Complete Report - University of New South Wales

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORT1. DIRECTORS’REPORTPhotovoltaics involves the direct conversion of light, normally sunlight, into electricitywhen falling upon devices known as solar cells. Silicon is the most common materialused to make these photovoltaic cells, similarly to its key role in microelectronics.The Australian Research Council (ARC) PhotovoltaicsCentre of Excellence commenced at the Universityof New South Wales (UNSW) on 13th June, 2003with funding initially until 31st December 2007.Upon review in 2005 and 2006, this funding wasextended until 31 st December 2010. Criteria forsuch extension included “progress of the Centretowards becoming independent of the ARC Centresof Excellence Scheme for funding”. With the rapidgrowth of the photovoltaics industry since theCentre’s commencement and the Centre’s strongand seminal links to this industry, the Centre wasclearly in a much stronger position for independentoperation beyond 2010 than many others in itsfunding cohort. Accordingly, while remaining underthe ARC Centres of Excellence umbrella beyond2010 in recognition of the quality of its research,the Centre will now be dependent upon industryrelatedfunding for research with near-termoutcomes and upon more academically-orientatedARC and international schemes for its longtermresearch.The Centre‘s mission remains to advance siliconphotovoltaic research on three separate fronts, aswell as to apply these advances to the related fieldof silicon photonics. The educational activities of theformer Key Centre for Photovoltaic Engineering arealso integrated into the Centre.Over the period of funding, photovoltaics hasbecome the world’s most rapidly growingenergy source, with markets increasing at acompounded rate of above 40%/year over thisperiod. The electricity generating capacity of newphotovoltaic product installed in 2006 exceedednew nuclear power capacity for the first time, withthe gap widening every year since. In Germany,photovoltaics has been the largest source of newelectricity generating capacity for each of thelast four years and became the second largestsuch source across Europe in 2010, ahead of windgenerators and behind only gas turbines. Reducingprices, at least partly traceable to past Centreinitiatives, means the technology is reaching “retailgrid parity”, where the cost of electricity generatedusing photovoltaics can compete at the point ofuse with normal retail electricity prices in manyparts of the world. This is expected to lead to aperiod of self-sustaining growth with photovoltaicspositioned to become one of the world’s majorindustries of the 21 st century. Even in Sydney where“cheap” electricity is supplied by power stationsbuilt on coal fields, photovoltaics is coming closeto being competitive with retail electricity priceswithout subsidies.Most present photovoltaic sales are of “firstgeneration”solar cells made from silicon wafers,similar to the wafers used in microelectronics.The Centre’s world-leadership with these “firstgeneration”devices, with international recordsfor the highest-performing silicon cells in mostmajor categories (including the outright highestperformingcells and modules) was recentlyextended with the demonstration of the first 25%efficient silicon solar cell. First-generation Centreresearch addresses the dual challenges of reducingcost and further improving efficiency. The rapidgrowth of the industry is generating widespreadinterest in ongoing innovations of the Centre’sfirst generation technology with several distincttechnologies now in large-scale production andadditional licences signed during 2010.Silicon is quite brittle so silicon wafers have to bereasonably thick, a fraction of a millimetre, to besufficiently rugged for processing into solar cellswith reasonable yield. Without this mechanicalconstraint, silicon would perform well even if verythin, even 100 times thinner than present wafers.Centre researchers have pioneered an approachwhere very thin silicon layers are deposited directlyonto a sheet of glass with the glass providingthe required mechanical strength. This “secondgeneration”approach gives enormous potential costsavings. Not only are the costly processes involvedin making wafers no longer required, but also thereis an enormous saving in silicon material. Cells alsocan be made more quickly over the entire area oflarge glass sheets. The Centre is at the forefrontof international research with such “secondgeneration”,silicon based approaches, with thefirst commercial product from “spin-off”, CSG Solar,2

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTFormer student and postdoctoralresearcher Dr Zhengrong Shi on thecover of Photon International asfounder and chief of the world’sNo. 1 solar company.Suntech HighlightsAn electron beam induced current (EBIC) image of thelaser doped region of a LDSE solar cell superimposedupon a scanning electron micrograph (SEM) of thesame region. The inset shows Dr Brett Hallam, a keymember of the team producing the new result.new cell efficiency recordNumber 4 on the list and another member of the“gigawatt” club with 1.1 GW produced in 2010 wasYingli Green Energy Holding. Dr Guoxiao Yao, whowas awarded his PhD from UNSW in 2005, was CTOat the time of the company’s listing on the NYSEin 2007. Yingli’s present CTO is Dr Dengyuan Song,who also received his PhD from the Centre in 2005.Number 5 on the list was Trina Solar with 1.1 GWproduced. Former PhD student, Dr MohanNarayanan, was CTO at the time of the company’slisting on NASDAQ in 2006. Dr Narayanan is nowCTO of Hanwha Solar One (formerly Solarfun), with0.5 GW of production in 2010. Another former PhDstudent, Fei Yun was CTO of Solarfun at the time ofits initial listing on NASDAQ in 2006. Fei is now Vice-President, Technology at LDK Solar, another majorplayer in the solar field.UNSW-linked companies therefore form an amazing4 of the top 5 in 2010. The remaining companymaking the list of the top 5 was former “NumberOne”, First Solar (USA, Malaysia, Germany).Numerous other of the world’s major photovoltaiccompanies have former Centre members in keytechnical positions. The Centre estimates thenumber as over 30, with the better known includingSunPower (USA, Philippines, Malaysia), Sunergy(China), SolarWorld (Germany), Hyundai Solar(Korea), Gintech (Taiwan), TopCell Solar (Taiwan),CSG Solar (Australia, Germany), Shinsung (Korea),Sulfurcell (Germany), LG Electronics (Korea),Innovalight (USA), Chinalight Solar (China), E-TonSolar (Korea), LDK Solar (China), Topco (Taiwan),MAGI Solar (China), Corum Solar (Taiwan), ApollonSolar (France), Sunrise Global (Taiwan) and RECSolar (Norway, Singapore).Dr Zhengrong Shi recently nicely summarisedAustralia’s impact in photovoltaics in his article “CanAustralia Save the World” (ATSE Focus, February2011) when he observed that, in the past, the UKhad shown the way to use coal for energy, theUSA had shown how to harness atomic power butAustralia and China were showing the world how tobest use solar.19.3% Laser Doped Solar CellA new solar cell efficiency record of 19.3%has been independently confirmed by theFraunhofer Institute in Germany for a large area(157 cm 2 ) laser doped, selective emitter (LDSE)cell fabricated by the Centre using cells largelyprocessed on a standard Centrotherm screenprintingproduction line. The only processvariation apart from a lighter junction diffusionwas to replace the front surface screen-printedmetal with the Centre’s laser doping process tosimultaneously pattern the cell antireflectioncoating and create localised doped regions for thesubsequent self-aligned process of light-inducedplating of Ni/Cu. Key members of the team fromthe Centre responsible for this result, Brett Hallam,Stuart Wenham, Chee Mun Chong, Matt Edwardsand David Jordan, plan to present the new results atthe 37 th IEEE Photovoltaics Specialists Conference inSeattle in June 2011.With further optimisation of the cell design, thelaser doped solar cell is expected to becomethe first low-cost cell capable of energyconversion efficiency above 20% to enter fullscaleproduction.Photoluminescence AwardsThe Centre has pioneered the use ofluminescence imaging technology for thecharacterisation of silicon ingots, wafers andcells with this technology now commercialisedby “spin-off” BT Imaging. This work continuesto attract substantial international attentionwith the work of the two original developersof the technology and the “B” and the “T” inBT Imaging, Deputy Director Associate ProfessorThorsten Trupke and Dr Robert Bardos, werehonoured in July by the award of the BertholdLeibinger Innovationspreis. The photo showsAssociate Professor Trupke accepting the prize inDitzingen, Germany. The Berthold Leibinger Stiftunggives this award every two years for “excellentresearch and development work on the application orgeneration of laser light”.Associate Professor Thorsten Trupkewith co-awardee Dr Robert Bardoson campus.PhotouminescenceawardsSeries resistance map of a siliconsolar cell showing features dueto non-uniform processingtemperatures. The image wascreated by the new techniquedeveloped by Dr HennerKampwerth (inset) during hisdoctoral thesis, resulting in hisselection as 2010 Student Inventorof the Year.2010 student inventor ofthe year5

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTLeft: Artist’s impressions of the new Solar IndustrialResearch Facility (SIRF) due to be completed by mid-2011which will showcase the Centre’s near-term technology.Above: Artist’s impressions of the new Tyree EnergyTechnologies Building that will house Centre research andteaching in the future.new photovoltaic buildingsDr Zhengrong Shi (left), CEO andChairman of Suntech Power withProfessor Martin Green and Plutomodule.450 MW Pluto capacityDoctoral student Nicole Kuepperin December accepting the 2010Energy Institute Awards in Londonon behalf of Suntech/UNSW aswell as the Individual AchievementAward on behalf of Professor MartinGreen. Earlier in the year, Nicolereceived the Australia Youth Trust’sTom Harvey Award for Citizenshipfor 2010.technology awardsLater in the year, Associate Professor Trupke wasalso chosen as the UNSW 2010 Inventor of theYear in recognition of the innovative features andcommercial significance of this work, following inthe footsteps of Centre Director Stuart Wenham,who won this title in 2009. Also recognised in the2010 Awards was Dr Henner Kampwerth as 2010Student Inventor of the Year, for his doctoral thesiswork on imaging series resistance in silicon cells.This work was supervised by Associate ProfessorTrupke and Professor Martin Green.Suntech Establishes 450 MWPluto CapacitySuntech has announced the company hasnow established 450 MW capacity of its highperformance Pluto cell, jointly developed with theCentre. This corresponds to close to $1 billion/yearin product in today’s market.The Pluto technology represents a low costapproach to fabricating the Centre’s world-record,25% efficient silicon cell. In 2009, earlyproduction cells were used in a solar modulethat surpassed the previous world recordof 15.5% for multicrystalline cells. In 2010,Suntech improved this module efficiency to17%. The strengths of the technology wererecognised at the 2010 Energy InstituteAwards in London when Suntech/UNSW werejointly awarded the Technology Award inrecognition of the favourable cost ratio of Plutotechnology “with manufacturing capacity nowwell above any other new silicon cell technologycommercialised worldwide”.Also at the same Awards, Centre ExecutiveResearch Director Martin Green received theAward for Individual Achievement for “unparalleledcontributions to solar cell design, incitement of majorinvestment in solar resulting in vastly improvedperformance and reduced costs, greatly increasingexpectations of what is commercially possible to meetgrowing energy demands”.Accepting the two awards on behalf of the Centre,Suntech and Professor Green, was Nicole Kuepper(photo), herself no stranger to receiving awards.In March, Nicole was awarded the Australia YouthTrust’s Tom Harvey Award for Citizenship for 2010 torecognise outstanding work of a young Australian indevelopment work.New Photovoltaic BuildingsDuring 2010, construction began on two newbuildings on the UNSW Kensington campusdestined for the Centre’s future use.The first of these is the Solar Industrial ResearchFacility (SIRF) on the upper campus, due to becompleted in mid-2011. This impressive, stylishbuilding will house a full-scale, state-of-art solarcell production line, although not with the level ofautomation required for commercial throughputs.A special area is also set aside for applying newprocessing steps to cells otherwise processed onthe main line.The SIRF facility will be used to demonstratethe next generation of the Centre’s wafer-basedtechnology at large scale, accelerating the transferof this technology into production and increasingthe scope and effectiveness of licensing to therapidly growing industry. The facility will also playa key role in Centre research and in the trainingof Centre students as well as, via short courses,the huge numbers of external engineers andtechnicians entering this industry.The second new Centre building is the Tyree EnergyTechnology Building (ETB) being purpose-builton the lower campus to house the Centre’s mainlaboratories. These laboratories are expected to bemoved from their present site during 2012. The new$155 million building is supported by Sir WilliamTyree as well as by a grant of $75 million from theAustralian Government’s Education InvestmentFund. This grant was awarded after a presentationby Centre researchers, Professors Stuart Wenhamand Martin Green as well as doctoral student, NicoleKuepper, to the Selection Committee emphasisingthe past achievements of the Centre and the impactof the new building on boosting Centre activities.Co-located in the building will be other UNSW6

Dr Zi Ouyang – Widely citedfirst author.Professor Martin Green receivingthe Eureka Prize for Leadership inResearch from Dr Megan Clark, ChiefExecutive, CSIRO.ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTDr Supriya Pillai – First author of topcited paper.other awardsProfessor Stuart Wenham (top)and Dr Richard Corkish (above)(accepting award on behalf ofChris McGrath) at the EngineeringExcellence Awards.energy-related activities such as the AustralianEnergy Research Institute (AERI).Other AwardsFollowing on from a best paper award at the 24 thEuropean Photovoltaic Solar Energy Conferencein 2009, then doctoral student Zi Quyang was firstauthor on another paper “Effective light trapping inpolycrystalline silicon thin-film solar cells by meansof rear localized surface plasmons”, published inJuly 2010 which was one of the “monthly top 20most-downloaded articles last year” in AppliedPhysics Letters.The Centre’s role in stimulating the currenthigh level of interest in applying plasmonics tophotovoltaics has been recognised by the intenseinterest in its seminal paper in this field, “Surfaceplasmon enhanced silicon solar cells”, authored bySupriya Pillai (pictured), Kylie Catchpole, ThorstenTrupke and Martin Green, published in the Journalof Applied Physics in May 2007. This has not onlybeen by far the Journal’s most highly cited paperof the 3,470 papers published in 2007, but also hasbeen on the monthly list of the Journal’s 20 mostdownloaded papers (all years) every month since,including each month of 2010.In August 2010, Executive Research DirectorProfessor Martin Green received the Eureka Prize forLeadership in Research, one of the most prestigiousawards in Australian science.In September, it was the turn of Centre Director,Professor Stuart Wenham who was named 2010Professional Engineer of the Year by the SydneyDivision of Australia’s peak engineering body,Engineers Australia. Chris McGrath, a graduate fromthe School, also won the student Engineer of theYear Award at the Sydney Division’s EngineeringExcellence Awards.Undergraduate AwardsIn March, the Dean of Engineering, ProfessorGraham Davies, awarded Dean’s Awards to eightof the Centre’s top undergraudate students fortheir academic achievements in 2009.Included in this group (see photo) was Bai Xuewho, this year, based on her achievements overthe entire photovoltaic engineering course wasawarded the University Medal, the University’saward for the highest academic achievement in thegraduating cohort.Another Centre student, Alice Lang, has beenawarded a prestigious Rhodes Scholarship. “Alicehas really worked hard in the Bachelor of Engineeringin Photovoltaics and Solar Energy Engineering degreeprogram and it is a real pleasure to see her rewardedlike this”, said Dr Richard Corkish, Head of School ofPhotovoltaic and Renewable Energy Engineering.“The engineering degrees that we offer attract somefantastic students. This Rhodes Scholarship couldn’thave gone to a more capable and deserving student”.New BooksFollowing on from the translation of the Centre’shighly regarded textbooks into simplifiedand traditional Chinese in 2008 and 2009, theCentre was invovled in publishing two moretranslated versions in 2010. Martin Green’stext “Solar Cells” was translated into simplifiedChinese during 2010 while his text was alsoreproduced in Japanese translation duringthe year. There are now nine different texts intotal published by the Centre. The translationinto Korean of “Applied Photovoltaics” is alsoin progress.Left to right: Professor GrahamDavies, Yu Feng, Shen-Long Ooi, XueBai, Professor Richard Henry, JohnDurrant, Benjamin Hughes, VincentAllen, Dr Richard Corkish, DongchenLan. Absent: Dong LinProfessor Stuart Wenham and BaiXue, winner of the University Medal.undergraduate awardsnew books publishedby centre7

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTDr Richard Corkish (left) and Professor Martin Green(middle) talking with Minister Wan in the laboratoryarea where the Centre’s world-record 25% efficientcells are fabricated (Credit: NSW Office for Science andMedical Research ).(Left to Right): Dr Gavin Conibeer, Dr Richard Corkish,Professor Martin Green, Dr Mary Archer,Dr Sergey Varlamov, Dr Anita Ho-Baillie, ProfessorStuart Wenham and Dr Ashraf Uddin.ranging from incremental efficiency improvementsin commercial technology to Third GenerationPV theory and its implementation. Dr Archer isworking on a revised edition of her monumentalreference work “Clean Electricity from Photovoltaics”in conjunction with Centre Executive ResearchDirector Professor Martin Green.Left to Right:• Dr Matt EdwardsUNSW, Technology Transfer Team• Mr Valantis VaisUNSW Co-op Scholar• The Hon. Kristina Keneally MPPremier of NSW• Prof. Stuart WenhamUNSW, Director, PhotovoltaicsCentre of Excellence• Ms Adeline SugiantoUNSW, Technology Transfer TeamNSW Premier Visits CentreHigh ProfileVisitorsNSW Premier Visits CentreFollowing on from visits to the Centre in 2009of the President of Korea, President MyungbackLee and the soon-to-be Prime Minister ofAustralia, Julia Gillard, the then-NSW Premier, TheHonourable Kristina Keneally, visited the Centre inMay to announce that the NSW Government haspledged $120 million to support and attract solarenergy projects to the State in conjunction withthe Commonwealth Government’s Solar FlagshipsProgram. “This is about leading the nation on solarenergy projects. Australia is a sun drenched nation andwe must harness this boundless energy,” Ms Keneallysaid during her announcement.Chinese Minister of Science andTechnology VisitAlso early in 2011, The Chinese Minister of Scienceand Technology, Dr Wan Gang, met with ProfessorMartin Green and Dr Richard Corkish. The Ministervisited SIRF, toured the Centre labs where theCentre’s world-record 25% efficient solar cell wasdeveloped and spoke with Chinese postgraduatestudents about their work. At a subsequent NSWGovernment-organised Innovative MaterialsRoundtable, Minister Wan said, “I am impressed bythe commitment shown by Australian Universities tothese new technologies and that so many Chinesegraduates have benefited from this expertise. Inparticular, I should point out that the pioneeringRenewable Energy program set up by UNSW over 10years ago and led by Prof Martin Green is an excellentexample of Australian innovation and shows howAustralia and China can work together to create newfields of technology and opportunity”.Also attending with Minister Wan was the newlyappointed Chinese Consul-General to Sydney,Duan Jielong, on his first official visit to anAustralian university.Baroness Dr Mary Archer Visits LabsEarly in 2011, Baroness Dr Mary Archer, a specialistin solar power conversion, visited the Centre.While here, the Centre’s Deputy Directors gavepresentations in the Centre’s research strands8

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORT

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORT3. STAFF LISTDirectorStuart Ross Wenham, BE BSc PhD UNSW, FTS, FIEEE, FIEAust(Scientia Professor)Executive Research DirectorMartin Andrew Green, BE MEngSc Qld., PhD DSc(Hon) McMaster, DEngUNSW, FAA, FTSE, FIEEE (Scientia Professor and Federation Fellow)Deputy Directors• Gavin Conibeer, BSc MSc London, PhD Southampton• Thorsten Trupke, PhD Karlsruhe• Ashraf Uddin, BSc (Hons) MSc Dhaka U, Bangladesh;PhD Osaka U, SMIEEE• Sergey Varlamov, PhD MoscowBusiness, Technology & Operations ManagerMark D. Silver, BE UNSW, GMQ AGSMUndergraduate Co-ordinatorAshraf Uddin, BSc (Hons) MSc Dhaka U, Bangladesh;PhD Osaka U, SMIEEPostgraduate Co-ordinator• Alistair B. Sproul, BSc (Hons) Sydney, PhD UNSWAssociate Professors• Gavin Conibeer, BSc MSc London, PhD Southampton• Alistair B. Sproul, BSc (Hons) Sydney, PhD UNSW• Thorsten Trupke, PhD KarlsruheSenior Lecturers• Stephen Bremner, BSc(Hons), PhD UNSW• Richard Corkish, BEng (Com) RMIT, PhD UNSW• Evatt Hawkes, BSc (Hons) BEng (Hons) W. Aust., PhD Cantab.• Alison Lennon, BSc(Hons) Sydney, PhD Sydney, PhD UNSW• Geoffrey J. Stapleton, BE UNSW• Ashraf Uddin, BSc(Hons) MSc Dhaka U, Bangladesh;PhD Osaka U, SMIEEE• Muriel Watt, BSc N.E., PhD MurdochLecturers• Anna Bruce, BE(Hons) PhD UNSW10

• Merlinde Kay BSc(Hons) PhD UNSW• Martha Lenio, B.A.Sc, Uni. of Waterloo(until February 2010)• Santosh Shrestha, PhD UNSW• Long Seng To, BE(Hons) BA UNSWSenior Fellows• Anita W.Y. Ho-Baillie, BE PhD UNSW (returnedMarch 2010)• Dirk König, Dipl. Ing. (EE) Dr. rer. nat. (Ph) Chemnitz• Sergey Varlamov, PhD MoscowResearch Fellows• Patrick Campbell, BSc BE PhD UNSW• Matthew Edwards, BE (Hons 1) PhD UNSW• Shujuan Huang, BE ME Tsinghua Uni.,PhD Hiroshima Uni.• Henner Kampwerth, MSc Freiburg, PhD UNSW(since July 2010)• Emily Mitchell, BE PhD UNSW(since November 2010)• Ivan Perez-Wurfl, PhD Uni. of Colorado, Boulder• Thomas Soderstrom, PhD Uni. Of Neuchatel,SwitzerlandPost-Doctoral Fellows• Xiaojing Hao, BE, MSEE Northeastern Uni, China• Supriya Pillai, B. Tech (India), PhD UNSW• Sangwook Park (until March 2010)Adjunct FellowDidier Debuf, BE ME PhD UNSWVisiting Professors/Fellows• Dr. Robert Bardos• A/Prof Chee Mun Chong• Dr Dong-Jin Kim, Presidential Committee, KoreanGovernment (December 2010 – May 2011)• A/Prof Soon Kweon, Chungju National University,Korea (January 2010 – January 2011)• Darcy Wentworth, BSc BE Sydney U., MSc ColoradoSch. of Mines, MSc UNSW• Prof. Bingqing Zhou, Inner Mongolia NormalUniversity, China (until July 2010)Visiting Researchers• Yukiko Kamikawa(JSPS Fellow, AIST, Tsukuba, Japan)• Mallar Ray, Bengal Eng. Sci. University(3 weeks in May 2010)Professional Officers• Mark Griffin, BE UNSW• Yidan Huang, BSc Anhui• Hamid R. Mehrvarz, PhD UNSW• Tom Puzzer, BSc PhD UNSW• Bernhard VoglResearch Assistants• Nino Borojevic• Catherine Chan• John Durrant• Brett Hallam• Ziv Hameiri• Phillip Hamer• Jianshu Han• Jialiang Huang• Guangyao Jin• Anahita Karpour• Omaki Kazuo• Nicole Kuepper• Pei Hsuan Lu• Ly Mai, BE UNSW• Berhard Mitchell• Zi Quyang (until November 2010)• James Rudd• Lei “Adrian“ Shi• Adeline Sugianto• Yuguo Tao• Budi Tjahjono• Kalaimagal Valliappan• Ned Western• Guangqi Xu• Yao Yu• Xiaohan Yang• Tian ZhangTechnical Support Staff• Kian Fong Chin, BE QUT, MEngSc QUT,MEngSc UNSW• Bany Jaya, BE UNSW• Robert Largent, AS USA• Nicholas Shaw, BE UNSW, PhD UNSW• Lawrence Soria, AssocDipCompAppl W’gong• Zhu Shi Yang• Alan Yee, BE UTSCasual Staff• Suprabhat Basnet• Angela Jin• Alex Lee• Clara Mazzone• Xian Ni, MPhil UNSW• Jonathan Pye• Alexandra Ryan• Erdinc Saribatir• Andrew See• Lei Shi• Yuguo Tao• Dean UtianStudent Administration ManagerKimberly Edmunds, MEdLead UNSWFinancial OfficerJulie KwanAdministrative Staff• Danny Lin Chen, BSc UNSW• Mathew Cheung• Sue Edwards• Joyce Ho• Jill Lewis• Melissa TangPhD Students• Pasquale Aliberti• Yael Augarten, BSc UNSW• Shelley Bambrook• Jose Bilbao• Nino Borojevic, BE UNSW• Kristen Casalenuovo• Catherine Chan• Obulesu Chatakonda• Jessie Copper• Hongtao CuiARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORT11

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORT• Dawei Di• Bonneville Eggleston• Jack Fan• Peter Gress• Brett Hallam• Ziv Hameiri, BSc BGU• Phillip Hamer• Jianshu Han• Andy Hsieh• Jialiang Huang• Guangyao Jin (until September 2010)• Craig Johnson• Nicole Kuepper, BE UNSW• Taraka Kukunuri• Martha Lenio, B.A.Sc, Uni. of Waterloo (until March2010)• Sammy Lee• Hua Li• Wei Li• Xiaolei Liu• Doris Lu• Sarah Macleod• Ly Mai, BE UNSW• Bernhard Mitchell• Amir Nashed• Zi Ouyang (until September 2010)• Shyam Pasunurthi• Robert Patterson• Yuanjiang Pei• Binesh Puthen Veettil• Fang Qi• John Rodriguez• Yong Heng So• Adeline Sugianto, BE UNSW• Yuguo Tao• Anthony Teal• Budi Tjahjono, BE UNSW• Long Seng To• Lara Treiber• Chao-Yang Tsao• Zhenyu Wan• Stanley Wang, BE, MEngSc UNSW• Kee Soon Wang• Qian Wang• Johnson Wong (until September 2010)• Bo Xiao• Guangqi Xu• Chaowei Xue• Xiaohan Yang• Yang Yang• Yu Yao• Jae Sung Yun• Bo Zhang• Haixiang Zhang• Haoyang ZhangMasters Students• Phillip Cook• Telia Curtis• Mark Griffin• Do Soon Jung• Miga Jung• Anahita Karpour• Kyung Hun Kim• Taekyun Kim• Kenneth McLean• Gerald Parnis• Shahla Zamani Javid• Haixiang Zhang• Tian ZhangPracticum Students• Stephan Lehnen, RWTH, Aachen, Germany (untilMarch 2010)• Stephan Michard, RWTH Aachen, Germany(September 2009 – February 2010)• Manuel Morgano, Milano-Bicocca University, Italy(October 2010 – January 2011)• Chandraprasad Rmachandran, UNSW (July –December 2010)• Dolf Timmerman, University of Amsterdam, TheNetherlands (September – December 2010)• Hsuan-Wen Wang, National Central University,Taiwan (March 2009 – March 2010)Taste of Research Summer Scholars• Xin Rui An• Cong Chen• David Cordiner12

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORT4. RESEARCH“First-generation” wafer-basedtechnology (BP Solar SaturnModule, the photovoltaic productmanufactured in the highestvolume by BP in Europe, usingUNSW buried-contact technology).Figure 4.1.1Example of “second-generation”thin-film technology (modulefabricated on CSG Solar’s Germanproduction line, based on thin-filmsof polycrystalline silicon depositedonto glass, again UNSW-pioneeredtechnology).Figure 4.1.24.1 Introduction to ResearchPhotovoltaics, the direct conversion of sunlight toelectricity using solar cells, is recognised as one ofthe most promising options for a sustainable energyfuture with the photoovltaics industry poisedto become one of the world’s largest of the 21 stcentury. The ARC Photovoltaics Centre of Excellencecommenced in mid-2003, combining previousdisparate strands of work supported under a varietyof programs, into a coherent whole addressingthe key challenges facing photovoltaics, as wellas “spin-off” applications in microelectronics andoptoelectronics. The Centre was funded by the ARCuntil December 2010 and has since been fundedunder a variety of other schemes.The Centre’s photovoltaics research isdivided into three interlinked strandsaddressing near-term, medium-termand long-term needs, respectively. Thepresent photovoltaic market is dominatedby “first-generation” product based onsilicon wafers, either single-crystallineas in microelectronics (Fig. 4.1.1) or alower-grade multicrystalline wafer. Thismarket dominance is likely to continue forat least the next decade. First-generationproduction volume is growing rapidly,with the technological emphasis uponstreamlining manufacturing to reduce costswhile, at the same time, improving theenergy conversion efficiency of the product. Alsoimportant is the reduction of the thickness of thestarting silicon wafer without losing performance,to save on material use.The Centre’s first-generation research is focussedon these key issues. Building upon the success of“buried-contact” solar cell, the first of the modernhigh-efficiency cell technologies to be successfullycommercialised (Fig. 4.1.1), the Centre hasdeveloped several other high-efficiency processesin commercial production or close to this, based onCentre innovations in laser and ink-jet processing.Wafers are expensive and need quite carefulencapsulation, since brittle and also thermallymismatched to the glass coversheet, makingfirst-generation technology reasonably materialintensive.Several companies worldwide arecommercialising “second-generation” thin-film celltechnology based on depositing thin layers of thephotoactive material onto supporting substratesor superstrates, usually sheets of glass (Fig. 4.1.2).Although materials other than silicon are of interestfor these films, silicon avoids problems that canarise with these more complex compounds dueto stability, manufacturability, moisture sensitivity,toxicity and resource availability issues. CSG Solar,a partner in the Centre, has commercialised anapproach pioneered by Centre researchers thatis unique in that it is based on the use of thesame high quality silicon used for first-generationproduction, but deposited as a thin layer onto glass.As well as its collaborative activities with CSGSolar, the Centre currently maintains a largelyindependent program addressing alternativesolutions to those adopted by CSG Solar forproducing high-performance “silicon-on-glass”solar cells. The main emphasis of both is thedevelopment of lower-cost approaches (such asdeposition by evaporation rather than PECVD). TheCentre also commenced activities on carbon-based,organic solar cells during 2009 and a program on14

Efficiency and cost projections forfirst-, second- and third-generationphotovoltaic technology (wafers,thin-films and advanced thinfilms,respectively).Second-generation thin-filmtechnology has a different coststructure as evident from this figure.Production costs per unit area area lot lower, since glass or plasticsheets are a lot less expensivethan silicon wafers. However, likelyenergy-conversion efficiencies arelower (6-15%). Overall, this trade-offproduces costs/watt estimated asabout 2 times lower than thoseof the wafer product, in largeproduction volumes.ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTFigure 4.1.3other “earth abundant” materials, particularly CZTS(Cu 2ZnSnS 4), in 2010.At the present time, second-generation thin-filmsare entering the market in increasing quantities.Large-scale commercialisation of thin-film productleads to a completely different manufacturing coststructure compared to the wafer-based product.However, costs again are increasingly becomingdominated by material cost as production increases,for example, by the cost of the glass sheet on whichthe cells are deposited.More power from a given investment in materialcan be obtained by increasing energy-conversionefficiency. This leads to the possibility of a thirdgenerationof solar cell distinguished by the factthat it is both high-efficiency and thin-film. Toillustrate the cost leverage provided by efficiency,Fig. 4.1.3 shows the relative cost structures of thethree generations being studied by the Centre. Thisfigure plots efficiency against manufacturing cost,expressed in US$/square metre. First-generationtechnology has relatively high production cost perunit area and moderate likely efficiencies at themodule level (13-20%). The dotted lines in Fig. 4.1.3show the corresponding cost/watt, the marketmetric. Values below US$1/watt seem increasinglyfeasible by improving the efficiency while reducingmanufacturing cost, but this is the likely limit of thefirst-generation approach.The third-generation is specified as a thin-filmtechnology, which therefore has manufacturingcosts per unit area similar to second-generation,but is based on operating principles that donot constrain efficiency to the same limits asconventional cells (31% for non-concentratedsunlight for the latter). Unconstrainedthermodynamic limits for solar conversion are muchhigher (74% for non-concentrated light, giving anidea of the scope for improvement). If a reasonablefraction of this potential can be realised, Fig. 4.1.3suggests that third-generation costs could be lowerthan second-generation by another factor of 2 to 3.Of the third-generation options surveyed by Centreresearchers, “all-silicon” tandem cells based onbandgap-engineering using nanostructureswas selected as the most promising forrelatively near-term implementation (Fig. 4.1.4).This involves the engineering of a new class ofmixed-phase semiconductor material basedon partly-ordered silicon quantum-dots inan insulating amorphous matrix. The generalCZTS material system may also have somepotential here due to the range of bandgapsaccessible. Even tandem cells built on siliconwafers could fall into the targetted cost zonein Fig. 4.1.3. Photon up-conversion as a way of“supercharging” the performance of relativelystandard cells forms a second line of research.A third is the investigation of schemes forimplementing hot-carrier cells. Both of thelatter may prove to be particularly well suitedto organic solar cells.The fourth Centre strand of silicon photonicsdraws upon elements of all three of thephotovoltaic strands. A by-product of this workhas been the development of techniques basedon silicon light-emission for characterising bothcompleted devices, particularly solar cells, as well assilicon wafers at different stages of processing (Fig.4.1.5). Developing this approach to its full potentialhas formed an increasingly large part of the Centre’sphotonics program.Conceptual design of an all-silicontandem cell based on Si-SiO 2(orSi-Si 3N 4or Si-SiC) quantum dotsuperlattices. Two solar cells ofdifferent bandgap controlled byquantum dot size are stackedon top of a third cell made frombulk silicon.Figure 4.1.4Schematic representation of aPL imaging system. An externallight source is used to illuminatethe silicon wafer or solar cellhomogeneously. The luminescentemission (red arrows) from thesample is captured with a sensitiveCCD camera. This technology wascommercialised by Centre “spin-off”BT Imaging during 2008.Figure 4.1.515

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTCentre of Excellence locationin Sydney.Figure 4.2.14.2 Facilities andInfrastructureThe ARC Photovoltaics Centre of Excellence islocated at the Kensington campus of the Universityof New South Wales, about 6 km from the heart ofSydney and close to its world famous beachesincluding Bondi, Coogee and Maroubra(Fig. 4.2.1).Organisationally, the Centre of Excellenceis located within the School of Photovoltaicand Renewable Energy Engineering (SPREE)within the Faculty of Engineering. The Centreof Excellence has a large range of laboratoryfacilities (Fig. 4.2.2). These include the BulkSilicon Research Laboratories, the DeviceCharacterisation Laboratory, the OptoelectronicResearch Laboratories, the Thin-Film CellLaboratory, the Industry CollaborativeLaboratory, Inkjet Processing Laboratory andOrganic Photovoltaic (OPV) Laboratory. Offcampus the Centre has a Thin-Film Cleanroomfacility at Botany, 5km south-west of the maincampus. Another important resource is theSemiconductor Nanofabrication Facility (SNF)jointly operated by the Faculty of Science andthe Faculty of Engineering.Additional equipment commonly used for solarcell work is found at the new MicroAnalyticalResearch Centre (MARC) and elsewhere on theUniversity campus. Included in this categoryare TEM/SEM electron microscopes, focused ionbeam (FIB) specimen preparation equipment,X-ray diffraction, Raman spectroscopy, AFM andsurface analysis equipment. TEM, ellipsometry anda femtosecond time resolved photoluminescencemeasurement system are also regularly accessed atSydney University.During 2010, the development and acquisition oflaboratory equipment and infrastructure continuedwith specific details about significant additionsfound under the laboratory headings which follow.The Centre of Excellence has three computernetworks. Both Kensington and Bay St Botany eachhave a research/admin network. The third networkis for students enrolled in the UndergraduateDegree Courses in Photovoltaics and RenewableEnergy. In 2010, the Kensington Research andAdministrative network grew by 20% to consist of5 file servers, 1 terminal server, 1 intranet server,2 Internet web-servers (hosted on Central ITinfrastructure), 2 licence servers (local and Central)and over 170 client workstations. At Kensington,an additional 31 computers are dedicated tothe computer control of laboratory and otherequipment. The Centre also has a 3 server computercontrolled SCADA and PLC network for equipmentcontrol and monitoring the research laboratoriesand related infrastructure. Bay St Botany has 9dedicated equipment controllers and the research/admin network supports 11 workstations.The computer resources are used for measurement,modelling/simulations, equipment control,16

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTLayout of laboratoriesand other facilitieswithin the ElectricalEngineeringBuilding, Kensington.Figure 4.2.2document control, laboratory design support,Internet access, general administrative purposesand maintaining the Centre’s presence on theInternet.Each postgraduate research student is allocateda dedicated personal computer and has access toshared computer resources.The Centre upgraded Linux (Beowulf) 64bitcluster has an estimated computational power of2 Teraflops to support the demanding Density-Functional-Hartree-Fock and molecular dynamicalcomputations. A second 128 core cluster hasbeen installed for complementary investigationsof energy efficiency and renewable energy viacomputational fluid dynamics techniques.The Student Computer network comprises oftwo computer laboratories of total 78 m 2 with 5servers, a UNSW custom CSE computer based IPQrouter and 34 workstations. Students enrolledin the Undergraduate Degree Courses use thesecomputers for computer-related coursework andInternet access. There is also an Internet capableweb-server that gathers and displays data collectedfrom solar arrays on the roof of the building. Thesedata can be viewed using a web browser and can bemade available for Internet access.Students also have access to over 200additional workstations made available throughcollaboration with the School of Computer Scienceand Engineering.The Laboratory Development and Operations Teamdevelops and maintains core Centre and laboratoryfacilities. During 2010, the team, under theleadership of Mark Silver, comprised of an additional5 equivalent full-time and 11 casual employees,including: electrical engineers, a computer/network manager, electronic/computer/laboratorytechnicians and administrative staff.Bulk Silicon Research LaboratoriesThe Centre houses the largest and mostsophisticated bulk silicon solar cell research facilityin Australia, incorporating both the High Efficiencyand Buried Contact Cell Laboratories. Totallaboratory processing space of over 580m 2 (including the Industry Collaborative,InkJet Processing Development, DeviceCharacterisation and OptoelectronicResearch Labs) is located in the ElectricalEngineering Building and is servicedwith filtered and conditioned air,appropriate cooling water, processinggases, de-ionized water supply, chemicalfume cupboards and local exhausts.There is an additional 819 m 2 area forthe accommodation of staff, researchstudents, school office and laboratorysupport facilities. Another 480 m 2 of combined roofspace accommodates fixed PV arrays and 361m 2of accessible outdoor experimental space. Off site,areas totalling 700 m 2 are used for the storage ofchemicals and equipment spare parts and a new100 m 2 area for a nickel sintering belt furnace facilityat 78 Bay St Botany.The laboratories are furnished with a range ofprocessing and characterisation equipmentincluding 23 tube diffusion furnaces, 6 vacuumevaporation deposition systems, a laser-scribingmachine, 2 laser doping machines, rapid thermalannealer, four-point sheet-resistivityprobe, quartz tube washer, silver/nickel and copper plating facility, visiblewavelength microscopes, 3 wafer maskaligners, spin-on diffusion system,photoresist and dopant spinners, electronbeam deposition system, metallizationbelt furnace, nickel sinter belt furnace,manual and automatic screen printersand a laboratory system control and dataacquisition monitoring system.The laser scribe tool, shown in Figure 4.2.3,has a 20 watt Nd:YAG laser for infraredoperation (1064 nm) and an optionalfrequency doubler for green operation(532 nm). The work stage is CNC controlled allowing1 micron positional accuracy and table speedsapproaching 25 cm/second across an area of 15cm by 15 cm. The tool is used primarily for BuriedCNC Laser Scribe Tool.Figure 4.2.3Wafer Mapping Tool.Figure 4.2.417

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTRemote plasma PECVD machine.Figure 4.2.7Flash Cell Tester.Figure 4.2.5Optical characterisation bench.Figure 4.2.6Contact solar cell fabrication, cutting 35-micronwide laser grooves as deep as 100 microns intosilicon wafers and scribing wafers in preparation forcleaving. It can also be used to cut other suitablematerials, such as stainless steel.Device Characterisation LaboratoryThis laboratory is located on the lower ground floorof the Electrical Engineering Building. Associatedwith it is the Optoelectronic Research Laboratories,reception area, seminar room, offices for Centre staffinteracting with the public and industry, includingthe Business & Technology Manager, FinanceOfficer, External Relations and Design AssistanceDivision Manager.The Device Characterisation Laboratory housescharacterisation equipment including “DarkStar”, the Centre’s station for temperaturecontrolled illuminated and dark current-voltagemeasurements, the Centre’s Fourier-transforminfrared spectroscopy system (FTIR), frequencydependent impedance analyser, ellipsometer,Sinton photoconductance lifetime equipment,wafer probing station, open circuit voltage versusillumination measurement system (Suns-V oc), 4 pointresistivity probe, spectral response system andspectrophotometer with integrating sphere.In 2008, a multi function wafer mapping tool, shownin Figure 4.2.4, was installed. This unit providesstate of the art capability for the measurementof carrier lifetime, bulk resistivity, emitter sheetresistance and Light Beam Induced Current (LBIC)on wafer samples.Additions in 2009 included a high speedcommercial flash cell tester with 156 mm squarewafer capability as shown in Fig 4.2.5, a replacementUV/VIS/NIR spectrophotometer and a commercialLuminescence Inspection System (LIS) from spin offcompany BT Imaging.Optoelectronic Research LaboratoriesThis facility has six optical benches and severalvisible and near-infrared semiconductor diodelasers along with other optical and electricalinstrumentation. The facility is used forphotoluminescence (PL) and electroluminescencemeasurements in the visible and infrared spectralrange up to wavelengths of 2500nm, luminescenceexperiments with simultaneous two-colourillumination, quasi steady state photoluminescencelifetime measurements and Sinton lifetime testingwith the conventional flash-light replaced by a highpowerlight emitting diode array.This facility also houses the Centre’s world-leadingfirst photoluminescence imaging system. Otherequipment includes a silicon CCD camera (forsensitive PL measurements), spectroscopic PLsystems, PL lifetime measurement unit, and LBICmeasurement system. Areas separate from theDevice Characterisation Laboratory were necessaryin order to meet stringent standards for safe laseruse. It shares cryogenic cooling equipment with theDevice Characterisation Laboratory.Thin-Film Cell LaboratoryThis 40 m 2 laboratory is equipped with a range ofequipment for thin-film deposition and patterning,including a plasma-enhanced chemical vapourdeposition (PECVD) system, a sputtering system,larger area plasma etcher, a reactive ion etcher (RIE),a resistively heated vacuum evaporator, heliumleak detector and an optical microscope withdigital image acquisition system. Also used by thelaboratory is an electron-beam vacuum evaporatorfor silicon which is physically located within theBulk Silicon Research Laboratory. This Si evaporatoris also equipped with an ionizer unit and a sampleheater, enabling fast-rate Si homoepitaxy attemperatures of about 500-600°C by means of ionassisteddeposition (IAD). The IAD is also equippedwith a residual gas analyser to permit real timeprocess monitoring. Other equipment of use in thinfilmprojects is located within the SemiconductorNanofabrication Facility.18

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTThermal and e-beam Evaporator.Figure 4.2.8The PECVD system, shown in Fig. 4.2.7, has a 40 x20 cm 2 process platen and can handle large-areasilicon wafers as well as smaller pieces. Two typesof plasma excitation (remote microwave and directRF) are available. The machine is used for thelow-temperature deposition of thin dielectric films(silicon nitride, silicon dioxide, silicon oxy-nitrides)and of amorphous silicon. The dual-cylinder,remote microwave plasma source producesexcellent-quality silicon nitride and silicon dioxidefilms, with precise control over the stoichiometryat temperatures up to 500°C. Amorphous andmicrocrystalline silicon films can also be depositedin the system.In 2008, the thermal evaporator was upgraded toalso support e-beam evaporation. This capabilitygreatly expands the range of materials which can bedeposited and is of great interest to 3 rd generationresearchers. In 2009 a new 4 point probing stationwas added to the laboratory. In 2010 the computersupport system for the optical microscope wasupgraded to improve the useability of the digitalcamera fitted.Industry Collaborative LaboratoryThis 120 m 2 laboratory houses equipment neededfor many of the industry-collaborative researchactivities including the Buried-Contact Solar Cellgroup. The laboratory equipment includes a beltfurnace; a state of the art laser micromachiningtool; a new PECVD deposition system (located inthe adjacent Thin Film Solar Cell Laboratory); aTiO 2spray deposition station; a high temperaturesemiconductor muffle furnace, manual screenprinter and a fully automatic production scalescreen printer. In 2008, a new 532nm green laserwas added and customised in-house to increase thecapability for processing laser doped solar cells.In 2009, new additions included a fast metallizationfiring furnace capable of processing 156 mm squarewafers, a light induced plating system, spin ondoper and mirror steered scanning laser for laserdoping. In 2010 a new wafer spinner was addedto the laboratory. Off site at 78 Bay St Botany, anitrogen environment belt furnace capable of nickelsintering 156 mm square wafers and larger on its 25inch wide belt was installed.Inkjet Processing Development LabThis laboratory houses the Centre’s inkjet printingdevelopment systems. The laboratory is used todevelop solar processing “inks” (chemical solutions)and for printing them onto a solar wafer undercomputer control. The aim is to develop low costprocessing techniques for creating fine structures insolar cells. It is anticipated that the inkjet printing iscapable of replacing processes such as laser scribingand photolithography for forming fine patternsfor contacting but at a cost of at least 10 timescheaper. Equipment includes two ink jet materialdeposition printing systems, capable of depositinga wide range of materials onto different substrates,a surface tension meter and viscometer. In 2009 acommercial aerosol deposition system capable ofvery fine scale patterning was purchased.Semiconductor Nanofabrication FacilityThe Centre also owns equipment within, and hasaccess to, the Semiconductor NanofabricationFacility (SNF) at the University. This is a joint facilityshared by the Faculties of Science and Engineeringand houses a microelectronics laboratory and ananofabrication laboratory for e-beam lithography.The SNF provides an Australian capability for thefabrication of advanced nanoscale semiconductordevices and their integration with microelectronics.SNF research projects form an integrated effort tofabricate innovative semiconductor nanostructuresusing the latest techniques of electron beampatterning and scanning probe manipulation.A major applied objective of the facility is thedevelopment of a prototype silicon nuclear spinquantum computer. The capabilities of this facilitywere expanded to house the, NCRIS funded,Australian National Nanofabrication Facility. In2010 the facility added a new large area e-beamlithography system and new e-beam evaporator fornon MOS-compatible metals deposition.Inkjet development systems.Figure 4.2.919

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTBay Street Clean Room Sputter Machine.Figure 4.2.10UHV Electron Beam Evaporator.Figure 4.2.11Thin-Film Clean Room facility inBay Street, BotanyIn 2003, the Centre added a 120 m 2 cleanroomfacility in Bay Street, Botany to its infrastructure,greatly improving its experimental capabilities inthe area of thin-films. This cleanroom is equippedwith several fume cupboards, two tube furnaces,an electron-beam vacuum evaporator, a thermalvacuum evaporator, a glass washing machine, arapid thermal processing (RTP) machine and a5-chamber cluster tool. The cluster tool presentlyfeatures four plasma-enhanced chemical vapourdeposition (PECVD) chambers and one lampheatedvacuum annealing chamber. The PECVDchambers enable the low-temperature depositionof dielectric films (silicon oxide, silicon nitride,etc) and amorphous silicon films (either n- orp-doped or undoped). Furthermore, samplescan be hydrogenated by PECVD using hydrogenplasma at substrate temperatures of up to 480°C.During 2004, the Centre purchased a low-pressurechemical vapour deposition (LPCVD) system, aninfrared NdYAG laser scriber and a box furnace forsample annealing. The LPCVD system is capable ofdepositing doped crystalline silicon on glass andwith its additional remote plasma source is currentlyengaged in hydrogenation work. The NdYAGlaser, located outside the clean room, is used forscribing silicon films and other suitable metal anddielectric materials.In 2006, a state of the art multi-target sputtermachine was installed. In 2008 this vacuum toolwas upgraded from four to five separate targets.Each target is able to be operated independentlyof one and other, allowing users to cosputter athin film from more than one target and depositmultilayers without breaking vacuum. Three powersupplies are available with substrate biasing. Thecustom made system can handle substrates up to150mm x 150mm. Excellent film purity is assuredas the system incorporates a load-lock. Computercontrol can be used for most operations, includingsubstrate heating, allowing precise multilayers to bedeposited repeatedly.In 2009, an optically assisted IV measurementsystem was installed for the determination of statichot carrier distributions. The LPCVD computercontrol system was also upgraded to provideenhanced throughput. A line beam diode lasersystem was also installed, outside the clean room,for work on low thermal budget defect annealing.In 2010, the first stage of a multi-chambersputter system has been ordered to allow our 3 rdGeneration Group to carry out significantly differentfunctions on a sample without exposing it to theatmosphere between each process. Processesinclude sputter-depositing different dopants inseparate chambers, vacuum annealing, also PECVDlayers. This purchase will be “Stage I” in that onechamber will be purchased; a loadlock chamberand other peripherals have been configured fromdecommissioned stock. Stage II will involve thepurchase of a second processing chamber and atransfer/loadlock chamber.The addition of computer control to ourhydrogenation tool provides more reliableprocess repeatability and accuracy of set processparameters. The processor inputs all processparameters for each “layer” defining each stepbefore starting the process. This upgrade alsofrees processors’ time in that they need not be inattendance to change settings when beginningeach process layer.Organic Photovoltaic Laboratory (OPV)In 2010 fit out of a new OPV laboratory in theChemical Sciences Building commenced andOPV equipment currently located in other Centrelaboratories is being integrated into the new facility.A recent arrival has been the inert atmosphere, fourglove boxes with integrated spin coater. The newlaboratory includes the Centre’s first fume cupboardwith integral hydrofluoric acid scrubber as well assupport space for vacuum evaporation, annealingand sample storage.During 2007, a new computer controlled multipocket, multi source UHV electron beam evaporatorsystem was delivered. This equipment provides thecapability for high “industrial” rate silicon and otherthin film deposition.20

