonas mobilis <strong>and</strong> a pentose utilizing strain, Pichia stipitis to convert the sugar mixture<br />
substrates including both synthetic media <strong>and</strong> hydrolysate media.<br />
Fermentation characteristics <strong>of</strong> the two organisms were optimized in media containing<br />
50 g/l <strong>of</strong> the corresponding sugars. The inoculum was a key factor in the optimization for the<br />
production <strong>of</strong> ethanol. A ten fold increase in the cell number in the inoculum <strong>of</strong> Z. mobilis<br />
reduced the fermentation time from 8 to 2.5 h, resulting in an improvement in the overall<br />
volumetric ethanol productivity from 2.93 to 10.54 g/l/h. Similarly, a five fold increase in the<br />
cell number in the inoculum <strong>of</strong> P. stipitis resulted in a decrease <strong>of</strong> the fermentation time from<br />
130 to 28 h, leading to an increase <strong>of</strong> the volumetric productivity from 0.14 to 0.73 g/l/h. This<br />
is the highest value <strong>of</strong> volumetric ethanol productivity for Pichia stipitis ever reported.<br />
Co-fermentation <strong>of</strong> sugar mixture simulating the bagasse hydrolysate (30 g/l glucose <strong>and</strong><br />
20 g/l xylose) was carried out with different fermentation schemes which included free cells,<br />
immobilized cells, co-culture <strong>and</strong> sequential culture. Co-culture with immobilized Z. mobilis<br />
<strong>and</strong> free cells <strong>of</strong> P. stipitis proved to be the best scheme. Complete utilization <strong>of</strong> the sugars<br />
within 19 h with an ethanol yield <strong>of</strong> 0.49 g/g was achieved, giving a volumetric productivity<br />
<strong>of</strong> 1.277 g/l/h. Application <strong>of</strong> this fermentation scheme on sugarcane bagasse hydrolysate<br />
resulted in a complete sugar utilization within 26 h; ethanol yield peaked at 40 h with a yield<br />
<strong>of</strong> 0.49 g/g. These results are the highest reported values for co-culture <strong>and</strong> compare favorably<br />
with best results reported with recombinant strains.<br />
62
Dual mode roll-up effect in a multicomponent near adiabatic adsorption<br />
process<br />
Gang Li, 1 * Dong Xu, 1,2 Penny Xiao, 1 Paul A Webley 1<br />
* Presenter<br />
1. Cooperative Research Centre for Greenhouse Gas Technologies, Department <strong>of</strong> Chemical<br />
Engineering, Monash University, Wellington Road, Clayton, Victoria 3800, Australia<br />
2. School <strong>of</strong> Materials & Metallurgy, Northeastern University, Shenyang 110004, P.R. China<br />
Adsorption technologies have been widely used in laboratorial analysis/separations, e.g.<br />
chromatography <strong>and</strong> portable life support systems, <strong>and</strong> also in industrial process, e.g. air<br />
separation <strong>and</strong> carbon capture. A common phenomenon in multi-component adsorption in a<br />
fixed bed is that the effluent concentration <strong>of</strong> some components may exceed their steady inlet<br />
values. This effect, <strong>of</strong>ten referred to as roll-up, is attributed to the displacement <strong>of</strong> the more<br />
weakly adsorbed component by the slower moving more strongly adsorbed component,<br />
according to the conventional explanation in the literature. This has also been applied to some<br />
practical operations in liquid chromatography <strong>and</strong> some large-scale chromatographic<br />
separations to desorb the pre-loaded components from a column by introducing a desorbent,<br />
which is more strongly adsorbed than any other component in the system. Many published<br />
works reported that in a system containing two adsorbed species, one could easily see a single<br />
roll-up <strong>of</strong> the weak component during the breakthrough process. However, our recent study<br />
showed that more than one roll-up could occur, in such a system using CO 2 <strong>and</strong> H 2 O vapour<br />
as the weakly <strong>and</strong> strongly adsorbed components respectively in double zeolite 3A/13X <strong>and</strong><br />
activated alumina/13X layered columns. It has also been found that a pure thermal wave was<br />
generated by H 2 O adsorption <strong>and</strong> it moved ahead <strong>of</strong> the H 2 O mass transfer zone, <strong>and</strong> the first<br />
roll-up coincided with the propagation <strong>of</strong> the thermal front to the 3A/13X layer interface<br />
while the second roll-up happened when H 2 O concentration front reached the 13X layer. In<br />
theory, enthalpy can be regarded as a direct equivalence to an extra adsorbed component so<br />
that temperature front will sweep <strong>of</strong>f the column before the propagation <strong>of</strong> the strongly<br />
adsorbed component (H 2 O in this case), but it only signifies under adiabatic or near adiabatic<br />
conditions. The interval <strong>of</strong> the two roll-ups reduces with the decrease <strong>of</strong> the length <strong>of</strong> 3A layer,<br />
<strong>and</strong> eventually they merge into a single pronounced roll-up. The results were further verified<br />
by computer simulations. Therefore, to the best <strong>of</strong> our knowledge, this is the first affirmative<br />
report about this unique dual mode roll-up effect, which will be essential in the design <strong>of</strong> near<br />
adiabatic adsorption/chromatography units.<br />
63
Effect <strong>of</strong> glucose <strong>and</strong> sodium chloride on the stability <strong>of</strong> aged glucose<br />
oxidase from Aspergillus niger<br />
Chun Ming (Eric) Wong 1* , Xiao Dong Chen 1 , Kwun Hei Wong 2<br />
* Presenter<br />
1. Biotechnology <strong>and</strong> Food Engineering Group, Department <strong>of</strong> Chemical Engineering,<br />
Monash University, Clayton Campus, Victoria 3800, Australia<br />
2. Auximedic Ltd, PO box 109-175, Newmarket, Auckl<strong>and</strong> 1149, New Zeal<strong>and</strong><br />
Glucose-1-oxidase (GOX) from Aspergillus niger is a well characterized enzyme used in<br />
many commercial applications <strong>and</strong> is found naturally in insects <strong>and</strong> fungi. Various properties<br />
<strong>of</strong> GOX had been well documented but information on the effects <strong>of</strong> GOX stabilizer, glucose<br />
<strong>and</strong> sodium chloride are sparse. This article presents the effectiveness <strong>of</strong> glucose <strong>and</strong> sodium<br />
chloride as GOX stabilizer against drying <strong>and</strong> heat denaturation in solution with a three year<br />
old food grade GOX preparation over time. The primary method used to measure GOX<br />
activity was to log the decrease in dissolved oxygen using clark type electrode. Drying<br />
experiments showed that the additions <strong>of</strong> thermo-stabilizers do not increase the stability <strong>of</strong><br />
dried GOX but actually accelerates its rate <strong>of</strong> degradation over time. After about a month,<br />
only 1/3 <strong>of</strong> the initial activity was left in the dried GOX containing glucose <strong>and</strong>/or sodium<br />
chloride. Heat denaturantion experiments showed that thermostability <strong>of</strong> GOX increases in<br />
the presence <strong>of</strong> sodium chloride <strong>and</strong>/or glucose. Stabilization due to dissolved sodium<br />
chloride <strong>and</strong> glucose are cumulative, postulated to work via different mechanism. The<br />
contradicting finding presented leave no doubt that there is implications for GOX<br />
manufacturers <strong>and</strong> users whether or not they are working towards more stable GOX<br />
preparations.<br />
64
Paper as a low-cost base material for diagnostic <strong>and</strong> environmental sensing<br />
applications<br />
Xu Li 1 * , Junfei Tian 1 , Wei Shen 1<br />
* Presenter<br />
1. Australian Pulp <strong>and</strong> Paper Institute, Department <strong>of</strong> Chemical Engineering, Monash<br />
University, Victoria 3800, Australia<br />
Paper has been used as a chromatographic substrate <strong>and</strong> for indicator strips for a long<br />
time. Recently, a novel application using patterned paper to make micr<strong>of</strong>luidic systems for<br />
diagnostic <strong>and</strong> environmental sensing has been developed. The patterned paper which can<br />
control liquid penetration within its hydrophilic channels has great potential to make low-cost<br />
<strong>and</strong> simple analytical devices for bioassays <strong>and</strong> environmental analysis. This article shows<br />
that basic papermaking <strong>and</strong> printing techniques can be applied to create paper-based<br />
micr<strong>of</strong>luidic analytical devices. High volume <strong>and</strong> high speed printing processes hold the key<br />
for the transformation <strong>of</strong> the novel concept <strong>of</strong> paper-based micr<strong>of</strong>luidic system into practical<br />
analytical devices that will find applications in many fields.<br />
65
Data Management in Cloud Scientific Workflow Systems<br />
Dong Yuan 1 *<br />
* Presenter<br />
1. <strong>Faculty</strong> <strong>of</strong> <strong>Information</strong> <strong>and</strong> Communication Technologies, Swinburne University <strong>of</strong><br />
Technology, Hawthorn, Melbourne, Australia 3122<br />
Data-intensive scientific applications are posing many challenges in distributed<br />
computing systems. In the scientific field, the application data are expected to double every<br />
year over the next decade <strong>and</strong> further. With this continuing data explosion, high performance<br />
computing systems are needed to store <strong>and</strong> process data efficiently, <strong>and</strong> workflow<br />
technologies are facilitated to automate these scientific applications. Scientific workflows are<br />
typically very complex. They usually have a large number <strong>of</strong> tasks <strong>and</strong> need a long time for<br />
execution. Running scientific workflow applications usually need not only high performance<br />
computing resources but also massive storage. The emergence <strong>of</strong> cloud computing<br />
technologies <strong>of</strong>fers a new way to develop scientific workflow systems. Scientists can upload<br />
their data <strong>and</strong> launch their applications on the scientific cloud workflow systems from<br />
everywhere in the world via the Internet, <strong>and</strong> they only need to pay for the resources that they<br />
use for their applications. As all the data are managed in the cloud, it is easy to share data<br />
among scientists. This kind <strong>of</strong> model is very convenient for users, but remains a big challenge<br />
to the system. This seminar will discuss several research topics <strong>of</strong> data management in<br />
scientific cloud workflow systems.<br />
Keywords<br />
component, formatting, style, styling, insert (key words)<br />
1. Introduction<br />
Data-intensive scientific applications are posing many challenges in distributed computing systems.<br />
In many scientific research fields, like astronomy, high-energy physics <strong>and</strong> bio-informatics, scientists<br />
need to analyse terabytes <strong>of</strong> data either from existing data resources or collected from physical<br />
devices. During these processes, similar amounts <strong>of</strong> new data might also be generated as intermediate<br />
or <strong>final</strong> products [9]. According to [20], in the scientific field, the application data are expected to<br />
double every year over the next decade <strong>and</strong> further. With this continuing data explosion, high<br />
performance computing systems are needed to store <strong>and</strong> process data efficiently, <strong>and</strong> workflow<br />
technologies are facilitated to automate these scientific applications. Scientific workflows are typically<br />
very complex. They usually have a large number <strong>of</strong> tasks <strong>and</strong> need a long time for execution. Running<br />
scientific workflow applications usually need not only high performance computing resources but also<br />
massive storage [9].<br />
Nowadays, popular scientific workflows are <strong>of</strong>ten deployed in grid systems [16] because they have<br />
high performance <strong>and</strong> massive storage. However, building a grid system is extremely expensive <strong>and</strong> it<br />
is normally not open for scientists all over the world. The emergence <strong>of</strong> cloud computing technologies<br />
<strong>of</strong>fers a new way to develop scientific workflow systems.<br />
Since late 2007 the concept <strong>of</strong> cloud computing was proposed [22], it has been utilised in many<br />
areas with some success. Cloud computing is deemed as the next generation <strong>of</strong> IT platforms that can<br />
deliver computing as a kind <strong>of</strong> utility [8]. Foster et al. made a comprehensive comparison <strong>of</strong> grid<br />
computing <strong>and</strong> cloud computing [11]. Some features <strong>of</strong> cloud computing also meet the requirements<br />
<strong>of</strong> scientific workflow systems. Cloud computing systems provide the high performance <strong>and</strong> massive<br />
storage required for scientific applications in the same way as grid systems, but with a lower<br />
infrastructure construction cost among many other features, because cloud computing systems are<br />
composed <strong>of</strong> data centres which can be clusters <strong>of</strong> commodity hardware [22]. Research into doing<br />
science <strong>and</strong> data-intensive applications on the cloud has already commenced [18], such as early<br />
experiences like Nimbus [14] <strong>and</strong> Cumulus [21] projects. The work by Deelman et al. [10] shows that<br />
67
cloud computing <strong>of</strong>fers a cost-effective solution for data-intensive applications, such as scientific<br />
workflows [13]. Furthermore, cloud computing systems <strong>of</strong>fer a new model that scientists from all over<br />
the world can collaborate <strong>and</strong> conduct their research together. Cloud computing systems are based on<br />
the Internet, <strong>and</strong> so are the scientific workflow systems deployed in the cloud. Scientists can upload<br />
their data <strong>and</strong> launch their applications on the scientific cloud workflow systems from everywhere in<br />
the world via the Internet, <strong>and</strong> they only need to pay for the resources that they use for their<br />
applications. As all the data are managed in the cloud, it is easy to share data among scientists. This<br />
kind <strong>of</strong> model is very convenient for users, but remains a big challenge for data management to<br />
scientific cloud workflow systems.<br />
Firstly, new data storage strategy is required in cloud scientific workflow systems. In a cloud<br />
computing, theoretically, the system can <strong>of</strong>fer unlimited storage resources. All the application data <strong>of</strong><br />
the scientific workflows can be stored, including the generated intermediate data, if we are willing to<br />
pay for the required resources. Storing all the application data in the cloud is obviously not costeffective,<br />
since some data are seldom used <strong>and</strong> huge in size. However, in scientific cloud workflow<br />
systems, a scientist can not decide whether a piece <strong>of</strong> application data should be stored or not, since<br />
the data are shared <strong>and</strong> he is not the only user.<br />
Secondly, new data placement strategy is also required, which means the cloud workflow systems<br />
must have the ability to decide where to store the application data. Cloud computing platform contains<br />
different cloud service providers with different pricing models, where data transfers between service<br />
providers also carry a cost. The cloud scientific workflows are usually distributed, <strong>and</strong> the data<br />
placement strategy will decide where to store the application data, in order to reduce the total system<br />
cost.<br />
Last but not least, new data replication strategy should also be designed for cloud scientific<br />
workflow systems. A good replication strategy can not only guarantee the security <strong>of</strong> application data,<br />
but also further reduce the system cost by replicating frequently used data in different locations.<br />
Replication strategy in the Cloud should be dynamic based on the application data’s usage rate.<br />
2. Related Works<br />
As cloud computing has become more <strong>and</strong> more popular, new data management systems have also<br />
appeared, such as Google File System [12] <strong>and</strong> Hadoop [1]. They all have hidden infrastructures that<br />
can store the application data independent <strong>of</strong> users’ control. Google File System is designed mainly for<br />
Web search applications, which are different from workflow applications. Hadoop is a more general<br />
distributed file system, which has been used by many companies, such as Amazon <strong>and</strong> Facebook.<br />
When you push a file to a Hadoop File System, it will automatically split this file into chunks <strong>and</strong><br />
r<strong>and</strong>omly distribute these chunks in a cluster. Furthermore, the Cumulus project [21] introduced a<br />
scientific cloud architecture for a data centre. And the Nimbus [14] toolkit can directly turn a cluster<br />
into a cloud <strong>and</strong> it has already been used to build a cloud for scientific applications.<br />
Comparing to the distributed computing systems like cluster <strong>and</strong> grid, a cloud computing system<br />
has a cost benefit [4]. Assunção et al. [5] demonstrate that cloud computing can extend the capacity <strong>of</strong><br />
clusters with a cost benefit. Using Amazon clouds’ cost model <strong>and</strong> BOINC volunteer computing<br />
middleware, the work in [15] analyses the cost benefit <strong>of</strong> cloud computing versus grid computing. The<br />
idea <strong>of</strong> doing science on the cloud is not new. Scientific applications have already been introduced to<br />
cloud computing systems. In terms <strong>of</strong> the cost benefit, the work by Deelman et al. [10] also applies<br />
Amazon clouds’ cost model <strong>and</strong> demonstrates that cloud computing <strong>of</strong>fers a cost-effective way to<br />
deploy scientific applications. In [13], H<strong>of</strong>fa conducted simulations <strong>of</strong> running an astronomy scientific<br />
workflow in cloud <strong>and</strong> clusters, which shows cloud scientific workflows are cost-effective. The above<br />
works mainly focus on the comparison <strong>of</strong> cloud computing systems <strong>and</strong> the traditional distributed<br />
computing paradigms, which shows that applications running on cloud have cost benefits. When it<br />
comes to how to reduce cost <strong>of</strong> running applications in clouds, Deelman et al. present in [10] that<br />
storing some popular intermediate data can save the cost in comparison to always regenerating them<br />
from the input data. Furthermore, in [2], Adams proposes a model to represent the trade-<strong>of</strong>f <strong>of</strong><br />
computation cost <strong>and</strong> storage cost in storing application data, but the authors have not given the<br />
specific method <strong>of</strong> managing the data.<br />
68
To the best <strong>of</strong> our knowledge, research in cost-effectively managing application data <strong>of</strong> cloud<br />
scientific workflow systems is still in blank. My research will focus on this spot <strong>and</strong> develop novel<br />
strategies <strong>of</strong> data storage, placement <strong>and</strong> replication in cloud scientific workflow systems.<br />
3. Research problems <strong>and</strong> methodologies<br />
A. Data storage strategy<br />
1) Problem analysis<br />
Traditionally, scientific workflows are deployed on the high performance computing facilities, such<br />
as clusters <strong>and</strong> grids. Scientific workflows are <strong>of</strong>ten complex with huge intermediate data generated<br />
during their execution. How to store these intermediate data is normally decided by the scientists who<br />
use the scientific workflows. This is because the clusters <strong>and</strong> grids only serve for certain institutions.<br />
The scientists may store the intermediate data that are most valuable to them, based on the storage<br />
capacity <strong>of</strong> the system. However, in many scientific workflow systems, the storage capacities are<br />
limited, such as the pulsar searching workflow we introduced. The scientists have to delete all the<br />
intermediate data because <strong>of</strong> the storage limitation. This bottleneck <strong>of</strong> storage can be avoided if we run<br />
scientific workflows in the cloud.<br />
In a cloud computing environment, theoretically, the system can <strong>of</strong>fer unlimited storage resources.<br />
All the intermediate data generated by scientific cloud workflows can be stored, if we are willing to<br />
pay for the required resources. However, in scientific cloud workflow systems, whether to store<br />
intermediate data or not is not an easy decision anymore.<br />
a) All the resources in the cloud carry certain costs, so either storing or generating an intermediate<br />
dataset, we have to pay for the resources used. The intermediate datasets vary in size, <strong>and</strong> have<br />
different generation cost <strong>and</strong> usage rate. Some <strong>of</strong> them may <strong>of</strong>ten be used whilst some others may be<br />
not. On one h<strong>and</strong>, it is most likely not cost effective to store all the intermediate data in the cloud. On<br />
the other h<strong>and</strong>, if we delete them all, regeneration <strong>of</strong> frequently used intermediate datasets imposes a<br />
high computation cost. We need a strategy to balance the generation cost <strong>and</strong> the storage cost <strong>of</strong> the<br />
intermediate data, in order to reduce the total cost <strong>of</strong> the scientific cloud workflow system.<br />
b) The scientists can not predict the usage rate <strong>of</strong> the intermediate data anymore. For a single<br />
research group, if the data resources <strong>of</strong> the applications are only used by its own scientists, the<br />
scientists may predict the usage rate <strong>of</strong> the intermediate data <strong>and</strong> decide whether to store or delete<br />
them. However, the scientific cloud workflow system is not developed for a single scientist or<br />
institution, rather, developed for scientists from different institutions to collaborate <strong>and</strong> share data<br />
resources. The users <strong>of</strong> the system could be anonymous from the Internet. We must have a strategy<br />
storing the intermediate data based on the needs <strong>of</strong> all the users that can reduce the cost <strong>of</strong> the whole<br />
system.<br />
Hence, for scientific cloud workflow systems, we need a strategy that can automatically select <strong>and</strong><br />
store the most appropriate intermediate datasets. Furthermore, this strategy should be cost effective<br />
that can reduce the total cost <strong>of</strong> the whole system.<br />
2) Methodology<br />
Scientific workflows have many computation <strong>and</strong> data intensive tasks that will generate many<br />
intermediate datasets <strong>of</strong> considerable size. There are dependencies exist among the intermediate<br />
datasets. Data provenance in workflows is a kind <strong>of</strong> important metadata, in which the dependencies<br />
between datasets are recorded. The dependency depicts the derivation relationship between workflow<br />
intermediate datasets. For scientific workflows, data provenance is especially important, because after<br />
the execution, some intermediate datasets may be deleted, but sometimes the scientists have to<br />
regenerate them for either reuse or reanalysis. Data provenance records the information <strong>of</strong> how the<br />
intermediate datasets were generated, which is very important for the scientists. Furthermore,<br />
regeneration <strong>of</strong> the intermediate datasets from the input data may be very time consuming, <strong>and</strong><br />
therefore carry a high cost. With data provenance information, the regeneration <strong>of</strong> the dem<strong>and</strong>ing<br />
dataset may start from some stored intermediated datasets instead. In the scientific cloud workflow<br />
system, data provenance is recorded while the workflow execution. Taking advantage <strong>of</strong> data<br />
provenance, we can build an IDG based on data provenance. All the intermediate datasets once<br />
generated in the system, whether stored or deleted, their references are recorded in the IDG. Based on<br />
the IDG, we can calculate the generation cost <strong>of</strong> every intermediate dataset in the scientific cloud<br />
69
workflows. By comparing the generation cost <strong>and</strong> storage cost, our strategy can automatically decide<br />
whether an intermediate dataset should be stored or deleted in the cloud system, no matter this<br />
intermediate dataset is a new dataset, regenerated dataset or stored dataset in the system.<br />
Data provenance is important to this strategy in building the IDG. Fortunately, due to the<br />
importance <strong>of</strong> data provenance in scientific applications, much research about recording data<br />
provenance <strong>of</strong> the system has been done. Some <strong>of</strong> them are especially for scientific workflow systems<br />
[6]. Some popular scientific workflow systems, such as Kepler [16], have their own system to record<br />
provenance during the workflow execution [3]. In [19], Osterweil et al. present how to generate a Data<br />
Derivation Graph (DDG) for the execution <strong>of</strong> a scientific workflow, where one DDG records the data<br />
provenance <strong>of</strong> one execution. Similar to the DDG, our IDG is also based on the scientific workflow<br />
data provenance, but it depicts the dependency relationships <strong>of</strong> all the intermediate data in the system.<br />
Hence we can build our IDG by taking advantage <strong>of</strong> these related works.<br />
B. Data placement strategy<br />
1) Problem Analysis<br />
Scientific applications are data intensive <strong>and</strong> usually need collaborations <strong>of</strong> scientists from different institutions [7], hence application<br />
data in scientific workflows are usually distributed <strong>and</strong> very large. When one task needs to process data from different data centres, moving<br />
data becomes a challenge. Some application data are too large to be moved efficiently, some may have fixed locations that are not feasible to<br />
be moved <strong>and</strong> some may have to be located at fixed data centres for processing, but these are only one aspect <strong>of</strong> this challenge. For the<br />
application data that are flexible to be moved, we also cannot move them whenever <strong>and</strong> wherever we want, since in the cloud computing<br />
platform, data centres may belong to different cloud service providers that data movement would result in costs. Furthermore, the<br />
infrastructure <strong>of</strong> cloud computing systems is hidden from their users. They just <strong>of</strong>fer the computation <strong>and</strong> storage resources required by users<br />
for their applications. The users do not know the exact physical locations where their data are stored. This kind <strong>of</strong> model is very convenient<br />
for users, but remains a big challenge for data management to scientific cloud workflow systems.<br />
2) Methodology<br />
In cloud computing systems, the infrastructure is hidden from users. Hence, for most <strong>of</strong> the<br />
application data, the system will decide where to store them. Dependencies exist among these data. In<br />
this paper, we initially adapt the clustering algorithms for data movement based on data dependency.<br />
Clustering algorithms have been used in pattern recognition since 1980s, which can classify patterns<br />
into groups without supervision. Today they are widely used to process data streams. In many<br />
scientific workflow applications, the intermediate data movement is in data stream format <strong>and</strong> the<br />
newly generated data must be moved to the destination in real-time. We adapt the k-means clustering<br />
algorithm for data placement in scientific cloud workflow systems. Scientific workflows can be very<br />
complex, one task might require many datasets for execution; furthermore, one dataset might also be<br />
required by many tasks. If some datasets are always used together by many tasks, we say that these<br />
datasets are dependant on each other. In our strategy, we try to keep these datasets in one data centre,<br />
so that when tasks were scheduled to this data centre, most, if not all, <strong>of</strong> the data they need are stored<br />
locally.<br />
Our data placement strategy has two algorithms, one for the build-time stage <strong>and</strong> one for the<br />
runtime stage <strong>of</strong> scientific workflows. In the build-time stage algorithm, we construct a dependency<br />
matrix for all the application data, which represents the dependencies between all the datasets<br />
including the datasets that may have fixed locations. Then we use the BEA algorithm [17] to cluster<br />
the matrix <strong>and</strong> partition it that datasets in every partition are highly dependent upon each other. We<br />
distribute the partitions into k data centres, where the partitions have fixed location datasets are also<br />
placed in the appropriate data centres. These k data centres are initially as the partitions <strong>of</strong> the k-means<br />
algorithm at runtime stage. At runtime, our clustering algorithm deals with the newly generated data<br />
that will be needed by other tasks. For every newly generated dataset, we calculate its dependencies<br />
with all k data centres, <strong>and</strong> move the data to the data centre that has the highest dependency with it.<br />
By placing data with their dependencies, our strategy attempts to minimise the total data movement<br />
during the execution <strong>of</strong> workflows. Furthermore, with the pre-allocate <strong>of</strong> data to other data centres, our<br />
strategy can prevent data gathering to one data centre <strong>and</strong> reduces the time spent waiting for data by<br />
ensuring that relevant data are stored locally.<br />
C. Data replication strategy<br />
1) Problem Analysis<br />
Data replication is also an important issue for cloud workflow systems as presented by the two<br />
points below:<br />
70
First, a good replication strategy can guarantee fast data access for the cloud workflow system. In<br />
the scientific workflows with many parallel tasks will simultaneously access the same dataset on one<br />
data centre. The limitation <strong>of</strong> computing capacity <strong>and</strong> b<strong>and</strong>width in that data centre would be a<br />
bottleneck for the whole cloud workflow system. If we have several replicas in different data centres,<br />
this bottleneck will be eliminated.<br />
Second, a good replication strategy can reduce data movement between data centres. For example,<br />
if tasks in one data centre always need to retrieve data from the same data set in a remote data centre,<br />
it is better to replicate that data set in the local data centre to reduce the data movement.<br />
Third, a good replication strategy can guarantee data reliability for the cloud workflow system.<br />
Because data centres in cloud workflow systems are built up with massive cheap commodity<br />
hardware, the breakdown <strong>of</strong> some hardware could happen any time. It is essential to keep several<br />
copies <strong>of</strong> each data in different data centres for reliability.<br />
However, at present, data replication strategies that utilised in cloud data management systems are<br />
usually static. For example, in Hadoop, users can manually set the number <strong>of</strong> replicas, <strong>and</strong> the system<br />
will automatically replicate the application data in different places (racks or clusters, depends on the<br />
scale <strong>of</strong> the system). Static replication can guarantee the data reliability, but in cloud environment,<br />
different application data have different usage rate, where we should have the dynamic strategy to<br />
replicate the application data based on their usage rate.<br />
2) Methodology<br />
The basic strategy for the replication could be as follow:<br />
a) Always keep fix number copies <strong>of</strong> each dataset in different data centres to guarantee reliability<br />
<strong>and</strong> dynamically add new replicas for each dataset to to guarantee data availability.<br />
b) Where to place the replicas is based on data dependency.<br />
c) How many replicas should a dataset have is based on usage rate <strong>of</strong> this dataset.<br />
Reference:<br />
[1] "Hadoop, http://hadoop.apache.org/", accessed on 25 November 2009.<br />
[2] I. Adams, D. D. E. Long, E. L. Miller, S. Pasupathy, <strong>and</strong> M. W. Storer, "Maximizing Efficiency<br />
By Trading Storage for Computation," in <strong>Workshop</strong> on Hot Topics in Cloud Computing<br />
(HotCloud'09), pp. 1-5, 2009.<br />
[3] I. Altintas, O. Barney, <strong>and</strong> E. Jaeger-Frank, "Provenance Collection Support in the Kepler<br />
Scientific Workflow System," in International Provenance <strong>and</strong> Annotation <strong>Workshop</strong>, pp. 118-<br />
132, 2006.<br />
[4] M. Armbrust, A. Fox, R. Griffith, A. D. Joseph, R. H. Katz, A. Konwinski, G. Lee, D. A.<br />
Patterson, A. Rabkin, I. Stoica, <strong>and</strong> M. Zaharia, "Above the Clouds: A Berkeley View <strong>of</strong> Cloud<br />
Computing," University <strong>of</strong> California at Berkeley,<br />
http://www.eecs.berkeley.edu/Pubs/TechRpts/2009/EECS-2009-28.<strong>pdf</strong>, Technical Report<br />
UCB/EECS-2009-28, accessed on 25 November 2009.<br />
[5] M. D. d. Assuncao, A. d. Costanzo, <strong>and</strong> R. Buyya, "Evaluating the cost-benefit <strong>of</strong> using cloud<br />
computing to extend the capacity <strong>of</strong> clusters," in 18th ACM International Symposium on High<br />
Performance Distributed Computing, Garching, Germany, pp. 1-10, 2009.<br />
[6] Z. Bao, S. Cohen-Boulakia, S. B. Davidson, A. Eyal, <strong>and</strong> S. Khanna, "Differencing Provenance<br />
in Scientific Workflows," in 25th IEEE International Conference on Data Engineering, ICDE<br />
'09., pp. 808-819, 2009.<br />
[7] R. Barga <strong>and</strong> D. Gannon, "Scientific versus Business Workflows," in Workflows for e-Science,<br />
pp. 9-16, 2007.<br />
[8] R. Buyya, C. S. Yeo, S. Venugopal, J. Broberg, <strong>and</strong> I. Br<strong>and</strong>ic, "Cloud computing <strong>and</strong> emerging<br />
IT platforms: Vision, hype, <strong>and</strong> reality for delivering computing as the 5th utility," Future<br />
Generation Computer Systems, vol. in press, pp. 1-18, 2009.<br />
[9] E. Deelman <strong>and</strong> A. Chervenak, "Data Management Challenges <strong>of</strong> Data-Intensive Scientific<br />
Workflows," in IEEE International Symposium on Cluster Computing <strong>and</strong> the Grid, pp. 687-692,<br />
2008.<br />
71
[10] E. Deelman, G. Singh, M. Livny, B. Berriman, <strong>and</strong> J. Good, "The Cost <strong>of</strong> Doing Science on the<br />
Cloud: the Montage example," in ACM/IEEE Conference on Supercomputing, Austin, Texas, pp.<br />
1-12, 2008.<br />
[11] I. Foster, Z. Yong, I. Raicu, <strong>and</strong> S. Lu, "Cloud Computing <strong>and</strong> Grid Computing 360-Degree<br />
Compared," in Grid Computing Environments <strong>Workshop</strong>, GCE '08, pp. 1-10, 2008.<br />
[12] S. Ghemawat, H. Gobi<strong>of</strong>f, <strong>and</strong> S.-T. Leung, "The Google file system," SIGOPS Oper. Syst. Rev.,<br />
vol. 37, pp. 29-43, 2003.<br />
[13] C. H<strong>of</strong>fa, G. Mehta, T. Freeman, E. Deelman, K. Keahey, B. Berriman, <strong>and</strong> J. Good, "On the Use<br />
<strong>of</strong> Cloud Computing for Scientific Workflows," in 4th IEEE International Conference on e-<br />
Science, pp. 640-645, 2008.<br />
[14] K. Keahey, R. Figueiredo, J. Fortes, T. Freeman, <strong>and</strong> M. Tsugawa, "Science Clouds: Early<br />
Experiences in Cloud Computing for Scientific Applications," in First <strong>Workshop</strong> on Cloud<br />
Computing <strong>and</strong> its Applications (CCA'08), pp. 1-6, 2008.<br />
[15] D. Kondo, B. Javadi, P. Malecot, F. Cappello, <strong>and</strong> D. P. Anderson, "Cost-benefit analysis <strong>of</strong><br />
Cloud Computing versus desktop grids," in IEEE International Symposium on Parallel &<br />
Distributed Processing, IPDPS'09, pp. 1-12, 2009.<br />
[16] B. Ludascher, I. Altintas, C. Berkley, D. Higgins, E. Jaeger, M. Jones, <strong>and</strong> E. A. Lee, "Scientific<br />
workflow management <strong>and</strong> the Kepler system," Concurrency <strong>and</strong> Computation: Practice <strong>and</strong><br />
Experience, pp. 1039–1065, 2005.<br />
[17] W. T. McCormick, P. J. Sehweitzer, <strong>and</strong> T. W. White, "Problem Decomposition <strong>and</strong> Data<br />
Reorganization by a Clustering Technique," Operations Research, vol. 20, pp. 993-1009, 1972.<br />
[18] C. Moretti, J. Bulosan, D. Thain, <strong>and</strong> P. J. Flynn, "All-Pairs: An Abstraction for Data-Intensive<br />
Cloud Computing," in IEEE International Parallel & Distributed Processing Symposium,<br />
IPDPS'08, pp. 1-11, 2008.<br />
[19] L. J. Osterweil, L. A. Clarke, A. M. Ellison, R. Podorozhny, A. Wise, E. Boose, <strong>and</strong> J. Hadley,<br />
"Experience in Using A Process Language to Define Scientific Workflow <strong>and</strong> Generate Dataset<br />
Provenance," in 16th ACM SIGSOFT International Symposium on Foundations <strong>of</strong> S<strong>of</strong>tware<br />
Engineering Atlanta, Georgia: ACM, 2008.<br />
[20] A. S. Szalay <strong>and</strong> J. Gray, "Science in an Exponential World," Nature, vol. 440, pp. 23-24, 2006.<br />
[21] L. Wang, J. Tao, M. Kunze, A. C. Castellanos, D. Kramer, <strong>and</strong> W. Karl, "Scientific Cloud<br />
Computing: Early Definition <strong>and</strong> Experience," in 10th IEEE International Conference on High<br />
Performance Computing <strong>and</strong> Communications, HPCC '08. , pp. 825-830, 2008.<br />
[22] A. Weiss, "Computing in the Cloud," ACM Networker, vol. 11, pp. 18-25, 2007.<br />
72
HC_AB: A New Heuristic Clustering Algorithm based on Approximate<br />
Backbone<br />
* Presenter<br />
1. Victoria University<br />
Yu Zong 1 * , Gu<strong>and</strong>ong Xu 1 , Yanchun Zhang 1 , Mingchu Li 1<br />
Clustering is an important research area with numerous applications in pattern<br />
recognition, machine learning, <strong>and</strong> data mining. Since the clustering problem on numeric data<br />
sets can be formulated as a typical combinatorial optimization problem, many research studies<br />
have addressed the design <strong>of</strong> heuristic algorithms for finding local suboptimal solutions in a<br />
reasonable period <strong>of</strong> time. However, a majority <strong>of</strong> heuristic clustering algorithms suffer from<br />
the problem <strong>of</strong> being sensitive to the initialization <strong>and</strong> not guaranteeing the high quality<br />
clusters. Recently, Approximate Backbone (AB), i.e., the commonly shared intersections <strong>of</strong><br />
several local suboptimal solutions, has been proposed in heuristic algorithm design to<br />
improve the performance <strong>of</strong> heuristic algorithms. In this paper, we aim to introduce the AB<br />
into heuristic clustering to overcome the drawbacks <strong>of</strong> conventional heuristic clustering<br />
algorithms. The main strength <strong>of</strong> the proposed method is the capability <strong>of</strong> restricting the<br />