4.3 First GenerationWafer Based ProjectsResearch TeamUniversity Staff• S. R. Wenham (Centre Director)• M. A. Green (ExecutiveResearch Director)• T. Trupke (Deputy Director –Photoluminescence Imaging)• A. Lennon (Group Leader – Inkjet andPlating Technologies)• A. Ho-Baillie (Group leader - HighEfficiency Cells)Postdoctoral Fellows• H. Mehrvarz (High efficiency Cells)• M. Edwards (Tech Transfer)• B. Tjahjono (Core Research)• L. Mai (Tech Transfer)Postgraduate Research Studentsand Research Assistants• M. Lenio (Research Assistant andPhD student)• Z. Hamieri (PhD student)• N. Kuepper (Research Assistant andPhD student)• S. Wang (PhD student)• N. Borojevic (Research Assistant andPhD student)• S. Zamani Javid (Research Assistant andMasters student)• P. Hamer (Research Assistant andPhD student)• A. Sugianto (Research Assistant andPhD student)• B. Hallam (Research Assistant andPhD student)• A. Karpour (Research Assistant andPhD student)• D. Wang (PhD student)• A. Han (Research Assistant)• C. Chan (Research Assistant andPhD student)• G. Xu (Research Assistant andPhD student)• K. Valliappan (Research Assistant)• L. Zhang (Research Assistant)• Y. Yao (PhD student and ResearchAssistant)• L. Gu (Research Assistant)• N. Western (Research Assistant)• Mattias Klaus Juhl (Research Assistant)• Xi Zhu (Research Assistant)• X. Bai (Research Assistant andundergraduate thesis)• Y. Yeung (Research Assistant)• Bo Xiao (PhD student)• J. Roderiguez (PhD student)• B. Mitchell (PhD Student)• Li Hua (PhD student)• D. Lu (PhD student)• Y. Yang (PhD student)Technology Transfer TeamManagers• D. Jordan (Team leader)• C.M. Chong (Deputy Team Leader– Devices)• M. Edwards (Deputy Team Leader –Program Manager)• S. Wenham (Technical Director)Team Members• A. Sugianto• N. Kuepper• B. Hallam• C. Chan• A. Lennon• B. Tjahjono• L. Mai• C. M. Chong• X. Bai• P. Hamer• E. MitchellNewSouth Innovations• N. Simpson• A. PribARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORT4.3.1 High Efficiency Silicon CellsAlternate Rear Reflectors for SiliconSolar CellsDuring 2010, the high efficiency group conductedstudies into improved planar rear reflectors and alsointo a scattering rear reflector for silicon solar cellsas shown in Fig. 4.3.1.1. The aim of this work is toincrease reflectance over a broad angular range andto enhance light trapping properties, respectively.The effect of varying refractive index of the middlelayer of a dielectric stack (DS) of SiO 2/SiNx/SiOSi wafer 2has been studied. The thicknesses of each layer inthe DS are kept at ¼ wavelength Dielectric thick stacks while therefractive index of the middle layer is varied fromMetalSi waferDielectric stacksMetalSi waferMetalPassivation layerPadding layerMetal(a) Planar rear reflector madeup of dielectric stacks and(b) scattering rear reflectormade up of silver plasmonicnanoparticles sandwichedbetween passivating andpadding dielectric layers.Figure 4.3.1.1Si waferPassivationPadding laye21

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTRear reflectance10.950.90.850.80.75Rb and IQE under perpendicular incident lightR (n=1.5)R (n=2.0)R (n=2.6)R (n=3.0)R (n=3.2)IQE (n=1.5)IQE (n=2.0)IQE (n=2.6)IQE (n=3.0)IQE (n=3.2)1.210.80.60.40.2IQERear reflectance0.990.970.950.930.910.890.87Rb and IQE under Lambertian light distributionR (SiO2 / Al)R (SiO2 / Ag)R (optimised DS / Al)R (optimised DS / Ag)IQE (SiO2 / Al)IQE (SiO2 / Ag)IQE (optimised DS / Al)IQE (optimised DS / Ag)1.210.80.60.40.2IQE0.70900 950 1000 1050 1100 1150 1200 1250Wavelength (nm)0.850900 950 1000 1050 1100 1150 1200 1250Wavelength (nm)R (left axis) and IQE (right axis) under normallyincident light for a planar silicon solar cell withdielectric stack plus metal rear reflector. Theeffect of varying the refractive index of themiddle layer of the dielectric stack (DS) is shown.Figure 4.3.1.2R and IQE under a Lambertian distribution of light fora silicon solar cell with various types of rear reflectorsincluding the standard SiO 2/metal and the optimiseddielectric stack (DS)/metal rear reflectors.Figure 4.3.1.31.00E+029.00E+01Measured EQE of un-textured solar cell(4.3.1.1)8.00E+01EQE (%)7.00E+016.00E+015.00E+014.00E+01Original solar cellEQE enhancement(4.3.1.2)3.00E+01Solar cell only with passivation layer2.00E+01Solar cell with passivation layer and Ag plasmon1.00E+010.00E+00300 400 500 600 700 800 900 1000 1100 1200wavelength (nm)EQE of a planar solar cell withstandard rear reflector (SiO 2/Al)structure and after the applicationof Ag nanoparticles (prior tothe application of padding andmetal layers).Figure 4.3.1.41.5 to 3.2. Such stacks with a rear contacting metallayer applied onto a planar cell result in internalreflectance (R) and internal quantum efficiencies(IQE) shown in Fig. 4.3.1.2. Figure 4.3.1.3 comparesthe R and IQE of standard SiO 2/metal reflectors tooptimised DS/metal reflectors when applied to cellswith textured front surface.(4.3.1.3)Current density Jsc (mA/cm 2 ) was calculatedby integrating the product of IQE and photonflux, from 300nm to 1300 nm (assuming zerofront surface reflection). The standard AM1.5Gspectrum has been used. Table 4.3.1.1 shows theimprovement of Jsc as a result of applying DS/metalrear reflectors (over the standard SiO 2/Al reflector)on both planar and textured solar cells.SunlightSpectrally sensitive mirrorsSolar cells ofdifferent band gapsMulti-junction cell concept usingspectral splitting.Figure 4.3.1.5The calculation of R and the IQE for Figures 4.3.1.2and 4.3.1.3 is carried out as follows. All photogeneratedcarries are assumed to be collectedby the solar cell (which is roughly the case inthe Centre’s PERL cells). A Bis the absorptionin the bulk region, A Trefers to the cell’s totalabsorption (A T=A B+A R, with A Rdefined as rearreflector absorption). IQE equals the ratio ofA Band A Tas indicated in Eq. 4.3.1.1. Formulaefor A Band A Tare given elsewhere [4.3.1.1].R is the wavelength dependent internalrear reflectance. It can be simply calculatedfor perpendicularly incident light, or be aweighted value of a Lambertian angular lightdistribution as shown in Eq. 4.3.1.3. T + is thefraction of light transmitted to the rear surfaceand T - is the fraction of light reflected from therear side transmitted to the front surface. Anapproximation for T + T - is used to simplified formulain Eq. 4.3.1.2, where a is the absorption coefficientwhich varies with wavelength and W is the waferthickness.(4.3.1.4)Another promising scattering rear reflectorstructure involves plasmonic scattering from Agnanoparticles. The idea is to scatter the unabsorbedlong wavelength light back into the silicon bulk,thereby randomising the reflected light directionsto enhance light trapping. This can be particularlyimportant from the first double-pass of light acrossa cell with pyramidal texture since a large fractionis lost when internally striking the surface of apyramid of opposite slope to that coupling it in.Light travelling past the Ag nanoparticles will bereflected back by the padding layer/metal interfaceas shown in Fig.4.3.1.1 (b), and be scattered forthe second time. Fig. 4.3.1.4 shows the preliminaryresults of applying such structure on a silicon solarcell prior to the application of the padding layer.The application of such a nanoparticle reflectorinvolves the removal of existing back metal layersand the thinning of the rear passivating SiO 2layer followed by the growth of Ag plasmonic22

Jsc improvement afterapplying dielectric stack (DS)/metal rear reflectors comparedto the standard SiO 2/Alreflector for both planar andtextured solar cells.Refractive index of middle layer in DSJsc improvement n=2.0 n=2.6 n=3.0 n=3.2Under normal incident light 1.08% 1.97% 2.50% 2.65%Under Lambertian incident light -0.03% 0.06% 0.15% 0.15%ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTTable 4.3.1.140.7% III-V cell stack on Ge showingthe I-V curves of constituentcells [4.3.1.2].Figure 4.3.1.6particles onto the thinned passivating layer. Whilethe EQE enhancement as shown in Fig. 4.3.1.4is encouraging, it is anticipated that more EQEgain is possible by adding the padding and metalreflector layers.Spectrum SplittingDuring 2010, the group continued its workon improving the efficiency of solar cells bysubdividing the broad solar spectrum into smallerenergy ranges and to convert each range witha cell of appropriately matched bandgap (seeFigure 4.3.1.5).The Centre is involved in a collaborative researchproject involving Solar Systems and Spectrolab infully exploring the potential of spectrum splittingin combination with the power tower concept,for higher efficiency conversion of concentratedsolar power (CSP). A working prototype CSP powertower system of 140kW rating at Waterbridge,Victoria has demonstrated the highest sustainedconversion efficiency to date for any solar to electricconversion system without yet the benefit of suchsplitting [4.3.1.3]. We plan to demonstrate theadditional potential by the construction of a small“power-cube” prototype capable of independentconfirmation of energy conversion efficiency above40%, including all optical losses.The work is based on combining the only twocommercial solar cell technologies for concentratedsunlight, silicon and III-V semiconductor stacks onGe substrates, with the latter shown in Fig. 4.3.1.6.The Ge substrate provides a good crystallographictemplate for subsequent layers but Ge has too low abandgap to be a good opto-electronic match to thetwo cells on top. From the current-voltage curvesshown, the Ge cell can be seen to be generatingmuch more current than the other two cells. Theextra current is essentially wasted as heat in theGe cell with the situation becoming worse as theairmass (AM) increases from the standard AM1.5test condition shown (particularly in the morningsand afternoons).The initial idea was to reflect some of the excessphotons reaching the Ge cell onto a separate andrelatively inexpensive silicon cell array using aninnovative “power cube” receiver as in Fig. 4.3.1.7(a).For the cell stack shown in Fig. 4.3.1.6, up to 0.085A/W of photon current could be reflected onto asilicon cell where it would be converted at about0.7V under concentration. This would add an extra6% absolute to the combined cell performancetaking it from 40.7% efficiency to above 46%.Additional advantages would be that theessentially lossless dicroic reflector used to give thiswavelength selective reflection could also be usedto control the angular emission from the III-V cells,increasing their voltage output (by 2-20 mV/cell forradiative efficiencies above 10%). Moreover, bycontrolling the precise wavelengths reflectedduring the day, spectral mismatch losses couldbe greatly reduced compared to the use of aIII-V array alone. An additional advantage is thereduced amount of waste heat to be removedfrom the receiver and the larger area availablefor its extraction.More detailed analysis of the receiver hasshown that the “demi-cube” approach of Fig.4.3.1.7(b) is even more attractive. By integratingthe array and reflector and switching theposition of the two arrays, the design of thereflector is greatly simplified. Low UV and far-IRabsorption in the filter to avoid excessive heatingR1.00(a) Original “Power Cube” receiverconcept. Light in approximately the890-1100nm wavelength range isreflected by a band-pass dielectricreflector onto a silicon array ratherthan being dissipated as heat inthe Ge cell; (b) Improved “Demi-Cube” concept. Thermal design isimproved by integrating one arrayand the reflector. By switching theposition of the silicon and III-V array,the filter design becomes far lesscritical as shown in Fig. 4.3.1.8.Figure 4.3.1.7Fig. 4.3.1.7 (b)Fig. 4.3.1.7 (a)900Simplification of filter design withnew configuration.Figure 4.3.1.8110023

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTFilter WavelengthCut-offs (nm)Jsc III-V Top Cell(Am-2)Jsc III-V MiddleCell (Am-2)Jsc III-V BottomCell (Am-2)Without Filter 152.41 147.27 199.24 38.8Design 1 932-1070 152.41 146.50 146.46 4.09Design 2 930-1068 152.31 146.89 144.96 6.41Design 3 930-1066 151.99 146.89 146.12 4.54CurrentMismatch (%)Effect of the use of dielectricreflectors on reducing currentmismatch within the III-V device.Table 4.3.1.3Seasonsin LaParguera24AnglebetweenIII-V & Sicells (°)Effect of varying angle betweenSi and III-V arrays to maximise celloutput from the various seasonalspectral contents shown inFig. 4.3.1.9(b).Table 4.3.1.2Seasonal averaged direct spectralcontents for (a) Canberra and (b)La Parguera.Figure 4.3.1.9Fixed Demi-Cube StructureFiltercut-offs(nm)III-V cellefficiency(%)Si cellefficiency(%)Total cellefficiency(%)Filtercut-offs(nm)Variable Demi-Cube StructureAnglebetweenIII-V & Sicells (°)III-V cellefficiency(%)at 500 suns (or higher) sunlight concentration is nolonger essential. Also, the requirements on reflectordesign are considerably relaxed due to the greaterease of designing for high reflection rather than forhigh transmission (Fig. 4.3.1.8), with design nowsimplifying to that of a noncritical band pass filter.The extra cell area required is not a factor since thelarger array is formed from lowcost silicon cells.Another advantage of the“demi-cube” concept is that theangle between the Si and III-Varrays can be varied to takeadvantage of the non-idealreflection and transmissionproperties of the dielectricreflector. As the incidenceangle on the dielectric reflectorchanges, the reflector cut-offwavelengths vary and so dothe slopes in the transitionregion between reflectionand transmission. This allowsthe reflector to capture thespectrum with the mostsuitable wavelength range forthe Si and III-V arrays as thespectral content of sunlightvaries daily and seasonally, asshown in Fig. 4.3.1.9.Table 4.3.1.2 shows the effect of varying anglebetween Si and III-V arrays to maximise output fromvarious spectral contents using the spectra of Fig.4.3.1.9 as an example. Note the spectra shown inFigure 4.3.1.9 and used for Table 4.3.1.2 are seasonalaverages only. Hence the advantage of variableSi cellefficiency(%)Summer 33.9 874-1060 41.0 4.1 45.1 874-1060 33.9 41.0 4.1 45.1Autumn 33.9 874-1060 40.2 4.1 44.3 894-1080 35.3 40.8 4.2 45.0Winter 33.9 874-1060 41.3 4.1 45.4 874-1060 33.9 41.3 4.1 45.4Spring 33.9 874-1060 40.3 4.1 44.4 894-1080 35.3 40.9 4.2 45.1Total cellefficiency(%)demi-cube structure will beeven greater if used within aday as the changes in dailyspectrum content will begreater than those observedin Figure 4.3.1.9. Table 4.3.1.3shows the effect of the useof dielectric reflectors onreducing current mismatchwithin the III-V device.References:4.3.1.1 M.A. Green, “Lambertian light trapping in texturedsolar cells and light-emitting diodes: analyticalsolutions”, Progress in Photovoltaics: Research andApplications 10(4), pp. 235-241, 2002.4.3.1.2. Richard King, 51st Electronic Materials Conference,June, 2009.4.3.1.3 J. Lasich et al., “World’s First Demonstration of a140kWp Heliostat Concentrator PV (HCPV) system”,34 th IEEE Photovoltaic Specialists Conference,Philadelphia, 7-12 June 2009.4.3.2 Industry Collaborative Research andCommercialisation4.3.2.1 IntroductionThe Centre’s Technology Transfer team (TTT) wasestablished in 2008 to accelerate the developmentof the Centre’s commercially significanttechnologies and to carry out technology transfersto industry. This includes assisting companies toestablish and optimise the large scale productionof UNSW photovoltaic technology. With the highdemand for UNSW technology and new licences,the TTT grew significantly during 2010 with theinclusion of Phil Hamer, Xue Bai and Emily Mitchell.The already strong academic record of the TTT wasalso further enhanced with an additional two teammembers having been awarded University Medalsfor being placed first in their respective universityprograms. This included Emily Mitchell who receivedthe University Medal for first place in PhotovoltaicEngineering at UNSW in 2007 prior to completingher PhD at the Fraunhofer Institute in Germany andXue Bai who received the University Medal for beingplaced 1 st in Photovoltaic Engineering at UNSW in

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTTTT photo June 2010 with Matt Edwards overseas atthe time and new comers Emily Mitchell, Phil Hamerand Xue Bai about to join the team.Figure 4.3.2.12010. This takes the number of University Medallistswithin the TTT to six at the end of 2010, but withanother university medallist Nicole Kuepper beingtemporarily on leave from the TTT while completingher PhD. Figure 4.3.1 shows the TTT in June 2010with members from left to right being D. Jordan, A.Sugianto, S. Wenham, A. Lennon, B. Hallam, L. Mai, B.Tjahjono and N. Kuepper.4.3.2.2 Industry Collaborators andLicensees of TechnologyThe Centre has a large number of collaboratorsand licensees including many cell manufacturersworking with the Centre to improve or develop newsolar cell technology for commercialisation. Someof these fund the Centre quite generously and areseen as being partners to NewSouth Innovations(NSi), the commercial arm of the University thatmanages the entire intellectual property portfoliofor the Centre. Interest from companies wantingto become commercial partners with the Centre orlicense technology grew rapidly during the last twoyears despite the international financial downturnin late 2008. This appears to be due in part to theperceived world leadership of UNSW in this area,but also due to the growing long-term importanceplaced on photovoltaics by a world struggling todeal with climate change, diminishing resourcesand increasing energy requirements. Collaboratorswith the Centre in the first generation photovoltaicsarea come predominantly from China, Germany,the United States, Australia, South Korea andTaiwan and include Suntech-Power, Roth and Rau,Guodian Solar, Centrotherm, Tianwei, Optomec,Global Sunrise Energy, BP Solar, Sunergy, CSG Solar,E-ton Solar, Spectraphysics, Schott Solar, HyundiHeavy Industries, China Light, Corum Solar, Advent,Shinsung, JA Solar and Yunnan Tianda. Several ofthese companies are amongst the world’s largestsolar cell and equipment manufacturers.4.3.2.3 Commercially RelevantTechnologies4.3.2.3.1 IntroductionWith screen-printed solar cells continuingto dominate commercial manufacturingwith well over 50% market share, thebroad aim of this work has been todevelop, in conjunction with severalindustry partners, the next generationof screen-printed solar cell. In particular,the fundamental limitations of theconventional screen-printed solar cellthat have limited its performance for thelast 30 years have been identified, and innovativeapproaches to redesigning the emitter and frontmetal contact have been devised, developed andanalysed in this work. In addition to overcomingthe current and voltage limitations imposed bythe design shown below, a further aim of this workhas been to retain compatibility with existingequipment and infrastructure currently used by ourindustry partners for the manufacture of screenprintedsolar cells.Despite the dominance of this technology,this solar cell design shown in Fig. 4.3.2.2 hassignificant performance limitations that limit thecell efficiencies to well below those achievablein research laboratories around the world. Inparticular, the front surface screen-printedmetallisation necessitates a heavily diffused emitterto achieve low contact resistance and also toachieve adequate lateral conductivity in the emittersince the metal lines need to be widely spacedcompared to laboratory cells to avoid excessiveshading losses. Such cells therefore typically haveemitters with sheet resistivities in the range of 50-60ohms per square, which inevitably give significantlydegraded response to short wavelength light.Good work in recent years by partners Dupontand Ferro has improved paste formulations,allowing ohmic contacts to be formed to morelightly doped emitters in the vicinity of 60 ohms/square while simultaneously lowering the bulkresistivity of the fired paste. To further raise this150-200 mp +p-typemetaln ++3 m m patterned metal contactStandard screen-printed solar cell.Figure 4.3.2.2phosphorusbulk of waferrear metal contact25

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTScreen-printed fingers runningperpendicular to the heavilydiffused semiconductor fingerswhere electrical contact ismade. A dielectric/AR coatingpassivates the top surface andisolates the metal from thelightly diffused top surface.Figure 4.3.2.3sheet resistivity to above 100 ohms per square, asrequired for near unity internal quantum efficienciesfor short wavelength light, serious resistive lossesare introduced, both in the emitter and throughthe contact resistance at the metal to n-typesilicon interface.Furthermore, the conventional design for screenprintedsolar cells has quite poor surface passivationin both the metallised and non-metallisedregions. Even if good ohmic contacts could bemade to more lightly doped emitters, the largemetal/silicon interface area would significantlylimit the voltages achievable due to the highlevels of recombination in these regions and thecorresponding contribution to the device darksaturation current. These voltage limitations arenot of major significance at the moment due to thelimitations imposed by the substrates. However,in the future as wafer thicknesses are reduced toimprove the device economics and improved rearsurface passivation is introduced, the cells will havethe potential for improved open circuit voltages,but only provided the surfaces, including under themetal, are well passivated.4.3.2.3.2 Semiconductor FingerSolar CellsTo accommodate a top surface emitter sheetresistivity of at least 100 ohms per square, metalfingers need to be spaced no more than 1mm apartto avoid excessive sheet resistivity losses. Due tothe large width of screen printed metal lines or100 microns or more, such a close spacing is notpossible without shading well over 10% of the cellsurface. The concept of semiconductor fingers istherefore introduced as shown in Figure 4.3.2.3.These semiconductor fingers are formed by laserdoping the silicon surface while simultaneouslypatterning the dielectric layer to expose theheavily doped silicon surface. Developmentalwork with the laser doping process allows sheetresistivities as low as 1 ohm/square to be achievedwhile simultaneously forming laser doped lines ofonly 8 microns width. Spacing such lines 0.8mmapart avoids significant resistive losses within thelightly doped emitter which can be diffused to100 ohms/square, ensuring excellent response tothe short wavelengths of light. In the laser dopedregions, approximately half of the incident light islost due to absorption within the heavily dopedsemiconductor fingers.This is defined as a 50% effective shading lossof the semiconductor fingers. Using this design,the effective shading loss of the semiconductorfingers is only 0.5%, while the effective emittersheet resistivity of the emitter is 50 ohms/square.This latter figure results from the emitter sheetresistance of 100 ohms/square in parallel with thelaser doped lines which cover 1% of the area with asheet resistivity of 1 ohm/square and therefore alsoeffectively contribute 100 ohms/square. This allowsthe screen-printed lines to retain their normalspacing, but with 99% of the emitter being lightlydoped and therefore able to achieve near unityinternal quantum efficiencies for short wavelengthsof light. In addition, the passivating dielectric layernot only passivates the lightly diffused surface soas to give near unity internal quantum efficienciesfor short wavelength light, but it also isolates themetal from these same regions to minimise thedevice dark saturation current. Importantly, thesilicon is only exposed at the semiconductor fingers,with the screen-printed metal having been shownto make excellent ohmic contact to the heavilyphosphorus diffused silicon in these regions.Both thick oxides and silicon nitride layers, whenused with appropriate pastes, appear to provideadequate protection to the lightly diffused surfaceregions, preventing the screen-printed metal fromcontacting the silicon.A typical cell design based on this concept usessemiconductor fingers 8 microns wide of sheetresistivity 1-2 ohms/square and spacing of 0.8mm.The top surface sheet resistivity is typically 100ohms/square, with the effective sheet resistivityin the direction parallel to the semiconductorfingers typically 50-60 ohms per square. The screenprinted metal lines are printed perpendicularto the semiconductor fingers with a width of110 microns and spacing of 2.2mm. The overall26

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTCross-sectional SEM photos show: (a) discontinuitiesin the Al-doped p + layer; (b) a deep and uniform Aldopedp + layer.Figure 4.3.2.5Photo of a hybrid screen-printedand plated solar cell showing thethree levels of metallisation. Each ofthe two busbars effectively collectcurrent from 8 small solar cells,each with a tapered screen-printedline that carries the current fromthe narrow plated lines to therespective busbar/interconnect.Figure 4.3.2.4b4.3.2.3.3 Advanced SemiconductorFinger Solar CellsAs part of the core research of the Centre ofExcellence funded by the ARC and UNSW, anenhanced version of the semiconductor finger solarcell was designed and developed to overcome thelimitations of the standard semiconductor fingersolar cell described in Section 4.3.2.3.2. This involvesplating 2 microns of silver (or Ni/Ag or Ni/Cu/Ag) tothe screen-printed metal at the end of processing,which simultaneously plates a similar thicknessto the semiconductor fingers when using lightinduced plating as shown in Figure 4.3.2.4B.The application of this plated metal has severalbenefits. Firstly, it makes good ohmic contactto both the heavily doped silicon and also thescreen-printed metal, overcoming the high contactresistance sometimes experienced by the standardsemiconductor finger solar cell. This makes high fillfactorsroutinely achievable while solving the yieldproblems of the original implementation. Secondly,the increased conductivity of the semiconductorfingers facilitates increasing the screen-printedmetal line spacing to typically 2-3cm. Although thecorresponding width has to be increased to 400microns, it has the benefit of allowing both taperingand increased height to above 50 microns.This reduces the shading loss of the screen-printedmetal fingers from 5% down to about 1%, withthe corresponding Jsc increase taking efficienciesto above 19%. Interestingly the effective shadingloss of the semiconductor fingers was expectedto increase by 0.5% although in reality, the platedlaser doped surface appears rough enough toscatter the reflected light so that almost half istotally internally reflected at the glass/air interfaceand returned to the solar cell surface. Thereforethe effective shading loss of the semiconductorfingers remains virtually unchanged. Importantly,by avoiding the contact between the screen-printedsilver and the lightly doped silicon, the top surfacedesign of this solar cell behaves very differently tostandard screen-printed contacts and appears tohave the potential to achieve open circuit voltagesapproaching 700mV.The development of this technology will form partof the ASI project involving Silex and Suntech. It isparticularly well suited to multicrystalline siliconwafers due to the avoidance of prolonged hightemperature thermal processes. Based on small areatest devices, the inclusion of rear surface passivationis expected to take efficiencies on multi material toover 18% with corresponding open circuit voltagesin excess of 650mV while on mono, efficienciesof 21% are anticipated with open circuit voltagesexceeding 670mV.4.3.2.3.4 Advanced N-Type Screen-Printed Solar Cells4.3.2.3.4.1 N-type Screen-Printed Cellswith Homogeneous Top Surface DiffusionIncreasing interest is being shown in n-typeCZ material as a means for avoiding the widelyreported defects associated with the high boronand oxygen concentrations in p-type CZ material.In particular, screen-printed aluminium has beenused as a simple and cost-effective way to create anAl-alloyed rear emitter for such n-type CZ material,especially in the n + np + cell design with rear junction.However low voltages for such structures reportedin the range 617-627mV appear to be a severelimitation of this approach since n-type CZ wafersshould be capable of achieving much higher opencircuit voltages.Discontinuities in the p + layer have beenidentified as the main cause for such performancedegradation of this device. These discontinuities,as seen in Fig.4.3.2.5 (a), are isolated points wherethe junction fails to form, usually created by nonuniformwetting of the silicon by the Al during thealloying process. Although their presence can beminimized by optimizing the firing process so as toallow uniform wetting of the surface to occur, theycannot be completely avoided. In small quantities,such non-uniformities have almost negligibleinfluence on the performance of the back surface28

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTPL images illustrate the improvement in uniformityand quality of the p + layer achieved by the newfiring process. (a) before and (b) after the lowtemperature treatment.Figure 4.3.2.6Inverted form of the PERL (Passivated Emitter and RearLocally diffused) solar cell developed at UNSW basedon the use of N-type silicon.Figure 4.3.2.7field in conventional p-type cells. However, they cansignificantly degrade the quality of the Al-alloyedemitter in n-type wafers by allowing Al to locallybypass the p + region and directly contact the n-typebulk via a Schottky barrier causing a non-linearshunt. A new and modified firing process hastherefore been developed to avoid the damagefrom such non-uniformities. In this method, apatented low temperature solid phase epitaxialgrowth process is employed after the conventionalstandard spike firing to minimize the impacts ofthese junction discontinuities so that a uniform andgood quality junction as illustrated in Fig. 4.3.2.5 (b)can be achieved.These improvements have facilitated a 15-20mVincrease in Voc relative to those reported in theliterature and efficiencies over 17% for standardscreen-printed solar cell technology withhomogeneous n-type emitter (front surface field)applied to n-type wafers. Figure 4.3.2.6 showsthe improvement in the photoluminescent (PL)response resulting from the improved firing andformation of the rear contact and junction.A common problem with the manufacture ofconventional screen-printed solar cells is minuteamounts of aluminium paste accidently cominginto contact with the cell front surface. This canhappen when wafers are face down during theprinting of the rear or during wafer transfer/handling. Aluminium paste contamination of suchsurfaces and transfer belts happens relatively easilyin a production environment such as through abroken wafer that has been Al printed, operators’contaminated gloves, tiny holes in the screenprintingscreen etc. However such contaminationwith the current rear junction n-type technologycreates unusual photoactivated shunts that are notpresent in the dark.This is because a p-n-p transistor structure is formedwhen the unwanted aluminium on the top surfaceis fired into the n-type surface (with the second p-njunction of course being at the rear of the device).In this phototransistor, the lightly doped waferforms the base of the transistor, and despite thebase being very thick, approaching200 microns, the high lifetime of theCZ n-type material allows the transistorto achieve moderate gain levels andtherefore conduct large currents whenilluminated by light that generatesthe necessary base current for thetransistor. The unusual consequenceis that the apparent shunt resistanceof the cell is very poor (low) whenilluminated brightly, increasing tohigh values in the dark or even lowillumination levels. The ramificationsof this for cell efficiency and fill-factorare that values fall with reducing lightintensity as is normally the case forshunted cells, but with values increasing againfor low illumination levels as the shunt problemsdisappear as the phototransistors are deactivated.Even without such phototransistor shunting,analysis of such homogeneous emitter n-typedevices indicates that even higher voltagesare potentially achievable if not for the largedark saturation current contribution from theheavily doped phosphorus diffused top surface.This emphasises the importance of moving tothe equivalent of a selective emitter design forthe front surface to facilitate both improvedshort wavelength response as well as increaseddevice voltages.4.3.2.3.4.2 N-type Screen-Printed Cellswith the Equivalent of a Selective EmitterUntil recently, the Centre held the world record(jointly with Stanford University) for the mostefficient n-type silicon devices with 22.7% efficiency.The cell design was based on the inverted formof the PERL (Passivated Emitter and Rear Locallydiffused) solar cell developed at UNSW and isshown below. In this work, the cell design hasbeen adapted to accommodate the use of low costscreen-printed solar cell processes involving thealignment of the screen-printed front metal lines tothe heavily doped n + regions to form the equivalentof a selective emitter on the front surface and theSchematic of a laser dopedAl-alloyed rear junction n-typesolar cell.Figure 4.3.2.829

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTReflectance (%)8070605040302025201510400 550 700 850 1000Photolith. inverted pyramidsI.J. inverted pyramidsI.J. v-grooves100300 450 600 750 900 1050 1200Wavelength (nm)(a) Optical Microscopephotograph of ink jetpatterned lines in a SiO 2layerand (b) Scanning ElectronMicroscope photograph ofthe grooves formed in thesilicon surface following KOHetching of (a).Figure 4.3.2.16Reflectance of silicon waferswith different surface finish (i)inverted pyramids formed byphotolithography, (ii) invertedpyramids patterned by ink jetprinting and (iii) v-groovespatterned by ink jet printing.Figure 4.3.2.1730-40 micron diameter holesformed by ink-jetting.Figure 4.3.2.14(a) Optical Microscope photographof ink jet patterned holes in a SiO 2layer and (b) Scanning ElectronMicroscope photograph of theinverted pyramids formed in thesilicon surface following KOHetching of (a).Figure 4.3.2.15The preferred implementation of the laser dopingselective emitter (LDSE) technology also uses anequivalent laser doping/plating combination onthe rear surface using a boron doping source. Inthis cell design, the majority of both the front andrear surfaces is well passivated using silicon nitridealthough the preferred rear surface passivationof the undiffused p-type surface uses somewhatdifferent deposition parameters for best results.Implied Voc values above 730mV at one-sundemonstrate the near perfect passivation achievedwith such surfaces. Following laser doping of therear surface, even though plated contacts can beused similarly to on the front surface, the preferredrear contact is achieved through depositingaluminium over the entire surface followed by alow temperature sinter used to form good ohmiccontact with the boron laser doped regions. Inthis cell design, the aluminium layer provides anexcellent rear surface reflector. Even with standardcommercial grade p-type CZ wafers, based onlaboratory results it appears this technology willachieve comfortably over 20% efficiency on fullsized commercial wafers when established in pilotproduction in early 2011 with impressive opencircuitvoltages in the vicinity of 670mV.The simpler version of the LDSE technologyequivalent to the pilot production established atSunrise with aluminium alloyed rear surface, is farsimpler to retrofit onto existing screen-printedproduction lines, with 90% of existing equipmentretained. For several manufacturers such asShinsung in South Korea, this new technologywhereby the laser doped selective emittercombined with light induced plating of the metalcontacts are used to replace the front surfacescreen-printed metal is achieving significantlyhigher efficiencies well in excess of 19% comparedto 18% for conventional screen-printed cells withminimal if any cost increase per cell. Roth and Rauhave also secured the rights from UNSW to developand sell turn-key production lines based on theLDSE technology world-wide This is expected tomake it far simpler for many companies, particularlynew manufacturers, to take up the new LDSEtechnology in large scale production.Two other popular implementations of the laserdoping technology with industry collaborators arethe bifacial structure using laser doped self-alignedcontacts of opposite polarity on both surfaces andthe interdigitated rear surface laser doped contactsfor rear junction n-type devices. Collaborativeresearch projects based on these cell designs havebeen established with the aim of developing thetechnologies to take cell efficiencies on n-typeCZ also to above 20% in large scale commercialproduction at some stage in the future.Another use of the laser doping technology wasdescribed in Section 4.3.2.3.4.2 as a replacementfor screen-printed contacts with n-type CZ wafersin conjunction with aluminium alloyed rearcontact and junction. As reported above, excellentefficiencies in the vicinity of 19% have also beenachieved with this cell design on 148.6cm 2 n-typeCZ wafers. A simpler patented version of thistechnology has also been developed which requiresno diffusion processes, no edge junction isolationand no thermal processes above 450 degreesCelsius except for the aluminium alloying processfor several seconds. Never-the-less, the technologyis still able to achieve efficiencies above 18% onfull-sized commercial substrates. The processingsequence used is:1. Wafer texturing2. Silicon nitride deposition3. Aluminium rear surface printing and firing4. Laser doping front surface5, Ni/Cu light induced plating4.3.2.3.7 Inkjet Technology for Solar CellFabrication4.3.2.3.7.1 Resist Based MethodInkjet technology has been an area of rapiddevelopment over the last decade, particularly forprinting. In recent years, its application has beenspreading to other fields, but as yet has had only32

Reflectance (%)8070605040302025201510400 550 700 850 1000Random upright pyramidsI.J. inverted pyramidsI.J. v-groovesARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORT100300 450 600 750 900 1050 1200Wavelength (nm)(a) Reflectance of silicon wafers with different surfacefinish (i) laboratory fabricated random, upright pyramids,(ii) inverted pyramids patterned by ink jet printing and (iii)v-grooves patterned by ink jet printing and (b) SEM of (iii).Figure 4.3.2.18minimal impact in photovoltaics. The first companyapparently to commercialise a photovoltaictechnology incorporating inkjet technology is CSGSolar who during 2006 commenced productionof a thin-film technology that uses inkjet printingof a corrosive material to etch patterns in aresist layer to facilitate metal contacting to theunderlying silicon.In the present work, the use of inkjet technologyhas been expanded to encompass a range of solarcell fabrication processes including texturing,grooving, patterning of dielectric layers for metalcontacting, localized diffusions, etc. The techniquesdeveloped to carry out these processes are uniquelydifferent to those used before. A non-corrosiveplasticizer is inkjet printed onto a low cost resistlayer, altering the chemical properties of the resistlayer in these localized regions to make thempermeable to etchants such as hydrofluoric acid(HF). This facilitates the patterning or etching ofunderlying dielectrics or semiconductor materialto facilitate a range of semiconductor processes.Importantly, the change in resist permeability isa reversible process making it feasible to returnthe resist to its original state after carryingout processes on the underlying material. Thisalso opens the option to partially reverse thepermeability to reduce the hole or feature sizeproduced in the underlying material.A particular exciting application of this inkjettechnology work is for very high efficiency siliconsolar cells. The University of New South Waleshas held the world record for silicon solar cellefficiencies for the last 15 years, initially with thePassivated Emitter solar cell (PESC) and morerecently with the Passivated Emitter and RearLocally diffused (PERL) solar cell. Despite theperformance and achievements of these twotechnologies, neither has been used commercially,apart from for space cells, primarily due to thesophistication, cost and complexity of the processesinvolved. The photolithographic based processing isprobably the main contributor to this. In this work,inkjet printing techniques have been developed forpatterning low cost resist layers as a simple, muchcheaper alternative to photolithographic basedprocessing.These new approaches appear capable of achievingsimilar device performance levels but with thegreatest challenge being to match the dimensionsof features in the resist patterning achievable withphotolithography. Test devices to date based oninkjet technology have achieved feature dimensionssuch as holes of 30-40 microns diameter as shownin the matrix of holes below.However, these dimensions needto be reduced to about 10 micronsdiameter to fully match theperformance levels demonstratedwith photolithographic basedprocessing. New and innovativeinkjet printing techniqueshave been recently developedfor further reducing the resistpatterning dimensions. This workis being greatly assisted by therecent availability of the new 1picolitre inkjet heads.The described dielectric patterningcapabilities via inkjet patterningcan not only be used for defining localized diffusionregions and locations for metal contacts, but also forthe formation if textured surfaces and light trappingschemes. Examples of the latter are shown below aswell as corresponding reflectance curves.Significant interest is also being shown indeveloping high efficiency approaches forHoneycomb texturing (a) andmacrogroove texturing (b) ofmulticrystalline silicon wafersthrough the use of inkjettechnology.Figure 4.3.2.19Schematic showing the principleof “Direct Etching” for patterningdielectric layers.Figure 4.3.2.2033

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORT(a) Optical image, and (b) Dektak profile of a grooveetched in a 260 nm thick thermal oxide grown ona polished Si wafer by deposition of 10 layers ofaerosolised 10% (w/v) NH 4F solution onto a polyacrylicacid polymer layer of thickness ~ 2.2 microns. Theaerosol and sheath gas flow rates were 21 and 50 cm 3 /min, respectively. The process velocity of the stagewas 10 mm/s and the platen was heated to 45ºC. Thegroove width at surface ~35 microns.Figure 4.3.2.21(a) Optical image of groovesetched in a 260 nm thick thermaloxide grown on a polished Siwafer by deposition of 25 layersof aerosolised 10% (w/v) NH 4Fsolution onto a polyacrylic acidpolymer layer of thickness ~ 2.2mm. The aerosol and sheath gasflow rates were 18 and 55 cm 3 /min,respectively. The process velocityof the stage was 10 mm/s and theplaten was heated to 45ºC. (b)Dektak profile of a section of thegrooves shown in (a).Figure 4.3.2.22Aerosol jet-etching of a siliconoxynitride layer to form lines of only15-20 microns.Figure 4.3.2.23multicrystalline silicon wafers. Record performancemulticrystalline silicon cells have in the pastbenefited from photolithographically defined“honeycomb” texturing using an acidic isotropicetch. Such structures have also been demonstratedrecently using inkjet technology as shown in Figure4.3.2.19. Such techniques appear to have significantcommercial appeal.4.3.2.3.7.2 Direct Etching ViaAerosol JettingThe key to high-efficiency silicon solar cells is theability to form small-area metal contacts to thesilicon through dielectric passivating layers. In thiswork, a new method for patterned etching of SiO 2,SiOxNy and SiN xdielectric layershas been developed and patented.The method uses Optomec’s M 3 Daerosol jetting device to deposit asolution containing fluoride ions,according to an etching pattern,onto an acidic water-solublepolymer layer formed over thedielectric layer as indicated in theschematic of Figure 4.3.2.20. Thedeposited solution reacts withthe polymer layer, at the locationswhere it is deposited, to form anetchant that etches the SiO 2and SiN xunder the polymer layer to form apattern of openings in the dielectriclayer. After the pattern of openings is formed, theacidic water-soluble polymer and the etch residueare easily removed by rinsing in water. The methodinvolves fewer steps than photolithography and issafer than existing immersion etching techniquesin that the corrosive etchant HF is only formedin-situ on the surface to be etched. Furthermore,the method uses small amounts of inexpensivechemicals and produces significantly less hazardousfluoride waste than existing immersion etchingmethods.Aerosol jet printing is a new deposition technologybeing pioneered by Optomec, Inc. The techniqueenables the finely-controlled deposition of anaerosol, which is generated from a liquid, by usinga sheath gas to constrict the aerosol into a fine jetwhich is directed to the substrate. The techniquehas been previously used in applications such asprinted electronics, fuel cells and displays.The aerosol etching method has been used to etchgroove structures in SiO 2which have been thermallygrown on polished silicon wafers. By varying theaerosol and sheath gas flow rates the geometryand depth of etching can be varied. For example,grooves (see Figure 4.3.2.21) were cleanly etched ina 260 nm thick SiO 2layer using aerosol and sheathgas flow rates of 21 and 50 cm 3 /min respectivelyand depositing 10 layers of aerosolized 10% NH 4Fsolution. These grooves were ~35 microns wide atthe surface but only ~ 20 microns wide at the Si/SiO 2interface.Even narrower structures can be etched by reducingthe aerosol flow rate and increasing the sheath gasflow rate. Figure 4.3.2.22 shows the formation ofcentral, more deeply, etched grooves which are 200nm deep and ~10 microns at the Si/SiO 2interface.It is anticipated that the use of an even lower flowrate for the first, wetting layer will largely eliminatethe wider, shallow etched region at the surface thusenabling very narrow grooves to be etched to adepth of at least 200 nm.We have also been able to etch grooves 20-25 microns wide in 75 nm thick SiN xlayers onchemically-textured silicon surfaces. The etchedgrooves can be used to form front or rear metalcontacts for silicon solar cells using metal plating,screen printing or evaporation deposition methods.34

luminescence imaging tools, with healthy sales toresearch groups and cell manufacturers throughoutthe world during 2010.In summary, photoluminescence techniques haveproven particularly useful for the characterisationof various aspects of the screen printed solarcell process. Photoluminescence imaging wasapplied to the study of belt furnace processing,while photoluminescence imaging with currentextraction was applied to the characterisation ofn-type solar cells with printed metallisation toidentify failure mechanisms in the metallisation. Theaccuracy, high speed and non-destructive natureof PL techniques make it an attractive candidatefor use in high-throughput screen printed solar cellproduction lines.4.3.2.4 Pilot ProductionMany of the above technologies have beenimplemented or are in the process of beingimplemented into pilot production. Table 4.3.2.1summarises the performance of each of thetechnologies in pilot production using standardcommercial wafers with commercial processesand equipment.4.3.2.5 Core ResearchThe distinguishing feature of core research is that itis funded by UNSW and the Centre of Excellence orother Government funding scheme. The prioritiesfor the work are therefore set by the Centre ratherthan by industry. Any IP is therefore wholly ownedby UNSW, but with any such developments madeavailable at no cost to industry partners of theCentre such as those funding collaborative researchprograms or else licensing Centre technology. Suchcore research is therefore of great importance in theinitial stages of development of a new technologyto take the development to the point whereindustry interest is gained and industry is willingto fund the ongoing development. An exampleof this is the Advanced Semiconductor FingerSolar Cell described in Section 4.3.2.3.3 wherecore research has facilitated the development anddemonstration of the concept, to the point wheretest structures can show that efficiencies above19% are achievable, making these the most efficientscreen-printed cells world-wide on standardcommercial grade p-type CZ wafers. This work isnow to continue as an industry funded project inconjunction with the ASI, with the industry partnersbeing Silex and Suntech.Other important areas of the core research arewhere problems are tackled and the solutionscan be used for the benefit of all or most industrypartners. Examples include improvements in thelight induced plating techniques, reducing the laserinduced defects during the laser doping process,reducing feature sizes during inkjet patterning ofdielectric layers, improving surface passivationtechniques, adapting processes to suit low costsilicon etc. The Centre has a large number ofvery talented PhD students contributing to thiscore research, the achievements from which aresignificantly accelerating the progress for mostindustry collaborative research projects. Thisrepresents a major benefit to companies thatengage in funding photovoltaic research at UNSW.Some core research however is conducted inconjunction with companies, but where thecompanies are not funding the work at UNSW andhave no claim over any resulting IP from the work atUNSW and cannot exercise any restriction regardingits use. An example of this is the collaboration withCentrotherm to evaluate the compatibility betweenthe LDSE technology and the standard Centrothermscreen-printing technology. Cells from a standardCentrotherm screen-printing line for producinghomogeneous emitter solar cells of 18% efficiencywere fabricated without front silver metal. Theemitter sheet resistivity was the only processingvariation, being raised to 120 ohms per square.After laser doping and plating at UNSW, the finished5 inch cells were sent to the Fraunhoffer Institute inGermany where they were confirmed to be 19.3%efficiency. This is believed to be the highest everindependently confirmed for industrial solar cellsfabricated on standard commercial grade p-type CZwafers. This record may however be shortly brokenthrough collaborative work with Shinsung wheresimilar devices fabricated using similar wafers havebeen measured to be 19.6% efficiency and havebeen sent to Fraunhofer for confirmation.ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTTechnologyARCJsc(mA/cm 2 )Voc(mV) FF (%)Effic(%)Area(cm 2 )Wafertype1 BCSC SiO 236.8 619 79.0 18.0 148.6 (p-type CZ)1 BCSC SiN 33.4 604 77.4 15.6 156 (p-type multi)2 IBBC SiN 37.1 675 79.4 19.8 46 (n-type CZ)3 Screen-print SiN 37.1 632 78.8 18.5 148.6 (p-type CZ)3 Screen-print SiN 33.3 614 75.5 16.0 156 (p-type multi)4 SCF SiN 36.7 635 79.0 18.4 148.6 (p-type CZ)5 LDSE (Sunrise) SiN 37.8 638 79.0 19.0 225 (p-type CZ)5 LDSE (UNSW) SiN 37.6 638 79.6 19.0 148.6 (p-type CZ)5 LDSE (Centrotherm) SiN 38.4 638 78.9 19.3 148.6 (p-type CZ)5 LDSE (Roth & Rau) SiN 37.8 635 78.4 18.8 225 (p-type CZ)5. LDSE (Shinsung) SiN 38.6 640 79.3 19.6 148.6 (p-type CZ)5 LDSE SiN 35.4 628 77.9 17.3 156 (p-type multi)5 LDSE SiN 36.0 645 79.9 18.6 148.6 (n-type CZ)6 Pluto SiN 38.4 638 79.5 19.5 148.6 (p-type CZ)6 Pluto SiN 38.9 638 81.9 20.4 4.0 (p-type CZ)6 Pluto SiN 35.7 633 78.6 17.7 156 (p-type multi)7 PERL ZnS/MgF2 41.7 704 82.5 24.2 46 (p-type FZ)Performance of each ofthe technologies in pilotproduction using standardcommercial wafers withcommercial processes andequipment.table 4.3.2.137

4.4 Second Generation:Silicon, Organic and other“Earth Abundant” Thin-FilmsARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORT38Research TeamUniversity Staff• Dr Sergey Varlamov (group leader)• Prof. Martin GreenResearch Fellows• Dr Thomas Soderstrom (since 6/2010)Technical Staff• Dr Patrick Campbell• Mark GriffinPostgraduate Research Students• Johnson Wong (PhD) (until 9/2010)• Zi Ouyang (PhD) (until 9/2010)• Guangyao Jin (PhD) (until 9/2010)• Jialiang Huang (PhD)• Yuguo Tao (PhD)• Peter Gress (PhD)• Hongtao Cui (PhD)• Do Soon Jung (Masters)• Bonne Eggleston (PhD)• Wei Li (PhD)• Qian Wang (PhD, since 3/2010)• Kyung Hun Kim (Masters, since 3/2010)• Anthony Teal (PhD, since 3/2010)• Jae Sung Yun (PhD, since 3/2010)• Chaowei Xue (PhD, since 9/2010)Visiting AcademicA. Prof. Soon Kweon (since 1/2010)4.4.1.1 SummaryThe focus of the silicon thin-film groupresearch has been on further improvementsin the technology of e-beam evaporatedpoly-crystalline (poly-Si) thin-film solar cells,which is conducted within an ARC Linkagegrant project with CSG Solar.The UNSW e-beam evaporator, the toolsupplying precursor Si films for theproject, underwent major refurbishmentduring 2010, allowing greatly improvedreliability and up-time of the tool, toreduce background contamination leveland to achieve more accurate control ofSi deposition rate and doping, allowingdeposition of high quality Si films. Afterprocessing the films into solar cells andapplying a newly developed etch-back Sitexturing process a new UNSW efficiencyrecord for evaporated cells of 7.1% (Voc458 mV, Jsc 26.6 mA/cm 2 , FF 58%) has beendemonstrated.Outstanding progress has been madein the area of light-trapping in poly-Sithin-film cells. Firstly, a comprehensivestudy of various back-surface reflectors(BSR) of practical significance has beenconducted. It allows choosing the best BSRdepending on the cell design and structure.For example, the best BSR for a planar cellis a thick pigmented resin (P150) with ahigh load of TiO 2, while the best BSR fora textured cell is a highly reflective whitepaint. Secondly, further development inplasmonic poly-Si thin-film cells led torecord-breaking Jsc enhancement of 50%.Thirdly, a simple and effective Si film etchbacktexturing process has been introducedto improve light-trapping in the e-beampoly-Si cells on planar glass, which allowedachieving a record current of 26.6 mA/cm 2 ,the highest ever reported for evaporatedpoly-Si solar cells.The first encouraging results have beenobtained for transient heating defectanneal of poly-Si films. The open-circuitvoltage improvement of about 58 mV wasdemonstrated after poly-Si cell structureswere exposed to diode laser or flash lampradiation for only a few milliseconds. Thediode laser was also successfully applied tolarge area boron diffusion to prepare backsurface field (BSF) layers.A large amount of data has been collectedabout effects of different annealingtreatments on solid phase crystallisation(SPC) of Si films and resulting poly-Si filmquality and cell performance. A focus wason application of higher temperaturesduring either the whole SPC process or

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTStructure of a p-type poly-Si solar cell on planar glass(layer thicknesses and grain size not to scale).Figure 4.4.1.1its stages, such as the incubation, to shorten thecrystallisation while maintaining or even improvingthe cell performance. The result is that, regardlessof when a higher temperature is used, it alwayscauses poorer poly-Si crystal and electronic quality.However, the degree to which the cell performancedeteriorates, depends on the Si film preparationconditions. E-beam cells and hybrid cells, i.e. thecells with PECVD emitter and e-beam absorber andBSF, can be crystallised at 640C within only twohours (as compared to 20 hrs at 600C) without asignificant loss in their performance. In the area ofthe SPC emitter seed-layer approach, the factorsleading to the best seed crystal quality are beingstudied.A comprehensive concept of performance limitingrecombination in poly-Si has been developed.According to the concept it is intragrain defects,such as dislocations that are responsible for limitingthe minority carrier lifetime in poly-Si material andthus the cell performance. In the practical range ofdopant concentrations in the quasi-neutral regionof the cell absorber (> ~5E15 cm -3 ), the lifetimelimiting recombination occurs via shallow levelslinked to dislocations. At dopant concentrationslower than ~5E15 cm -3 the lifetime is limited byrecombination via deep levels introduced bycharged impurities at dislocations and possiblygrain-boundaries. The dislocation density in poly-Sithin-film cells was estimated from the TEM imagesand found to be of an order of 1E10 cm -2 , whichis consistent with experimental voltages of about500 mV.4.4.1.2 IntroductionThe technology of thin-film poly-Si on glass solarcells is developing steadily and it is expectedultimately to become cost-competitive withother thin-film technologies [4.4.1.1, 4.4.1.2]. Thepresently best developed poly-Si on glass PVtechnology has been commercialised by CSG SolarAG where an approximately 1.5 mm thick PECVDamorphous silicon (a-Si) precursor diode is solidphase-crystallisedto form a poly-Si film, which isannealed and hydrogenated to activate dopants,remove and passivate defects, and thenprocessed into metallised modules[4.4.1.3]. The highest achieved efficiencyfor 94 cm 2 mini-modules produced byCSG technology is 10.5% [4.4.1.4], and fullscale 7-8% efficient modules have beenmanufactured in Thalheim, Germanysince 2006.Research in the thin-film group atUNSW focuses on critical cell fabricationprocesses and explores a range ofadvanced approaches to improve theperformance and manufacturabilityof poly-Si thin-film cells. The group’swork over the past years has led toinnovative solutions to the key steps ofthe cell fabrication process, including Al inducedglass texturing (AIT) [4.4.1.5] and localised surfaceplasmons (SP) in metal nanoparticles [4.4.1.6] forenhanced light trapping, silicon deposition byhigh rate PECVD [4.4.1.7] and e-beam evaporation[4.4.1.8], defect anneal and passivation, and cellmetallisation and interconnection [4.4.1.9]. Thefollowing sections summarise the thin-film cellresearch projects and their results.4.4.1.3 Solar cell fabrication sequenceThe thin-film group research in 2010 focused ontwo different solar cell types: cells deposited byPECVD and by e-beam evaporation. A schematicrepresentation of a poly-Si thin film cell is shown inFigure 4.4.1.1 and Figure 4.4.1.2 describes the cellfabrication sequence.Additionally hybrid cells combining a PECVD emitterwith an e-beam absorber and BSF were producedin collaboration with CSG Solar. The substrateused for all cells is 3.3 mm thick borosilicate glass(Schott Borofloat33). Both planar and textured glasssubstrates are investigated: typically e-beam andhybrid cells are fabricated on the planar glass whilePECVD cells are fabricated on textured glass. Thetexture on the Si facing glass surface aims to reducethe reflection losses and to enhance the lighttrapping in a weakly absorbing poly-Si thin-film. TheProcess sequence of the fourtypes of poly-Si thin-film on glasssolar cells under development atUNSW. All cells are designed for thesuperstrate configuration (i.e., thesunlight enters the cell throughthe glass).Figure 4.4.1.239

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTQuv [%]110100680 C 640 C 600 C9080706050403020E-beam10Hybrid0PECVD-105 50 500Annealing Time [min]Calculated random scatter absorption limit (blackdots) and measured optical absorption in 2 μm thickpoly-Si thin films on planar, AIT, and SAIT glass with airas the BSR.Figure 4.4.1.4Normalised UV crystallinity factor (Q uv) of PECVD,E-beam and Hybrid films as a function of the annealingtime at 600ºC, 640ºC and 680ºC.Figure 4.4.1.5AFM images of bare AIT (a) and SAIT(b) glass and SEM image of 2 μmthick poly-Si film on AIT glass (c).Figure 4.4.1.3texture is prepared by the UNSW-developed AITmethod [4.4.1.10] described below and consistingof an irregular rough array of sub-micron sizeddimples.The next step is the deposition of anapproximately 75 nm thick SiN film (refractiveindex ~2.1 at 633 nm) by PECVD at about 400°Cor by reactive sputtering at 200°C. The SiN filmacts both as an AR coating and a barrier layerfor contaminants from the glass. An a-Si n +/ p/p + precursor diode is then deposited by eithere-beam evaporation or PECVD on planar or AITglass respectively, followed by SPC at 600°C. Theresulting as-crystallised poly-Si diodes have Vocof only about 150-180 mV. A high temperaturetreatment is then performed at ~1000°C for 30-60s to activate the dopants and to anneal defects,which increases Voc up to 220-250 mV. The finalstep of the material preparation is hydrogenationin a remote hydrogen plasma to passivate theremaining defects resulting in Voc of 450~500 mVtypical for the finished cells.The cells are metallised using a combinationof photolithographic patterning with dielectricand metal deposition processes to create aninterdigitated electrode structure on the rear (Siside) of the cells. Typically, the BSF electrode of thePECVD cells covers the whole rear cell surface andthe cells can only be illuminated from the glassside [4.4.1.11, 4.4.1.12]. The BSF electrode of thee-beam cells consists of thin Al fingers covering only4% of the rear surface, thus allowing bifacial (i.e.illumination from both glass and Si side) operationof the cells [4.4.1.13, 4.4.1.14, 4.4.1.15]. The detailsof the cell metallisation are described in respectivesections below.The hybrid cells are made on planar glass with theSiN layer and the emitter deposited in the CSG Solarlarge area KAI PECVD tool and the absorber and BSFlayers in the UNSW e-beam evaporator. The filmsare processed into either individual 2 cm 2 large solarcells using the UNSW bifacial metalisation or solarmini-modules of 36 cm 2 area using the proprietaryCSG Solar metallisation technology describedelsewhere [4.4.1.4].4.4.1.4 Glass TexturingAn advanced glass texturing method wasdeveloped at UNSW and it is referred to as“aluminium induced texture” (AIT) [4.4.1.5, 4.4.1.10].A thin sacrificial Al film is deposited onto planarglass, followed by annealing at about ~600°C inan inert atmosphere. The anneal initiates a red-oxreaction whereby Al is oxidised to Al 2O 3and SiO 2from the glass is reduced to silicon, as follows:A surface texture is imparted to the glass accordingto the nucleation conditions provided during thisanneal. Subsequent wet-chemical etching removesthe reaction products from the glass surface andreveals the glass texture (Figure 4.4.1.3a). Furtherimprovement in the AIT process conditions led todevelopment of so-called “sub-micron AIT” (SAIT)glass with smaller and steeper feathers (Figure4.4.1.3b), which ensure even better light absorptionin a poly-Si film prepared on such texture (Figure4.4.1.4) [4.4.1.16].Typical atomic force microscope (AFM) images ofthe bare AIT and SAIT glass surfaces with the RMSroughness of 800 and 150 nm respectively and anSEM image of the AIT glass coated with a 2 μm thickpoly-Si film are shown in Figure 4.4.1.3.As shown in Figure 4.4.1.4, the absorption inthe 2 μm thick poly-Si films on the best AITand particularly SAIT glass in the 500~1000 nmwavelength interval is very close to the calculatedrandom scattering absorption limit, based onMonte Carlo ray tracing. It should be noted thatthe significantly higher absorption than thetheoretical limit at 1000 nm and above can beattributed to the enhanced parasitic absorption inthe glass and/or measurements errors as discussedelsewhere [4.4.1.11].40

Films Method Incubation Time [min] Crystallisation Time [min]600°C 640°C 680°C 600°C 640°C 680°COTM 300 50 16 780 140 29Raman 360 50 12 720 125 28XRD 270 62 16 1020 120 32UV-R 270 50 14 900 125 32PECVD Average 300 53 14 855 128 30OTM 300 35 11 720 90 23Raman 300 32 10 660 90 24XRD 270 31 12 1020 105 24UV-R 270 31 9 900 90 25E-beam Average 285 32 10 825 94 24OTM 300 45 14 780 130 28Raman 360 40 12 690 120 28XRD 270 60 15 1020 135 30UV-R 270 45 12 900 120 30Hybrid Average 300 48 13 848 126 29Generic equation [4.4.1.19] 660 108 22Incubation andcrystallisation timesfor Si films (solar cellstructures) deposited bydifferent techniques.Table 4.4.1.1ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORT4.4.1.5 Solid Phase CrystallisationSolid phase crystallisation (SPC) is a thermallyactivated process of transformation of a-Si into poly-Si, which is used for poly-Si thin-film cell fabrication.The SPC parameters, such as the exact temperatureprofile, as well as the film structure, such as thecomposition and thickness of the individuallayers, have effects on the resulting crystal andelectronic quality of poly-Si film and thus on the cellperformance. A detailed study of the SPC kineticswas conducted to clarify such effects. Severaltechniques – optical transmission microscopy(OTM), Raman, UV reflection (UV-R), and X-raydiffraction (XRD) spectroscopes, scanning electronmicroscopy (SEM) – were used to characterise theSPC kinetics and the crystal quality of the films.External Quantum Efficiency (EQE) and Suns-Vocmeasurements were used to characterise the cellperformance [4.4.1.17, 4.4.1.18].Representative SPC kinetics for Si films depositedby different techniques (PECVD, e-beam, andhybrid) at different temperatures based on the UV-Rmeasurements are shown in Figure 4.4.1.5. Similarso-called “S-curves” were obtained by the othercharacterisation techniques mentioned above. Thesummary of the incubation and full crystallisationtimes based on all film characterisation methods isgiven in Table 4.4.1.1.Based on the data from the table it is possibleto estimate the activation energies (Ea) for theincubation and crystal growth processes usingthe Arrhenius law. The Arrhenius plots are shownin Figure 4.4.1.6. The Ea calculated from the plotsare 3.0 and 3.3 eV (e-beam), and 2.7 and 3.2eV (PECVD), and 2.8 and 3.2 eV (hybrid) for theincubation and crystal growth respectively. TheEa for the crystal growth falls within the reported3.1-3.4 eV range, but the Ea of less than 3.0 eV forthe incubation is significantly smaller than ~3.4 eVpreviously found [4.4.1.20]. Most likely it is becausethe nucleation rate, to which the incubation timeis closely related, is enhanced by particular filmdeposition conditions and/or the particular filmstructure, such as the presence of the heavilydoped layers and interfaces.The average grain size in the e-beam poly-Si filmscrystallised at different temperatures was estimatedfrom the SEM images shown in Figure 4.4.1.7, wherethe grain boundaries were enhanced by Seccoln(1/time) [s -1 ]ln(1/time) [s -1 ]-6-7-8-9-10-7-8-9-10Annealing Temperature [ C]680 640 600IncubationCrystal growth-1112.0 12.5 13.0 13.51/kT [eV -1 ]Arrhenius plots for the incubationand crystal growth processes inPECVD (♦), E-beam (■) and Hybrid(●) films with the n+/p-/p+ solarcell structure.Figure 4.4.1.6E-beam = 3.01 eVHybrid = 2.82 eVPECVD = 2.72 eVE-beam = 3.31 eVHybrid = 3.17 eVPECVD = 3.18 eVFilmsPECVDE-beamHybridV ocJscAbsorber Doping Diffusion Length [nm]T [°C] [mV] [mA/cm 2 ] p.FF [%] p.Eff [%] [/cm 3 ]Emitter Absorber600 474 12.7 76 4.6 5.4×10 15 160 2400640 466 11.5 70 3.8 5.8×10 15 140 1800680 456 10.8 63 3.1 4.6×10 15 135 1500600 474 13.4 76 4.8 9.8×10 15 180 2150640 462 13.3 75 4.6 8.6×10 15 155 2100680 432 12.9 74 4.1 9.5×10 15 120 2000600 459 14.4 75 4.9 2.9×10 15 235 4000600 455 14.1 75 4.8 3.2×10 15 210 4000640 433 13.6 73 4.2 2.5×10 15 170 4000Performance parameters ofthe poly-Si thin film solar cellsobtained by integration of theEQE curves over the AM1.5Gsolar spectrum, from Suns-V ocmeasurements, and C-V analysis.All cells have an area of 2 cm 2 .Table 4.4.1.2(a)(b)(c)3 m3 m3 mSEM images of the fullycrystallized and Secco etchede-beam poly-Si films for differentSPC temperatures.Figure 4.4.1.741

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTAverage Grain Size [µm]2.72.4E-beam2.1PECVD1.8Hybrid1.51.20.9Normallised intensityTc-Si600 C640 C680 C0.6560 600 640 680 720SPC Temperature [℃]508 512 516 520 524 528 532Shift [/cm]a) The average grain size of PECVD (♦), E-beam (■) andHybrid (●) films obtained from respective SEM imagesas a function of the SPC temperature.b) Raman spectra of the E-beam films crystallised atdifferent temperatures (c-Si as a reference). The peakintensities are normalised.Figure 4.4.1.8Thickness, nmThe Raman peak shift andFWHM for emitter seed layersof different thickness and withdifferent P doping.Table 4.4.1.3Hall mobilities for emitter seedlayers of different thickness and withP doping of 3.3E20 cm -3 .Table 4.4.1.4UV reflection peaks at 274 nm forSPE poly-Si cell structures and c-Siwafer as a reference.Figure 4.4.1.9Intrinsic P 8E18, cm -3 P 1e20, cm -3 P 3E20, cm -3Shift, cm -1 FWHM, cm -1 Shift, cm -1 FWHM, cm -1 Shift, cm -1 FWHM, cm -1 Shift, cm -1 FWHM, cm -150 517.7 7.46 517.6 7.43 517.7 6.86 517.4 7.87100 518.4 7.42 517.3 6.92 517.8 6.72 517.0 7.23200 518.5 6.75 518 6.90 518.5 6.94 518.5 6.76Thickness, nm Carrier conc., cm -3 Hall coeff., cm 3 /C Mobility, cm 2 /V*s50 8.22E19 -9.12E-08 63100 8.66E19 -8.67E-08 53200 8.76E19 -8.56E-08 35etching. The grain size decreases from 2.3 µm to 1.5µm and 1.0 µm as the SPC temperature rises from600C to 640C and 680C. Similar trends are observedfor PECVD and hybrid Si films as shown in Figure4.1.1.8. The Raman peaks for Si films of all types areslightly shifted to the left of the c-Si peak, and theirFWHM are larger for thehigher SPC temperaturessimilarly indicating a slightlypoorer crystal qualityof the films crystallisedat higher temperatures(Figure 4.4.1.9).All performance parameters of the poly-Si thin-filmsolar cells degrade for higher SPC temperaturesbut to a different extent (Table 4.4.1.2). The PECVDcells are most affected, with the pseudo efficiencydecreasing from 4.6% to 3.1% for SPC temperaturesof 600C and 680C respectively. The e-beam andhybrid cell pseudo-efficienciesdecrease more modestly, from4.9% to 4.2% for the same SPCtemperatures. There is hardly anyloss in the e-beam and hybrid cellperformance after SPC at 640C but itallows shortening the SPC time fromtypical 20 hrs to only 2 hrs. It makes640C a preferred SPC temperaturefor such cells, particularly for processdevelopment purposes when 0.1-0.2% loss in the efficiency is notimportant.The research continued in the areaof the SPC emitter seed layer. It waspreviously reported that a heavilyphosphorus doped (>1e20 cm -3 ) Si layer can form ahigh crystal and electronic quality poly-Si materialafter SPC [4.4.1.21]. If it isused as a seed layer followedby solid-phase epitaxy (SPE)of the absorber and BSFlayers, it can help to improvethe quality of the cell poly-Simaterial and to lead tobetter cell performance. Tofind preparation conditions that result in betterseed layer properties a series of films was fabricatedwith different P concentrations and thickness. Theseed layer quality was assessed by Raman (Table4.4.1.3) and UV-R spectroscopy and the Hall mobility(Table 4.4.1.4). The best quality is found for thethickest seed layer of 200 nm and intermediate Pconcentration of 1e20 cm -3 . Using the optimisedemitter seed layers high crystal quality poly-Si cellstructures were grown by SPE as confirmed by the98% Quv quality factor (Figure 4.4.1.9).4.4.1.6 Transient heating defect annealAs was mentioned in section 4.4.1.3, the defectannealing in poly-Si films after SPC is a veryimportant process for improving cell Voc. Findingthe best conditions for this process is complicateddue to conflicting trends. On one hand, higherannealing temperatures and longer times arepreferred because of better dopant activationand more complete defect dissolution. On theother hand, the same conditions cause excessivedopant smearing and glass distortion, which needto be avoided. Typically, for any given annealingtemperature, there is an optimum time balancingfavourable processes with unfavourable ones. Forexample, at 950ºC such an optimum time is about3 min, while a similar Voc effect can be achieved at1050ºC for only 50 s. Following this trend, it can beestimated that, at temperatures near the Si meltingpoint (1414 ºC), the optimum annealing timescan be of order of a few milliseconds, because ofwhich this type of processes are generally called“millisecond annealing” [4.4.1.22]. Very importantly,for such short times the thermal diffusion lengthin glass is of an order of tens of microns only[4.4.1.23], which means the bulk of the glass stays42

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTMetallisation on cell levelcompleted: top view of a 5x5 cm 2sample showing 4 individual cellsplus cross sectional view (SiN andsilica layers not shown) with theemitter and air side busbars.Figure 4.4.1.15A schematic air-side top view of amini-module using wire-bondedcell interconnection.Figure 4.4.4.161.210.80.60.40.2Deposition rate of a-Si:H as afunction of power for the SPCoptimisedPECVD process.Figure 4.4.1.13Raman spectra (normalised) of thepoly-Si films obtained by SPC of thestandard (pink) and high (blue) ratePECVD a-Si:H. The spectrum of c-Siis (black) shown for reference.Figure 4.4.1.14Absorber deposition rate,nm/minSuns-Voc results (the average of fivepoints) after RTA and hydrogenationof SPC poly-Si diodes on AITglass with PECVD a-Si:H absorberdeposited at 35 nm/min (“low rate”)250 nm/min (“high rate”).Table 4.4.1.71-Sun Voc,mVhi gh r at estd. rateS i wa f e r0505 510 515 520 525 530 535n=1 Voc,mVHigh rate PECVD of a-Sifilms for poly-Si cellsThe manufacturing of poly-Si thin-filmsolar cells is a commercially viable process[4.4.1.1]. However, the deposition ofthe precursor a-Si films by currentlyconventional PECVD has been identifiedas having high cost and a low throughputdue to its low deposition rate (25-35 nm/min) and the high equipment cost [4.4.1.3].Si evaporation by electron beam (e-beam)has been investigated as a higher rate andlower cost alternative. A challenge with e-beamevaporation is its directional nature (“line-of-sight”deposition), leading to non-conformal coatingof textured surfaces, which results inmicrostructural defects and the poortextured cell performance [4.4.1.25,4.4.1.26]. Another high rate process, whichin contrast to the evaporation can providea conformal coating of textures, is hot-wireCVD. However, no performance results havebeen reported for the SPC hot-wire CVDsolar cells, and it is also known that thehot-wire CVD a-SiH has very high nucleationrates [4.4.1.23], which may lead to poorcrystal and electronic quality of SPC poly-Si.A major part of research conducted on PECVD ofa-Si:H has been traditionally focused on minimisingthe defect density for a-Si:H solar cell applications.However, device-grade a-Si:H, which performs wellfor a-Si:H cells, is not necessarily the best precursormaterial for making poly-Si cells byn=2 Voc,mV35 494.6 ± 8.9 496 666250 512.4 ± 4.5 513 711SPC and a higher rate PECVD a-Si:Hmay also be suitable for producinga good electronic quality poly-Sifilms. Using a conventional 13.56MHz PECVD system and a processoptimised for the SPC application,much higher deposition rates up to 265 nm/min have been demonstrated as shown in Figure4.4.1.13 [4.4.1.7].Raman microprobe characterisation of the poly-Sifilms obtained after SPC from the high rate PECVDa-Si:H material confirms an identical crystal qualityto that of the standard rate poly-Si as shown inFigure 4.4.1.14.The poly-Si cell structures were fabricated withthe thin emitter and BSF layers deposited atthe standard rate and the ~2 µm thick absorberdeposited at the high rate of 250 nm/min in adifferent PECVD chamber. After the standardmaterial processing sequence described above, theVoc of the high rate cell was higher than the Voc ofthe standard rate cell as shown in Table 4.4.1.7.The current research in PECVD poly-Si cells isfocusing on developing advanced cells, whichcombine the best AIT glass for enhanced lighttrapping with the high rate PECVD Si films toachieve efficiencies exceeding 10%.PECVD cell metallisation – cellinterconnectionDuring fabrication of large size thin-film PVmodules, the as-deposited films are dividedinto smaller area unit cells, which are theninterconnected in series to reduce the resistiveohmic losses. A novel approach allowing a lowloss interconnection of individual thin-filmcells into a minimodule using wire-bondingwas recently developed in the UNSW thin filmgroup [4.4.1.9]. Figure 4.4.1.15 shows a samplecontaining four individual PLASMA cells prior to theinterconnection process.The interconnection method consists of laserscribing an isolation groove along the glass sidealuminium busbar to define the cells. Wire-bondsare then placed over this groove, connecting theair-side BSF busbar to the glass side emitter busbarof the adjacent cell as shown in Figure 4.4.4.16. Thismethod can be expanded by applying a conductivematerial (with a non-conductive adhesive) to thesurface of the cell, and wire-bonding the first andthe last cell busbars to the surface of the conductivematerial as also shown in the figure.44

I-V curves of the individual cells andthe wire-bonded minimodule.ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTFigure 4.4.1.17Schematic the of e-beam Sievaporation tool with boronand phosphorus effusion cellsfor in-situ doping.Figure 4.4.1.18Table 4.4.1.8 and Figure 4.4.1.17 show the exampleelectrical results from wire-bonding of two adjacentPECVD cells. As can be seen, the interconnectionscheme improves the Jsc and, thus, the overallmodule efficiency compared to the individual cells,which is primarily due to the large reduction inlateral series resistance along each busbar to thenearest series or module wire-bond connectionpoint. All I-V measurements were done withaperture to define the cell area.4.4.1.8 E-Beam evaporated Si cellsE-Beam cell structure, fabrication, andperformanceDeposition of Si using an e-beam evaporator canbe performed at a very high rate of up to 1 mm/min. Other advantages of the method includeabsence of toxic gases, good Si source materialusage and compatibility with a continuous in-linedeposition mode. Thus, if performed in a non-UHVenvironment (base pressure > 1×10 -8 Torr, pressureduring Si evaporation > 1×10 -7 Torr), e-beamevaporation is a cost effective Si deposition methodfor the thin-film PV applications (Figure 4.4.1.18).The thin-film group at UNSW has been researchinge-beam Si thin-film cells since 2004, and significantprogress has been achieved in advancing thecell technology from a few square millimetrearea mesa devices at the start to recent fullyfunctional 2 cm 2 area 6% efficient cells.Figure 4.4.1.19 illustrates the progresswith the efficiencies of the e-beamcells and Table 4.4.1.9 summarises thedesign features of the cells typicallyfabricated at UNSW. The cells are alsomade with the reverse polarity structure,i.e. the p-type glass side emitter, and then-type absorber and air-side BSF layers.During previous years the focus of thee-beam cell research was on improvingthe poly-Si material quality and optimisingthe cell structure and post deposition treatments(RTA, hydrogenation) to achieve better voltages.Respectable Voc in the range up to 500 mV weremeasured by the Suns-Voc technique on the nonmetallisedcells. More recently the focus has shiftedto developing a working metallisation scheme,improving the cell currents and fill factors (FF), i.e.producing functional e-beam cells with appreciableconversion efficiencies.Cell A Before Cell B BeforeABinterconnectedVoc (mV) 508.7 507.9 1010.6Isc (mA) 96.4 97.6 101.3FF (%) 69.78 68.45 70.9Eff (%) 7.8 7.7 8.3Parameter DetailsGlass3.3 mm planar Borofloat33AR coating SiN (~75 nm, n ~2.1)Emittern + (~100 nm, P, 5e19~1e20 cm -3 , 400~500 W/sq)Base p (~2 µm, B, 1e16~1e17 cm -3 )BSFp + (~150 nm, B, up to 1~5x10 19 cm -3 , ~1000 Ω/sq)RTA 10~20 s at 1000°C or ~4 min at 900°CHydrogenation 15~20 min at ~600°C, remote plasmaMetal0.5-1.5 μm thick interdigitated Al fingers on front & rearBack reflector Diffuse while paintEvolution of the PECVD and e-beamcell efficiencies at UNSW.Figure 4.4.1.19PECVD thin-film cell(prior to interconnection,encapsulation) and module(after interconnection,encapsulation)electrical properties.Table 4.4.1.8Typical design parameters ofthe UNSW E-beam Si cells.Table 4.4.1.945

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTV oc(mV) J sc(mA/cm 2 ) FF (%) Eff (%) Rs (Ωcm 2 )479 20.1 66 6.3 4Measured 1 sun J-V results of1.8 µm thick planar e-beamcells with aligned bifacialmetallisation ([4.4.1.13,4.4.1.14]). The shunt resistanceof the cells is > 2000 Ωcm 2and has a negligible effect onthe performance.Table 4.4.1.10Schematic (left) and actual (right) views of a metallised e-beam cell (whitepaint BSR not shown).Figure 4.4.1.20Doping (cm -3 ) V oc(mV) FF (%) J sc(mA/cm 2 ) R s(Ωcm 2 ) Eff (%) pEff (%)1.7×10 15 413 64.1 19.2 2.0 5.09 5.55.7×10 15 418 61.9 17.6 3.4 4.67 5.31.0×10 17 449 62.5 6.5 7.9 2.16 2.58.3×10 17 403 54.4 4.0 19 1.04 1.3Performance parameters of e-beamcells with different absorberdoping densities.Table 4.4.1.11Measured values for Voc, pFF,J sc(EQE), and pEff as a functionof the absorber doping. J sciscalculated from EQE results.Jsc(EQE) and pEff were obtainedunder two different back surfacereflector (BSR) schemes: air (blackclosed squares) and white paint (redopen squares).Figure 4.4.1.21E-Beam cell metallisationThe e-beam cell metallisation scheme, illustratedin Figure 4.4.1.20, was developed in 2008 and allowbifacial illumination and versatile back surfacereflector (BSR) study. It features interdigitated Al linecontacts for both heavily doped layers, the emitterand BSF. The BSF electrodes only cover about 4% ofthe area. The emitter electrode is formed in a waythat avoids contact with the absorber layer entirely.Instead of plasma etching the groovesthrough the whole Si thickness, a timed etchstop is used such that a thinned emitter layerstill remains at the bottom of the grooves.Then, the glass-side comb-like electrode iscentred in the grooves leaving significantspace (>10 μm) between the edges of theelectrodes and the groove sidewalls. Aftermetallisation the entire cell rear surface isthen coated with a layer of a white paint asa diffuse BSR. The typical cell performanceparameters obtained with the describedmetallisation are listed in Table 4.4.1.10.E-Beam cell absorberdoping densityThe electronic properties of the absorberlayer, mainly the effective minority carrierdiffusion length and life-time, are expectedto depend strongly on the layer doping level,which thus can have a large effect on the cellperformance. After introducing a workingmetalisation scheme, an absorber dopingoptimisation experiment was conducted in orderto estimate this effect. All cells in the experimentwere intended to be nominally identical but withthe different doping densities in the absorber(about 2 μm thick, made on the SiN coated planarglass, and with the metallisation and the whitepaintBSR as described in the previous section).The dopant concentration was determined usingthe Z-analysis technique described elsewhere[4.4.1.27]. This technique has an advantage ofyielding the electrically active dopant density,which is most relevant to the cell performance. Thecells were characterised by Suns-Voc, EQE, lightJ-V measurements and the results are shown inTable 4.4.1.11. To account for the spectral mismatchbetween the light I-V tester and the AM1.5G solartest spectrum, which causes measurement artifacts,a conservative approach was used where the shortcircuitcurrent density during the J-V measurementswas adjusted to match those determined fromEQE measurements.The major observed effect of the doping densityis a sharp reduction of the cell current withincreased doping. It dominates a weaker effecton the cell voltage, where V ocrises slightly up todopant density of about 1e17/cm 3 before fallingsharply at the higher densities. The best cellefficiencies are obtained at the lowest active dopantconcentrations in the range between 1e15/cm 3 and5e15/cm 3 . Figure 4.4.1.21 shows the effect of thedoping densities on the key cell parameters.Role of back surface filedThe BSF is very important for the poly-Si solar cell.Besides transporting the holes to the metalliccontacts, it also repels the electrons from the rearcell surface thus reducing recombination at thisinterface. In fact, this effect is crucial for the poly-Sithin-film device as shown in Figure 4.4.1.22. Withoutthe BSF the V ocof the cell is reduced from 435 mV toonly 270 mV. Increasing the BSF thickness from 30to 90 nm decreases the sheet resistance by a factor3 (as expected) but does not impact the Voc of thedevice as presented in Figure 4.4.1.22. Increasingthe deposition rate of the BSF has a positive effecton the Voc. The Voc increases by 40 mV when thedeposition rate increases from 5 A/s to 10 A/s.The best efficiency reported previously for thepoly-Si cell with the identical metallization was5.2 % with a voltage of 435 mV [4.4.1.13]. Thedevelopment of the high rate BSF clearly improved46

Voc [mV]2 m thick film480BSF deposition rate465 5 [A/s]10 [A/s]450 2.5 [A/s]435420405+165 mV+40 mVJ sc[mA/cm 2 ]ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORT2702550 30 60 90BSF thickness [nm]20.017.515.012.510.07.55.02.5 No BSR 482 mV 68% 14.8mA/cm2 4.9%White paint 479mV 66% 20.1mA/cm2 6.3%0.00 50 100 150 200 250 300 350 400 450 500V oc[mV]Voc of 2 µm thick evaporated poly-Si solar cells withvarious BSF layers thickness and deposition rate.Figure 4.4.1.22I-V curves of 2 µm thick planar poly-Si thin film solarcells with and without white paint as back reflector.Figure 4.4.1.23the voltage. The I-V curves data of the solar cellswith and without a BSR are presented in Figure4.4.1.23. The poly-Si solar cells with a white paintBSR has a conversion efficiency of 6.3% withVoc=479 mV, FF=66% and Jsc=20.1 mA/cm 2 . TheBSR increases the Jsc by 5.3 mA/cm 2 and decreasesthe FF by 2% due to the increase of Jsc and (increaseof ohmic losses in the metallization).Light- trapping in e-beam poly-Sithin‐film cellsPoly-Si thin-film solar cells need effective lighttrappingto compensate for the moderateabsorption. This is typically achieved by glasssubstrate texturing for PECVD cells. However,evaporated poly-Si cells are not compatible withtextured glass due to low density material grown ontextured substrates as shown in Figure 4.4.1.24. Across-sectional FIB investigation of the evaporatedSi films on the textured glass reveals noticeablemorphological differences from the films on theplanar glass. On relatively smoother textures the Sifilm has apparently vague areas of lower materialdensity, or high defect density, extending from theinflexions in the texture features. Such defectiveareas further develop into microcracks or voidseither after the thermal treatment (SPC, RTA) oron the rougher textures. When the glass texturehas vertical or overhanging features, extendeddiscontinuities in the evaporated Si film are found.Most described defects are believed to be relatedto the columnar microstructure, which is typical forevaporated films. Clear experimental evidence forthe columnar microstructure and the low densityvoid network present in the SPC poly-Si films wasrecently reported [4.4.1.26].Such a defective structuralmorphology is formed due totwo inherent characteristicsof the evaporation, highdirectionality of the arrivingatomic flux and the adatom lowsurface mobility. The depositiononly takes place on the surfaceareas in direct line-of-sight ofthe evaporation source resulting in discontinuitiesin the areas shaded from the source either by thesurface topographical features or by the previouslydeposited atoms themselves. The earlier arrivedadatoms cast “shadows” for subsequently arrivingadatoms, thus creating a network of parallel highdensity columns surrounded by lower densitymaterial, as it is schematically shown in Figure4.4.1.25. Besides, in contrast to PECVD deposition,the evaporated adatoms have a sticking coefficientclose to unity; they do not have or receive theexcessive energy to move in any way fromthe place of their initial landing to find a morethermodynamically favorable position, thus leadingto morphological and possibly electronic defects.Possible approaches to reducing an extent of thedefect formation during deposition by evaporationis using smoother glass texturing and/or to texturethe Si film surface instead of, or in addition tosmooth glass textures or a completely differentapproach, the localized surface plasmon (LSP)enhanced light-trapping. The different structuresare presented in Figure 4.4.1.26.FIB cross section of an evaporated Sifilm on textured glass substrate. (a)is a grey-scale image, and (b) is thesame image but set to a black andwhite mode.Figure 4.4.1.24Schematic representation offormation of low density materialand voids during evaporation,where already adsorbed adatomscast shadows for newly arrivingadatoms thus forming parallelchannels filled with no or lowdensity material.Figure 4.4.1.25Schematic representationdifferent light trappingscheme. A) planar, B)textured glass, C) backtexture, D) nanoparticles.Figure 4.4.1.2647

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTLight-scattering by surfaceplasmons (top-left); SEM imageof Ag nanoparticles formedfrom 16 nm thin precursor film(top-right); AFM micrograph ofSi cell surface coated with Agnanoparticles (left).Figure 4.4.1.27AFM micrograph of etch-backtextured poly Si film surface.Figure 4.4.1.28EQE, 1-R [a.u.]100908070605040302010Schematic structure of plasmonicpoly-Si thin-film solar cell with 44%Jsc enhancement.Figure 4.4.1.29J sc[mA/cm 2 ]Ag nano 23.5white paint 21.9Si back text 24.20300 400 500 600 700 800 900 1000 1100wavelength [nm]EQEs of a cell with planar interfaces(black), a plasmonic cell, and a cellwith Si-back texture.Figure 4.4.1.30The plasmonic light-trapping relies on scatteringby Ag nanoparticles as shown in Figure 4.4.1.27[4.4.1.28]. Such nanoparticles can be relativelysimply formed by depositing a 5-30 nm thin Agfilm on already metallised and characterisedpoly-Si e-beam cells followed by annealing atabout 200ºC leading to formation of 50-200 nmwide irregularly shaped particles. The AFM profileand SEM image of the Ag nanoparticles on apoly-Si film are presented in Figure 4.4.1.27.The silicon back texture can be simply made byetching the silicon film in a KOH solution. Aftera few second, pyramids formation appears aspresented in Figure 4.4.1.28. The roughness istypically around 100 nm whereas the roughnessof the nanoparticles is around 15nm. Thisdemonstrates the different mechanism of lightscattering involved into the two methods.The optimisation of the two approaches is currentlyongoing at UNSW. The best developed process forplasmonic light-trapping uses a Ag nanoparticlearray formed from 16 nm thick precursor Ag filmdirectly on the poly-Si cell coated with a350 nm thick MgF 2layer, with a white paintBSR on the top, a spacing layer between thenanoparticles and the back reflector, typically350 nm of MgF 2(Figure 4.4.1.29) [4.4.1.6]. TheJsc enhancement due to the plasmonic lighttrappingis 44%.The Si-back texture can be made by wetchemical etching, plasma etching, or acombination of both. The current gainachieved due to the Si-back texture is about66%, significantly higher than the gain due tothe plasmonic light-trapping (Figure 4.4.1.30).However, both approaches have further potentialfor improvement, currently being evaluated.4.4.1.9 Back-Surface Reflector studyMost thin-film solar cells, and poly-Si cells inparticular, are not thick enough to fully absorbincident light even if the cells are textured toimprove light-trapping. A back-surface reflector(BSR) is a necessary part of cell structure whichcollects transmitted light, which otherwise wouldbe lost, and redirects it back to the cell thusincreasing the cell optical thickness at least twice.Additionally, most practically used BSRs are diffuseand thus scatter the light, contributing to betterlight- trapping. Depending on particular celldesign, different BSRs are used: UNSW PECVD cellsuse a combination of silicon dioxide layer with Al;UNSW e-beam cells use a diffuse white paint or Agnanoparticles (or both); CSG Solar cells (both PECVDand e-beam) use a combination of the diffusepigmented resist coated with Al [4.4.1.1]. However,until recently there was no systematic study of BSRsto know which one is best suited particular poly-Sicell design.A comprehensive comparative study has beenconducted during 2010, which characterised andcompared the performance of different BSRs onpoly-Si solar cells. The cell material prepared byPECVD was provided by CSG Solar and processedinto metallised cell using the UNSW bifacialmetallisation described in the previous section(Figure 4.4.1.30). To ensure fair comparison, allBSRs were tested on the same cell, which wascharacterised prior to and after application of eachBSR. The following BSRs were examined: acrylicwhite paint (WP); P150 resist loaded with titaniaparticles (P150W); P150W coated with ~ 150 nmAl (P150W-Al); P150 with a double load of titania(P150DW); twice thicker P150DW (DP150DW); Agnanoparticles. The results are summarised in Table4.4.1.12 and also plotted in Figure 4.4.1.31.For the planar cells the Ag nanoparticles providethe best Jsc enhancement of about 48% on average(Jsc ~21 mA/cm 2 ) (the best cell enhancement is50%) followed by the double P150DW and thewhite paint. For the textured cell one has to note alot higher Jsc (~24 mA/cm 2 ) compared to the planarcells even without any BSR due to excellent lighttrappingproviding by the random glass texture.The enhancement due to BSRs are very modestwith the white paint performing the best. The Agnanoparticle BSR is not tested for the textured cellsbecause nanoparticles cannot be formed on thetextured surface.48

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTCurrent enhancement in planar (left) and textured (right)poly-Si solar cells dur to different back-surface reflectors.Note that the reference Jsc is 14.25 and 24. 18 mA/cm 2for the planar and textured cell respectively with no BSR.Figure 4.4.1.314.4.1.10 Recombinationprocesses in thin-film Si onBSRglass solar cellsThe current output of poly-Si thin-filmsolar cells is approaching the socalledLambertian limit, which servesas a gauge for quantum efficiencies.However, there is little understandingabout how high the Voc could be– the lowest foreseeable practicallimit is the V ocof multicrystalline Si wafer basedsolar cells. Thus, the voltage of poly-Si cells as anacademic topic is still a research frontier where thefundamental limits are unknown and speculationsabound. In previous years many researchers haveattributed the lower lifetimes of poly-Si solar cells totheir small grain size, a view which holds merit if thegrain boundaries are not well passivated [4.4.1.29],but in more recent years it is becoming clear thatintragrain defects are also potent enough to limitthe lifetime to the nanosecond range [4.4.1.30].While the grain boundary recombination can besuccessfully mitigated by enlarging the grain size oroptimising the passivation of dangling bonds withinthe grain boundaries, so far there has not been aneffective method to circumvent the detrimentaleffects of intragrain defects. The goal of research atUNSW is to elucidate the dominating recombinationpathway in the poly-Si thin-film solar cells so thatthe attainment of higher Voc becomes a moretractable problem.As a result of a comprehensive study of the poly-Sion glass solar cell characteristics (Suns-Voc, EQE)and film properties over a range of temperaturesPlanarTexturedJsc, mA/cm 2 Gain, % Jsc, mA/cm 2 Gain, %No BSR 14.25 24.18 (80%)WP 18.86 32.4 26.19 8.3P150W 17.16 20.4 24.37 0.8P150W/Al 17.75 24.6 24.58 1.7P150DW 18.56 30.2 25.35 4.9DP150DW 19.09 34.0 25.44 5.2Ag nanoparticles 21.51 50.9and absorber dopant concentrations, a conceptconsistent with experimental observations, whichdescribes the dominant recombination in poly-Sihas emerged [4.4.1.31, 4.4.1.32, 4.4.1.33, 4.4.1.34].According to the concept, the carrier recombinationin poly-Si on glass solar cells can be modelled asa superposition of two processes. The first oneinvolves the electronic transition between shallowstates which are 0.05-0.07 eV below the conductionband and 0.06-0.09 eV above the valence band,respectively, which are consistent withthe shallow bands at silicon dislocations(Figure 4.4.1.32a). The second processoccurs via deep levels at charged defects.The shallow band recombinationdictates the solar cell properties overthe entire practical absorber dopingrange of 5E15~1E17 cm -3 ; the deep levelrecombination originates from eithercharged dislocations or grain boundariesand only becomes influential at dopantconcentration lower than ~5E15 cm -3(Figure 4.4.1.32b).EEEcFvJsc enhancement by differentback-surface reflectors onplanar and textured poly-Sithin-film solar cells.Table 4.4.1.121qΦ(a) The dominant recombinationpathways in poly-Si solar cellsinvolving shallow defect levels: 1)direct transition between shallowbands; 2)transition betweendeep level and shallow band; (b)Simulation of the lifetime at V ocvsdopant concentration. The shortdashed line represents shallowbands recombination, and the longdashed line represents deep levelrecombination at grain boundaries.2Figure 4.4.1.3249

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTlifetime (s)1.E-081.E-091.E-10PECVD 2.2e15PECVD 5.1e15e-beam 6.4e15ALICE 2e16ALICE 4.5e16ALICE 5e16lifetime (s)1.E-081.E-091.E-10PECVD 3e15PECVD 1e16PECVD 2.4e16PECVD 1e16e-beam 1.56e17e-beam 7.2e15e-beam 6.9e16e-beam 2.5e161.E-111.E-111.E-1230 40 50 60 70 80q/kT1.E-1230 40 50 60 70 80q/kTLifetime vs q/kT plots for a) p-type, and b) n-typepoly-Si thin-film cells.Figure 4.4.1.35Arrhenius plots for the diodesaturation current J 01/(J L/I) (a) andfor short-circuit current Jsc (c) forpoly-Si and c-Si solar cells; b) Vocplots extrapolated to 0K for poly-Siand c-Si solar cells.Figure 4.4.1.33Typical sequence of EQE curvesat different temperatures rangingfrom 120K to 320KFigure 4.4.1.34In the ideal sc-Si the diode saturation current J 01follows the Arrhenius law with the activation energyEa equal to the Si bandgap when extrapolated to0K. The short-circuit current density, Jsc, in sc-Si isonly very weakly dependent on the temperature(due to the effect of band-gap variation). However,poly-Si solar cells exhibit different characteristics[4.4.1.31]. The Arrhenius plots for J 01of poly-Sisolar cells have distinctively a smaller slope, whichreflects a lower Ea associated with J 01by 0.15-0.18 eV. As a consequence, the Voc plots versus Textrapolated to 0K have a lower interceptin the range 1.0~1.1 eV as comparedto 1.27 eV for sc-Si solar cells (Figure4.4.1.33a, b). The Arrhenius plots for Jscof poly-Si cells whose absorber diffusionlength is shorter than the absorberthickness have steeper slopes indicatingstronger temperature dependence of thepoly-Si cell current (Figure 4.4.1.33c).Although the smaller Ea and theintercept could in principle be due to thenarrower band gap in poly-Si material ascompared to c-Si, it cannot explain thetemperature sensitivity of Jsc. On theother hand, both effects are consistentwith existence of shallow sub-bandgapstates acting as either minority carrier traps orrecombination centres.Further confirmation of the shallow-level relatedrecombination in poly-Si solar cells comes fromthe study of the bulk lifetimes as a functionof temperature which allows a more detailedunderstanding of the behaviour of these shallowlevels in the device quasi-neutral, bulk regions.Minority carrier diffusion lengths in the absorberlayers of various poly-Si thin-film cells wereextracted by fitting the device EQE simultaneouslywith its optical reflectance, a procedure whichhas become a routine for poly-Si thin-film cellswhose short lifetimes precludes quantificationby the standard techniques such as the quasisteady state photoconductance (QSSPC). The EQEmeasurement and fitting procedure were repeatedfor each sample at different temperatures rangingfrom 120K to 320K. Figure 4.4.1.34 superposes thefront-side EQE curves of the same cell at differenttemperatures (note that the heights of the curvesare reduced slightly due to additional reflectanceby a quartz window on the cryostat in which thesample is placed). Clearly the quantum efficiency ofthe cell varies strongly with temperature.Figures 4.4.1.35 a) and b) plot the extracted lifetimeτ against q/kT for p- and n-type cells respectively.Although the actual lifetime values at a giventemperature differ greatly from cell to cell, all thelifetime curves exhibit Arrhenius law with theactivation energy (Ea) of 0.17-0.21 eV near roomtemperature [4.4.1.32]. These activation energies50