initial search space around the global optimal results by defining the AB, <strong>and</strong> in turn,<br />
reducing the impact <strong>of</strong> initialization on clustering, eventually improving the performance <strong>of</strong><br />
heuristic clustering. Experiments on synthesis <strong>and</strong> real world data sets are performed to<br />
validate the effectiveness <strong>of</strong> the proposed approach.<br />
73
Web Page Prediction Based on Conditional R<strong>and</strong>om Fields<br />
* Presenter<br />
1. The University <strong>of</strong> Melbourne<br />
Yong Zhen Guo 1 * , Yuan Miao 1<br />
Web page prefetching is used to reduce the access latency on the Internet. However, if<br />
most prefetched Web pages are not visited by the users in their subsequent accesses, the<br />
limited network b<strong>and</strong>width <strong>and</strong> server resources will not be used efficiently <strong>and</strong> may even<br />
worsen the access delay problem. Therefore, enhancing the Web page prediction accuracy is a<br />
key issue <strong>of</strong> Web page prefetching. In this talk, a Web page prediction method based on the<br />
powerful sequential learning model, Conditional R<strong>and</strong>om Fields (CRFs), is proposed to<br />
improve the Web prediction accuracy. We also show how to scale the CRF-based Web<br />
prediction method to large-size websites by using the ECOC (Error Correcting Output<br />
Coding) technique. Moreover, because the limited class information provided to the binarylabel<br />
sub-classifiers in ECOCCRFs will also lead to inferior accuracy when compared to the<br />
multi-label CRFs, in this talk, we introduce the grouping method which allows us to obtain a<br />
prediction accuracy closer to that <strong>of</strong> multi-label CRFs while maintaining the advantage <strong>of</strong><br />
ECOC-CRFs. The experimental results show that the Web prediction method based on the<br />
grouped ECOC-CRFs is highly accurate <strong>and</strong> scalable, <strong>and</strong> is ready for use in large-scale<br />
websites to perform predictions.<br />
74
Provenancing Qualifications Using MEASUR within Higher Education<br />
Institutions: An Australian Case<br />
Xiaochen Li 1 * , Bruce Calway 1 , <strong>and</strong> Suku Sinnappan 1<br />
* Presenter<br />
1. Swinburne University <strong>of</strong> Technology<br />
Individuals move across educational institutions, learning modes <strong>and</strong> borders for further<br />
study <strong>and</strong> employment purpose. Properly provenanced learning credentials can be carried<br />
across institutes <strong>and</strong> borders <strong>and</strong> be recognised, thus facilitate individuals’ mobility.<br />
Provenance is the origin or source from which an item comes, <strong>and</strong> the history <strong>of</strong> subsequent<br />
owners or pro<strong>of</strong> <strong>of</strong> authenticity. Provenance data has been collected <strong>and</strong> used in areas such as<br />
online learning, social networks, <strong>and</strong> e-science research for a range <strong>of</strong> uses. This workinprogress<br />
paper models the existing workflows adopted by higher education institutions on<br />
assessing qualifications using MEASUR method. The workflow model has implications on<br />
the collection <strong>of</strong> provenance data <strong>of</strong> learning credentials for authentication <strong>and</strong> verification<br />
purpose.<br />
75
A Multi-modal Gesture Recognition System in a Human-Robot Interaction<br />
Scenario<br />
* Presenter<br />
1. Monash University<br />
Zhi Li 1 * <strong>and</strong> Ray Jarvis 1<br />
Recognition <strong>of</strong> non-verbal gestures is essential for robots to underst<strong>and</strong> a user’s state <strong>and</strong><br />
intention in a Human-Robot Interaction (HRI) scenario. In this paper a multi-modal system is<br />
proposed to recognize a user’s h<strong>and</strong> gestures <strong>and</strong> estimate body poses from a single viewpoint.<br />
A newly emerged active sensing technology is employed to derive depth data at a high frame<br />
rate. Depth data is useful for objects detection <strong>and</strong> localization in 3D spaces <strong>and</strong> for image<br />
segmentation. A pair <strong>of</strong> stereo cameras is used to sense the head gestures <strong>of</strong> the user, which<br />
provides useful information about the user’s attention. Both h<strong>and</strong> shapes <strong>and</strong> h<strong>and</strong> trajectories<br />
are recognized. Full configurations <strong>of</strong> body poses are estimated using a model-based<br />
algorithm. Poses are tracked by a Particle Filter, <strong>and</strong> refined by a gradient-based searching<br />
method in the neighborhood <strong>of</strong> the particles.<br />
76
Distributed Agent based Interoperable Virtual EMR System for Healthcare<br />
System Integration<br />
* Presenter<br />
1. Vicotoria University<br />
Xuebing Yang 1 * , Yuan Miao<br />
One <strong>of</strong> the major problems in health care system integration is the formidable cost <strong>of</strong><br />
mediating between myriad vendors <strong>and</strong> policy makers for updating existing heterogeneous<br />
systems to support a great variety <strong>of</strong> st<strong>and</strong>ards or interfaces. To provide cost-effective<br />
healthcare system integration solution, this paper presents a Graphical User Interface state<br />
model (GUISM) for automatically exchanging information with existing healthcare s<strong>of</strong>tware<br />
through their GUIs with no modifications needed to them. This can save the huge cost <strong>of</strong><br />
upgrading, testing <strong>and</strong> redeploying the existing systems. By using the GUISM model,<br />
distributed agents are deployed to the client computers interacting with the local electronic<br />
medical system (EMR) for communicating with other EMR systems. The whole system is<br />
called virtual EMR system <strong>and</strong> each client in this system can request needed patient<br />
healthcare information without knowing the actual location <strong>of</strong> the data.<br />
77
Searching for Fair Joint Gains in Agent-based Negotiation<br />
* Presenter<br />
1. Swinburne University<br />
Minyi Li 1 *<br />
In multi-issue negotiations, autonomous agents can act cooperatively to benefit from<br />
mutually preferred agreements. However, empirical evidence suggests that they <strong>of</strong>ten fail to<br />
search for joint gains <strong>and</strong> end up with inefficient results. To address this problem, we propose<br />
a novel mediated negotiation procedure to support the negotiation agents in reaching an<br />
efficient <strong>and</strong> fair agreement in bilateral multi-issue negotiation. At each stage <strong>of</strong> negotiation,<br />
the mediator searches for the compromise direction based on a new E-DD (Equal Directional<br />
Derivative) approach <strong>and</strong> computes the new tentative agreement. We experimentally evaluate<br />
the proposed approach with different kinds <strong>of</strong> utility functions. The experimental result<br />
demonstrates that the proposed approach not only guarantees Pareto efficiency, but also<br />
produces fairer improvements for two negotiating agents compared with other existing<br />
methods.<br />
78
Mesoporous Silica-Templated Assembly <strong>of</strong> Luminescent Polyester Particles<br />
[1]<br />
Jiwei Cui, 1 2 * Yajun Wang, 1 2 Jingcheng Hao, 2 <strong>and</strong> Frank Caruso 1<br />
* Presenter<br />
1. Centre for Nanoscience <strong>and</strong> Nanotechnology, Department <strong>of</strong> Chemical <strong>and</strong> Biomolecular<br />
Engineering, The University <strong>of</strong> Melbourne, Parkville, Victoria 3010, Australia.<br />
2. Key Laboratory <strong>of</strong> Colloid <strong>and</strong> Interface Chemistry <strong>of</strong> Ministry <strong>of</strong> Education, Sh<strong>and</strong>ong<br />
University, Jinan 250100, P. R. China.<br />
We report the template assembly <strong>of</strong> luminescent poly-3-hydroxybutyrate particles doped<br />
with rare-earth complexes. The hydrophobic polymer, poly-3-hydroxybutyrate, has been<br />
infiltrated into the nanopores <strong>of</strong> mesoporous silica particles in organic solvent. Owing to the<br />
van der Waals interaction <strong>of</strong> the polymer chains, poly-3-hydroxybutyrate loaded in the<br />
nanopores yields replicated particles following removal <strong>of</strong> the mesoporous silica template. To<br />
prevent aggregation <strong>of</strong> the hydrophobic poly-3-hydroxybutyrate particles in aqueous media,<br />
the poly-3-hydroxybutyrate-loaded mesoporous silica particles were coated with a<br />
polyelectrolyte multilayer shell through the layer-by-layer assembly <strong>of</strong> poly(allylamine<br />
hydrochloride) <strong>and</strong> poly(sodium 4-styrenesulfonate). Following removal <strong>of</strong> the silica core, the<br />
polyelectrolyte multilayer-coated poly-3-hydroxybutyrate replicas were used to effectively<br />
coordinate rare-earth complexes (europium β-diketone complexes). The rare-earth-loaded<br />
poly-3-hydroxybutyrate replicas coated with polyelectrolyte multilayer emit intense<br />
luminescence over a wide range <strong>of</strong> pH (3~11) <strong>and</strong> for at least several months in aqueous<br />
solution, which is due to the intramolecular energy transfer from the lig<strong>and</strong> to the luminescent<br />
center in the rare-earth complexes. The poly-3-hydroxybutyrate replicas with stable <strong>and</strong><br />
intense luminescence may find application in diagnostics <strong>and</strong> drug delivery.<br />
Introduction<br />
Mesoporous silica (MS) materials have attracted considerable attention in the areas <strong>of</strong> chemistry,<br />
physics, biology, <strong>and</strong> materials science since their introduction in the early 1990s. 1 Owing to their high<br />
surface areas (up to ~1200 m 2 g -1 ), unique pore structures (in the range <strong>of</strong> 2-50 nm), <strong>and</strong> tunable<br />
particle morphology, MSs have been widely used to load <strong>and</strong> encapsulate various species, <strong>and</strong> for the<br />
template synthesis <strong>of</strong> diverse nanostructured materials. 2-4 Replication <strong>of</strong> MSs with metals, 5 metal<br />
oxides, 6 carbon, 7 <strong>and</strong> polymers 8 have been achieved by filling the nanopores with small molecule<br />
precursors such as metal alkoxides, sucrose, <strong>and</strong> organic monomers, followed by hydrolysis,<br />
carbonization, or polymerization <strong>of</strong> the precursors <strong>and</strong> removal <strong>of</strong> the template.<br />
Recently, we introduced a method to prepare nanoporous, polymer-based spheres (NPS) by<br />
templating MS spheres. This approach involves infiltration <strong>of</strong> preformed macromolecules (i.e.,<br />
polymers or proteins) into the nanopores <strong>of</strong> the MS particles, followed by crosslinking <strong>of</strong> the polymer<br />
chains, <strong>and</strong> removal <strong>of</strong> the MS particle templates. 9 The general applicability <strong>of</strong> this technique<br />
facilitates the design <strong>of</strong> polymeric particles with tunable properties by varying the macromolecules<br />
used, which can include synthetic polyelectrolytes, proteins, polypeptides, <strong>and</strong> polymer-drug<br />
conjugates. 10 Furthermore, this method permits tuning <strong>of</strong> the morphology <strong>and</strong> size <strong>of</strong> the nanoporous<br />
particles. Our previous studies have focused on infiltrating water-soluble macromolecules from<br />
aqueous solution to prepare particles that are dispersed in water. The preparation <strong>of</strong> similar but<br />
hydrophobic polymeric particles is also fundamentally interesting, as such particles can provide<br />
significantly different physicochemical properties to their hydrophilic counterparts. This may provide<br />
new opportunities for the loading <strong>and</strong> release <strong>of</strong> hydrophobic drugs, <strong>and</strong> for the protection <strong>and</strong><br />
stabilization <strong>of</strong> molecules <strong>and</strong> macromolecules that are sensitive to aqueous<br />
[1]. Published at Chemistry <strong>of</strong> Materials (2009) 21:4310-4315
environments. However, challenges associated with preparing hydrophobic particles from<br />
hydrophobic constituents include a lack <strong>of</strong> control over the size <strong>of</strong> the particles <strong>and</strong> difficulties in<br />
dispersing them in aqueous media for water-based applications.<br />
In this work, we report a facile method to prepare hydrophobic poly-3-hydroxybutyrate (PHB)<br />
particles through templating MS spheres. PHB is hydrophobic polyester that has similar physical<br />
properties to those <strong>of</strong> polypropylene. Compared with previous reports, a significant difference <strong>of</strong> the<br />
current work is that no crosslinking <strong>of</strong> the infiltrated polymer chains is required to obtain intact replica<br />
particles, owing to the hydrophobic PHB chains, which can strongly associate <strong>and</strong> maintain the<br />
template particle morphology after MS template removal in aqueous solution. Furthermore, PHB is<br />
biocompatible <strong>and</strong> biodegradable, <strong>and</strong> has been evaluated as a material for use in tissue engineering<br />
scaffolds <strong>and</strong> controlled drug-release carriers. 11 Recently, it was reported that PHB can also be<br />
coordinated with rare-earth complexes to construct luminescent films. 12 We demonstrate that the<br />
obtained PHB replicas can be effectively used to coordinate rare-earth complexes. The rare-earth<br />
compounds (e.g., rare-earth β-diketone complexes) have unique spectral characteristics, including<br />
sharp emission peaks with a small peak width at half-height, long lifetimes, high fluorescence<br />
quantum yields, <strong>and</strong> a large Stokes shifts, 13 which makes them <strong>of</strong> interest in clinical diagnostic assays,<br />
genomic screening, <strong>and</strong> fluorescence immunoassays. 14 However, coordination <strong>of</strong> −OH groups<br />
significantly reduces the luminescence emission intensity <strong>and</strong> decay time <strong>of</strong> the rare-earth complexes<br />
because <strong>of</strong> nonradiative dissipation <strong>of</strong> energy <strong>of</strong> the high-energy −OH vibrations. 15 This has largely<br />
limited the application <strong>of</strong> rare-earth complexes in biology <strong>and</strong> related fields, in which aqueous media<br />
are typically required. Hence, the design <strong>of</strong> rare-earth materials with both strong luminescence <strong>and</strong><br />
good stability in aqueous media is fundamentally important. 16 The rare-earth immobilized in the PHB<br />
replicas reported here can emit intense luminescence over a wide range <strong>of</strong> pH (3~11) <strong>and</strong> for at least<br />
several months in aqueous solution. The polyelectrolyte multilayer (PEM) coating (poly(allylamine<br />
hydrochloride) (PAH) <strong>and</strong> poly(sodium 4-styrenesulfonate) (PSS)) on the PHB replicas provides<br />
stable <strong>and</strong> dispersed particles in aqueous solution.<br />
Experimental Section<br />
Materials. The MS particles with a bimodal pore structure (particle diameter, 2-4 μm; pore<br />
diameters, 2-3 nm <strong>and</strong> 10-40 nm) were synthesized via the protocol reported by Schulz-Ekl<strong>of</strong>f et al. 17<br />
Poly(allylamine hydrochloride) (PAH, M w 70 000), poly(sodium 4-styrenesulfonate) (PSS, M w 70<br />
000), europium(III) chloride hexahydrate (EuCl 3·6H 2 O), 2-thenoyltrifluoroacetone (TTA), chlor<strong>of</strong>orm,<br />
<strong>and</strong> hydr<strong>of</strong>luoric acid (HF) were obtained from Sigma-Aldrich. Poly-3-hydroxybutyrate (M w 10 000)<br />
was purchased from Polysciences, Inc. The water used in all experiments was prepared in a three-stage<br />
Millipore Milli-Q Plus 185 purification system <strong>and</strong> had a resistivity greater than 18 MΩ·cm.<br />
Synthesis <strong>of</strong> the Rare-Earth Complex. The europium β-diketone complex, Eu(TTA) 3·2H 2 O<br />
(denoted as EuC), was synthesized according to published methods. 18 Briefly, 266 mg <strong>of</strong> TTA was<br />
dissolved in a solution <strong>of</strong> 6 mL ethanol <strong>and</strong> 1.2 mL <strong>of</strong> 1 M NaOH. Then, EuCl 3·6H 2 O (146 mg) in 42<br />
mL <strong>of</strong> water was added to the above solution, <strong>and</strong> the mixture was heated at 60 ºC for 30 min. The<br />
yellow complex <strong>of</strong> EuC precipitated during cooling the solution to room temperature. The precipitate<br />
was dried in vacuum after removal <strong>of</strong> the solution through filtration <strong>and</strong> water washing.<br />
PHB Replicas from MS Particles, Surface Modification <strong>and</strong> EuC Loading. PHB was firstly<br />
loaded in the MS particles. Approximately 5 mg <strong>of</strong> MS particles were dispersed in 0.5 mL <strong>of</strong> the PHB<br />
solution (concentration <strong>of</strong> 35 mg mL -1 ) in chlor<strong>of</strong>orm at room temperature for 16 h. After removing<br />
the supernatant with centrifugation (1500 g for 2 min), the PHB-loaded MS particles were dried in a<br />
vacuum desiccator to remove the chlor<strong>of</strong>orm. The PHB-loaded MS particles were dispersed in water<br />
with brief sonication. After removing the supernatant with centrifugation (1000 g for 2 min), 500 μL<br />
<strong>of</strong> 2.5 M HF was added into the PHB-loaded MS particles. PHB replicas were obtained after three<br />
washing with water. PHB replicas coated with PAH/PSS were also prepared. Before removal <strong>of</strong> the<br />
MS templates, one PAH/PSS bilayer was coated on the PHB-loaded MS particles using the layer-bylayer<br />
(LbL) technique. 19 The PHB replicas coated with PAH/PSS multilayers were obtained after<br />
dissolving the silica particles. Details <strong>of</strong> this process are given in Figure 1. For the EuC loading, the<br />
PAH/PSS-coated PHB replicas were then dispersed in an ethanol solution <strong>of</strong> EuC for 16 h. After<br />
removal <strong>of</strong> the supernatant with centrifugation (2000 g for 2 min), three washings with water were<br />
carried out to remove the unadsorbed EuC.<br />
81
Figure 1. Schematic representation <strong>of</strong> the formation <strong>of</strong> EuC-loaded PHB particles obtained by<br />
templating MS spheres. A polyelectrolyte multilayer (PEM) coating is used to stabilize the PHB<br />
particles, thus avoiding aggregation in water <strong>and</strong> to impart functional groups to the surface <strong>of</strong> the<br />
particles.<br />
Instruments. Thermogravimetric analysis (TGA) experiments were conducted on a Mettler<br />
Toledo/TGA/SDTA851e Module analyzer. The samples were heated from 25 to 120 ºC with a heating<br />
rate <strong>of</strong> 5 ºC per min <strong>and</strong> kept at 120 ºC for 20 min under nitrogen (30 mL min -1 ). Then, the samples<br />
were heated from 120 to 550 ºC with a heating rate <strong>of</strong> 10 ºC per min under oxygen (30 mL min -1 ). ζ-<br />
potential measurements were performed on a Malvern 2000 Zetasizer. Fourier transform infrared<br />
(FTIR) experiments were conducted on a Varian 7000 FT-IR spectrometer. Transmission electron<br />
microscopy (TEM, Philips CM120 BioTWIN, operated at 120 kV) <strong>and</strong> scanning electron microscopy<br />
(SEM, FEI Quanta 200 FEG, operated at 5 kV) were used to examine the morphology <strong>of</strong> the PHB<br />
replicas. The TEM samples (2 μL) were placed onto Formvar-coated copper grids <strong>and</strong> allowed to airdry.<br />
The SEM samples (2 μL) were placed onto silicon wafers <strong>and</strong> allowed to air-dry prior to gold<br />
sputter-coating. Luminescence emission spectra were recorded on a Horiba Yvon-Jacob FL3-22<br />
fluorimeter using a quartz cuvette (excitation wavelength, 350 nm; emission wavelength range, 500-<br />
670 nm; increment size, 1 nm; excitation <strong>and</strong> emission slit widths, 5 nm). 40 μL <strong>of</strong> EuC-doped PHB<br />
replicas coated with PAH/PSS (ca. 1.5 × 10 9 particles mL -1 ) was dispersed in 160 μL <strong>of</strong> 5 mM<br />
phosphate buffer solution <strong>of</strong> different pH for the luminescence measurements. Fluorescence<br />
microscopy images were taken with an Olympus IX71 inverted fluorescence microscope. The samples<br />
(1.5 μL) were placed onto glass slides <strong>and</strong> viewed using a 60× oil immersion objective.<br />
Results <strong>and</strong> Discussion<br />
As illustrated in Figure 1, PHB was first infiltrated into the nanopores <strong>of</strong> the MS particles in<br />
chlor<strong>of</strong>orm solution. After removal <strong>of</strong> the chlor<strong>of</strong>orm solvent, the PHB-loaded MS particles were<br />
dispersed in water by brief sonication. TEM images reveal that the PHB-loaded MS particles (Figure<br />
2a) have the same morphology <strong>and</strong> diameter (2-4 μm) as the pristine MS template particles (data not<br />
shown). No signs <strong>of</strong> PHB clustering were observed either in the solution or on the MS particle surface.<br />
TGA was employed to estimate the PHB loading in the MS particles. The thermogravimetric analysis<br />
<strong>of</strong> the PHB-loaded particles showed a mass loss <strong>of</strong> 31.5% at 215 ºC (Figure S1). Hence, it can be<br />
estimated that 1 g <strong>of</strong> MS particles can load ca. 0.46 g <strong>of</strong> PHB. Based on the TEM <strong>and</strong> TGA data, we<br />
82
conclude that the hydrophobic PHB can be effectively loaded in the nanopores <strong>of</strong> the MS particles<br />
from organic solvent (chlor<strong>of</strong>orm).<br />
Figure 2. (a) TEM image <strong>of</strong> PHB-loaded MS particles. (b) Optical micrograph <strong>of</strong> PHB replicas coated<br />
with PAH/PSS dispersed in water. (c) TEM images <strong>of</strong> PHB replicas coated with one PAH/PSS bilayer.<br />
The inset is a higher magnification image. (d) SEM images <strong>of</strong> the PHB replica coated with one<br />
PAH/PSS bilayer. A higher magnification image <strong>of</strong> the surface <strong>of</strong> the replica is shown in the inset.<br />
PHB replicas were obtained after removal <strong>of</strong> the MS template particles, indicating that the van der<br />
Waals interaction among the hydrophobic polymer chains plays a significant role in maintaining the<br />
morphology <strong>of</strong> the particles in aqueous media. This is in contrast to our previous reports that use water<br />
soluble macromolecules, where cross linking <strong>of</strong> the polymer chains in the porous particles is required<br />
to obtain a stable polymer network in solution. 9,10 However, the replica PHB particles easily<br />
aggregated in water due to the hydrophobic nature <strong>of</strong> PHB, hampering their application as colloidal<br />
particles in aqueous media. To avoid such aggregation, a PEM (PAH/PSS) coating was assembled on<br />
the surface <strong>of</strong> the PHB-loaded MS particles before MS template removal. The PEM shell provides the<br />
dual role <strong>of</strong> stabilizing the PHB replicas from aggregation <strong>and</strong> functionalizing the particle surface.<br />
The polyelectrolyte coating was assembled through alternately depositing PAH <strong>and</strong> PSS. ζ-potential<br />
measurements were used to monitor the variation in surface charge <strong>of</strong> the MS particles after PHB<br />
infiltration <strong>and</strong> polyelectrolyte coating (Figure S2). The PHB-loaded MS particles had a ζ−potential <strong>of</strong><br />
−25mV in water. After deposition <strong>of</strong> each layer <strong>of</strong> PAH <strong>and</strong> PSS, the ζ−potential <strong>of</strong> the particles<br />
alternated between ca. 30 mV (corresponding to PAH deposition) <strong>and</strong> ca. −40 mV (corresponding to<br />
PSS deposition), indicating the consecutive deposition <strong>of</strong> PAH <strong>and</strong> PSS, respectively.<br />
83
Figure 3. (a) Transmission <strong>and</strong> (b) fluorescence microscopy images <strong>of</strong> EuC-doped PHB replicas<br />
coated with one PAH/PSS bilayer. The insets in (a) <strong>and</strong> (b) are photographs <strong>of</strong> the particles after<br />
centrifugation in water <strong>and</strong> exposed to sunlight <strong>and</strong> ultraviolet light, respectively. The EuC incubating<br />
concentration was 0.5 mg mL -1 .<br />
After removal <strong>of</strong> the silica template, PAH/PSS-coated PHB replicas readily dispersed in water, as<br />
observed by optical microscopy (Figure 2b). The particles have a diameter <strong>of</strong> 2-4 μm, similar to the<br />
size <strong>of</strong> the MS template spheres. Figure 2c shows a TEM image <strong>of</strong> the PHB replicas. The diameter <strong>of</strong><br />
the PHB replicas decreased to 1-2 μm <strong>and</strong> a rough particle surface was observed under high<br />
magnification conditions (insert in Figure 2c). Figure 2d shows the SEM images <strong>of</strong> the PHB replicas<br />
coated with PAH/PSS, which is a free-st<strong>and</strong>ing, noncollapsed sphere. The porosity on the surface is<br />
apparent at higher magnification (Figure 2d inset). This noncollapsed structure is in contrast to what is<br />
typically observed for polyelectrolyte capsules prepared via the sequential assembly <strong>of</strong> polymers on<br />
the surface <strong>of</strong> MS particles without PHB loading, where collapsed, hollow structures are observed. 20<br />
FTIR was also used to monitor the composition <strong>of</strong> the particles. For the PHB-loaded MS particles, a<br />
strong <strong>and</strong> broad absorption b<strong>and</strong> at 1065 cm -1 was observed (Figure S3). The b<strong>and</strong> can be assigned to<br />
the Si-O stretching vibration. The peak at 1723 cm -1 is attributed to the C=O stretching vibration <strong>of</strong><br />
PHB. 21 No absorption b<strong>and</strong> at 1065 cm -1 was observed for the replicas, suggesting complete removal<br />
<strong>of</strong> the silica. The spectrum <strong>of</strong> the PAH/PSS-coated PHB replicas is similar to the spectrum <strong>of</strong> pure<br />
PHB. The absorption peak at 1007 cm -1 originates from the SO 3 - symmetric vibration <strong>of</strong> PSS. 22<br />
Luminescent materials are widely employed in fields as diverse as biology <strong>and</strong> information<br />
technology. Commonly sought properties for luminescent materials are brightness <strong>and</strong> stability. 23 EuC<br />
has emerged as a novel luminescent material because <strong>of</strong> its unique spectral characteristics, including<br />
sharp emission peaks, long lifetimes, high fluorescence quantum yields, <strong>and</strong> a large Stokes shift.<br />
However, the luminescence <strong>of</strong> EuC is not stable in aqueous solution because coordination <strong>of</strong> −OH<br />
groups extensively reduces its luminescence emission intensity <strong>and</strong> decay time. Our aim was to<br />
immobilize the EuC into the PHB replicas to construct particles with stable <strong>and</strong> intense luminescence.<br />
84
Figure 4. (a) Energy transfer model from the lig<strong>and</strong> to the Eu III luminescent center in a PHB replica<br />
doped with EuC. PHB serves as an efficient cosensitizer for the Eu III ion. (b) Luminescence emission<br />
spectra <strong>of</strong> EuC-doped PEM capsules composed <strong>of</strong> three bilayers <strong>of</strong> PAH/PSS (black line) <strong>and</strong> PHB<br />
replicas coated with one PAH/PSS bilayer (red line) dispersed in water (λ excitation = 350 nm). The EuC<br />
incubating concentration was 0.05 mg mL -1 .<br />
The EuC-doped PHB replicas coated with PAH/PSS were visualized in situ by fluorescence<br />
microcopy (Figure 3). From the images, the replicas are observed to be spherical with a diameter <strong>of</strong> 2-<br />
4 μm, <strong>and</strong> no aggregation was observed in both transmission <strong>and</strong> fluorescence modes. The intact<br />
replicas show no shrinkage or swelling after loading EuC in ethanol, suggesting the PHB replicas have<br />
good stability in both aqueous <strong>and</strong> alcohol solutions. The bright luminescent spheres confirm the<br />
effective loading <strong>of</strong> EuC in the replicas (Figure 3b). The insets in Figure 3a <strong>and</strong> 3b show the images <strong>of</strong><br />
the EuC-doped PHB replicas coated with PAH/PSS in water after centrifugation <strong>and</strong> exposed to<br />
sunlight <strong>and</strong> ultraviolet light, respectively. The intense monochromatic red luminescence <strong>of</strong> the EuCdoped<br />
replicas is easy to detect by a laboratory ultraviolet lamp at 365 nm (Figure 3b inset).<br />
The intense luminescence observed from the EuC-doped PHB replicas originates from the loaded<br />
EuC. Figure 4a displays an energy transfer model for the lig<strong>and</strong>s, PHB, <strong>and</strong> the central rare-earth ions<br />
in the replica. Firstly, the β-diketone lig<strong>and</strong> TTA is excited by ultraviolet light. Efficient<br />
intermolecular energy transfer from the lig<strong>and</strong> to the luminescent center <strong>and</strong> subsequent highly<br />
efficient emission from the excited state <strong>of</strong> Eu III to the low level state endow the EuC-doped replicas<br />
with luminescent properties. This is the so-called “antenna effect”. 24 A detailed energy level diagram<br />
for Eu III along with the singlet <strong>and</strong> triplet levels <strong>of</strong> TTA is given in Figure S4. Here, the PHB also<br />
serves as an efficient cosensitizer for the Eu III ions. 12<br />
85
Figure 4b shows the luminescence emission spectrum <strong>of</strong> EuC-doped PHB replicas coated with<br />
PAH/PSS There are four characteristic peaks (580 nm, 595 nm, 613 nm <strong>and</strong> 650 nm), corresponding<br />
to the electric dipole transitions <strong>of</strong> 5 D 0 → 7 F J (J = 0, 1, 2 <strong>and</strong> 3), which is similar to that <strong>of</strong> EuC in<br />
ethanol solution. The red luminescence observed is mainly attributed to the 5 D 0 → 7 F 0 transition<br />
around 613 nm.<br />
25<br />
Intensity (/10 6 ) / a.u.<br />
20<br />
15<br />
10<br />
5<br />
0<br />
0.00 0.05 0.10 0.15 0.20 0.25 0.30<br />
C EuC<br />
/ mg mL -1<br />
Figure 5. Luminescence intensity <strong>of</strong> EuC-doped PHB replicas coated with one PAH/PSS bilayer<br />
dispersed in water as a function <strong>of</strong> the EuC incubating concentration. The luminescence intensity was<br />
recorded at 613 nm (λ excitation = 350 nm).<br />
The EuC content in the PHB particles was studied at different EuC incubating concentrations. PHB<br />
replicas coated with PAH/PSS were dispersed in ethanol solutions <strong>of</strong> EuC <strong>of</strong> different concentration<br />
for 16 h. After removal <strong>of</strong> the ethanol supernatant, <strong>and</strong> three water washing cycles, the luminescence<br />
intensity <strong>of</strong> the particles was measured in water. Figure 5 plots the relationship between the<br />
luminescence intensity <strong>and</strong> the EuC incubating concentration. The luminescence intensity linearly<br />
increases at EuC incubating concentrations in the range 0.01−0.05 mg mL -1 . The luminescence<br />
intensity <strong>of</strong> the PHB replicas obtained using 0.05 mg mL -1 EuC incubating solutions is equivalent to<br />
that <strong>of</strong> neat EuC in bulk ethanol with a concentration <strong>of</strong> 6.05 × 10 -3 mg mL -1 . According to the TGA<br />
results, 1 g <strong>of</strong> MS can adsorb 0.46 g <strong>of</strong> PHB, <strong>and</strong> hence the EuC doping is ca. 0.5% (EuC/PHB, w/w).<br />
At this incubating concentration, the luminescence intensity <strong>of</strong> the EuC-doped PHB replicas is more<br />
than 20 times higher than that <strong>of</strong> the EuC-doped PAH/PSS multilayer capsules (3 bilayers) templated<br />
by the same MS particles (Figure 4b). With further increasing the EuC concentration, the<br />
luminescence intensity increases only marginally. This is likely caused by concentration quenching in<br />
the EuC-doped particles.<br />
86
20<br />
Intensity (/10 6 ) / a.u.<br />
15<br />
10<br />
5<br />
0<br />
0 2 4 6 8 10 12 14<br />
pH<br />
Figure 6. Luminescence intensity <strong>of</strong> EuC-doped PHB replicas coated with one PAH/PSS bilayer <strong>and</strong><br />
dispersed in a 5 mM phosphate buffer solution <strong>of</strong> different pH. The EuC incubating concentration was<br />
0.05 mg mL -1 . The luminescence intensity was recorded at 613 nm (λ excitation = 350 nm).<br />
The stability <strong>of</strong> EuC in the PHB particles was also investigated. The saturated complexes can<br />
effectively preserve the rare-earth ions from being attacked by −OH groups in solution. 25 Here, PHB is<br />
used to coordinate the EuC to form saturated EuC complexes in the PHB particles. Figure 4a<br />
illustrates the EuC coordinated to the PHB backbone <strong>and</strong> the lig<strong>and</strong>s. The formation <strong>of</strong> saturated<br />
complexes in the hydrophobic PHB particles protects <strong>and</strong> stabilizes the EuC in aqueous solution.<br />
Figure 6 shows the luminescence intensity <strong>of</strong> EuC-doped PHB replicas coated with PAH/PSS in<br />
phosphate buffer solution <strong>of</strong> different pH. EuC-doped PHB replicas can emit intense luminescence<br />
under ultraviolet radiation over a wide pH <strong>of</strong> 3~11. No measurable decrease <strong>of</strong> the luminescence <strong>of</strong> the<br />
EuC-doped PHB particles was observed after several months. However, the luminescence <strong>of</strong> EuCdoped<br />
PEM (e.g., PAH/PSS) capsules dispersed in water becomes weaker <strong>and</strong> weaker, suggesting that<br />
the PHB stabilizes the EuC through coordination. As the EuC is water insoluble, uncharged <strong>and</strong> has a<br />
low molecular weight, it has similar physical properties to some hydrophobic drugs. Therefore, the<br />
PHB replicas are interesting systems for the loading <strong>of</strong> hydrophobic (e.g., anticancer) drugs. This is an<br />
important property for the EuC-doped PHB replicas to find application in various applications.<br />
Conclusions<br />
We have demonstrated the preparation <strong>of</strong> PHB replicas through templating MS particles, <strong>and</strong> their<br />
subsequent surface coating with PEMs <strong>and</strong> doping with rare-earth complexes to obtain bright <strong>and</strong><br />
stable luminescent PHB particles dispersed in aqueous media. Hydrophobic PHB was effectively<br />
assembled in MS particles <strong>and</strong> intact PHB replicas are obtained after removal <strong>of</strong> the silica template.<br />
The broad range <strong>of</strong> MS materials available with tunable size, morphology <strong>and</strong> porosity will enable the<br />
preparation <strong>of</strong> PHB materials with tailored dimensions <strong>and</strong> morphologies. The surface properties <strong>of</strong><br />
the PHB replicas can be easily tuned by assembly <strong>of</strong> a PEM coating. The PEM provides the dual role<br />
<strong>of</strong> preventing aggregation <strong>of</strong> the PHB replicas in aqueous solution <strong>and</strong> also imparting functionality<br />
through the polymer coating. The biodegradable PHB replicas can be effectively used to coordinate<br />
rare-earth complexes. The EuC-doped PHB replicas emit intense luminescence over a wide range <strong>of</strong><br />
pH (3~11) <strong>and</strong> for at least several months in aqueous solution, making them potentially useful for a<br />
range <strong>of</strong> bioapplications.<br />
Acknowledgements: This work was supported by the Australian Research Council under the<br />
Federation Fellowship <strong>and</strong> Discovery Project schemes. Jiwei Cui acknowledges the China Scholarship<br />
Council for a Joint-Education PhD Program Scholarship. Philipp Senn <strong>and</strong> Cameron R. Kinnane are<br />
thanked for the SEM measurements.<br />
87
Supporting <strong>Information</strong> Available: Thermogravimetric analysis <strong>of</strong> PHB-loaded MS particles; ζ-<br />
potential <strong>of</strong> the PAH/PSS-coated PHB particles as a function <strong>of</strong> the number <strong>of</strong> polyelectrolyte layers<br />
deposited; FTIR spectra <strong>of</strong> PHB-loaded MS, PHB replicas coated with PAH/PSS, <strong>and</strong> neat PHB; <strong>and</strong><br />
energy level diagram for Eu III <strong>and</strong> the singlet <strong>and</strong> triplet levels <strong>of</strong> TTA.<br />
References:<br />
(1) (a) Kresge, T.; Leonowicz, M. E.; Roth, W. J.; Vartuli, J. C.; Beck, J. S. Nature 1992, 359, 710-<br />
712.<br />
(2) For reviews see: (a) Stein, A. Adv. Mater. 2003, 15, 763-775. (b) Yang, H.; Zhao, D. J. Mater.<br />
Chem. 2005, 15, 1217-1231. (c) Hartmann, M. Chem. Mater. 2005, 17, 4577-4593. (d) Trewyn,<br />
B. G.; Slowing, I. I.; Giri, S.; Chen, H.-T.; Lin, V. S.-Y. Accounts Chem. Res. 2007, 40, 846-853.<br />
(e) Balas, Vallet-Regi, F.; Arcos, D. Angew. Chem. Int. Ed. 2007, 46, 7548-7558. (f) Wang, Y.;<br />
Angelatos, A. S.; Caruso, F. Chem. Mater. 2008, 20, 848-858. (g) Wang, Y.; Price, A. D.; Caruso,<br />
F. J. Mater. Chem. 2009, DOI: 10.1039/b901742a.<br />
(3) (a) Wang, Y.; Caruso, F. Chem. Commun. 2004, 1528-1529. (b) Wang, Y.; Caruso, F. Chem.<br />
Mater. 2005, 17, 953-961. (c) Deng, Y.; Qi, D.; Deng, C.; Zhang, X.; Zhao, D. J. Am. Chem. Soc.<br />
2008, 130, 28-29. (d) Price, A. D.; Zelikin, A. N.; Wang, Y.; Caruso, F. Angew. Chem. Int. Ed.<br />
2009, 48, 329-332. (e) Johnston, A. P. R.; Lee, L.; Wang, Y.; Caruso, F. Small 2009, 5, 1418-<br />
1421.<br />
(4) (a) Yu, C.; Fan, J.; Tian, B.; Zhao, D.; Stucky, G. D.; Adv. Mater. 2002, 14, 1742-1745. (b) Dong,<br />
A.; Ren, N.; Tang, Y.; Wang, Y.; Zhang, Y.; Hua, W.; Gao, Z. J. Am. Chem. Soc. 2003, 125,<br />
4976-4977. (c) Wang, Y.; Caruso, F. Adv. Funct. Mater. 2004, 14, 1012-1018.<br />
(5) Crowley, T. A.; Ziegler, K. J.; Lyons, D. M.; Erts, D.; Olin, H.; Morris, M. A.; Holmes, J. D.<br />
Chem. Mater. 2003, 15, 3518-3522.<br />
(6) Tian, B.; Liu, X.; Yang, H.; Xie, S.; Yu, C.; Tu, B.; Zhao, D. Adv. Mater. 2003, 15, 1370-1374.<br />
(7) Ryoo, R.; Joo, S. H.; Kruk, M.; Jaroniec, M. Adv. Mater. 2001, 13, 677-681.<br />
(8) (a) Kageyama, K.; Tamazawa, J.-I.; Aida, A. Science 1999, 285, 2113-2115. (b) Yilmaz, E.;<br />
Ramsrtöm, O.; Möller, P.; Sanchez, D.; Mosbach, K. J. Mater. Chem. 2002, 12, 1577-1581.<br />
(9) Wang, Y.; Yu, A.; Caruso, F. Angew. Chem. Int. Ed. 2005, 44, 2888-2892.<br />
(10) (a) Wang, Y.; Caruso, F. Adv. Mater. 2006, 18, 795-800. (b) Wang, Y.; Caruso, F. Chem. Mater.<br />
2006, 18, 4089-4100. (c) Wang, Y.; Angelatos, A. S.; Dunstan, D. E.; Caruso, F.<br />
Macromolecules 2007, 40, 7594-7600. (d) Wang, Y.; Bansal, V.; Zelikin, A. Z.; Caruso, F. Nano<br />
Lett. 2008, 8, 1741-1745.<br />
(11) Köse, G. T.; Ber, S.; Korkusuz, F.; Hasirci, V. J. Mater. Sci. - Mater. Med. 2003, 14, 121-126.<br />
(12) Kai, J.; Parra, D. F.; Brito, H. F. J. Mater. Chem. 2008, 18, 4549-4554.<br />
(13) (a) Bekiari, V.; Lianos, P. Adv. Mater. 1998, 10, 1455-1458. (b) Cui, J.-W.; Zhang, R.-J.; Lin, Z.-<br />
G.; Li, L.; Jin, W.-R. Dalton Trans. 2008, 895-899.<br />
(14) (a) Mathis, G. J. Biomol. Screen 1999, 4, 309-313. (b) Mathis, G. Clin. Chem. 1993, 39, 1953-<br />
1959. (c) Saha, A. K.; Kross, K.; Kloszewski, E. D.; Upson, D. A.; Toner, J. L.; Snow, R. A.;<br />
Black, C. D. V.; Desai, V. C. J. Am. Chem. Soc. 1993, 115, 11032-11033.<br />
(15) (a) Maas, H.; Currao, A.; Calzaferri, G. Angew. Chem. Int. Ed. 2002, 41, 2495-2497. (b) Petoud,<br />
S.; Cohen, S. M.; Bünzli, J.-C. G.; Raymond, K. N. J. Am. Chem. Soc. 2003, 125, 13324-13325.<br />
(16) Selvin, P. R. Nat. Struct. Biol. 2000, 7, 730-734.<br />
(17) Schulz-Ekl<strong>of</strong>f, G.; Rathouský, J.; Zukal, A. Int. J. Inorg. Mater. 1999, 1, 97-102.<br />
(18) (a) Melby, L. R.; Rose, N. J.; Abramson, E.; Caris, J. C. J. Am. Chem. Soc. 1964, 86, 5117-5125.<br />
(b) Charles, R. G.; Ohlmann, R. C. J. Inorg. Nucl. Chem. 1965, 27, 255-259.<br />
(19) (a) Decher, G.; Hong, J. D. Ber. Bunsen-Ges. Phys. Chem. 1991, 95, 1430-1433. (b) Decher, G.<br />
Science 1997, 277, 1232-1237. (c) Donath, E.; Sukhorukov, G. B.; Caruso, F.; Davis, S. A.;<br />
Möhwald, H. Angew. Chem. Int. Ed. 1998, 37, 2201-2205. (d) Cui, J.; Liu, Y.; Hao, J. J. Phys.<br />
Chem. C 2009, 113, 3967-3972.<br />
(20) Yu, A.; Wang, Y. Barlow, E.; Caruso, F. Adv. Mater. 2005, 17, 1737-1741.<br />
(21) (a) Li, X.; Loh, X. J.; Wang, K.; He, C.; Li, J. Biomacromolecules 2005, 6, 2740-2747. (b) Zhang,<br />
J.; Sato, H.; Furukawa, T.; Tsuji, H.; Noda, I.; Ozaki, Y. J. Phys. Chem. B 2006, 110, 24463-<br />
24471.<br />
88
(22) (a) Cho, J.; Quinn, J. F.; Caruso, F. J. Am. Chem. Soc. 2004, 126, 2270-2271. (b) Caruso, F.;<br />
Furlong, D. N.; Ariga, K.; Ichinose, I.; Kunitake, T. Langmuir 1998, 14, 4559-4565.<br />
(23) Burns, A.; Ow, H.; Wiesner, U. Chem. Soc. Rev. 2006, 35, 1028-1042.<br />
(24) (a) Sabbatini, N.; Guardigli, M.; Lehn, J.-M. Coord. Chem. Rev.1993, 123, 201-228. (b) Bredol,<br />
M.; Kynast, U.; Ronda, C. Adv. Mater. 1991, 3, 361-367.<br />
(25) Gago, S.; Fern<strong>and</strong>es, J. A.; Rainho, J. P.; Sá Ferreira, R. A.; Pillinger, M.; Valente, A. A.; Santos,<br />
T. M.; Carlos, L. D.; Ribeiro-Claro, P. J. A.; Gonçalves, I. S. Chem. Mater. 2005, 17, 5077-5084.<br />
Supporting <strong>Information</strong>:<br />
100<br />
95<br />
Mass loss / %<br />
90<br />
85<br />
80<br />
75<br />
70<br />
31.5 %<br />
65<br />
0 100 200 300 400 500 600<br />
Temperature / o C<br />
Figure S1. Thermogravimetric analysis <strong>of</strong> PHB-loaded MS particles.<br />
40<br />
PAH<br />
PAH<br />
ζ-potential / mV<br />
20<br />
0<br />
-20<br />
-40<br />
PHB<br />
PAH<br />
PSS<br />
PSS<br />
PSS<br />
1 2 3 4 5 6 7<br />
Layer number<br />
Figure S2. ζ-potential <strong>of</strong> the PAH/PSS-coated PHB particles as a function <strong>of</strong> the number <strong>of</strong><br />
polyelectrolyte layers deposited.<br />
89
a<br />
b<br />
Absorbance<br />
c<br />
1800<br />
1650<br />
1500<br />
1350<br />
1200<br />
1050<br />
900<br />
Wavenumber / cm -1<br />
Figure S3. FTIR spectra <strong>of</strong> (a) PHB-loaded MS, (b) PHB replicas coated with PAH/PSS, <strong>and</strong> (c) neat<br />
PHB.<br />
S 1 5D 1<br />
S 1<br />
T 1<br />
5D 1<br />
5D 0<br />
7F 2<br />
S 0<br />
TTA<br />
Eu Ⅲ<br />
7F 0<br />
Figure S4. Energy level diagram for Eu III <strong>and</strong> the singlet <strong>and</strong> triplet levels <strong>of</strong> TTA.<br />
90
Grain refinement <strong>of</strong> pure magnesium by back pressure equal channel<br />
angular pressing at room temperature<br />
Ji-zhong Li 1 *<br />
* Presenter<br />
1. School <strong>of</strong> Materials <strong>and</strong> Metallurgy, Northeastern University, Shanyang 110004, China<br />
Systematic investigation on feasibility <strong>of</strong> refining grain structure in commercial pure<br />
magnesium was conducted by means <strong>of</strong> back pressure equal channel angular processing (BP-<br />
ECAP) at temperatures from 200 °C to room temperature. It was clearly shown that the grain<br />
refinement <strong>of</strong> commercial pure Mg casting ingot can be successfully achieved from greater<br />
than 800 µm down to submicron scale with 90° channel die, 100 MPa back pressure <strong>and</strong><br />
2mm/min pressing speed at room temperature. The temperature effect on the microstructure<br />
<strong>and</strong> mechanical behaviour in compression was discussed.<br />
Keywords Back Pressure, Equal Channel Angular Pressing, Pure Magnesium, Mechanical<br />
properties, Twinning<br />
Introduction<br />
Materials with ultra-fine <strong>and</strong> nano-scaled grain structures have attracted growing interest during<br />
the past decade because <strong>of</strong> their unique mechanical properties in comparison <strong>of</strong> their coarse grained<br />
ones [1-4]. A popular method to produce such a material is severe plastic deformation (SPD) as grain<br />
refinement <strong>of</strong> a metallic material can be achieved through large amount <strong>of</strong> plastic deformation.<br />
Although there are several SPD techniques, such as equal channel angular processing (ECAP), high<br />
pressure torsion (HPT) <strong>and</strong> accumulated roll bonding, ECAP is an effective practice to refine large<br />
grains <strong>of</strong> a bulk metallic material into ultra-fine or nano-scaled level though one or multi-pass in an<br />
angular die [5, 6]. To date there are significant ECAP works on materials with the fcc <strong>and</strong> bcc<br />
structures [7-10], but limited researches on hcp structural materials that have been claimed as hard-todeformed<br />
metals, such as Mg <strong>and</strong> Ti [11,12].<br />
Mg <strong>and</strong> its alloys, having the low density <strong>and</strong> workability, are very attractive for automotive,<br />
aviation <strong>and</strong> electronic industries [13, 14]. However, their applications have been restricted because <strong>of</strong><br />
their low ductility determined by the hcp structure, in which only a limited number <strong>of</strong> slip systems<br />
exists. To improve their ductility while retaining or enhancing their strength, refining their gain<br />
structures has been an objective in recent studies. ECAP has recently been effectively used in refining<br />
grains <strong>of</strong> Mg <strong>and</strong> its alloys with enhanced ductility, strength <strong>and</strong> super plasticity [5-6, 15-18].<br />
However, in the most works, the processing temperature used was at or higher than 200 ˚C. Otherwise,<br />
during ECAP processing the material cracked into many segments along shearing plan which was<br />
about 45˚ to the longitudinal direction [11, 12]. At such a temperature, grain size could not be refined<br />
smaller than 5 µm as recrystallisation <strong>and</strong> grain growth occurred. By means <strong>of</strong> BP-ECAP, submicron<br />
grain size in AZ31 was obtained after 4 passes at 150 ˚C plus another 4 passes at 100 ˚C following the<br />
Bc route [6]. In the current work, refinement <strong>of</strong> gain structures in commercial purity Mg was initiated<br />
using BP-ECAP at a range <strong>of</strong> temperatures from 200 ˚C to room temperature. For comparison purpose<br />
a specimen was also ECAP pressed at 250 °C without back pressure. The microstructures <strong>and</strong> the<br />
mechanical properties at room temperature were also investigated.<br />
Experimental procedure<br />
Specimens <strong>of</strong> 9 mm diameter with a length <strong>of</strong> 70 mm were machined from a commercial-purity<br />
Mg ingot (purity >99.5%). ECAP die was consisting <strong>of</strong> two round channels with the intersect-angle <strong>of</strong><br />
90˚. The die was heated to a pressing temperature after a specimen was inserted into the entrance<br />
channel. Graphite lubricant was used between the specimen <strong>and</strong> the channel surface. During ECAP<br />
processing, a back pressure <strong>of</strong> 100 MPa was applied <strong>and</strong> low speed <strong>of</strong> 2 mm/min was used. Each<br />
specimen was pressed for 4 passes with route C at a given temperature.<br />
91
Microstructures in the as-cast <strong>and</strong> BP-ECAPed materials were observed using optical<br />
microscopy (OM) <strong>and</strong> scanning electron microscopy (SEM). The samples for OM <strong>and</strong> SEM were cut<br />
perpendicular to the longitudinal direction <strong>and</strong> prepared by st<strong>and</strong>ard metallographic procedures<br />
followed by chemical etching using a solution <strong>of</strong> 10 ml nitric acid, 30 ml acetic acid, 40 ml distilled<br />
water <strong>and</strong> 120 ml ethanol. Average grain sizes <strong>of</strong> the as-cast <strong>and</strong> BP-ECAPed materials were obtained<br />
following the linear method.<br />
Compression specimens <strong>of</strong> 3 mm x 3 mm x 6 mm were cut along the longitudinal direction <strong>of</strong><br />
the BP-ECAPed samples. Compression tests were performed at room temperature with an initial strain<br />
rate <strong>of</strong> 1×10 -3 s -1 . For each material at least three compression tests were performed. Microstructures on<br />
the samples after the compression tests were also observed along cross section.<br />
Results <strong>and</strong> discussion<br />
Fig. 1 showed that the majority grains in the as-cast Mg were equiaxed <strong>and</strong> very coarse with<br />
irregular morphology. The average grain size was larger than 800 µm. Fig. 2 (a) displayed the<br />
appearances <strong>of</strong> a sample after 1 pass through the die at 250 °C without back pressure. Fracture<br />
segments distributed uniformly along the sample <strong>and</strong> the cracks went though the bottom to top along<br />
shearing direction which was about 45° to the extruded direction. In contrast, all the samples after BP-<br />
ECAP at temperatures ranging from 200 °C down to room temperature with 23 °C interval showed no<br />
cracks at all. Fig. 2 (b) exhibited the sample after 4 passes at room temperature. It demonstrated that<br />
applying back pressure was significantly helpful in ECAP processing <strong>of</strong> hard-to-deform metals at a<br />
relatively low temperature or even at room temperature as an appropriate back pressure can effectively<br />
prevent the material from cracking during ECAP processing.<br />
Fig.1 Optical microstructure in pure cast ingot<br />
Fig. 2 Appearance <strong>of</strong> pure Mg billets: (a) 250 ºC<br />
without back pressure; (b) 23 ºC withbackpressure<br />
Fig. 3 (a) through (c) presented the typical microstructures under OM for the materials after 4<br />
passes at temperatures <strong>of</strong> 200, 150 <strong>and</strong> 100 °C, respectively, <strong>and</strong> Fig. 3 (d) revealed the representative<br />
microstructure under TEM for the material after 4 passes at room temperature. The grain refinement<br />
was significant. Especially, the grain size was refined into submicron level when the BP-ECAP<br />
process was carried out at room temperature. The average grain sizes <strong>of</strong> the materials were listed in<br />
Table 1.<br />
92
Fig. 3 The optical <strong>and</strong> SEM microstructure <strong>of</strong> pure Mg: (a) 200 ºC; (b) 150 ºC; (c) 100 ºC; (d) room<br />
temperature (TEM).<br />
Table 1 average grain size <strong>of</strong> different temperature<br />
Temperature (°C) 200 150 100 23<br />
Average grain size (µm) ~8 ~6 ~2 ~0.8<br />
Another clear feature was that twining existed in both as-cast <strong>and</strong> the materials BP-ECAPed at<br />
200 °C with the grain size <strong>of</strong> 10 µm. But no twining was detected when the materials BP-ECAPed at<br />
the temperature lower than 200 °C <strong>and</strong> having the grain size <strong>of</strong> 5 µm or less. Mechanism behind this<br />
phenomenon was not clear yet <strong>and</strong> further investigation was necessary.<br />
Fig. 4 showed the true stress-strain curves <strong>of</strong> the materials. It can be seen that the as-cast<br />
material showed almost no ductility <strong>and</strong> the low yield strength (0.2% pro<strong>of</strong> stress) <strong>of</strong> about 40 MPa.<br />
The yield strength was increased as the grain size refined into 5 to 1 µm <strong>and</strong> then decreased when<br />
grain size was within submicron level. The material after 4 passes at 200 °C displayed good ductility<br />
<strong>and</strong> highest ultimate strength among those materials. Both the as-cast <strong>and</strong> the 200 °C processed<br />
materials had high strain hardening rate. On the other h<strong>and</strong>, the other BP-ECAPed materials showed<br />
almost no strain hardening. The compression behaviours <strong>of</strong> those materials also displayed obviously<br />
that the plastic deformation after yielding increased to about 15% as the grain size reduced to<br />
submicron scale. This indicated that grain boundary sliding <strong>and</strong> texture s<strong>of</strong>tening played dominant role<br />
<strong>and</strong> contributed to plastic deformation extensively.<br />
Another interesting point is that the material processed at 200 °C showed double yield strengths,<br />
as indicated by arrows. It can be explained that the first yielding was attributed to twins as the twins<br />
occurred only in the grains at low strain, <strong>and</strong> the second one was due to the combination effect <strong>of</strong><br />
twins <strong>and</strong> dislocation movement which had contribution to further strain hardening <strong>and</strong> some ductility.<br />
93
Fig. 4 The true stress-strain curves for cast state <strong>and</strong> ECAPed with 4 passes at different temperatures.<br />
Fig. 5 (a) though (c) showed the OM microstructures <strong>of</strong> the materials processed at 200, 150 <strong>and</strong><br />
100 °C after compression tests <strong>and</strong> Fig. 5 (d) exhibited the microstructure <strong>of</strong> the material processed at<br />
room temperature. Fig. 5 demonstrated that there were no twins when the grain size was less than 5<br />
µm, but some twins existed in the grains about 5 µm or larger, as in Fig. 5 (b). Combining this feature<br />
with the compression behaviours in Fig. 4, it might conclude that strain hardening was, at least<br />
partially, attributed to mechanical twining. In addition, the high yield stress <strong>and</strong> low ductility for the<br />
materials ECAPed at 150 <strong>and</strong> 100 ˚C may due to formation <strong>of</strong> texture during the BP-ECAP. Cisar [10]<br />
<strong>and</strong> Kim [19] etc have reported that ECAP processing makes basal plan parallel to extrusion direction<br />
results in Schmid factor closed to 0, which is also makes yield strength increase <strong>and</strong> ductility decrease.<br />
Fig. 5 The microstructure <strong>of</strong> compressive deformed samples: (a) 200 ºC; (b) 150 ºC; (c) 100 ºC; (e)<br />
room temperature (SEM).<br />
94
Conclusions<br />
(1) The large grains in the as-cast pure Mg were successfully refined into submicron sized grains<br />
using BP-ECAP at room temperature without cracks.<br />
(2) The compressive tests at room temperature showed that the yield strength might decrease as the<br />
grain size was reduced into submicron level. However, the ductility was significantly enhanced<br />
due to submicron grain size.<br />
(3) Strong strain hardening was attributed to large grain sizes, which was larger than 10 µm.<br />
Deformation twining, might contribute to the yield strength for the pure Mg materials having<br />
grains larger than 5 µm.<br />
References:<br />
[1] Y.T. Zhu, J.Y. Huang, J. Gubicza, T. Ungár, Y.M. Wang, E. Ma, R.Z. Valiev, J. Mater. Res. 18<br />
(2003) 1908.<br />
[2] D. Jia, Y.M. Wang, K.T. Ramesh, E. Ma, Y.T. Zhu, R.Z. Valiev, Appl. Phys. Lett. 79 (2001) 611.<br />
[3] R. Valiev, Nature Mater. 3 (2004) 511.<br />
[4] Z. Horita, K. Ohashi, T. Fujita, K. Kaneko, T.G. Langdon, Adv. Mater. 17 (2005) 1599.<br />
[5] K. Matsubara, Y. Miyahara, Z. Horita, T. G. Langdon. Acta Mater. 51 (2003) 3073.<br />
[6] K. Xia , J. T. Wang, X. Wu, G. Chen, M. Gurvan. Mater. Sci. Eng. A 410-411 (2005)324.<br />
[7] C. Mallikarjuna, S. M. Shashidhara, U. S. Mallik. Mater. Design. 30 (2009) 1638.<br />
[8] P. L. Sun, P. W. Kao, C. P. Chang. Mater. Sci. Eng. A. 283 (2000) 82.<br />
[9] M. Furukawa, Y. Ma, Z. Horita, M. Nemoto, R. Z. Valiev, T. G. Langdon. Mater. Sci. Eng. A. 241<br />
(1998) 122.<br />
[10] L. Cisar, Y. Yoshida, S. Kamado, Y. Kojima, F. Watanabe. Mater. Sci. Forum. 419-4 (2003) 249.<br />
[11] E. J. Kwak, C. H. Bok, M. H. Seo, H. S. Kim. Mater. Trans. 49 (2008) 1006.<br />
[12] X. Zhao, W. Fu, X. Yang, T. G. Langdon. Scripta Mater. 59 (2008) 542.<br />
[13] J. Koike, R. Ohyama. Acta Mater. 53 (2005) l963.<br />
[14] N. Ogawa, M. Shiomi, K. Osakada. Int J Mach Tool Manu. 42 (2002) 607.<br />
[15] W. M. Gan, M. Y. Zheng, H. Chang, B. Schwebke. J. Alloys. Comp. 470 (2009) 256.<br />
[16] A. Yamashita, Z. Horita, T. G. Langdon. Mater. Sci. Eng. A 300 (2001) 142.<br />
[17] K. Matsubara, Y. Miyahara, Z. Horita, T. G. Langdon. Metall. Mater. Trans. A 35 (2004) 1735.<br />
[18] S. Y. Chang, S. W. Lee, K. M. Kang, S. Kamado, Y. Kojima. Mater. Trans. 45 (2004) 488.<br />
[19] W. J. Kim, S. I. Hong, Y. S. Kim, S. H. Mind, H. T. Jeong, J. D. Lee. Acta Mater. 51 (2003) 3293.<br />
95
Dielectrophoretic Separation <strong>of</strong> Carbon Nanotubes <strong>and</strong> Polystyrene<br />
Microparticles [1]<br />
C. Zhang 1 * , K. Khoshmanesh 2 , F. J. Tovar-Lopez 1 , A. Mitchell 1 , W. Wlodarski 1 , K. Klantarzadeh<br />
1<br />
* Presenter<br />
1. School <strong>of</strong> Electrical <strong>and</strong> Computer Engineering, RMIT University, Australia<br />
2. School <strong>of</strong> Engineering <strong>and</strong> IT, Deakin University, Australia<br />
The separation <strong>of</strong> multi-walled carbon nanotubes (MWCNTs) <strong>and</strong> polystyrene<br />
microparticles using a dielectrophoresis (DEP) system is presented. The DEP system consists<br />
<strong>of</strong> arrays <strong>of</strong> parallel microelectrodes patterned on a glass substrate. The performance <strong>of</strong> the<br />
system is evaluated by means <strong>of</strong> numerical simulations. The MWCNTs demonstrate a<br />
positive DEP behaviour <strong>and</strong> can be trapped at the regions <strong>of</strong> high electric field. However, the<br />
polystyrene microparticles demonstrate a negative DEP behaviour at a certain range <strong>of</strong><br />
frequencies <strong>and</strong> migrate to the regions <strong>of</strong> low electric field. Experiments are performed on the<br />
microparticles at the frequencies between 100 Hz <strong>and</strong> 1 MHz to estimate their crossover<br />
frequency <strong>and</strong> select the range <strong>of</strong> separation frequencies. Further, experiments are conducted<br />
at the obtained range <strong>of</strong> separation frequencies to separate the MWCNTs <strong>and</strong> polystyrene<br />
microparticles.<br />
Keywords Diekectrophoresis, Particle separation, Carbon nanotube, Polystyrene<br />
microparticle<br />
1. Introduction<br />
When a neutral particle is placed in a non-uniform electric field, it experiences a translational<br />
force due to the polarization effect <strong>of</strong> the particle [1]. Pohl [2] was one <strong>of</strong> first who recognized <strong>and</strong><br />
explored the applications <strong>of</strong> this force, known as dielectrophoretic (DEP) force, for manipulating <strong>and</strong><br />
alignment <strong>of</strong> particles.<br />
DEP force could be positive or negative. Positive DEP force traps particles into the regions with<br />
high electric field gradient (e.g. the corners <strong>of</strong> electrodes), <strong>and</strong> negative DEP force pushes particles<br />
into weak field regions (e.g. the spaces between electrodes) [3]. Due to the different trapping effects <strong>of</strong><br />
positive <strong>and</strong> negative DEP forces, particles with different properties (i.e. dielectric constant, size, <strong>and</strong><br />
conductivity) could respond differently to the applied electric fields with different oscillation<br />
frequencies. DEP can be employed in a number <strong>of</strong> micr<strong>of</strong>luidic technology applications including<br />
particle separation <strong>and</strong> sorting [1, 4-6], particle trapping <strong>and</strong> assembling [7-11], as well as patterning<br />
[12-14].<br />
Due to their unique characteristics, one dimensional nanostructures have been extensively studied<br />
<strong>and</strong> applied for a number <strong>of</strong> novel nanotechnological applications [15]. They have been employed in<br />
developing electronic devices such as sensors [16, 17], oscillators [18] <strong>and</strong> transistors [19].<br />
As the manipulation <strong>of</strong> nanostructures is most effectively conducted within liquid suspensions, it<br />
is important to realize the methods for separating the desired nanostructures from the liquid mixture.<br />
The application <strong>of</strong> DEP to particle discrimination, separation, <strong>and</strong> DEP field-flow fractionation was<br />
reviewed in [1]. Numerical scheme based on the distributed Lagrange multiplier method was used for<br />
studying the motion <strong>of</strong> the nano-sized particles <strong>of</strong> dielectric suspensions subjected to uniform <strong>and</strong><br />
nonuniform electric fields [3]. Cells with different size were separated using direct current DEP in [20].<br />
The separation <strong>of</strong> polystyrene beads <strong>and</strong> cells by different dielectric fields was reported in [4], where<br />
opposite DEP forces were used for accurately focusing a stream <strong>of</strong> beads <strong>and</strong> yeast cells at different<br />
locations across the channel by adjusting potentials <strong>and</strong> frequencies. A microparticle<br />
filter (i.e. polystyrene beads, yeast cells, spores <strong>and</strong> bacteria) was reported to be successfully<br />
developed <strong>and</strong> characterized by Li et al [21] where DEP force was employed to capture desired<br />
particles. Live <strong>and</strong> nonviable cells were separated due to their significant differences in DEP mobility<br />
[1]. Published at Micr<strong>of</strong>luidics <strong>and</strong> Nan<strong>of</strong>luidics (2009) 7:633-645
[22-25]. Sorting <strong>of</strong> normal <strong>and</strong> Babesia bovis infected erythrocytes was reported using DEP by<br />
exploiting the higher ionic membrane permeability <strong>of</strong> Babesia bovis infected cells [5]. Separating <strong>and</strong><br />
sorting <strong>of</strong> other bioparticles including DNA [26, 27], proteins [28] <strong>and</strong> viruses [29-31] were also<br />
reported. Additionally, the possibility <strong>of</strong> separating metallic <strong>and</strong> semi-conducting single-walled carbon<br />
nano-tubes (CNTs) using DEP was theoretically investigated in [32] by applying Brownian dynamics<br />
method, where conductivity difference produces different DEP behaviours for semiconducting <strong>and</strong><br />
conducting CNTs.<br />
The DEP separations have been reported extensively. However, there is rarely any report<br />
concerning the separation <strong>of</strong> nanostructures from microstructures. This technique can be implemented<br />
in novel applications including: purifying cell suspensions with nanoscale impurities, separation <strong>of</strong><br />
organelles, <strong>and</strong> nanostructure assembling <strong>and</strong> separation for microelectronic fabrications. For instance,<br />
nanotechnology approaches are recently applied for drug delivery to target cells. In these approaches,<br />
drug materials are encapsulated within or coated around different nanoscale containers <strong>and</strong> vehicles<br />
such as nanospheres [33], lipid coated nanoparticles [34], magnetic nanoparticles [35] <strong>and</strong> carbon<br />
nano-tubes [36] to be delivered to the target cells. For assessing whether the delivery <strong>of</strong> drugs to cells<br />
is achieved successfully or their therapeutic functionalities are effective, the cells are lysed <strong>and</strong> their<br />
ingredients are released for analysis. Consequently, DEP techniques can play a major role in<br />
separating the nanoscale particles <strong>and</strong> microscale organelles (such as nucleus <strong>and</strong> mitochondria) from<br />
the lysed cells.<br />
Both micro <strong>and</strong> nano structures can be employed as sensing elements in sensor development <strong>and</strong><br />
fabrication. Due to the large surface area to volume ratio <strong>of</strong> such materials, the developed devices are<br />
highly sensitive; however, their selectivity should be improved. DEP trapping <strong>and</strong> separation<br />
techniques enable the deposition <strong>of</strong> multiple <strong>and</strong> dissimilar materials on a single sensing device at<br />
different locations, which can be used for the development <strong>of</strong> an array <strong>of</strong> sensors with multiple<br />
sensitive layers.<br />
In this work, MWCNTs were separated from polystyrene microparticles in stationary liquid by<br />
using DEP. The polystyrene microparticles were mixed with MWCNTs in suspension. The mixture<br />
was then placed on a micro-gapped interdigital transducer (IDT) <strong>and</strong> sealed with a<br />
polydimethylsiloxane (PDMS) block. An alternating voltage with amplitude 10 V peak-to-peak <strong>and</strong><br />
frequencies from 100 Hz to 1 MHz was applied across the electrodes. An inverted microscope was<br />
used to observe the behaviour <strong>of</strong> the particles. The observed DEP separation performance was<br />
compared with computational fluid dynamic predictions.<br />
2. DEP Theory<br />
2.1 DEP Force<br />
If damping <strong>and</strong> viscous forces are negated, the DEP force equation can be expressed as [37]:<br />
* * 2<br />
FDEP<br />
= Γ ⋅ εm<br />
⋅ Re[fCM<br />
(ε<br />
p,<br />
εm<br />
)] ⋅ ∇E<br />
(1)<br />
,<br />
where, Γ is the geometric factor <strong>of</strong> the particles, ε m is the dielectric constant <strong>of</strong> suspending medium,<br />
ε * m <strong>and</strong> ε * p are the complex permittivities <strong>of</strong> suspending medium <strong>and</strong> particles, respectively, f CM is the<br />
Claussius-Mossotti (CM) factor, (also known as the complex polarisation factor), <strong>and</strong> E refers to the<br />
alternating electric field. The complex permittivity is defined as:<br />
σ<br />
(2)<br />
ε * = ε − i<br />
ω ,<br />
where, ω is angular frequency <strong>of</strong> the electric field, σ is electric conductivity <strong>and</strong> i=√-1.<br />
Γ depends on the geometry <strong>of</strong> the particles that is expressed distinctly for cylindrical [20, 21] <strong>and</strong><br />
spherical structures [3] as below:<br />
2<br />
πr l<br />
(3)<br />
Γ = cylindrica l<br />
6 , (4)<br />
3<br />
Γ<br />
spherical<br />
= 2 πr<br />
,<br />
where, r indicates the radius <strong>of</strong> the sphere or the cross section <strong>of</strong> the cylinder <strong>and</strong> l represents the<br />
length <strong>of</strong> the cylindrical structures.<br />
97
The CM factor also depends on the geometry <strong>of</strong> structures <strong>and</strong> is expressed differently for<br />
cylindrical [38] in (5) <strong>and</strong> spherical [3] structures in (6) as below:<br />
* *<br />
ε<br />
(5)<br />
p<br />
− ε<br />
* *<br />
m<br />
f<br />
CM<br />
(ε<br />
p<br />
,ε<br />
m<br />
) =<br />
*<br />
ε<br />
m ,<br />
* *<br />
ε<br />
p<br />
− ε<br />
* *<br />
m<br />
f<br />
CM<br />
(ε<br />
p<br />
,ε<br />
m<br />
) =<br />
* *<br />
(6)<br />
ε<br />
p<br />
+ 2ε<br />
m<br />
,<br />
2.2 Characterisation <strong>of</strong> the CM factor for MWCNT <strong>and</strong> polystyrene microparticles<br />
The DEP force takes the same sign <strong>of</strong> the real part <strong>of</strong> the CM factor, which can be either positive<br />
or negative. To explore the ability <strong>of</strong> DEP force to sort polystyrene microparticles <strong>and</strong> WMCNTs,<br />
calculations <strong>of</strong> the CM factor for various electric field frequencies were conducted using Equations<br />
(1)-(6) for each <strong>of</strong> these particles suspended in DI water. Because <strong>of</strong> the significant size difference<br />
between MWCNTs <strong>and</strong> polystyrene microparticles, the geometric factor Γ was also calculated <strong>and</strong><br />
taken into the consideration.<br />
Firstly, the value <strong>of</strong> Г×Re[f CM ] <strong>of</strong> MWCNTs was calculated. The relative permittivity <strong>of</strong><br />
MWCNTs was taken as 2.5 while the conductivity was assumed between 1000 S/m <strong>and</strong> 150 S/m<br />
according to [38, 39]. The diameter <strong>and</strong> length <strong>of</strong> MWCNTs were assumed to be 50 nm <strong>and</strong> 2 μm<br />
respectively. The relative permittivity <strong>of</strong> the DI water is taken 78 <strong>and</strong> the conductivity is measured as<br />
0.2 mS/m.<br />
Figure 1 presents the value <strong>of</strong> Г×Re[f CM ] as a function <strong>of</strong> applied electric field frequency,<br />
calculated for MWCNTs suspended in deionised (DI) water. Since the value stays above zero for all<br />
frequencies considered in this investigation (1 kHz-100 MHz), MWCNTs always experience a<br />
positive DEP force.<br />
Next, Г×Re[f CM ] <strong>of</strong> polystyrene microparticles was calculated. This calculation is more<br />
complicated than that <strong>of</strong> MWCNT as the conductivity <strong>of</strong> polystyrene must be carefully defined.<br />
Polystyrene has a very low conductivity, <strong>and</strong> thus the inherent conductivity <strong>of</strong> material is zero.<br />
However, placing the microparticles in DI water results in the inducing <strong>of</strong> surface conductance [40].<br />
The total surface conductance can be modelled as two separate components: a conductance due to the<br />
movement <strong>of</strong> charges in the diffuse double layer <strong>and</strong> another due to charge movements in the Stern<br />
layer. As a result, the microparticle has an non-zero effective conductivity σ p , which could be<br />
expressed as [41]:<br />
2K<br />
(7)<br />
S<br />
σ<br />
p<br />
=<br />
r<br />
where K s represents the surface conductance <strong>and</strong> r is the radius <strong>of</strong> the microparticle. K s can be<br />
calculated as [41]:<br />
r<br />
(8)<br />
⎜⎛<br />
2<br />
2<br />
K = − + − ( − )( + )( ⋅ ) ⎟⎞<br />
S<br />
σ<br />
m<br />
9σ<br />
m<br />
4 ε<br />
p<br />
ε<br />
m<br />
ε<br />
p<br />
2ε<br />
m<br />
f<br />
0<br />
2π<br />
4 ⎝<br />
⎠<br />
Here, f 0 is known as the crossover frequency where DEP force is zero. Therefore, at the crossover<br />
frequency the value <strong>of</strong> K s has a linear relationship with the particle diameter if permittivity <strong>of</strong> the<br />
medium <strong>and</strong> particles are constant.<br />
The surface conductance has been calculated for different dimensions <strong>of</strong> latex beads [41].<br />
According to the results, the value <strong>of</strong> K s is found to be 0.25 nS. In this case, the microparticle<br />
conductivity is calculated to be 1 mS/m. The relative permittivity <strong>of</strong> microparticle is assumed to be 2.5.<br />
Regarding these values, Г×Re[f CM ] was calculated as a function <strong>of</strong> frequency <strong>and</strong> presented in Figure 2.<br />
It is evident that in this case, the microparticles experience positive DEP forces below 200 kHz, <strong>and</strong><br />
negative DEP force above 200 kHz.<br />
98
Figure 1. Г×Re[fCM] vs. frequency for the MWCNTs in DI water; (up): log to log scale; (down):<br />
normal scale.<br />
By comparing Figures 1 <strong>and</strong> 2, it is evident that MWCNTs experience strong positive DEP force<br />
for frequencies smaller than 1 MHz while polystyrene microparticles experience neutral or moderate<br />
negative DEP force for frequencies larger than 200 kHz. As a result, it should be possible to use DEP<br />
to separate the MWCNT <strong>and</strong> polystyrene microparticles, particularly by applying an electric field with<br />
frequency in the range <strong>of</strong> 200 kHz to 1 MHz.<br />
In this paper, the used MWCNTs were metallic <strong>and</strong> their conductivity was assumed to lie in a<br />
range between 150 to 1000 S/m <strong>and</strong> the calculations <strong>and</strong> simulations were based on this assumption.<br />
Semi-conductive carbon nanotubes (CNTs) have also been used in DEP experiments by other groups<br />
[42]. However, due to their dielectric properties, they experience negative DEP force at high<br />
frequencies. As, the crossover frequencies <strong>of</strong> polystyrene microparticles <strong>and</strong> the semi-conductive<br />
MWCNTs are different, it is possible to perform DEP separation <strong>of</strong> these two materials for ranges<br />
between the two crossovers frequencies.<br />
99
Figure 2. Г×Re[fCM] vs. frequency for polystyrene microparticles in DI water, as the existence <strong>of</strong><br />
negative value, plot in log to log scale is not applied.<br />
2.3 Particle Geometry <strong>and</strong> Conductivity<br />
According to Equation (3) to (6), the DEP force is a strong function <strong>of</strong> particle geometry, as well<br />
as particle conductivity. In order to identify the influence <strong>of</strong> these two parameter on the DEP force, the<br />
Г×Re[f CM ] value <strong>of</strong> cylindrical <strong>and</strong> spherical particles with similar volumes, for both similar <strong>and</strong><br />
starkly different conductivities were calculated <strong>and</strong> compared as shown in Figure 3. DI water (relative<br />
permittivity <strong>of</strong> 78, conductivity <strong>of</strong> 0.2 mS/m) was selected to be the particle suspending medium. In<br />
Figure 3, it is assumed that cylindrical particles have the same dimensions as the tested MWCNTs<br />
(diameter <strong>of</strong> 50 nm, length <strong>of</strong> 2 μm) <strong>and</strong> the spherical particle have a diameter <strong>of</strong> 0.2 μm. The relative<br />
permittivity <strong>of</strong> both particles was assumed to be 2.5. Calculations were conducted for two different<br />
sets <strong>of</strong> conductivities: 1 S/m <strong>and</strong> 1000 S/m.<br />
Figure 3 shows that the DEP spectra <strong>of</strong> cylindrical particles with different conductivities follow<br />
similar trends. However, the DEP force magnitudes are proportional to conductivity. The crossover<br />
frequency is the same for both conductivities in the cylindrical structure. On the contrary, the spherical<br />
particles with conductivity <strong>of</strong> 1000 S/m experience stable positive DEP force for the whole range <strong>of</strong><br />
investigated frequencies. This is because in equation (6), the CM factor converges to one for large<br />
particle conductivities. However, the DEP force, applied on spherical particles with conductivity <strong>of</strong> 1<br />
S/m, decreases when frequency is higher than 10 MHz. As a conclusion, the conductivity <strong>of</strong> spherical<br />
particle plays a more important role in DEP separation, as it alters the crossover frequency.<br />
3. Design <strong>and</strong> Simulation<br />
The DEP platform consists <strong>of</strong> 8 arrays <strong>of</strong> microelectrodes, each containing 20 interdigitated<br />
microgaps, as shown in Figure 4. The gap separation is 10 μm, <strong>and</strong> gaps are placed at a distance <strong>of</strong><br />
100 μm from each other.<br />
100
Figure 3. The DEP spectrum <strong>of</strong> cylindrical <strong>and</strong> spherical particles with conductivities <strong>of</strong> 1000 S/m<br />
<strong>and</strong> 1 S/m. The calculation was based on comparable particle size <strong>and</strong> relative permittvities.<br />
Figure 4. Design <strong>of</strong> the DEP platform. (A): top overview <strong>of</strong> the DEP platform; (B): the arrays <strong>of</strong><br />
micro electrode gap; <strong>and</strong> (C): the close view <strong>of</strong> a gap section.<br />
The 10 μm gap was chosen because it was less likely to be damaged when high potential AC<br />
voltages were applied. The purpose to separate each gap by 100 μm is to maximize the difference<br />
between strong field <strong>and</strong> weak field regions, in order to clearly distinguish between positive <strong>and</strong><br />
negative DEP behaviours. It also provides enough space for polystyrene microparticles to have<br />
constant motions without colliding with the MWCNTs trapped in the gap. The distance between arrays<br />
was designed to be 260 μm to minimize the field interference between arrays. Similar designs have<br />
been extensively used in aligning <strong>and</strong> assembling <strong>of</strong> nanowires [37, 43, 44] for micro <strong>and</strong> nano<br />
electronics applications.<br />
The Computational Fluid Dynamics (CFD) method is applied to estimate the performance <strong>of</strong> the<br />
DEP system. The CFD method enables us to analyse the distributions <strong>of</strong> the electric potential as well<br />
as the electric field that are generated by the applied signals around electrodes.<br />
101
The electric potential field is governed by the Poisson’s equation as follows:<br />
2<br />
∇ ϕ = 0 ,<br />
(9)<br />
where, φ is the electric potential <strong>and</strong> ∇ is the gradient operator. The electric field is obtained by<br />
calculating the derivative <strong>of</strong> the electric potential field as follows:<br />
E r = −∇ϕ , (10)<br />
where, E r 2<br />
is the electric field that is a vector variable. Having this variable, the value <strong>of</strong> ∇E was<br />
obtained throughout the field <strong>and</strong> substituted into Equation (1) to calculate the DEP force in different<br />
locations <strong>of</strong> the device.<br />
The simulations were conducted in a two-steps process. First, the geometry <strong>of</strong> the device was<br />
created <strong>and</strong> then divided into small elements using the Gambit-2.3 s<strong>of</strong>tware package (Fluent USA,<br />
Lebanon, NH). Along the electrode edges, fine structured quadrilateral elements were applied,<br />
enabling the control <strong>of</strong> the thickness <strong>of</strong> these elements. Selecting the proper thickness <strong>of</strong> these<br />
elements was crucial to predict the sharp gradients <strong>of</strong> the electric field along the electrodes <strong>and</strong> to<br />
assure the convergence <strong>of</strong> the simulation. In the interior regions <strong>of</strong> the system, unstructured<br />
quadrilateral elements were applied due to their flexibility.<br />
Next, the Fluent-6.3 s<strong>of</strong>tware package (Fluent USA, Lebanon, NH) was applied to solve the<br />
governing Equations (9)-(10). The s<strong>of</strong>tware applied the finite volume method to discretise the<br />
governing equations across each element. The s<strong>of</strong>tware was developed to solve the differential Navier-<br />
Stokes equations in fluidic systems <strong>and</strong> predict the velocity, pressure <strong>and</strong> other flow variables.<br />
However, other differential equations can be solved using the UDS (user-defined scalars) module <strong>of</strong><br />
the s<strong>of</strong>tware.<br />
The boundary conditions applied were as follows: the electrodes on one side <strong>of</strong> the device were<br />
set at 5 V while the electrodes on the opposite side are grounded. The bottom, top <strong>and</strong> side walls <strong>of</strong> the<br />
device were regarded as insulators, where the gradient <strong>of</strong> the electric potential was zero.<br />
Figures 5-A, B <strong>and</strong> C show the gradient <strong>of</strong> the electric field square along x, y <strong>and</strong> z directions.<br />
Figures 5-A <strong>and</strong> B show that the field has a much higher gradient at the corners <strong>of</strong> electrode tips.<br />
Figure 5-C shows that the field gradient is positive in region between the electrode tips <strong>and</strong> is negative<br />
at electrode edges. Therefore, particles with a positive CM factor are trapped when they are close to<br />
the electrode tips, <strong>and</strong> are strongly levitated when are located within the intermediate region.<br />
The simulation results assert that the design <strong>of</strong> this DEP platform is appropriate for the separation<br />