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTTypical TEM DFWB image of a poly-Si film on glass.0.5 μmFigure 4.4.1.37are much larger than those reported for plasticallydeformed Si and multicrystalline Si [4.4.1.35],where the dominant recombination is thoughtto be via transitions between the shallow anddeep levels. Assuming direct transitions betweenshallow levels, the Ea is expected to be (Ec-E de)+(E dh-Ev)+2.5kT=180~120 meV (where E deand E dhareshallow band energies near the conduction and thevalence bands respectively), which is in remarkableagreement with the experimental values.Furthermore, as a consequence of electrontransitions between the shallow bands, therecombination rate becomes proportional to theconcentrations of both minority and majoritycarriers, as in the case of radiative recombination,in contrast to recombination via deep levels,where the rate is proportional to the minoritycarrier concentration only. It leads to minoritycarrier lifetime inversely proportional to themajority carrier concentration, or roughly to thedopant concentration [4.4.1.32]. The experimentalconfirmation of this effect is shown in Figure4.4.1.36 which plots the minority carrier lifetimesin a number of poly-Si solar cells versus the dopantdensity. There is a clear inverse relationshipbetween the lifetime and dopant density.Amongst the n-type cells in Figure 4.4.1.35b, threeare made from small grained (~1 μm) poly-Si (circleor square data points), and three are made fromlarge grained (~5 μm) poly-Si epitaxially grown onan Al induced crystallised (AIC) seed layer (triangledata points). The similarity in the temperaturebehaviour of the lifetime, regardless of the grainsize, indicates that large and small grain materialshave a common lifetime limiting mechanism. Also,in Figure 4.4.1.36 one readily sees that the AIC cells(triangles) can take on the lifetime values eitherabove or below the power law trend line, indicatingthat they do not have consistently superior lifetimesjust by virtue of their larger grain size. It was foundthat the predominant grain orientation plays agreater role in determining the Voc of AICcells in fact, of the two rightmost triangledata points representing AIC cells atabout 5E16 cm -3 doping in Figure 4.1.1.36,the higher lifetime point originates from(100) preferentially oriented material andthe lower lifetime point comes from (111)preferentially oriented material.1.E-091.E-10All the above evidence points atintragrain defects, as a likely limiton poly-Si thin-film cell lifetimes. In1.E-11particular, dislocations are the most likelycandidates because they provide thestrain fields necessary to create split-offstates from the band edges that can actas shallow levels [4.4.1.36]. To estimate dislocationdensity in poly-Si solar cells a TEM study wasconducted using the Weak-Beam-Dark-Field (DFWB)technique. An example image is shown in Figure4.4.1.37 and the dislocation density deduced by theimage analysis is in the order of 1E10 cm -2 . Thereis very rough correlation between the dislocationdensity found in poly-Si solar cells and the cell Voc.The best cells with Voc > 500 mV have dislocationdensity of 6~8x10 9 cm -2 , while for the cell with verypoor Voc < 450 mV the dislocation density is about1.5x10 10 cm -2 [4.4.1.37].As the dopant density decreases, the shallowband recombination gradually becomes less andless significant and eventually, at sufficiently lowdopant densities, less than about 5x10 15 cm -3 therecombination involving deep levels starts todominate (process 2 in Figure 4.4.1.32a). Such deeplevels are associated with defects concentrating indislocations and grain boundaries. The defects areusually charged to the same polarity as the majoritycarriers thus creating potential barriers whichattract the minority carriers. The lifetime in this casedepends on the potential barrier heights, whichbecome greater with lowering dopant density.1.00E+15 1.00E+16 1.00E+17 1.00E+18dopant concentration (cmlifetimeat 232K (s)-0.9619y = 462576xR 2 = 0.8161Series1AIC ALICE (n-type)e-beam (n-type)PECVD (ne-beam(p type)PECVD (p-type)Power (Series1)Lifetimes at 232K for various poly-Sithin-film cells plotted againstdopant concentration.Figure 4.4.1.3651

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTSimulated Voc of a poly-Si solarcell (2.2 µm thick n-type absorber,Jsc 27 mA/cm 2 ) in which bothshallow bands recombination anddeep level recombination at grainboundaries (with electrostaticfluctuations) occur.Figure 4.4.1.38The effect of recombination via deep levelsat charged defects is to place an upperlimit on the minority carrier lifetime (thusthe cell Voc). Figure 4.4.1.38 shows Vocsimulation results of poly-Si solar cellswhere two recombination processes inthe absorber, both via shallow bands andvia deep levels, taking place in parallel.Above 1x10 16 cm -3 doping, the Voc is fairlyconstant and approaches the 485 mVlimit corresponding to the case whereonly shallow band recombination occurs.Below 1x10 16 cm -3 doping, deep levelrecombination becomes influential to thedevice performance eroding the Voc by about 21mV when the doping reaches 1x10 15 cm -3 . Thus, inthe practically significant dopant concentrationrange of 5x10 15 ~5x10 16 cm -3 , an immediatechallenge to consistently raising the Voc above 500mV is to reduce the dislocation density to belowthe 1x10 9 cm -2 level. If and when this is achievedthe next step to even higher Voc is minimisation ofcharged defects likely associated with impurities.4.4.1.11 References4.4.1.1 M.A. Green, “Polycrystalline silicon on glassfor thin-film solar cells”, Applied Physics A: 96,p. 153, 2009.4.4.1.2 A. G. Aberle, “Fabrication and characterisation ofcrystalline silicon thin-film materials for solar cells”,Thin Solid Films, 26, p. 511, 2006.4.4.1.3 P. A. Basore, Proceedings of the 21st EUPVSEC, p.544, Dresden, Germany, 2006.4.4.1.4 M. Keevers, T.L. Young, U. Schubert, R. Evans,R.J. Egan and M.A. Green, “10% Efficient Csgminimodules:, Proceedings of the 22nd EUPVSEC,p. 1783, Milan, Italy, 2007.4.4.1.5 G. Jin, P. I. Widenborg, P. Campbell, S. Varlamov,“Lambertian matched absorption enhancementin pecvd poly-Si thin film on aluminiuminduced textured glass superstrates for solar cellapplications”, Progress in Photovoltaics: Researchand Application, 2010, v. 18, pp. 482-589.4.4.1.6 Z. Ouyang, S. Pillai, F. Beck, O. Kunz, S. Varlamov, K.R. Catchpole, P. Campbell, M. A. Green, “EffectiveLight Trapping in Polycrystalline Silicon Thin-filmSolar Cells by Means of Rear Localised SurfacePlasmons”, Appl. Phys. Lett., 2010, v. 96, 261109.4.4.1.7 P. I. Widenborg, G. Jin, P. J. Gress, and S. Varlamov,“High rate 13.56MHz PECVD a-Si:H depositions forSPC poly-Si thin film solar cells”, Proceedings of the24th EUPVSEC, Hamburg, 2009.4.4.1.8 Z. Ouyang, O. Kunz, A. B. Sproul, S. Varlamov,“Influence of the absorber doping for p-typeevaporated polycrystalline silicon thin-film solarcells on glass, J. Appl. Phys., 2011, 109, 060104.4.4.1.9 P.J. Gress, P.I. Widenborg, S. Varlamov, A.G. Aberle,Wire bonding as a cell interconnection techniquefor polycrystalline silicon thin-film solar cells onglass”, Progress in Photovoltaics: Research andApplications, 2010, v. 18, pp. 221-228.4.4.1.10 P.I. Widenborg, N. Chuangsuwanich, A.G. Aberle,Glass texturing, Patent AU2004228064A1.4.4.1.11 P.I. Widenborg, A.G. Aberle, “PolycrystallineSilicon Thin-Film Solar Cells on AIT-Textured GlassSuperstrates”, Advances in Optoelectronics, 2007,article ID 24584.4.4.1.12 P.I. Widenborg, S.V. Chan, T. Walsh, and A.G. Aberle,“Thin-Film Poly-Si Solar Cells on AIT-textured Glass– Importance of the Rear Reflector”, Proceedings ofthe 33rd IEEE PVSC, San-Diego, 2008.4.4.1.13 O. Kunz, Z. Ouyang, S. Varlamov, and A.G. Aberle,“5% efficient evaporated solid-phase crystallisedpolycrystalline silicon thin-film solar cells”, Progressin Photovoltaics: Research and Applications, v. 17,p. 567, 2009.4.4.1.14 O. Kunz, Z. Ouyang, J. Wong, and A.G. Aberle,“Advances in evaporated solid-phasecrystallizedpoly-Si thin-film solar cells on glass(EVA)”, Advances in Optoelectronics, 2008,article ID 532351.4.4.1.15 O. Kunz, J. Wong, J. Janssens, J. Bauer, O.Breitenstein, A.G. Aberle, “Shunting Problems Dueto Sub-Micron Pinholes in Evaporated Solid-PhaseCrystallised Poly-Si Thin-Film Solar Cells on Glass”,Progress in Photovoltaics, 2009, v. 17, pp.35-46.4.4.1.16 H. Cui, G. Jin, M. Wolf, P. Campbell, M. Geen,“Enhanced Absorption of SPC Poly-Si Thin Filmson Aluminium Induced Textured (AIT) Glass withSubmicron Texture Feature Size”, Proceedings ofthe 25th EUPVSEC, Valencia, Spain, 2010.4.4.1.17 S. Varlamov, Y. Tao, J. Wong, O. Kunz, R. Egan,“Crystallisation kinetics of e-beam evaporated Sifilms and its effects on poly-Si film and solar cellproperties”, 25th EUPVSEC, Valencia, Spain, 2010.4.4.1.18 Y. Tao, S. Varlamov, J. Wong, O. Kunz, R. Egan,“Effects of SPC temperature on properties ofevaporated poly-si thin films and soalr cells”, 35thIEEE PVSC, Hawaii, USA, 2010.52

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORT4.4.1.19 A. Blum, C.J. Feldmann, J. Non-crystall. Solids, 1972,v. 11, 242.4.4.1.20 C. Spinella, F. Priolo, “Crystal grain nucleation inamorphous silicon”, Appl. Phys. Rev., 1998, v.84,pp.5383-5414.4.4.1.21 T.Matsuyama, M. Tanaka, S. Tsuda, S. Nakano, Y.Kuwano, “Improvement of n-type poly-Si filmproperties by solid-phase crystallisation method”,Jap. J. Appl. Phys., 1993, v. 32, pp. 3720-3728.4.4.1.22 P. Timans, J. Gelpey, S. McCoy, W. Lerch, S. Paul,“Millisecond Annealing: Past, Present and Future”,Mat. Res. Soc. Symp. Proc., 2006, v. 912,4.4.1.23 K. Ohdaira, T. Fujiwara, Y. Endo, S. Nishizaki, H.Matsumura, “Formation of Several-Micrometer-Thick Polycrystalline Silicon Films on Soda LimeGlass by Flash Lamp Annealing”, Jap. J. Appl. Phys.,47, p. 8239, 2008.4.4.1.24 K. Ohdaira, T. Fujiwara, Y. Endo, S. Nishizaki,H. Matsumura, “Explosive crystallisation ofamorphous Si films by flash lamp annealing” J.Appl. Phys. 106, 044907, 2009.4.4.1.25 Z. Ouyang, O. Kunz, M. Wolf, P. Widenborg, G.Jin and S. Varlamov, “Challenges of EvaporatedSolid-Phase-Crystallised Poly-Si Thin-Film SolarCells on Textured Glass”, Proceedings of the18th International Photovoltaic Science andEngineering Conference & Exhibition, Kolkata,India, 2009.4.4.1.26 M. Werner, U. Schubert, C. Hagendorf, J. Schneider,M. Keevers, R. Egan, “Thin film morphology, growthand defect structure of e-beam deposited siliconon glass”, Proceedings of the 24th EUPVSEC,Hamburg, Germany, 2009.4.4.1.27 A. Straub, R. Gebs, H. Habenicht, S. Trunk, R. A.Bardos, A. B. Sproul, and A. G. Aberle, “Impedanceanalysis: A powerful method for the determinationof the doping concentration and built-in potentialof nonideal semiconductor p-n diodes”, J. Appl,Phys. 97, p. 083703, 2005.4.4.1.28 K.A. Catchpole, A. Polman, “Plasmonic solar cells”,Optics Express, 2008, v. 16, pp. 21794-21800.4.4.1.29 P. Altermatt and G. Heiser, “Development of athree-dimensional numerical model of grainboundaries in highly doped polycrystalline siliconand applications to solar cells”, J. Appl. Phys., 91,p. 2002.4.4.1.30 L. Carnel, I. Gordon, D. Van Gestel, G. Beaucarne,J. Poortmans, A. Stesmans, “High open-circuitvoltage values on fine-grained thin-film polysiliconsolar cells”, J. Appl. Phys., 100, p. 063702, 2006.4.4.1.31 J. Wong, J. L. Huang, O. Kunz, Z. Ouyang, S. He,P. I. Widenborg, A. G. Aberle, M. Keevers, M. A.Green, “Anomalous temperature dependence ofdiode saturation currents in polycrystalline siliconthin-film solar cells on glass”, J. Appl. Phys., 105, p.103705, 2009.4.4.1.32 J. Wong, J. Huang. B. Eggleston, M. Green, O.Kunz, M. Keevers, R. Egan, “Lifetime limitingrecombination pathway in thin-film polycrystallinesilicon on glass solar cells”, J. Appl. Phys. V. 107,123705.4.4.1.33 J. Wong, J. Huang, M. Keevers, M. Green, “Voc-Limiting recombination mechanism in thin filmsilicon on glass solar cells”, Proc. 35th IEEE PVSC,Hawaii, USA, 2010.4.4.1.34 J. Wong, J. Huang, S. Varlamov, M. Green. R. Evans,M. Keevers, R. Egan, “Structural inhomogeneitiesin polycrystalline silicon on glass solar cells andtheir effects on device characteristics”, Progress inPhotovoltaics, 2011, DOI: 10.1002/pip.1089.4.4.1.35 M. Kittler, W. Seifert, K.W. Schroeder, “Temperaturedependent EBIC diffusion length measurementsin silicon”, Phys. Stat. Sol. (a), 2986, v. 93,pp.K191-K104.4.4.1.36 V. Kveder, M. Kittler, W. Schroter, “Recombinationactivity of contaminated dislocations in silicon:a model describing electron-beam inducedcurrent contrast behaviour”, Phys. Rev. B, 63,p. 115208 2001.4.4.1.37 J. Huang, J. Wong, M. Keevers, M. Green,“Cell performance limitation investigation ofpolycrystalline silicon thin-film solar cells on glassby transmission electron microscopy”, Proc. 25thEUPVSEC, Valencia, Spain, 2010, pp.3565-3567.4.4.1.38 J. Wong, J. Huang, S. Varlamov, M. Green, “The rolesof shallow and deep levels in the recombinationbehaviours of polycrystalline silicon solar cells”,Progress in Photovoltaics, 2011, in press.53

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTRecent improvements in independently confirmedefficiency for small area (≤ 1cm 2 ) organic solar cells(extracted from Green et al. [4.4.2.2]).Figure 4.4.2.1University Staff• Dr Ashraf Uddin• Prof. Martin Green• A/Prof. Gavin ConibeerSenior FellowDr Dirk KönigResearch FellowDr Shujuan HuangResearch AssistantXiaohan YangPostgraduate Research Students• Lara Treiber• Xiaohan Yang• Miga Jung• Xiaolei Liu• Kim Taekyun• Fred Oi• Matthew WrightUndergraduate Thesis Students• Thoo Chee Hong• Matthew Wright• Xinda Yang• Jizhong Yao• Adam Sen Ann Ong• Lihui Song• Buyi Yan• Viren NathooTaste of Research StudentSolomon Freer4.4.2.1 IntroductionA new organic photovoltaic (OPV) researchlab has been established at Level-10 inthe UNSW Chemical Science Buildingto accommodate work on organicsemiconductor materials and OPV devices.Today this lab is equipped with basicfacilities organised with support from anARC Discovery Grant, UNSW equipmentgrant, a family grant from the Green family,a Faculty grant and with the support ofthe ARC Photovoltaic Centre of Excellence.Currently, the research team consists ofthree academics, two research fellows, oneresearch assistant, 6 PhD, one M.Eng. and 8honours students.We have already produced some meaningfuldata from our research work. Two journalpapers, three conference proceedingpapers and one patent are already inpreparation. The nanoscale texture or thinfilm morphology of the donor/acceptorblends used in most OPV devices is acritical variable that can dominate boththe performance of new materials beingoptimized in the lab and efforts to movefrom laboratory-scale to factory-scaleproduction. Although OPV conversionefficiencies (up to 8.3%) have improvedsignificantly in recent years, progress inmorphology optimization still occurs largelyby trial and error, in part because much ofour basic understanding of how nanoscalemorphology affects the optoelectronicproperties of these heterogeneous organicsemiconductor films has to be inferredindirectly from macroscopic measurements.The morphology of the active film is avery important factor in producing highefficiency devices.Films spin-coated from blend solutionsundergo separation of the donor/acceptorphases. The scale of the phase separationdepends on the solvent, solubility of thematerials and parameters of the spincoatingprocess such as the spin speedand temperature. If the morphologycould be controlled on a molecularscale the efficiency of charge separationand transport could be expected to besubstantially higher. Now we are workingfor the morphology optimization oforganic films for OPV devices in respect tofundamental issues such as light trapping,control of electronic structure at filmsinterfaces, exciton dissociation and carriertransport for photovoltaic operation toachieve our goals.The detailed physical understanding of OPVdevices still lags behind their application,on account of fundamental differences inthe optoelectronic properties of organicmaterials compared to conventionalsemiconductors. This lack of understandinglimits the scope of material and devicedesign. Among those features thatdistinguish OPV devices from inorganicsemiconductor based devices are: (i) thecharges and excited states are localizedon individual molecules or molecularsegments, with the result that charge andenergy transport processes are relativelyslow; (ii) the dielectric permittivity is low,leading to stronger space charge effects;(iii) the organic semiconductor materialsare electronically disordered, dispersingthe rates of charge transfer and transportprocesses; (iv) the active layers are oftenheterogeneous, either as multi-componentfilms or because of nonuniform molecularordering; and (v) the organic semiconductoris usually not doped, thus precluding theconditions that allow charge dynamics tobe linearized in the description of devicephysics. The need for appropriate device54

(a) The schematic band energystructure of bulk-heterojunctionorganic cell used to produce theresults of Fig. 4.4.2.1; (b) (left)a schematic showing randomdistribution of donor (P3HT) andacceptor (PCBM) regions in theblended bulk-heterojunctionorganic solar cell, (right) aschematic diagram of systematicalignment of donor andacceptor layers.ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTFigure 4.4.2.2physics for organic semiconductor materialsis of particular relevance to the organic bulkheterojunction solar cell.4.4.2.2 OPV Conversion EfficiencyInterest in OPV with increasing conversion efficiencyhas grown exponentially over recent years. Thekey development in the OPV device is the bulkheterojunction cell [4.4.2.1] obtained by blendingdonor and acceptor layers, a device structure wellsuited to the associated short excitonic diffusionlengths. The workhorse for OPV research hasbecome polymer-fullerene bulk heterojunctiondevices fabricated by coating a thin layer ofPEDOT:PSS on the top of cleaned ITO substratesfollowed by a blend of PCBM/P3HT (or other donorpolymer), then by vacuum evaporation of Al andannealing. Both packaging and cell efficiency hadimproved by 2006 to the stage where independentmeasurement of cell efficiency was both feasibleand warranted, with subsequent efficiencyimprovements well documented, after reviewby a team that includes one of the present teammembers [4.4.2.2]. Using the above structure,Konarka, USA established 4.8% efficiency for a tiny0.14 cm 2 cell in July 2005. Konarka increased thisto 5.15% in December 2006 with 5.24% postedfor a 0.7 cm 2 cell in July 2007. Companies such asPlextronics have reported roughly comparableresults using similar materials with 5.4% efficiencyconfirmed in July 2007 for a 0.1 cm 2 device, 6.0% inAugust 2008 for an even smaller 0.04 cm 2 cell and2.0% as recently as 28 January 2009 for a 224 cm 2module. OPV device developer Solarmer Energyachieved 7.9% efficiency in December 2009 [4.4.2.3].In November 2010, Konarka achieved their highestOPV efficiency 8.3%. The history of OPV conversionefficiency is shown for small area (≤ 1cm 2 ) cells inFigure 4.4.2.1.4.4.2.3 Standard OPV StructuresThe fullerene based acceptor material (PCBM) isinterspersed with a donor polymer, commonlyP3HT, with line-bond diagrams for these materialsand the energy band diagram of the standard“bulk heterojunction” cell structure shown in Fig.4.4.2.2(a). From this baseline device structure, theCentre believes it has several strengths that willallow it to contribute to the ongoing cell efficiencyevolution documented in Fig. 4.4.2.1 and to thedevelopment of more stable and durable devices.The challenge in the OPV is that absorbing light inan organic donor material produces coulombicallybound excitons that require dissociation at thedonor/acceptor interface. The energy-level offsetof the heterojunction (donor/acceptor interface)is also believed to play an important role for thedissociation of bound excitons in OPV cells. Anefficient exciton separation into free charge carriersat the interface of an acceptor and donor materialswill increase the photocurrent of the solar cell[4.4.2.4]. The transport of excitons to the interfacealso limits the conversion efficiency of OPV. Thedetails of the exciton transport mechanism in OPVare still not well understood. There is a need for moresystematic study to gain in-depth understandingof the mechanisms of light absorption, excitondissociation and charge transport, particularly theirrelationships with molecular and morphologicalstructures of the materials as well as the nano-scalearchitectural design of the devices. Figure 4.4.2.2(b)shows the two examples of donor/acceptor blendingdesign (left) and systematic alignment of donorand acceptor layers for bulk-heterjunction organicsolar cells.We believe that the conversion efficiency of OPV canbe increased to over 10% for possible commercialapplications through in-depth understanding ofthe light harvesting behaviour of organic materialsand by selecting suitable materials and new devicearchitectures.4.4.2.4 Active Layer AnnealingThermal annealing is an effective method thatincreases the polymer crystallinity and improvesthe phase segregation of donor and acceptordomains. The surface morphology is also affectedby annealing as shown in Fig. 4.4.2.3. Usually,as-cast P3HT:PCBM films have a featureless and55

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORT(a) Annealing effects on theabsorbance of P3HT/PCBMblended layers on glass substrate(temperatures from 125°C to 145°Cfor 5 to 15 minutes duration). (b)Average absorbance compared fordifferent annealing temperaturesand duration. The optimumannealing conditions were foundas 125°C for 15 minutes for the bestquality layer.Figure 4.4.2.4AFM images of surface morphologyof P3HT layer on glass substrate (a)as spin coated; and (b) annealedat 140°C for 10 minutes. Afterannealing the layer structurebecome more clearly visible. Thesurface roughness is also reducedby annealing.Figure 4.4.2.3Voc and fill-factor (FF) of our OPVcells, fabricated with differentdonor type materials of P3HT.Table 4.4.2.1homogeneous surface,while annealing produces acoarser textured surface withobvious phase segregation.The majority of device studieson P3HT:PCBM blends utilizeannealing to increase deviceefficiency. This is believed tobe due to changes in boththe morphology of the activelayer whereby P3HT and PCBMself-organize into nanoscaledomains as well as the betterintermolecular alignment ofthe P3HT giving higher holemobilities.In our studies, we havefound that annealing thedevice after the applicationof the cathode improvesthe efficiency of the devicemuch more significantly thatannealing before the cathodedeposition. The reason forthis is believed to be reduced mobility of the PCBMmolecules under the increased confinement of thecathode. In our studies we have explored a rangeof anneal temperatures and find that the optimumannealing temperature depends upon the blendratio, and in general lies in the range of 120°C to150°C. The optimum anneal temperature for ourblend appears to be ~125°C (for 15 minutes) – avalue subsequently used in the remainderof the optimization experiments detailedbelow. An improved transport propertyof the active layer is also observed whichleads to a reduced series resistance and inturn a better device performance. Recently,microwave annealing has been proposedas an alternative thermal treatmentmethod [4.4.2.5].Figure 4.4.2.4 shows the absorbancespectra of the blended active layer of P3HT/PCBM, annealed at different temperature and fordifferent durations. The absorbance spectra almostremain the same under all annealing temperaturesand times. However, the intensity of the spectrachanges. As mentioned, the optimum annealingtemperature was found as 125°C for 15 minutes forthe best quality layer.We have started to fabricate our OPV cells withaverage Voc ≈ 0.6 V, FF ≈ 70 % and Jsc ≈ 10 mA/cm 2 .The average conversion efficiency is around 3 - 4%.The measured Voc of our OPV devices as a functionof light intensity is shown in Figure 4.4.2.5.We have used different types of P3HT to fabricateour OPV cells. Their corresponding Voc and FF arerecorded in Table 4.4.2.1.4.4.2.5 Light Trapping in OPVGiven the high absorption coefficient of organicmaterials, 100 nm thick organic layers are typicallysufficient to absorb most of sunlight. The diffusionlength of excitons in the organic materials istypically in the range of 5–10 nm. A partialsolution to the problem of low charge carrier andexciton mobility is the reduction of the activelayer thickness, that also implies the loss of a largefraction of the impinging light. The use of a lighttrapping systems has been actively investigated toincrease the total path length of light into the activematerial without needing to increase its physicalthickness. The concept is to induce sunlight to travela longer distance in the active layer by a multi-passpath in order to exploit most of the radiation.Many different light trapping schemes have beenproposed to enhance the quantity of light absorbedin OPV cells such as metal gratings [4.4.2.6], buriednanoelectrodes [4.4.2.7] and multireflectionstructures [4.4.2.8]. However, the requirement thatthe feature sizes of the scattering structures becomparable to the film thickness of the OPV cells,i.e., 50–300 nm, and at the same time larger than thewavelength of light to effectively alter the photonpropagation direction is intrinsically contradictory.The electrical properties can be negatively affectedby the introduction of light scattering elements indirect contact with such thin active layers: defectsand shorts may easily occur. Alternate approaches56

light intensity (Suns)100.0010.001.000.100.01DataVocMPPV_n1 fitV_n1VxFitV_0.1RshV_n2 fitV_n2ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORT0.000.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 0.55 0.60 0.65 0.70 0.75 0.80V oc (V)Open circuit voltage (Voc) of OPV devices as a functionof light intensity. The Voc increases logarithmicallywith light intensity as for conventional silicon solarcells.Figure 4.4.2.5in improving absorption efficiency have exploitedthe properties of surface plasmons (SPs). SPs areelectromagnetic surface waves confined to a metaldielectricinterface by coupling to the free electronplasma in metals [4.4.2.9]. The evanescent nature ofSPs permits the manipulation and enhancement ofoptical fields below the diffraction limit, allowingthe use of thin OPV layers without sacrificing theirabsorption potential. The integration of plasmonicstructures with OPVs has been previously examinedin the context of metallic nanoparticles, which canbe used to increase optical absorption. We have alsostarted working on the plasmonic light trappingsystem to increase the efficiency of our OPV cells.4.4.2.6 Simulation WorkThe Centre has purchased an upgrade of theGaussian Density Functional Theory (DFT) programpackage to Gaussian09. We can model largersystems in the so-called ONIOM model whichconsists of up to three model shells of differentaccuracy. The highest accuracy is limited to themolecule of interest, while the lower accuracyregions are bigger and cover the electronicenvironment of the species under investigation.This can include neighbouring molecules for chargetransfer as well as electrodes. An improved tool forcalculating infrared- and Raman-spectra is includedas well, providing an important link to experimentalobservations at the organic species. The presentstrengths include a dedicated ab-initio molecularsimulation capability (Fig. 4.4.2.7) presently rated at1080 gigaflops but with upgrading to 1.6 teraflopsplanned for the near-term. To date, this facilityhas been used for ab-initio modelling of siliconquantum dots in a dielectric environment. Theextended capability for excited state modellingshould allow the screening of new candidate OPVmaterials for both performance and durability.Other strengths include past experience with anumber of engineered vertical junction structuresand with light-trapping, not yet fully exploitedin OPV devices. The Centre has pioneered lighttrappingin both wafer-based and silicon thinfilmdevices, as well as the use of plasmonicsfor light-trapping in thin, plane-parallelphotovoltaic structures [4.4.2.10].The Centre is also particularly interested inhybrid organic/inorganic systems as a way ofimproving both performance and stability.Some work in this area has already commencedas reported below. Hybrid solar cells are a mixerof nanostructures of both organic and inorganicmaterials. They combine the unique propertiesof inorganic semiconductor nanoparticles withproperties of organic/polymeric materials.Inorganic semiconductor nanoparticles or quantumdots may have high absorption coefficients andparticle size induced tunability of the optical bandgap.Band-gap tuning in inorganic nanoparticleswith different nanoparticle sizes can be used forrealization of device architectures, such as tandemsolar cells in which the different bandgaps canbe obtained by modifying only one chemicalcompound. Thus, the organic/inorganic hybridconcept for photovoltaic solar cells is becomingattractively interesting in recent years. Thesolubility of the n-type and p-type componentsis an important parameter in the construction ofhybrid solar cells processed from solutions.4.4.2.7 Ordered Nanoparticle ArraysFor Hybrid Organic/InorganicSolar CellsExperimental work on fabrication of highlyordered arrays of nanoparticles as absorbermaterials originally for hot carrier solar cellshas been initiated as a potential means to realiseordered superlattice structures. In this work, the aimis to establish a fabrication system for depositingsequential monolayers of nanoparticles withuniform shells.We have installed a Langmuir-Blodgett (LB) systemfor fabrication of highly ordered nanoparticlemonolayers as shown in Fig. 4.4.2.8. The LBtechnique leads to the development of orderedmonolayers at an air-water interface whileexploiting the self-organization mechanism ofAbsorbance spectra of conventionalOPV cell (ITO/PEDOT:PSS/P3HT:PCBM/Al without TiOx spacerlayer and with TiOx layer.Figure 4.4.2.6Dedicated Linux 64 bit cluster forab-initio molecular simulationspresently rated at 1080 gigaflops.Figure 4.4.2.757

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTSchematic of the LB apparatus used to fabricate amonolayer of encapsulated nanoparticles.Figure 4.4.2.8Si nanoparticle encapsulated with organosilanes.Figure 4.4.2.9is very attractive for organic and inorganic hybriddevice applications. Organic/inorganic heterojunctionshave generated a great interest for theirpotential in developing hybrid devices as well asto understand organic semiconductors and theirinterfacial properties. We are also investigatingthe possibility of poly-Si/organic hybrid solarcells. Figure 4.4.2.10 shows the absorption spectraof poly-Si/PEDOT:PSS/P3HT on glass substratecompared with the absorption spectra of c-Si/PEDOT:PSS/P3HT structure.Absorption spectra of poly-Si/PEDOT:PSS/P3HT on glass substratecompared with absorption spectraof c-Si/PEDOT:PSS/P3HT structure.Figure 4.4.2.10colloidal dispersion. Compression of the monolayeris monitored via measurements of surface pressureand then controlled by a feedback loop. This uniquetechnique allows transfer of this ordered monolayeronto a wide range of solid substrates such asglass or Si wafers. By controlling the interspacingbetween adjacent particles, i.e. the shell thickness,by varying the molecular weight of capping species,we can control the periodicity of the film - leadingto new optical and electrical properties.Silicon (Si) nanoparticles are being used as corematerials. In order to control the interspacingbetween the particles, the termination of Sinanopaticles is carried out using organosilanes ofvarying alkyl chain lengths, as shown in Fig. 4.4.2.9.Progress to date is reported elsewhere [4.4.2.11].Once the assembly approach is mastered, itshould be possible to build up device structuresincorporating layers of quantum dots with thedoping in each layer individually controlled.4.4.2.8 Poly-Si/Organic Hetero-Junction CellRecently organic/inorganic hetero-junction thinfilms solar cells have gained importance for theircost-effective potential applications [4.4.2.12]. Thecombination of hydrogenated amorphous (a-Si:H)and/or poly-silicon (poly-Si:H) and organic materials4.4.2.9 References4.4.2.1 G. Yu, K. Pakbaz and A. J. Heeger, Appl. Phys. Lett.64 (1994), p3422.4.4.2.2 M.A. Green, K. Emery, Y. Hishikawa and W. Warta,“Solar Cell Efficiency Tables”, Versions 26-34,Progress in Photovoltaics, 2006-2009.4.4.2.3 http://www.solarmer.com4.4.2.4 A. Holzhey, C. Uhrich, E. Brier, E. Reinhild, P. Bauerle,K. Leo and M. Hoffmann, J. Appl. Phys. 104 (2008),p064510.4.4.2.5 K. Lee, J. Y. Kim and A. J. Heeger, SPIE OrganicPhotonic Materials and Devices VIII, San Jose, CA,January 23-26, 2006.4.4.2.6 L.S. Roman, O. Inganäs, T. Granlund, T. Nyberg,M. Svensson, M.R. Andersson, J.C. Hummelen,Advanced Materials 12 (3) (2000), p189–195.4.4.2.7 M. Niggemann, M. Glatthaar, A. Gombert, A.Hinsch, V. Wittwer, Thin Solid Films 451–452 (2004),p619–623.4.4.2.8 K. Tvingstedt, V. Andersson, F. Zhang, O. Inganas,Applied Physics Letters 91 (2007) p123514.4.4.2.9 N.C. Lindquist, A. Lesuffleur, and S. Oh, AppliedPhysics Letters 91 (2007), p253105.4.4.2.10 S. Pillai, K.R. Catchpole, T. Trupke, and M.A. Green,“Surface Plasmon Enhanced Silicon Solar Cells”,Journal of Applied Physics 101 (2007), p093105.4.4.2.11 L. Treiber, C. Bumby, S. Huang, G. Conibeer, 23rdEuropean Photovoltaic Solar Energy Conference,Valencia, Spain (2008).4.4.2.12 R. A. Halton, et al., Organic Electronics 10(2009), p388.58

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTSynthetic silicon. Portion of the PeriodicTable showing pathways to engineeringsemiconductors with 4 valence electrons/atom.Figure 4.4.3.14.4.3 Earth-Abundant PhotovoltaicDevices (CZTS)University Staff• Dr Xiaojing Hao• Prof. Martin GreenPostgraduate Research StudentXiaolei Liu4.4.3.1 Strand OutlineAll successfully commercialised non-concentratingphotovoltaic technologies to date are based onsilicon or the chalcogenides (semiconductorscontaining Group VI elements, specifically Te, Seand S). As indicated by Figure 4.4.3.1, the successfulchalcogenide materials, CdTe and CuInSe 2, can beregarded as “synthetic silicon” where the balancebetween atoms in these materials provides thesame average number of valence band electrons asin silicon.Unfortunately Cd and Se are toxic “heavy metals”while Te and In are amongst the 12 most rareelements in the Earth’s crust, factors that wouldseem to limit the long-term potential of theestablished chalcogenide technologies. However,as indicated in Figure 4.4.3.1, by investigating moredeeply into the Periodic Table, the compoundCu 2ZnSnS 4is uncovered with the same number ofvalence band electrons on average but involvingearth-abundant, non-toxic elements.Although the potential of this material is relativelyunexplored for photovoltaics, initial results havebeen promising with a group at IBM reporting 9.7%efficiency for small cells based on related materials(alloy of the above sulphide and the correspondingselenide). The appearance of the cells involved(Fig. 4.4.3.2) is quite close to the appearance ofcells made using CIGS technology (CuInSe 2plusCuGaSe 2alloy).The Centre’s work in this new strand of activitytakes a different direction from most of the presentinternational work in this area. The starting point isa viable high throughput manufacturing process.Efforts are to develop technology suitable for thedesired process rather than persevering with solutiongrowth processes that have given the best laboratoryresults to date. Materials selection would beguided by stable compositions formed in naturallyoccurring minerals such as kesterite and stannites.Another key area that is to be investigated in detailis epitaxial relationships to silicon to investigate thesuitability for tandem cell stacks, including stacksinvolving silicon as the lowermost layer.4.4.3.2 Simplified fabrication of CZTSthin film cellsThough various approaches have been appliedto fabricate CIGS solar cell, CIGS with the highestefficiency (up to 20.3% so far) have been achievedby the physical vapour deposition (PVD) approach,with the first 10% efficient cell demonstrated in1980, the first 15% cell in 1993 and the first 20% cellin 2008. As a substitute for CIGS thin film solar cells,CZTS has many similarities to CIGS as mentionedabove. As such, it is logical that the full potentialof high efficiency CZTS thin film solar cells will beextracted using the PVD approach.Though a major milestone was reached at theend of 2009 with the demonstration of 9.7% CZTSefficiency (some Se also in device) by a researchteam at IBM by a “hydrazine-based solution”approach [4.4.3.1], solution growth is believedinherently incapable of giving the material controlrequired for best-possible performance. Evidencefor this arises in the similar case of CIGS where thesame IBM team holds the record of 12.2% efficiencyfor solution-deposited CIGS (again hydrazine–based) [4.4.3.2], while in contrast the PVD approachgives 20.3% efficiency.Evaporation and sputtering are the two mostimportant PVD techniques. The disadvantage ofevaporation is that it is very difficult to control theCu evaporation source; and the lack of commerciallyavailable equipment for large area thermalevaporation. In contrast, magnetron sputtering is59

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTSEM cross-section of CZTS-type cells (Todorov et al., Adv. Mater. 22,E156, 2010).Figure 4.4.3.2a more attractive process because large area/highrate sputtering is a better developed techniquethan evaporation. In addition, sputtering hascommercially available equipment that gives gooduniformity over large areas. UNSW has considerableexperience in PVD techniques, such as theachievement of the first silicon quantum dot solarcell on quartz [4.4.3.3], a recent patent relating tocrystal Ge on Si wafer by a sputtering technique andthe polycrystalline silicon thin film solar cell on glassby an evaporation technique [4.4.3.4].Almost all of the methods that have beenattempted for CZTS fabrication involve two steps;deposition of a metal/metal-sulfide precursorand a subsequent anneal at high temperaturein a hydrogen sulphide or elemental sulphurenvironment to incorporate more sulfur into thefilm and evolve the correct phase [4.4.3.5-4.4.3.7].It is reported that the H 2S/sulphur anneal is adestructive process that can potentially introducedefects into the film due to the out-diffusion ofmetals. In addition, for large scale production ofsolar cells, a direct one-step, large area depositionprocess is preferred.Fortunately, reactive sputtering is a well suitedtechnique because of the above-addressed featureof precise control on the stoichiometry of depositedfilms over a large area with high uniformity at arelatively low cost. The experience documented inour patent of “A Method of Forming A GermaniumLayer On A Silicon Substrate And A PhotovoltaicDevice Including A Germanium Layer”, where theintroduction of H 2gas into the in-situ depositedgermanium thin film by magnetron sputteringresults in the significant improvement of theperformance of germanium thin film, appearshighly relevant.As such, the idea of one-step reactive sputteringby in-situ introducing H 2S gas into the depositedCZT (Cu 2ZnSn) film is a logical step to attempt abetter CZTS film quality. However, there have beenfew attempts to date at such a simplified one-stepprocess by introducing the sulphide into the CZTSfilm during in-situ deposition. In addition, thisprocess is not well understood and significantlymore effort is required in order to understandgrowth mechanisms involved in the reactivesputtering process. Comparison of reactivelysputtered films with 2-step sulphidisation processeswill provide the early emphasis of the Centre’s work.4.4.3.3 References4.4.3.1 T.K. Todorov, K.B. Reuter, D.B. Mitzi, AdvancedMaterials 2010, 22, E156.4.4.3.2 W. Liu, D.B. Mitzi, M. Yuan, A.J. Kellock, S.J. Chey,and O. Gunawan, Chem.Mater., 2010, 22(3), 1010-1014.4.4.3.3 I. Perez-Wurfl, X.J. Hao, A .Gentle, D-H. Kim, G.Conibeer, M.A. Green, Applied Physics Letters 95(2009), 153506.4.4.3.4 Z. Ouyang, S. Pillai, F. Beck, O. Kunz, S. Varlamov, etal., Applied Physics Letters, 2010, 96 (26), 261109.4.4.3.5 J.J. Scragg et al., Thin Solid Films, 517 (2009), pp.2481-2484.4.3.3.6 R. Schuee et al., Thin Solid Films, 517 (2009) pp.2465-2468.4.4.3.7 Hironori Katagiri et al., Thin Solid Films, 480-481(2005), pp. 426-432.60

4.5 Third Generation Strand –Advanced ConceptsUniversity Staff:• A/Prof. Gavin Conibeer (group leader)• Dr Richard Corkish• Prof. Martin GreenSenior Research Fellows:• Dr. Dirk König• Bo Zhang• Andy Hsieh• Dawei DiARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11• Zhenyu “Wayne” WanANNUAL REPORT• Craig Johnson• Sammy Lee• Haixiang ZhangResearch Fellows:• Dr Shujuan Huang• Dr. Ivan Perez-WurflLecturers:• Dr Santosh ShresthaPostdoctoral Fellows:• Dr Supriya Pillai (part time)• Dr Xiaojing “Jeana“ Hao• Dr Sangwook Park (to Apr 2010)Professional officers:• Dr Tom Puzzer (part time)• Dr Patrick Campbell (shared with ThinFilm)• Mark Griffin (shared with Thin Film)Research Associates:• Dr Didier Debuf (adjunct fellow)• Yidan Huang• Lei “Adrian” Shi (from Sept 2009)Higher Degree Students:• Lara Treiber• Pasquale Aliberti• Yong-Heng So• Robert Patterson• Binesh Puthen Veettil• Chao-Yang “Jeff” TsaoCasual Staff:James Rudd (part time)Visiting Researchers:• Dr. Yukiko Kamikawa (AIST, Tsukuba,Japan, June 2009 – Dec 2012)• Prof. Bingqing Zhou (Inner MongoliaNormal University, China, Sept 2009 –Aug 2010)• Dr. Mallar Ray (Bengal Eng. and Sci.University, May 2010)Undergraduate Students:4 th Year Thesis Projects:• Yu Feng• Tao Zhan• Chen Pan• Xi Li• Yao Yao• Linzhi Ma• Lin DongMasters Project:Chandraprasad RamachandranVisiting Practicum students:• Stephan Michard (RWTH Aachen,Germany, Aug 2009 – Feb 2010)• Dolf Timmerman (AmsterdamUniversity, Oct – Dec 2010)AbstractThere has been approximately equal work onthe two major projects in Third Generation in2010. These are the Group IV nanostructuretandem cells project - the “all-Si” tandem cell- and the Hot Carrier solar cell project, withits continuing funding from GCEP (GlobalClimate and Energy Project). There has alsobeen further work on Up-conversion and onPlasmonics.The Si nanostructure work has seen increasedunderstanding of the mechanisms fortransport and quantum confinement. Moresophisticated modelling of both has beentied more directly to improved interpretationof experimental results. This has led toestablishment of a predictive ‘equivalent circuitmodeller’ which will allow optimisation of SiQD device parameters to maximise transportand performance in the photovoltaic devices.Improved models for the understandingof doping effects in these materials havealso been established. Work on alternatematrices for Si quantum dots, in both siliconnitride and carbide, has seen development ofcomposite structures which have improvedtransport in the growth direction, whilstmaintaining quantum confinement in theplane, but which also increase the uniformityof QD sizes. Work on Ge nanostructures hasalso improved with high electrical p-typeconductivity established for Ge quantumdots and excellent pseudo single crystallinegrowth quality for Ge quantum wells in anitride matrix. Heterojunction photovoltaicdevices combining the advantages of two ofthese different material types are now beinginvestigated.Hot Carrier cells have seen very significantimprovement in demonstrated resonancein energy selective contacts using Sinanostructure layers, as well as a developmentof 2 and 3D modelling of transport in thesestructures. Modelling of Hot Carrier efficienciescontinues to get more sophisticated withapplication to real material systems such asIII-nitrides and inclusion of Auger processeswhich become significant at high carrierconcentrations. Work on absorbers hasallowed modelling of the phononic propertiesof a range of bulk materials, in conjunctionwith time resolved photoluminescencemeasurement of carrier cooling in some ofthese materials. Also modelling of coherentnanoparticle nanostructures, which emulatethe phononic properties required, hasdeveloped into direct application to structuresgrown directly. These structures include the61

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTLoss processes in a standard solarcell: (1) non-absorption of belowband gap photons; (2) latticethermalisation loss; (3) and (4)junction and contact voltage losses;(5) recombination loss.Figure 4.5.1colloidal dispersion of Si nanoparticles, whichhas seen coherent arrays successfully grown,and growth of III-V quantum dot structures withcollaborators, with modelling and characterisationindicating their usefulness as absorbers. Theparameters needed for a real hot carrier solar cellare now better defined, and candidates for itsstructure are now being modelled in more detail.This will feed into fabrication and characterisationof some candidate structures in the near future.The up-conversion project in 2009 reported theapproach of use of porous-Si as a host material foran Erbium upconverter. This has the advantageof much greater versatility in fabrication andelectrochemical doping with Er, compared toother host phosphors. Work in 2010 has furtherdemonstrated up-conversion of below band gapphotons. The ease of control of porous-Si growth atdifferent refractive indices has also been exploitedto create fully integrated multiple Distributed BraggReflectors. These allow a tuning of the forbiddenphotonic band such that regions just outsidethe bandgap (which have enhanced density ofphotonic states) lie at the 1500nm absorptionwindow of Er. This effectively concentrates lightinto this absorbance window and boosts upconversionquadratically. Enhancements of upto 80 are possible with multiple levels, with aconsequent narrowing of the wavelength rangeenhanced. Enhancements of 40 are very practical.This represents a big potential improvement for upconversionefficiency.Plasmonics has previously been applied to bothfirst generation and second generation Si cells withvery significant enhancements in absorbance ofnear band gap photons. Work in 2010 has appliedplasmonic silver nanoparticles to nanostructuredSi QD layers. Again significant enhancements ofphotoluminescence (reciprocal with absorption)have been absorbed, up to 16x for wavelengthsjust shorter than the effective band gap. Alsoapplication to the rear of the layer is seen to bebetter than the front because the reflection of thesilver particles themselves is beneficial rather thanparasitic. Just as with thin film second generationcells, light trapping by these non-texturing methodsis very important for third generation devices, whichconsist of very thin layers of material.This progress in all the main Third Generationproject areas is improving understanding andallowing optimisation of modelling, structuresand devices. Developments to come in thenext year will see significant advancement inthese areas, with excellent prospects for gooddemonstration devices.4.5.1 Third Generation PhotovoltaicsThe “Third Generation” photovoltaic approach isto achieve high efficiency whilst still using “thinfilm” second generation deposition methods. Theconcept is to do this with only a small increasein areal costs and to use abundant and non-toxicmaterials and hence reduce the cost per Watt peak[4.5.1]. Thus these “third generation” technologieswill be compatible with large scale implementationof photovoltaics. The aim is to decrease costs to wellbelow US$0.50/W, towards US$0.20/W or better, bydramatically increasing efficiencies but maintainingthe economic and environmental cost advantagesof thin film deposition techniques (see Fig. 4.1.3showing the three PV generations) [4.5.1, 4.5.2]. Toachieve such efficiency improvements such devicesaim to circumvent the Shockley-Queisser limit forsingle band gap devices that limits efficiencies tothe “Present limit” indicated in Fig. 4.1.3 of either31% or 41% (depending on concentration ratio).This requires multiple energy threshold devicessuch as the tandem or multi-colour solar cell. TheThird Generation Strand is investigating severalapproaches to achieve such multiple energythreshold device [4.5.1, 4.5.3].The two most important power loss mechanismsin single-band gap cells are the inability to absorbphotons with energy less than the band gap (1 inFig. 4.5.1), and thermalisation of photon energyexceeding the band gap, (2 in Fig. 4.5.1). Thesetwo mechanisms alone amount to the loss ofabout half of the incident solar energy in solar cellconversion to electricity. Multiple energy threshold62