experiments. The region between electrodes has a weak <strong>and</strong> uniformly distributed electric field, which<br />
is suitable for producing negative DEP while the tips <strong>of</strong> electrodes have a sharp electric field, which is<br />
suitable for producing positive DEP. Hence spatial separation <strong>of</strong> the particles will occur if some <strong>of</strong><br />
them experience negative DEP <strong>and</strong> the others experience the positive DEP forces.<br />
4. Experimental<br />
The DEP platform was micr<strong>of</strong>abricated using photolithography on glass substrates. The glass<br />
substrate was first metallised (Electron-beam evaporation technique) using Cr <strong>and</strong> Au with the<br />
thickness <strong>of</strong> 50 nm <strong>and</strong> 150 nm, respectively. The AZ 1512 photoresist was then spin coated at 3000<br />
rpm (acceleration <strong>of</strong> 1000 rpm/sec) for 25 seconds, <strong>and</strong> this procedure was followed by s<strong>of</strong>t baking for<br />
20 minutes at 90 °C. A mask aligner (Suss MJB3) was used to expose the mask pattern on to the<br />
photoresist. The photoresist was developed <strong>and</strong> then the metal film was etched using reactive ion<br />
etching. After patterning the sample was cleaned using isopropyl alcohol <strong>and</strong> acetone <strong>and</strong> then dried<br />
with nitrogen.<br />
102
A<br />
B<br />
C<br />
Figure 5. The gradient <strong>of</strong> the electric field square along x, y <strong>and</strong> z directions obtained by the CFD<br />
∂ ∂<br />
simulations, which depicts the intensity <strong>of</strong> the DEP forces. (A): , (B): , (C):<br />
∂<br />
∂x<br />
∂y<br />
∂z<br />
A PDMS block (30 mm×30 mm×10 mm) with a micr<strong>of</strong>luidic channel (20 mm ×1 mm×80 μm)<br />
was used to immobilise the particle suspension in the vicinity <strong>of</strong> the electrodes. To fabricate the<br />
PDMS block, an 80 μm (channel height) KMPR1025 photoresist layer was coated onto a polished<br />
silicon wafer, <strong>and</strong> photolithography was used for transferring the channel pattern onto the photoresist<br />
layer. After developing <strong>and</strong> hard baking, liquid PDMS was applied on the surface, <strong>and</strong> cured on<br />
hotplate at 100 ºC for 2 minutes. After PDMS was cured, 2 holes were punched in for liquid inlet <strong>and</strong><br />
E 2<br />
E 2<br />
E 2<br />
.<br />
103
outlet. The channel was assembled on to the DEP platform such that the interdigital electrodes were<br />
sealed within the channel but with contacting pads exposed, as shown in Figure 6.<br />
Figure 6. PDMS channel (80 μm) block for liquid sealing<br />
MWCNTs (Sigma Aldrich, 636649), 20-50 nm in diameter, 0.5-2 μm in length, with single-wall<br />
thickness <strong>of</strong> 1-2 nm were dispersed into DI water with an initial concentration <strong>of</strong> 0.1 mg/ml. The<br />
suspension was stabilized with the aid <strong>of</strong> Triton X-305 surfactant (Sigma Aldrich, approximately 40 μl<br />
surfactant per 10 ml suspension). The liquid was then mixed with polystyrene microparticles (Duke<br />
Scientific, 1μm in diameter, <strong>and</strong> concentration <strong>of</strong> 1.8×10 10 ) in the ratio 1:1. The mixture was then<br />
placed in a 125-W ultra-sonicator (Branson) bath for 45 minutes <strong>and</strong> left intact for 6 hours to allow<br />
visible agglomerates to precipitate from the suspension. The suspension was then ready for DEP<br />
experiments.<br />
After the liquid suspension was injected into the micr<strong>of</strong>luidic channel, alternating<br />
voltages with amplitudes <strong>of</strong> 10 V peak-to-peak <strong>and</strong> frequencies from 100 Hz to 1 MHz was<br />
applied across the electrodes. An inverted microscope was used to observe the behaviour <strong>of</strong><br />
the particles.<br />
5. Result <strong>and</strong> Discussion<br />
5.1 Selection <strong>of</strong> Voltage Amplitude<br />
During DEP separation experiments, an AC voltage <strong>of</strong> 10 V peak to peak was employed as the<br />
optimum amplitude. According to our preliminary tests, only AC voltages with the amplitudes higher<br />
than 1 V can produce visible DEP behaviour for 1 μm polystyrene microparticles, <strong>and</strong> higher<br />
amplitudes resulted in higher separation rates. However, applying AC voltages <strong>of</strong> more than 10 V<br />
damaged <strong>and</strong> delaminated the electrodes, as shown in Figure 7. It is seen that the Au electrodes were<br />
etched <strong>and</strong> the gap between them was extended by an extra 10 μm. According to this limitation, the<br />
amplitude <strong>of</strong> 10 V peak to peak appeared to be the optimum voltage.<br />
104
Figure 7. SEM image <strong>of</strong> damaged electrodes which used for performing DEP trapping with AC<br />
voltage <strong>of</strong> over 10 V peak to peak. The tip <strong>of</strong> electrode was etched (circled part) during the experiment.<br />
5.2 Verification <strong>of</strong> the Practical Crossover Frequencies<br />
To verify the value <strong>of</strong> K s <strong>and</strong> hence the crossover frequencies obtained in Section 2.2, the DEP<br />
experiments <strong>of</strong> polystyrene microparticles in DI water was carried out for frequencies in the range <strong>of</strong><br />
100 to 150 kHz. An inverted microscope was used for observing the particle behaviour.<br />
Figure 8 present the DEP behaviour for different frequencies <strong>of</strong> applied electric field, the subfigures<br />
show 100 Hz (A), 500 Hz (B), 1 kHz (C), 5 kHz(D), 10 kHz (E), 50 kHz (F), 150 kHz (G), 500<br />
kHz (H) <strong>and</strong> 1 MHz (I). Positive DEP behaviour can be observed in Figures 8-A to 8-D, as the<br />
microparticles are trapped at the edge <strong>of</strong> electrode. The decrease in the number <strong>of</strong> microparticles<br />
trapped while frequency <strong>of</strong> applied field was increased from 100 Hz to 5 kHz is due to the decrease <strong>of</strong><br />
the DEP force magnitude. Figures 8-E <strong>and</strong> 8-F show the behaviour <strong>of</strong> microparticles at the frequencies<br />
close to the crossover frequency, as the microparticles lost the tendency to get trapped at the edges <strong>of</strong><br />
the microelectrodes. At 150 kHz (as shown in Figure 8-G), a typical negative DEP behaviour <strong>of</strong><br />
microparticles is evident, as the microparticles are repelled from the microelectrodes <strong>and</strong> are clustered<br />
in the spaces between them; chaining <strong>of</strong> microparticles is due to the polarization effect in the electric<br />
field. There is not much difference in microparticle behaviours in Figures 8-H <strong>and</strong> I (500 kHz <strong>and</strong> 1<br />
MHz) compared to that <strong>of</strong> Figures 8-G (150 kHz).<br />
According to the theoretical calculations <strong>of</strong> Section 2.2, the crossover frequency should be<br />
approximately 200 kHz, however, the results <strong>of</strong> Figure 8 suggest the crossover frequency was between<br />
10 kHz to 50 kHz. This mismatching might be due to the fact that the assumed values for surface<br />
conductance K s were not accurate. In [41], the author only calculated K s for the particles with diameter<br />
in the range <strong>of</strong> 2 μm to 10 μm, K s was found to be equal to 1 nS from extrapolation . Additionally, the<br />
measurement environment is different, the DI water was obtained from a different purifier,<br />
microparticles are purchased from different supplier, <strong>and</strong> experiment temperatures could be varied<br />
(both nominally room temperature). As small changes in surface conductance can give rise to large<br />
changes in the DEP spectrum, the crossover frequency could be tuned by change <strong>of</strong> K s [40]. And when<br />
microparticle conductance σ p was adjusted to 0.3 mS/m in the model <strong>of</strong> Section 2.2, theoretical <strong>and</strong><br />
practical results were matched.<br />
105
Figure 8. Particle behaviours at frequencies from 100 Hz to 1 MHz. (A-D): positive DEP observed; (E<br />
& F): close to crossover frequency. (G): typical negative DEP behaviour. (H & I): DEP behaviour <strong>of</strong><br />
particles at 500 kHz <strong>and</strong> 1 MHz is similar to that at 150 kHz.<br />
5.3 Particle Separation<br />
The ability to use DEP to sort particles was investigated by applying electric fields with<br />
frequencies <strong>of</strong> 100 Hz, 1 kHz, 100 kHz <strong>and</strong> 150 kHz <strong>and</strong> examining the system for spatial separation<br />
<strong>of</strong> the particles. The separation progress in the stationary liquid is shown in Figures 9 <strong>and</strong> 10.<br />
At 100 Hz (Figure 9-A), strong positive DEP was observed. Large numbers <strong>of</strong> polystyrene<br />
microparticles <strong>and</strong> MWCNTs were in constant motion in the region near the electrode tips. This was<br />
due to the existence <strong>of</strong> large levitating <strong>and</strong> trapping forces described in Section 3. When the frequency<br />
<strong>of</strong> applied field was increased to 1 kHz (Figure 9-B), microparticles in motion could no longer be<br />
observed. As this frequency increased to values closer to the crossover frequency (approximately 30<br />
kHz), the DEP force became weaker <strong>and</strong> the number <strong>of</strong> assembled microparticles decreased. Figure 9-<br />
C illustrates initial separation behaviour <strong>of</strong> polystyrene microparticles <strong>and</strong> MWCNTs at 100 kHz, as<br />
all polystyrene microparticles were repelled from strong electric field region, while MWCNTs stayed<br />
on the edge <strong>of</strong> the electrodes. At 150 kHz (Figure 10-A), the separation <strong>of</strong> MWCNTs <strong>and</strong> polystyrene<br />
microparticles was complete. Since the magnitude <strong>of</strong> DEP forces affecting MWCNTs <strong>and</strong> polystyrene<br />
microparticles is not changing for frequencies larger than 100 kHz, applied signals in the range <strong>of</strong> 100<br />
kHz to 1 MHz resulted in similar DEP behaviour (Figure 9-D illustrates the particle behaviour at 1<br />
MHz).<br />
After the polystyrene microparticles were removed <strong>and</strong> the liquid was evaporated, the substrate<br />
was characterised by a scanning electron microscope (SEM) to observe the trapping effect <strong>of</strong><br />
MWCNTs; <strong>and</strong> as can be in Figure 10-B, MWCNTs were trapped on the tip <strong>and</strong> edges <strong>of</strong> electrodes.<br />
106
Figure 9. Particle separation progress observed from the frequency <strong>of</strong> 100 Hz to 100 kHz. (A): at 100<br />
Hz, both microparticles <strong>and</strong> MWCNTs were experiencing strong positive DEP forces; (B): at 1 kHz,<br />
positive DEP forces were weaker, (C): at 100 kHz, MWCNTs were still experiencing positive DEP<br />
forces, while microparticles were influenced by negative ones. (D): at 1 MHz, particle behaviours are<br />
similar to that at 150 kHz.<br />
A<br />
B<br />
Assembled<br />
MWCNTs<br />
Assembled<br />
MWCNTs<br />
Assembled<br />
MWCNTs<br />
20 µm<br />
Slide 1<br />
Figure 10. A: At 150 kHz, complete particle separation achieved, polystyrene microparticles were<br />
repelled from electrode tips, while MWCNTs trapped at tips. B: SEM image <strong>of</strong> assembled MWCNTs<br />
on the tip <strong>and</strong> edge <strong>of</strong> electrodes.<br />
5.4 Collection <strong>of</strong> Separated Particles<br />
According to the calculations presented in Section 2.2, the magnitude <strong>of</strong> the DEP force affecting<br />
MWCNTs is much larger than polystyrene microparticles. Therefore, maintaining a proper flow in the<br />
micr<strong>of</strong>luidic channel, the polystyrene microparticles can be washed away <strong>and</strong> removed. MWCNTs can<br />
be released when the alternating voltage is switched <strong>of</strong>f. In this case, the collection <strong>of</strong> polystyrene<br />
microparticle can be achieved by maintaining the flow in positive direction; <strong>and</strong> after that, the<br />
107
MWCNTs can be collected by reversing the flow direction. The collecting process is shown in Figure<br />
11.<br />
Figure 11. The collecting process for polystyrene microparticles <strong>and</strong> MWCNTs. (A) Particles injected<br />
in micr<strong>of</strong>luidic channel. (B): DEP trapping when the electric field is applied on the electrodes. (C):<br />
Collection <strong>of</strong> polystyrene microparticles with the fluidic flow. (D): Releasing captured MWCNTs by<br />
switch <strong>of</strong>f the applied electric field. (E): Collection <strong>of</strong> MWCNTs by reversing the flow direction.<br />
5.5 More Discussions<br />
As in this work, all particles used for experiment were non-bioparticles, there are no limitations.<br />
However, causation should be experienced when the electric field stresses the live cells membrane<br />
[45]. Experimentally, strong electric field (large amplitude <strong>of</strong> alternating voltage) can generate large<br />
DEP force, which improves the separation efficiency. However, cells are damaged just in a short time<br />
after they are exposed to a very high electric field level [20] specially near the electrode tips.<br />
108
The separation efficiency can be improved by optimising the system design. A multi-stage<br />
separation system could be one <strong>of</strong> the solutions <strong>and</strong> much work need to be conducted in order to<br />
minimise the detrimental effects <strong>of</strong> the applied electric field on live tissues.<br />
6. Conclusion<br />
A platform for separating particles according to their dielectric response to alternating electric<br />
fields at specific frequencies has been presented. The performance <strong>of</strong> the device was simulated using<br />
CFD method. The dielectrophoretic spectrums for both polystyrene microparticles <strong>and</strong> MWCNTs<br />
were studied <strong>and</strong> compared to obtain the optimal frequency for particle separation. Experiments were<br />
carried out <strong>and</strong> MWCNTs were successfully separated from polystyrene microparticles for frequencies<br />
over 100 kHz. Particle behaviours at different frequencies were characterized <strong>and</strong> analysed. The<br />
developed device has potential novel applications such as: purifying cell suspensions with nanoscale<br />
impurities, tissues <strong>and</strong> organelles separation, as well as microelectronic <strong>and</strong> nanoelectronic device<br />
fabrications.<br />
Reference:<br />
[1] P. R. C. Gascoyne, <strong>and</strong> J. Vykoukal, “Particle separation by dielectrophoresis,”<br />
Electrophoresis, vol. 23, no. 13, pp. 1973-1983, Jul, 2002.<br />
[2] H. A. Pohl, Dielectrophoresis: The Behaviour <strong>of</strong> Neutral Matter in Nonuniform Elctric Field,<br />
Cambridge: Cambridge University Press, 1978.<br />
[3] A. T. J. Kadaksham, P. Singh, <strong>and</strong> N. Aubry, “Dielectrophoresis <strong>of</strong> nanoparticles,”<br />
Electrophoresis, vol. 25, no. 21-22, pp. 3625-3632, Nov, 2004.<br />
[4] N. Demierre, T. Braschler, R. Muller et al., “Focusing <strong>and</strong> continuous separation <strong>of</strong> cells in a<br />
micr<strong>of</strong>luidic device using lateral dielectrophoresis,” Sensors <strong>and</strong> Actuators B-Chemical, vol.<br />
132, no. 2, pp. 388-396, 2008.<br />
[5] E. M. Nascimento, N. Nogueira, T. Silva et al., “Dielectrophoretic sorting on a<br />
micr<strong>of</strong>abricated flow cytometer: Label free separation <strong>of</strong> Babesia bovis infected erythrocytes,”<br />
Bioelectrochemistry, vol. 73, no. 2, pp. 123-128, 2008.<br />
[6] T. Yasukawa, M. Suzuki, T. Sekiya et al., “Flow s<strong>and</strong>wich-type immunoassay in micr<strong>of</strong>luidic<br />
devices based on negative dielectrophoresis,” Biosensors & Bioelectronics, vol. 22, no. 11, pp.<br />
2730-2736, May 15, 2007.<br />
[7] S. Grilli, <strong>and</strong> P. Ferraro, “Dielectrophoretic trapping <strong>of</strong> suspended particles by selective<br />
pyroelectric effect in lithium niobate crystals,” Applied Physics Letters, vol. 92, no. 23, 2008.<br />
[8] X. Xiong, A. Busnaina, S. Selvarasah et al., “Directed assembly <strong>of</strong> gold nanoparticle<br />
nanowires <strong>and</strong> networks for nanodevices,” Applied Physics Letters, vol. 91, no. 6, 2007.<br />
[9] A. Kuzyk, B. Yurke, J. J. Toppari et al., “Dielectrophoretic trapping <strong>of</strong> DNA origami,” Small,<br />
vol. 4, no. 4, pp. 447-450, 2008.<br />
[10] B. H. Lapizco-Encinas, S. Ozuna-Chacon, <strong>and</strong> M. Rito-Palomares, “Protein manipulation with<br />
insulator-based dielectrophoresis <strong>and</strong> direct current electric fields,” Journal <strong>of</strong><br />
Chromatography A, vol. 1206, no. 1, pp. 45-51, Oct 3, 2008.<br />
[11] G. B. Salieb-Beugelaar, J. Teapal, J. van Nieuwkasteele et al., “Field-dependent DNA<br />
mobility in 20 nm high nanoslits,” Nano Letters, vol. 8, no. 7, pp. 1785-1790, Jul, 2008.<br />
[12] M. Yang, <strong>and</strong> X. Zhang, “Electrical assisted patterning <strong>of</strong> cardiac myocytes with controlled<br />
macroscopic anisotropy using a micr<strong>of</strong>luidic dielectrophoresis chip,” Sensors <strong>and</strong> Actuators a-<br />
Physical, vol. 135, no. 1, pp. 73-79, 2007.<br />
[13] M. Suzuki, T. Yasukawa, H. Shiku et al., “Negative dielectrophoretic patterning with colloidal<br />
particles <strong>and</strong> encapsulation into a hydrogel,” Langmuir, vol. 23, no. 7, pp. 4088-4094, Mar 27,<br />
2007.<br />
[14] A. Rosenthal, <strong>and</strong> J. Voldman, “Dielectrophoretic traps for single-particle patterning,”<br />
Biophysical Journal, vol. 88, no. 3, pp. 2193-2205, 2005.<br />
[15] E. T. Thostenson, Z. F. Ren, <strong>and</strong> T. W. Chou, “Advances in the science <strong>and</strong> technology <strong>of</strong><br />
carbon nanotubes <strong>and</strong> their composites: a review,” Composites Science <strong>and</strong> Technology, vol.<br />
61, no. 13, pp. 1899-1912, 2001.<br />
109
[16] T. Ueda, M. M. H. Bhulyan, H. Norimatsu et al., “Development <strong>of</strong> carbon nanotube-based gas<br />
sensors for NOx gas detection working at low temperature,” Physica E-Low-Dimensional<br />
Systems & Nanostructures, vol. 40, no. 7, pp. 2272-2277, 2008.<br />
[17] J. H. Lee, J. Kim, H. W. Seo et al., “Bias modulated highly sensitive NO2 gas detection using<br />
carbon nanotubes,” Sensors <strong>and</strong> Actuators B-Chemical, vol. 129, no. 2, pp. 628-631, 2008.<br />
[18] A. A. Pesetski, J. E. Baumgardner, S. V. Krishnaswamy et al., “A 500 MHz carbon nanotube<br />
transistor oscillator,” Applied Physics Letters, vol. 93, no. 12, 2008.<br />
[19] P. Stokes, <strong>and</strong> S. I. Khondaker, “Local-gated single-walled carbon nanotube field effect<br />
transistors assembled by AC dielectrophoresis,” Nanotechnology, vol. 19, no. 17, 2008.<br />
[20] Y. J. Kang, D. Q. Li, S. A. Kalams et al., “DC-Dielectrophoretic separation <strong>of</strong> biological cells<br />
by size,” Biomedical Microdevices, vol. 10, no. 2, pp. 243-249, Apr, 2008.<br />
[21] H. B. Li, Y. N. Zheng, D. Akin et al., “Characterization <strong>and</strong> modeling <strong>of</strong> a micr<strong>of</strong>luidic<br />
dielectrophoresis filter for biological species,” Journal <strong>of</strong> Microelectromechanical Systems,<br />
vol. 14, no. 1, pp. 103-112, Feb, 2005.<br />
[22] B. H. Lapizco-Encinas, B. A. Simmons, E. B. Cummings et al., “Dielectrophoretic<br />
concentration <strong>and</strong> separation <strong>of</strong> live <strong>and</strong> dead bacteria in an array <strong>of</strong> insulators,” Analytical<br />
Chemistry, vol. 76, no. 6, pp. 1571-1579, 2004.<br />
[23] M. D. Vahey, <strong>and</strong> J. Voldman, “An equilibrium method for continuous-flow cell sorting using<br />
dielectrophoresis,” Analytical Chemistry, vol. 80, no. 9, pp. 3135-3143, May 1, 2008.<br />
[24] Z. Y. Wang, O. Hansen, P. K. Petersen et al., “Dielectrophoresis microsystem with integrated<br />
flow cytometers for on-line monitoring <strong>of</strong> sorting efficiency,” Electrophoresis, vol. 27, no. 24,<br />
pp. 5081-5092, Dec, 2006.<br />
[25] J. T. Huang, G. C. Wang, K. M. Tseng et al., “A chip for catching, separating, <strong>and</strong><br />
transporting bio-particles with dielectrophoresis,” Journal <strong>of</strong> Industrial Microbiology &<br />
Biotechnology, vol. 35, no. 11, pp. 1551-1557, Nov, 2008.<br />
[26] R. Krishnan, B. D. Sullivan, R. L. Mifflin et al., “Alternating current electrokinetic separation<br />
<strong>and</strong> detection <strong>of</strong> DNA nanoparticles in high-conductance solutions,” Electrophoresis, vol. 29,<br />
no. 9, pp. 1765-1774, May, 2008.<br />
[27] G. O. F. Parikesit, A. P. Markesteijn, O. M. Piciu et al., “Size-dependent trajectories <strong>of</strong> DNA<br />
macromolecules due to insulative dielectrophoresis in submicrometer-deep fluidic channels,”<br />
Biomicr<strong>of</strong>luidics, vol. 2, no. 2, Apr-Jun, 2008.<br />
[28] L. L. Jia, S. G. Moorjani, T. N. Jackson et al., “Microscale transport <strong>and</strong> sorting by kinesin<br />
molecular motors,” Biomedical Microdevices, vol. 6, no. 1, pp. 67-74, Mar, 2004.<br />
[29] I. Ermolina, J. Milner, <strong>and</strong> H. Morgan, “Dielectrophoretic investigation <strong>of</strong> plant virus<br />
particles: Cow Pea Mosaic Virus <strong>and</strong> Tobacco Mosaic Virus,” Electrophoresis, vol. 27, no.<br />
20, pp. 3939-3948, Oct, 2006.<br />
[30] H. Morgan, M. P. Hughes, <strong>and</strong> N. G. Green, “Separation <strong>of</strong> submicron bioparticles by<br />
dielectrophoresis,” Biophysical Journal, vol. 77, no. 1, pp. 516-525, Jul, 1999.<br />
[31] N. G. Green, H. Morgan, <strong>and</strong> J. J. Milner, “Manipulation <strong>and</strong> trapping <strong>of</strong> sub-micron<br />
bioparticles using dielectrophoresis,” Journal <strong>of</strong> Biochemical <strong>and</strong> Biophysical Methods, vol.<br />
35, no. 2, pp. 89-102, Sep 25, 1997.<br />
[32] M. J. Mendes, H. K. Schmidt, <strong>and</strong> M. Pasquali, “Brownian dynamics simulations <strong>of</strong> singlewall<br />
carbon nanotube separation by type using dielectrophoresis,” Journal <strong>of</strong> Physical<br />
Chemistry B, vol. 112, no. 25, pp. 7467-7477, 2008.<br />
[33] P. Blasi, S. Glovagnoli, A. Schoubben et al., “Solid lipid nanoparticles for targeted brain drug<br />
delivery,” Advanced Drug Delivery Reviews, vol. 59, pp. 454-477, 2007.<br />
[34] A. J. Almeida, <strong>and</strong> E. Souto, Solid lipid nanoparticles as a drug delivery system for peptides<br />
<strong>and</strong> proteins, Advanced Drug Delivery Reviews,Volume 59, Issue 6, 10 July 2007, Pages 478-<br />
490<br />
[35] M. Arruebo, R. Fernández-Pacheco, M. Ricardo Ibarra <strong>and</strong> J. Santamaría, Magnetic<br />
nanoparticles for drug delivery, Nanotoday, Volume 2, Issue 3, June 2007, Pages 22-32<br />
[36] C. Klumpp, K. Kostarelos, M. Prato et al., “Functionalized carbon nanotubes as emerging<br />
nanovectors for the delivery <strong>of</strong> therapeutics,” Biochimica Et Biophysica Acta-Biomembranes,<br />
vol. 1758, no. 3, pp. 404-412, 2006.<br />
110
[37] A. Motayed, M. Q. He, A. V. Davydov et al., “Realization <strong>of</strong> reliable GaN nanowire<br />
transistors utilizing dielectrophoretic alignment technique,” Journal <strong>of</strong> Applied Physics, vol.<br />
100, no. 11, Dec 1, 2006.<br />
[38] M. Dimaki, <strong>and</strong> P. Boggild, “Dielectrophoresis <strong>of</strong> carbon nanotubes using microelectrodes: a<br />
numerical study,” Nanotechnology, vol. 15, no. 8, pp. 1095-1102, Aug, 2004.<br />
[39] D. J. Yang, S. G. Wang, Q. Zhang et al., “Thermal <strong>and</strong> electrical transport in multi-walled<br />
carbon nanotubes,” Physics Letters A, vol. 329, no. 3, pp. 207-213, Aug 23, 2004.<br />
[40] M. P. Hughes, H. Morgan, <strong>and</strong> M. F. Flynn, “The dielectrophoretic behavior <strong>of</strong> submicron<br />
latex spheres: Influence <strong>of</strong> surface conductance,” Journal <strong>of</strong> Colloid <strong>and</strong> Interface Science,<br />
vol. 220, no. 2, pp. 454-457, 1999.<br />
[41] L. Cui, D. Holmes, <strong>and</strong> H. Morgan, “The dielectrophoretic levitation <strong>and</strong> separation <strong>of</strong> latex<br />
beads in microchips,” Electrophoresis, vol. 22, no. 18, pp. 3893-3901, Oct, 2001.<br />
[42] R. Krupke, Hennrich, F., Kappes, M. M., Lohneysen, H., “Surface Conductance Induced<br />
Dielectrophoresis <strong>of</strong> Semiconducting Single-Walled Carbon Nanotubes,” NANO LETTERS,<br />
vol. 4 no. 8, pp. 1395–1399, 2004.<br />
[43] J. Kim, Y. H. Shin, J. H. Yun et al., “A nickel silicide nanowire microscopy tip obtains<br />
nanoscale information,” Nanotechnology, vol. 19, no. 48, Dec 3, 2008.<br />
[44] J. W. Lee, K. J. Moon, M. H. Ham et al., “Dielectrophoretic assembly <strong>of</strong> GaN nanowires for<br />
UV sensor applications,” Solid State Communications, vol. 148, no. 5-6, pp. 194-198, Nov,<br />
2008.<br />
[45] J. Voldman, “Electrical forces for microscale cell manipulation,” Annual Review <strong>of</strong><br />
Biomedical Engineering, vol. 8, pp. 425-454, 2006.<br />
111
Use <strong>of</strong> a Radially Polarized Beam for Ultra-low Energy Threshold<br />
for Cancer Photothermal Therapy with Gold Nanorods [1]<br />
Hong Kang 1 * , Jingliang Li 1 , Baohua Jia 1 , Dru Morrish 1 <strong>and</strong> Min Gu 1<br />
* Presenter<br />
1. Centre for Micro-Photonics , Swinburne University <strong>of</strong> Technology, VIC, Austraila,<br />
3122<br />
We demonstrate that the use <strong>of</strong> a radially polarized beam is more efficient for<br />
both imaging <strong>and</strong> therapy for cancer cells labelled with gold nanorods, compared with<br />
that <strong>of</strong> a linearly polarized beam.<br />
OCIS codes<br />
(180.0180) Microscopy, (170.0170) Medical optics <strong>and</strong> biotechnology<br />
1. Introduction<br />
Advances in the development <strong>of</strong> nanosized materials have created new opportunities <strong>and</strong><br />
applications in biomedical <strong>and</strong> biological fields. Amongst these nanomaterials, metallic nanoparticles<br />
have attracted more interest owing to their flexible synthesis, enhanced collective dipole oscillation<br />
(surface plasmon resonance) <strong>and</strong> their strong luminescence emission when they are exposed to laser<br />
irradiation [1-3]. The luminescence property makes them a novel contrast agent for diagnosing <strong>and</strong><br />
detecting specific cells. Furthermore, it enables plasmonic photothermal therapy (PPTT) in which the<br />
photon energy is converted into heat to induce hyperthermia in adjoining cells. Due to its localised <strong>and</strong><br />
minimum invasive feature, PPTT has become a promising method for cancer treatment.<br />
Distinguished from spherically shaped particles, nanorods exhibit two absorption b<strong>and</strong>s due to the<br />
surface plasmon resonance as shown in Fig. 1. The absorption peak located in the shorter wavelength<br />
region can be attributed to the oscillation along the transverse axis, which experiences relatively<br />
insignificant change with the aspect ratio (length divided by width) <strong>of</strong> nanorods. The other much<br />
stronger absorption peak resulting from the longitudinal oscillation is highly sensitive to the size<br />
variation <strong>and</strong> shifts to a longer wavelength as the aspect ratio increases. The absorption spectra for the<br />
longitudinal mode can be tuned to the near infrared (NIR) region, which is highly desirable for in vivo<br />
applications as NIR excitation experiences less absorption, scattering <strong>and</strong> a larger penetration depth<br />
[4-6]. Fig. 1 illustrates the absorption b<strong>and</strong>s <strong>of</strong> the gold nanorods used in this paper. The transverse<br />
<strong>and</strong> longitudinal absorption peaks at wavelengths 517 nm <strong>and</strong> 780 nm, respectively, can be clearly<br />
identified from Fig. 1.<br />
Due to the surface plasmon property, gold nanorods are most efficiently excited by light polarized<br />
parallel to their longitudinal axis. However, when cells are labelled with gold nanorods, the nanorods<br />
are r<strong>and</strong>omly dispersed <strong>and</strong> orientated. This makes the excitation <strong>of</strong> the nanorods by a linearly<br />
polarized beam less efficiency because only a small portion <strong>of</strong> the r<strong>and</strong>omly oriented nanorods<br />
aligning with the incident polarization direction can be excited. Higher incident laser power is required<br />
to compensate for the less efficient nanorod excitation. Too high an incident power can kill healthy<br />
cells outright. To overcome this challenge, a highly localized focal spot polarized in all three<br />
directions is highly desired. By utilizing such a radially polarized beam, nanorods in every orientation<br />
can be more efficiently excited, which subsequently reduces the required laser power <strong>and</strong> leads to an<br />
ultra-low energy threshold for cancer photothermal therapy. In this paper, a radially polarized beam is<br />
employed for this purpose.<br />
[1]. Published at the Conference <strong>of</strong> Novel Techniques in Microscopy, Vancouver Canada,<br />
April 30, 2009, @2009 Optical Society <strong>of</strong> America
Fig. 1. The transverse <strong>and</strong> longitudinal surface plasmon absorption spectra <strong>of</strong> gold nanorods. The solid<br />
arrows indicate the direction <strong>of</strong> the surface plasmon resonance modes within the nanorod. The dashed<br />
arrows indicate the transverse <strong>and</strong> longitudinal absorption peaks at wavelength 517 nm <strong>and</strong> 780 nm,<br />
respectively.<br />
As shown in Fig. 2, a radially polarized beam possesses even electric field aligning in a radial<br />
direction in free space. Once focused by an objective with high numerical aperture (NA), a strong<br />
longitudinal component is produced due to the depolarization effect [7,8]. As a result, a light field<br />
polarized in all directions is generated composing these three orthogonal electric fields. Thus by using<br />
such a tightly focused radially polarized beam in PPTT nanorods <strong>of</strong> arbitrary orientation can be<br />
efficiently excited. Therefore, the required energy fluence for imaging <strong>and</strong> treatment can be greatly<br />
reduced to a level below the medical st<strong>and</strong>ard for clinical applications.<br />
Fig. 2. Experimental intensity distribution <strong>of</strong> a radially polarized beam. The arrows indicate the<br />
polarization direction <strong>of</strong> the light.<br />
2. Experiments<br />
The radially polarized beam was generated by a radial polarization converter (Arcoptix S.A.). Fig.<br />
2 shows the experimental intensity distribution <strong>of</strong> a typical hollow radial beam in free space. To<br />
examine the polarization pr<strong>of</strong>ile, an analyser is placed rotated to three orientations within the beam.<br />
Fig. 3 shows the obtained intensity patterns at the three different directions. The two high-intensity<br />
113
lobes align with the analyser direction with approximately equal intensities, demonstrating the good<br />
polarization uniformity <strong>and</strong> the good beam quality <strong>of</strong> the converted radially polarized beam.<br />
Fig. 3. Intensity distributions at the output <strong>of</strong> the analyser at three polarization angles (a) 0° (b) 45° (c)<br />
90° (polarization angle is defined as the angle between the incident polarization direction <strong>and</strong> the<br />
polarization direction <strong>of</strong> the analyser). The arrows indicate the polarization direction <strong>of</strong> the analyser.<br />
The in vitro two-photon imaging <strong>and</strong> treatment were both carried out on a confocal scanning<br />
microscopic system (Olympus: IX70). HeLa (cervical cancer) cells mediated with poly(4-<br />
styrenesulfonic acid) (PSS)-coated gold nanorods were used for all experiments [4]. The samples were<br />
imaged <strong>and</strong> treated with femtosecond laser pulses at the wavelength <strong>of</strong> 780 nm generated with an<br />
ultrafast laser with a repetition rate <strong>of</strong> 80 MHz <strong>and</strong> a pulse width <strong>of</strong> approximately 80 fs (Spectra<br />
Physics: Mai Tai). A water immersion objective <strong>of</strong> NA 1.2 was used both for imaging <strong>and</strong> treatment.<br />
Figs. 4(a) <strong>and</strong> 4(b) show the fluorescent images <strong>of</strong> gold nanorods directly on a cover slip excited with<br />
the linearly polarized beam (Fig. 4(a)) <strong>and</strong> the radially polarized beam (Fig. 4(b)), respectively, under<br />
the same power level.<br />
The luminescence intensity under linear polarization excitation is much weaker than that under<br />
radial polarization illumination. In Fig. 4(c), the combined luminescence <strong>of</strong> the nanorods <strong>and</strong><br />
transmission images <strong>of</strong> HeLa cells incubated with PSS coated gold nanorods is presented. It has been<br />
found that by using the radially polarized beam, the energy fluence needed for PPTT was just one fifth<br />
<strong>of</strong> the clinical safety level <strong>of</strong> 100 mJ/cm 2 [9].<br />
Fig. 4. Luminescence <strong>of</strong> gold nanorods excited with (a) a linearly polarized beam, (b) a radially<br />
polarized beam <strong>and</strong> (c) combined luminescence <strong>of</strong> nanorods <strong>and</strong> transmission images <strong>of</strong> HeLa cells<br />
incubated with PSS coated gold nanorods. The scale bars are 10 μm.<br />
3. Summary<br />
In this paper, we have demonstrated that a radially polarized beam is an effective source for PPTT<br />
<strong>of</strong> cancer cells labelled with gold nanorods. By using a tightly focused radially polarized beam,<br />
nanorods r<strong>and</strong>omly oriented on HeLa cells can be efficiently excited leading to a significantly reduced<br />
energy threshold for both imaging <strong>and</strong> treatment.<br />
114
References:<br />
[1] Xiong Liu, Mark C. Lloyd, Inna V. Fedorenko, Priya Bapat, Tatyana Zhukov <strong>and</strong> Qun<br />
Huo, ”Enhanced imaging <strong>and</strong> accelerated photothermalysis <strong>of</strong> A549 human lung cancer cells by<br />
gold nanospheres,” Nanomedicine 3, 617-626 (2008).<br />
[2] Xiaohua Huang, Prashant K. Jain, Ivan H. El-Sayed <strong>and</strong> Mostafa A. El-Sayed,, “Plasmonic<br />
photothermal therapy (PPTT) using gold nanoparticles,” Lasers Med Sci 23, 217-228 (2008).<br />
[3] Jingyi Chen, Danling Wang, Jiefeng Xi, Leslie Au, Andy Siekkinen, Addie Warsen, Zhi-Yuan Li,<br />
Hui Zhang, Younan Xia <strong>and</strong> Xingde Li, “Immuno gold nanocages with tailored optical properties<br />
for targeted photothermal destruction <strong>of</strong> cancer cells,” Nano Lett 7, 1318-1322 (2007).<br />
[4] Jingliang Li, Daniel Day <strong>and</strong> Min Gu, “Ultra-low energy threshold for cancer photothermal<br />
therapy using transferrin-conjugated gold nanorods,” Adv. Mater, 20, 3866-3871 (2008).<br />
[5] Ling Tong, Yan Zhao, Terry B. Huff, Matthew N. Hansen, Alex<strong>and</strong>er Wei <strong>and</strong> Ji-Xin Cheng,<br />
“Gold nanorods mediate tumor cell death by compromising membrane integrity,” Adv Mater 19,<br />
3136-3141 (2007).<br />
[6] Takuro Nidome, Masato Yamagata, Yuri Okamoto, Yasuyuki Akiyama, Hironobu Takahashi,<br />
Takahito Kawano, Yoshiki Katayama <strong>and</strong> Yasuro Niidome, “PEG-modified gold nanorods with a<br />
stealth character for in vivo applications,” Journal <strong>of</strong> Controlled Release 114, 343-347 (2006).<br />
[7] Min Gu, Advanced Optical Imaging Theory (Springer, 2000), Chap. 6.<br />
[8] Baohua Jia, Xiaosong Gan, <strong>and</strong> Min Gu, “Direct measurement <strong>of</strong> a radially polarized focused<br />
evanescent field facilitated by a single LCD,” Optics Express 13, 6821–6827 (2005).<br />
[9] American National St<strong>and</strong>ard for Safe Use <strong>of</strong> the Lasers ANSI Z136.1 (American Laser Institute,<br />
2000).<br />
115
Electrospinning Thermoplastic Polyurethane-contained Collagen<br />
Nan<strong>of</strong>ibers for Tissue Engineering Applications [1]<br />
Rui Chen 1 * , Qinfei Ke 1 , Pujang Shi 3 , Yos Morsi 3 <strong>and</strong> Xiumei Mo 2<br />
* Presenter<br />
1. College <strong>of</strong> textiles, Donghua University, Shanghai, 201620, China<br />
2. College <strong>of</strong> chemistry <strong>and</strong> chemical engineering <strong>and</strong> biological engineering, Donghua<br />
University, Shanghai, 201620, China<br />
3. Biomechanics <strong>and</strong> Tissue Engineering Group, <strong>Faculty</strong> <strong>of</strong> Engineering <strong>and</strong> Industrial<br />
Science, Swinburne University <strong>of</strong> Technology, Hawthorn, VIC3122, Australia<br />
Electrospinning is a new method used in Tissue Engineering field. It can spin fibers in<br />
nanoscale by electrostatic force. A serious <strong>of</strong> thermoplastic polyurethane (TPU)/collagen<br />
blend nan<strong>of</strong>ibrous membranes were prepared with different weight ratios <strong>and</strong> concentrations<br />
via electrospinning. The two biopolymer were all used 1,1,1,3,3,3,-hexafluoro-2-<br />
propanol(HFP) as solvent. The electrospun thermoplastic polyurethane-contained collagen<br />
nan<strong>of</strong>ibers was characterized using scanning electron microscopy (SEM), XPS spectroscopy,<br />
atomic-force microscopy, apparent density <strong>and</strong> porosity measurement, contact angle<br />
measurement, mechanical tensile testing <strong>and</strong> viability <strong>of</strong> Pig iliac endothelial cells (PIECs) on<br />
blended nan<strong>of</strong>iber mats. Our date indicate that fiber diameter was influenced by both polymer<br />
concentration <strong>and</strong> blend weight ratio <strong>of</strong> collagen to TPU. The average diameter <strong>of</strong> nan<strong>of</strong>ibers<br />
gradually decreases with increasing collagen content in the blend. XPS analysis indicates that<br />
collagen is found to be present at the surface <strong>of</strong> blended nan<strong>of</strong>iber. The results <strong>of</strong> porosity <strong>and</strong><br />
contact angle measurement suggest that with the collagen content in the blend system, the<br />
porosity <strong>and</strong> hydrophicity <strong>of</strong> the nan<strong>of</strong>iber mats is greatly improved.We have also<br />
characterized the molecular interactions in TPU/collagen complex by fourier transform<br />
(FTIR). The result could demonstrate that there were intermolecular bonds between the<br />
molecules <strong>of</strong> TPU <strong>and</strong> collagen. The ultimate tensile stress <strong>and</strong> strain were carried out <strong>and</strong> the<br />
dates could also prove the analysis <strong>of</strong> FTIR. The TPU/collagen blend nan<strong>of</strong>ibrous mats were<br />
further investigated as promising scaffold for PIECs culture. The cell proliferation <strong>and</strong> SEM<br />
morphology observations showed that the cells could not only favorably grow well on the<br />
surface <strong>of</strong> blend nan<strong>of</strong>iberous mats, but also able to migrate inside the scaffold with 24 hours<br />
<strong>of</strong> culture. These results suggest tha the blend nan<strong>of</strong>ibers <strong>of</strong> TPU/collagen are designed to<br />
mimic the native extracellular matrix for tissue engineering <strong>and</strong> develop functional<br />
biomaterials.<br />
Keywords<br />
Engineering<br />
Electrospinning, Thermoplastic polyurethane /collagen blend, Nan<strong>of</strong>iber, Tissue<br />
1 Introduction<br />
The technology <strong>of</strong> tissue engineering (TE) aims to generate new or substitute or malfunctioning<br />