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTA tandem photovoltaic cell using quantum confined QDs or QWsto engineer the band gap of the top cell and potentially also thelower cells. Short wavelength light is absorbed in the top cell andlonger wavelengths in lower cells, thus boosting the overall voltagegenerated and hence efficiency.Figure 4.5.2approaches can utilise some of this lost energy.Such approaches do not in fact disprove the validityof the Shockley-Queisser limit, rather they avoid itby the exploitation of more than one energy levelfor which the limit does not apply. The limit whichdoes apply is the thermodynamic limit shown inFig. 4.1.3, of 68.2% or 86.8% (again dependingon concentration).In the Third Generation Strand, we aim to introducemultiple energy levels by fabricating a tandem cellbased on silicon and its oxides, nitrides and carbidesusing reduced dimension silicon nanostructures toengineer the band gap of an upper cell material.We are aiming to collect photo-generated carriersbefore they thermalise in the “Hot Carrier” solar cell.Also we are investigating absorption of two belowbandgap photons to produce an electron-holepair in the cell by up-conversion in a layer behindthe Si cell using erbium doped host materials.In order to optimise the requisite properties,all these structures are likely to be thin hencemaximising absorption of light in thin structuresthrough light trapping is very important. Hence weare also investigating localised surface plasmonenhanced coupling of light into these ThirdGeneration devices.4.5.2 Si nanostructure solar cells4.5.2.1 The ‘‘all-Si’’ Tandem cellResearchers:Shujuan Huang, Ivan Perez-Wurfl, Dirk König,Tom Puzzer, Xiaojing Hao, Sangwook Park, BoZhang, Dawei Di, Yong-Heng So, Zhenyu Wan,Sammy Lee, Yidan Huang, Gavin Conibeer,Martin GreenWe are developing a material based on Si (or othergroup IV) quantum dot (QD) or quantum well(QW) nanostructures, from which we can engineera wider band gap material to be used in tandemphotovoltaic cell element(s) positioned above a thinfilm bulk Si cell, see Fig. 4.5.2.Previously we have demonstrated the ability tofabricate materials which exhibit a blue shift in theeffective band gap as the QD or QW size is reduced,using photouminescence [4.5.4] and absorption[4.5.5] data [4.5.6]. A thin film deposition of a selforganisedQD nanostructure is achieved through asputtered multi-layer of alternating Si rich materialand stoichiometric dielectric [4.5.4]. On annealingthe excess Si precipitates into small nanocrystalswhich are limited in size by the layer thickness,thus giving reasonable size uniformity, as firstdemonstrated by Zacharias [4.5.7]. Demonstrationof doping of these layers with both phosphorusand boron to create a rectifying p-n junction hasresulted in devices with a photovoltaic open circuitvoltage of 490mV [4.5.6, 4.5.7, 4.5.8].Formation of Si (or Ge or Sn) QDs through layeredthin film deposition of Si rich material whichcrystallises into uniform sized QDs on annealing.A cell based entirely of Si, or other group IVelements, and their dielectric compounds withother abundant elements (i.e. silicon oxide, nitrideor carbide) fabricated with thin film techniques, isadvantageous in terms of potential for large scalemanufacturability and in long term availability ofits constituents. Such thin film implementationimplies low temperature deposition withoutmelt processing, it hence also involves imperfectcrystallisation with high defect densities. Hencedevices must be thin to limit recombination due totheir short diffusion lengths, which in turn meansthey must have high absorption coefficients.For photovoltaic applications, nanocrystal materialsmay allow the fabrication of higher band gap solarcells that can be used as tandem cell elements ontop of normal Si cells [4.5.11, 4.5.12]. For an AM1.5solar spectrum the optimal band gap of the topcell required to maximize conversion efficiency is~1.7 to 1.8eV for a 2-cell tandem with a Si bottomcell [4.5.13]. To date, considerable work has beencompleted on the growth and characterization of Sinanocrystals embedded in oxide [4.5.7, 4.5.14] andnitride [4.5.15, 4.5.16] dielectric matrices. However,little has been reported on the experimental63

4.5.2.2.2 Modelling of quantumconfinement and tunnelling transportResearchers:Binesh Puthen-Veettil, Robert Patterson,Dirk KönigBand gap engineering for these Si nanostructurematerials requires a large number of almostidentical quantum dots (QDs), evenly distributedin a dielectric layer. As the proximity of dotsincreases, confined wavefunctions in the QDsinteract with one another to form mini-bands. Itis also seen that higher Si content in the Si richlayer gives rise to bigger QDs, due to the tendencyof QDs to merge together and form bigger dotsduring annealing [4.5.12, 4.5.20]. If the dielectriclayer is a good diffusion barrier for Si (such as thenitride layer in Si/Si 3N 4or Si 3N 4interlayer samples,see sections 4.5.2.3.1 and 4.5.2.3.2), this effect isstrongly reduced. For relatively thick layers in whichdiffusion is not constrained, the shapes of theseQDs can differ greatly from an ideal spherical shape.Calculation of confined energy in QD structures is ofinterest since their electronic and optical propertiescan be determined by determining the confinedenergy levels in the structure.Since these calculations are extremely difficult toperform using ab-initio methods due to the largememory and computation time requirements,we have developed a model to quantitativelyanalyse electronic states and their interactions ina QD array, in the framework of the effective massapproximation (EMA) calibrated by ab-initio DFTcalculations. This model is realized by solving thethree dimensional time independent discretizedSchrödinger’s equation in the EMA, employing amodified Full Multi Grid method (FMG) [4.5.21] toovercome the difficulty of computational intensity.By using exact solution methods to solve the Eigenequation for the coarsest grid, the non convergenceproblems [4.5.22] arising from oscillatoryeigenvalues are eliminated. Since initial guessesof the solution are very close to the approximatesolution, this method proves to be much faster thanother iterative methods. The flow diagram for thisprocess is shown in Fig. 4.5.5.In order to demonstrate the conformity of resultsfrom this model with experimental results,we compared it with experimental resultsfor Silicon QDs in SiO 2dielectric grown usingsputtering methods and characterized usingphotoluminescence (PL) [4.5.20]. Figure 4.5.6 showsthe dependency of confined energy on the sizeof QDs for Si/SiO 2QD structures. Confined energyincreases exponentially as QD size decreases.Comparison shows that for larger QDs experimentaland theoretical values are similar, but for smallerdots the experimental results for the confinedenergy shows a lower degree of confinementthan the theoretical result. This difference can beattributed to interface defects between Si and SiO 2and to defects within the SiO 2matrix but near theinterface. For QDs below a diameter of about 4 nmthese interface effects have a larger influence on theconfined energy level. This is because the groundstate electron density, which is concentrated atthe centre of the QD, is closer to these interfacedefects for the smaller QDs and hence morestrongly influenced.As an extreme example of irregularlyshaped QD, we have investigatedelectronic states in aribitrarilyshaped QDs. Here we have used the“horseshoe” morphology [4.5.24],formed by lateral growth of Si QDinside SiO 2dielectric shown inFigure 4.5.7 (left). These states arefound to be very different from theexpected electronic energy levelsthat are associated with sphericalQDs as shown in Fig. 4.5.7 (right).The degeneracy in spherical QDs (3in first excited level and 5 in secondexcited level) is lost due to the asymmetric shape.The deviation from the expected confinement levelbecomes significant at higher confined levels.This thus indicates the importance of knowingthe regime of QD growth relevant for the materialconditions. As shown in the TEM tomography ofSection 4.5.2.2.3, under appropriate conditions ofSi density and annealing, a spherical morphologycan be maintained and QD merging avoided,thus giving greater control over the confinedenergy levels.ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTSimulation results – multigridmethod and Density FunctionalTheory (DFT) calculations [4.5.23]compared with the experimentalresults for Silicon QDs inSiO 2dielectric.Figure 4.5.6First three modes of electronicwave-functions in a “horseshoe”shaped QD with high surface tovolume ratio (far left). The groundstate is located near the centre ofthe dot and is approximately in aspherical shape with zero nodes.The number of nodes increaseswith the increase in the energylevel. Energy levels associated witha horseshoe shaped QD comparedwith that of a 6nm diameterspherical QD (left). The twomorphologies give energies whichare most similar for their groundstates, with the horseshoe shapegiving significantly lower energiesthan spherical for higher confinedlevels. The degeneracy in energylevels is lost in the horseshoeshaped dot.FigurE 4.5.765

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTEnergy filtered TEM tomography images of Si QDsin a SiO 2matrix formed as multilayers (left), andtop view of one Si QD layer (right).Figure 4.5.8The spectra fittings of the R and T data. The insetshows the corresponding employed optical model inthe fitting of a 20 bilayer structure composted of 5 nmSRN / 3 nm Si 3N 4.Figure 4.5.9Real (ε1) and imaginary (ε2) partsof the complex dielectric functionof Si QDs samples with various sizesobtained from spectral fittings.Figure 4.5.104.5.2.2.3 3D EF-TEM tomography on SiQDs embedded in oxide matrix formedas multilayers (Collaboration withCornell University)Researchers:Xiaojing Hao, Shujuan Huang, Lena FittingKourkoutis (Cornell University), IvanPerez‐Wurfl, Tom PuzzerA significant challenge to the design of nano-scalematerials and devices arises from the difficultyof characterising complex three-dimensionalstructures on small length scales, which isnonetheless critical in understanding the materialperformance. Whilst anumber of techniques, suchas conventional transmissionelectron microscopy,provide sufficient twodimensional(2-D) resolution,they have insufficientdepth sensitivity to obtaininternal three-dimensional(3-D) structure. Electrontomography can be appliedto obtain such informationby using high-angle darkfieldscanning transmissionelectron microscopic(STEM) tomography andenergy filtered transmissionelectron microscopy (EFTEM)tomography. We have appliedthis EFTEM tomography toinvestigate Si QDs embeddedin an oxide matrix formedas multilayers, in order tobetter understand the dotsize, shape and distribution for further optimizationof Si QD device performance. Fig. 4.5.8 illustratesthe ability of electron tomography to reconstructthe morphology and internal structure of a Si QDnanostructure. Fig. 4.5.8 (left) shows the aligned“on-axis” reconstructed EFTEM tomography, andFig. 4.5.8 (right) shows a top-view of the Si QD sizeand shape distribution in one Si QD layer. With suchinformation, we can confirm that inside the Si QDlayer, most of the Si QDs are reasonably sphericaland well separated from each other. This indicatesthat individual QDs have not merged to form largerparticles under these conditions. This is importantin designing the optimum concentrations togive minimum QD spacing whilst maintainingindividual QDs.4.5.2.3. Different materials forQD nanostructuresAs described in section 4.5.2.2.1, different matricesare useful to modify both the tunnelling probabilitybetween adjacent quantum dots and the energylevels in the quantum dots themselves. Thisapproach has been carried out for both Si QDs inSiN xand in SiC. It is also possible to have asymmetricsuperstructures in which quantum confinementand hence enhanced band gap is maintained inthe plane, but in which tunnelling probability ismaximised in the direction normal to the plane suchthat transport of carriers to contacts is maximised.Interlayers use for such asymmetric structurescan also act as diffusion barriers enhancing sizeuniformity of QDs.Alternative group IV materials such as Ge offer thepossibility of lower temperature precipitation ofnanostructures and the potential for band gapslower than that of silicon should these be requiredfor tandem cell elements under a silicon cell. Wehave investigated both QDs and QWs of Ge in eitherSiO 2or SiN xmatrices.66

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORT(a) Normalized PL spectra of Si QDs samples (S3-S5)with various sizes. (b) Comparison of band gapexpansion (ΔE g) of Si QD as a function of QD size.Figure 4.5.114.5.2.3.1 Silicon QDs embedded in siliconnitride matrixResearchers:Yong-Heng So, Shujuan Huang(i) Size-dependent optical properties of SiQDs in Si-rich nitride/Si 3N 4superlatticeAn approach similar to spectroscopic ellipsometryanalysis has been proposed to determine the sizedependent optical properties of silicon quantumdots (Si QDs) in Si-rich nitride/silicon nitridesuperlattice structure (SRN/Si 3N 4-SL) [4.5.26]. Theoptical properties of Si QDs are modelled usingthe Tauc-Lorentz (TL) model and Bruggemaneffective medium approximation that can yield theenergy bandgap of the Si QDs based on spectralfitting of the reflection (R) and transmission (T) ofcharacterized samples, as shown in Fig. 4.5.9. A fourphaseoptical model as shown in the inset of Fig.4.5.9 was employed for the spectral fitting.Figure 4.5.10 shows that the dielectric functionsof Si QDs are strongly size dependent. Sample S3,S4 and S5 correspond to SRN/Si 3N 4thicknessesof 3/3 nm, 4/3 nm and 5/3 nm, respectively. Thesuppressed imaginary dielectric function of Si QDsexhibit a broad peak centred between transitionenergies E 1and E 2of bulk crystalline Si and whichblue shift towards E 2as the QD size reduces.Figure 4.5.11(b) shows that the band gapexpansion indicated by the TL model when thesize of Si QD reduces is in good agreement withPL measurements (Fig. 4.5.11(a)). The bandgapexpansion with the reduction of Si QD size is wellsupported by the ab-initio calculations of confinedenergy levels from [4.5.25].(ii) n-type conductivity of nanostructuredthin film composed of antimony-doped Sinanocrystals in silicon nitride matrixHighly conductive thin films composed of antimony(Sb)-doped Si nanocrystals (Si-NCs) embeddedin Si 3N 4matrix were prepared by a co-sputteringtechnique. Results from structural characterizationssuggest that doping with Sb concentration of 0.54at. % has negligible affect on the crystallizationproperties of Si-rich nitride (SRN) films. The X-rayPhotoelectron (XPS) data of Fig. 4.5.12(a) and (b)show the Si 2p and Sb 3d 3/2spectra of the undopedand Sb-doped SRN films. The slight asymmetric Sb3d 3/2peak was analyzed using a curve-fitting routineand could be decomposed into two components,δ 1(537.0 eV) and δ 2(537.8 eV) as shown in Fig.4.5.12(c). The two components δ 1and δ 2(~0.8 eVapart) can be assigned to neutral Sb and positivelycharged donors, respectively.Thus, the appearance of δ 2suggests the presenceof Sb-Si bonds which implies that Sb stoms wereeither incorporated within the Si-NCs or located atthe interface between the NCs and Si 3N 4matrix.XPS spectra of (a) Si 2p and (b) Sb3d 3/2for the undoped and 0.54at.% Sb-doped Si-NC. (c) Fittingof Sb 3d 3/2spectrum by two peaksseparated by ~0.8 eV in energy (δ 1and δ 2).Figure 4.5.1267

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORT(a) Conductivities measuredthroughout the entire voltagerange. (b) ln(σ) vs 1/T 1/2 plotfrom 120 to 220 K, for the 0.54at.% Sb-doped sample. Thesolid line is least square fitto the data. (c) Temperaturedependentconductivities forthe undoped and 0.54 at.%Sb-doped samples.Figure 4.5.13The temperature dependent conductivity dataof Fig. 4.5.13(a) indicate that the conductivity of0.54 at.% Sb-doped SRN film (~2.8×10 -2 S/cm)is six orders of magnitude higher than that forundoped material (~7.3×10 -8 S/cm), which could beattributed to an increase in carrier concentration.Furthermore, n-type electrical behaviour withcarrier concentration of 2.4×10 16 cm -3 and mobilityof 2.94 cm 2 /V s in the doped films as observed fromHall measurements was attributed to free carriergeneration due to the effective Sb doping.The temperature-dependent conductivities ofboth undoped and Sb-doped samples are shownin Fig. 4.5.13(b). As can be seen, doping of theSi-NCs strongly influences the electronic transportproperties of the films. Arrhenius-like temperaturedependence is observed in the T range between220 and 320 K, attributable to thermally activatedconduction. From the slope of the Arrhenius plot,we found that the activation energy E Adecreases to0.182 eV for the Sb-doped Si-NCs film, suggestingeffective n-type doping of the Si-NCs. Nevertheless,the extracted E Ais much larger than that of bulk Siwith the same dopant. One possible explanation forthe large E Aobserved would be the deeper donorlevel expected in Si-NCs. Also, it is possible that thenumber of free carriers available for conductionmay be limited due to a trap density distributedwithin the bandgap that is comparable to thedoping density.Interestingly, the conductivity for the Sb-dopedsample at T < 220 K deviates from Arrheniusbehavior and is best described by the expression, σ= σ 0exp [-(T 0/T) 1/2 ] as illustrated in Fig. 4.5.13(c). Thecharge transport mechanism can be explained wellby a percolation-hopping model.4.5.2.3.2 Si QDs in SiO 2/Si 3N 4hybrid matrixReserachers:Dawei Di, Ivan Perez-Wurfl, Gavin ConibeerTo improve the current transport properties in thevertical direction and to obtain better size control ofSi quantum dots, we proposed a new design basedon Si QDs embedded in a SiO 2/Si 3N 4hybrid matrix[4.5.27, 4.5.28]. By replacing the SiO 2tunnel barrierswith the Si 3N 4layers, the new material managesto constrain the growth of doped Si quantumdots effectively and enhances the apparent bandgap. Also electrical characterisation on Si QD/c-Si hetero-interface test structures indicates thenew material possesses improved vertical carriertransport properties.Doped and undoped samples with differentbarrier dielectrics (SiO 2and Si 3N 4) are compared.Samples were fabricated using the co-sputteringmultilayer technique with layers containing excessSi consisting of SRO for both materials, with an SROthickness of 4 nm. Samples were either undoped orco-sputtered with either P 2O 5or B for n- and p-typedoping respectively. After annealing at 1100⁰Cthe crystalline properties were characterised byX-ray diffraction (XRD), as shown in Fig. 4.5.14. Forsamples with Si 3N 4barriers, the measured Si NCsizes are 4.3 nm when undoped, 3.5 nm when Bdoped and 5.0 nm when P 2O 5doped. For sampleswith SiO 2barriers, grain sizes are 7.7 nm whenundoped, 7.2 nm when B doped, and 15 nm whenP 2O 5doped. The Si NC sizes vary significantlyalthough all samples have the same silicon-tooxygenratio in the SRO layers and the same asdepositedlayer thickness.It can be observed from the experiment thatsamples with Si 3N 4barriers contain generallysmaller Si QDs than samples with SiO 2barriers.This is because Si 3N 4, a denser and stiffer materialthan SiO 2, acts as a better diffusion barrier thanSiO 2[4.5.27]. It could also be thermodynamicallyrelated to the interface free energy between68

Intensity (a.u.)100009000800070006000500040003000200010000(111)(220)Samples with SiO2 barriers(311)P2O5 dopedundopedB doped15 20 25 30 35 40 45 50 55 60 652 Theta (degrees)ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORT10000Samples with Si3N4 barriers9000Intensity (a.u.)800070006000500040003000200010000(111)(220)(311)P2O5 dopedundopedB doped15 20 25 30 35 40 45 50 55 60 652 Theta (degrees)XRD patterns of (a) samples with SiO 2barriers, (b)samples with Si 3N 4barriers.Figure 4.5.141.0E+02B doped samples1.0E+01P2O5 doped samplesCurrent density (A/cm 2 )1.0E+011.0E+001.0E-01nitride barrieroxide barrierCurrent density (A/cm 2 )1.0E+001.0E-01nitride barrieroxide barrier1.0E-020.0 0.2 0.4 0.6 0.8 1.0 1.2Bias voltage (V)1.0E-020.0 0.2 0.4 0.6 0.8 1.0 1.2Bias voltage (V)the Si NC and the matrix dielectrics during thegrowth of the Si NCs. As a result, Si 3N 4is able toconstrain the growth of Si NCs very effectively,within ±1 nm of the intended diameter (i.e. 4 nm,the deposition thickness of SRO). In contrast, SiNCs with SiO 2barriers are significantly larger thanthe as-deposited thickness of SRO - over 180%larger for undoped and B doped and 375% for P 2O 5doped. Hence, it can be noted that samples withphosphorus (P 2O 5) dopants tend to form larger sizedSi NCs than undoped samples, while introductionof boron doping tends to suppress this size increaseand slightly reduces NC size. This phenomenon ismore pronounced for samples with SiO 2barriers(Fig. 4.5.14 (a)) than those with Si 3N 4barriers (Fig.4.5.14 (b)).Si 3N 4barriers have lower band gaps (~5.3 eV)than SiO 2(~9 eV) therefore they should be moretransparent to charge carrier transport becauseof the larger tunnelling probability for electronsand holes. To verify this hypothesis in practice,we compared the current transport of SiO 2barrier structures with Si 3N 4barrier structuresvia the measurement of vertical currents under abias voltage.The current-voltage characteristics of these teststructures are shown in Fig. 4.5.15. It should benoted that these samples are not PV devices.Therefore, the dark currents measured are primarilyaffected by the materials’ conductivities and arenot related to carrier recombination. It can beclearly seen from the measurement that for Bdoped materials, Si 3N 4barriers result in a currentenhancement of approximately an order ofmagnitude over those with SiO 2barriers (Fig. 4.5.15(a)). This can be qualitatively explained by [4.5.19]:(4.5.1)where T eis the tunnelling probability betweenquantum dots, d is the barrier thickness or theseparation between dots, m * is the effective mass ofelectron, ħ is the reduced Plank constant, ΔE is theenergy difference between the conduction bandedge of the barrier material and the conductionband edge of the confined QDs.Although the difference in Si NC sizes would play animportant role in the tunnelling event, the variationof ΔE is still the dominant factor for B doping. Onthe other hand, for the P 2O 5doping, samples withSi 3N 4and SiO 2barriers have similar currents (Fig.4.5.15 (b)). These apparently contradictory resultscan be explained by referring back to the XRDresults. The P 2O 5doped sample with SiO 2barrierscontains Si NCs with an average size of 15 nm, thissuggests that a single Si NC is physically penetratingthree thin layers on average (inset of Fig. 4.5.15 (b)).Vertical current densitymeasurements on (a) B dopedsamples, (b) P 2O 5doped samples.Insets show how barrier type andNC dimension influence the verticalcurrent transport process (showingdominant effect only in each case,not to scale).Figure 4.5.1569

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORT(a) Raman spectra of SRC samples after RTA and (b) thesame samples with additional furnace annealing. (BulkSi wafer peak at 520.5 cm -1 for reference.)Figure 4.5.16HRTEM image of annealed 2.0nmUT-Si 3N 4sample. The maintenanceof a layered structure and thepresence of Si-QDs can be seen.Figure 4.5.17XRD analysis and correspondinggrain sizes from the Scherrerformula for all the Si3N4 barrierthicknesses. Insert: Si (111) and SiC(111) peaks of each sample.Figure 4.5.18This greatly reduces the number of barrier layersthat the carriers need to tunnel through. This sizeeffect becomes so significant that it effectivelycompetes with and largely cancels the increasedconductivity due to lower ΔE for the Si 3N 4barriermaterial.This result is very promising for both QD sizecontrol and increased vertical currnt transport inthese nitide barrier interlayer Si QD materials.4.5.2.3.3 Silicon QD nanocrystalsembedded in silicon carbide matrixResearchers:Zhenyu Wan, Shujuan Huang, Gavin ConibeerSilicon carbide (SiC) has a lower barrier heightthan either Si 3N 4or SiO 2. Therefore it seems likelythat it should offer a higher tunnelling probabilitybetween QDs and hence a higher conductivitymatrix than either. In previous work, we havedemonstrated that both Si and SiC have beencrystallised by high temperature annealing of asingle thick Si-rich SiC (SRC) layer or of a Si 1-xC x/SiC multilayer structure [4.5.28, 4.5.29]. Webelieve that the formation of β-SiC nanocrystalsmay hinder the formation of Si QDs. Also theymay cause current leakage via the SiC grainboundary traps to increase the shunt current inthe solar cell. Therefore, the work in 2009 focusedon studying the mechanism of the crystallisationof SiC and optimising the materials by comparingthe annealing methods of rapid thermalannealing (RTA) and conventional furnaceannealing. We found that RTA annealed samplesrevealed a better degree of crystallisationof Si nanocrystals with a smaller residue ofamorphous Si [4.5.30]. The work in 2010 has been tofurther improve the structure by reducing the stresscaused by RTA and by replacing SiC barrier layerswith thin silicon nitride layers in order to suppressthe growth of SiC crystals.(i) Study of Rta induced stressTwo different annealing processes have beenapplied on all SRC amorphous samples: furnaceannealing at 1100 o C/1hr and RTA at 1100 o C/30sec. After annealing, Si and SiC nano-crystals wereclearly observed in TEM and XRD. The peak positionsof Raman spectra are down-shifted to lowerwavenumbers for RTA samples (Δ=5.4cm -1 frombulk Si) as compared to furnace annealed samples(Δ=2cm -1 from bulk Si), indicating greater tensilestress in the Si nano-crystals, as shown in Fig. 4.5.16(a). This is because of a different thermal expansioncoefficient for Si and SiC crystals and the fasttemperature ramping rate in RTA. In order to releasethe stress in RTA samples an additional annealingprocess was carried out. Raman analysis indicatesthat after an additional furnace anneal at 1100°C for30 min, all samples with different Si concentrationscould release most of their residual stress, as shownin Fig. 4.5.16 (b). However, an additional RTA couldnot release residual stress in low Si concentrationsamples due to insufficient duration [4.5.31].(ii) Si QDs embedded in SiC matrix withSi 3N 4barrier layersSi 3N 4is also considered preferable to SiO 2in termof carrier transport. It has been proven as a gooddiffusion barrier to suppress inter-diffusion betweensilicon rich layers in multilayer structures duringannealing [4.5.27]. In 2010, we have successfullyintroduced ultra-thin Si 3N 4(UT-SiN) barriers (0.2nm-2.0nm) into a Si-NC in SiC matrix structure usingsputtering followed by RTA. Crystallisation of theSiC matrix has been greatly suppressed and a clearlayered superlattice structure can be observed asshown in the TEM of Fig. 4.5.17. The nanocrystalsize of all samples was calculated using XRD peakanalysis to quantitatively investigate the optimumthickness of the Si 3N 4barrier layer required toproduce confined crystalline Si-QDs, as shown inFig. 4.5.18. It is seen that as the Si 3N 4barrier layerthickness increases, the β-SiC peaks decrease, suchthat β-SiC-NCs almost disappear when the Si 3N 4barrier thickness is over 0.8nm.70

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTAbsorption edge fitting for UT-SiN2.0 and UT-SiN0.2samples respectively, (insert: band gap energy fordifferent Si 3N 4thicknesses).Figure 4.5.19Temperature-depentent conductitity of different Si 3N 4barrier thickness.Figure 4.5.20Optical characterization has been performedto investigate quantum confinement effects fordifferent nanocrystal sizes. Optical analysis reveals ablue shift in the strong absorption edge, consistentwith quantum confinement effects in small Si NCsmatching qualitatively calculated results usingeffective mass theory from other groups, as shownin Fig. 4.5.19.Temperature dependent I-V measurements werecarried out to investigate the carrier transportmechanism. The conductivity is best describedby the expression, σ = σ 0exp [-(T 0/T) 1/2 ] whenthe Si 3N 4barrier thickness is greater than 0.5 nm,as illustrated in Fig. 4.5.20. The charge transportmechanism can be explained well by thepercolation-hopping model, similar to the resultfound in Section 4.5.2.3.1 for Sb doped SiQDs inSi 3N 4. Finally, we conclude that in future, a 0.5-0.8nm thick layer of Si 3N 4barrier layer would beoptimum to achieve both good Si-NC confinementand minimal film resistivity for a candidate materialfor photovoltaic applications.4.5.2.3.4.1 Quantum Dots in SiO 2Researchers:Santosh Shrestha, Bo Zhang, Pasquale Aliberti,Gavin Conibeer(i) Fabrication of Germaniumquantum dotsThe current work builds on earlier work on Genanocrystals carried out in our group [4.5.32]. Thenew work uses a different approach to annealing ofthe samples which gives a wider range of controlof Ge NC formation. Ge NC samples have beenfabricated by RF magnetron sputtering using acombination of a combination target comprisinga fused quartz disc and high purity Ge strips.Sputtering of the Ge-rich oxide (GeRO) layer isachieved by sputtering with just an Ar background,Cross-section HRTEM image of atypical sample containing GeROlayers between GeO 2/SiO 2layersfollowing annealing.Figure 4.5.214.5.2.3.4 Germanium NanostructuresAlthough the main focus has been on Sinanocrystals during the last decade, other group IVnanocrystals have also been studied. Since Ge hassmaller electron and hole effective masses and alarger dielectric constant than Si, the excitonic Bohrradius of bulk Ge is larger than that of Si. This leadsto a more prominent quantum confinement effectin Ge NCs. Furthermore the lower melting point ofGe at 938.3 o C implies that Ge NCs should be ableto form at lower temperatures than Si NCs. This isindeed a significant advantage both for processingcompatibility and for processing costs, althoughset against this are the greater cost and lowerabundance of Ge as compared to Si.whereas sputtering of the oxide layers, GeO 2/SiO 2, is performed by reactive sputtering with O 2.These layers are alternately deposited to obtainthe desired multilayered structure. A thick oxidecapping layer is deposited to prevent oxidation ofthe GeRO layers during subsequent annealing. Postdeposition annealing of the samples was performedunder a low pressure (in-situ in the growthchamber) during which Ge NCs are formed. Results(a) Raman spectra and (b) GIXRDimage of the as-deposited andannealed films.Figure 4.5.2271

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTRaman spectra of the samples annealed fordifferent durations.Figure 4.5.23Raman spectra of the samples annealed atdifferent temperatures.Figure 4.5.24Room temperature PL of amultilayer film containing Ge NCs inGeO 2/SiO 2.Figure 4.5.25Intensity (arb. units)Growth Temp.III 445 o C420 o C400 o CII 380 o C350 o C320 o CI RT150 200 250 300 350 400R aman S hift (cm -1 )Raman spectra from Ge-NCs insingle layer samples grown atdifferent substrate temperatures.Figure 4.5.26from structural characterization using TEM,Raman, XRD and PL have been obtained.Figure 4.5.21 shows a high resolution TEMimage of a section of a typical samplecontaining GeRO layers between GeO 2/SiO 2. The sample was annealed at 650°C.Ge nanocrystals of about 5nm diameterare evident. Although there is a slightvariation, the NCs appear to be sphericalin shape. The clear lattice fringes observedin the TEM image give direct evidenceof the formation of the Ge NCs. It shouldbe noted that the distance between thelattice fringes is 3.3 Å, which is consistentwith the lattice spacing of the {111} planes of the Gediamond structure.Fig. 4.5.22 (a) shows Raman spectra of the asdepositedand the annealed samples, includingthose for bulk Ge as a reference. For these samples,the sputtering times for each GeRO layer andGeO 2/SiO 2layer were 8 minutes and 6 minutes,respectively, and the post depositionannealing was performed at 650°C for 40minutes. A broad hump at around 270cm -1 is observed in the spectra of theas-deposited film which is attributed tothe non-crystalline Ge phase. But in theannealed film, it is replaced by a sharppeak at 300.5 cm -1 , which is very close tothe Ge-Ge optical phonon mode for bulkGe (300.2 cm -1 ), indicating the formationof Ge NCs with good crystallinity. Peakbroadening and an asymmetric shoulderon the lower frequency side can beinterpreted by the model of the opticalphonon confinement effect in nanocrystals[4.5.33]. However, it seems that in our Ge system thehigh frequency side of the Raman peak does notshow a shift to lower phonon frequencies which isusually the case for NCs.Figure 4.5.22 (b) shows GIXRD patterns for thissample. In the as-deposited film there are noobvious peaks but only two broad bands at around2θ = 26° and 2θ = 49°. After annealing, the sampleshows three sharp peaks at 27.12°, 45.13° o and53.21°, corresponding to the groups of planes {111},{220} and {311} of crystalline Ge, respectively. Thisobservation confirms good crystallinity of the Gephase in the film and agrees well with our Ramanresults. The average size of the Ge NCs, calculatedfrom the {111} peak broadening using the Schererequation, is about 4.5 nm. This value is slightlysmaller than that estimated from the HRTEM image.The difference in Ge NCs sizes obtained from thesetwo methods may possibly be due to spatial nonuniformityof the Ge NCs size; with TEM probinga much smaller sample region compared to XRDmeasurement. In addition, the penetration depthof the incident X-rays is larger than the thicknessof our film, thus information obtained from XRD isaveraged throughout the whole film.The crystallisation of Ge NCs with annealingduration has been investigated with Ramanspectroscopy. Fig. 4.5.23 shows Raman spectrafor identical multilayer samples with each GeROand GeO 2/SiO 2layer deposited for 6 minutes. Thesamples were annealed at 685°C for differentdurations as indicated in the diagram. Thecrystallization of Ge is found to take place withinthe first few minutes of annealing. However, thenoticeable broad hump, which is attributed to smallNCs, suggests an early stage of the crystallizationprocess at this annealing duration. As annealingduration increases, Raman peaks become sharperand narrower indicating an increase in Gecrystallinity. It is also observed that the peaks shownegligible difference for annealing durations longerthan 10 minutes.The average size of NCs, as calculated from XRDdata, increases with annealing duration. This isconsistent with the Raman results and agreeswell with the growth dynamics by diffusion ofneighbouring Ge atoms. Most importantly, theincrease in the NCs size begins to level off after 15minutes of annealing in this particular case. This issimilar to the size confinement effect observed forthe growth of Si NCs in SiO 2matrix [4.5.7, 4.5.34].In our superlattice structure, Ge NCs are confinedwithin the GeRO layers between GeO 2/SiO 2layers,72

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTA High resolution TEM image ofmultilayered Ge NCs in SiO 2matrix.The sample was grown at 380°C.Figure 4.5.27which work as barrier layers to the crystal growth.Hence, Ge tends to precipitate with a diameterapproximately equal to the thickness of the GeROlayers. This control mechanism is more significantfor thin layers with thickness of a few nanometers,within which there is 2D rather than a 3D growth.Figure 4.5.24 shows the Raman spectra of thesamples annealed at different temperatures for 40minutes, which is enough to complete the growthprocess as discussed above. As expected, the smallnanocrystalline asymmetric hump is reduced withthe increase in annealing temperature within therange of temperatures studied here, just as inthe case of annealing durations. The results alsoillustrate the formation of Ge NCs in SiO 2matrix at atemperature as low as 620 °C.A sample with the identical process sequence asthose for structural characterizations was used forphotoluminescence measurements, except thatthe film was deposited on quartz. Fig. 4.5.25 showsresults of room-temperature PL measurement.The PL spectrum consists of a single broad bandcentred at 1.77eV (corresponding to a wavelengthof 700 nm) which can be fitted with three Gaussiandistributions. Measurement on GeO 2/SiO 2filmdeposited on quartz under similar conditionsdid not show any observable PL in the range ofwavelengths concerned. Thus the PL signal from themultilayer sample can be attributed to Ge NCs in theGeRO layers. Furthermore, a blue shift of PL energywith NCs size has been shown in previous work[4.5.32]. We tentatively consider that this is due tothe band gap increase induced by the quantumconfinement effect in Ge NCs [4.5.35].(ii) Low temperature growth ofGermanium quantum dotsIn this section, a low temperature growth of Genanocrystals is discussed. This is advantageousbecause of the potential for reduced processingcost and suitability for cheaper substrates suchas soda-lime glass. Samples were grown with RFmagnetron sputtering as described earlier, exceptin this case the substrate was heated during the filmgrowth and no post annealing process was used[4.5.36].Figure 4.5.26 shows Raman spectra for a single layerGeRO film deposited at different temperatures asindicated in the diagram. For clarity, the graph hasbeen divided into three different regimes. In regimeI, the samples grown below 350°C are shown.In this case, the spectra show no Raman peaksrelated to nanocrystalline Ge, instead broad bandscentred around 280 cm -1 are observed. The bandsare shifted toward higher frequency by about 10cm -1 compared with that of amorphous Ge (α-Ge)located at around 270 cm -1 . This indicates that α-Gecoexists with very small particles (1~2 nm) [4.5.37].For the sample grown at T g= 320°C, the observationof a hump at ~ 280 cm -1 as well as a shoulder at ~298 cm -1 implies an increase in the number andsize of small particles. The rapid growth of Ge NCsis found to occur in Regime II, in which samplesgrown at temperatures between 350°C and 400°Care shown. The Raman peak corresponding to theTO phonon mode of the crystalline Ge (c-Ge), near300.4 cm -1 , appears at T g= 350°C and becomessharper with increase in the growth temperature.This range of temperature is very close to the onsettemperature for Ge crystallisation reported in[4.5.38]. However, when the growth temperature T greaches 420°C (Regime III), a drastic degradation ofGe-Ge peak intensity and shape is observed. Thesephenomena indicate the absence of Ge crystallinityin this temperature range.Based on the Raman results discussed above, it isexpected that the growth temperature window isbetween 350 o C and 420 o C, in which Ge atoms orclusters can accumulate and eventually grow intohighly crystallised nanocrystals. These temperaturesare much lower than the usual post-depositionannealing temperatures used for Ge NCs fabrication.A HRTEM image of a multilayered sample consistingof alternate layers of GeRO and GeO 2/SiO 2films isshown in Fig. 4.5.27. In this case the sample wasgrown at T g= 380 o C without further post-depositionthermal treatment. The TEM image shows close tospherical Ge NCs with fairly uniform size, separated73

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTXRD spectra of Ge QDs of various Ge content in Ge-rich nitride layer with a fixed substrate temperature of 600 ºC (left) andTEM image of 40 vol% Ge self-assembled at 600 ºC (right).Figure 4.5.30encouraging as a starting point for using Ge-NC intrue photovoltaic applications.4.5.2.3.4.2 Germanium QW and QDnanostructures embedded in siliconnitride matrixResearchers:Sammy Lee, Jian Chen, Shujuan Huang,Martin GreenAs discussed above in section 4.5.2.3.1, a siliconnitride matrix (Si 3N 4) should increase tunnelingprobability and hence increase conductivity of Genanostructure materials. In addition a nitride willstrongly suppress the possibility of oxidation of Ge.Ge QDs and QWs were fabricated by co-sputteringfollowed by post-annealing and self-organisation bysubstrate heating.(i) Ge QDs embedded in siliconnitride matrixUsing a very similar approach to our previous workon Ge QDs in oxide and Sn QDs in nitride matrices,alternating layers of germanium-rich nitride (GRN)layers and Si 3N 4layer were alternately co-sputteredon a heated substrate of either a silicon waferor quartz slide. The Ge QDs crystallise due tominimisation of surface energy and the Si 3N 4layerdeposited on top of the GRN layer truncates crystalgrowth. The temperature of the substrate and theGe content in GRN layers were the dominant factorsin controlling the size of the NCs. The structure ofthe NCs was studied by GIXRD, Raman spectroscopyand TEM, and the optical properties were studiedwith transmittance and reflectance measurementsanalysed using a Tauc-plot.Figure 4.5.30 shows the XRD spectra of samples ofvarious Ge content in GRN. The Bragg peaks of {111},{220} and {311} planes sharpen with increasing Gecontent, indicating the increase of the nanocrystalsizes. Using the Scherrer equation the size of thenanocrystals was estimated as shown in Table 1,with sizes in agreement with the NC size observedin TEM images.Figure 4.5.31 summarises the band gap engineeringtrend with the size of Ge QDs. The absorption edgeshows a tendency to shift to high photon energywhen Ge nanocrystal size decreases. The absorptioncoefficient was calculated by a simplified equation;where T and R are the transmission and reflectionand d is the thickness of the film. Withdecrease in Ge content in the GRN layer,and hence in the diameter of the QDs, theabsorption edge appears to increase by upto ~1.0 eV, as estimated with the Tauc-plotmethod. However, these data should beconsidered as indicative of the trend onlyas there is error in quantitative values forthe shift in absorption edge due touncertainty in extrapolating back to theenergy axis.(ii) Ge QWs with silicon nitridebarriersWith a similar approach to the Ge QD fabrication,Ge QWs were fabricated by co-sputtering ofalternating GRN and Si 3N 4layers and followed byfurnace annealing or RTA. The Ge content in GRNwas higher in this case in order to give continuouslayers. It was varied between 60vol% to 100vol%and the annealing temperature in N 2was also variedfrom 600ºC to 900ºC for 1 hour. The structural andoptical properties were studied by GIXRD, Ramanspectroscopy, TEM and transmission and reflectancemeasurements.Fig. 4.5.32 shows the continuous Ge layer withits QW-like-structure between amorphous Si 3N 4barriers. Highly ordered lattice planes can clearlybe seen in these Ge QW, indicating formationof large single crystal type structures. Based onRaman and XRD characterisation, the crystallisationtemperature shows a significant dependenceon the thickness of either Ge QW or Si 3N 4barrierGe vol% ∆(2Ө) of (111) NC Size (nm)30 5.62 NA40 3.85 2.2250 2.41 3.5460 2.38 3.5970 1.92 4.45Ge NC size estimation from XRD.Table 4.5.1Absorption edge observed fromtransmission and reflectancemeasurements.Figure 4.5.31.TEM image of the samplecontaining 60 vol% Ge and furnaceannealed at 900ºC.Figure 4.5.3275

mesa, depicted in Fig. 4.5.33, can be modelledadding a current dependence on the seriesresistance. This series resistance, R S, can beexpressed as the sum of a current independent, R extand a current dependent series resistance R intarisingfrom current crowding:(4.5.3)The value of this resistance can be found bynumerically solving the following transcendentalequation:(4.5.4)Within this mathematical framework it is thenpossible to extract the value of the seriesresistance, remove its effect from the measuredI-V characteristics to extract the actual ideality ofthe diode current. We normally observe an idealityfactor of 3. Based on the proposed model of at leasttwo diodes in series, without any assumptions, itis only possible to establish that the value of theband gap extracted from temperature dependentI-V measurements corresponds to a sum of the bandgaps (or activation energies) of these diodes.An interesting behaviour observed in thesedevices was the apparent lack of correspondencebetween the dark and light I-V characteristics. Theseries resistance extracted from the diode dark I-Vcharacteristic is too small to explain the limitedshort circuit current as well as the low fill factormeasured under a simulated 1-sun condition. Amore complete circuit model is necessary to explainthis discrepancy. We have proposed a model wherethe observed behaviour is due to two distinctareas in the fabricated devices. The photocurrent isproduced only in a small area of the device, this areabeing proportional to a fraction of the normalizeddiode area. In a Spice circuit model this area is givena value smaller than 1, that we denote as fraction.This will be a fitting parameter to reproduce themeasured dark and illuminated I-V characteristics.Figure 4.5.34 shows the circuit proposed.A relatively large percentage of the device areais responsible for the measured characteristics inthe dark. The current in this area is caused by thediffusion of minority carriers caused by the appliedvoltage (V1 in Fig. 4.5.34) from the p or n side tothe opposite region. This current is expected tobe large due to the low lifetime of the minoritycarriers (mostly recombination current in thedepletion region). The observed series resistancewill be proportional to the total series resistance,R tot, and inversely proportional to the area, R2=R tot/(1-fraction), where the diffusion current occurs. Thedark I-V behaviour is modelled by the top branch ofthe circuit depicted in Fig. 4.5.34. D_QD1 and D_Schrepresent the series connection of diodes whoseband gaps are extracted using Eq. (4.5.2). The seriesresistance, R2, has the temperature and currentdependence detailed in Eqs. (4.5.3) and (4.5.4).Only a small part of the diode area may have a largeenough lifetime to produce a photocurrent. As thisphotocurrent flows only through this fractionalarea, the series resistance is inversely proportionalto this fraction: R1=R tot/fraction. Since thephotocurrent, Iph, flows through R1, the illuminatedI-V characteristics are limited by a larger resistancethan that observed in the dark condition, as long asthe fraction of the diode area is smaller than onehalf. The simulations depicted in Fig. 4.5.35 showthe reduction of I SCas the fraction, f, is varied from99% to 1% of the total diode area.With the circuit model described, it will be possibleto extract complementary information from darkand illuminated I-V measurements.In view of these simulations, it is clear that theelectrical characterisation of our devices needsto take into consideration previously overlookedlimitations. For example, great care should betaken when interpreting the Quantum Efficiencyextracted from a spectral response measurementas the assumed condition of short circuit currentmay be incorrect even if the device is externallyshort circuited (internally, the diodemay be forward biased). Moreover,as the current is proportional to thephoton flux, and the flux is generallydifferent at each wavelength tested,the internal bias of the device can bedifferent at each point of the spectruminvestigated.4.5.2.4.2 Demonstrator programfor Si nanostructure devicesUsing the approach described inSection 4.5.2.4.1, a device simulatorhas been developed in an MS Excel spreadsheet.The Demonstrator is a 2D simulator of the diodesfabricated from Si QD nanostructures. It is based ona simple diode model with a few fitting parametersand uses the method in section 0 and in [4.5.9].There are three areas which the user can modify:1. Measured parameters: these come frommeasured properties of the fabricated diodes.The base resistivity and its temperaturedependence are measured. Thedevice geometry is known. Thedefault values are representativeones for the devices.2. J 0, ideality factor, band gap and XTOcan be varied to give flexibility infitting to a specific diode properties.3. The variation parameters are Currentand Temperature.The outputs are:1. The I-V curves for the range oftemperatures chosen, in both log-linand lin-lin output.2. The variation of the idealityfactor with current forvarious temperatures.3. The series resistancewith temperature.4. The extracted band gap, using either acalculation incorporating a series resistancecorrection or with both a series resistance and acurrent modification term.f=0.01f=0.03ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTEquivalent Spice circuitrepresentation of thefabricated devices.Figure 4.5.34f=0.1f=0.31f=0.99Spice simulations of 1-sun I-Vcharacteristics based on the circuitdepicted in Fig. 4.5.34. The fraction,f, represents the normalized area ofthe diode where the photocurrentis produced.Figure 4.5.3577

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTDark I-V curves.Figure 4.5.36Local ideality factor.Figure 4.5.37Some example outputs, using the default parametersettings, follow:Figure 4.5.36 shows that a reasonable diodecharacteristic is shown, with good turn-on andrectification.Figure 4.5.37 shows the increase in the idealityfactor with current is as expected with adiode transiting from domination by radiativerecombination (n=1) to being dominated byShockley Read Hall recombination (n=2). But thelarge regime in which n=3 and then region whereit goes to much higher values, are indicative ofa non-physical aspect to the model. The mostlikely explanation is that rather than two diodesin series there are actually several, probably avariable number in different parts of the device, thecombination of which lead to a composite idealityfactor.The demonstrator is useful as an iterative tool usedto simulate the real measured data from a device.The parameters used for such a fit then define thevalues of the circuit elements in the EC model.Attachment of physical meaning to these elementswill then require further development and interfacewith the EMA/quantum mechanical model.4.5.2.5 Doping in Si QD nanostructuresFabrication of a PV device from the Si QD materialsrequires a control of work function such as to allowseparation of photogenerated electron-hole pairs.Methods by which this can be achieved includefabrication of a grown or diffused p–n junctionor p–i–n junction with the Si QD multilayers asthe i-region. These require careful control ofthe work functions of the p and n-regions. P-njunction devices have been fabricated and dopingdemonstrated. Progress on the theory explainingthese doping effects using conventional dopantsand the possibility of other modulation dopingapproaches has been made.As discussed in Section 4.5.2.4, rectifying p-n andp-i-n structures can be fabricated with ‘conventional’doping by incorporation of B and P duringsputtering growth [4.5.9, 4.5.10]. Also, formation ofp- and n-type materials is clearly demonstrated inthe Si QD nanostructure materials with Si nitrideinterlayers doped with either P or B discussed inSection 4.5.2.3.2; in the Sb doped n-type Si QDs inSi 3N 4in Section 4.5.2.3.1; in the p-type Ge QD inSiO 2material in section 4.5.2.3.4; and in MOS typedevices doped with either P or B by a diffusionanneal discussed in [4.5.43].However, the doping mechanisms taking placein these structures are not well understood.Theoretical work has shown that direct doping ofthe QDs is prevented by segregation of impurityatoms from the perfectly crystalline QDs [4.5.44,4.5.23]. Experimentally this is supported by dataon the free electron density in Si nanocrystalsusing Electron Paramagnetic Resonance (EPR),which show that 95% of P atoms are segregatedto the surface of the nanocrystals and that theircontribution to doping is at least an order ofmagnitude lower than the atomic concentration[4.5.45]. Further evidence that doping causeselectronic changes comes from the quenching ofluminescence on the incorporation of P (or Au) inSi nanostructures [4.5.46] and also from a smallenhancement of luminescence observed at lowP doping levels followed by quenching at higherlevels [4.5.47]. This latter being explained by thepassivation of QD surface states by low levels of Pincreasing luminescence and then the saturation ofthis mechanism by excess P. This also explains thenon-monotonic behaviour of activation energy seenin P doped nanostructures [4.5.10].If direct doping of the QDs is not occurring,then another possibility is modulation dopingof the matrix SiO 2, as is commonly used in III-Vnanostructures; but this is ruled out because ofthe very high ionisation energies of about 5eV thatwould be required for the resultant deep defectsin a dielectric matrix [4.5.48]. These clearly willnot occur at room temperature, or even at highertemperatures.78

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTStoichiometricdielectric, 2-5nm a) b)Two models for doping of Si QD nanostructures. (a)An extended sub-oxide region surrounding the Si QDsallows doping by P or B at relatively shallow levels. Thesethen can provide carriers to be captured by the QDsthrough a modulation doping mechanism. (b) QD sizeenhancement (suppression) by P (B) doping results in atype I heterojunction between P and B doped material.This facilitates carrier separation due to the much greatermobility of electrons – the Dember effect [4.5.49].Figure 4.5.404.5.2.5.1 Doping mechanisms inSi QD nanostructuresResearchers:Dawei Di, Xiaoming Hao, Ivan Perez‐Wurfl,Gavin ConibeerAs there is clear evidence for doping by P and B,a third possibility is that it is the sub-oxide QD/matrix interface region, with its shallower dopingdefect levels, that is doped by the P or B atoms, thusproviding free carriers to be captured by the QDs[4.5.49]. As shown in Fig. 4.5.38, ab-initio calculationof the levels associated with B atoms associateddirectly with the interface region on the surface ofa Si QD indicate that there are no shallow or deeplevels introduced by the B within the effective bandgap (HOMO-LUMO gap). The reason is that thestrong affinity of B for O results in very strong B-Obonds, effectively splitting any available OMOs andUMOs a long way apart and well inside the alreadyexisting density of states.However, this approach considers the interfacebetween Si QD and SiO 2matrix to be very abrupt.In fact there is highly likely to be an extendedtransition region of a sub-oxide around the QD. Ifthis region is a sufficiently extended silicon suboxide(SiO x) it could provide flat Bloch bands ableto be doped in a pseudo bulk-like regime. In orderthat B and P doping levels are shallow enough tobe ionised in such a region, x would need to be lessthan about 0.5. Such a region would be amorphousand hence not able to be modelled with ab-initiomethods. But the presence of such regions can bedetermined from absorption measurements.There is some support for this theory in theabsorption data for Si QDs in SiO 2[4.5.5] shown inFig. 4.5.39. This shows a region of strong absorptionwith an absorption edge which blue shifts asthe O to Si ratio increases, which translates to anincrease in QD size. But there is also an additionalweakly absorbing tail which extends well into theeffective band gap region. This tail is likely to bedue to an amorphous region around the Si QDs. Anexplanation for the band gap of this being largerthan bulk Si is if this region consists of sub-oxidematerial. If this region were doped with B or P, itcould provide free carriers to be captured by theQDs in a modulation doping mechanism, thusgiving rise to the p- or n-type behaviour observed.(This is illustrated schematically in Fig. 4.5.40 (a).)An alternative explanation for the rectifyingcharacter of the junctions obtained canbe based on the modification to the Si QDcrystalisation discussed in Section 4.5.2.3.2,for P and B doping of Si QDs in oxide withnitride interlayers [4.5.49]. The fact that P (B)doped material produces larger (smaller)QDs means that its confined energy levelsand hence effective band gap will be smaller(larger) than undoped material underotherwise similar conditions. Thus a junctionbetween a P and a B doped region willactually be a type-I heterojunction betweensmall and large band gap materials,respectively, as illustrated in Fig. 4.5.40 (b).It is also very likely that the mobility of electronsin these materials, whilst not big, will still be muchgreater than that of holes. Hence photogeneratedelectrons near the junction would experience adrift field sweeping them into the P doped materialwith its lower conduction band edge, whilst holeswith their very limited mobility would be immobile.This would result in electrons accumulating in the PAbsorption (10 6 cm -1 )252015105Ab-initio calculation usingGAUSSIAN: (a) the density of statesfor a Si 83/84OH 62oxide terminated QDwith and without a B atom at theinterface; (b) the optimised Si QD,consisting of Si 83B-OH 62, with the Batom at top centre (in pink).Figure 4.5.3805.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5Energy (eV)Room temperature absorptionco-efficient of annealed SiO x/SiO 2multilayer films with variousx (1100 o C, 1hour). The inset isan estimate of the approximateoptical band gap (filled squares:strong absorption; empty circles:absorption tail) [4.5.5].Figure 4.5.39Wavelength (nm)300 400 500 600 700 800Energy (eV)4.84.44.03.62.42.22.01.8m=2m=1/20.8 0.9 1.0 1.1 1.2 1.3O/Si ratioSiO 0.86/SiO 2SiO 1.00/SiO 2SiO 1.30/SiO 279

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTElectronic density of states (DOS) of Si 10QD in 2 MLSiO 2(dotted black), fully oxidised nanocrystal (grey)and Si 10QD in 2 ML SiO 2with Sc replacing Si atom inthe SiO 2shell at the outermost position (top). DOS ofSc-doped approximants split into α (spin up, red) andβ (spin down, blue) DOS because of non-even numberof electrons within the approximant due to acceptorstate of Sc. The b-highest occupied molecular orbital(HOMO) 1193 indicated in magenta can take up 0.07electrons from the Si 10QD over two 2 ML (6 Å) of perfectSiO 2. The β-lowest unoccupied MO (LUMO) 1194 shownin green can take up a much bigger proportion of anelectron if the Si QD is slightly bigger (e.g. Si 35, d QD=11 Å). The Sc-doped nanocrystals are shown with theiso-density plots of 1.35×10 −3 e/Å 3 for β-HOMO 1193(lower left) and β-LUMO 1194 (lower right). The atomsof the Si 10QD and the Sc atom are highlighted in cyan.Si atoms in SiO 2are grey, O atoms are red, H-atomsterminating the outermost O-bonds are white. All DFTcomputations carried out with a Hartree-Fock 3-21G(d)optimisation and B3LYP/6-31G(d) electronic structurecalculation [4.5.23].Figure 4.5.42Spatial distribution of 4s and 3delectronic states of Sc (upper) andenergetic arrangement of 4s and 3dstates (lower).Figure 4.5.41material and being collected in the external circuitas if from an n-type material, with electrons injectedback into the device through the B doped materialwhere they would recombine with the immobileholes, thus making the B material appear p-type.This qualitatively mirrors the behaviour observed,but will require corroboration. As it relies on thedifference in electron and hole mobilities, it isessentially the same as the Dember effect, whichis known to produce a photovoltaic effect throughcarrier separation, albeit not very efficiently.It is also possible that both the sub-oxide dopingeffect and the mismatched QD size effects operatein the device. The fact that both mechanisms actin the same direction is fortuitous as both couldtherefore contribute to the observed rectifyingphotovoltaic behaviour. Differentiation betweenthese or a determination of the relative strengthswill require further study on the effects of doping,interlayers and band alignments. These will alsofurther optimisation of the devices and lead tobetter photovoltaic performance from these Si QDnanostructure devices.4.5.2.5.2 Ab-initio modelling ofmodulation doping possibilities inSi QDs nanostructuresResearchers:Dirk König, James Rudd, Daniel Hiller (IMTEK,University of Freiburg, Germany)As discussed in Section 4.5.2.5, direct doping ofQDs by dopants is thermodynamically very unlikely.The alternative of modulation doping of the SiO 2matrix is also not feasible due to the strong anionicnature of O leading to very deep ionisation energies[4.5.48]. We have concentrated recently on possiblecandidates for more efficienct acceptor dopingof Si QDs. Modulation acceptors must have anenergetic position below the confined hole level,they must not introduce defect levels in the QDband gap if located at the interface as an activedopant and as a completely saturated foreignatom, and they must have one single oxidationnumber. The transition metal scandium (Sc) fulfilsthe last two requirements. It also has a very suitablevalence state configuration which consists of afull (doubly occupied) 4s shell and one electronin the 3d 1 state located within the 4s shell. On theenergy scale, the 4s electrons have a higher bindingenergy as compared to the 3d 1 and 3d 2 states. Thisprovides the 3d 1 and 3d 2 states with an electrostaticscreening which results in a higher binding energyof these states. Thereby an unoccupied 3d 2 stateexists at a high binding energy which is able to takeup an electron from a nearby source such as a Si QD,see Fig. 4.5.41.Ab-initio Density Functional Theory calculationson a Si 10QD in 2 mono layers (MLs) of SiO 2havebeen carried out, with a Si atom at the outermostposition with full O termination replaced by a Scatom. The Si 10atom cluster is the smallest size whichstill behaves as a Si QD [4.5.23]. Such a nanoparticleallows us to directly observe the electronic structureof Sc as a modulation acceptor in SiO 2in theproximity of a Si QD, see Fig. 4.5.42.Figure 4.5.42 shows that 0.07 electrons stemmingfrom the Si 10QD (d QD= 7.3Å) are localized atthe Sc atom over two 2 ML (6 Å) of perfect SiO 2.Technological relevant QD sizes start at d QD≈ 20Å,with a much lesser degree of quantum confinementas compared to a 7.3Å QD. It is thus reasonable toassume that a Si QD with d QD≥ 20Å will be positivelyionized to a much greater extent, ensuring thatholes are majority carriers in a Si QD super lattice(SL) embedded in SiO 2. Experimental verificationof this doping mechanism is being carried out. Thedifference to conventional acceptor modulationdoping as used in III-V electronic devices is theway the Si or Ge QDs are ionised by Sc in SiO 2. AsSc induces a deep acceptor level in SiO 2it cannotbe thermally ionized from the SiO 2valence band.Instead, the Sc acceptor state takes up the electrondirectly from the initially occupied state at a lowerbinding energy presented by a nearby QD, see Fig.4.5.43. This requires the Sc acceptor to be within50Å of the Si QD in order to exploit field emission80

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTAcceptor modulation doping:Conventional example with GaSQD in AlAs matrix and electronthermally emitted into acceptor(Mg), such that holes are capturedby the QD (left). In SiO 2, chargetransfer occurs directly as theSc state is too deep within theband gap to be thermally ionized.Instead, the acceptor state isionised directly from the QD byfield emission of an electron intothe Sc acceptor state.Figure 4.5.43which is a consequence of the energy differencebetween the QD HOMO and the Sc acceptor state.Acceptor modulation doping of SiO 2can beaccomplished by incorporation of a small amountof Sc, say one monolayer of Sc, into the SiO 2barriersbetween the SRO QD array layers. During thesegregation anneal, Sc is incorporated into theSiO 2matrix and thereby activated as an acceptor.Ab-initio calculation indicates that Sc attachedto the Si QD does not create any levels withinthe HOMO-LUMO gap of a Si 84(OH) 64QD. We cantherefore assume that a Sc atoms bonded onto aSi QD does not result in a detrimental change inits electronic structure. This concept is now beinginvestigated experimentally.4.5.2.6 Summary of Group IVnanostructures for tandem cell elementsIn 2010, significant progress has been made onunderstanding the transfer of growth parametersof Si QDs to other matrices. The use of anisotropicstructures both to provide diffusion barriersto excessive QD growth and to provide higherconductivity paths in the transport direction, hasshown the validity of these approaches both forSi 3N 4and for SiC matrices. Knowledge of Ge QDnanostructures has improved significantly withclear p-type behaviour of Ge QD s in SiO 2and verysuccessful growth of Ge QWs in Si 3N 4matrix. Theincrease in sophistication of modelling of confinedenergy in these structures now matches reasonablywell with experimental data.Improved modelling of confined levels inirregular shaped particles also matches wellwith experimental data of other groups. Higherlevel device modelling is now able to describethe anomalous effects due to high lateral seriesresistance in devices and consequent currentcrowding. This is now being used to optimisenanostructure growth and device design, so as topush beyond the current open circuit voltages of490mV. Part of this is an understanding of doping.This has progressed with greater knowledge andcontrol over the effects on QSD size of introductionof P and B dopants, and modelling of possibledoping mechanisms based on modified modulationdoping and these size effects. Alternativemodulation dopants are being modelled withscandium as an acceptor looking like a strongcandidate. Further work will build on this increasedunderstanding of anisoptropic structures, growth ofdifferent materials and understanding of doping toimprove materials and devices further.4.5.3 Hot Carrier cellsResearchers:Shujuan Huang, Santosh Shreshtha, DirkKönig, Robert Patterson, Pasquale Aliberti,Binesh Puthen Veettil, Lara Treiber, IvanPerez-Wurfl, Andy Hsieh, Yu Feng, JamesRudd, Stephan Michard, Martin Green,Gavin ConibeerHot carrier solar cells offer the possibility of veryhigh efficiencies (limiting efficiency above 65% forunconcentrated illumination) but with a structurethat could be conceptually simple compared toother very high efficiency PV devices – such asmulti-junction monolithic tandem cells. For thisreason, the approach lends itself to ‘thin film’deposition techniques, with their attendant lowcosts in materials and energy usage and facility touse abundant, non-toxic elements.An ideal Hot Carrier cell would absorb a wide rangeof photon energies and extract a large fractionof the energy to give very high efficiencies byextracting ‘hot’ carriers before they thermalise tothe band edges. Hence an important property of ahot carrier cell is to slow the rate of carrier coolingto allow hot carriers to be collected whilst they arestill at elevated energies (“hot”), and thus allowinghigher voltages to be achieved from the cell andhence higher efficiency. A Hot Carrier cell must alsoonly allow extraction of carriers from the devicethrough contacts which accept only a very narrowrange of energies (energy selective contacts or81

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTBand diagram of the Hot Carrier cell. The device hasfour stringent requirements:Figure 4.5.44(a) Schematic representationof energy and particle fluxesinteractions used in the model(Particle fluxes - full line arrow,energy fluxes- dotted line arrow).(b) HCSC efficiency as a functionof carrier extraction energy level.Parameters used are: thermalisationtime = 100 ps, concentration =1000, lattice temperature = 300Kand absorber layer thickness = 50nm. [J, F and E are current densityand particle and energy fluxesas denoted by subscripts for Aabsorption, E emission and IAAuger processes.]Figure 4.5.45Carriers transmission probabilityversus energy for (a) ideal ESC, (b)non-ideal ESC.Figure 4.5.46ESCs). This is necessary in order to prevent coldcarriers in the contact from cooling the hot carriers,i.e. the increase in entropy on carrier extraction isminimized [4.5.50]. The limiting efficiency for thehot carrier cell is over 65% at 1 sun and 85% atmaximum concentration – very close to the limitsfor an infinite number of energy levels [4.5.1, 4.5.51,4.5.52]. Fig. 4.5.44 is a schematic band diagram of aHot Carrier cell illustrating these two requirements.a) To absorb a wide range of photon energies;b) To slow the rate of photogenerated carriercooling in the absorber;c) To extract these ‘hot carriers’ over a narrow rangeof energies, such that excess carrier energy is notlost to the cold contacts;d) To allow efficient renormalisation of carrierenergy via carrier-carrier scattering.In 2010 modelling of Hot Carrier efficiencies hasprogressed with implementation of realmaterial properties to give more realisticefficiencies for InN which include Augerprocesses and more realistic contactstructures. Significant progress has beenmade on demonstrating resonance indouble barrier selective energy structures.Further work on triple barrier double SiQW structures has been carried out forrectifying ESCs. This is complemented byimprovements in 2/3D modelling of transport inthese ESC structures. For absorbers, modelling ofnanocrystals superlattice arrays hasbee applied to real material systems.The growth of such systems in bothIII-V QD superlattices with collaboratorsand with colloidal Langmuir-Blodgettdispersion of Si nanocrsystals hasproduced structures which arenow being characterised for theirmodulation of phononic properties.Also meausurement of carrier coolingrates has been extended to other largephononic gap bulk materials includingInN, demonstrating the importance ofmaterial quality. Design of structuresfor hot carrier cells which should bepractical and realisable has developed, with thedevice properties more carefully specified and plansfor fabricating such structures in real devices.4.5.3.1 Modelling of Hot Carrier Solar CellEfficiencyResearchers:Pasquale Aliberti, Yu Feng, Santosh Shrestha, GavinConibeer, Martin GreenCollaboration with:Yasuhiko Takeda (Toyota Central ResearchLaboratories, Nagoya)Previous work was focused on developing a newmodel to calculate limiting efficiency of a real HCSCbased on an Indium Nitride (InN) absorber layer. InNhas been chosen as a potential material because ofits narrow electronic band gap for absorption of awide range of photon energies, whilst also having awide phonon band gap. This is good for suppressionof phonon decay, and hence for slowing carriercooling [4.5.53, 4.5.54].Calculation of limiting efficiency was performedtaking into account real optical and electronicproperties of InN, removing most of the idealityassumptions used in other models [4.5.51, 4.5.52,4.5.55, 4.5.56]. The detailed band structure82

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORT(a) HCSC efficiency as a function of extraction widthof ESCs for different extraction energies ΔE. (b) HCSCefficiency as a function of extraction energy ΔE fordifferent ESCs energy width. Thermalisation time is100 ps, lattice temperature is 300 K. Absorber layerthickness is 50 nm.Figure 4.5.47of wurtzite bulk InN has been considered inperforming computation of carrier densities,pseudo-Fermi potentials and II-AR time constants[4.5.57]. Results have been calculated consideringideal energy selective contacts for the HCSC, whichmeans that contacts have a very high conductivityand a discrete energy transmission level.Recently the limiting efficiency for the hot carrierInN solar cell has been calculated consideringnon-ideal ESCs. In this case the carriers are notextracted on a single energy level, but in a finiteenergy window. Calculations have been performedtaking into account contact resistance and entropygeneration effects.The flux of current travelling through the ESCstowards the cold metal electrodes can be describedusing the following relation.(4.5.5)The current density in this case is proportional tothe occupation probability at the two sides of theESC. Equation (4.5.5) has been derived assuming nocorrelation of energy of electrons in three differentdirections as shown in (4.5.6). This assumption isacceptable if there is a parabolic dispersion relationat minimum energy point along the three differentdirections.(4.5.6)Based on the energy and carrier conservation, Δμand TC at steady state are calculated.(4.5.7)[quantities as defined in Fig. 4.5.45 (a).]The maximum efficiency has been found for a ΔEbetween 1.15 eV and 1.2 eV with a transmissionenergy window δE of 0.02 eV. The value of limitingefficiency is 39.6% compared to 43.6% calculatedin the previous section using ideal ESCs. The dropin efficiency is mostly due to the decrease of opencircuit voltage related to the decreased extractionlevel, equation (4.5.7). This is partially compensatedby an increase in extracted current due to increasedII rate.Figure 4.5.47 (a) shows calculated efficiency as afunction of for several values of extraction energy. Inall the curves two different trends can be identified.If the value of δE is too close to zero, the efficiencyis very low due to low carrier extraction, thus a verysmall value of short circuit current. The conductivityof the contact in this case is indefinitely large.Enlarging δE, the number of carriers available forextraction increases, improving J SC, and so themaximum efficiency. In general the efficiency peakhas been found for values of δE from 0.02 eV to 0.1eV depending on the extraction energy ΔE. For theconfigurations which show higher efficiencies, ΔE< 1.35 eV, the optimum value of δE goes from 0.02eV to 0.05 eV. This optimum value for δE of betweenis very close to kT RT. This represents the variationin energy in the contacts such that approximatelythe kT RTwill inevitably be lost anyway by carriersthermalising within the contacts. Thus it sets alower limit on a reasonable δE. Therefore this resultindicates that the transmission energy range hasto be very small and confirms once again the highselectivity requirements of ESCs for HCSC [4.5.58].In Fig. 4.5.47 (b) the value of maximum efficiency asa function of ΔE is reported for different values ofδE. It can be observed that for small transmissionenergy window the extraction energy which allowsmaximum efficiency is lower compared to the onecalculated using ideal ESCs. This effect is relatedto the higher occupancy at lower energies, whichincreases the value of J SCfor contacts with a smalltransmission window.83

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTPeak at 1.205eV of the conductance vs. energy plot fordifferent values of σ for configurational disorder (left).Conductance vs. energy plot with peak at 1.573 eV fordifferent values of σ for morphological disorder (right).Conductance given in units of 2e 2 /h.Figure 4.5.48Planar view of the double barrierstructure formed by Si QDs in SiO 2matrix, with SiC barriers of 2nmwidth (left). Mean diameter of QDsis 1.8nm. I-V characteristics for thedouble barrier structure with SiO 2and SiC barriers at temperature 10Kand 300K (right).Figure 4.5.494.5.3.2 Energy Selective ContactsThe requirement for a narrow range of contactenergies can be met by an energy selectivecontact (ESC) based on double barrier resonanttunnelling. Tunnelling to the confined energy levelsin a quantum dot layer embedded between twodielectric barrier layers, can give a conductancesharply peaked at the line up of the Fermi level onthe ‘hot’ absorber side of the contact with the QDconfined energy level. Conductance both belowthis energy and above it should be very significantlylower. This is the basis of the current work ondouble barrier resonant tunnelling ESCs.4.5.3.2.1 Modelling of QD structuresResearchersBinesh Puthen-Veettil, Dirk König,Gavin ConibeerWe developed a robust 2 dimensional modelfor describing the transport properties throughquantum dot structures and have used this modelto understand the filtering characteristics of EnergySelective Contacts (ESCs). In this way we are ableto compute the effective filtering in 2 dimensionsby running numerical simulations. The model isdeveloped from a discretized Schrödinger equationby considering the sample volume as a collectionof discrete points and using an effective massapproximation method over the entire volume.During fabrication of the quantum dots in adielectric matrix for selective energy contacts,different kinds of irregularities can be presentin the structure, the major disorders beingconfigurational (disorders in the position of thedots) and morphological (disorders in the size ofthe dots). The extent to which configurational andmorphological disorders determine the electricalproperties of the overall structure is investigatedusing simulation runs of the model. The disordersare assumed to follow a normal distribution fromthe mean position and size. The results show theoutcome of an average of 1000 simulation runs withdifferent standard deviation (σ) values.Figure 4.5.48 shows the simulation results forresonance in 2.6nm Si dot in SiO 2matrix underdifferent orders of configurational disorders. Asthe disorders increases from σ =0 to σ =1 theconductance decreases by 53%, but the resonantenergy remains the same at 1.205eV. This showsthe confined energy in the QDs does not changeas their size is fixed but the effective filteringreduces dramatically. Figure 4.5.49 shows thesimulation results for a 2.2 nm Si dot in a SiO 2matrix under different orders of morphologicaldisorders. As the disorder increases from σ =0 toσ = 1, the conductance decreases by 60% and theresonant peak remains the same at 1.573eV. Butthe morphological disorders cause major impactcompared to configurational disorders because ofthe widening of the energy selection window. Thisis due to the distribution in size of the QDs, sinceQDs with different sizes have different resonantenergies which are slightly different from the meanresonant energy, the average of them all increasethe spread of the resonant peak thus reducing theefficiency of the double barrier structure as energyselective contacts.A planar representation of the double barrierstructure is shown as in Fig. 4.5.49 (left). The barriersare usually high band gap dielectric materials likeSiO 2, Si 3N 4or SiC. SiC barriers have advantagesover SiO 2barriers that SiC barriers in double barrierstructure makes a very good diffusion barrierfor silicon during processing, which can yield a84

250 – 500 nm 500 – 550 nm 550 – 600 nmARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORT600 – 650 nm 650 – 700 nm 700 – 750 nmPrinciple of oa-IV, shown for electron injection into the sample(left). If free carriers obtain additional energy by a massivephoton flux significantly above the band gap, a smaller bias fieldis required to reach a resonant condition. Sample stage of oa-IV(right); wavelength range indicated. Every 50 nm spectral rangehas a photon flux of approximately 50 Suns.Figure 4.5.50structure with uniform QD size in the transportdirection thus ensuring sharp resonances at theresonant energies of the structure. Since theconductance of the structure is more in the case ofSiC barriers because of the lower barrier height ofSiC, thicker barriers can be used to give the sameconductance as SiO 2, for which the very thin barriersare difficult to fabricate.Figure 4.5.49 (right) shows the I-V characteristics forthe double barrier QD structure with SiO 2dielectricand SiC barriers at temperature 10K and 300K.As can be seen, the structures with SiC potentialbarriers have higher current density at both lowtemperature and at room temperature. As thetemperature increases, the current density as well asthe width of the resonant peak increases. Also, thenegative differential resistances appear at slightlylower voltage in structures with SiC barrier thanthat in structures with SiO 2barrier. This is due to theincreased leakage and thus the lower confinementenergy for the SiC.4.5.3.2.2 Characterization of EnergySelective Contacts using ‘Opticallyassisted I-V’Researchers:Dirk König, Stephan Michard (RWTH Aachen,Germany), Binesh Puthen-VeettilCollaboration with:Daniel Hiller (IMTEK, University ofFreiburg, Germany)Optically assisted IV (oa-IV) is a new characterisationmethod which investigates an energy selectivecontact (ESC) fabricated on a Hot Carrier Absorbermaterial with optical excitations in a pre-definedwave length range. If a very large optical generationrate G optexists adjacent to the ESC, it can probe thisHC population immediately next to the contact.Alternatively, if the HC population within theabsorber is known, the ESC can be tested for itsenergy selectivity. The basic principle is to providefree carriers with additional energyby optical means, see Fig. 4.5.50. Theresonant level of the ESC can thenbe reached by carriers at a lower biasfield, shifting the tunnelling resonanceof the ESC – here a QD array – to lowerbias voltages V bias.Samples based on Si were providedby the University of Freiburg, seeFig. 4.5.51. Si is very far from beingan ideal Hot Carrier Absorber due tothe complete lack of a gap betweenoptical and acoustic phonon modesand hence no propensity to slowcarrier cooling. Furthermore, Sihas an indirect band gap in thespectral range of optical excitationwhich renders the interpretationof measured data rather complexdue to the absorption coefficient a Sichanging over two orders of magnitude. However,for initial proof of concept it is adequate providingthe excitation of hot electron-hole pairs in thematerial is sufficiently large and collection can bemade from very close to the ESC.A Xenon (Xe) arc lamp is a continuous light sourcewith one of the most constant photon fluxes F hninthe spectral range 1000 to 500 nm. Nevertheless,there are significant deviations in particular in thelong wavelength region (> 800 nm). We thereforecompare oa-IV curves which have the same V biasfor the current transition from forward to reversedirection, V bias(j → 0), indicating a constant carrierproduct (constant quasi-Fermi level positions) atthe ESC. Under these conditions, the current densityas a function of quasi-Fermi level position can beassumed to be identical for these curves and thiscurrent density can be treated as a constant offsetfor each V biasvalue.For liquid nitrogen (LN 2, T = 77 K) measurements,resonant tunnelling behaviour (negative differentialresistance – NDR) was detected for all spectralranges. With the method mentioned above, onlythe oa-IV curves from 800 – 750 nm and 650 –Sample structure for oa-IV (upper).Contact area on QD array is 1mm 2 , with ≈ 10 10 QDs. Opticalgeneration rate G optwithinballistic transport rangeto the ESC and its relativechange for photon wavelength intervals used inoa-IV (middle). For these, theabsorption coefficient of c-Sia Siand the photon flux of theXe arc lamp F hnwere used(lower). Full green rectanglesshow range of measurementswhich can be compared.Dashed green rectangles showmeasurements which cannotbe compared, but approachthe high optical injection limit.Please note breaks on leftordinate in both graphs.Figure 4.5.5185

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTSchematic representationof time resolvedphotoluminescencearrangement.Figure 4.5.53oa-IV results of a typical samplewith optimum barrier thicknessesand QD size at LN 2temperature(upper) and at room temperature(lower). Curves in the wave lengthintervals from 800 to 600 nm canbe compared due to their identicaloffset of V biasto absolute zero V bias.,showing the free carrier product tobe constant. Curves for 550 – 250nm and 1000 to 500 nm showmaximum optical injection limit.Both curve groups printed in bold.Figure 4.5.52600 nm with F hn≈ 50 Suns foreach interval could be directlycompared to each other. Inaddition, we evaluated the curveswith optical excitations in therange of 1000 – 500 nm (F hn≈ 600Suns) and 550 – 250 nm (F hn≈140 Suns). The former is near thepractical high concentration limitfor solar cells of 1000 Suns.As shown in Fig. 4.5.52, for thecurves in the 800 to 600 nmregion at LN 2temperature, thepotential difference between V bias(j→ 0) and the average voltage ofthe resonance (V Res,avg) shows adecrease of 48 mV with increasingphoton energy (decreasingλ range). As carrier densities(quasi-Fermi level positions) areidentical for these curves, thispotential drop corresponds to anincrease in free hole energy of48 meV. As this energy differenceis equal to the effective kT, it canbe interpreted as an increasein carrier temperature of DT = 560 K when goingfrom 800 – 750 nm (1.55 – 1.65 eV) to 650 – 600 nm(1.91 – 2.06 eV). Even at F hn≈ 50 Suns, this effectis small. This is due to Si being a very unsuitableabsorber material, further corroborated by the ratioof peak to valley current densities (quality factor ofNDR – QF NDR) dropping from 2.2 to 1.5. Si does nothave a phononic band gap so that optical phononsundergo a rapid decay which tremendouslyaccelerates carrier cooling.For the 1000 – 500 nm and 550 – 250 nm ranges,the QF NDRincreases significantly to 4.1 and 3.2,respectively. The potential difference V bias(j → 0) –V Res,avgis shifted by 365 and 200 mV, respectively;a quantitative comparison of both curves to eachother or to the curves in the spectral range from800 – 600 nm is thus not feasible. However, it showsthat the energy selectivity may increase under highconcentration F hn.For all oa-IV curves, a tunnelling feature for electronextraction can be seen at 760 mV forward bias.However, the background current density is muchhigher than the tunnelling portion of the currentdensity in this forward bias range, so no NDR canbe seen.At room temperature (T = 300 K), NDR was detectedonly for λ ranges ≤ 600 nm, while the impact ofresonant electron injection is still visible for curvesin the 800 – 600 nm range. The background currentdensity in the bias range around the tunnellingevent increases by about an order of magnitudewhen going from LN 2to room temperature.This is due to thermal activation of trap statesand increased thermal scattering during carriertransport through the QD array. It has the sameeffect on NDR suppression as explained above forelectron injection under forward bias.The QD arrays were processed by segregationanneals, comprising some 10 10 QDs as ESCsunder the contact pad. It is very encouraging tosee resonant tunnelling transport even at roomtemperature in a material system which was notgrown epitaxially. At LN 2temperature (77 K), anincrease in carrier temperature of DT = 560 K wasdetected when increasing the photon energy from1.6 ± 0.05 eV to 1.985 ± 0.075 eV under a constantphoton flux of F hn≈ 50 Suns. More suitable HotCarrier Absorber materials will be investigated.4.5.3.3 Hot Carrier Absorbers: slowing ofcarrier coolingResearchers:Robert Patterson, Gavin Conibeer, SantoshShreshtha, Dirk König, Pasquale Aliberti,Shujuan Huang, Yukio Kamikawa, Lara Treiber,Martin GreenCarrier cooling in a semiconductor proceedspredominantly by carriers scattering their energywith optical phonons. This builds up a nonequilibrium‘hot’ population of optical phononswhich, if it remains hot, will drive a reverse reaction86

to re-heat the carrier population, thus slowingfurther carrier cooling. Therefore the criticalfactor is the mechanism by which these opticalphonons decay into acoustic phonons, or heat inthe lattice. The principal mechanism by which thiscan occur is the Klemens mechanism, in which theoptical phonon decays into two acoustic phononsof half its energy and of equal and oppositemomenta [4.5.59].The build up of emitted optical phonons isstrongly peaked at zone centre both for compoundsemiconductor due to the Frölich interactionand for elemental semiconducotrs due to thedeformation potential interaction. The strongcoupling of the Frölich interaction also means thathigh energy optical phonons are also constrainedto near zone centre even if parabolicty of the bandsis no longer valid [4.5.60]. This zone centre opticalphonon population determines that the dominantoptical phonon decay mechanism is this pureKlemens decay.4.5.3.3.1 Suppression of phonon decay inbulk materialsIn some bulk semiconductors, with a largedifference in their anion and cation masses, therecan be a large gap between the highest acousticphonon energy and the lowest optical phononenergy, possibly large enough to block operationof this Klemens mechanism, which can be termeda ‘phononic band gap’. Work previously presented[4.5.61] using a simple 1D force constant model andcomplemented by high accuracy DFT computation[4.5.62] indicated that GaN, InN and InP all havelarge phononic band gaps, which are close to thosefound experimentally [4.5.63]. We are using timeresolved photoluminescence (tr-PL) to investigatethe carrier cooling rates in these materials.4.5.3.3.2 Time resolvedphotoluminescence measurements ofbulk phononic band gap materialsResearchersPasquale Aliberti, Dirk König, SantoshShrestha, Gavin Conibeer, Martin GreenCollaboration with:Raphael Clady, Murad Tayebjee, Tim Schmidt(University of Sydney), Nicholas Ekins-Daukes(Imperial College)The potential efficiency boost, which can beachieved by Hot Carrier solar cells, is directlyrelated to the possibility of extracting highenergy carriers from the absorber layer beforethermalisation, increasing the voltage and hencethe conversion efficiency. The poor conversionefficiency of photons with energies above the bandgap of the absorber is the main loss mechanismin conventional single junction solar cells. Theinvestigation of thermalisation time constants ofhot carriers is a crucial step towards the engineeringof Hot Carrier cells. The efficiency of an InN basedhot carrier solar cell has been calculated using acomplex theoretical model, see Section 4.5.3.1. Iwas found that the limiting efficiency is stronglyrelated to hot carriers relaxation velocity in theabsorber [4.5.57].A comparison of femtosecond time resolvedphotoluminescence (tr-PL) spectroscopy betweenInP and GaAs was reported in last year’s annualreport and is also now published [4.5.64]. Thisshowed a distinctly longer carrier cooling timeconstant for the wide phononic gap InP ascompared to almost zero phononic gap of GaAs.It also showed further evidence for excitationinto higher side valleys for both GaAs and InP forappropriate excitation wavelengths. In 2010 thehot carrier cooling in InN has been investigatedusing tr-PL. The wide gap between optical andacoustic branches in the InN phononic dispersionrelation (wider than that for InP) prevents theKlemens decay of optical phonon into acousticphonons. This can lead to slower carrier cooling dueto “Hot Phonon Effect” [4.5.65]. The decay of hotcarriers for different excitation wavelengths InN hasbeen investigated.Tr-PL experiments have been performedon InN samples using the measurementconfiguration shown in Fig. 4.5.53. In thistechnique a laser pulse acts as a switchinggate relating the photoluminescencesignal to the time domain. The PL signalis collected from the sample, after afemtosecond laser excitation, and focusedin a non linear crystal. The gate signalis generated from the same laser and isfocused on the same crystal after passingthrough an optical delay stage. The signalis detected using a monochromator anda PMT. Our system configuration provides150 fs pulses with tunable wavelengthover a range of 256 nm (4.84 eV) to 2.6 µm (0.48 eV).Figure 4.5.54 is a three dimensional representationof PL as a function of time for all the probedwavelengths. It can be observed that the PL sharplyrises when the carriers are photo-excited by thelaser pulse. The fast decay of the PL shows thethermalisation of carriers towards respective bandedges. The decay is faster for highly energeticcarriers compared to carriers closer to the bandgap.Thus the carrier population quickly degeneratestowards the band edges during the thermalisationprocess. In InN the thermalisation is most probablydue to interaction of highly energetic electrons andholes with LO phonons [4.5.66].To investigate the rate of the carrier coolingprocess, the effective temperature of the carrierpopulation has been calculated fitting the highenergy tail of the PL spectrum for every singletime during the cooling transient. The PL has beenfitted assuming that carriers form a Boltzmann-likedistribution in a femtosecond time scale using thefollowing equation.L(ε) represents the measured PL intensity atenergy ε, α(ε) is the measured sample absorptioncoefficient, E Gis the InN energy gap, 0.7 eV in thisARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORT3D representation of InN timeresolved PL data.Figure 4.5.5487

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTSome images of the structure as grown and modelled.(Left) the near-perfect stacking of the structure isshown. (Centre) in-plane InAs QD arrangement.(Right) simple hexagonal superlattice structureshowing stacking. The QDs are similar to flatteneddisks with in plane dimensions of 4050 nm andheights of 7 nm.Figure 4.5.57Carrier relaxation curves for InN.The photoexcited carrier densityis 1.5 x 10 19 cm -3 . The blue dashedcurve is a single exponential fit.Figure 4.5.55Schematic of the simple hexagonalsuperlattice of InAs QD disks,showing their relative dimensions.The red square indicatesorthorhombic repeat unit for thesimple hexagonal system.Figure 4.5.56case, and k Bis the Boltzmann constant.T Cis the fitted parameter and representsthe hot carrier temperature. Figure 4.5.55shows the carrier temperature transient,which seems to follow an exponentialbehaviour quite well (blue line – fit).The fitting of the calculated temperaturedata has been performed using asingle exponential.Here τ THrepresents the carriers thermalisationtime constant, whereas C and K are two constantparameters. The fitted value for τ THis 7 ps. Thislong cooling constant can be attributed to the hotphonon effect due to the long lifetime of the A1(LO)phonon [4.5.64,4.5.66]. This hot carrier relaxationvelocity is still too high to achieve a considerableefficiency gain in an InN hot carrier solar cell.However, it has been demonstrated that, for InN,the carrier cooling velocity is strictly related to thequality of the material and slower carrier coolingconstants compared to the ones calculated in thischapter have been reported in the literature [4.5.67].4.5.3.3.3 Nanostructures for absorbersNanostructures offer the possibility of modificationof the phonon dispersion of a composite material.III-V compounds or indeed most of the cubic andhexagonal compounds can be considered as veryfine nanostructures consisting of ‘quantum dots’of only one atom (say In) in a matrix (say N) withonly one atom separating each ‘QD’ and arrangedin two interpenetrating fcc lattices. Modelling ofthe 1D phonon dispersion in this way gives a closeagreement with the phonon dispersion for zincblendeInN extracted from real measured data forwurtzite material.Similar ‘phonon band gaps’ should appear ingood quality nanostructure superlattices, throughcoherent Bragg reflection of modes such that gapsin the superlattice dispersion open up [4.5.61]. Thereis a close analogy with photonic structures in whichmodulation of the refractive index in a periodicsystem opens up gaps of disallowed photonenergies. Here modulation of the ease with whichphonons are transmitted (the acoustic impedance)opens up gaps of disallowed phonon energies.4.5.3.3.3.1 Force constant modelling ofIII-V QD materials by SK growthResearchers:R. Patterson, Y. Kamikawa, G. ConibeerCollaboration with:Yoshitaka Okada – University of Tokyo3D force constant modelling, using the reasonableassumption of simple harmonic motion of atomsin a matrix around their rest or lowest energyposition, reveals such phononic gaps [4.5.68].The model calculates longitudinal and transversemodes and can be used to calculate dispersions ina variety of symmetry directions and for differentcombinations of QD sub-lattice structure and superlatticestructure.III-V Stranski-Krastinov grown QD arrays of InAs inInGaAs and InGaAlAs matrices are fabricated at theUniversity of Tokyo using MBE. We are investigatingthese for evidence of phonon dispersionmodulation and potential slowed carrier cooling.In order to understand the expected phonondispersions these are being modeeled using the 3Dforce constant technique.Lattice matched and strain compensated materialpairs that may produce large phonon bandgaps areof interest. Previous iterations in the design of thesestructures indicated the importance of separating“light” and “heavy” atoms to different parts of thenanostructure. Initially, the lightest atom in thesystem, As, was present in both the QD and thematrix. This meant that the reduced mass of bothregions (proportional to the sum of the inverses88

Dispersion curves for InAs QDs ina simple hexagonal SL. The matrixis In 0.5Ga 0.5-xAl xAs with (x = 0.4).There is a large phonon bandgappresent in the system due to thepresence of Al. The gap seemsvery tolerant to Al location. (Farleft): , (Centre): , (Left): directions.Figure 4.5.58Densities of states (DOS) for thedispersion curve immediately to theleft. Note small changes in the DOScould be due to pseudo-randomlocations of the Ga and Al particlesor to differences in crystallinedirection. (Far left): ,(Centre): , (Left): directions respectively.Figure 4.5.59of each atomic mass) was very similar as the lightelement dominates in this case.On this iteration structures with an In 0.5Ga 0.5-xAl xAsmatrix (with x=0.4) and InAs QDs were grown.Significant Al content was introduced into thesestructures with the expectation that this lightelement, segregated to the matrix material,might produce appreciable phonon bandgaps. Aschematic of the structure is shown in Fig. 4.5.56.Some images derived from characterisation ofthe structure are presented in Fig. 4.5.57. Thesuperlattice of QDs has a simple hexagonalstructure. Extraordinary periodic out-of-planestacking is achievable and largely defect freestructures can be grown on the order of microns.Force constant modelling of this structure predictsan appreciable phonon bandgap, as shown inFig. 4.5.58 and Fig. 4.5.59. This bandgap is duealmost entirely to the presence of the Al in thesystem. A small bandgap is present due to themass difference between In, Ga and As, but it isless than a quarter of the size shown in Fig. 4.5.58and Fig. 4.5.59. Due to computational constraintsthe size of the QDs is quite small, only about ananometre in diameter. While actual sizes for thesestructures are too computationally intensive tomodel without extreme effort, recent modellingwith gradually increasing size suggests that thedispersion relations scale linearly with size. Thatis, once the discrete distances (bond lengths) aresmall relative to the superlattice unit cell dimension,the dispersion should look exactly the samewhen scaled such that the relative dimensions arepreserved. This will be investigated in greater detailin further work.4.5.3.3.3.2 Fabrication andcharacterisation of highly orderednanoparticle arrays for HotCarrier absorberResearchers:Lara Treiber, Shujuan Huang,Gavin ConibeerAs demonstrated by the work onmodelling phonon modulation,periodic core-shell QD arrays offera way to significantly change thephonon modulation in a superlatticebecause the core and shell can be of materials ofvery different force constant, directly leading to astrong phonon confinement. Deposition methodsare being investigated to fabricate such highlyordered QD arrays.The Langmuir-Blodgett (LB) technique fabricatesthin films from colloidal dispersions of quantumdots (QDs). The LB technique allows for controllable,uniform film formation and subsequent transferonto a solid substrate, such as quartz [4.5.69].Through surface passivation of the QDs,interparticle spacing can also be controlled. Figure4.5.60 illustrates the set-up of the LB apparatus andthe functionalized QDs.In 2010 work has focused on fabrication of periodicthin films of Si nanoparticles (NPs) provided bythe University of Minnesota. The Si NPs underinvestigation are surface passivated with dodecane(C 12H 26), a long chain hydrocarbon rendering thesurface hydrophobic. The particles were dispersedin chloroform and investigated using TEM, UV-visspectroscopy and photoluminescence (PL). UsingTEM imaging the average diameter was found to be3.0 ±0.5 nm. Fig. 4.5.61 shows a droplet of solution(dispersed in chloroform) and the inset highlightsthe lattice fringes confirming the presence of silicon- a 3.14 Å lattice spacing corresponding to the{111} planes.Langmuir-Blodgett set-up forfilm fabrication with surfacefunctionalized Si NPs.Figure 4.5.60(Upper) TEM images of Si NPs driedon a grid. (Lower) TEM image ofSi NP monolayer deposited at 50mN/m of SP.Figure 4.5.6189

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTabsorption (a.u)40mN/m60mN/m65mN/mcollapse0.5 1.5 2.5 3.5 4.5Energy (eV)(Left) Optical absorption of Si monolayers deposited atvarious surface pressures (SP). (Right) PL of Si NP filmsat varying SP, as compared to the Si NP solution.Figure 4.5.62Conceptual Integrated DeviceFigure 4.5.63External contactSelective energy contact:double (or triple) barrier QD resonant tunnellingSlowed carrier cooling absorber layer:§ Periodic QD array tuned to block optical phonondecay by Klemens mechanism;§ Core shell QD for ease of tuning;§ Must also allow renormalization of carrier energy bye-e & h-h scattering.Proof of concept with MOVPE/MBE III-V QD array.Transfer to thin film or self-assembled QD array.Selective energy contact:selecting opposite carrier typeExternal contactMonolayer structures of Si NPs have beensuccessfully deposited using the LB method andcharacterisation has been on-going using TEM, UVvis,PL and Raman spectroscopy. Figure 4.5.61 (right)illustrates TEM image of the Si films fabricated at 50mN/m surface pressures (SP) of the water trough.As noted, high contrast is difficult to achieve due tothe small density difference of Si and the carbonpassivatingchain. As a result, to further quantifythe coverage and regularity, computer recognitionsoftware is required; on which work is on-going. It isdifficult to see for this reason, but Fig. 4.5.61 (right)appears to show a dense packing of Si NPs.UV-Vis spectroscopy was used to study the opticalabsorption of the films. As illustrated in Fig.4.5.62 (left) for single layers of Si NPs at varyingcompression surface pressures (SP), the absorptionclearly increases when SP increases, correspondingto higher packing densities of Si NPs. The PL spectraof the above films show narrow and symmetricpeaks at 1.6 ~1.7 eV, as shown in Fig. 4.5.62 (right),which are in good agreement with the estimatedbandgap using the ‘effective mass approximation’.The red shift in emission energy from Si NP filmsas compared to the solution corresponds to theelectronic energy transfer from small particles tolarger ones when they are closely packed [4.5.70].Investigation of the electronic and phononicproperties of the monolayer and multilayerstructures are on-going using low-wavenumberRaman and time resolve PL spectroscopy.4.5.3.3.3.3 Implementation ofnanostructures for Hot Carrier cellsIn order to use a nanoparticle array as a Hot Carrierabsorber, and hence utilise slowed carrier coolingdue to modified phonon dispersion, then a wayto fabricate a complete structure with ESCs mustbe established. A structure in which nanoparticlesare arranged in a uniform 3D array should give therequired phononic band gap. The degree to whichthis needs to be ordered is still under investigation.Certainly a large difference between masses isbeneficial between the nanoparticles and thematrix and this can be best achieved in an array of‘core shell QDs’ in which the core and shell have verydifferent acoustic impedance in order to promotecoherent reflection and hence confinement ofphonons [4.5.61]. The matrix in which such an arrayis embedded needs to allow transport of carriers tocontacts and also electron-electron and preferablyhole-hole scattering to renormalise carrier energies.Such a structure should also have a narrowelectronic band gap so as to absorb a wide rangeof photon energies. This combination of propertiesis challenging but not mutually exclusive becausephononic properties are largely independent ofelectronic properties. Energy selective contactsare also required for such a structure. These wouldmost likely be arranged at the top and bottom ofthe absorber. (A hole contact might not need to beselective because the hole population only containsa small fraction of the hot carrier energy and hencethermalisation of holes is less important.) TheseESCs would be QD or QW double barrier structures.(Addition of an extra layer to give double QD/QWtriple barrier structures should also give rectificationat the contact.) Such a QD array / double barrier QDESC structure is shown schematically in Fig. 4.5.63.90

Physical picture of the ND HC SolarCell (upper) with a carrier diffusionfield (core/shell thickness rationot to scale), embedded into awide band gap solar cell [4.5.71].Iso-valent impurities provide anenergy selective level for electrons.Band diagram of the ND HC SolarCell matched to the physicalstructure (lower). Direct generationof hot exciton (1), energy selectiveextraction of hot electron from HotCarrier absorber (2), thermionicemission/diffusion of holes overthe barrier (3), electron-electronscatteringwith renormalizationof carrier energies and possibleenergy loss by phonon emission(4), free carrier re-absorption (5),impact ionization (6) and carriergeneration in the wide band gapmatrix (7).ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTFigure 4.5.64Nanoparticle cell design incorporatingESCs and absorberResearcher:Dirk KönigA full Hot Carrier cell requires control of optical,electronic and phononic properties in the samestructure. One possible approach is the ‘nano-dot(ND) Hot Carrier cell’ embedded into the i-region ofa wide band gap solar cell as shown in Fig. 4.5.64.Hot Carrier solar cells must have a continuouselectronic density of states (DOS) to maximisemaximum photon absorption and elastic electronelectronscattering. The latter refills the depletedelectronic energy levels from which electrons areextracted through the ESC. On the other hand, ifthey do not have a phononic band gap themselvesthe NDs must not be much bigger than the ballisticmean free path in order to prevent inelasticscattering by which carriers cool down rapidly.However, optical phonons can be confined in thenanocrystals as mentioned in Section 4.5.3.3.3by arranging for a mismatch between opticalphonon energies in the two materials. Core shellnanocrystals can be used to realise this [4.5.61,4.5.62]. In addition the shell can be doped withiso-valent impurities which form split-off subbandswithin the electronic band gap of the material,such as boron replacing gallium in gallium nitride(GaN) [4.5.71]. Such an ESC would only work forone carrier type as co-doping of a thin nano-shellappears not to be technologically feasible. For someIII-V compounds with a very high effective massratio between holes and electrons such that most ofthe excess energy during optical generation is takenup by the electron, thermalisation of holes does notrepresent a large loss in energy. Hence thermionicemission of holes over a low barrier can be usedto transport holes into the bulk of the device to becollected by the junction with little loss in efficiency.4.5.3.3.4 Hot Carrier Absorber:Choice of MaterialsResearchers:G. Conibeer, R. Patterson, P. Aliberti, S. Huang,Y. Kamikawa, D. König, Binesh Puthen-Veettil,S.Shrestha, M.A. GreenThe properties required for a good hot carrierabsorber material are listed below in orderof priority.1. Large phononic band gap (E O(min)- E LA) - inorder to suppress Klemens decay of opticalphonons this must be at least as large as themaximum acoustic phonon energy. Hence alarge mass difference (or large force constantdifference) between constituent elements isrequired [4.5.61].2. Narrow optical phonon energy dispersion (E LO-E O(min)) – in order to minimise the loss of energyto TO phonons by Ridley decay. This requiresa high symmetry atomic or nano-structure,preferably cubic with degenerate opticalphonon energies at zone centre.3. Small electronic band gap – to allow a widerange of photon absorption. This should be lessthan 1eV. For 1 sun concentration the optimumis 0.7eV - as a band gap below this energygives no advantage in the balance betweenabsorption and emission. As the concentrationratio increases this optimum band gap decreasesto zero at maximum concentration [4.5.1].4. A small LO optical phonon energy (E LO). Thisreduces the amount of energy lost per LOphonon emission, requiring a greater numberto be emitted for a given energy loss [4.5.62].However, it is difficult to have both a small E LOand the large phononic gap required in 1. Thiscondition requires further investigation, butdoes still argue for a small E LO.5. An absolute small maximum acoustic phononenergy (E LA). This maximises the phononic gap ifE LOis also small. A small E LArequires a large massfor the heavy atom and/or small force constantfor its bonds.6. Good renormalisation rates in the material,i.e. good e-e and h-h scattering (e=electron,91