<strong>and</strong> could well become an alternative method to whole organ transplantation [1, 2]. So far as we<br />
know, the TE methods have been applied to different types <strong>of</strong> tissue <strong>and</strong> organ such as skin[3],<br />
bone[4], liver[5], intestine[6], heart valve[7], muscle[8] <strong>and</strong> tongue[9]. One <strong>of</strong> the most important<br />
factor in TE is the three-dimensionally porous scaffold, which allows biological activities such as<br />
cell adhering, migrating, growing <strong>and</strong> differentiating to attain a proper integration between cells <strong>and</strong><br />
scaffold for synthesizing a<br />
[1]. Published at Journal <strong>of</strong> Biomaterials Science (2009) 20:1513-1536
new tissue[10]. Most <strong>of</strong> these human organs deposited on fibrous structures with the fibril/fiber size<br />
realigning from nanometer realigning from nanometer to millimeter scale. So nan<strong>of</strong>ibers have now<br />
been extensively used to mimic these natural tissue matrixes. At present, electrospinning is the most<br />
prevalent process that can creates nan<strong>of</strong>ibers through an electrically charged jet <strong>of</strong> polymer solution<br />
or polymer melt. Different processing parameters such as kind <strong>of</strong> polymer, viscosity, surface tension,<br />
jet charge density, temperature <strong>and</strong> humidity control the electrospinning process, especially the<br />
diameter <strong>and</strong> morphology <strong>of</strong> the resulting fibers[11]. Recently, researchers have found that the<br />
nan<strong>of</strong>ibrous structure formed by electrospinning method will improve the function in vitro tissue<br />
regeneration <strong>and</strong> decrease the formation <strong>of</strong> scar tissue[12]. So the scaffolds prepared from<br />
electrospinning method can be considered to mimic the native extracellular matrix (ECM).<br />
Representative polymers including synthetic ones such as poly(lactic-acid)(PLA)[13,<br />
14],poly(glycolic-acid)(PGA)[15], poly(lactic-co-glycolic acid)(PLGA)[16], poly(ε-caprolactone)<br />
(PCL)[17, 18]<strong>and</strong> natural ones such as collagen[19], chitosan[20], gelatin[21] <strong>and</strong> silk[22]have been<br />
electrospun into nan<strong>of</strong>ibers.<br />
Native ECM is the complex <strong>of</strong> polyprotein <strong>and</strong> polysaccharide with nan<strong>of</strong>ibrous structure.<br />
Among the natual biopolymers, collagen as polyprotein has been widely used in TE for its excellent<br />
biocompatibility <strong>and</strong> nonimmunogenicity, but the mechanical property can’t achieve the request <strong>of</strong><br />
native ECM. Thermoplastic polyurethanes (TPUs) are a widely used class <strong>of</strong> polymer with excellent<br />
mechanical properties <strong>and</strong> good biocompatibility, <strong>and</strong> have been evaluated for a variety <strong>of</strong><br />
biomedical applications such as coating materials for brest implants, catheters, <strong>and</strong> prosthetic heart<br />
valve leaflets[23]. Conventional TPU are among biomaterials not intended to degrade but are<br />
susceptible to hydrolytic, oxidative <strong>and</strong> enzymatic degradation in vivo. While the susceptibility <strong>of</strong><br />
TPU to such degradation is a problem for long lasting biomedical implants, it can be deliberately<br />
exploited to design biodegradable polyurethane[24]. The TPU used in this research one kind <strong>of</strong><br />
medical-grade, aliphatic, polyether-based TPUs. It can biodegrade <strong>and</strong> the biostability is better than<br />
poly(ester urethane).<br />
As c<strong>and</strong>idate materials, pure TPU <strong>and</strong> collagen have already been electrospun into nan<strong>of</strong>ibers<br />
as biomaterials respectively[19, 25]. And they all have biological benefits used as TE scaffolds. The<br />
promising study <strong>of</strong> electrospun TPU <strong>and</strong> collagen complex has not been well understood. Especially,<br />
the relationship between processing parameters <strong>and</strong> microstructures in the electrospun nan<strong>of</strong>bers has<br />
not been reported.<br />
In this study, collagen <strong>and</strong> <strong>and</strong> TPU were co-electrospun to produce collagen/TPU composite<br />
nan<strong>of</strong>ibrous scaffolds. We created these electrospun hybrid scaffolds that combined synthetic<br />
material with natural proteins to overcome limitations seen with scaffolds constructed with either<br />
one alone. We characterized the properties <strong>of</strong> these novel scaffolds <strong>and</strong> optimized for porosity,<br />
strength, cell seeding. Collagen/TPU hybrid fibrous scaffolds were determined to provide optimal<br />
fiber diameter, pore size <strong>and</strong> mechanical strength, leading to enhance the seeding <strong>of</strong> the electrospun<br />
scaffolds with cells. The object <strong>of</strong> this project is to study the electrospinning <strong>of</strong> TPU/collagen<br />
complex <strong>and</strong> to develop the biomimetic extracellular matrix for TE application.<br />
2 Materials <strong>and</strong> methods<br />
2.1 Materials<br />
The polymer <strong>of</strong> TPU (Tec<strong>of</strong>lex EG-80A) was purchased from Noveon, Inc.(USA) <strong>and</strong> collagen<br />
(mol.wt, 0.8-1×105Da) was purchased from Sichuan Ming-rang Bio-Tech Co. Ltd (China). The two<br />
materials all used 1,1,1,3,3,3,-hexafluoro-2-propanol (HFP) as solvent, which was brought from<br />
Daikin Industries Ltd( Japan). Pig iliac endothelial cells (PIECs) were obtained from institute <strong>of</strong><br />
biochemistry <strong>and</strong> cell biology (Chinese Academy <strong>of</strong> Sciences, China). Except specially explained, all<br />
culture media <strong>and</strong> reagents were purchased from Gibco Life Technologies CO, USA.<br />
2.2 Scaffold fabrication<br />
The polymers <strong>of</strong> TPU <strong>and</strong> Collagen solutions were dissolved in HFP separately. When they were<br />
prepared already, the two solutions were blended at different weight ratios <strong>and</strong> volume ratios with<br />
sufficient stirring at room temperature before electrospinning.<br />
The pure <strong>and</strong> blend polymer solutions were placed into a 2.5 ml plastic syringe with a bluntended<br />
needle with an inner diameter <strong>of</strong> 0.21mm. The needle was located at a distance <strong>of</strong> 130~150mm<br />
from the grounded collector. A syringe pump (789100C, cole-pamer, America) was employed to feed<br />
117
solutions to the needle tip at a feed rate <strong>of</strong> 1.6~2.0 ml/h. A high electrospinning voltage was applied<br />
between the needle <strong>and</strong> ground collector using a high voltage power supply (BGG6-358,<br />
BMEICO.LTD, China). The applied voltage was between 16 <strong>and</strong> 20KV. The electric field generated<br />
by the surface charge caused the solution drop at the tip <strong>of</strong> the needle to distort into a Taylor cone.<br />
2.4 Characterization <strong>of</strong> collagen-blended TPU nan<strong>of</strong>ibers<br />
The morphologies <strong>and</strong> diameters <strong>of</strong> the nan<strong>of</strong>ibers electrospunned by pure <strong>and</strong> different weight<br />
ratio concentration <strong>of</strong> TPU to collagen were determined with SEM (JEOL, JSM-5600, Japan) at a<br />
accelerated voltage <strong>of</strong> 15KV. The diameter range <strong>of</strong> the fabricated nan<strong>of</strong>ibers was measured based on<br />
the SEM images using an image visualization s<strong>of</strong>tware ImageJ 1.34s (National Institutes <strong>of</strong> Health,<br />
USA).<br />
Surface chemistry analysis <strong>of</strong> the electrospun scaffolds were also analyzed using X-ray<br />
photoelectron spectroscopy (XPS) (Escalab 250; Thermo Scientific Electron, East Grinstead, UK)<br />
equipped with Mg Kα at 1,486.6 eV <strong>and</strong> 150W power at the anode. A survy scan spectrum was taken<br />
<strong>and</strong> the surface elemental compositions relative to carbon were calculated from the peak height with a<br />
correction for atomic sensitivity.<br />
Surface propeties <strong>of</strong> the nan<strong>of</strong>ibers were examined using a a Nanoscope Ⅳ atomic-force<br />
microscope (Digital Instruments), in the tapping mode <strong>and</strong> expressed as height <strong>and</strong> phase images.<br />
2.5 Contact angle measurement<br />
The electrospun mats made from pure TPU <strong>and</strong> TPU/collagen were characterized by the water<br />
contact angle measurement to find the effect <strong>of</strong> addition <strong>of</strong> collagen on the hydrophicity <strong>of</strong> the<br />
materials. The images <strong>of</strong> the droplet on the membrane were visualized through the image<br />
analyzer(OCA40, Datephysics Co, German) <strong>and</strong> the angles between the water droplet <strong>and</strong> the surface<br />
were measured. The measurement used distilled water as the reference liquid <strong>and</strong> was automatically<br />
dropped onto the electrospun mats. To confirm the uniform distribution <strong>of</strong> blend nan<strong>of</strong>iber mats, the<br />
contact angle was measured 3 times from different positions on each mat <strong>and</strong> an averge value was<br />
calculated by statistical method.<br />
2.6 The porosity <strong>of</strong> electrospun mats<br />
The electrospun mats <strong>of</strong> TPU/collagen with different blend ratio at a concentration <strong>of</strong> 8wt%. The<br />
thickness <strong>of</strong> the nan<strong>of</strong>iber mesh was measured with a micrometer (Shanghai, China). The apprent<br />
density <strong>and</strong> porosity were calculated accroding to the following equations [26]:<br />
Apparent density <strong>of</strong> nan<strong>of</strong>iber mats(g/cm 3 )=<br />
Mass <strong>of</strong> nan<strong>of</strong>ibrous mast (g)<br />
2<br />
Thickness <strong>of</strong> nan<strong>of</strong>ibrous mats (cm) × Nan<strong>of</strong>iber mats area (cm )<br />
Porosity <strong>of</strong> nan<strong>of</strong>iber mats (%)=<br />
3<br />
Naonfiber mats apparent density (g/cm )<br />
) × 100%<br />
3<br />
bulk density fo raw TPU/collagen (g/cm )<br />
2.6 FTIR spectra<br />
Electrospun TPU nan<strong>of</strong>ibers, collagen nan<strong>of</strong>iber <strong>and</strong> TPU/collagen with different weight ratios<br />
were prepared for the FTIR test on AVATAR 380 FTIR instrument (Thermo Electron, Waltham, MA).<br />
All spectra were recorded by an absorption mode in the wave length range <strong>of</strong> 4000-500 cm -1 .<br />
2.7 Mechanical measurement<br />
Mechanical measurements were carried out by applying tensile test loads to samples which were<br />
prepared from electrospun ultra fine non-woven fiber mats. The specimens were electrospun with<br />
different blend ratios <strong>of</strong> collagen to TPU (0:1, 1:3, 1:1, 3:1). In this study, the test was performed in<br />
ambient temperature at 20 ℃ <strong>and</strong> relative humidity <strong>of</strong> 65%. Four specimens <strong>of</strong> each sample were<br />
prepared according to the method described by Huang et al [21]. First, a white paper was cut into<br />
templates with a planner dimension <strong>of</strong> width × gauge length=10mm×30mm <strong>and</strong> then double side tapes<br />
were glued onto the top <strong>and</strong> bottom areas <strong>of</strong> one side. Secondly, the aluminum foil was carefully<br />
scraped <strong>of</strong>f <strong>and</strong> single side tapes were applied onto the griping areas as end-tabs. Finally, the resulting<br />
specims were conducted on a commercial materials testing machine(H5K, Hounsfield, Engl<strong>and</strong>) with<br />
a load cell <strong>of</strong> 10N, <strong>and</strong> the elongation speed is 10mm/min.<br />
118
2.8 Viability <strong>and</strong> morphology study <strong>of</strong> PIEC on nan<strong>of</strong>iber mats<br />
Pig iliac endothelial cells(PIECs) were cultured in DMEN medium with 10% fetal serum, <strong>and</strong><br />
100 units/ml <strong>and</strong> 100units/ml streptomycin in humidified incubator under st<strong>and</strong>ed culture conditions<br />
(5% CO 2 , content at 37℃), <strong>and</strong> the medium was replaced every three days. Mats <strong>of</strong> blend nan<strong>of</strong>ibers<br />
with different ratios <strong>of</strong> TPU <strong>and</strong> collgen were electrospun on circular 14mm glass coverslips. After<br />
the coverslips with nan<strong>of</strong>ibers were prepared already, they were dried in vacuum for over one week to<br />
release the residua solvents. Then the mats were placed in desiccator to crosslink using glutaraldehyde<br />
(25%water solution) steam for two days. Whereafter the blend nan<strong>of</strong>ibers mats were dried in vacuum<br />
for over two weeks to release the residual glutaradehyde. And the next step is to fixed the coverslips<br />
into 24-well plates with stainless ring. Before seeding cells, fiber scaffolds were sterilized with 75%<br />
alcohol solution, which were placed with phosphate-buffered saline solution (PBS) after 2 hours.<br />
Cells viability on the nan<strong>of</strong>ibers were determined by MTT method. Briefly, the cell <strong>and</strong><br />
nan<strong>of</strong>iber complex was incubated with 5mg/ml 3-(4,5)-dimethylthiahiazo (-z-y1)-3,5-diphenytetrazoliumromide<br />
(MTT) for 4h. Thereafter the culture media were extracted <strong>and</strong> added 150ul<br />
Dimethyl Sulfoxide (DMSO) to have low speed surge for about 10 minutes. When the crystal was<br />
sufficient resolved, aliquots were pipetted into the wells <strong>of</strong> a 96-well plate <strong>and</strong> placed into a Enzymelabeled<br />
Instrument (MK3, Thermo, USA), <strong>and</strong> the absorbance at 490nm for each well was measured.<br />
For the cell viability test, endothelium cells were seeded onto nan<strong>of</strong>iber mats (n=3) with<br />
different weight ratio <strong>of</strong> collagen to TPU <strong>and</strong> control glass coverslips at a density <strong>of</strong> 5×10 4 cells/cm 2 .<br />
On days 1, 3, 5 <strong>and</strong> 7 after cell seeding, untached cells were quantified by MTS kit ((C0009, Beyotime<br />
Institute if Biotechnology, China) <strong>and</strong> Enzyme-labeled Instrument (MK3, Thermo, USA).<br />
After 24 hours <strong>of</strong> culturing, the electrospun fibrous scaffolds with cells (density is 1.5×10 5<br />
cells/cm 2 ) were examined by SEM. To prepare the cell-cultured samples for SEM observation, the<br />
scaffolds were rinsed twice with PBS in the first place, followed by fixation with 4% glutaraldehyde<br />
water solution for 2h <strong>and</strong> then the samples were again rinsed twice with PBS. Thereafter dehydrated in<br />
graded concentrations <strong>of</strong> ethanol (30, 50, 70, 80, 90, 95 <strong>and</strong> 100). Finally, they were dried in vacuum<br />
overnight. The dry cellular constructs were coated with gold sputter <strong>and</strong> observed under the SEM at a<br />
voltage <strong>of</strong> 15KV.<br />
2.9 Statistics analysis<br />
Statistics analysis was performed using origin 7.5 (Origin lab Inc, USA). Values (at least<br />
triplicate) were averaged <strong>and</strong> expressed as means ± st<strong>and</strong>ard deviation (SD). Statistical differences<br />
were determined by the anaylsis <strong>of</strong> One-Way ANOVA <strong>and</strong> differences were considered statistically<br />
significant at p
Blend nan<strong>of</strong>iber diameters distribution <strong>of</strong> different blend ratios were further determined <strong>and</strong> the<br />
results were shown in Fig.2. Nan<strong>of</strong>iber diameters <strong>of</strong> every weight ratios were calculated from the<br />
diameter <strong>of</strong> 100 nan<strong>of</strong>ibers each sample which were directly measured from SEM photographs. From<br />
numerical statement <strong>of</strong> average diameters <strong>and</strong> st<strong>and</strong>ard deviation we can see that the fiber diameter<br />
gradually deceased with increasing collagen content in the blend. This phenomenon could be<br />
explained as the conductivity increase <strong>of</strong> the blend solution. It has been known that in electrospinning,<br />
the force that causes the stretching <strong>of</strong> the solution is due to the repulsive forces between the charges on<br />
the electrospinning jet[29]. Collagen is typical amphiprotic macromolecule electrolyte. When collagen<br />
was affiliated, more ions were formed in the blend solution. So a higher discharge density could be<br />
carried by the electrospinning jet. The conductivity <strong>of</strong> the solution could be increased by the addition<br />
<strong>of</strong> ions. On the other h<strong>and</strong>, the increased charge carried by the solution will increase the stretching <strong>of</strong><br />
the solution. And the increased in the stretching <strong>of</strong> the solution also could tend to yield fibers <strong>of</strong><br />
smaller diameter.<br />
In order to investigate the factor <strong>of</strong> solution concentration, we selected one weight ratio (80/20,<br />
TPU/collagen) to prepare this blend solution with six different concentrations.<br />
Fig 1 <strong>and</strong> 2 SEM images <strong>of</strong> nan<strong>of</strong>iber mats at concentration <strong>of</strong> 6wt% with different blend ratios <strong>of</strong><br />
TPU to Collagen(T/C) <strong>and</strong> their fiber diameter distribution<br />
Then the mixed solutions were subjected to the electrospinning process under the same<br />
processing condition. Fig.3 shows SEM images <strong>of</strong> TPU/collagen fibers with the concentration which<br />
varied from 1.5% to 9% at equal interval. Their fiber diameter distributions were calculated in the<br />
same method <strong>and</strong> the results were shown in Fig.4. From the SEM images we could observe that the<br />
smooth fibers could be obtained when the concentrations <strong>of</strong> the solutions were located between 4.5%<br />
<strong>and</strong> 7.5%. At the same time, we could find that either the concentrations were too low or too high,<br />
more beads structure could appear. It is established that chain entanglements in the solution plays an<br />
important role in the fiber diameter <strong>and</strong> morphology. The increase <strong>of</strong> the solution concentration will<br />
result in great polymer chain entanglements. Thus, if the concentration is to low, bead-only structure<br />
will be produced due to a lack <strong>of</strong> chain entanglements in the solution[30]. On the contrary, if the<br />
concentration is too high, that will make it very difficult to pump the solution through the syringe<br />
needle. Morevoer, when the concentration is too high, the solutions may dries at the tip <strong>of</strong> the needle<br />
before the electrospinning can be initiated.<br />
120
Fig 3 <strong>and</strong> 4 SEM images <strong>of</strong> electrospun TPU/collagen (80/20) blend nan<strong>of</strong>ibers as a functions <strong>of</strong><br />
different concentrations.(A: 1.5wt%; B: 3wt%; C: 4.5%; D: 6%; E: 7.5%; F: 9% ) <strong>and</strong> their fiber<br />
diameter distribution<br />
The surface chemistry changes <strong>of</strong> blended nan<strong>of</strong>ibers was verified by XPS spectroscopy. Fig.5<br />
shows the XPS survey scans <strong>of</strong> nan<strong>of</strong>ibrous scaffold surfaces. Table 1 shows the atomic ratios <strong>of</strong><br />
carbon, nitrogen <strong>and</strong> oxygen on pure <strong>and</strong> blended nan<strong>of</strong>ibers with weight ratio <strong>of</strong> 50:50. As expected,<br />
the blend-electrospun scaffold collagen/TPU shows three peaks corresponding to C1s (binding energy,<br />
286 eV), N1s (binding energy, 400 eV) <strong>and</strong> O1s (binding energy, 532 eV). The atomic ratios <strong>of</strong> carbon,<br />
nitrogen <strong>and</strong> oxygen on nan<strong>of</strong>ibrous scaffolds calculated from XPS survey scan spectra are shown in<br />
Table 1. The oxygen content (20.29%) <strong>of</strong> TPU nan<strong>of</strong>iber surface was not changed by the<br />
incorporation <strong>of</strong> collagen (collagen/TPU, 19.46). On the other h<strong>and</strong>, nitrogen (7.4%) found on the<br />
blend nan<strong>of</strong>iber surface was much higher than TPU nan<strong>of</strong>iber (1.77%). This result illustrate that<br />
collagen was found to be present on the surface <strong>of</strong> blended nan<strong>of</strong>ibers. However, it is hard to say that<br />
collagen was located on the surface <strong>of</strong> blended nan<strong>of</strong>ibers, because both TPU <strong>and</strong> collagen have<br />
Nitrogen content in their chemical structures. The existence <strong>of</strong> collagen molecules <strong>of</strong> collagen on the<br />
surface <strong>and</strong> inside the nan<strong>of</strong>ibers provides sustained cell recognition sigals with polymer degradation,<br />
which is important for cell function development .<br />
Fig 5 X-ray photoelectron Spectroscopy survy scan spectra <strong>of</strong> (a) collagen nan<strong>of</strong>iber, (b) TPU<br />
nan<strong>of</strong>iber, (c)collagen/TPU nan<strong>of</strong>iber(50:50,w/w)<br />
To study the surface morphologies <strong>of</strong> TPU, collagen <strong>and</strong> TPU/collagen nan<strong>of</strong>iber surface, AFM<br />
image was examined using a height mode (Fig. 6). From the analysis <strong>of</strong> AFM, we found that the<br />
121
surface <strong>of</strong> TPU nan<strong>of</strong>ibers is much smooth than collagen nan<strong>of</strong>ibers. The blend nan<strong>of</strong>biers wers also<br />
even, but there were a lot <strong>of</strong> lines on the surface at the surface <strong>of</strong> blend nan<strong>of</strong>ibers.<br />
Fig 6 AFM images represented by height mode:(a) TPU fiber (b) collagen fiber (c) TPU/collagen<br />
fiber(50:50,w/w)<br />
3.2 Water contact measurements<br />
The hydrophobic (or hydrophilic) nature <strong>of</strong> a substrate has a direct impact on the avenue <strong>of</strong> its<br />
usage. Fig.7 gives the contact-angle dates variation <strong>and</strong> the water convex shape on different<br />
electrospun fibrous mats measured with a surface analysis instrument. As expected, collagen showed<br />
much better wettability than the TPU because <strong>of</strong> its hydrophilic group. The electrospun pure TPU mats<br />
showed an angle around 134.1° which could seemed to be as lyophobic. With the increasing collagen<br />
component in the blend nan<strong>of</strong>iber, the various electrospun mats exhibited the contact angle around 60-<br />
80°. The result indicated that the hydrophilicity <strong>of</strong> the membranes based on TPU was significantly<br />
improved by adding the collagen ingredient. Such a property will be extremely useful for creation <strong>of</strong><br />
tissue with cellularity throughout the scaffold by cell culture.<br />
Fig 7 Water contact angle <strong>of</strong> electrospun TPU/collagen nan<strong>of</strong>ibrous mats with different blend ratios.<br />
(Inset this figure shows the variation shapes <strong>of</strong> contact angle on different mats. (A) T/C =1:0; (B) T/C<br />
= 3:1; (C) T/C =1:1; (D) T/C =1:3; (E) T/C =0:1; (F) T/C =0:1.Date are means±SD (n=3)<br />
The hydrophilicity <strong>of</strong> tissue engineering scaffolds is very important for their application. Many<br />
works have reported that cells attached <strong>and</strong> spread more easily <strong>and</strong> effectively on surfaces with proper<br />
hydrophilicity than on hydrophobic surfaces[31]. From the cell culture result, we could find that there<br />
the improved surface hydrophilicity was expected to facilitate cell attachment <strong>and</strong> proliferation. With<br />
collagen ingredient increasing in the blend nan<strong>of</strong>ibers, the hydrophilicity was greatly improved. The<br />
blend ratio <strong>of</strong> TPU/collagen from 100:0 to 75:25, the contact angle had a great change from 134 to 82°.<br />
Although there were no sigificant difference among these blend nan<strong>of</strong>ibrous mats in the first 24 hours,<br />
they all showed better proliferation rate at 72, 120 <strong>and</strong> 144 hours. Furthermore, Cell attachment to the<br />
substratum is almost always mediated by extracellular matrix (ECM) proteins absorbed on the culture<br />
surface. ECM proteins are present in the serum used in most cell culture applications. Numerous<br />
122
studies have indicated that surface properties such as surface hydrophilicity <strong>and</strong> charge could affect<br />
cell attachment by influencing the ability <strong>of</strong> the substratum to absorb protein <strong>and</strong>/or by altering the<br />
conformation <strong>of</strong> the absorbed protein[32, 33].<br />
3.3 Porosity <strong>of</strong> nan<strong>of</strong>ibrous mats<br />
During the process <strong>of</strong> electrospinning, the thickness <strong>of</strong> the blended nan<strong>of</strong>ibers was controlled by<br />
the deposition time if the electrospinning parameters were all fixed. We can see that all the apparent<br />
density <strong>of</strong> blended nan<strong>of</strong>ibers was in the range <strong>of</strong> 0.30-0.60g/cm 3 <strong>and</strong> it slowly decreased with the<br />
collagen increasing in the blend.<br />
As the electrospun TPU blended collagen nan<strong>of</strong>ibous mats are having highly pore size, using the<br />
apprent density <strong>of</strong> all different mats <strong>and</strong> the bulk density <strong>of</strong> TPU/collagen blends, we could calculate<br />
the porosity <strong>of</strong> the blended nan<strong>of</strong>ibers. However, it was not easy to accurately measure the bulk<br />
density <strong>of</strong> the TPU/collagen blend for it was a high porous material <strong>and</strong> contained a mixture <strong>of</strong><br />
different ingredient, so the bulk density <strong>of</strong> the blends were estimated to be 1.04g/cm 3 on the basis <strong>of</strong><br />
the bulk density <strong>of</strong> TPU. From Fig.8 we can see that the porosity <strong>of</strong> TPU/collagen nan<strong>of</strong>ibers had a big<br />
increase with the increasing weight ratio <strong>of</strong> collagen to TPU <strong>and</strong> we can also got the conclusion from<br />
Fig.1 that the pore size <strong>of</strong> D, E, F were larger than A, B, C. The porosity increase <strong>of</strong> nan<strong>of</strong>iber mats by<br />
adding collagen in the blend can also be explained as the increase <strong>of</strong> solution conductivity[34]. From<br />
the fiber diameter analysis we could know that with the increase <strong>of</strong> the conductivity <strong>of</strong> the solution,<br />
the finer <strong>of</strong> the fiber would be. On the other h<strong>and</strong>, with more electronic charges carried by the<br />
nan<strong>of</strong>iber, there would be stronger repulsive forces among fibers during depositing to the collector.<br />
Fig 8 The porosity <strong>of</strong> TPU/collagen blend nan<strong>of</strong>ibrous mats with various blend ratios.<br />
Pores play an important role in determining the physical <strong>and</strong> chemical properties <strong>of</strong> porous<br />
substrates <strong>and</strong> have a deterministic effect on the performance <strong>of</strong> membranes, catalysts, adsorbents etc.<br />
On the other side, the mats must have a large pore volume fraction as well as an interconnected pore<br />
work to permit the transport, multiplication <strong>and</strong> metabolites. High porosity is propitious to cell<br />
adhesion on the mat, promotes extracellular matrix regeneration. Thus, porosity is an essential factor<br />
for materials to minic extracellular matrix in tissue engineering. From the analysis <strong>of</strong> the blended<br />
nan<strong>of</strong>ibrous mats, they all showed high porosity <strong>and</strong> adequate mechanical property for cell ingrowth.<br />
Date are representatives <strong>of</strong> three independent experiments <strong>and</strong> all date points are plotted as<br />
means±SD (n=3)<br />
3.4 FTIR spectra<br />
Fig 9 depicts the spectra <strong>of</strong> TPU/ collagen blend with different weight ratio. It can be seen that<br />
although the blends prepared in different ratios, these spectras are quite simlar to each other <strong>and</strong> the<br />
peaks appear at the same b<strong>and</strong>s. The difference among the three spectra is the intensity <strong>of</strong> some peaks.<br />
The characteristic peaks <strong>of</strong> thermoplastic polyurethane at 3290 cm-1, 2930 cm-1 <strong>and</strong> 2860 cm-1 can<br />
be found in the range <strong>of</strong> 3500-2500 cm-1,while the peaks between 2000-1000 cm-1, such as those at<br />
123
1640 cm-1, 1540 cm-1 <strong>and</strong> 1240 cm-1, can be assigned to the characteristic absorption <strong>of</strong> collagen. It<br />
is noticed that the characteristics absorption peak <strong>of</strong> TPU at 1530 cm-1 <strong>and</strong> 1220 cm-1 were overlaped<br />
with characeristic b<strong>and</strong> <strong>of</strong> collagen, <strong>and</strong> the peak at 1710 cm-1 is belong to the stretching vibration <strong>of</strong><br />
carbonyl group. The spectras show both characteristics peaks <strong>of</strong> electrospun collagen <strong>and</strong> TPU. So this<br />
is the evidence to illuminate that there may be no reaction between collagen <strong>and</strong> TPU.<br />
Fig 9 FTIR spectra <strong>of</strong> TPU <strong>and</strong> collagen blend nan<strong>of</strong>ibrous membranes (a: TPU/collagen=75:25; b:<br />
TPU/collagen=50:50; c: TPU/collagen=25:75).<br />
3.5 Mechanical test<br />
The mechanical properties <strong>of</strong> blend nan<strong>of</strong>ibers are important for their successful applications in<br />
tissue engineering. Collagen <strong>and</strong> TPU blend nan<strong>of</strong>ibers were electrospun into 0.5mm thick fiber mats<br />
to measure their mechanical properties. Fig.11 shows the tensile stress-strain curve <strong>of</strong> collagen/TPU<br />
nan<strong>of</strong>iber mats with various blend ratios. The electrospun TPU material gives a characteristic response<br />
for elastomeric materials-sigmoidal in shape. It showed a very s<strong>of</strong>t <strong>and</strong> flexible characteristic with low<br />
Young’s modulus <strong>and</strong> the high elongation at break <strong>of</strong> 365%. With the increase <strong>of</strong> collagen content in<br />
the blend, the initial modulus <strong>of</strong> the mats became large. This phenomenon indirect that electrospun<br />
collagen has a plastic property which is different from TPU. From the stress-strain curves, we can see<br />
that with collagen content increasing, the materials mechanical properties have changed from<br />
elastomeric to plasic. Therefore, we can adjust the mechanical property to meet the requirement in<br />
practice through changing the blend ratio <strong>of</strong> collagen to TPU.<br />
To design an ideal scaffold, various factors should be considered such as pore size <strong>and</strong><br />
morphology, mechanical properties versus porosity, surface properties <strong>and</strong> appropriate<br />
biodegradability. Of these factors, the importance <strong>of</strong> mechanical properties on cell growth is<br />
particularly obvious in tissues such as bone, cartilage, blood vessels, tendors, heart valve <strong>and</strong> muscles.<br />
Different natual tissues have different mechanical propertiies. For example, heart valve <strong>of</strong> human <strong>and</strong><br />
porcine have different mechanical tensile stress <strong>and</strong> ultimate strain. TPU has high tensile strength,<br />
good tear <strong>and</strong> abrasion resistance. But its Young’s modulus is very low. Collagen has high Young’s<br />
modulus <strong>and</strong> excellent biocompatiblity, but its tensile strenght <strong>and</strong> strain are poor. In order to mimic<br />
the mechanical properties in the radial <strong>and</strong> circuit direction, we need to adjust the blend ratios <strong>of</strong> TPU<br />
<strong>and</strong> collagen. From the analysis, we can find that with the collagen ingredient increasing, the Young’s<br />
modulus, tensile strength <strong>and</strong> tensile strain changed. Blended nan<strong>of</strong>ibers may have improved<br />
mechanical strength compared with pure non-crosslinked collagen nan<strong>of</strong>ibers, which could combine<br />
the advantages <strong>of</strong> both synthetic <strong>and</strong> natual materials.<br />
124
Fig 10 Tensile Stress curve <strong>of</strong> electrospun TPU <strong>and</strong> TPU/collagen blend nan<strong>of</strong>ibrous mats (A)<br />
electrospun TPU, (B) electrospun T/C=75:25, (C) electrospun T/C=50:50, (D) electrospun T/C=25:75.<br />
3.6 Viablity <strong>of</strong> cells on collagen-blended TPU nan<strong>of</strong>ibers<br />
The viability <strong>of</strong> PIECs on days 1, 3, 5 <strong>and</strong> 7 after seeding on blend nan<strong>of</strong>ibers with various<br />
weight ratios <strong>of</strong> TPU to collagen is shown in Fig.11. The viability <strong>of</strong> cells cultured on blend<br />
nan<strong>of</strong>ibers was compared with that <strong>of</strong> cells cultured on coverslips (control), pure collagen nan<strong>of</strong>ibers<br />
<strong>and</strong> TPU nan<strong>of</strong>ibers. It was revealed that all the nan<strong>of</strong>iber mats had good cell viability than coverslips<br />
<strong>and</strong> cell viability had no obvious difference among the blend nan<strong>of</strong>iber mats at 24 hours, but cell<br />
proliferation was very fast <strong>and</strong> the highest MTT absorbance index could reach 0.7. On days 3, among<br />
the different blend nan<strong>of</strong>ibers, the one with the TPU to collagen blend weight ratios <strong>of</strong> 3:1 had showed<br />
better cell viability than the others, at the same time blend ratios <strong>of</strong> 3:1 had most excellent cell<br />
viability on days 5 <strong>and</strong> 7. Meanwhile, the MTT absorbance index on day 7 descended a little could be<br />
ascribed that there was no more available area for cells to proliferate. The fibers diameters, porosity<br />
<strong>and</strong> mehchanical property are very important for cell growing <strong>and</strong> migration. Among the blending<br />
system, the diameters, porosity <strong>and</strong> mechanical stress changed along with the addition <strong>of</strong> collagen.<br />
Thus, compared with pure collagen <strong>and</strong> TPU, blend nan<strong>of</strong>ibers could provide better growth condition<br />
for cell proliferation. In our studies, the one with TPU to collagen weight ratios <strong>of</strong> 3:1 might <strong>of</strong>fer the<br />
most suitable qualification for cell culture.<br />
Fig12 shows SEM micrographs <strong>of</strong> PIECs on TPU(A), TPU/collagen=75:25(B),<br />
TPU/collagen=50:50(C), TPU/collagen=25:75(D) mats after seeding for 24hours. As SEM images<br />
shown, cells could spread well both on on TPU nan<strong>of</strong>ibers <strong>and</strong> TPU/collagen blend nan<strong>of</strong>ibers. Fig.9<br />
(B) <strong>and</strong> (C) behave better cell proliferation than others with TPU/collagen blend ratios <strong>of</strong> 3:1 <strong>and</strong> 1:1.<br />
125
Fig 11 Viability <strong>of</strong> PIECs cultured on TPU, collagen/TPU <strong>and</strong> collagen nan<strong>of</strong>ibers. PIECs cultured on<br />
the coverslips acted as a negative control. Date are representative <strong>of</strong> three independent experiments<br />
<strong>and</strong> all date points are ploted as means±SD,*P
well as on a usual nan<strong>of</strong>ibrous scaffold, especially collagen/TPU blend ratio <strong>of</strong> 1:3 is the most<br />
appropriate for endothelial cell proliferation. Blended nan<strong>of</strong>ibers can improve bioactivity relative to<br />
pure polymer nan<strong>of</strong>ibers <strong>and</strong> possess the potential to refine the composition <strong>of</strong> nan<strong>of</strong>ibers readily by<br />
adjusting ingredients according to the cell type. All these dates strongly suggested the potential<br />
application <strong>of</strong> blended nan<strong>of</strong>ibers as vascular engineering scaffolds with good endothelialization.<br />
Acknowledgement<br />
This research was supported by the National High-Tech Research <strong>and</strong> Development Programme <strong>of</strong><br />
China (2008AA03Z305), Open foundation <strong>of</strong> State Key Laboratory for Modification <strong>of</strong> Chemical<br />
Fibers <strong>and</strong> Polymer Materials, National Nature Science Foundation <strong>of</strong> China (30570503), Natural<br />
Science Foundation <strong>of</strong> Shanghai (07ZR14001) <strong>and</strong> the 111 project (B07024)"<br />
References:<br />
1. Langer R, Vacanti JP. TISSUE ENGINEERING. Science 1993;260(5110):920-926.<br />
2. Wein<strong>and</strong> C, Pomerantseva I, Neville CM, Gupta R, Weinberg E, Madisch I, et al. Hydrogel-beta-<br />
TCP scaffolds <strong>and</strong> stem cells for tissue engineering bone. Bone 2006;38(4):555-563.<br />
3. Bannasch H, Fohn M, Unterberg T, Bach AD, Wey<strong>and</strong> B, Stark GB. Skin tissue engineering.<br />
Clinics in Plastic Surgery 2003;30(4):573-+.<br />
4. Li CM, Vepari C, Jin HJ, Kim HJ, Kaplan DL. Electrospun silk-BMP-2 scaffolds for bone tissue<br />
engineering. Biomaterials 2006;27(16):3115-3124.<br />
5. Feng QL. Materials selection <strong>and</strong> scaffold construction for liver tissue engineering. Pricm 5: The<br />
Fifth Pacific Rim International Conference on Advanced Materials <strong>and</strong> Processing, Pts 1-5<br />
2005;475-479:2391-2394.<br />
6. Chen MK, Beierle EA. Animal models for intestinal tissue engineering. Biomaterials<br />
2004;25(9):1675-1681.<br />
7. Neuenschw<strong>and</strong>er S, Hoerstrup SP. Heart valve tissue engineering. Transplant Immunology<br />
2004;12(3-4):359-365.<br />
8. Levenberg S, Rouwkema J, Macdonald M, Garfein ES, Kohane DS, Darl<strong>and</strong> DC, et al.<br />
Engineering vascularized skeletal muscle tissue. Nature Biotechnology 2005;23(7):879-884.<br />
9. Bunaprasert T, Hadlock T, Marler J, Kobler J, Cowan D, Faquin W, et al. Tissue engineered<br />
muscle implantation for tongue reconstruction: A preliminary report. Laryngoscope<br />
2003;113(10):1792-1797.<br />
10. Zhang YZ, Venugopal J, Huang ZM, Lim CT, Ramakrishna S. Characterization <strong>of</strong> the surface<br />
biocompatibility <strong>of</strong> the electrospun PCL-collagen nan<strong>of</strong>ibers using fibroblasts.<br />
Biomacromolecules 2005;6(5):2583-2589.<br />
11. Nair LS, Bhattacharyya S, Laurencin CT. Development <strong>of</strong> novel tissue engineering scaffolds via<br />
electrospinning. Expert Opinion on Biological Therapy 2004;4(5):659-668.<br />
12. Webster TJ, Waid MC, McKenzie JL, Price RL, Eji<strong>of</strong>or JU. Nano-biotechnology: carbon<br />
nan<strong>of</strong>ibres as improved neural <strong>and</strong> orthopaedic implants. Nanotechnology 2004;15(1):48-54.<br />
13. Kim K, Yu M, Zong XH, Chiu J, Fang DF, Seo YS, et al. Control <strong>of</strong> degradation rate <strong>and</strong><br />
hydrophilicity in electrospun non-woven poly(D,L-lactide) nan<strong>of</strong>iber scaffolds for biomedical<br />
applications. Biomaterials 2003;24(27):4977-4985.<br />
14. Yang F, Murugan R, Wang S, Ramakrishna S. Electrospinning <strong>of</strong> nano/micro scale poly(L-lactic<br />
acid) aligned fibers <strong>and</strong> their potential in neural tissue engineering. Biomaterials<br />
2005;26(15):2603-2610.<br />
15. Bol<strong>and</strong> ED, Wnek GE, Simpson DG, Pawlowski KJ, Bowlin GL. Tailoring tissue engineering<br />
scaffolds using electrostatic processing techniques: A study <strong>of</strong> poly(glycolic acid)<br />
electrospinning. Journal <strong>of</strong> Macromolecular Science-Pure <strong>and</strong> Applied Chemistry<br />
2001;38(12):1231-1243.<br />
16. Ayutsede J, G<strong>and</strong>hi M, Sukigara S, Ko F. Carbon nanotube reinforced Bombyx mori nan<strong>of</strong>iber<br />
composites by the electrospinning process. Mechanical Properties <strong>of</strong> Bioinspired <strong>and</strong> Biological<br />
Materials 2005;844:281-286.<br />
127
17. Li WJ, Tuli R, Okafor C, Derfoul A, Danielson KG, Hall DJ, et al. A three-dimensional<br />
nan<strong>of</strong>ibrous scaffold for cartilage tissue engineering using human mesenchymal stem cells.<br />
Biomaterials 2005;26(6):599-609.<br />
18. Yamane S, Iwasaki N, Majima T, Funakoshi T, Masuko T, Harada K, et al. Feasibility <strong>of</strong><br />
chitosan-based hyaluronic acid hybrid biomaterial for a novel scaffold in cartilage tissue<br />
engineering. Biomaterials 2005;26(6):611-619.<br />
19. Matthews JA, Wnek GE, Simpson DG, Bowlin GL. Electrospinning <strong>of</strong> collagen nan<strong>of</strong>ibers.<br />
Biomacromolecules 2002;3(2):232-238.<br />
20. Desai K, Kit K, Li J, Zivanovic S. Morphological <strong>and</strong> surface properties <strong>of</strong> electrospun chitosan<br />
nan<strong>of</strong>ibers. Biomacromolecules 2008 Mar;9(3):1000-1006.<br />
21. Huang ZM, Zhang YZ, Ramakrishna S, Lim CT. Electrospinning <strong>and</strong> mechanical<br />
characterization <strong>of</strong> gelatin nan<strong>of</strong>ibers. Polymer 2004;45(15):5361-5368.<br />
22. Min BM, Lee G, Kim SH, Nam YS, Lee TS, Park WH. Electrospinning <strong>of</strong> silk fibroin nan<strong>of</strong>ibers<br />
<strong>and</strong> its effect on the adhesion <strong>and</strong> spreading <strong>of</strong> normal human keratinocytes <strong>and</strong> fibroblasts in<br />
vitro. Biomaterials 2004;25(7-8):1289-1297.<br />
23. Pedicini A, Farris RJ. Mechanical behavior <strong>of</strong> electrospun polyurethane. Polymer<br />
2003;44(22):6857-6862.<br />
24. Tatai L, Moore TG, Adhikari R, Malherbe F, Jayasekara R, Griffiths I, et al. Thermoplastic<br />
biodegradable polyurethanes: The effect <strong>of</strong> chain extender structure on properties <strong>and</strong> in-vitro<br />
degradation. Biomaterials 2007;28:5407-5417.<br />
25. Demir MM, Yilgor I, Yilgor E, Erman B. Electrospinning <strong>of</strong> polyurethane fibers. Polymer 2002<br />
May;43(11):3303-3309.<br />
26. He W, Yong T, Teo WE, Ma ZW, Ramakrishna S. Fabrication <strong>and</strong> endothelialization <strong>of</strong><br />
collagen-blended biodegradable polymer nan<strong>of</strong>ibers: Potential vascular graft for blood vessel<br />
tissue engineering. Tissue Engineering 2005;11(9-10):1574-1588.<br />
27. Pham QP, Sharma U, Mikos AG. Electrospinning <strong>of</strong> polymeric nan<strong>of</strong>ibers for tissue engineering<br />
applications: A review. Tissue Engineering 2006;12(5):1197-1211.<br />
28. Heydarkhan-Hagvall S, Schenke-Layl<strong>and</strong> K, Dhanasopon AP, R<strong>of</strong>ail F, Smith H, Wu BM, et al.<br />
Three-dimensional electrospun ECM-based hybrid scaffolds for cardiovascular tissue<br />
engineering. Biomaterials 2008;29(19):2907-2914.<br />
29. Yagi K, Michibayashi N, Kurikawa N, Nakashima Y, Mizoguchi T, Harada A, et al.<br />
Effectiveness <strong>of</strong> fructose-modified chitosan as a scaffold for hepatocyte attachment. Biological &<br />
Pharmaceutical Bulletin 1997;20(12):1290-1294.<br />
30. Eda G, Liu J, Shivkumar S. Flight path <strong>of</strong> electrospun polystyrene solutions: Effects <strong>of</strong> molecular<br />
weight <strong>and</strong> concentration. Materials Letters 2007;61(7):1451-1455.<br />
31. Vanwachem PB, Beugeling T, Feijen J, Bantjes A, Detmers JP, Vanaken WG. INTERACTION<br />
OF CULTURED HUMAN-ENDOTHELIAL CELLS WITH POLYMERIC SURFACES OF<br />
DIFFERENT WETTABILITIES. Biomaterials 1985;6(6):403-408.<br />
32. Dhiman HK, Ray AR, P<strong>and</strong>a AK. Characterization <strong>and</strong> evaluation <strong>of</strong> chitosan matrix for in vitro<br />
growth <strong>of</strong> MCF-7 breast cancer cell lines. Biomaterials 2004;25(21):5147-5154.<br />
33. Chatelet C, Damour O, Domard A. Influence <strong>of</strong> the degree <strong>of</strong> acetylation on some biological<br />
properties <strong>of</strong> chitosan films. Biomaterials 2001;22(3):261-268.<br />
34. Zong XH, Kim K, Fang DF, Ran SF, Hsiao BS, Chu B. Structure <strong>and</strong> process relationship <strong>of</strong><br />
electrospun bioabsorbable nan<strong>of</strong>iber membranes. Polymer 2002;43(16):4403-4412.<br />
35. Jiang Zhi YK-j, Li Shu-fen, CHOW W.K. Study <strong>of</strong> FTIR spectra <strong>and</strong> thermal analysis <strong>of</strong><br />
polyurethane. Spectroscopy <strong>and</strong> Spectral Analysis 2006;26:624-628.<br />
36. Jackson M, Choo LP, Watson PH, Halliday WC, Mantsch HH. BEWARE OF CONNECTIVE-<br />
TISSUE PROTEINS - ASSIGNMENT AND IMPLICATIONS OF COLLAGEN<br />
ABSORPTIONS IN INFRARED-SPECTRA OF HUMAN TISSUES. Biochimica Et Biophysica<br />
Acta-Molecular Basis <strong>of</strong> Disease 1995;1270(1):1-6.<br />
37. Chen ZG, Mo XM, Qing FL. Electrospinning <strong>of</strong> collagen-chitosan complex. Materials Letters<br />
2007;61(16):3490-3494.<br />
128
Uniform Coating <strong>of</strong> WO x on TiO 2 Nanotubes for Enhanced Electrochromic<br />
Performance<br />
Jian Zhen Ou 1 * , Hai Dong Zheng 1 , Kourosh Kalantar-zadeh 1<br />
* Presenter<br />
1. School <strong>of</strong> Electrical <strong>and</strong> Computer Engineering, RMIT University, Australia<br />
Both WOx <strong>and</strong> TiO 2 are very good electrochromic materials due to their strong electronphonon<br />
interaction capability. Thus, the combination <strong>of</strong> these two materials could possibly<br />
enhance the electrochromic performance. In our work, WOx was uniformly coated on TiO 2<br />
nanotubes using electrodeposition method from the solution <strong>of</strong> 0.5M Na 2 WO 4. . SEM <strong>and</strong><br />
EDX characterisation techniques were used to determine the surface morphology <strong>and</strong> film<br />
composition respectively. Electrochromic measurement was also conducted <strong>and</strong> the device<br />
performance was evaluated based on optical power reflective density. It was also compared<br />
with the performance <strong>of</strong> conventional RF sputter WO3 thin film coated on FTO glass.<br />
129
Nanoparticle trapping by surface plasmon resonance assisted thermal<br />
forces<br />
Jingzhi Wu 1 * , Xiaosong Gan 1<br />
* Presenter<br />
1. Centre for Micro-Photonics, Swinburne University <strong>of</strong> Technology, VIC, Austraila, 3122<br />
Surface plasmon resonance <strong>of</strong> metal nanostructures allows strong field intensity<br />
enhancement <strong>and</strong> field gradients, which is promising for trapping <strong>and</strong> manipulation <strong>of</strong><br />
nanoparticles. However, metallic nanostructures generate heat in the presence <strong>of</strong><br />
electromagnetic radiation especially under plasmon resonance. Here, we report an approach<br />
for nanoparticles trapping explores thermal forces induced by temperature gradients in the<br />
medium surrounding the nanostructure. Simulation results show that light absorptions <strong>of</strong> the<br />
nanostructure generate great temperature gradients corresponding to strong thermal forces.<br />
This suggests stable nanometric trapping can be realized with thermal forces in nanostructures.<br />
130
Union as a Social Regulator <strong>of</strong> Market Risk?<br />
--Empirical Evidence from Pr<strong>of</strong>essional Union Leader Program in B City<br />
Youqing Fan 1 * , Peter Gahan 1<br />
* Presenter<br />
1. Work <strong>and</strong> Employment Rights Research Centre, Department <strong>of</strong> Management, Monash<br />
University, Caulfield East Victoria 3145, Australia<br />
Polanyi <strong>and</strong> his followers summarized a ‘double movement’ theory to underst<strong>and</strong> the<br />
interaction <strong>of</strong> marketization <strong>and</strong> social protectionism during Engl<strong>and</strong>’s transformation in 18th<br />
century. In the recent 30 years, China is also experiencing a similar marketization expansion,<br />
according to Polanyi; it is supposed to be a social protectionism developed as a<br />
countermovement.<br />
Despite Chinese unions’ influence <strong>and</strong> membership has been seriously undermined<br />
during the SOE reform, there has been an attempt by some regional <strong>of</strong>fices <strong>of</strong> the All-China<br />
Federation <strong>of</strong> Trade Unions (ACFTU) to recruit pr<strong>of</strong>essional leaders, <strong>and</strong> establish unions<br />
independently from management. This attempt does have its potential to make Chinese union<br />
a social regulator in the rising private sector.<br />
This paper draws empirical evidence from pr<strong>of</strong>essional union leader program in B city to<br />
examine the question that ‘to what extent does this pr<strong>of</strong>essional union leader program has<br />
made union as a social regulator to protect workers from market risks?’<br />
1. Introduction<br />
Writing near the mid-point <strong>of</strong> the last century, Karl Polanyi wrote his seminal work, The Great<br />
Transformation (1944[1957]), which sought to provide an underst<strong>and</strong>ing <strong>of</strong> Engl<strong>and</strong>’s transformation.<br />
Central to his analysis was the Polanyi’s analysis <strong>of</strong> these processes have proved enduring,<br />
particularly over the last decade, as a useful approach to examining processes <strong>of</strong> globalisation (e.g.,<br />
Munck, 2002; St<strong>and</strong>ing, 2007) <strong>and</strong> rise <strong>of</strong> new transnational social movements seeking to reverse or<br />
ameliorate the new risks associated with globalisation (Edelman, 2005; Evans, 2008).<br />
To what extent have Chinese unions developed an effective countermovement to the expansion <strong>of</strong><br />
market? There are, <strong>of</strong> course, stark differences between China’s economic transformation <strong>and</strong><br />
Engl<strong>and</strong>’s industrial revolution. But we would suggest that Polanyi’s conceptual framework for<br />
underst<strong>and</strong>ing the economic processes associated with industrialisation <strong>and</strong> subsequent responses to it<br />
– notably the idea <strong>of</strong> the double movement <strong>and</strong> social embeddedness <strong>of</strong> economic institutions –<br />
provide a useful approach for thinking about China’s economic transformation <strong>and</strong> the social<br />
responses to this transformation. Here, we do not provide any attempt to analyse the transition from a<br />
Polanyian perspective (for an early attempt to incorporate Polanyi ideas into such an analysis, see Nee,<br />
1992; 1996; <strong>and</strong> more recently Deyo <strong>and</strong> Ağarton, 2007). Our aim is to sketch briefly Polanyi’s key<br />
ideas <strong>and</strong> examine the role <strong>of</strong> pr<strong>of</strong>essional union leader program in protecting workers against market<br />
risks in China’s new economy.<br />
Thus, the paper will: firstly, introduce Polanyi’s ‘Double Movement’ theory; secondly, briefly<br />
summarize China’s economic transition <strong>and</strong> changing role <strong>of</strong> Chinese trade union; thirdly, discuss the<br />
role <strong>of</strong> pr<strong>of</strong>essional union leader program in protecting workers against market risks in China’s new<br />
economy; <strong>and</strong> <strong>final</strong>ly make some concluding comments.<br />
2. Polanyi <strong>and</strong> the Concept <strong>of</strong> “Double Movement”<br />
In his book, The Great Transformation (1944), Karl Polanyi described the impact <strong>of</strong><br />
industrialisation on established institutions that had regulated production <strong>and</strong> labour markets.<br />
Industrialisation brought with it an extension <strong>of</strong> market-based regulation to new areas <strong>of</strong> economic <strong>and</strong><br />
social relations, including the labour market. Consequently, the regulation <strong>of</strong> labour through the<br />
132
system <strong>of</strong> guilds <strong>and</strong> government regulation gave way to a growing reliance on markets <strong>and</strong> price<br />
signals, permitting employers to hire <strong>and</strong> fire when required, without any broader obligation for<br />
worker welfare. While industrialisation <strong>and</strong> market liberalisation may have improved overall<br />
economic welfare, it subordinated all other relationships to economic concerns, increasing worker<br />
exposure to risk. The unleashing <strong>of</strong> market capitalism, Polanyi posited, gave rise a ‘countermovement’,<br />
a movement for greater social protection against market forces:<br />
“It can be personified as the action <strong>of</strong> two organising principles <strong>of</strong> society... The one was the<br />
principle <strong>of</strong> economic liberalism, aiming at the establishment <strong>of</strong> a self-regulating market... using<br />
largely laisez-faire <strong>and</strong> free trade as its methods; the other was the principle social protection<br />
aiming at the conservation <strong>of</strong> man <strong>and</strong> nature...using protective legislation, restrictive<br />
associations, <strong>and</strong> other instruments <strong>of</strong> intervention as its methods (Polanyi, 1944[1957]: p. 132).<br />
Labour unions were one form that the second <strong>of</strong> these organising principles manifested itself as a<br />
‘countermovement’ to the expansion <strong>of</strong> markets <strong>and</strong> free trade as institutions regulating economic life.<br />
Unions, along with other groups, sought to regulate the terms on which employers could hire <strong>and</strong> fire<br />
workers, as well as to agitate for legislation that afforded greater social protection. This<br />
countermovement reflected the inevitably process by which the extension <strong>of</strong> market relations became<br />
“dis-embedded” socially, disconnected from its social <strong>and</strong> political institutions, thereby generating<br />
insecurity <strong>and</strong> social anxiety (Munck, 2002: p. 18). Counter movements formed a reaction to these<br />
effects; <strong>and</strong> were attempts by social actors to re-embed markets into a more sustainable framework<br />
(Dale 2008). These opposing forces, Polanyi argued, co-exist in an uneasy tension, both having the<br />
potential to undermine the logic <strong>of</strong> the other. Just as ‘blind market relations’ resulted in dis-embedded<br />
social institutions, social protectionism undermined the efficiency <strong>and</strong> effectiveness <strong>of</strong> markets.<br />
Scholars seeking to translate Polanyi’s logic <strong>of</strong> the double movement to the twenty first century<br />
have typically begun with the notion <strong>of</strong> globalisation (Edelman, 2005; St<strong>and</strong>ing, 2007; Evans, 2008),<br />
although as Evans (2008, p.273) in particular has pointed out Polanyi’s critique <strong>of</strong> industrialism is<br />
squarely centred on the role <strong>of</strong> countermovements within relatively autonomous nation states.<br />
Consequently, the scope for countermovement <strong>and</strong> new forms <strong>of</strong> protectionism under globalization<br />
necessitate social actors mobilising at a transnational level. As such, countermovements against<br />
globalisation are more problematic <strong>and</strong> are difficult to sustain.<br />
3. China’s Economic Transition <strong>and</strong> the Changing Role <strong>of</strong> Unions<br />
Transition to a market economy in China has now extended over a thirty year period. This process<br />
has been widely documented by a number <strong>of</strong> scholars in the research on China’s economic transition<br />
(e.g., Naughton, 1995; 2007; Wu, 2005; Qian, 2003). It has consisted <strong>of</strong> successive waves <strong>of</strong> reforms<br />
<strong>and</strong>, in contrast to the ‘shock therapy’ administered to the comm<strong>and</strong> economies <strong>of</strong> Eastern Europe<br />
(McMillan 2004), the Chinese approach to transition has been a gradualist one (Child <strong>and</strong> Tse, 2001).<br />
The transition process has typically been characterized as consisting <strong>of</strong> two distinct phases<br />
(Naughton, 2007). In the first phase, reformers established a ‘dual track’ economic system in which<br />
reforms allowed for an increasing level <strong>of</strong> market activity alongside the planned economy (Lin et al.,<br />
1998; Lau et al., 2000). While this phase allowed for growing competition in product markets through<br />
the formation <strong>of</strong> town <strong>and</strong> village enterprises (TVEs), domestic private enterprises (POEs), joint<br />
venture enterprises (JVEs) with foreign multinationals, <strong>and</strong> wholly foreign owned enterprises (FOEs),<br />
the activities <strong>of</strong> SOEs remained virtually untouched <strong>and</strong> the development <strong>of</strong> competitive labor markets<br />
was weak. Then, from 1992, a series <strong>of</strong> reforms were introduced intended to integrate these dual tracks<br />
through the strengthening <strong>of</strong> market institutions, particularly in the area <strong>of</strong> labor <strong>and</strong> capital markets,<br />
<strong>and</strong> exposing SOEs to greater market competition through restructuring, privatization <strong>and</strong> corporate<br />
governance reforms (Qian, 2003).<br />
As a consequence <strong>of</strong> these reforms, China has experienced a transformation in the institutional,<br />
technological <strong>and</strong> organizational foundations <strong>of</strong> industry (Naughton, 2007). New industries have been<br />
created, <strong>and</strong> existing ones have undergone major structural reform. These reforms included policies<br />
<strong>and</strong> institutional changes intended to promote great labour mobility across sectors <strong>and</strong> regions,<br />
incentives for greater direct foreign investment (DFI), the privatization <strong>of</strong> state-owned enterprises<br />
(SOEs) <strong>and</strong> the promotion <strong>of</strong> domestic private firms. These developments have, in turn, been<br />
133
associated with a reinvention <strong>of</strong> institutions governing industrial relations <strong>and</strong> labor management,<br />
which have inevitably involved the creation <strong>of</strong> more conventional labour market (Shen, 2008; Frenkel<br />
<strong>and</strong> Kuruvilla, 2002). Consistent with the more general policy orientation toward market<br />
liberalization, these policy shifts were aimed at creating a free labour market <strong>and</strong> enabling individual<br />
enterprises to determine a broad range <strong>of</strong> issues at the workplace level, including recruitment <strong>and</strong><br />
dismissal, wages <strong>and</strong> other forms <strong>of</strong> remuneration, <strong>and</strong> investments in training –see Table 1 for a<br />
summary <strong>of</strong> major labour law reforms. These reforms have not only altered the structure <strong>and</strong> dynamics<br />
<strong>of</strong> China’s economy, but have also pr<strong>of</strong>oundly altered the exposure <strong>of</strong> workers to market risk. In this<br />
context our aim is to consider the emergence <strong>of</strong> new forms <strong>of</strong> union organisation as a<br />
‘countermovement’ against the extension <strong>of</strong> the market, providing workers with new forms <strong>of</strong> social<br />
protection.<br />
Table 1. Summary <strong>of</strong> Main Labour Law Reforms in China, 1992-2008<br />
Law/Regulations<br />
Major Contribution<br />
Labour Union Law (1992)<br />
Stipulating the status, obligations <strong>and</strong> rights <strong>of</strong> trade union in<br />
new economic system<br />
Regulations On Enterprises Labour<br />
Disputes Settlement(1993)<br />
Labour Law (1994)<br />
Labour Union Law (2001)<br />
Labour Contract Law (2008)<br />
Labour Dispute, Mediation <strong>and</strong><br />
Arbitration Law (2008)<br />
Setting up framework for labour dispute resolution in new<br />
economic system<br />
Basic legal rules for labour market <strong>and</strong> <strong>of</strong> labour relations<br />
mediation<br />
More emphasis on unions’ role <strong>of</strong> worker representation <strong>and</strong><br />
participation in labour process<br />
More pro-labour provisions:<br />
-Workplace <strong>and</strong> occupational health provisions;<br />
-Work time <strong>and</strong> holiday provisions;<br />
-Compensative payment for pre-maturity contract<br />
termination.<br />
Provisions to make arbitration procedures more accessible:<br />
-Arbitration terminated in the first instance;<br />
-Free <strong>of</strong> charge for arbitration;<br />
-Employers have more ‘burden <strong>of</strong> pro<strong>of</strong>’ responsibilities in<br />
arbitration.<br />
The Role <strong>of</strong> Unions in Socialist China<br />
The peak union organization, the All-China Federation <strong>of</strong> Trade Union (ACFTU) was established<br />
as the formal union movements under the leadership <strong>of</strong> the Communist Party <strong>of</strong> Chinese (CPC<br />
hereafter), <strong>and</strong> incorporated other union movements. Officially, “the trade unions <strong>of</strong> China are a<br />
voluntarily organized mass organizations <strong>of</strong> Chinese working class led by the CPC <strong>and</strong> formed by the<br />
staff <strong>and</strong> members voluntarily, are the bridges <strong>and</strong> transmission belt linking the Party <strong>and</strong> the masses<br />
<strong>of</strong> workers <strong>and</strong> staff members, are the important social pillar <strong>of</strong> the state power, <strong>and</strong> are the<br />
representatives <strong>of</strong> the interests <strong>of</strong> the trade union members workers <strong>and</strong> staff” (Chinese Union Charter,<br />
2003).<br />
The ACFTU formally functions as both a representative institution <strong>of</strong> workers <strong>and</strong> a state<br />
institution. In private workplaces, grassroots unions are largely affiliated to the management (Shen,<br />
2007, p.236-237). Chinese unions assumed three objectives:<br />
1) to assist state administrative function;<br />
2) to collaborate with enterprise management to improve production efficiency; <strong>and</strong><br />
3) to represent <strong>and</strong> protect the interests <strong>of</strong> employees (Ge, 2007, p.2).<br />
All trade union organisations at the provincial, local <strong>and</strong> workplace levels are required by<br />
law to be affiliated to the ACFTU.<br />
134
Chinese Unionism in Reform Period<br />
As noted above, the economic reforms over the last 30 years have been associated with significant<br />
changes in the industrial relations framework in China. The major legal reforms have formally<br />
provided workplace unions with an exp<strong>and</strong>ed role in collective bargaining <strong>and</strong> greater capacity to<br />
represent members in the workplace. What is not clear, however, is whether these developments have<br />
been associated with any substantial change in the role <strong>of</strong> unions at the workplace, or the nature <strong>of</strong> the<br />
relationship between unions <strong>and</strong> their members.<br />
The available evidence indicated that the restructure <strong>of</strong> Chinese industry has created a number <strong>of</strong><br />
adverse effects for unions, most notably in relation to union membership. Figure 1 graphs union<br />
membership for the period 1980 to 2005. This figure shows that from the mid 1990s, membership fell<br />
in absolute terms before increasing again after 1999. A second major decline in membership is then<br />
evident in 2003, after which membership increased again. This decline in membership during the<br />
1990s has been attributed to the effects <strong>of</strong> the economic reforms, particularly the restructuring <strong>of</strong> state<br />
owned enterprises (SOEs). Taylor <strong>and</strong> Li (2005) estimate that between 1998 <strong>and</strong> 2005, SOEs<br />
dismissed approximately 60 percent <strong>of</strong> their workforce, or around 30 million employees. Although<br />
there are no reliable statistics, the evidence also indicates that unions have found it difficult to recruit<br />
<strong>and</strong> retain members in foreign invested forms (Chan, 2006).<br />
Following this period <strong>of</strong> restructuring, the ACFTU sought to reverse the decline member by<br />
organise the emerging private sector (Chan, 2006). This is evident in both the union membership<br />
figures <strong>and</strong> the decline <strong>and</strong> subsequent growth in workplace union organisations. Both <strong>of</strong> the two<br />
indicators are summarised in Figure 1. This data likewise reveals the decline in union organisation<br />
during the mid 1990s <strong>and</strong> subsequent recovery after the turn <strong>of</strong> the century.<br />
Figure 1. Chinese Trade Union Organization <strong>and</strong> Membership Change Trend (1980-2000)<br />
Given the limited capacity <strong>of</strong> organizing foreign enterprises rapidly, Chan (2006) reports that the<br />
ACFTU has no choice but accept management compliant unionism. The main pattern <strong>of</strong> workplace<br />
union branches organizing strategy in foreign-funded enterprises is “top-down”, <strong>of</strong> which process <strong>and</strong><br />
basic features has been summarized by Chan (2006) as follows:<br />
“The district-level union would have sought management approval <strong>and</strong> cooperation to set up<br />
a union branch. Once an agreement was struck, management <strong>and</strong> the local union would have<br />
decided together on a mid-level PRC Chinese manager to serve as the union chair, without a<br />
union election. After the fact, an announcement would have been made to the employees about<br />
the formation <strong>of</strong> a new union branch, or in some cases, no announcement would have been made<br />
at all… [Such union chairs] have no power or independence from management. More <strong>of</strong>ten than<br />
not, this “union” does not even perform the traditional welfare functions it does in state-owned<br />
135
enterprises…There obviously is no collective bargaining or other actions that we associate with<br />
unions.” (Chan, 2006)<br />
This subordinate role has been reinforced by unions’ financial <strong>and</strong> personnel dependence on the<br />
enterprise (Chan, 1998, pp.134-140; Shen, 2007, p.236). As a consequence, unions have not been<br />
capable or willing to represent workers, but have generally taken a pro-management or at least not<br />
pro-labour when labour disputes happen (Taylor <strong>and</strong> Li, 2005, pp.3-4). As a consequence, the fast<br />
growing in private sector unionism has not been associated with effective representation <strong>of</strong> workers<br />
interests in the workplace.<br />
4. Empirical Evidence--Pr<strong>of</strong>essional Union Organizer/Leader Program in B City<br />
Faced with a new type <strong>of</strong> capitalism -- neo-liberal globalization -- the surge <strong>of</strong> unionism in China<br />
shows signs <strong>of</strong> taking different forms from the past to counter the new risks. There are some reports<br />
<strong>of</strong> a recent attempt by regional <strong>of</strong>fices <strong>of</strong> the ACFTU to recruit pr<strong>of</strong>essional union leaders <strong>and</strong> form<br />
region-based unions, which is independent from management. In particular, this paper examines the<br />
role <strong>of</strong> these unions in protecting labour from market risks.<br />
Fieldwork Method<br />
Empirical evidence <strong>of</strong> this case is mainly originated from a major fieldwork. The fieldtrip involves:<br />
1) a major interview with district level, sub-district level union leaders <strong>and</strong> pr<strong>of</strong>essional union leaders;<br />
2) a special interview with 2 pr<strong>of</strong>essional leaders; <strong>and</strong> 3) a half-day long participative observation in<br />
pr<strong>of</strong>essional union leaders’ daily work. Government policy documents are also used as secondary data<br />
source.<br />
Case Description<br />
Since 2002, a number <strong>of</strong> regional branches <strong>of</strong> the ACFTU started to try out two top-down trials<br />
aiming at organizing workers <strong>and</strong> form independent leadership from management in workplace/local<br />
unions. Among these attempts, the practice in P sub-district, S district, B city is a unique combination<br />
<strong>of</strong> recruiting pr<strong>of</strong>essional union organizers <strong>and</strong> leaders, which represents characteristics <strong>of</strong> both the<br />
pr<strong>of</strong>essional leader <strong>and</strong> organizer program. The critical innovations are:<br />
1. Recruit organizers outside the firms to organize rank-<strong>and</strong>-file workers <strong>and</strong> establish<br />
local united unions;<br />
2. Get the organizers elected as union leaders, <strong>and</strong> get them paid by various levels <strong>of</strong><br />
unions rather than by management.<br />
P sub-district is a suburb right lies on the border <strong>of</strong> rural <strong>and</strong> urban area <strong>of</strong> B city with 110<br />
thous<strong>and</strong> populations. Among all the 900 employers in this sub-district, around 600 are small business.<br />
In recent years, urban expansion transfers all the former farmers in the sub-district into urban residents.<br />
Some <strong>of</strong> them set up their own business in various occupations with 2-8 employees.<br />
Faced with the complex situation, P sub-district union leader initiated an innovative form <strong>of</strong><br />
grassroots union, which is called “Small Firms United Grassroots Union”. The sub-district level union<br />
leader explains why they adopt this new form <strong>of</strong> grassroots union:<br />
We have lots <strong>of</strong> petty shops in our area. These shops are too tiny to be organized as an<br />
individual grassroots union, since at least 25 employees are required by the Trade Union Law<br />
for a workplace to be eligible for setting up an individual grassroots union. If we set up unions<br />
in these petty firms, it could leads to two different situations. One possibility is that the boss<br />
holds a concurrent post as union president, which will undermine union’s role <strong>of</strong> maintaining<br />
workers’ interests. Especially, these petty unions are mostly run by former farmers, who lack<br />
<strong>of</strong> sense <strong>of</strong> conforming to labour legislations. The other possibility is rank-<strong>and</strong>-file workers<br />
take the position <strong>of</strong> workplace union president, which will put themselves in danger <strong>of</strong> being<br />
fired by their boss when they try to bargain with boss over labour relations issues, which is<br />
simply because they are hired by, <strong>and</strong> get remuneration from, their boss. Both <strong>of</strong> the<br />
possibilities will not be favourable for union’s worker representation.<br />
136
Obviously, the trade union leaders realize that if they want P sub-district union to represent<br />
workers, firstly, they should organize those small firms to cover the workers; <strong>and</strong> secondly, make<br />
grassroots union leader independent from management both in personnel <strong>and</strong> in finance. Consequently,<br />
after consulting with district level union leaders, they worked out a solution: first, recruit union<br />
organizers outside the firms <strong>and</strong> get them paid by upper level union; second, dispatch them to organize<br />
the tiny firms, <strong>and</strong> then sign collective contract with the firm employers; third, during the organizing<br />
campaign, union member representatives are selected from the firms <strong>and</strong> convened to attend a ‘small<br />
firm grassroots union establishment congress’ <strong>and</strong> directly elect the organizers as united grassroots<br />
union leader; <strong>final</strong>ly, use revisit as main method to detect labour complaint <strong>and</strong> maintaining workers’<br />
rights.<br />
The above blueprint forms a framework for the pr<strong>of</strong>essional union leader practice. In the following<br />
paragraphs, we will describe the implementation <strong>of</strong> the practice in detail.<br />
Personnel <strong>and</strong> Financial Sources <strong>of</strong> Organizing Campaign<br />
To secure personnel independence, initially, P sub-district level union leader intended to recruit<br />
‘union assistants’ from society. However, there is no special funding for recruiting such union<br />
assistants within the union system. As a result, the city <strong>and</strong> district level union leaders requested<br />
Municipal Labour Bureau for using the Social Welfare Fund, which is administrated by the Bureau, to<br />
fund recruitment <strong>of</strong> union leaders. Finally, the Bureau approved the request on the condition that the<br />
recruited assistant must be the “4050” unemployed. 1<br />
Since 2007, 27 “4050” unemployed party member have been selected <strong>and</strong> recruited by sub-district<br />
level unions as union assistants. Municipal Labour Bureau pays salary (at the minimum wage level in<br />
B city) <strong>and</strong> social insurance contributions for them. P sub-district also immediately recruited 2 union<br />
assistants after securing the financial source. The recruited leaders are both SOE laid-<strong>of</strong>f workers.<br />
Organizing Process<br />
P sub-district is divided by an ancient river into two sub-areas. The numbers <strong>of</strong> small businesses in<br />
each sub-area are nearly the same. Therefore, before the proposed organizing campaign, sub-district<br />
level union leader divided the whole sub-district into two zones, ‘river east’ <strong>and</strong> ‘river west’. The<br />
assistants are separately responsible for organizing <strong>and</strong> setting up united grassroots union in 1 zone.<br />
The main organizing targets are small businesses with less than 8 employees.<br />
Formal organizing campaign was initiated in May 2007 with a focus on organizing the river east<br />
zone. The campaign adopted a door to door organizing method, which involves ask each employer in<br />
the zone to participate in the proposed united grassroots union. In the process, normally, union<br />
assistants firstly try to talk to bosses <strong>and</strong> explain the function <strong>of</strong> Chinese unions <strong>and</strong> benefits <strong>of</strong><br />
participating in union. After gaining the permission from bosses, assistants would require bosses to<br />
sign an application form for participating union <strong>and</strong> a prepared collective agreement, <strong>and</strong> then acquire<br />
a copy <strong>of</strong> their business licence. After 3 months organizing, about 75 firms have been organized in this<br />
manner.<br />
Union Election<br />
In order to convene union establishment congress <strong>and</strong> elect union leaders, union assistants also<br />
require each unionized firm to nominate a rank-<strong>and</strong>-file employee as union member representative<br />
during organizing campaign. Finally, in the already organized 75 firms, 50 representatives were<br />
nominated. Nearly 4/5 <strong>of</strong> them are rank-<strong>and</strong>-file employees. The rest 8-9 representatives are bosses,<br />
who are all party members selected by the union assistants. This measure secures quality <strong>of</strong> member<br />
representative team.<br />
1 The “4050” unemployed mainly refers to those middle-aged former SOE workers who got unemployed in SOE<br />
reform. Normally, the female workers are in their 40s, <strong>and</strong> male workers are in their 50s. Basically, they are<br />
associated with low job skills except their original job skills, while still have 1-10 years to be eligible for pension<br />
benefits. Thus, it is quite critical while hard for them to find a job in labour market. Thus, the government<br />
prioritizes the 4050 unemployed to take casual positions in public sector, such as union assistant, traffic assistant,<br />
community assistant etc.<br />
137
Union assistants participate in organizing campaign, thus they are quite familiar with union<br />
members. Once they become union leaders, they are more likely to win trust <strong>of</strong> members.<br />
Union Funding<br />
Lacking <strong>of</strong> enough funding becomes the most serious problem for the new united grassroots unions.<br />
The main financial source <strong>of</strong> the united grassroots unions comes from employers’ contribution <strong>of</strong><br />
union expenditure, which is 2% <strong>of</strong> the total payroll. However, nearly all the participant firms don’t<br />
transfer union dues to their union accounts. S district level union leader attributes this to the weak<br />
Trade Union Law, which doesn’t stipulate any penalty for union expenditure arrears. Other sources <strong>of</strong><br />
funding are almost unavailable: unions are not allowed to receive any donation; financial support from<br />
the government is quite limited, <strong>and</strong> only focuses on support city <strong>and</strong> district level unions.<br />
Union Activities<br />
The primary activities <strong>of</strong> the united grassroots union mainly include signing collective agreements<br />
<strong>and</strong> revisiting small firm workplaces.<br />
Firstly, during the organizing campaign, collective agreements were signed between the proposed<br />
union <strong>and</strong> employers, which includes collective contract, wage agreement <strong>and</strong> female worker<br />
agreement. The content is all the same for different firms. Terms <strong>and</strong> conditions in the agreement is<br />
almost a repeat <strong>of</strong> Chinese Labour Law <strong>and</strong> Municipal Wage Bylaw <strong>of</strong> B City over minimum wage <strong>and</strong><br />
working condition issues. Sub-district level union leader explained:<br />
The union is just a mass organization, it doesn’t have any power to force firms to sign<br />
collective contract. So, in private firms, a feasible way is to lower the collective contract terms<br />
in exchange for bosses’ willingness <strong>of</strong> signing collective contracts. After signing the contract,<br />
at least, we can monitor the implementation <strong>of</strong> collective contract <strong>and</strong> guarantee the minimum<br />
wage <strong>and</strong> basic working conditions for our members.<br />
Secondly, the union presidents/union assistants pay revisits to workplace, <strong>and</strong> have informal<br />
discussion with workers. The revisit bears multiple roles: first, to get to know their members’ needs<br />
<strong>and</strong> concerns at their workplace; second, to monitor the implementation <strong>of</strong> collective agreements; third,<br />
do some labour legislation propag<strong>and</strong>a.<br />
Normally, if there is a labour dispute detected during revisit, united grassroots union president<br />
would call on employer or manager for a mediation meeting. If dispute cannot be mediated, then<br />
united grassroots union president will report this to the district level legal aid centre, <strong>and</strong> lawyers there<br />
will help the member out by representing him/her at labour tribunal or court. Some labour disputes<br />
have been successfully mediated <strong>and</strong> solved by the union. For instance, one union member wanted to<br />