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTIIA IIIA IB IIB IIIB IVB VB VIBBe B C N OMg Al Si P SCa Sc Cu Zn Ga Ge As SeSr Y Ag Cd In Sn Sb TeBa La Au Hg Tl Pb Bi PoErUse of the periodic table to analyse possible analogue compounds of InNbased on atomic mass combination and electro-negativity.Figure 4.5.65h=hole). This requires a reasonable DOS atall energies above E g, or at least only verynarrow gaps between energy levels, ≤kT 300K. This in turn requires a good overlapof wavefunction for carriers through thematerial, i.e. poor electronic confinementor a reasonable conductor. This condition ismet in all inorganic semiconducotrs quiteeasily, with e-e scattering rates of less than100fs for reasonable carrier concentrations.It may not be met in organicsemiconductors or in nanostructures withlarge barrier heights.7. Good carrier transport in order to allowtransport of hot carriers to the contacts.This is similar to 6, except that it onlyneed be in the direction of the contacts,probably the z growth direction. Areasonably low resistance is probably good.8. It should be possible to make good quality,highly ordered, low defect material.Preferably it should be easy and cheap todo this.9. Earth abundant and readily available andprocessable constituent elements andprocesses.10. No or low toxicity of elements, compoundsand processes.[(E LO, E O(min)& E LAare the maximum optical,minimum optical and maximum acousticphonon energies respectively.]InN has most of these properties, except 4, 8 &9, and is therefore a good model material for ahot carrier cell absorber.Analogues of InNAs InN is a model material, but has theproblems of abundance and bad materialquality, another approach is to use analoguesof InN to attempt to emulate its near idealproperties. These analogues can be II-IV-nitridecompounds, large mass anion III-Vs, group IVcompounds/alloys or nanostructures.II-IV-Vs: ZnSnN; ZnPbN; HgSnN; HgPbNWith reference to Fig 4.5.65, it can be seen thatreplacement of In on the III sub-lattice withII-IV compounds is analogous and is now quitewidely being investigated in the Cu 2ZnSnS 4analogue to CuInS 2[4.5.72].ZnGeN can be fabricated [4.5.73] and is mostdirectly analogous with Si and GaAs. However,its band gap is large at 1.9eV. It also has a smallcalculated phononic band gap [4.5.74]. ZnSnNhas a smaller electronic gap (1 eV) and largercalculated phononic gap [4.5.74]. It is howeverdifficult to fabricate, and also its phononic gapis not as large as the acoustic phonon energymaking it difficult to block Klemens decaycompletely. HgSnN or HgPbN should bothhave smaller E gand larger phononic gaps.These materials have not yet been fabricated[4.5.75].Large mass cation:The Bi and Sb compounds have largepredicted phononic gaps and Bi is a relativelyabundant material, with only low toxicity[4.5.75]. BiB has the largest phononic gapbut AlBi, Bi 2S 5, Bi 2O 3(bismuthine) are alsoattractive. Similarly SbB has a large predictedphononic gap. That for AlSb is the same sizeas the acoustic phonon energy and its bandgap is 1.5eV, making it marginal as an absorbermaterial and similar to InP.Group IIIA III-VsLaN and YN both have large phononic gapswhilst that for ScN is too small.The Lanthanides can also form III-Vs. ErN andother RE nitrides can be grown by MBE. Thephononic band gaps of the Er compounds arepredicted to be large, because of the heavyEr cation, but its discrete energy levels makeit not useful as an absorber, although thecombination of properties in a nanostructurecould be advantageous.Group IV alloy/compounds:All of the combinations Si/Sn, Ge/C or Sn/Clook attractive with large gaps predicted in1D models. However being all group IVs theyonly form weak compounds. UnfortunatelySiC, whether 3C, 4H or 6H, has too narrowa phononic gap. Nonetheless GeC doesform a compound and is of significantinterest [4.5.76].There are also several other inherentadvantages of group IV compounds/alloys allof which are associated with the four valenceelectrons of the group IVs which result inpredominantly covalent bonding:a) The elements form completely covalentlybonded crystals primarily in a diamondstructure (tetragonal is also possible as inβSn). However for group IV compoundsthe decreasing electronegativity down thegroup results in partially ionic bonding.This is not strong in SiC and whilst it tendsto give co-ordination numbers of 4, cannonetheless result in several allotropes ofdecreasing symmetry: 4c, 4h, 6h. However,as the difference in period increases for theas yet theoretical group IV compounds, sotoo does the difference in electronegativityand hence also the bond ionicity and thedegree of order. For a hot carrier absorberthis is ideal because it is just such a largedifference in the period which is neededto give the large mass difference andhence large phononic gaps. All of GeC,SnSi, SnC (and the Pb compounds) havecomputed phononic gaps large enoughto block Klemens decay, and should alsotend to form ordered diamond structurecompounds.b) Because of their covalent bonding, thegroup IV elements have relatively smallelectronic band gaps as compared to theirmore ionic III-V and much more ionic II-VIanalogues in the same period: e.g. Sn0.15eV, InSb 0.4eV, CdTe 1.5eV. In fact toachieve approximately the same electronicband gap one must go down one periodfrom group IV to III-V and down anotherperiod from III-V to II-VI: e.g. Si 1.1eV, GaAs1.45eV, CdTe 1.5eV. This means that forgroup IV compounds there is greater scopefor large mass difference compounds92

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTCross-sectional electron micrograph of a 30-bilayerporous silicon distributed Bragg reflector.Figure 4.5.66whilst still maintaining small electronicband gaps. A small band gap of coursebeing important for broadband absorptionin an absorber - property 3 in the desirableproperties for hot carrier absorbers listedabove.c) The smaller E gwould tend to be for thelarger mass compounds of Pb or Sn. Which,to give large mass difference, would becompounded with Si or Ge. This trendtowards the lower periods of group IV alsomeans that the maximum optical phononand maximum acoustic phonon energieswill be smaller for a given mass ratio - thedesirable properties 4 and 5.d) Furthermore, unlike most groups, thegroup IV elements remain abundant for thehigher mass number elements – desirableproperty 9. Property 10 is also satisfiedbecause the group IVs have low toxicity.Nanostructures:As discussed in section 4.5.3.3.3 QDnanostructures can be viewed in the sameway as compounds. Their phononic propertiescan be estimated from consideration of theircombination force constants. Hence it ispossible to ‘engineer’ phononic properties ina wider range of nanostructure combinations.Of the materials discussed above the Group IVslend themselves most readily to formation ofnanostructures instead of compounds due totheir predominantly covalent bonding, whichallows variation in the coordination number.Therefore the nanostructure approaches ofsection 4.5.3.3.3 are consistent with a similardescription as analogues of InN, whether it beIII-V QDs, colloidally dispersed QDs or for coreshell QDs.4.5.3.3.5 Summary of Hot Carriersolar cell research2010 has seen significant development inmost areas of Hot Carrier solar cell work.The modelling of efficiencies not only nowincludes real material parameters for highlypromising absorber materials such as InN,but also is now extended to Energy Selectivecontacts of finite width. The transport acrosssuch contacts is further modelled for a rangeof QD matrix combinations. Work on ESCs hasseen further more detailed demonstrationof the necessary resonance in double barrierresonant tunnelling structures, with additionalevidence for hot carrier populations, albeitvery small, from illuminated I-V measurementswith the ‘optically assisted I-V’ technique.Measurement of InN with time-resolved PLhas indicated some evidence for slowedcarrier cooling, further corroborating theimportance of a large phonon band gapto block optical phonon decay, but alsohighlights the importance of materialquality. Modelling of nanostructures forabsorber materials has focussed on real III-VQD structures, showing phonon bandgapswhich will soon be measured in phonondispersion measurements. Progress on theLangmuir-Blodgett approach to orderednanoparticle arrays has seen development ofordered single layer arrays of Si nanoparticles.The potential application of nanostructuresto fully integrated devices has started tobe investigated conceptually, with variousdesigns considered. Similarly the possibilityof absorber materials which are analogousto InN is also being investigated. These manyaspects of Hot Carrier cells will see furtherdevelopment and consolidation in 2011with recent success in significant additionalfunding.4.5.4 Up-conversionResearchers:Craig Johnson, Gavin ConibeerCollaboration with:Peter Reece (Physics, UNSW)Up-conversion in novel silicon-basedmaterialsUp-conversion (UC) in erbium-dopedphosphor compounds (particularly NaYF 4:Er)has been shown to be a promising meansof enhancing the sub-band-gap spectralresponse of conventional Si solar cells withoutmodification of the electrical properties of thecell [4.5.77]. In this scheme, a layer containingthe phosphor is applied to the rear of a highefficiencybifacial cell. After absorbing twolong-wavelength (~1500nm) photons - whichare transmitted by the cell - the excited Erions can relax by emitting a photon with anenergy greater than the Si band gap, therebyincreasing the current that can be extractedfrom the cell.While phosphors have demonstratedhigh-efficiency UC behaviour, their usepresents particular challenges with regardto fabrication and cost. Our work in the lastyear has focused on the development ofEr-doped porous Si (PSi:Er) as an alternativeUC material. PSi is unique in that its porosity -and hence the material refractive index - canbe varied as a function of depth, allowing forthe elaboration of high-quality monolithic Sioptical structures such as distributed Braggreflectors (DBRs). A cross-sectional electronmicrograph of such a structure is shown in Fig.4.5.66. Its porous substructure also allows fordeep infusion of dopant atoms via techniquessuch as electroplating.93

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTSimulated reflectivity spectra ofPSi DBRs with varying numbersof bilayers; layer thicknesseswere adjusted to fix the shortwavelengthstop band edgeat 1550nm (marked with reddashed vertical line).Figure 4.5.68Band structure in the neighborhoodof the fundamental band gap ofan infinite 1D photonic crystal(left) and group velocity in thematerial (right).Figure 4.5.67We have previously reported room-temperature UCluminescence from Er-doped nanoporous Si whenexcited at very high power densities with a 1550nmlaser. In the past year we have focused on increasingthe efficiency of this UC luminescence by modifyingthe PSi structure as described above.The dependence of the efficiency of the UC processon irradiance is non-linear in the range of “sunlike”incident power. That is, the concentrationof the incident radiation into a small Er-dopedregion results in greater efficiency than for nonconcentratedradiation over a proportionally largeareadevice. Typically, concentration is achievedusing lenses or non-imaging optics, but we haveexamined an analogous electromagnetic fieldenhancement in the vicinity of optically-active Erions by the excitation of slow-light modes in thesePSi:Er DBRs.Field enhancement in 1D PSi:Erphotonic crystalsThough their use preceded the advent of thephotonic crystal concept, DBRs can be describedas one-dimensional photonic crystals. The regularvariation of the optical thickness of multiplealternating layers results in wavelength-dependentinterference behaviour manifesting as discretebands of allowed and disallowed electromagneticmodes, as shown in the left panel of Fig. 4.5.67. The“edge” of a photonic band is a spectral region inwhich the effect of coherent scattering processeschanges abruptly between transmission andreflection of incident photons as determined bythe superposition of Bloch waves. Insofar as thegroup velocity v gof an incident wave can be saidto transition between “positive” and “negative”values - implying energy propagation into and outof the structure, respectively, in bands of strongtransmission and strong reflection - there exists aninversion point at which v gis zero, corresponding toa standing wave established by the superposition ofscattered Bloch components. As can be seen in theright panel of Fig. 4.5.67, group velocity is drasticallysuppressed across a broad region near the zeropoint.For steady-state equilibrium conditions,this slowing of energy propagation requires aproportional increase in energy density, that is,an enhancement of the electromagnetic fieldstrength within the structure. It can be shown thatthe efficiency increase for a two-step UC process isproportional to (c/nv g– 1)I 0, where c is the speed oflight in vacuum, n is the average refractive index ofthe multilayer material, and I 0is the incident fieldintensity [4.5.78]. In this way the slow-light modeacts to augment the interaction that convertsenergy from the field into atomic potential energy,resulting in a boost in the efficiency of the processuntil saturation is reached.Simulation of field enhancement in 1DPC structuresUsing a 1D transfer matrix calculation, we simulatedthe reflectivity characteristics of a series of PSi DBRswith 10, 20, 30 and 40 high-/low-porosity bilayers.As shown in Fig. 4.5.68, the reflectivity characteristicof a DBR contains a wide photonic stop-band.The structural parameters were tuned slightly ineach case as the number of bilayers increased toproduce a short-wavelength band edge minimumat 1550nm, the excitation wavelength of interest(marked with a dashed red line).Maximum transmission of ~1550nm light into thestructure is clearly required for efficient couplinginto the Er ions. From Fig. 4.5.68 it is evident thatan increased number of layers “compresses” thereflectivity characteristic, resulting in a steeper94

Intensity-valued mode profile in DBRswith varying numbers of high/lowporositybilayers for a leading band edgefixed at ~1550nm. All intensities arenormalized to the incident field intensityand each plot is independently colorscaledto show maximum detail. The airand substrate interfaces with the DBR areindicated by the dashed black (upper)and red (lower) lines, respectively.ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTFigure 4.5.69band edge and a narrower transmission windowbetween the stop band and the neighboringinterference fringe.Figure 4.5.69 shows the calculated field intensityprofile for these DBRs in the spectral regionbetween 1500 and 1600nm. The results havebeen normalized to the incident field intensityand each plot is independently colour-scaled toshow maximum detail. They clearly demonstratethat considerable field enhancement is possiblethroughout the depth of such a multilayer structure,with peak relative intensities of more than 11 forthe 30-bilayer structure and more than 18 for the40-bilayer structure. However, it is also clear that theregions of enhancement are increasingly narrowlyconcentratedas the number of bilayers increases.This is due to the additional interference fringesin the reflectivity characteristic of the DBR thatarise with additional layers, as is apparent from Fig.4.5.69. Calculations for more than 40 layers showthat enhancements of up to 80 are possible for alarge number of bilayers though the peak narrowsto less than 1nm wide for 100 bilayers.ConclusionThese simulations show that drastic field intensityenhancement is in principle achievable in Er-dopedPSi DBRs. Our current work is focused on thefabrication of suitable samples by anodic chemicaletching, electroplating, and high-temperatureannealing. Angular-dependent room temperaturephotoluminescence studies are underway toprobe the effect of the photonic band edge on theefficiency of UC in the optically-active Er ions withinthe structures. This Si-based approach to efficientup-conversion is conceptually promising and byoptimising our fabrication techniques to improvethe quality of our optical structures we expect it willprove to be valuable work.4.5.5 Plasmonics for 3 rdGeneration structuresResearcher:Supriya PillaiPlasmonics is the study of the interaction of lighton a thin metal dielectric interface resulting inthe collective oscillation of the free electrons inthe metal. This gives rise to exciting propertiesthat are very different from the bulk metal. It isan emerging area of technology for photovoltaicapplications with potential for light managementat the nanoscale. Ever since our proof-ofconcept results were published in 2007 [4.5.79],there has been a huge increase in interest inincorporating nanoparticles on solar cells dueto their demonstrated light trapping properties.The large optical polarisation associated with thenanoparticles results in a strong enhancement ofthe electric field around the vicinity of the particlesalong with strong scattering. This optical property isbeing utilised for improving light trapping in thinfilmsolar cells.The scattering or near field enhancement isstrongest at the surface plasmon resonance (SPR)frequency and this can be tuned by changing thesize, shape, dielectric medium around the particles.SPR tunability offers control in manipulating lightin the regime where the absorption of the solarcells are weak. Our recent work (see section 4.4Thin Film) has shown the benefits of having theparticles on the rear as opposed to previous workwith front located particles [4.5.80]. By having theparticles on the back of the cell any potential lossdue to absorption in the metal or losses due tosub-resonance photocurrent suppression resultingfrom destructive interference of the scattered andtransmitted light into the semiconductor layercan be avoided [4.5.81]. An additional advantageis that because the scattering plasmonics layeris electrically decoupled from the cell, theantireflection layer on the front and the plasmons95

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTIncreasing order of wavelengthDecreasing order of bandgap energyof the semiconductorIncreasing order of particle sizeSchematic representation showingan example of surface plasmonresonance tunability by varying theof size of the metal nanoparticlesfor wavelength dependent lighttrapping in a tandem solar cell.Figure 4.5.70can be optimised independently of each other. Ourrecent work has also made use of an additionalreflecting layer (paint) along with the nanoparticlelayer on the rear side of the cell to maximise therandomisation of light and to ensure no light islost [4.5.80].In this study we focus our interest on the use ofplasmonics for low-dimensional structures likequantum dot structures for possible applicationsin tandem solar cells. PL measurements wereused to characterise the samples. The analysisis based on the reciprocity of light that a goodabsorber can be a good emitter and the proofofconcept established with EL measurementsin an earlier study [4.5.82]. Hence the PLenhancement would be synonymous to anincrease in photocurrent for an optimised QD cell.The tunability of plasmon resonance allows thepossibility of increasing the photocurrent of eachcell in a tandem cell configuration. Optimisingthe nanoparticle parameters can help achievewavelength dependent light trapping which willbe an encouraging step towards increasing theperformance of a tandem solar cell. Figure 4.5.70shows one way of achieving the SPR tunability byvarying the size of the nanoparticles in a tandemcell configuration.The sample under study is a 4nm single layer siliconQD structure in an oxide matrix with a 6nm cappingoxide layer with an emission wavelength of 930nm(~1.3eV). Mass thicknesses of 10, 14, 18 and 22 nmsilver were deposited and the particles annealedat 200°C in nitrogen for an hour. A 532nm, 10mWNd:YAG laser was used to illuminate the samplesfor both the front (incident on QD layer) and rearside (incident on the quartz slide) locations of thesilver nanoparticles and the PL plots studied. PLmeasurements were carried out on these samplesand the effect of different size particles wereinvestigated both before and after the deposition ofthe nanoparticles.The self assembly techniques of nanoparticlefabrication give a varied size and shape distributionvery suitable for a broadband response. This isparticularly noted for particles corresponding tothicker Ag layers as can be seen from the PL plots inFig. 4.5.71. The red-shift of the enhancement peakswith larger particles are also clearly evident. Moredetails of this work has been reported in [4.5.83].Rear located nanoparticles perform better thanfront located nanoparticles, consistent with theresults from thin-film cells. We believe this isbecause of the change in scattering cross-section ofthe nanoparticles due to the changes in the drivingfield and also due to the change in the mode ofexcitation (from air for front located nanoparticlesand from Si for rear located nanoparticles) forthe two different locations on the samples. Morework need to be done to better understand themechanism. Further work will also concentrateon using alternate approaches for tuning theresonance position close to the band-edge(emission wavelength for PL) to increase thescattering properties and hence absorption in theweakly absorbing range.4.5.6 Concluding remarks for the ThirdGeneration sectionIn 2010 work has proceeded significantly in all theareas of Third Generation research, with improvedfabrication and characterisation of materials andcomplexity of modelling which together give anoverall better understanding and optimisationof devices.Group IV based nanostructure materials have seensignificant improvement in device design. Withinterlayers, in both SiN xand SiC matrices, used toeffectively control both the uniformity of size of SiQDs and carrier transport in the growth direction.Modelling now allows the confined energy levelsof complex QD shapes to be calculated and theiroverlap to form mini-bands estimated in moredetail. At the same tiome improved characterisationis giving a much clearer idea on the shape anddistributions of Si QDs and is now allowing choiceof growth parameters to give uniform size sphericalQDs to be grown. Ge nanostructure materials haveseen improved control of p-type conductivity in96

PL enhancementPL Enhancement76no MNP (a) Front10nm Ag14nm Ag18nm Ag522nm Ag43210600 700 800 900 1000Wavelength (nm)20no MNP (b) Rear18Ag reflector10nm Ag1614nmAg18nm Ag1422nm Ag121086420600 700 800 900Wavelength (nm)10001.00.90.80.70.60.50.40.30.20.10.03.22.82.42.01.61.2PL (a.u)PL (a.u)0.80.40.0PL plots from a 4nm singlelayer QD structure in an oxidematrix with a 6nm cappingoxide layer (right axis) alongwith the correspondingenhancement plots (left axis)with silver nanoparticlescorresponding to 10, 14, 18and 22 nm mass thickness ofsilver for (a) silver nanoparticleson the front and (b) silvernanoparticles on the rear.Figure 4.5.71ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTGe QDs and very highly crystalline growth of GeQWs. Homojunction Si nanostructure deviceshave demonstrated improved conduction, therehas been continued demonstration of rectifyingproperties for heterojunctions with c-Si andheterojunctions between two nanostructuredmaterials are now starting to exploit the advantagesof both.Hot carrier cells have seen further developmentof efficiency modelling which now not onlyincludes material parameters such as Augercoefficients for real promising absorber materialssuch as InN, but also includes the effects of realnon-ideal energy selective contacts. There hasbeen significant improvement in measurementof negative differential resistance in Si QD basedenergy selective contacts is now allowing and thisoverlaps well with continuing improvement inmodelling the transport through such structuresin various matrices. Carrier cooling measurementson the candidate absorber material InN haveshown evidence for phonon bandgaps building onthe slowed carrier cooling seen in InP previously.Modelling of analogues of these materials innanostructures now predicts the phonon propertiesof rea lIII-V QD arrays grown by collaborators. Inaddition colloidally dispersed Si nanoparticleshave now been fabricated as coherent single layernanoparticle arrays and are progressing towardsmulti-layer structures. Theoretical work on thepotential materials for absorbers with controlledphon dispersion and hence slowed carrier coolinghas progressed on approapratie analogues forInN and devices involving various nanostructureconfigurations are being analysed conceptually.Up-conversion has seen a further significantdevelopment of the use of porous-Si as a host for Erup converting species. The porosity of porous-Si canbe controlled to give layers of alternating refractiveindex which can be used as distributed Braggreflectors to modifiy the photonic density of statessuch that there is strong enhancement of photonicmodes at the critical Er absorption at 1500nm.This has been demonstrated and can lead directlyto large increase in the up-conversion efficiency,thus tackling one of the key problems of rare earthup-conversion that it is inherently non-linear anddependent on concentration.Plasmonics, whic has previously been shown toenhance emission and absorption from both 1 stand 2 nd generation cells, has now been appliedto the Si annostructured layers of 3 rd generatrionmaterials. This has shown significant enhancementin luminescence demonstrating the strong potentialto achieve local concentration with a global 1 sunillumination. This is particularly useful for several ofthe non-linear 3 rd generation approaches.The development of all the 3 rd generation projectsin 2010 now allows much greater understandingof the materials and devices. Work in 2011 will seeconsolidation of this into improved devices. Severalnew areas of funding will contribute to this and alsoallow development of new project areas.References:4.5.1 M. A. Green, “Third Generation Photovoltaics:Ultra-High Efficiency at Low Cost” (Springer-Verlag, 2003).4.5.2 J. Nelson, “The Physics of Solar Cells”, ImperialCollege press, 2003.4.5.3 G. Conibeer, “Third Generation Photovoltaics”,Materials Today, 10 (11) (2007) 42-50.4.5.4 E-C. Cho, S. Park, X. Hao, D. Song, G. Conibeer,S-C. Park, M.A. Green, “Silicon quantum dot/crystalline silicon solar cells”, Nanotechnology, 19(2008) 2452014.5.5 Hao X J, Podhorodecki A, Shen Y S, Zatryb G,Misiewicz J, Green M A, “Effects of Si-rich oxidelayer stoichiometry on the structural andoptical properties of Si QDs/SiO 2multilayer film”,Nanotechnology 20, 485703 (2009).4.5.6 M.A. Green, G. Conibeer, D. König, I. Perez-Wurfl,A. Gentle, S. Huang, D. Song, X.J. Hao, F. Gao, S.Park, C. Flynn, Y. So, B. Zhang, D. Di, P. Campbell,Y. Huang and T. Puzzer, “Progress with All-SiliconTandem Cells On Glass Using Silicon QuantumDots in Silicon-Based Dielectric Matrices”, 24thEuropean Photovoltaic Solar Energy Conference,Hamburg, Germany, 21-25 September 2009.4.5.7 M. Zacharias, J. Heitmann, R. Scholz, U. Kahler, M.Schmidt, Appl. Phys. Lett. 80(2002) 661.4.5.8 X.J. Hao, I. Perez-Wurfl, G. Conibeer, M.A. Green,19th PVSEC, Korea, Nov 2009.97

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORT4.5.9 I. Perez-Wurfl, X. Hao, A. Gentle, D. Kim, G.Conibeer, M.A. Green, “Si nanocrystal p-i-ndiodes fabricated on quartz substrates for thirdgeneration solar cell applications,”Applied PhysicsLetters, 95 (2009) 153506.4.5.10 X.J. Hao, E-C. Cho, G. Scardera, Y.S. Shen, E.Bellet-Amalric, D. Bellet, G. Conibeer, M.A. Green,“Phosphorus-doped silicon quantum dots forall-silicon quantum dot tandem solar cells”, SolarEnergy Materials and Solar Cells, 93 (2009) 1524.4.5.11 G. Conibeer, M.A. Green, R. Corkish, Y.-H. Cho, E.-C.Cho, C.-W. Jiang, T. Fangsuwannarak, E. Pink, Y.Huang, T. Puzzer, T. Trupke, B. Richards, A. Shalav,K.-L. Lin, Thin Solid Films 511/512 (2006) 654.4.5.12 E.-C. Cho, Y.H. Cho, T. Trupke, R. Corkish, G.Conibeer and M.A. Green, Proc. 19th EuropeanPhotovoltaic Solar Energy Conference, Paris, (June2004), p. 235.4.5.13 F. Meillaud, A. Shah, C. Droz, E. Vallat-Sauvain,C. Miazza, Sol. Energy Mater. Sol. Cells 90,29522 (2006).4.5.14 T. Arguirov, T. Mchedlidze, M. Kittler, R. Rölver, B.Berghoff, M. Först, B. Spangenberg, Appl. Phys.Lett. 89 (2006) 053111.4.5.15 T.-W. Kim, C.-H. Cho, B.-H. Kim, S.-J. Park, Appl. Phys.Lett. 88 (2006) 123102.4.5.16 Y.-H. Cho, E.-C. Cho, Y. Huang, C.-W. Jiang, G.Conibeer, M.A. Green, Proceeding of 20thEuropean Photovoltaic Solar Energy Conference,Barcelona, Spain, June 6-10, 2005, p.47.4.5.17 Y. Kurokawa, S. Miyajima, A. Yamada and M.Konagai, Jpn. J. Appl. Phys. 45, L1064 (2006).4.5.18 C.-W. Jiang, M.A. Green, J. Appl. Phys. 99(2006) 114902.4.5.19 K. Boer, “Survey of Semiconductor Physics”, vanNostrand Reinhold, 1990.4.5.20 Aliberti, P., et al., “Study of silicon quantum dotsin a SiO 2matrix for energy selective contactsapplications”, Solar Energy Materials and SolarCells. 94(11): p. 1936-1941 (2010).4.5.21 E.A.B.Cole, “Mathematical and NumericalModelling of Heterostructure SemiconductorDevices: From Theory to Programming”. 2009:Springer. 406.4.5.22 Grinstein, F.F., H. Rabitz, and A. Askar, “TheMultigrid Method for Accelerated Solution ofthe Discretized Schrodinger Equation”. Journal ofComputational Physics, 1983. 51: p. 423-443.4.5.23 D. König, J. Rudd, M.A. Green, G. Conibeer, “Role ofthe interface for the electronic structure of siliconquantum dots”, Physical Review B, 2008. 78(3):p. 035339.4.5.24 Yurtsever, A., M. Weyland, and D.A. Muller,“Three-dimensional imaging of nonsphericalsilicon nanoparticles embedded in silicon oxideby plasmon tomography”, Applied Physics Letters,2006. 89(15): p. 151920-151920-3.4.5.25 L.-W. Wang and A. Zunger, J. Phys. Chem. 98,2158 (1994).4.5.26 Y-H. So, M.A. Green, G. Conibeer, S. Huang, A.Gentle, “Size dependent optical properties of Siquantum dots in Si-rich nitride/Si3N4 superlatticesynthesized by magnetron sputtering”, AppliedPhysics Letters, accepted Feb 2011.4.5.27 D. Di, I. Perez-Wurfl, G. Conibeer, M.A. Green, SolarEnergy Materials & Solar Cells 94 (2010) 2238-2243.4.5.28 G. Conibeer, M.A. Green, D. Konig, I. Perez-Wurfl,S. Huang, X. Hao, D. Di, L. Shi, S. Shrestha, B.Puthen-Veettil, Y. So, B. Zhang and Z. Wan, “Siliconquantum dot based solar cells: addressing theissues of doping, voltage and current transport”,Progress in Photovoltaics: Res. Appl. 18 (2010) Inpress, Published online in Wiley Online Library(wileyonlinelibrary.com). DOI: 10.1002/pip.1045,accepted Sept 20104.5.28 D. Song, E. Cho, Y. Cho, G. Conibeer, Y. Huang,S. Huang, MA. Green, “Evolution of Si (and SiC)nanocrystal precipitation in SiC matrix”, ThinSolid Films 516 (2008) 3824–3830, doi:10.1016/j.tsf.2007.06.1504.5.29 D. Song, E-C. Cho, G. Conibeer, Y. Huang, C. Flynn,MA. Green, “Structural characterization of annealedSi1-xCx/SiC multilayers targeting formation of Sinanocrystals in a SiC matrix”, Journal of AppliedPhysics, 103 (2008) 083544.4.5.30 Z. Wan, S. Huang, M.A. Green, G. Conibeer, PhysicaStatus Solidi C, “Residual stress study of siliconquantum dot in silicon carbide matrix by Ramanmeasurement”, EXCON ‘10 Proceedings specialissue, accepted September 2010.4.5.31 Zhenyu Wan, Shujuan Huang, Martin Greenand Gavin Conibeer, Rapid thermal annealingand crystallization mechanisms study of siliconnanocrystal in silicon carbide matrix, NanoscaleResearch Letters, 7541289134788882, acceptedJanuary 2011.4.5.32 F. Gao, M.A. Green, G. Conibeer, E-C. Cho, Y. Huang,I. Perez-Wurfl, C. Flynn, “Fabrication of multilayeredGe nanocrystals by magnetron sputtering andannealing”, Nanotechnology, 19 (2008) 455611.4.5.33 Richter et al., Solid State Commun., 39 (1981), 625;I.H. Campbell, P.M. Fauchet, Solid State Commun.58 (1986) 739; Y. Sasaki, C. Horie, Phys. Rev. B 47(1993) 3811.4.5.34 G. Conibeer, M. Green, E-C. Cho, D. König, Y-H. Cho,T. Fangsuwannarak, G. Scardera, E. Pink, Y. Huang, T.Puzzer, S. Huang, D. Song, C. Flynn, S. Park, X. Hao,D. Mansfield, “Silicon quantum dot nanostructuresfor tandem photovoltaic cells”, Thin Solid Films, 516(July 2008) 6748.4.5.35 B. Zhang, S.K. Shrestha, P. Aliberti, M.A. Green, G.J.Conibeer, Characterisation of size-controlled andred luminescent Ge nanocrystals in multilayeredsuperlattice structure, Thin Solid Films 518 (2010)5483-5487.4.5.36 B. Zhang, S.K. Shrestha, M.A. Green, G.J. Conibeer,“Size controlled synthesis of Ge nanocrystals inSiO 2at temperatures below 400 degree C usingmagnetron sputtering”, Applied Physics Letters, 96,26901 (2010).4.5.37 Y. Sasaki, C. Horie, Phys. Rev. B 47 (1993) 3811;Kartopu et al., J. Appl. Phys. 103 (2008) 113518.4.5.38 Karmous et al., Phys. Rev. B 73 (2006) 075323.4.5.39 B. Zhang, S. Shrestha , M.A. Green , G. Conibeer,Surface states induced high p-type conductivityin nanostructured thin film composed of Genanocrystals in SiO 2matrix, Applied PhysicsLetters, 97, 132109 (2010).4.5.40 X.J. Hao, E-C. Cho, C. Flynn, Y.S. Shen, S.C. Park,G. Conibeer, M.A. Green, Solar Energy Materials& Solar Cells, “Synthesis and characterization ofboron-doped Si quantum dots for all-Si quantumdot tandem solar cells”, 93 (2009) 273–279.4.5.41 I. Perez-Wurfl, A. Gentle, X. Hao, M.A. Green, G.Conibeer and D-H. Kim, Proceedings of 24thEPVSEC, Hamburg, Germany, (2009).4.5.42 J.M. Shah, Y.-L. Li, Th. Gessmann, E. F. Schubert.,Journal of applied physics. 94 (2003) 2627.4.5.43 L. Ma, D. Lin, G. Conibeer, I. Perez-Wurfl,“Introducing dopants by diffusion to improve theconductivity of silicon quantum dot materialsin 3rd generation photovoltaic devices”, PhysicaStatus Solidi, accepted Sept104.5.44 G. Cantele, E. Degoli, E. Luppi, R. Magri, D. Ninno,G. Iadonisi, S, Ossicini S, “First-principles study of n-and p-doped silicon nanoclusters”, Physical ReviewB, 72 (2005) 113303; G.M. Dalpian, J.R. Chelikowsky,“Self-Purification in Semiconductor Nanocrystals”,Physical Review Letters, 96 (2006) 226802.4.5.45 A.R. Stegner, R.N. Pereira, R. Lechner, K. Klein,H. Wiggers, M. Stutzmann, Brandt MS, “Dopingefficiency in freestanding silicon nanocrystals fromthe gas phase: Phosphorus incorporation anddefect-induced compensation”, Physical Review B,80 (2009) 165326.98

4.5.46 A.L. Tchebotareva, M.J.A. de Dood, J.S. Biteen, H.A.Atwater, A. Polmana, “Quenching of Si nanocrystalphotoluminescence by doping with gold orphosphorous”, Journal of Luminescence, 114(2005) 137-144.4.5.47 M. Fujii, A. Mimura, S. Hayashi, K. Yamamoto, C.Urakawa, H. Ohta, Journal of Appl. Phys., 87 (2000)1855-1857.4.5.48 D. König, S. Shrestha, J. Rudd, G. Conibeer,M. A. Green, “Doping of Si-Based Dielectricsfor Providing Majorities to Si Quantum Dots:Acceptors in SiO 2”, Proc. 24 th European PVSEC,(Hamburg 2009).4.5.49 G. Conibeer, M.A. Green, D. Konig, I. Perez-Wurfl,S. Huang, X. Hao, D. Di, L. Shi, S. Shrestha, B.Puthen-Veettil, Y. So, B. Zhang and Z. Wan, “Siliconquantum dot based solar cells: addressing theissues of doping, voltage and current transport”,Progress in Photovoltaics: Res. Appl. 18 (2010) Inpress, Published online in Wiley Online Library(wileyonlinelibrary.com). DOI: 10.1002/pip.1045,accepted Sept 2010.4.5.50 Santosh K. Shrestha, Pasquale Aliberti, Gavin J.Conibeer, Energy selective contacts for hot carriersolar cells, Solar Energy Materials and Solar Cells,94 (Sept 2010) 1546-1550.4.5.51 P. Würfel, Sol. Energy Mats. and Sol. Cells. 46(1997) 43.4.5.52 R. Ross, A.J. Nozik, J. Appl Phys (1982) 53, 3318.4.5.53 V.Y. Davydov, et al., “Absorption and emission ofhexagonal InN: Evidence of narrow fundamentalband gap”, Physica Status Solidi (B), 229 (2002) p.r1-r3.4.5.54 D. Fritsch, H. Schmidt, M. Grundmann, “Banddispersion relations of zinc-blende and wurtziteInN”, Physical Review B, 69 (2004) 165204.4.5.55 Y. Takeda, T. Ito, T. Motohiro, D. König, S.K. Shrestha,G. Conibeer, “Hot carrier solar cells operating underpractical conditions”, Journal of Applied Physics,105 (2009) 074905-10.4.5.56 Wurfel, P., A.S. Brown, T.E. Humphrey, and M.A.Green, Particle conservation in the hot-carrier solarcell”, Progress in Photovoltaics, 13 (2005) 277-285.4.5.57 P. Aliberti, Y. Feng, Y. Takeda, S.K. Shrestha, M.A.Green, G.J. Conibeer, “Investigation of theoreticalefficiency limit of hot carrier solar cells with bulkInN absorber”, Journal of Applied Physics, 108(2010) 094507-10.4.5.58 G. Conibeer, C.W. Jiang, D. König, S.K. Shrestha, T.Walsh, M.A. Green, “Selective energy contacts forhot carrier solar cells”, Thin Solid Films, 216 (2008)6968-6973.4.5.59 P.G. Klemens, Phys. Rev. 148 (1966) 845.4.5.60 G. Conibeer, R. Patterson, P. Aliberti, L. Huang, J-F.Guillemoles, D. König, S. Shrestha, R. Clady, M.Tayebjee, T. Schmidt, M.A. Green “Hot Carrier solarcell Absorbers”, 24th European PVSEC, Hamburg(Sept. 2009).4.5.61 G. Conibeer, N.J. Ekins-Daukes, J-F. Guillemoles, D.König, E-C. Cho, C-W. Jiang, S. Shrestha and M.A.Green, “Progress on Hot Carrier Cells”, Solar EnergyMaterials and Solar Cells, 2009, 93, 713.4.5.62 D. König, K. Casalenuovo, Y. Takeda, G. Conibeer, J.F.Guillemoles, R. Patterson, L.M. Huang, M.A. Green,“Hot Carrier Solar Cells: Principles, Materials andDesign”, Physica E, 42 (2010) 2862-2866.4.5.63 H. Bilz, W. Kress, “Phonon dispersion relations inInsulators”, Springer-Verlag (1979).4.5.64 R. Clady, M.J.Y. Tayebjee, P. Aliberti, D. Konig, N.J.Ekins-Daukes, G.J. Conibeer, T.W. Schmidt, M.A.Green, “Interplay between Hot phonon effect andIntervalley scattering on the cooling rate of hotcarriers in GaAs and InP”, Progress in Photovoltaics,accepted March 2011.4.5.65 K.T. Tsen, J.G. Kiang, D.K. Ferry, H. Lu, W.J. Schaff,H.W. Lin, S. Gwo, “Direct measurements of thelifetimes of longitudinal optical phonon modesand their dynamics in InN”, Applied Physics Letters,90 (2007) 3.4.5.66 Y.C. Wen, Y.C., C.Y. Chen, C.H. Shen, S. Gwo, C.K. Sun,“Ultrafast carrier thermalization in InN”, AppliedPhysics Letters, 89 (2006) 3; M.D. Yang, et al.,“Hot carrier photoluminescence in InN epilayers”,Applied Physics A - Materials Science & Processing,90 (2008) 123-127.4.5.67 F. Chen, A.N. Cartwright, H. Lu, W.J. Schaff, “Ultrafastcarrier dynamics in InN epilayers”, Journal ofCrystal Growth, 269 (2004) 10-14.4.5.68 R. Patterson, M. Kirkengen, B. Puthen Veettil, D.Konig, M.A. Green, G. Conibeer, Sol. Ener. Mats. AndSol. Cells, 94 (2010) 1931-1935.4.5.69 S. Huang, K. Minami, H. Sakaue, S. Shingubara, T.Takahagi, “Effects of the surface pressure on theformation of Langmuir-Blodgett monolayer ofnanoparticles”, Langmuir, 20 (2004), 2274.4.5.70 C. R. Kagan, C. B. Murray, M. Nirmal, M. G. Bawendi,Phys. Rev. Lett. “Electronic Energy Transfer in CdSeQuantum Dot Solids”, 76 (1996) 1517.4.5.71 D. König, “Photovoltaic Device Physics at theNanoscale”, Chapter 3 (pp. 73 - 146) in L. Tsakalakos(Ed.), “Nanotechnology for Photovoltaics”, Taylor &Francis, April 2010; ISBN-13: 97814200767454.5.72 T.K. Todorov, K.B. Reuter, D.B. Mitzi, AdvancedMaterials, 22 (21010) 1–4.4.5.73 W.R. L. Lambrecht, “Structure and phonons ofZnGeN 2”, Phys. Rev. B 72 (2005) 1552024.5.74 T.R. Paudel, W.R.L. Lambrecht, Phys. Rev. B, 79(2009) 245205.4.5.75 S.C. Erwin, I. Zutic, Nature Materials, 3 (2004)410‐414.4.5.76 Z.T. Liu, J.Z. Zhu, N.K. Xu, X.L. Zheng, Jap. J. Appl.Phys., 36 (1997) 3625.4.5.77 A. Shalav, B.S. Richards, M.A. Green, “Luminescentlayers for enhanced silicon solar cell performance:Up-conversion”, Sol. Energy Mats. and Solar Cells,91 (2007) 829-842.4.5.78 C.M. Johnson, P.J. Reece, G.J. Conibeer, “Upconversionluminescence enhancement inerbium-doped porous silicon photonic crystals forphotovoltaics”, Photonics West Conference, SanFrancisco, Jan 2011.4.5.79 S. Pillai, K. R. Catchpole, T. Trupke, and M. A. Green,J. Appl. Phys., 101 (2007) 093105.4.5.80 Z. Ouyang, S. Pillai, F. Beck, O. Kunz, S. Varlamov, K.R. Catchpole, P. Campbell, and M. A. Green, App.Phys. Lett., 96 (2010) 261109.4.5.81 S. H. Lim, W. Mar, P. Matheu, D. Derkacs, and E. T. Yu,J. Appl. Phys., 101 (2007) 104309.4.5.82 S. Pillai, K. R. Catchpole, T. Trupke, G. Zhang, J. Zhao,and M. A. Green, Appl. Phys. Lett. 88 (2006) 161102.4.5.83 S. Pillai, I. Perez-Wurfl, G. J. Conibeer, and M. A.Green, physica status solidi (c) 8 (2011) 181.ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORT99

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORT4.6 Photonics and deviceCHARACTERISATION4.6.1 Photoluminescence basedcharacterisation of siliconUniversity StaffA/Prof. Thorsten TrupkeProject Scientists andTechniciansAllen YeeUndergraduate StudentsMatthias JuhlPostgraduate Students• Yael Augarten• Henner Kampwerth• Bernhard Mitchell• Li HuaExternal Collaborators• Ron Sinton, Sinton Consulting• BT Imaging Pty Ltd• Otwin Breitenstein, Jan Bauer, MaxPlanck Institute for MicrostructurePhysics, Halle, Germany4.6.1.1 BackgroundIn photoluminescence (PL) imaging an intenselight source is used to illuminate large areasamples such as silicon bricks, wafers orcomplete solar cells homogeneously withtypically one-sun equivalent illuminationintensity. A CCD camera captures a highresolution picture of the luminescent lightthat the sample emits. Figure 4.6.1.1 shows aschematic diagram of a luminescence imagingsystem. The power supply that is shown onthe bottom left can be used as an excitationsource in electroluminescence (EL) imagingexperiments on finished cells [4.6.1.3] and alsofor applications in which a solar cell is heldat specific operating points under externalillumination, such as the series resistanceimaging methods that were introduced by ourgroup [4.6.1.4-4.6.1.6].High resolution (typical one megapixel)luminescence images can be capturedwithin seconds or even fractions of a second,allowing very fast and detailed studies ofspecific material and device properties.Over the last five years PL imaging has seenthe development of a range of specificapplications and also rapid adoption by thePV community. Specific applications thatare applicable across the PV value chainare now available and the number of theseapplications is growing rapidly. Examplesinclude series resistance imaging anddiffusion length imaging on fully processedcells. The research of the PL group at UNSWcontinues to contribute significantly to thesedevelopments. This year the research by ourgroup was focussed on the application of thephotoluminescence intensity ratio methodfor measurements of the bulk lifetime onsilicon bricks (Section 4.6.1.2), proof of conceptstudies on an application of luminescenceimaging for measuring the emitter sheetresistance on diffused wafers (4.6.1.4) and onanalysing local shunt or defect currents fromPL images on metallised and non-metallisedsamples in a quantitative fashion (4.6.1.5).The commercialisation of PL imaging byUNSW spin-off company BT Imaging PtyLtd has also made significant progress, witha growing number of leading wafer andcell manufacturers and research institutesworldwide now using the BTi-R1 R&D toolthat was introduced into the market in 2008.In addition, as-cut wafer inspection with PLimaging (see Section 4.6.1.3) is increasinglyrecognised as an ideal tool for incomingwafer inspection in solar cell productionor for outgoing quality control in wafermanufacturing [4.6.1.2, 4.6.1.7-4.6.1.8].New in-line PL imaging tools that are fastenough to keep up with typical productionline throughput at full one megapixel image100

Schematic diagram of aluminescence imaging system. Inphotoluminescence (PL) imaging anintense laser is used to illuminatethe entire sample and a sensitiveIR camera then captures theluminescent emission from thesample. In PL imaging the samplecan be a brick prior to wafering,an as-cut or partially processedwafer or a fully processed solarcell (as shown). A power supplyis used in electroluminescenceimaging on finished solar cellsand in combination with thelight source in series resistanceimaging techniques.ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTFigure 4.6.1.1resolution were introduced by BT Imaging in 2010and are currently entering the market.4.6.1.2 PL imaging on Si bricksPhotoluminescence imaging is an ideal tool forfast characterisation of silicon bricks, giving instantinformation about the position of low lifetimeregions that are commonly observed near thebottom and top of a brick and also exposing areasof high structural defect density. This informationcan be used as valuable feedback during theproduction of silicon wafers. Initial work onluminescence imaging on bricks showed excellentqualitative correlation between the as measured PLintensity and for example μ-PCD effective minoritycarrier lifetime maps [4.6.1.9]. More recently wealso demonstrated a quantitative interpretationof the PL signal in terms of the bulk minoritycarrier lifetime m b.The bulk lifetime is a more usefulparameter than the effective lifetime, especially inunpassivated samples, where the measurementresult is strongly influenced by surfacerecombination. PL imaging thus has advantagescompared to other measurement techniques suchas μ-PCD, which have commonly been used to datefor the characterisation of bricks in production, andwhich report only effective lifetime values.In our previous approach to measure bulk lifetimefrom PL images we converted the measured PLintensity form a single PL image into bulk lifetimeusing a pre-determined tool specific transferfunction. While that approach proved to be areliable approach for obtaining bulk lifetime ona number of samples and without the need foradditional calibration, one disadvantage is thefact that it requires a separate doping densitymeasurement. The PL intensity in low injectionconditions is given as PL~Dn∙N D, i.e. it is proportionalto both the effective lifetime and the backgrounddoping density N D. Normalising the measurementfor doping variations is required in order to obtaininformation on Dn and thereby the lifetime.Recent work in the Photoluminescence Groupfocussed on a modified approach for bulk lifetimemeasurements on silicon bricks, which avoidsthe need for a separate resistivity measurement.The method is based on analysing thephotoluminescence intensity ratio (PLIR) betweentwo PL images taken with different spectral filtersmounted in front of the camera [4.6.1.10-4.6.1.11].It is equivalent to the diffusion length imagingmethod proposed and first demonstrated atUNSW by Peter Würfel on finished solar cells usingelectroluminescence imaging [4.6.1.12]. In fact, theelimination of voltage variations from a diffusionlength image that is achieved in the PLIR methodon cells is in perfect analogy to the eliminationof doping density variations when applying thismethod to PL images on bricks. Silicon bricksrepresent an ideal test case for the PLIR method,since uncertainties associated with variations insurface texture and with the optical and electronicproperties of the rear surface are almost completelyavoided. In cells and wafers these uncertainties canlead to large experimental errors.Figure 4.6.1.2 shows a comparison of twobulk lifetime images obtained from a dopingnormalised PL image and from the PL intensityratio, respectively. The latter image was obtainedusing a transfer function that was calculatedtaking into account sample temperatureand the exact spectral properties of themeasurement system.Deviations are observed near the lowlifetime edges at the bottom and top,where the PLIR overestimates the bulklifetime (see also Fig.4.6.1.3, showing acomparison of cross sections throughthe two bulk lifetime images). Defectluminescence contributions, which becomemore noticeable in the low lifetime regionswith reduced band to band luminescencesignal is currently believed to cause thosedeviations. Spectral PL measurementscurrently under way will shed more light on thisissue. Several approaches to mitigate the impactof the above measurement errors in the PLIRnear the bottom and top regions are currentlyunder investigation.Bulk lifetime images obtained froma single doping density normalisedPL image (upper) and from the PLIR(lower), see text for details.Figure 4.6.1.2Cross sections through the bulklifetime images from Fig. 4.6.1.2,showing good agreement in thecentre and artefacts near thebottom and top regions.Figure 4.6.1.3101