get medical reimbursement <strong>of</strong> his kid, the union coordinated this issue with local residence community,<br />
<strong>and</strong> <strong>final</strong>ly got the reimbursement for the member.<br />
At present, few members go to grassroots union to address their problems on their own initiative.<br />
However, the grassroots leaders believe that if they continue the revisit practice, members will get to<br />
know them, trust them <strong>and</strong> are willing to address their problems to them.<br />
The next two tasks <strong>of</strong> the united grassroots union are setting up a union service centre <strong>and</strong> ten<br />
service hubs based on the centre, with an aim to provide special service to all current members<br />
according to members’ interests. For example, sports hub could organize training <strong>and</strong> matches for<br />
members who like doing sports. The service centre is located in the sub-district government building,<br />
equipped with reading room <strong>and</strong> meeting spaces.<br />
5. Discussion: Pr<strong>of</strong>essional Leader as a Regulator <strong>of</strong> Market Risks<br />
After reviewing double movement theory <strong>and</strong> pr<strong>of</strong>essional union leader practice, we back to our<br />
question: ‘to what extent has the practice developed an effective countermovement to the expansion <strong>of</strong><br />
market risks in China’s new economy? In other words, what is the effectiveness <strong>of</strong> the new form in<br />
protecting their members from labour market risks?’ In this section, we will identify the evidence <strong>of</strong><br />
worker protection in the practice.<br />
138
In terms <strong>of</strong> union internal structure, firstly, this practice organizes membership in small firms,<br />
which are almost not organized at all before; secondly, the initiative innovatively creates a form <strong>of</strong><br />
‘united grassroots union’ to cover as many petty shop employees as possible in the area.<br />
In terms <strong>of</strong> union-management relations, through the pr<strong>of</strong>essional leader programs, grassroots<br />
leaders gain the personnel <strong>and</strong> financial independency from the management, which enables them to<br />
genuinely represent worker to balance management extend union’s influence into workplace.<br />
In reference to union-member relations, firstly, the organizing campaign led by pr<strong>of</strong>essional union<br />
leaders extends unions’ influence to both small private businesses <strong>and</strong> precarious employees within<br />
those small shops; secondly, after the establishment <strong>of</strong> united grassroots union, rural migrant workers<br />
working in P sub-district can find a place to address their labour relations concerns; thirdly, the union<br />
plays the role <strong>of</strong> workplace labour dispute mediator <strong>and</strong> contact person for district level legal aid<br />
centre; Finally, the grassroots union leaders’ “revisit” has formed a kind <strong>of</strong> surveillance to firm bosses<br />
on their conformity to labour legislations <strong>and</strong> collective contract.<br />
In terms <strong>of</strong> union democracy, mutual-trust has been established to some extent between<br />
pr<strong>of</strong>essional leaders/organizers <strong>and</strong> rank-<strong>and</strong>-file members during the union organizing campaign <strong>and</strong><br />
following ‘revisit’ process. Based on the mutual-trust, more member participation on union<br />
governance <strong>and</strong> union activities could be expected.<br />
Besides, the practice, together with union pr<strong>of</strong>essional leader trials in other cities, influences the<br />
revision <strong>of</strong> the ACFTU regulations, which will get the practice implemented throughout all the<br />
branches <strong>of</strong> the ACFTU. In October 2008, an amendment <strong>of</strong> the Constitution <strong>of</strong> the All-China<br />
Federation <strong>of</strong> Trade Unions, Chapter 6, provision 34, two more clauses were added to further stipulate<br />
the personnel <strong>and</strong> financial independence from the management, which are, “county level unions <strong>and</strong><br />
union branches above the county level should set up a fund for union leadership rights <strong>and</strong> interests<br />
protection” <strong>and</strong> “county level unions <strong>and</strong> union branches above the county level may select <strong>and</strong><br />
employ union <strong>of</strong>ficials for grassroots unions”.<br />
6. Concluding Comments<br />
The union leader/organizer pr<strong>of</strong>essionalization program emerged within Chinese <strong>of</strong>ficial unions<br />
suggests a clear response to the market liberalization process in the last 30 years reform in China. This<br />
response manifests itself by involving organizing more membership, more independency <strong>of</strong> grassroots<br />
unions from management <strong>and</strong> more formal way <strong>of</strong> worker presentation, <strong>and</strong> expected closer <strong>and</strong> more<br />
interactive union-member relationships.<br />
All these changes have not only paved a way for unions to represent their members, but also<br />
formed a relatively formal institution to protect their union members from market risks. The increase<br />
<strong>of</strong> union membership <strong>and</strong> influence helps unions get involved in the new industrial relations<br />
framework; independency from management enables union leaders to balance interests <strong>of</strong><br />
management <strong>and</strong> workers; the promotion <strong>of</strong> union-member relationships makes union rank-<strong>and</strong>-file<br />
members participate more in union activities.<br />
To sum up, despite its limitations, the pr<strong>of</strong>essional union leader practice does have its subtle<br />
implications that Chinese unions do have their potential to act as social actor being able to set up a<br />
countermovement against market risks exerted by globalized capitalism, like other social actors<br />
mentioned in the Polanyi’s theory, e.g. farmers <strong>and</strong> consumer groups.<br />
References:<br />
[1] ACFTU (2003). Chinese Union Charter (Vol. General Principle, Paragraph 1).<br />
[2] Chan, A. (1998). Labour Relations in Foreign-funded Ventures. In G. O'Leary (Ed.), Adjusting to<br />
Capitalism: Chinese Workers <strong>and</strong> the State (pp. xviii, 181 p.). Armonk, N.Y. London: M.E.<br />
Sharpe.<br />
[3] Chan, Anita (2006) "Organizing Wal-Mart: The Chinese Trade Union at a Crossroads," Japan<br />
Focus, September 8, 2006, http://japanfocus.org/products/to<strong>pdf</strong>/2217<br />
[4] Dale, Gareth (2008) “Karl Polanyi’s The Great Transformation: Perverse Effects, Protectionism<br />
<strong>and</strong> Gemeinschaft.” Economy <strong>and</strong> Society, 37(4): 495-524.<br />
139
[5] Deyo, Frederic C <strong>and</strong> Ağartan, Kaan (2007) “Reforming East Asian Labor Systems: China, Korea<br />
<strong>and</strong> Thail<strong>and</strong>.” In Ayşe Buğra <strong>and</strong> Kaan Ağartan, eds., Reading Karl Polanyi for the Twenty-First<br />
Century: Market Economy as a Political Project, Palgrave Macmillan, New York, pp. 191-219.<br />
[6] Edelman, Marc (2005) “Bringing the Moral Economy Back in…to the Study <strong>of</strong> 21st-Century<br />
Transnational Peasant Movements.” American Anthropologist, 107(3): 331-345.<br />
[7] Evans, Peter (2008) “Is an Alternative to Globalization Possible?” Politics <strong>and</strong> Society, 36(2):<br />
271-305.<br />
[8] Ge, Y. (2007). What Do Unions Do in China? : SSRN Working Paper.<br />
[9] Huang, X., <strong>and</strong> Ji, L. (2006, May 11st) Report on Trade Unions in Foreign-invested Companies<br />
(Waiqi Gonghui Shengcun Diaocha), Southern Weekly (Nanfang Zhoumo), May 11st, 2006.<br />
[10] Ma, H., Wu, D. (2008) Guangzhou Trade Union Organizers Officially Appear Today (Guangzhou<br />
Gonghui Zuzhiyuan Jinzao Liangxiang) Yangcheng Evening News (Yangcheng Wanbao), August<br />
6th, 2008.<br />
[11] Munch, Ronaldo (2002) “Globalization <strong>and</strong> Democracy: A New “Great Transformation”?”<br />
Annals <strong>of</strong> the American Academy <strong>of</strong> Political <strong>and</strong> Social Science, 581(May): 10-21.<br />
[12] Nee, Victor (1992) “Organizational Dynamics <strong>of</strong> Market Transition: Hybrid Forms, Property<br />
Rights <strong>and</strong> Mixed Economy in China.” Administrative Science Quarterly, 37: 1-27.<br />
[13] Nee, Victor (1996) “The Emergence <strong>of</strong> a Market Society: Changing Mechanisms <strong>of</strong> Stratification<br />
in China.” American Journal <strong>of</strong> Sociology, 101(4): 908-949.<br />
[14] Ngok, K.L. (2007) Chinese Labor Policy: A Market <strong>and</strong> Globalization Perspective, Beijing:<br />
Social Sciences Academic Press.<br />
[15] Polanyi, K (1944[1957]) The Great Transformation: The Political <strong>and</strong> Economic Origins <strong>of</strong> Our<br />
Time, Beacon Press: Boston.<br />
[16] Sang, W. (2007) Guangzhou: 100 Union Organizer Entering the Firms (Guangzhou: Baiming<br />
Gonghui Zuzhiyuan Jin Qiye) Retrieved 18th Oct, 2008, from<br />
http://www.acftu.net/template/10004/file.jsp?cid=731&aid=78071<br />
[17] Shen, J. (2007). Labour disputes <strong>and</strong> their resolution in China. Oxford: Ch<strong>and</strong>os.<br />
[18] St<strong>and</strong>ing, Guy (2007) “Labor Recommodification in the Global Transformation.” In Ayşe Buğra<br />
<strong>and</strong> Kaan Ağartan, eds., Reading Karl Polanyi for the Twenty-First century: Market Economy as<br />
a Political Project, Palgrave Macmillan, New York, pp. 67-94.<br />
[19] Taylor, B., <strong>and</strong> Li, Q. (2005). Review on ACFTU’s Union Organizing Practice. Paper presented<br />
at the Globalization <strong>and</strong> Labour Turnover in India <strong>and</strong> China, Monash University, Australia.<br />
[20] Taylor, B. (2008). Does China Have a Labour Movement? Prospects for Industrial Relations<br />
Reform in China. Montreal: CRIMT.<br />
[21] Taylor, B., <strong>and</strong> Li, Q. (2007). Is the ACFTU a Union <strong>and</strong> Does it matter? Journal <strong>of</strong> Industrial<br />
Relations, 49(5), 701-715.<br />
[22] Unger, J., <strong>and</strong> Chan, A. (1995). China, Corporatism <strong>and</strong> the East Asian Model. Australian Journal<br />
<strong>of</strong> Chinese Affairs, 33, 29-53.<br />
[23] Wang, W. (2007) Union Organizer System Initiated in Our Province (Wosheng Zai Quanguo<br />
Shouchuang Gonghui Zuzhiyuan Zhidu), Hebei Daily (Hebei Ribao), June 20th, 2007.<br />
140
Developing Dynamic Capability through Partnership: The Role <strong>of</strong><br />
Capabilities<br />
Wei Jiang 1 * , Felix Mavondo 1<br />
* Presenter<br />
1. Department <strong>of</strong> Marketing, Monash University, Clayton Compus, Wellington Rd, Clayton<br />
3800, Australia<br />
Partnerships have been adopted by most organizations as a major source <strong>of</strong> competitive<br />
advantage <strong>and</strong> to manage technological turbulence <strong>and</strong> dynamic market environments. The<br />
paper aims to investigate some key organizational capabilities in partnerships that were<br />
proposed to contribute to the improvement <strong>of</strong> the organization’s agility—a dynamic capability<br />
for managing capricious environments. The role <strong>of</strong> trust between partnering organizations as a<br />
moderating variable is examined. A samples (n=300) from Chinese managers in<br />
manufacturing industries was used for this study. The results indicate that trust is an important<br />
moderator <strong>of</strong> the relationship between market orientation <strong>and</strong> learning orientation <strong>and</strong><br />
organisational agility.<br />
1. Introduction<br />
The business environment has been described in the literature as hyper turbulent, unpredictable,<br />
hypercompetitive [39] [42] [45]. Environmental changes force acquired skills <strong>and</strong> capabilities to<br />
become obsolete or create new opportunities, both <strong>of</strong> which may require firms to build new<br />
capabilities [3]. Failure or slowness <strong>of</strong> building new capabilities endangers firm’s prosperity or even<br />
survival. Such environments, thus, have been recognized as the cause <strong>of</strong> most organizations’ failures<br />
[38]. Dynamic capability theory contends that dynamic capabilities underlie the source <strong>of</strong> sustainable<br />
competitive advantage in such market environments; they are the important mechanisms for<br />
reconfiguring, recombining <strong>and</strong> deleting ordinary organizational resources to achieve a fit with the<br />
environment <strong>and</strong> strategic imperatives [43] [50] [28]. Despite its importance to firm development <strong>and</strong><br />
the scholarly attention devoted to it, dynamic capability remains underspecified [3], <strong>and</strong> empirical<br />
work is still in its infancy [36] [55]. It has been generally acknowledged in the literature that firms can<br />
develop dynamic capabilities through long-term partnerships [53], but ‘the how question’ has not been<br />
systematically conceptualized or empirically grounded. To address this gap, this research seeks to<br />
examine the implications <strong>of</strong> organizational capabilities for the development <strong>of</strong> dynamic capabilities<br />
<strong>and</strong> to identify any boundary limiting conditions. Since trust is an important relational variable this is<br />
conceptualized as a possible boundary condition—a moderator <strong>of</strong> the various relationships.<br />
The paper is organized as follows: first, the conceptual framework is presented. Second, relevant<br />
literatures <strong>and</strong> propositions are presented, followed by a discussion <strong>of</strong> methodological issues. Finally,<br />
the results <strong>and</strong> discussion are presented with limitations <strong>of</strong> the study <strong>and</strong> its potential implications for<br />
managers <strong>and</strong> academics.<br />
2. Literature review<br />
2.1. Dynamic capability<br />
Dynamic capability theory provides a compelling explanation for the ability <strong>of</strong> some companies to<br />
continuously create, define, discover <strong>and</strong> exploit entrepreneurial opportunities [55]. It proposes that<br />
possessing strategic resources <strong>and</strong> capabilities is not a sufficient condition to generate rents, but that<br />
the dynamic capabilities which grow <strong>and</strong> evolve in response to the external environment are the real<br />
source <strong>of</strong> rent generation [44]. Capabilities are the skills <strong>and</strong> routines that combine both tangible <strong>and</strong><br />
intangible resources together coherently in a synergistic manner to enable firms to produce efficiently<br />
<strong>and</strong>/or effectively valued market <strong>of</strong>ferings [5]. While dynamic capabilities can be any capabilities<br />
which enable firm to modify its resources <strong>and</strong> routines to changing environment, so it can continue<br />
producing market <strong>of</strong>ferings efficiently <strong>and</strong>/or effectively [28] [55].<br />
141
Organizational Capabilities<br />
• Learning Orientation<br />
• Market Orientation<br />
• Manufacturing Capability<br />
• Managerial Capability<br />
Trust<br />
Dynamic Capability<br />
• Agility Capability<br />
Agility, which emerged since 1990s in the literature is considered a major capability that<br />
emphasizing the ability to adjust <strong>and</strong> change [39]. Agility is the ability to survive <strong>and</strong> prosper in a<br />
competitive environment with continuous <strong>and</strong> unpredictable changes by reacting quickly <strong>and</strong><br />
effectively [12]. Some important attributes <strong>of</strong> agility emphasized in all definitions include: speed,<br />
effective responsiveness; proactiveness; <strong>and</strong> availability <strong>of</strong> slack resources [57]. Speed is the firm’s<br />
ability to accomplish tasks in the shortest possible time, such tasks includes quick new product<br />
development, fast operation, quick learning <strong>of</strong> new technology, <strong>and</strong> fast adaptation to change, etc. [39].<br />
Responsiveness is the ability to identify changes <strong>and</strong> opportunities, respond reactively or proactively<br />
to them, <strong>and</strong> recover from them [38]. Proactiveness is the capability to act proactively, taking initiative<br />
in improving current circumstances or creating new favorable ones [2]. Organizational slack is defined<br />
by Bourgeois (1981, p. 29) as a cushion <strong>of</strong> actual or potential resources, which firms can use to initiate<br />
changes in strategies for environment adapting; they are accumulated for pursuing market <strong>and</strong><br />
competitive opportunities in the future [43].<br />
2.2. Organizational capabilities<br />
Four key organizational capabilities are identified from the literature as significant for firm’s<br />
competitive advantages. They are market orientation, learning orientation, managerial capability <strong>and</strong><br />
manufacturing capability. Market orientation has been defined as an important organizational culture<br />
that creates the necessary behaviors for underst<strong>and</strong>ing <strong>and</strong> fulfilling customers’ expressed needs [32],<br />
<strong>and</strong> also latent needs [35]. The concept <strong>of</strong> market orientation is focusing continuously on discovering<br />
new opportunities for target-customers, especially under dynamic market or competitive environments<br />
[34]. Managerial capability is innate <strong>and</strong> learned abilities, <strong>and</strong> expertise <strong>of</strong> managers in a firm [1] [14]<br />
[24]. It directly affects the firm’s performance by integrating, reintegrating firms’ resources <strong>and</strong><br />
capabilities [29], <strong>and</strong> the partnership performance by the managers’ coordination <strong>and</strong> control <strong>of</strong><br />
operation [8]. Learning orientation “is comprised by the continually evolving knowledge stocks<br />
existing in individuals, groups <strong>and</strong> the organization, which flow to continuously exploit <strong>and</strong> explore<br />
knowledge in accordance with the environmental conditions” [37] (p.501). It allows the transfer <strong>of</strong><br />
tacit knowledge, resources <strong>and</strong> capabilities in the partnership, thus providing opportunities to develop<br />
new capabilities [7]. Manufacturing capability directly influences a firm’s success because such a<br />
capability has the potential to support <strong>and</strong> shape corporate strategy [11]. Various studies have<br />
indicated the direct impact <strong>of</strong> manufacturing capability on the manufacturing <strong>and</strong> organizational<br />
performance [25]. A core manufacturing capability is a multifaceted complex concept <strong>and</strong> is multidimensional<br />
<strong>and</strong> implies cost, quality, delivery, <strong>and</strong> flexibility [11] [25]. Based on the above, we<br />
develop the following hypotheses:<br />
H1: Market Orientation is positively associated with Agility.<br />
H2: Learning Orientation is positively associated with Agility.<br />
H3: Managerial Capability is positively associated with Agility.<br />
142
H4: Manufacturing Capability is positively associated with Agility.<br />
2.3. Trust<br />
In economic exchange, trust implies a general expectation <strong>of</strong> good faith efforts by parties to honor<br />
commitments, to be honest in negotiations, <strong>and</strong> to decry opportunistic behavior. Trust has been argued<br />
as playing a critical role in creating synergistic outcomes in partnership [2] [32]. First, trust is regarded<br />
as one <strong>of</strong> the most widely acknowledged means for governing <strong>and</strong> coordinating inter-organizational<br />
exchange [18] [31]. Second, partnerships <strong>of</strong> two or more companies produce a strong potential for<br />
dysfunctional conflicts, where trust is a key factor to make cooperation smooth by alleviating the<br />
functional conflicts, facilitating mutual underst<strong>and</strong>ing <strong>and</strong> bilateral communication [2] [48]. Third,<br />
trust facilitates knowledge exchange, both in quantity <strong>and</strong> quality, <strong>and</strong> learning in alliances [48]. Thus,<br />
the hypotheses are stated as:<br />
H5: Trust is positively associated with Agility.<br />
H6: Trust moderates the relationship between Market Orientation <strong>and</strong> Agility.<br />
H7: Trust moderates the relationship between Learning Orientation <strong>and</strong> Agility.<br />
H8: Trust moderates the relationship between Managerial Orientation <strong>and</strong> Agility.<br />
H9: Trust moderates the relationship between Manufacturing Orientation <strong>and</strong> Agility.<br />
3. Research methodology<br />
Inspired from the literature in typologies <strong>of</strong> strategic alliances [4], the long-term partnership defined<br />
in this study includes: (1) long-term sourcing agreements (major customer or major supplier), (2) joint<br />
R&D, (3) joint marketing <strong>and</strong> promotion, <strong>and</strong> (4) joint manufacturing. In this research, the unit <strong>of</strong><br />
analysis is individual manufacturing firm in long-term partnership where at least one partner is a<br />
Chinese firm. In the partnership which involves more than two companies, the main partner will be<br />
only one considered. Senior managers in various manufacturing industries were selected as<br />
respondents. Mail-out questionnaires were employed in seven manufacturing provinces in China,<br />
including major cities representing each part <strong>of</strong> China. A usable sample <strong>of</strong> 300 was received<br />
representing a 35% response rate.<br />
4. Results <strong>and</strong> discussions<br />
Table 1 shows the correlations <strong>and</strong> reliabilities <strong>of</strong> the constructs in the model. In Table 2 the results<br />
<strong>of</strong> hypothesis testing are presented. In Model 2 Market orientation is significantly related to marketing<br />
agility (p
Table 1. Correlations, Reliabilities <strong>of</strong> the variables in the Model<br />
n=300 1 2 3 4 5 6<br />
1 Market orientation 0.932<br />
2 Learning orientation 0.488 0.941<br />
3 Managerial capability 0.540 0.668 0.885<br />
4 Manufacturing capability 0.477 0.540 0.618 0.891<br />
5 Trust 0.532 0.402 0.418 0.351 0.881<br />
6 Agility 0.776 0.524 0.547 0.481 0.486 0.936<br />
Mean 5.335 5.274 5.609 5.449 5.845 5.331<br />
St<strong>and</strong>ard deviation 0.895 0.986 0.816 0.900 0.870 0.857<br />
Note: The diagonal (in italics <strong>and</strong> bold) shows the Cronbach’s α for each construct<br />
Table 2: Regression models for the antecedents <strong>of</strong> Marketing Agility<br />
Variables<br />
Model 1 Model 2<br />
β t-value β t-value<br />
H1: Market Orientation 0.622*** 13.379 -0.644 -1.103<br />
H2: Learning Capability 0.118** 2.402 1.072*** 2.557<br />
H3: Managerial Capability 0.074 1.396 -0.924 -1.471<br />
H4: Manufacturing Capability 0.055 1.186 0.507 1.322<br />
H5: Trust 0.058 1.361 -0.963* -1.763<br />
H6: Trust x Market Orientation 2.014* 2.134<br />
H7: Trust x Learning<br />
-1.405* -2.284<br />
orientation<br />
H8: Trust x Managerial<br />
1.768* 1.682<br />
Capability<br />
H9: Trust x Manufacturing<br />
-0.713 -1.223<br />
Capability<br />
R 2 0.640 0.656<br />
Adj R 2 0.634 0.646<br />
F-Ratio 104.454 61.565<br />
Table 3: Investigation <strong>of</strong> Slopes <strong>of</strong> Market <strong>and</strong> Learning Orientation at different levels <strong>of</strong> Trust<br />
Low (t-value) Medium (t-value) High (t-value)<br />
Market Orientation 0.6530 (11.1114) 0.6870 (16.5933) 0.7211 (14.0508)<br />
Learning Orientation 0.3103 (5.4074) 0.3413 (7.7685) 0.3724 (6.4487)<br />
5. Conclusions<br />
In this paper, we identified the four key organizational capabilities, decomposed their effects on the<br />
development <strong>of</strong> agility respectively, <strong>and</strong> examined the moderating role <strong>of</strong> trust in such relationships.<br />
The proposed model showed that all the four capabilities have positive <strong>and</strong> significant impacts on<br />
agility. Trust moderates the impacts <strong>of</strong> capabilities on agility, but only for market orientation, learning<br />
orientation. That means in a partnership with high level <strong>of</strong> trust, the contribution <strong>of</strong> market orientation<br />
<strong>and</strong> learning orientation to the development <strong>of</strong> agility capability is higher than in a partnership with<br />
lower trust. Market orientation <strong>and</strong> learning orientation are the most critical input to marketing agilitydynamic<br />
capability in a partnership. Very surprisingly, data from Chinese manufacturing industries<br />
showed that manufacturing capability does not influence the development <strong>of</strong> agility in the partnership<br />
no matter whether the level <strong>of</strong> trust that exists.<br />
This study has several academic contributions in terms <strong>of</strong> empirically examining dynamic<br />
capability, its operationalisation, <strong>and</strong> the relationship between ordinary capabilities <strong>and</strong> dynamic<br />
capabilities. It extends studies in dynamic capability theory into inter-organizational partnerships.<br />
Moreover, the study contributes to the literature related to business issues in China.<br />
144
6. Managerial implication <strong>and</strong> limitations<br />
This research also <strong>of</strong>fers a number <strong>of</strong> implications that could add to our underst<strong>and</strong>ing <strong>of</strong><br />
competence-centered practices <strong>and</strong> their contribution to firm agility. China has been considered as one<br />
<strong>of</strong> the biggest emerging markets in the world for its rapid <strong>and</strong> sustainable development, after the<br />
transforming from centralized economy to a more market-orientated economy [51]. Firm agility has<br />
become extremely important to keep firms energetic enough to survive in the dynamic global<br />
environment.<br />
This study has tested one moderator—trust, which is only one aspect <strong>of</strong> the partnership connection,<br />
in the cooperation process. Other factors, such as type <strong>of</strong> partnership, length <strong>of</strong> the partnership,<br />
conflict management, commitment, et al., may indicates various degree <strong>of</strong> influences on the<br />
capabilities’ contribution to dynamic capability development.<br />
References:<br />
[1]. Castanias, R. P., & Helfat, C. E. (2001). The managerial rents model: theory <strong>and</strong> empirical<br />
analysis. Journal <strong>of</strong> Management, 27(6), 661-678.<br />
[2]. Crant, J. M. (2000). Proactive behaviour in organizations. Journal <strong>of</strong> Management, 26(3), 435-462.<br />
[3]. Cullen, J. B., Johnson, J. L., & Sakano, T. (2000). Success through commitment <strong>and</strong> trust: the s<strong>of</strong>t<br />
side <strong>of</strong> strategic alliance management. Journal <strong>of</strong> World Business, 35(3), 223-241.<br />
[4]. Danneels, E. (2008). Organizational antecedents <strong>of</strong> second-order competences. Strategic<br />
Management Journal, 29(5), 519-543.<br />
[5]. Das, T. K., & Teng, B.-S. (2000). A resource-based theory <strong>of</strong> strategic alliances. Journal <strong>of</strong><br />
Management, 26(1), 31-61.<br />
[6]. Day, G. S. (1994). The capabilities <strong>of</strong> market-driven organizations. Journal <strong>of</strong> Marketing, 58(1),<br />
37-52.<br />
[7]. Eisenhardt, K. M., & Martin, J. A. (2000). Dynamic capabilities: what are they? Strategic<br />
Management Journal, 21(10/11), 1105-1121.<br />
[8]. Eisenhardt, K. M., & Schoonhoven, C. B. (1996). Resource-based view <strong>of</strong> strategic alliance<br />
formation: strategic <strong>and</strong> social effects in entrepreneurial firms. Organization Science, 7(2), 136-<br />
150.<br />
[9]. Fahy, J., Hooley, G., Cox, T., Beracs, J., Fonfara, K., & Snoj, B. (2000). The development <strong>and</strong><br />
impact <strong>of</strong> marketing capabilities in central Europe. Journal <strong>of</strong> International Business Studies,<br />
31(1), 63-81.<br />
[10]. Griffith, D. A., & Harvey, M. G. (2001). A resource perspective <strong>of</strong> global dynamic<br />
capabilities. Journal <strong>of</strong> International Business Studies, 32(3), 597-607.<br />
[11]. Grobler, A., & Grubner, A. (2006). An empirical model <strong>of</strong> the relationships between<br />
manufacturing capabilities. International Journal <strong>of</strong> Operations & Production Management, 26(5),<br />
458-485.<br />
[12]. Gunasekaran, A. (1999). Agile manufacturing: A framework for research <strong>and</strong> development.<br />
International Journal <strong>of</strong> Production Economics, 62(1/2), 87-105.<br />
[13]. Helfat, C. E., & Peteraf, M., A. (2003). The dynamic resource-based view: capability<br />
lifecycles. Strategic Management Journal, 24(10), 997-1010.<br />
[14]. Harris, D., & Helfat, C. E. (1997). Specificity <strong>of</strong> CEO human capital <strong>and</strong> compensation.<br />
Strategic Management Journal, 18(11), 895-920.<br />
[15]. Hitt, M. A., Ahlstrom, D., Dacin, M. T., Levitas, E., & Svobodina, L. (2004). The institutional<br />
effects on strategic alliance partner selection in transition economies: China vs. Russia.<br />
Organization Science, 15(2), 173-185.<br />
[16]. Hitt, M. A., Bierman, L., Shimizu, K., & Kochhar, R. (2001). Direct <strong>and</strong> moderating effects <strong>of</strong><br />
human capital on strategy <strong>and</strong> performance in pr<strong>of</strong>essional service firms: a resource-based<br />
perspective. Academy <strong>of</strong> Management Journal, 44(1), 13-28.<br />
[17]. Hitt, M. A., Dacin, M. T., Levitas, E., Arregle, J.-L., & Borza, A. (2000). Partner selection in<br />
emerging <strong>and</strong> developed market contexts: resource-based <strong>and</strong> organizational learning perspectives.<br />
Academy <strong>of</strong> Management Journal, 43(3), 449-467.<br />
145
[18]. Jap, S. D. (1999). Pie-expansion efforts: collaboration processes in buyer-supplier<br />
relationships. Journal <strong>of</strong> Marketing Research, 36, 461-475.<br />
[19]. Kale, P., Singh, H., & Perlmutter, H. (2000). Learning <strong>and</strong> protection <strong>of</strong> proprietary assets in<br />
strategic alliances: building relational capital. Strategic Management Journal, 21, 217-237.<br />
[20]. Kale, P., Dyer, J., & Singh, H. (2001). Value creation <strong>and</strong> success in strategic alliances:<br />
alliancing skills <strong>and</strong> the role <strong>of</strong> alliance structure <strong>and</strong> systems. European Management Journal,<br />
19(5), 463-471.<br />
[21]. Kogut, B., & Z<strong>and</strong>er, U. (1992). Knowledge <strong>of</strong> the firm, combinative capabilities, <strong>and</strong> the<br />
relationship <strong>of</strong> technology. Organization Science, 3(3), 383-397.<br />
[22]. Lambe, C. J., Spekman, R. E., & Hunt, S. D. (Spring2002). Alliance competence, resources,<br />
<strong>and</strong> alliance success: conceptualization, measurement, <strong>and</strong> initial test. Academy <strong>of</strong> Marketing<br />
Science, 30(2), 141-158.<br />
[23]. Lee, D. Y., & Dawes, P. L. (2005). Guanxi, trust, <strong>and</strong> long-term orientation in Chinese<br />
business markets. Journal <strong>of</strong> International Marketing, 13(2), 28-56.<br />
[24]. Li, H., & Zhang, Y. (2007). The role <strong>of</strong> managers' political networking <strong>and</strong> functional<br />
experience in new venture performance: evidence from china's transition economy. Strategic<br />
Management Journal, 28(8), 791-804.<br />
[25]. Li, L. L. X. (2000). Manufacturing capability development in a changing business<br />
environment. Industrial Management & Data Systems, 100(6), 261-270.<br />
[26]. Lin, H. (2006). Inter-organizational collaboration, social embeddedness, <strong>and</strong> value creation: a<br />
theoretical analysis. International Journal <strong>of</strong> Management, 23(3), 548-558.<br />
[27]. Luo, Y. (2000). Partnering with Chinese firms: lessons for international managers: Ashgate.<br />
[28]. Madhavaram, S., & Hunt, S. D. (2008). The service-dominant logic <strong>and</strong> a hierarchy <strong>of</strong> operant<br />
resources: developing masterful operant resources <strong>and</strong> implications for marketing strategy.<br />
Journal <strong>of</strong> the Academic Marketing Science, 36(1), 67-82.<br />
[29]. Mahoney, J. T. (1995). The management <strong>of</strong> resources <strong>and</strong> the resource <strong>and</strong> management.<br />
Journal <strong>of</strong> Business Research, 33(1), 91-101.<br />
[30]. Makadok, R. (2001). Toward a synthesis <strong>of</strong> the resource-based <strong>and</strong> dynamic-capability views<br />
<strong>of</strong> rent creation. Strategic Management Journal, 22(5), 387-401.<br />
[31]. Morgan, R. M., & Hunt, S. D. (1994). The commitment-trust theory <strong>of</strong> relationship marketing.<br />
Journal <strong>of</strong> Marketing, 58(3), 20-38.<br />
[32]. Muthusamy, S. K., White, M. A., & Carr, A. (2007). An empirical examination <strong>of</strong> the role <strong>of</strong><br />
social exchanges in alliance performance. Journal <strong>of</strong> Managerial Issues, 11(1), 53-75.<br />
[33]. Narver, J. C., & Slater, S. F. (1990). The effect <strong>of</strong> a market orientation on business<br />
pr<strong>of</strong>itability. Journal <strong>of</strong> Marketing, 54, 20-35.<br />
[34]. Narver, John C.; Slater, S. F., & MacLachlan, D. L., (2004) Responsive <strong>and</strong> Proactive Market<br />
Orientation <strong>and</strong> New-Product Success. Journal <strong>of</strong> Product Innovation Management, 21 (5) 334-<br />
347.<br />
[35]. Nasution, H., N.,, & Mavondo, F., T. (2008). Organisational capabilities: antecedents <strong>and</strong><br />
implications for customer value. European Journal <strong>of</strong> Marketing, 42(3/4), 477-501.<br />
[36]. Newbert, S. L. (2007). Empirical research on the resource-based view <strong>of</strong> the firm: an<br />
assessment <strong>and</strong> suggestions for future research. Strategic Management Journal, 28(1), 121-146.<br />
[37]. Prieto, I. M., & Revilla, E. (2006). Assessing the impact <strong>of</strong> learning capability on business<br />
performance: empirical evidence from Spain. Management Learning, 37(4), 499-522.<br />
[38]. Sharifi, H., & Zhang, Z. (1999). A methodology for achieving agility in manufacturing<br />
organizations: An introduction. International Journal <strong>of</strong> Production Economics, 62(1/2), 7-22.<br />
[39]. Sherehiy, B., Karwowski, W., & Layer, J. K. (2007). A review <strong>of</strong> enterprise agility: concepts,<br />
frameworks, <strong>and</strong> attributes. International Journal <strong>of</strong> Industrial Ergonomics, 37(2), 445-460.<br />
[40]. Slater, S. F., & Narver, J. C. (1994). Does the competitive environment moderate the marketorientation<br />
performance relationship. Journal <strong>of</strong> marketing, 58, 46-55.<br />
[41]. Sirman, D., G.;, & Hitt, M. A. (2003). Managing resources: linking unique resources,<br />
management, <strong>and</strong> wealth creation in family firms. Entrepreneurship: Theory & Practice, 27(4),<br />
339-358.<br />
[42]. Tan, J. (2003). Curvilinear relationship between organizational slack <strong>and</strong> firm performance:<br />
evidence from Chinese state enterprises. European Management Journal, 21(6), 740-749.<br />
146
[43]. Tan, J. & Peng, M.W., (2003) Organizational Slack <strong>and</strong> Firm Performance during Economic<br />
Transition: Two Studies from an Emerging Economy. Strategic Management Journal, 24 (4)<br />
1249-1263.<br />
[44]. Teece, D. J., Pisano, G., & Shuen, A. (1997). Dynamic capabilities <strong>and</strong> strategic management.<br />
Strategic Management Journal, 18(7), 509-533.<br />
[45]. Thoumrungroje, A., & Tansuhaj, P. (2004). Globalization effects, co-marketing alliances, <strong>and</strong><br />
performance. Journal <strong>of</strong> American Academy <strong>of</strong> Business, 5(1/2), 495-502.<br />
[46]. Van Den Bosch, F. A. J., Volberda, H. W., & Boer, M. d. (1999). Coevolution <strong>of</strong> firm<br />
absorptive capacity <strong>and</strong> knowledge environment: organizational forms <strong>and</strong> combinative<br />
capabilities. Organization Science, 10(5), 551-568.<br />
[47]. Vorhies, D. W., & Morgan, N. A. (2005). Benchmarking marketing capabilities for<br />
sustainable competitive advantage. Journal <strong>of</strong> Marketing, 68, 80-94.<br />
[48]. Voss, K. E., Johnson, J. L., Cullen, J. B., Sakano, T., & Takenouchi, H. (2006). Relational<br />
exchange in US-Japanese marketing strategic alliances. International Marketing Review, 23(6),<br />
610-635.<br />
[49]. Wang, C. L. (2007). guanxi vs. relationship marketing: exploring underlying differences.<br />
Industrial Marketing Management, 36(1), 81-86.<br />
[50]. Wang, C. L., & Ahmed, P. K. (2007). Dynamic capabilities: a review <strong>and</strong> research agenda.<br />
International Journal <strong>of</strong> Management Reviews, 9(1), 31-51.<br />
[51]. Wang, Y., Lo, H., & Yang, Y. (2004). The constituents <strong>of</strong> core competencies <strong>and</strong> firm<br />
performance: evidence from high-technology firms in China. Journal <strong>of</strong> Engineering <strong>and</strong><br />
Technology Management, 21, 249 - 280.<br />
[52]. Xin, K. R., & Pearce, J. L. (1996). Guanxi: connections as substitutes for formal institutional<br />
support. Academy <strong>of</strong> Management Journal, 39(6), 1641-1658.<br />
[53]. Yusuf, Y. Y., Sarhadi, M., & Gunasekaran, A. (1999). Agile manufacturing: The drivers,<br />
concepts <strong>and</strong> attributes. International Journal <strong>of</strong> Production Economics, 62(1/2), 33-43.<br />
[54]. Zahra, S. A., & George, G. (2002). Absorptive capacity: a review, reconceptualization, <strong>and</strong><br />
extension. Academy <strong>of</strong> Management Review, 27(2), 185-203.<br />
[55]. Zahra, S. A., Sapienza, H. J., & Davidsson, P. (2006). Entrepreneurship <strong>and</strong> dynamic<br />
capabilities: a review, model <strong>and</strong> research agenda. Journal <strong>of</strong> Management Studies, 43(4), 917-<br />
955.<br />
[56]. Zhang, Y., & Zhang, Z. (2006). Guanxi <strong>and</strong> organizational dynamics in China: a link between<br />
individual <strong>and</strong> organizational levels. Journal <strong>of</strong> Business Ethics, 67(5), 375-392.<br />
[57]. Zhang, Z., & Sharifi, H. (2007). Towards theory building in agile manufacturing strategy-a<br />
taxonomical approach. IEEE Transactions on Engineering Management, 54(2), 351-370.<br />
[58]. Zollo, M., Reuer, J. J., & Singh, H. (2002). Inter-organizational routines <strong>and</strong> performance in<br />
strategic alliances. Organization Science, 13, 701-713.<br />
[59]. Zollo, M., & Winter, S., G. (2002). Deliberate learning <strong>and</strong> the evolution <strong>of</strong> dynamic<br />
capabilities. Organization Science, 13(3), 339-351.<br />
147
Early Childhood Education Matters: Functions, Situations, <strong>and</strong> Perceptions<br />
Beibei Pan 1 *<br />
* Presenter<br />
1. Department <strong>of</strong> Management, Monash University, Clayton Compus, Wellington Rd,<br />
Clayton 3800, Australia<br />
This paper is aiming at providing an overall picture <strong>of</strong> former <strong>and</strong> current views on early<br />
childhood education. The general developmental tendency <strong>of</strong> this field is that early education<br />
for young children has increasingly received attention <strong>of</strong> parents, educationists, <strong>and</strong> scholars<br />
from diverse research backgrounds. After clarifying the educational <strong>and</strong> socioeconomic<br />
functions, the focus <strong>of</strong> this paper shifts to digging out the underpinning determinants which<br />
make quality services <strong>of</strong> preschool programs.<br />
Keywords<br />
early childhood education; development; essential factors<br />
Introduction<br />
During the first decades <strong>of</strong> the nineteenth century, industrialization brought changes to family<br />
structure as well as to the initial concept <strong>of</strong> early childhood care <strong>and</strong> education. Since then, early<br />
childhood education (ECE) has become an international interest (Tietze, Cryer, Bairrão, Palacios, &<br />
Wetzel, 1996). ‘In the international arena, the collocation <strong>of</strong> a variety <strong>of</strong> integrated, flexible services<br />
for children has gained momentum’ (Noble, 2007, p. 51, as cited in OECD, 2001). Literally, ECE has<br />
developed as an association which concerns individuals, families, <strong>and</strong> societies, <strong>and</strong> also a conjunction<br />
which involves education, economy, <strong>and</strong> policy. For its impacts on a rich variety <strong>of</strong> fields, ECE<br />
attracts attentions <strong>of</strong> scholars from differing research backgrounds.<br />
Significant Influence <strong>of</strong> Early Childhood Education<br />
The importance <strong>of</strong> ECE in both individual development <strong>and</strong> social progress has been clarified by<br />
recent studies. Traditionally, child care centres were established to propel women’s participation in<br />
labour market; later on, the educational functions for the sake <strong>of</strong> human development has been<br />
attached to ECE as well (Nyl<strong>and</strong>, Nyl<strong>and</strong>, & Maharaj, 2009).<br />
Contemporary scholars have well acknowledged the favourable outcomes <strong>of</strong> early years’<br />
education. Researchers hold the view that investment in early education should be encouraged for<br />
reasons summarized as follows (Heckman, 2006):<br />
• The formation <strong>of</strong> both cognitive <strong>and</strong> socio-emotional abilities begins at early age, which<br />
greatly influences a person’s whole life.<br />
• This is an affair <strong>of</strong> social justice. Children from less advantaged families need more financial<br />
support to get access to ECE resources.<br />
• Early intervention can effectively guarantee better school outcomes. Correspondingly,<br />
promoted quality <strong>of</strong> workforce will lead to increased productivity in economy <strong>and</strong> society at<br />
large.<br />
• Quality child care can reduce crime rate, teenage pregnancy <strong>and</strong> welfare dependency.<br />
From an educational perspective, early years are a sensitive period which is essentially influential<br />
to overall competency development during a life time according to behavioural research (Knudsen,<br />
Heckman, Cameron, & Shonk<strong>of</strong>f, 2006, as cited in Shonk<strong>of</strong>f & Phillips, 2000; Heim & Nemer<strong>of</strong>f,<br />
2001). An array <strong>of</strong> studies have demonstrated that child care quality is positively related to the<br />
cognitive development <strong>and</strong> social competence, with other factors generating variations in early<br />
education effects, such as socioeconomic status or family structure (Peisner-Feinberg & Burchinal,<br />
1997). Peisner-Feinberg <strong>and</strong> Burchinal’s (1997) research also provides evidence to the positive<br />
relation for children from all backgrounds. Furthermore, they observed that the ECE influence is<br />
stronger on children from less advanced families. And it has been reiterated by Knudsen et al. (2006, p.<br />
10155)as ‘that the quality <strong>of</strong> the early childhood environment is a strong predictor <strong>of</strong> adult<br />
148
productivity (As cited in Carneiro & Heckman, 2003) <strong>and</strong> that early enrichment for disadvantaged<br />
children increases the probability <strong>of</strong> later economic success (As cited in Currie & Blau, 2005)’. In the<br />
light <strong>of</strong> those positive research results, now it is widely agreeable that ECE matters in a large range <strong>of</strong><br />
aspects <strong>of</strong> human competency development.<br />
From a socioeconomic perspective, Cunha et al. (2005) referred to relevant literatures <strong>and</strong><br />
integrated main findings by presenting the rates <strong>of</strong> return to educational funding according to critical<br />
stages <strong>of</strong> life circle in figure 1. It is noticeable that a remarkable return rate appears in the stage <strong>of</strong><br />
preschool programs. Along with the growth <strong>of</strong> age, human capital investment sees a dramatic decline<br />
in the rate <strong>of</strong> return. This finding turns out to be a persuasive evidence to convince the essential<br />
significance <strong>of</strong> ECE.<br />
Figure 1. Rates <strong>of</strong> return to human capital investment initially setting investment to be equal across all<br />
ages<br />
Source: Cunha et al. (2005) Interpreting the evidence on life cycle skill formation.<br />
As observed in those studies above, it is the high economic return <strong>of</strong> investing in preschool<br />
programs that make early education a crucial issue to the development <strong>of</strong> human capacities <strong>and</strong> the<br />
progress <strong>of</strong> contemporary societies. Responsively, educational requirements extend to early years<br />
spontaneously. The mode <strong>of</strong> contemporary ECE is seeking to better meet the criteria for quality human<br />
resources.<br />
Taking the points above into account, we can safely draw a conclusion that ECE features a series<br />
<strong>of</strong> critical characteristics in relation to a complexity <strong>of</strong> education <strong>and</strong> socio-economy. This explains<br />
the particular attentions paid by parents, educationists, economists, <strong>and</strong> policy-makers on ECE.<br />
Overview <strong>of</strong> Early Childhood Education Development<br />
With all the endeavour <strong>of</strong> educators <strong>and</strong> researchers, ECE has been theoretically <strong>and</strong> practically<br />
well developed during recent decades. ECE systems have undergone regulations in every aspect that<br />
could possibly be improved across the world. The UK launched the Sure Start <strong>and</strong> other programmes;<br />
the USA initiated the Head Start <strong>and</strong> family support approaches; great subsidies are commonly <strong>of</strong>fered<br />
by governments to support child care <strong>and</strong> education among OECD countries (Kamerman, 2000).<br />
Whereas, with a diversity <strong>of</strong> historical <strong>and</strong> socio-economic circumstances, child care in different areas<br />
share similarities as well as disparities.<br />
Due to the active interactions between countries in Europe for centuries, early education<br />
approaches share ‘a remarkable degree <strong>of</strong> similarity among ECE systems in western industrialized<br />
countries’ (Tietze, et al., 1996, p. 449). Take Germany for example. According to the point <strong>of</strong> Cryer et<br />
al. (2002), Germany takes measures to make ECE services affordable to most families based on a<br />
subsidized sliding scale. Thus cost is not a strong consideration when German parents make their<br />
149
decisions on selection <strong>of</strong> child care centre. And the educational functions <strong>and</strong> availability <strong>of</strong> ECE<br />
institutions are taken into serious consideration (Cryer, et al., 2002). Cyrer et al. (2002) also claimed<br />
that the range <strong>of</strong> ECE programme quality in Germany is relatively restricted with little variation which<br />
leaves parents not many options as well.<br />
While in the USA, subsidies are <strong>of</strong>fered to disadvantaged families, but considering the high fees<br />
for child care parents normally have very limited choices <strong>of</strong> both affordable <strong>and</strong> accessible centrebased<br />
ECE services (Cryer, et al., 2002). The availability <strong>of</strong> certain types <strong>of</strong> ECE provisions becomes<br />
the limitation <strong>of</strong> parents’ choices (Peyton, Jacobs, O'Brien, & Roy, 2001). Another outst<strong>and</strong>ing feature<br />
is the considerable variation existing in both sponsorship <strong>and</strong> quality <strong>of</strong> American ECE system. The<br />
arrangements <strong>of</strong> child care are <strong>of</strong> great variety to meet differing American families. However, the<br />
provisions <strong>of</strong> high quality child care, in particular the affordable ones, are scarce resource (Cryer, et al.,<br />
2002).<br />
China’s case is more like the situation in the USA. The types <strong>of</strong> ECE services range widely with<br />
certain constraints in fact confining parents’ options to few choices. ‘Parental choice is limited in each<br />
country, but in different ways‘ (Cryer, et al., 2002, p. 263).<br />
Overall, progressive development has been made in many aspects across various regions, but<br />
new progresses always deserve continuous efforts to make it better <strong>and</strong> better.<br />
Access to Early Childhood Education<br />
The aim <strong>of</strong> investigations relating to this topic is to figure out what factors really matter in ECE<br />
systems to make high quality services for child care <strong>and</strong> education. The concept <strong>of</strong> ECE programme<br />
quality varies in accordance with values, beliefs, needs etc. (Cryer, et al., 2002, as cited in Bernhard &<br />
Gonzalez- Mena, 2000; Dhalberg, Moss, & Pence, 1999; Moss, 1994), whereas ‘when quality<br />
definition are closely inspected, the themes <strong>of</strong> these core elements appear repeatedly, with only the<br />
details differing’ (Cryer, et al., 2002, p. 261).<br />
Various research methods <strong>of</strong> social science have been adopted to examine each <strong>and</strong> every aspect<br />
in this field. A range <strong>of</strong> elements related to quality characteristics have influenced developmental<br />
outcomes to various extent (Burchinal & Cryer, 2003). Features <strong>of</strong> quality programming are to be<br />
examined ‘including st<strong>and</strong>ards for curriculum <strong>and</strong> assessment, health <strong>and</strong> safety, class size <strong>and</strong><br />
teacher/child ratios, parent involvement, pr<strong>of</strong>essional development <strong>and</strong> staff qualifications' (Warash,<br />
Ward, & Rotilie, 2008, p. 646). In brief, four categories consisting <strong>of</strong> quality, safety, convenience <strong>and</strong><br />
availability are main reasons decide the result (Davis & Connelly, 2005). Normally, the demographic<br />
characteristics <strong>of</strong> family also affect the process <strong>of</strong> parents decision-making (Peyton, et al., 2001).<br />
Some scholars (Taylor, Dearing, & McCartney, 2004) find that family economic status matters in<br />
children’s development, especially for children from low income families. Through their own survey,<br />
Tietze <strong>and</strong> Cryer (1999) view availability <strong>and</strong> affordability as major criteria for quality ECE<br />
programmes. Availability represents provision <strong>of</strong> services, acceptable distance, convenient schedule,<br />
suitable programmes etc. Affordability is the principle in the light <strong>of</strong> European original idea to make<br />
ECE services available to parents. (Tietze & Cryer, 1999)<br />
However, the needs <strong>of</strong> children <strong>and</strong> parents have received a comprehensive awareness <strong>of</strong> both<br />
pr<strong>of</strong>essionals <strong>and</strong> the public nowadays in the European countries (Tietze & Cryer, 1999). In most<br />
cases, parents take full responsibility to select ECE services for their children <strong>and</strong> family based on<br />
their own criteria as child care consumers (Cryer & Burchinal, 1997). Consequently, the opinion <strong>of</strong><br />
parents, the decision-makers <strong>of</strong> early education selection, will be valued in the process <strong>of</strong> assessing the<br />
quality ECE services. Cryer et al. (2002) claimed that factors influencing parents perceptions <strong>of</strong> ECE<br />
quality, such as affordability, accessibility, traditions related to ECE etc., is not fixed in different<br />
countries. By rating, ranking <strong>and</strong> conjoint analysis, Rose <strong>and</strong> Elicker (2008) find that warmth <strong>and</strong><br />
educational level <strong>of</strong> caregivers score very high in mothers’ considerations. Grace <strong>and</strong> O’cass (2003)<br />
mention that ethics, service delivery, <strong>and</strong> service failure response are the main reasons when parents<br />
choose <strong>and</strong> switch child care centres.<br />
Another point made by Cryer <strong>and</strong> Burchinal (1997) is that child care area has been viewed as a<br />
market obeying the basic principle <strong>of</strong> supply <strong>and</strong> dem<strong>and</strong>. Supply is provided by all types <strong>of</strong> ECE<br />
programmes, as dem<strong>and</strong>, in most cases, is the needs <strong>of</strong> parents having their children receive quality<br />
care <strong>and</strong> education. According to market rules, the magnitude <strong>of</strong> child care supply is inevitably<br />
affected by the preference <strong>of</strong> parents. They pointed out that parents are not well informed with<br />
150
adequate <strong>and</strong> accurate information <strong>of</strong> the services their children have received. They cited economist<br />
James Walker (1991) as ‘The lack <strong>of</strong> perfect information is the most striking difference between the<br />
child care market <strong>and</strong> the idealized perfect market…. Consumers do not know the quality <strong>of</strong> care<br />
<strong>of</strong>fered by providers once identified. Even after a long period <strong>of</strong> use, consumers will not be fully<br />
informed about the behaviour <strong>of</strong> the provider’ (p. 65).<br />
Since parent attitudes matter, what are the determinants in their decisions? Noble (2007) set four<br />
sub categories to cover the major influencing elements in parents’ decisions (p. 53):<br />
1. Parent relationship with child<br />
2. Influence <strong>of</strong> significant others<br />
3. Parent underst<strong>and</strong>ings <strong>of</strong> childhood<br />
4. Maximising the child’s potential<br />
She even emphasises the essential relationship between the needs <strong>of</strong> family <strong>and</strong> children (Noble,<br />
2007). Moreover, ‘ As evidenced from the pilot study, adults’ needs (such as hours <strong>of</strong> operation, cost,<br />
<strong>and</strong> location) were rated as influential parents’ child care choices’ (Seo, 2003, p. 648). Interestingly,<br />
survey result indicates that parents consider convenience factors (cost, location etc.) more than quality<br />
indicators (safety, pedagogy, interaction etc.) when weighing determinants <strong>of</strong> ECE services (Cryer &<br />
Burchinal, 1997, as cited in Kisker & Maynard, 1991). It shows that parents place their own needs<br />
prior to children’s which may lead to a hasty decision (Gable & Cole, 2000). Nevertheless, Gable <strong>and</strong><br />
Cole (2000) give reasonable explanations to those phenomena: if the service is too expensive to afford,<br />
too far to deliver their child, or difficult to cooperate in time schedule, the decision is actually<br />
unrealistic. ‘In some ways, the priorities <strong>of</strong> parents, the state <strong>and</strong> experts in the field are not the same’<br />
("Parents pleased with child care options <strong>and</strong> quality," 2008, p. 1). It is uncertain that to what extent<br />
the quality factors in parental values match those <strong>of</strong> pr<strong>of</strong>essionals (Peyton, et al., 2001).<br />
In summary, up until now, the importance <strong>of</strong> ECE has been acknowledged all over the world.<br />
Supportive forces come from all fields <strong>of</strong> society to attribute to this career. ECE has become an<br />
integrated concept which involves complex factors concerning the prospect <strong>of</strong> entire human<br />
community. Accompanying with accelerated development <strong>of</strong> socio-economy <strong>and</strong> human capacity,<br />
ECE is looking forward to contribution <strong>of</strong> everybody concerned to meet the dem<strong>and</strong>s <strong>of</strong> both societies<br />
<strong>and</strong> individuals.<br />
References:<br />
Burchinal, M. R., & Cryer, D. (2003). Diversity, child care quality, <strong>and</strong> developmental outcomes.<br />
Early Childhood Research Quarterly, 18(4), 401-426.<br />
Cryer, D., & Burchinal, M. (1997). Parents as child care consumers. Early Childhood Research<br />
Quarterly, 12(1), 35-58.<br />
Cryer, D., Tietze, W., & Wessels, H. (2002). Parents' perceptions <strong>of</strong> their children's child care: a crossnational<br />
comparison. Early Childhood Research Quarterly, 17(2), 259-277.<br />
Cunha, F., Heckman, J. J., Lochner, L., & Masterov, D. V. (2005). Interpreting the evidence on life<br />
cycle skill formation. In E. Hanushek & F. Welch (Eds.), H<strong>and</strong>book <strong>of</strong> the Economics <strong>of</strong><br />
Education. North Holl<strong>and</strong>.<br />
Davis, E. E., & Connelly, R. (2005). The influence <strong>of</strong> local price <strong>and</strong> availability on parents' choice <strong>of</strong><br />
child care. Population Research <strong>and</strong> Policy Review, 24, 301-334.<br />
Gable, S., & Cole, K. (2000). Parents' child care arrangements <strong>and</strong> their ecological correlates. Early<br />
Education & Development, 11(5), 549-572.<br />
Grace, D., & O'cass, A. (2003). Child care services: an Exploratory study <strong>of</strong> choice, switching <strong>and</strong><br />
search behaviour. European Journal <strong>of</strong> Marketing, 37(1/2), 107-132.<br />
Heckman, J. J. (2006). The economics <strong>of</strong> investing in early childhood. Paper presented at the 2006<br />
NIFTeY Conference - Prevention: Invest Now or Pay Later, Sydney, Australia.<br />
Kamerman, S. B. (2000). Early childhood education <strong>and</strong> care: an overview <strong>of</strong> developments in the<br />
OECD countries. International Journal <strong>of</strong> Educational Research, 33(1), 7-29.<br />
Knudsen, E. I., Heckman, J. J., Cameron, J. L., & Shonk<strong>of</strong>f, J. P. (2006). Economic, neurobiological,<br />
<strong>and</strong> behavioral perspectives on building America's future workforce. Paper presented at the<br />
The National Academy <strong>of</strong> Science <strong>of</strong> the United States <strong>of</strong> America, Washington.<br />
151
Noble, K. (2007). Parent choice <strong>of</strong> early childhood education <strong>and</strong> care services. Australian Journal <strong>of</strong><br />
Early Childhood, 32(2), 51-57.<br />
Nyl<strong>and</strong>, B., Nyl<strong>and</strong>, C., & Maharaj, E. A. (2009). Early childhood education <strong>and</strong> care in urban China:<br />
the importance <strong>of</strong> parental choice. Early Child Development <strong>and</strong> Care, 179(4), 517-528.<br />
Parents pleased with child care options <strong>and</strong> quality. (2008) & A. Dickman (Vol. Ed.): Vol. 96. Survey<br />
results suggest contrasting child care priorities (pp. 1-11). Wisconsin Public Policy Forum.<br />
Peisner-Feinberg, E. S., & Burchinal, M. R. (1997). Relations between preschool children's child-care<br />
experiences <strong>and</strong> concurrent development: The cost, quality, <strong>and</strong> outcomes study. Merrill -<br />
Palmer Quarterly, 43(3), 451-477.<br />
Peyton, V., Jacobs, A., O'Brien, M., & Roy, C. (2001). Reasons for chosing child care: associations<br />
with family factors, quality, <strong>and</strong> satisfaction. Early Childhood Research Quarterly, 16, 191-<br />
208.<br />
Rose, K. K., & Elicker, J. (2008). Parental decision making about child care. Journal <strong>of</strong> Family Issues,<br />
29(9), 1161-1184.<br />
Seo, S. (2003). Early child care choices: a theoretical model <strong>and</strong> research implications. Early Child<br />
Development <strong>and</strong> Care, 173(6), 637-650.<br />
Taylor, B. A., Dearing, E., & McCartney, K. (2004). Incomes <strong>and</strong> outcomes in early childhood. The<br />
Journal <strong>of</strong> Human Resources, 39(4), 980-1007.<br />
Tietze, W., & Cryer, D. (1999). Current trends in European early child care <strong>and</strong> education. The Annals<br />
Of The American Academy Of Political And Social Science, 563(1), 175-193.<br />
Tietze, W., Cryer, D., Bairrão, J., Palacios, J., & Wetzel, G. (1996). Comparisons <strong>of</strong> observed process<br />
quality in early child care <strong>and</strong> education programs in five countries. Early Childhood<br />
Research Quarterly, 11(4), 447-475.<br />
Walker, J. (1991). Public policy <strong>and</strong> the supply <strong>of</strong> child care services. New York: Russel Sage<br />
Foundation.<br />
Warash, B. G., Ward, C., & Rotilie, S. (2008). An exploratory study <strong>of</strong> the application <strong>of</strong> early<br />
childhood environment rating scale critera. Education, 128(4), 645-658.<br />
152
Five Dimensions <strong>of</strong> Entrepreneurship: A Study <strong>of</strong> First <strong>and</strong> Second<br />
Generations Chinese Entrepreneurs in Melbourne<br />
Henri Lee 1 *<br />
* Presenter<br />
1. <strong>Faculty</strong> <strong>of</strong> Business <strong>and</strong> Enterprise, Swinburne University <strong>of</strong> Technology<br />
The primary objective is to investigate the success achieved by first <strong>and</strong> second<br />
generation Chinese entrepreneurs so as to improve underst<strong>and</strong>ing <strong>of</strong> their impact <strong>and</strong><br />
contribution in the development <strong>of</strong> Melbourne. Ultimately, this study seeks to provide an<br />
insight into their entrepreneurial activities <strong>and</strong> their social <strong>and</strong> economic contributions to<br />
themselves <strong>and</strong> the nation, leading to Chinese being recognized as an important source <strong>of</strong><br />
economic growth. Five dimensions <strong>of</strong> Chinese entrepreneurs from the first <strong>and</strong> second<br />
generations were identified for investigation <strong>and</strong> exploration. These are their background,<br />
their motivations for business ownership, traditional Confucian values, difficulties <strong>and</strong><br />
hurdles, <strong>and</strong> their overseas expansionary vision. Ultimately, the study concentrates on their<br />
successes.<br />
153
An Exploration <strong>of</strong> Country <strong>of</strong> Origin Effect on Union Policies in Chinese<br />
Multinational Enterprises<br />
Shuqin Zhu 1 *<br />
* Presenter<br />
1. Department <strong>of</strong> Management, Monash University, Clayton Compus, Wellington Rd,<br />
Clayton 3800, Australia<br />
There are an increasing number <strong>of</strong> studies investigating country <strong>of</strong> origin effect on the<br />
management practices <strong>of</strong> multinational enterprises (MNEs). On the basis <strong>of</strong> the research on<br />
multinational corporations from developed countries, these studies argue that country <strong>of</strong><br />
origin effect has a significant impact on MNEs’ operation in other countries. However, few <strong>of</strong><br />
them have investigated country <strong>of</strong> origin effect on the management practices <strong>of</strong> MNEs from<br />
developing country. Using qualitative methodology, this research aims to explore whether<br />
<strong>and</strong> how country <strong>of</strong> origin effect exemplifies itself in the union policies <strong>of</strong> Chinese<br />
multinational enterprises. This paper first reviews the relevant literature on country <strong>of</strong> origin<br />
effect. Then, it proceeds to examine the studies on the union policies <strong>of</strong> multinational<br />
corporations. Following this, the paper discusses the industrial relations system in China.<br />
Finally, the paper proposed some questions for future research.<br />
154
Authors Title Affiliation<br />
Wei Tao, Bao<br />
Hongchun, Dru<br />
Morrish, & Min Gu<br />
Gang Li, Ranjeet<br />
Singh, Dan Li, Chunxia<br />
Zhao, Liying Liu, &<br />
Paul A. Webley<br />
Shan Liu, Charles Ma,<br />
Lina Wng, Ross<br />
Coppel, & Gareth M.<br />
Forde<br />
Xin Gao & Philip J.<br />
Marriott<br />
Junfei Tian, Xu Li, &<br />
Wei Shen<br />
Supercontinuum generation for fiber-optic nonlinear<br />
microscopy<br />
Synthesis <strong>of</strong> biomorphic zeolite honeycomb<br />
monoliths with 16000 cells per square inch<br />
Synthesis <strong>and</strong> delivery <strong>of</strong> Malaria DNA vaccines via<br />
biodegradable polymer carrier systems<br />
Gas chromatographic retention indices <strong>and</strong> mass<br />
spectra <strong>of</strong> trimethylsilyl derivatives <strong>of</strong> flavonoids on<br />
capillary columns with non-polar stationary phase<br />
Micr<strong>of</strong>luidic systems on paper<br />
Swinburne<br />
University <strong>of</strong><br />
Technology<br />
Monash<br />
University<br />
Monash<br />
University<br />
RMIT<br />
University<br />
Monash<br />
University<br />
Yan-e Gao, TG Johns,<br />
E Dimitriadis, E<br />
Menkhorst, Joanne<br />
Mockler, BRG<br />
Williams, & S Tong<br />
Ying Lu, Ramanie<br />
Samarathunge, &<br />
Charmine Hartel<br />
Dan Wang, Susan<br />
Freeman & Kate<br />
Hutchings<br />
Mingqiong Zhang,<br />
Chris Nyl<strong>and</strong>, &<br />
Cherrie J. Zhu<br />
Liang B. Y. Du &<br />
Arthur J. Lowery<br />
Liang B. Y. Du &<br />
Arthur J. Lowery<br />
Liang B. Y. Du &<br />
Arthur J. Lowery<br />
Haidong Zheng, Abu<br />
Sadek, David Yao,<br />
Michael Breedon, &<br />
Kourosh Kalantarzadeh<br />
MicroRNA expression changes detected by<br />
microarray with forskolin syncytialisation <strong>of</strong> BeWo<br />
cells<br />
The Impact <strong>of</strong> Acculturation on Pr<strong>of</strong>essional Chinese<br />
Immigrants in the Australian Workplace<br />
Outward Foreign Direct Investment from China<br />
The Institution <strong>of</strong> Mingongzhi in Contemporary<br />
China <strong>and</strong> the Strategies <strong>of</strong> MNEs: An Institutional<br />
Analysis<br />
Fiber Nonlinearity Precompensation for Long-hual<br />
links using Direct-Detection Optical OFDM<br />
Improving Nonlinear Precompensation in Direct-<br />
Detection Optical OFDM Communications Systems<br />
Improved Nonlinearity precompensation for Longhual<br />
High-date-rate Transmission using Coherent<br />
Optical OFDM<br />
Rapid Formation <strong>of</strong> Thick <strong>and</strong> Transparent Anodic<br />
TiO 2 Nanotubular Films from Sputtered Ti<br />
Monash<br />
Institute <strong>of</strong><br />
Medical<br />
Research<br />
Monash<br />
University<br />
Monash<br />
University<br />
Monash<br />
University<br />
Monash<br />
University<br />
Monash<br />
University<br />
Monash<br />
University<br />
RMIT<br />
University<br />
156
Supercontinuum generation for fiber-optic nonlinear microscopy<br />
Wei Tao 1 , Bao Hongchun 1 , Dru Morrish 1 , & Min Gu 1<br />
1. Centre for Micro-Photonics, Swinburne University <strong>of</strong> Technology<br />
We report the experimental investigation <strong>of</strong> the transmission characteristics <strong>of</strong> a<br />
supercontinuum through double-clad photonic crystal fiber for fiber-optic nonlinear<br />
microscopy.<br />
157
Synthesis <strong>of</strong> biomorphic zeolite honeycomb monoliths with 16000 cells per<br />
square inch<br />
Gang Li, 1,2 Ranjeet Singh, 1,2 Dan Li, 1 Chunxia Zhao, 1,3 Liying Liu 1,2,4 <strong>and</strong> Paul A. Webley 1,2<br />
1. Department <strong>of</strong> Chemical Engineering, Monash University, Wellington Road, Clayton, VIC<br />
3800, Australia<br />
2. Cooperative Research Centre for Greenhouse Gas Technologies, Grnd Flr NFF House, 14-<br />
16 Brisbane Ave, Barton, ACT 2600, Australia<br />
3. State Key Laboratory <strong>of</strong> Advanced Technology for Materials Synthesis <strong>and</strong> Processing,<br />
School <strong>of</strong> Materials Science <strong>and</strong> Engineering, Wuhan University <strong>of</strong> Technology, Wuhan<br />
430070, PR China<br />
4. School <strong>of</strong> Materials & Metallurgy, Northeastern University, Shenyang 110004, PR China<br />
Zeolite NaX <strong>and</strong> silicalite honeycomb monoliths were synthesized via hydrothermal<br />
growth <strong>of</strong> zeolites on the interior surface <strong>of</strong> cuttlefish bone by a novel flow coating technique.<br />
The resultant samples were characterized by SEM, XRD, EDX, N2 adsorption <strong>and</strong> TGA. The<br />
loading <strong>of</strong> nano-crystalline silicalite (36.4 wt%) was higher than for NaX type zeolite. The<br />
biomorphic honeycomb monoliths were found to have an exceptionally high cell density <strong>of</strong><br />
16000 cells per square inch which is ten times higher than the best synthetic ones. The<br />
honeycombs were evaluated for their hydrodynamic <strong>and</strong> kinetic properties, showing<br />
substantial improvement <strong>of</strong> mass transfer rate with lower pressure drops than conventional<br />
zeolite packing.<br />
158
Gas chromatographic retention indices <strong>and</strong> mass spectra <strong>of</strong> trimethylsilyl<br />
derivatives <strong>of</strong> flavonoids on capillary columns with non-polar stationary<br />
phases<br />
Xin Gao 1 2 & Philip J. Marriott 2<br />
1. Health Science Centre, Peking University<br />
2. School <strong>of</strong> Applied Sciences, RMIT University<br />
Flavonoids are one <strong>of</strong> the most important groups <strong>of</strong> compounds occurring in plants,<br />
which exhibit a wide range <strong>of</strong> biological effects <strong>and</strong> act as natural antioxidants. They<br />
comprise the main class <strong>of</strong> polyphenols in diet <strong>and</strong> Chinese herbs, <strong>and</strong> usually exist in the<br />
presence <strong>of</strong> complex matrix, which considerably hinders the identification <strong>of</strong> these analytes.<br />
Chemical analysis <strong>of</strong> flavonoids in samples allows for their precise compositional assignment,<br />
<strong>and</strong> a number <strong>of</strong> advanced methods exist for this task. Analysis by using gas chromatography<br />
(GC) <strong>and</strong> comprehensive two-dimensional gas chromatography (GC×GC) may be regarded as<br />
very promising methods for studying flavonoids in diet <strong>and</strong> herbs due to the ability to speciate<br />
different types <strong>of</strong> flavonoids. In the present work, it was <strong>of</strong> interest to analyse a large database<br />
<strong>of</strong> flavonoids by GC <strong>and</strong> GC×GC in order to present the GC <strong>and</strong> GC×GC characteristics <strong>of</strong> 29<br />
compounds. These data include basic retention index information on high-performance<br />
capillary columns with non-polar stationary phases. A number <strong>of</strong> previously unreported<br />
flavonoids have been included in this study, <strong>and</strong> so this report presents for the first time mass<br />
spectral <strong>and</strong> retention data for a series <strong>of</strong> trimethylsilyl derivatives <strong>of</strong> those flavonoids.<br />
159
The Impact <strong>of</strong> Acculturation on Pr<strong>of</strong>essional Chinese Immigrants in the<br />
Australian Workplace<br />
Ying Lu 1 , Ramanie Samarathunge 1 , & Charmine Hartel 1<br />
1. Department <strong>of</strong> Management, Building 11, Monash University, Clayton Campus,<br />
Wellington Road Clayton Victoria 3800 AUSTRALIA<br />
Pr<strong>of</strong>essional Chinese immigrants (PCIs) encounter many adjustment difficulties arising<br />
from cultural <strong>and</strong> social differences after entering into the Australian workplace (Birrell &<br />
Healy, 2008; Ho, 2006). Evidence shows that Australian employers do not pay adequate<br />
attention to the integration <strong>of</strong> PCIs, which has already caused adjustment problems in the new<br />
homel<strong>and</strong> <strong>and</strong> has negatively impacted on their career development <strong>and</strong> success (Birrell &<br />
Healy, 2008). Given the important role <strong>of</strong> PCIs in Australia with regards to creating a diverse<br />
workforce, boosting economic growth <strong>and</strong> narrowing the gap in the skills shortage <strong>of</strong> the<br />
country, their integration process into the Australian workplace deserves due consideration.<br />
The research looks at this issue from acculturation, which is an inevitable process that every<br />
immigrant has to experience. To sum up, the aim <strong>of</strong> this research is to examine the individual<br />
<strong>and</strong> workgroup contextual factors which may affect PCIs’ acculturation strategy as well as the<br />
effect <strong>of</strong> various acculturation strategies on selected individual work outcomes in the<br />
workgroup. In particular, it explores the possible mediating function <strong>of</strong> acculturation strategy<br />
in a workplace context.<br />
160
Outward Foreign Direct Investment from China<br />
Dan Wang 1 , Susan Freeman 1 & Kate Hutchings 1<br />
1. Department <strong>of</strong> Management, Building N , 27 Sir John Monash Drive, Caulfield East<br />
Victoria 3145, AUSTRALIA<br />
This poster examines the unique motivations <strong>and</strong> strategies for outward foreign direct<br />
investment (OFDI) undertaken by Chinese multinationals (MNEs), <strong>and</strong> addresses the<br />
challenges associated with these unique motivations <strong>and</strong> strategies. We find that the activities<br />
<strong>of</strong> Chinese MNEs are strongly influenced by the institutional environment <strong>of</strong> China, <strong>and</strong> that<br />
despite the unique resources Chinese MNEs possess, institutional theory rather than the<br />
resource based view is best placed as a theoretical framework when examining Chinese OFDI.<br />
161
The Institution <strong>of</strong> Mingongzhi in Contemporary China <strong>and</strong> the Strategies <strong>of</strong><br />
MNEs: An Institutional Analysis<br />
Mingqiong Zhang 1 , Chris Nyl<strong>and</strong> 1 , & Cherrie J. Zhu 1<br />
1. Department <strong>of</strong> Management, Building 11, Monash University, Clayton Campus,<br />
Wellington Road Clayton Victoria 3800 AUSTRALIA<br />
From the early 1980s, China has entered into an age <strong>of</strong> rural-urban migration. Millions <strong>of</strong><br />
farmers have flooded into cities <strong>and</strong> towns for work <strong>and</strong> significantly reshaped the character<br />
<strong>of</strong> Chinese society. Confronted with the flood <strong>of</strong> rural-urban migrants, the central government,<br />
city governments, private <strong>and</strong> public companies <strong>and</strong> urban residents have responded on the<br />
various bases. Consequently, a combination <strong>of</strong> linked rules, values, norms, <strong>and</strong> patterned<br />
practices that structure the way that rural migrants should be treated in urban areas has<br />
emerged <strong>and</strong> rural migrants are being socially excluded. From the perspective <strong>of</strong><br />
neoinstitutionalism, these linked hukou-based regulative, normative <strong>and</strong> cognitive elements<br />
demonstrate that a new institution has emerged, which is termed in this study the institution <strong>of</strong><br />
mingongzhi. Like the terms slavery <strong>and</strong> apartheid, the institution <strong>of</strong> mingongzhi refers to a<br />
social-economic system under which rural migrant workers are socially excluded in urban<br />
areas <strong>of</strong> contemporary China based on their household registration (hukou) status. As a major<br />
institution <strong>of</strong> the labour market in contemporary China, mingongzhi has had a pr<strong>of</strong>ound<br />
influence on China’s urban employment environment, industrial relations (IR) <strong>and</strong> human<br />
resource management (HRM) practices. It is an ideal variable to conduct analysis <strong>of</strong> the<br />
strategic response <strong>of</strong> MNEs to the institutional characteristics <strong>of</strong> a host country. This is the<br />
more so as international business scholars have accorded little systematic attention to the<br />
relationship between institutions <strong>and</strong> multinational enterprises (MNEs).<br />
162
Fiber Nonlinearity Precompensation for Long-hual links using Direct-<br />
Detection Optical OFDM<br />
Liang B. Y. Du 1 * <strong>and</strong> Arthur J. Lowery 1<br />
1. Electrical <strong>and</strong> Computer Systems Engineering, Monash University, Clayton, VIC 3800,<br />
Australia<br />
The use <strong>of</strong> nonlinearity precompensation in direct-detection optical orthogonal frequency<br />
division multiplexed links is investigated by simulation. Because <strong>of</strong> the presence <strong>of</strong> a strong<br />
optical carrier its performance is poorer than for coherent systems: with compensation the<br />
signal quality is found to vary almost periodically across the signal b<strong>and</strong>. We propose <strong>and</strong><br />
explain the operation <strong>of</strong> two optical, one electrical <strong>and</strong> one computational method <strong>of</strong><br />
removing this periodic variation. Optical filtering <strong>of</strong> one sideb<strong>and</strong> at the receiver is most<br />
effective, but a substantial improvement can be obtained by a simple modification to the<br />
precompensation algorithm.<br />
163
Improving Nonlinear Precompensation in Direct-Detection Optical OFDM<br />
Communications Systems<br />
Liang B. Y. Du 1 * <strong>and</strong> Arthur J. Lowery 1<br />
1. Electrical <strong>and</strong> Computer Systems Engineering, Monash University, Clayton, VIC 3800,<br />
Australia<br />
Carrier boosting at the receiver enables direct detection optical OFDM (DDO-OFDM) to<br />
outperform coherent OOFDM in the nonlinear limit. Boosting also improves the effectiveness<br />
<strong>of</strong> nonlinearity precompensation substantially.<br />
164
Improved Nonlinearity precompensation for Long-hual High-date-rate<br />
Transmission using Coherent Optical OFDM<br />
Liang B. Y. Du 1 * <strong>and</strong> Arthur J. Lowery 1<br />
1. Electrical <strong>and</strong> Computer Systems Engineering, Monash University, Clayton, VIC 3800,<br />
Australia<br />
We show that the performance <strong>of</strong> precompensation <strong>of</strong> fiber nonlinearity in coherent<br />
optical OFDM systems operating at up to 60 Gbps/ polarization can be improved by electrical<br />
filtering the precompensation signal. The optimal filter b<strong>and</strong>width is related to the FWM<br />
efficiency spectrum when dispersion is considered.<br />
165
Rapid Formation <strong>of</strong> Thick <strong>and</strong> Transparent Anodic TiO2 Nanotubular<br />
Films from Sputtered Ti<br />
Haidong Zheng 1 , Abu Sadek 1 , David Yao 1 , Michael Breedon 1 , & Kourosh Kalantarzadeh<br />
1<br />
1. School <strong>of</strong> electrical <strong>and</strong> computer engineering, RMIT University.<br />
This poster reports the new insights found in fast anodization <strong>of</strong> sputtered Ti .<br />
166