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTPL images of two multicrystallinesilicon wafers. (a) Centre wafer withlow dislocation density, (b) Centrewafer with high dislocation density,(c) bottom wafer with low lifetimeacross the entire wafer area, (d)corner wafer with low dislocationdensity [4.6.1.2].Figure 4.6.1.4Correlation of cell performancedata (normalised) with thedefect density obtained fromthe PL images on the as-cutwafers [4.6.1.1].Figure 4.6.1.5Very good quantitative agreementis observed between the twomeasurements in the central part,i.e. across the majority of the brick.This region is the section of the brickthat is supposed to be used for wafermanufacturing (the very low lifetimetop and bottom regions are supposedto be removed from the ingot).4.6.1.3 Raw wafer inspectionWafer quality in industrialmanufacturing of silicon solar cellsvaries substantially and can have astrong impact on the efficiency ofsolar cells, particularly in the caseof multicrystalline silicon. As-cutwafer inspection of incoming qualitycontrol (IQC) in solar cell productionis therefore indispensable in orderto maintain high efficiencies at highyield in production. To date IQC inproduction typically includes opticalinspection for cracks and large chips, thicknessand total thickness variation, surface morphology,resistivity measurements and in some casesμ-PCD lifetime scanning. A problem with thesetechniques is that the electrical cell performancedoes not correlate sufficiently strongly with thedata resulting from these measurements. One mainreason is surface recombination dominating totalrecombination (and thus effective lifetime) in as-cutwafers. Only wafers with very low bulk lifetime(

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTBT-Imaging W1 PL imaging system,with integrated automated imageprocessing algorithms and a currentthroughput of up to 2400 wafersper hour with full one megapixelimage resolution.Figure 4.6.1.6response of as-cut wafers makes inline applicationswith a throughput of up to 3600 wafers per hour,a significant experimental challenge. BT Imagingrecently presented the W1 (see Fig. 4.6.1.6), thefirst inline PL imaging system capable of takingPL images on as-cut wafers with a throughput of2400 wafers per hour and using automated imageprocessing to identify a range of relevant specificdefect metrics. Using these tools, application inproduction of the above concepts, which so far hasonly been proven off-line on relatively small samplesets, will demonstrate the value of using PL imagingfor IQC.4.6.1.4 Emitter sheet resistance imagingIn solar cell production, the value of the emitterresistance typically needs to be maintained withintight specifications in order to avoid subsequentproblems with contact formation such as junctionshunting or high contact resistance. Previous workin the PL group focussed on luminescence basedemitter sheet resistance imaging measurementson fully processed metallised solar cells. In a newapproach, the aim is to measure the emitter sheetresistance in a contactless fashion from PL imageson non-metallised samples, which would allowapplication of that method immediately after theemitter diffusion in production. Since the emittersheet resistance is not expected to exhibit lateralvariations over short distances, a method withcomparatively low spatial resolution is sufficient.Here we investigated the principle of using a PLimage for a coarse measurement of the emittersheet resistance, providing a single average datapoint for a large area.The measurement principle is to take a PL imagewith non-uniform illumination and obtain theemitter sheet resistance information from theresulting PL intensity variation, in combinationwith theoretical modelling. A 1-inch circular longpass filter is placed in close proximity to the sampleduring the PL imaging measurement. The filterblocks effectively all (>99.9 %) incident light formthe area underneath the filter and transmits >98%of the luminescence that is emitted from that area.The luminescence from the non-illuminated areais generated by lateral flow of excess carriers fromthe illuminated into the non-illuminated parts ofthe wafer, in other words, electroluminescenceis generated in a contactless fashion from thatregion, as reported previously elsewhere [4.6.1.14].The illumination intensity and the emitter sheetresistance have a strong impact on the carrierredistribution via lateral carrier flows and therebyon the steady state excess carrier concentration,the latter reflected in the luminescence intensity.Combined with theoretical modelling this allowsinterpretation of the luminescence intensitydistribution in terms of the emitter sheet resistance.An example of a PL image taken on a diffusedwafer with 60 Ω/sq is shown in Figure 4.6.1.7. Thecircular filter is visible in the centre of the wafer.The luminescence signal from underneath thefilter represents contactless electroluminescenceas described above. Figure 4.6.1.8 shows threePL intensity profiles from the centre of the filterto the outside region taken on three differentsamples with emitter sheet resistivities rangingbetween 30 Ω/sq and 100 Ω/sq. With increasingemitter resistance the PL intensity decays with asteeper slope towards the centre ofthe non-illuminated region, which isa result of the limited ability of theemitter to transport current laterally.An analytical model for that relativePL profile was developed and usedto fit the measured data within thenon-illuminated part of the image.Resulting emitter sheet resistancevalues were found to agree with eddycurrent measurements of the emittersheet resistance.The analytical model so far describesonly the PL profile in the nonilluminatedfilter region. Further workwill aim to extend that model to the illuminatedpart of the wafer, resulting in increased accuracy.The impact of the wafer background doping densityon the measurement results is another topic forfurther work.PL image of a diffused siliconwafer taken with a 1-inch circularlong pass filter located in closeproximity to the wafer surface. Theluminescence signal emitted fromunderneath the filter is causedby lateral current injection fromthe illuminated wafer area viathe emitter.Figure 4.6.1.7Cross section of the normalisedPL intensity from the centre of thefilter (x=0) to the outside regionthrough PL images as shown in Fig4.6.1.7 for different samples withvariable emitter sheet resistance.Figure 4.6.1.8103

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORT0.5 cm(a)(b)PL images taken on a test structure. An externalresistor is connected in parallel to the structure using acontact probe (visible in the bottom right). The imagesshow the impact of a large resistance (mild shunt,left) and a small resistance (strong shunt, right) on thePL image.Figure 4.6.1.9Extracted current as calculated from PL imagesusing a new analysis methodology as a functionof the measured current extracted over a resistor,demonstrating that PL imaging allows a quantitativeanalysis of the extracted current from localised shuntsor defects.Figure 4.6.1.104.6.1.5 Luminescence based shuntimagingQualitative shunt detection from luminescenceimages has been demonstrated in the past by theUNSW PL group (amongst others) [4.6.1.15-4.6.1.17].The vicinity of shunted areas typically appears as ablurred region of reduced luminescence intensityin both PL and EL images. This blurring is caused byvoltage drops associated with lateral current flowthrough the emitter and the front surface grid.A luminescence imaging based method fordetermining quantitative shunt values in siliconsolar cells from open circuit photoluminescenceimages has been developed and demonstratedin collaboration with the Max Planck Institute forMicrostructure Physics, Halle, Germany. The methodis based on interpretation of the luminescenceintensity around a local shunt in terms of theextracted current density [4.6.1.18]. Under theassumption of a unity ideality factor the localreduction in PL count rate in the vicinity of a shuntor local defect is proportional to the current beingextracted. The theoretical framework for quantifyingthe total current extracted by a local defect byintegrating the reduction of the PL signal in thesurrounding of the shunt or defect and by applyingsuitable calibration procedures was developed.Experimental verification of the method wasachieved using a specifically prepared test structure,i.e. a solar cell structure with single point contacton the front surface. That structure allowedgenerating local shunts with variable and welldefined external shunt resistance. Figure 4.6.9shows two PL images of that structure with a high(left) and low (right) Ohmic resistor connected inparallel with that structure. Figure 4.6.1.10 showsthe extracted current, as obtained from the abovemethodology as a function of the current flowingacross the resistor, as measured with a multimeter.Excellent agreement is observed. Combining theabove current measurement technique with aninterpretation of the local luminescence signal interms of the diode voltage near a shunt site allowsquantification of individual local shunts.Applications of this method to finished cellsare affected by current extraction via the metalgrid, which causes experimental errors. Theimpact of this effect was assessed experimentallyusing specific test structures, where the relativeposition with respect to the nearest grid finger ofartificially introduced shunts could be varied. Theseexperiments indicate that the above experimentalerrors in fully processed cells can at least partiallybe corrected using empirically determined lookup tables.When applied to partially processed cells prior tometallisation the method can be used to quantifylocal shunts, to determine the total recombinationcurrent in local and distributed defects and alsoto quantify recombination channels that areintroduced by specific processing steps such aslaser processing.4.6.1.6 SummaryAn exceptional variety of material and solar cellparameters can be measured on silicon bricks,silicon wafers and silicon solar cells with highlateral spatial resolution and short measurementtime using luminescence imaging techniques.The range of applications continues to grow, withnew applications for PL imaging being developedat UNSW, and increasingly also in other researchinstitutes and by R&D groups in PV companies.PL imaging, introduced at UNSW only five yearsago, has now been broadly adopted as a standardcharacterisation method, with PL tools now inuse at virtually all leading research institutesworldwide and also by leading wafer and solar cellmanufacturers. A range of opportunities for in-linequality control, process monitoring and processcontrol are enabled by the high resolution andspeed of PL imaging. Demonstration by UNSWspin-off company BT Imaging of those applicationsin beta trials at several company sites is currentlyin progress.104

References4.6.1.1 McMillan, W. et al., In-line monitoring of electricalwafer quality using photoluminescence imaging.Proceedings of 25th EPVSC, Valencia, Spain,September, 2010.4.6.1.2 Trupke, T., Nyhus, J. and Haunschild, J.,Luminescence imaging for inline characterisation insilicon photovoltaics. Phys.Stat.Solidi RRL, 2011: p.published online February 2011.4.6.1.3 Fuyuki, T. et al., One shot mapping of minoritycarrier diffusion lenth in polycrystalline Si solar cellsusing electroluminescence. 31st IEEE PhotovoltaicSpecialists Conference, Orlando, USA, 2005.4.6.1.4 Trupke, T. et al., Fast photoluminescence imaging ofsilicon wafers. WCPEC-4, Waikoloa, USA, 2006.4.6.1.5 Trupke, T., et al., Spatially resolved series resistanceof silicon solar cells obtained from luminescenceimaging. Appl. Phys. Lett., 2007. 90: p. 093506.4.6.1.6 Kampwerth, H. et al., Advanced luminescence basedeffective series resistance imaging of silicon solarcells. Appl. Phys. Lett., 2008. 93: p. 202102.4.6.1.7 Haunschild, J. et al., Quality control of ascutmulticrystalline silicon wafers usingphotoluminescene imaging for solar cell production.Solar Energy Materials and Solar Cells, 2010.94(12): p. 2007-2012.4.6.1.8 You, D. et al., The dislocation distributioncharacterisitics of multi-crystalline silicon ingotand its impact on cell efficiency. 35th IEEE PVSC,Honululu, USA, June, 2010.4.6.1.9 Trupke, T., Bardos, R.A. and Nyhus, J.,Photoluminescence characterisation of silicon wafersand solar cells. 18th workshop on Crystalline SiliconSolar cells & Modules, 2008.4.6.1.10 Trupke, T. et al., Bulk minority carrier lifeitmefrom luminescence intensity ratios measuredon silicon bricks. 25th EPVSC, Valencia, Spain,September 2010.4.6.1.11 Mitchell, B. et al., Bulk minority carrier lifetimes anddoping of silicon bricks from photoluminescenceintensity ratios. J. Appl.Phys., 2011: p. acceptedfor publication.4.6.1.12 Würfel, P. et al., Diffusion lengths of silicon solarcells obtained from luminescence images. Journal ofApplied Physics, 2007. 101: p. 123110.4.6.1.13 Demant, M. et al., Analysis of luminescenceimages applying pattern recoginition techniques.Proceedings of 25th EPVSC, Valencia, Spain,September, 2010.4.6.1.14 Sinton, R.A., Contactless electroluminescence forshunt-value measurement in solar cells. Proceedingsof the 23rd EPVSC, Valencia, September, 2008:p. 1157.4.6.1.15 Trupke, T. et al., Luminescence imaging for fastshunt localisation in silicon solar cells and siliconwafers. International Workshop on Science andTechnology of Crystalline Silicon Solar Cells,Sendai, Japan, 2-3 October, 2006.4.6.1.16 Kasemann, M. et al., Shunt detection capabilities ofluminescence imaging on silicon solar cells. 22ndEuropean Photovoltaic Solar Energy Conference,Milan, Italy, 2007.4.6.1.17 Breitenstein, O. et al., On the detection of shunts insilicon solar cells by photo- and electroluminescenceimaging progress in Photovoltaics, 2008. 16: p. 325.4.6.1.18 Augarten, Y. et al., Luminescence shunt imaging:qualitative and quantitative shunt images usingphotoluminesccence imaging. 24th EuropeanPhotovoltaic Solar Energy Conference, Hamburg,September, 2009.4.6.2 General CharacterisationUniversity StaffDr Henner KampwerthProject Scientists and TechniciansAlan YeeThe range of research undertaken by the Centreinvolves a variety of measurement techniquesfor the characterisation of photovoltaic materialsand devices. Understanding and utilising thesetechniques appropriately, and improving themwhere possible, is crucial in order to maintain theworld-class reputation of our research output.Improvements to measurement techniques with thegoal of producing more accurate and meaningfulresults will have a direct impact both on researchactivities and on the optimization of proceduresused in manufacturing. The time and numberof experiments needed to understand a certainphenomena can be greatly reduced. The need tomeasure parameters of interest more precisely,without the possibility of misinterpretation, istherefore high. The General Characterisation groupwas formed in July 2010 to address this need.The group primarily focuses on opticalcharacterization methods such asphotoluminescence spectroscopy and absorptionspectroscopy. In recent years, it has been shownthat these optical measurements can be used toextract a large number of material and deviceproperties. In addition the ability to performcontactless non-destructive probing of small areasdown to a few ten micrometers in diameter makesoptical measurement techniques very attractive forsamples that cannot be measured otherwise.In addition, the group will review somecommonly used measurement methods withthe aim of automating parameter extractionfrom measurement data. It is common that lessexperienced researchers misinterpret data and drawincorrect conclusions. Computer programs thatassist in various levels of interpretation thereforecan prove their value. Some computer programswritten in the course of this work will be madeavailable for public use.ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORT105

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORT5. EDUCATIONSummaryThe ARC Photovoltaics Centre of Excellence incorporates the activitiesof the former Key Centre for Photovoltaic Engineering. The formerKey Centre started in 1999, after the award of special funding fromthe Australian Government to promote teaching and research in thearea of photovoltaics. A major initiative of this Key Centre was theestablishment of the world’s first undergraduate degree in Photovoltaicsand Solar Energy. The Centre was one of only eight such Key Centresawarded Australia-wide across all disciplines, demonstrating theGovernment’s understanding of the importance of the field ofrenewable energy. In 2003 the Key Centre and its activities wereincorporated into this Centre of Excellence, awarded to the same teamat the University of New South Wales. On 1 January 2006 UNSW officiallyformed a new School within the Faculty of Engineering, the School ofPhotovoltaic and Renewable Energy Engineering, which includes theARC Photovoltaics Centre of Excellence.The School offers undergraduate, postgraduate and research programsencompassing a range of aspects relating to the photovoltaic and otherrenewable energy industries. These programs have been developed inconsultation with representatives from industry to ensure graduatesare appropriately qualified to enter the field upon completion of theirstudies. The undergraduate programs have experienced continuingstrong demand growth since 2007. Since 2003, the School has seen atotal of 160 students graduate from its undergraduate programs, and126 students from its postgraduate degrees. These graduates are nowtaking advantage of a range of opportunities being created in thebooming photovoltaics and renewable energy industries.The higher media profile of global warming, greenhouse and energyissues in the general Australian community is helping to raise theSchool’s profile locally and the industry growth in Asia raised theSchool’s profile internationallyThe granting of a large number of sponsorships was finalised in August2007 for up to $5.2m support from the Australian Government forstudy at the School, as part on the Asia-Pacific Partnership on CleanDevelopment and Climate. In 2010, these scholarships attracted 21 newundergraduate and 48 (29 in Semester 1 and 19 in Semester 2) newpostgraduate coursework students to the School. The number of formalagreements with Chinese universities established as a result of thePartnership grew to seven during 2010.At the end of 2010, there were 437 undergraduate students, 88postgraduate coursework students and 63 research students (11Masters by research and 52 PhD) enrolled in the School’s programs.106

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORT700600Total No. of students50040030020010002002 2003 2004 2005 2006 2007 2008 2009 2010 2011Year5.1 UndergraduateEducational ProgramsOverall enrolment numbers insecond semester, by year.The School offers two undergraduate coursework engineeringprograms, each of which can be taken in combination with a Bachelorof Arts (5 years), Bachelor of Science (5 years), Bachelor of Commerce(5.5 years) or a Bachelor of Laws (6 years). These degrees aim to developwell-educated graduates with the basic skills, attributes and knowledgerequired to practise as professional engineers in the boomingrenewable energy industries.The Photovoltaics and Solar Energy Engineering program draws onthe long history of expertise of Centre staff in this field and was thefirst of its kind internationally. Two new programs were established in2009 in China, at Nanchang University in Jianxi Province and at NorthChina Electric Power University in Beijing. The UNSW program includeseducation in technology development, manufacturing, quality control,reliability, life cycle assessment, system design, maintenance and faultdiagnosis, marketing, policy development, energy efficiency and otherrenewable energy technologies. It has two unique features: a groupproject in the second year of the program, and a ‘strand’, which is aminor area of specialisation. Since 2003, a total of 128 students havegraduated from the program, 24 graduated in 2010 and another 34 areexpected to graduate in March 2011 after completing their studies in2010. At the end of 2010, a total of 310 students were enrolled.The second, broader degree program, in Renewable Energy Engineering,was introduced in 2003. In addition to photovoltaic devices, studentsin this program study solar architectural technologies, wind energy,biomass, solar thermal and renewable energy policy as part of their corecurriculum. At the end of 2009, a total of 127 students were enrolled.2010 saw 15 new graduates of this program, bringing the total to 32,and 5 more are expected to graduate in March 2011. At the end of 2010,24% of enrolled undergraduate students were women.Graduates of these programs are gaining employment within a varietyof organisations, including in the areas of energy efficiency andsustainable design. Each year a number of graduates start on a researchprogram at the Centre, elsewhere at UNSW or at other institutionswhile others are employed in industry by solar cell and equipmentmanufacturers, system design and integration companies, electricityutilities, energy efficiency and environmental design companies andmajor end users of products.107

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTMeasurement of lightdistribution in a horizontalplane. A lantern is shown inthe inset.Top side of Glowstar PCB showingseveral features such as fusesand microprocessor.(Upper) Student-built current-tovoltageconverter. (Lower) Studentbuiltphototransistor circuit.5.1.1 The StrandThe strand is unique within UNSW to thePhotovoltaics and Solar Energy Engineeringprogram, and is essentially a minor area ofspecialisation which complements a student’sstudy of photovoltaics. Students can take strandscovering a wide variety of areas includingcomputing, electronics, mathematics, physics,mechanical engineering, civil engineering, andarchitecture. The aim of the strand is to providestudents with broader engineering backgroundsimportant for the cross-disciplinary nature ofphotovoltaic applications.In 2011 strands will be offered to RenewableEnergy Engineering students as well. These will beless formally structured than in the PhotovoltaicEngineering program and take the form of setsof electives associated with various renewableenergy technologies.5.1.2 Undergraduate Second-YearStudent Group ProjectsIn the second year of the Photovoltaics and SolarEnergy program, students undertake a year-longgroup project. The main emphasis of the secondyear group project course is hands-on projectengineering. The course has a lecture componentcovering project engineering, report writing andpresentation skills and each project has a researchcomponent, a planning and design component, ahands-on component and a presentation/reportingcomponent. This course helps to prepare studentsfor their fourth year thesis, which is undertaken byall students enrolled in both the Photovoltaics andSolar Energy and Renewable Energy Engineeringundergraduate programs.In 2010, second-year project groups worked on thefollowing projects: silicon and cuprous oxide thinfilm solar cells, photoluminescence analysis of solarcells, solar thermal water heating, Sunswift solar racecar, solar water pumping, solar lanterns, cookers anda “solar suitcase” for developing countries and iceair conditioning.5.1.2.1 Developing CountriesThe School has been involved with projects in thedeveloping countries for the last eight years. Theapplication of photovoltaics and other renewableenergy technologies can make a great difference topeople’s lives and living standards in these places.Students involved with this project in the past haveinstalled and maintained photovoltaic lightingsystems in rural locations in Nicaragua, Vanuatuand Nepal and photovoltaics powered waterpurification in Sri Lanka. Skills that the students haddeveloped throughout the year were put to gooduse in the field as invaluable lessons were gainedregarding project management of activities indeveloping countries. Students faced and overcamemany technical and non-technical issues in projectimplementation. With the rapidly growing studentpopulation the School is no longer involved indirectly supervising international journeys forgroups of second year students but, instead,suggests interested students involve themselves inthe UNSW Chapter of Engineers Without Borders.Solar Lanterns ProjectEnergy is fundamental to alleviate poverty inhealth, education, communication, economicdevelopment and social equity. Electricity impactsthe lifestyles and living conditions, primarilythrough electric lighting. However, approximately1.5 billion people in developing countries haveno access to electricity. The majority of thesepeople use kerosene lamps for lighting. Kerosenelamps are expensive to run, have health impacts,are potentially dangerous and are a source ofgreenhouse gases.Solar lanterns have potential to improve the livingconditions of people in developing countries. Itcan provide much needed study hours for schoolchildren as well as support after-hour incomegenerating activities. One study has found that solarlanterns have significantly benefited school-agedchildren in India. The extended study hours madepossible with solar lanterns shown to have a verypositive impact on their performance.108

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTThermal image of experiments determining theefficiency and heat loss to the environment of variouscooking devices.Image of one groups insulated potcooking food.There are many solar lanterns are available in themarket. The performance of these lanterns varieswidely making it difficult for consumers to selecta product.In this project students have studiesa range of solar lanterns commonly available indeveloping contries. Using a set of tools theycompared performance of solar panel, ballastelectronics, charging-discharging circuits, lightoutput and efficiency.Solar Cookers ProjectStand alone PV systems (PV/diesel/battery) supplypower to many households in remote parts ofthe world. Remote households in Australia and inparticular Aboriginal communities utilise such PVsystems. Cooking is an energy load that needs asubstantial amount of energy. In many instancesdesigners for stand-alone PV systems advocate useof gas or microwave cookers for cooking. However,often gas bottles are a very expensive way to getenergy for cooking.Is it possible that as the price of PV falls that someform of electric cooking could be the best option? Ismicrowave cooking the most efficient technology?To answer this question we asked students: Whatis the most efficient way to cook? This projectexplored low energy approaches to cooking and PVsystem design to supply the energy to power suchsystems. One approach investigated “slow cookers”.Typically they are cooking systems that utilise lowenergy but for long periods of time. Typically theyare uninsulated, so at present not very efficient!Students investigated various approaches toinsulate their cooking devices and used PV panelsto charge a battery and then use the energy to heata pot of food. Energy flows (current and voltage)were monitored, as well as the temperature of thefood. Students investigated, designed, built andmonitored whether small PV systems are capable ofsupplying such cooking systems.5.1.2.2 Silicon Thin-Film SolarProcessing Equipment ProjectThis project aims to give students a broadexperience in enhancing and maintaininga world-class thin-film solar cell researchlaboratory. This is achieved through a twostageprocess. Firstly, students develop afundamental understanding of equipmentand systems within the Centre’s thin-filmlaboratory, including vacuum systems, gasdelivery systems, leak testing equipment,and plasma processing machines. Thesecond stage involves the studentsintegrating what they have learnt, by applying it toa specific project in the thin-film laboratory.In 2010, six second-year students undertook thisproject. A group of three students worked onupgrading a vacuum test chamber to redesign thevacuum control interface and build it. A secondgroup of three students worked on halogen lampvacuum annealing stage, with the aim of beingable to anneal samples at up to 1000°C. Thisgroup successfully designed and built a halogenlamp assembly that was both functional and easyto maintain.Both groups benefited greatly from the courseas they learnt to integrate project managementand engineering problem solving to successfullycomplete their projects.5.1.2.3 Cuprous Oxide Thin-Film SolarCells ProjectStudents this year revived a decades-oldphotovoltaics technology that was once theglobally leading technology, before silicon came tothe fore in the 1950s. The manufacturing process forcuprous oxide cells is relatively simple and studentsmade and tested their own cells from scratch.5.1.2.4 Water Pumping ProjectA group of students was charged with the taskof demonstrating the pumping of water withphotovoltaic power from ground level to theCake successfully cooked!109

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTSeventy SPREE students illustrated their final year thesis work on 22 October2010 in the largest yet of the school’s graduating classes.roof of the UNSW Electrical Engineering building.The students worked on realistic conceptualapplications such as toilet flushing in the buildingwith excess stormwater, water supply for a ruralvillage or stock watering on a cattle station. Fortheir demonstration project, they were requiredto provide nominal flow rate and storage todemonstrate the concept. The skills learnt aredirectly applicable to solar water pumping fora wide range of applications, including in ruralvillages around the world.5.1.2.5 Solar Thermal Water HeatingProjectIn our daily life, we are using lots of hot water fordifferent purposes, needing a considerable amountof electrical or other energy sources. In this projectthe students learned how to use solar thermalenergy to heat cold water. In this project studentstudied the basic principle of solar heating systemand designed a new water heating system. Theyhave also learned what materials should be usedin the systems and investigated the possibilitiesto improve the design for better conversionefficiency. They built their designed systems usingtheir selected materials, as shown in the followingfigures. They tested their systems and collected andanalysed data to find conversion efficiency, costevaluation, etc.5.1.2.6 UNSW Sunswift ProjectOver the past year, second year SPREE studentshave helped to design and build several majorcomponents of the UNSW solar car, which placedthe first in the silicon class in the 2009 Global GreenChallenge (GGC), the premier international solar carrace. The result was the best result achieved by theteam in its 14 year history.Students were involved in two solar car relatedprojects – design and construction of thephotovoltaic array that powers the car, andconstruction of the lightweight compositecomponents that reduce the weight, and hence theenergy usage of the car.ArrayThe solar array is essentially an off-grid solarsystem, with the added complications of veryfragile modules, and the requirement that it mustwithstand the conditions of driving 3000km ofdusty, bumpy road.The students designed the module and array sizingto conform to the race regulations, and selectedthe appropriate cabling and bypass diodes. Theyalso performed array calculations and simulationsin order to calculate the optimum wiring schemethat allowed for maximum output power, themost efficient battery charging, and maximumrobustness against shading.Two types of cells were used for maximumefficiency, standard industrial cells, and UNSWproducedTopcell cells. The team characterisedthe Topcells, initially by open-circuit voltagemeasurements, and later by full IV sweeps. The cellswere then encapsulated, with assistance from CSGSolar, and assembled into modules.Finally, the array team and other Sunswift membersassembled the array, mounting and soldering thefragile modules onto the solar car.CompositesCarbon fibre composites are extremely strong,lightweight materials, which are currently used,e.g. in aeroplanes and wind farms, to increase theenergy conversion efficiency of the systems.The composites team had the substantial job ofbuilding the carbon fibre composite componentsfor a car which was almost entirely carbon fibre.These included the top shell, which housed thesolar array, the bottom shell, which formed thestructural chassis to safely house the driver, andthe structural components which held the braking,steering and suspension systems. The team alsosuccessfully constructed the first new carbon fibrewheels for the car in 8 years. The wheels are themost complicated composite part to construct,require a great deal more care and attention, andmust be thoroughly tested. The new wheels were110

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTUpper: Copper pipes mounted on a copper sheet for water heating.Lower: A complete system with glass covered on top.Student-built thermal water heating system for domestic use in their 2nd year project.manufactured extremely well, and performedbeautifully over the 3000km of rough terrain.The construction of all of the components took oversix months, and was done at the Boeing Aerospacefacilities in Bankstown. Boeing sponsored theteam the carbon fibre and other materials, andthe students learned how to work with compositematerials from the experts.The carbon work on the car by these studentswas easily some of the highest quality work in afield of 40 international solar cars, including semiprofessional teams, and also suffered no issuesduring the race.5.1.3 Fourth Year ThesisThe thesis project is usually completed in the lasttwo sessions of an engineering undergraduatestudent’s studies. Students may start their project ineither semester. They undertake directed laboratoryand/or other research work on an approvedsubject under guidance of the School’s academicstaff, sometimes with industry participation orco-supervision. Typically, the thesis involves thedesign and construction of experimental apparatustogether with practical tests. Each student isrequired to present a seminar in their first semesterand submit a written report and present an OpenDay poster in their second semester.Some students finished and presented theirprojects at the end of Semester 1, 2010. ThePhotovoltaic and Solar Energy Thesis Poster Prizewas won by Lei Zhang (Laser Doping TechniqueInvestigation and Optimisation for High EfficiencyLaser Doped Solar Cells Fabrication). The RenewableEnergy Engineering Thesis Poster Prize was won byNicholas Boerema (Renewable Energy Integrationinto the National Electricity Market). Other projectscoming to completion at that time were concernedwith wind energy and its grid integration, annealingeffects on silicon thin film solar cells, determinationof carrier diffusion length and saturation currentdensity in silicon solar cells, laser doped emittersand back surface fields, metalization effects on theoptical properties of nanostructure silicon, highefficiency dye sensitised solar cells, optimization ofhigh temperature ITO, time of flight measurementsetup, sustainable transport, building integratedPV system design and testing, evaluation of pv/diesel mini grid applications in the NT, solarpower for a weeding robot, silicon quantum dots,design of a domestic solar water heating system,performance assessment and evaluation of theUNSW Quad building’s rooftop PV System, feasibilityof electric cars in Sydney, management strategiesfor building sustainability, and deploying energyefficiency policy to offset cost of meeting MRET inthe residential sector, energy efficiency in buildings& Greenstar.The Poster Prize selection in November 2010,for students completing at end of Semester 2,resulted in a wins for Hua Fan for his poster about“Solar Electrolysis System for Remote Productionof Disinfectant Solutions” (Photovoltaic andSolar Energy) and Kah Howe Chan for his posteron “Wind and other renewable energies at theUNSW Wellington property” (Renewable EnergyEngineering). The other topics on which theseswere completed included: study of carrier transportmechanism in organic semiconductor for differentelectrode metals for organic solar cells, laserdoping of thin film silicon solar cells, wind energyforecasting, impact of electric vehicles on LVelectrical network, optical fibre/nanowire hybridstructures for dye-sensitized solar cells, design ofmetamorphic tandem solar cells, the 46% efficientcell project: cell modelling & testing, comparisonof properties of Si QD film treated by rapid thermalannealing and converntional furnace annealing,feasibility study of biogas/electric hybrid cars,benefits of NSW wind power for the NEM, electricalcharacterisation of Ge nanocrystals embeddedin amorphous silica matrix, 46% efficient splitspectrum cells, biogas production on a Wisconsinfarm, Investigation of n/p-type doping of GeNcs in SiO 2towards tandem photovoltaic cells,optimisation of thermal annealing parameters onplanarisation and electrical properties of thin-filmpoly-crystalline silicon solar cells on glass, homecharging for electric vehicles, reduction in office111

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTtenant energy, doping concentration effects onmetal ohmic contact formation to silicon wafer,energy efficiency on UNSW campus, opticalcharacterisation of silicon quantum dots in SiOx/Si 3N 4for optical detection of defects, design andconstruction of PV solar electrolysis system formedical aid kit, zero energy buildings, fabricationand characterisation of Si DQs nanostructuresgrown by PECVD, adapting direct etching methodfor texturing of silicon, design of light trappingstructure for organic solar cells, low energy waterpump systems, PV, pumping, micro-hydro or windat UNSW’s Wellington property, passive PV coolingusing stack effect, investigation of phosphorousdoping of Si-QD materials by furnace diffusion,sustainable energy in Willoughby city councilbuildings, design of light trapping structure fororganic solar cell, boron and phosphorus dopingof silicon rich oxide layers by diffusion of dopants,solar water pump testing, design of high efficiencyorganic solar cells, investigation of emitter sheetresistance using photolumescence imaging,aerosol jet etching for silicon solar cells, zero energybuildings, study of temperature and doping effectson silicon wet-etching, low contact area laserdoping, inverters on unstable grids, on-demandstreet light, electric motorcycle, investigations intoover-floating of laser-doped cells, energy selectivecontact formed by thermal oxidation, growth of SiQD at low temperature, determination of carrierdiffusion length and saturation current density insilicon solar cells (I and II generation solar cells),zero energy building for China, optics of theheliostat field and how this impacts uniformityat the receiver, novel approach to plating for lowcost solar cells, size controlled growth of Ge QDin SiO 2, solar access shading survery of UNSWKensington campus, optimisation of rear localisedsurface plasmon enhanced light trapping inpolycrystalline silicon thin-film solar cells, hybridbulk heterojunction solar cells, carrier transportmechanism in organic semiconductor for differentelectrode metals for organic solar cells, aerosol jetetching for silicon solar cells, PV partial shading andminimisation of its impact, novel techniques forlow cost high efficiency solar cells, temperature anddoping effects on silicon wet etching, innovativerear point contact scheme for silicon solar cells,PV and RE at Wellington campus, efficiencyprediction on advanced photovoltaic converterconcepts, modelling of recombination activities inpolycrystalline silicon thin film solar cells, propertiesof Sb-doped polycrystalline germanium thin film bya magnetron sputtering method, and boron dopingof Si-QD materials by diffusion.The $500 Photovoltaics Thesis Prize for 2010 waswon by Xiang Zhao and Eleanor Wood was awardedthe Renewable Energy Thesis Prize. These prizesare awarded for the highest marks in the final yearthesis subjects.5.1.4 Dean’s AwardsEach year the Dean of Engineering presents awardsfor outstanding students in each non-final year ofstudy in each Engineering School. Prizes awardedduring 2010 for academic performance in theprevious year went to Dongchen Lan (stage 1),Benjamin Hughes (stage 1), John Durrant (stage2), Vincent Allen (stage 2), Xue Bai (stage 3), DongLin (stage 3), Yu Feng (stage 3) and Shen-Long Ooi(stage 4).5.1.5 N&M Radiant Design PrizeThe N&M prize is awarded to one student groupundertaking the School’s project in the firstyearFaculty of Engineering course, ENGG1000,Engineering Design. The project is to design, buildand demonstrate a solar powered “space elevator”to climb a suspended tether in the final weeks ofthe semester. The Prize is awarded to the fastestteam. The winners in 2010 were “Captain Planet”,team members being Mingyang Jiang, DanielMoctezuma-Baker, Syed Syed Ihsan, Edric Verbeek-Martin, Jack Wong and Heng Wu.N&M Suntastic Project PrizeThis prize, donated by the same sponsors as theN&M Prize, was offered again in 2010. It recognisesbest performance in SOLA2052 second year projectcourse. It was won by Jarred Rudman.Photovoltaics PrizeThis prize honours best performance in SOLA3540,Applied Photovoltaics, and was awarded in 2010 toYan, Xia and Yang, Chuanxi.Steve Robinson Memorial PrizeThis prize is in memory of a former PhD student,Steve Robinson, and recognises the bestpostgraduate performance in SOLA9002, SolarCells and Systems. It was awarded in 2010 to Pei-Chieh Hsiao.5.1.6 Engineers Australia Student of theYear (Sydney Division)SPREE student Chris McGrath, won the StudentEngineer of the Year Award at the Sydney Division’sEngineering Excellence Awards in September 2010.Chris finished his studies in 2009 with First ClassHonours with a thesis on designing and installing anano- hydro electric power system for the regionalprimary school in a remote village in Vanuatu. Hecontinues to support this work with a 2010 UNSWfinal year undergraduate student and to promotefurther sustainable energy development projectsin Vanuatu.5.1.7 University MedalsThree students from the School are to be awardedthe University Medal when they graduate in early2011 following completion of their studies in theSchool’s programs in 2010. The students are Xue Baiand Yu Feng (Photovoltaics and Solar Energy) andEleanor Wood (Renewable Energy).112

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTChris McGrath and Imakivillage workers marking thecompletion of the penstockinlet for the hydro turbine(Image: D. McGrath).Engineers AustraliaStudent of the Year(Sydney Division)5.2 Postgraduate ProgramsThe School offers three postgraduate coursework programs, a GraduateCertificate of Engineering Science, a Graduate Diploma of EngineeringScience and a Master of Engineering Science, and two researchprograms, Masters by Research and a Doctor of Philosophy. Thesedegrees are intended to provide students with an exceptional educationin advanced concepts and research in the photovoltaics or otherrenewable energy areas.The 1.5-year Master of Engineering Science coursework program wasdeveloped to build on the prior education of engineers from otherengineering disciplines who are attracted to the photovoltaics andrenewable energy industries. Students study courses chosen from theareas of photovoltaic devices, photovoltaic systems and applications,energy efficiency and renewable energy technologies. Enrolments inthe program have grown significantly from the previous small numbers,with total enrolments of 89 at the end of 2010. There were 37 graduatesin 2010 and 21 more are expected in early 2011.At the end of 2010, the School had 63 students enrolled in postgraduateresearch degrees, 52 in a Doctor of Philosophy and 11 in Mastersby Research. Research topics available to these students cover theentire photovoltaic sector and building energy efficiency, but withgreatest emphasis on device theory, design and production, moduledesign, balance of system components, and photovoltaic systemsand applications. Research students play a pivotal role throughout allthe Centre’s activities. In 2010 the Centre produced six PhDs, and twonew PhD and one Masters by research are expected to graduate inearly 2011.All the Faculty’s postgraduate coursework programs include anemphasis on engineering management and all students will be requiredto take some faculty-based “Engineering and Technical Management”courses of general importance. The School introduced its ManagingEnergy Efficiency (GSOE9017) course into this set of generally availablecourses in 2010 and it proved to be very popular.5.3 ScholarshipsThe undergraduate programs at the Centre attract very brightstudents from across the world. The UAI cut off for local studentsfor these programs was lifted to 91.7 for the 2010 intake. The Co-OpScholarship Program and the Faculty of Engineering’s Rural ScholarshipProgram have helped to attract these high-achieving students, whilethe Taste of Research Summer Scholarship Program and occasionalcasual employment opportunities provide experience in the School’slaboratories with research leaders and encourage them to pursueresearch careers. The Centre’s very high international profile alsoallows it to attract high quality research students, evidenced by thenumber doctoral students undertaking research programs at theCentre who have been awarded either of an APA (local students) or IPRS(international students) scholarships.5.3.1 Asia-Pacific Partnership onClean Development and ClimateThe School applied during 2006 for sponsorship funding from theAustralian Government under the Asia-Pacific Partnership on CleanDevelopment and Climate (APP) programme to bring students fromthe Asia-Pacific region, specifically China, India and South Korea,to study photovoltaics engineering at UNSW. In January 2007, APPProgramme investment of $5.2m was approved by the APP Ministersand the first students were able to come in 2008. This significant grantsupports study at UNSW by international PhD students (full fees),MEngSc coursework students (50% fees), and third and fourth yearundergraduate students (100% fees). The undergraduate arrangementof shared education requires detailed agreements between UNSWand the selected source universities, which are restricted to China.These scholarships will address a key need for additional expertise andknowledge in the rapid expansion of the photovoltaics industry in theregion and their award acknowledges UNSW’s importance in the field.In Semester 1 of 2008, these scholarships attracted five undergraduate,one postgraduate coursework and three PhD research students to theSchool. In Semester 2 of 2008, these scholarships brought in a furthereight undergraduate, eleven postgraduate coursework and two PhDresearch students. Enrolments for 2009 were much stronger and the APPstudents comprised a significant proportion of the year’s intake. Twentyoneundergraduate and forty-eight MEngSc APP students commencedtheir study in 2010.Agreements, known as “2+2” agreements, have been signed with sevenleading Chinese universities: Sun-Yat-Sen University and South ChinaUniversity of Science and Technology in Guangzhou; Nankai Universityand Tianjin Universities in Tianjin; Zhejiang University in Hangzhou;Nanchang University in Nanchang, Beijing Jiaotong Universityin Beijingu.113

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTSome of the students attractedto the School by the sponsorshipoffered by the AustralianCommonwealth Department ofthe Environment, Water, Heritageand the Arts, through the AsiaPacific Partnership for CleanDevelopment and Climate. Alsoshown in the group photo are MrDemin Zheng and Ms Feng Junying,Consuls, Education Office of theConsulate - General of the PeoplesRepublic of China in Sydney, MrMatthew Hopkins, Department ofEnvironment, Water, Heritage andArts and UNSW staff.Six MEngSc students completed the APP programat end of Semester 1 2009 and a further tenMEngSc and five undergraduate completed theAPP program at end of Semester 2 2009. In 2010,10 MEngSc students completed the APP programat end of Semester 1 and 12 at the end of Semester2 and 8 and 26 BEng students finished their studiesin Semesters 1 and 2, respectively. Several of thegraduating students have intended to enter theSchool’s PhD program but the numbers able todo so has been restricted by the availability ofscholarships.5.3.2 Co-Op Scholarship ProgramThe Co-Op Program is an industry-linkedscholarship program where students obtainextended work experience with industry sponsorsas part of their undergraduate studies, throughtwo summer placements, and two six monthplacements (prior to final year of coursework),with sponsor companies. In addition to theiroutstanding academic achievements, studentsare selected based on their involvement in schooland community activities, their demonstratedleadership and teamwork skills and their abilityto communicate. Participation in this programenables students to apply the knowledge theyhave gained during their studies in real industryprojects, as well as opportunities for networking,recruitment and professional development throughleadership camps. This program is also highlybeneficial to industry sponsors, who have access toexcellent students, can observe these students intheir workplace, and have the first opportunity ofrecruiting the best new graduates.Co-Op sponsors currently include, BP Solar, a globalcompany that has been supplying PV systems toAustralian markets for 30 years, CSG Solar, a newtechnology startup with an innovative crystallinesilicon on glass PV product, Suntech Power Co. Ltd.,the leading Chinese photovoltaic manufacturerand Renewable Energy Corp. (REC), a Norwegianbased silicon and photovoltaics company thatis establishing photovoltaics manufacturingin Singapore.A new sponsor, Infigen, awarded its first Co-Opscholarship and the School’s first for the RenewableEnergy Engineering Program in 2010.5.3.3 Rural Scholarship ProgramThe Faculty of Engineering established the RuralScholarship Program in 2001 to encourage highachievingstudents living in rural and isolated areasto study engineering. The scholarships are valued atapproximately $9,500 per annum for four years offull-time study which eases the financial hardshipof relocating to and living in Sydney. Two new ruralscholars began their studies in the School in 2010.5.3.4 Thyne Reid FoundationScholarshipsDuring 2008 two new undergraduate four-yearscholarships were awarded to the School by TheThyne Reid Foundation, one for Photovoltaics andSolar Energy Engineering and one for RenewableEnergy Engineering. The two students beganstudies in Semester 1 of 2009 and will proceed intotheir third year in 2011.5.3.5 Taste of Research SummerScholarship ProgramThe Taste of Research Summer Scholarship Programis primarily for high achieving 3 rd year studentsand, in exceptional cases, 2 nd year students may beconsidered. As part of the program, engineeringschools offer 10 week projects for students tocomplete during their summer break. Theseprojects provide students with scholarship supportto gain experience working as part of a researchteam, for example in the laboratories of theCentre. In addition to providing an opportunityfor checking whether research is their ideal careerpath, participation in these projects helps studentsfurther develop their technical skills and theirwritten and oral communication through writtenreports and poster presentation.From December 2009 to February 2010, theCentre hosted 12 students, taking advantage ofthe growing interest in the it’s research programsamong outstanding undergraduates, particularlyfrom those in this School. Again, in the 2010/2011summer the Centre was keen to retain for researchtraining as many as possible of the outstandingsenior undergraduate students and took on 10students through the Taste of Research program.5.3.6 Research ScholarshipsSowbaranigha Chinnusamy Jayanthi, a studentfrom India, was awarded an e8 Sustainable EnergyDevelopment Masters Scholarships for oneacademic year so that she may study for a Master ofEngineering Scienceion in Photovoltaics and SolarEnergy, at the School. The e8 is an organizationof leading electricity companies from the globalelectricity sector that wish to help in protectingthe global environment, and to promote theefficient generation and use of electricity. Thee8 has established a scholarship program tosupport outstanding students from developingcountries who plan to study at the Masters level114

in areas directly related to sustainable energydevelopmeant.Through thesis projects and scholarships, includingthe Taste of Research Scholarship program, theCentre encourages some of the best undergraduatestudents, particularly those from the School, intoresearch. Being an internationally recognisedresearch organisation, the Centre attracts veryhigh quality students to its research programswhere students have the opportunity to workwith leaders in this field. The fact that 21 of ourdoctoral students are currently in receipt of the verycompetitive Australian Postgraduate Award (APA)for local students or the Australian Government’sInternational Postgraduate Research Scholarship(IPRS) for international students is testament to thequality of these students. To be awarded an APA,the student must be a local student with first classhonours graduate at Bachelor degree level, whilethe IPRS is available to international students andare much more competitive. Suntech Power alsosponsors a PhD scholarship in the Centre. Interestin and applications for postgraduate research inthe School boomed during 2009 and the Schoolexpects to have four new IPRS recipients begintheir research programs in 2011. The Centre alsohas five PhD scholars sponsored by the AustralianGovernment through the Asia-Pacific Partnership onClean Development and Climate scheme.5.3.7 Rhodes ScholarshipsThe Rhodes Scholarship, arguably the mostprestigious scholarship in the world, is forpostgraduate study at Oxford University. Latein 2010 Alice Lang, a combined undergraduatePhotovoltaic Engineering/Arts degrees studentof SPREE with significant research experience inthe Centre’s laboratories, learned that she wouldbecome a Rhodes Scholar. Alice will complete herArts courses in Semester 1 of 2011 and start herstudies at Oxford in Semester 2.ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORT5.4 Educational Resoruces5.4.1 New Book TranslationA new book in Japanese language, “Fundamentalsand Applications of Solar Cells”, was publishedin 2010, including five chapters authored by theCentre’s Prof. Martin Green. A simplified Chinesetranslation of Silicon Solar Cells: AdvancedPrinciples and Practice” will be published byShanghai Jiao Tong University Press in early 2011,bringing to three the number of the Centre’s booksavailable in that language.5.5 PVSOCOne characteristic of the Centre which has led tothe success of the educational programs is thefriendly atmosphere that is engendered by beinga small (but strongly growing!) school with highlymotivated students and academic and generalstaff. Students appreciate being able to formfriendships and support networks with fellows,as well as feeling comfortable and familiar withacademic, technical and administrative staff.PVSOC (Photovoltaics Society) is a social committeeestablished by the students which fosters thisatmosphere with organised social events andactivities to encourage student interaction. In 2010,the committee organised several events, includingbarbeques, student participation in SustainableHouse Day (organised by ANZSES and held inSeptember) and the PVSOC Annual Dinner. Studentsand staff enjoyed this party at Coogee in September2010 (held earlier in the year than previously toavoid distraction from final week assignments andexam preparations) where final year students werefarewelled, staff thanked, and the 2011 PVSOCcommittee was elected.A new aspect of PVSOC’s activities began in2010 with the advent of regular consultations bythe PVSOC President, as a representative of thestudent body, with the Head of School and theUndergraduate Coordinator to work with the staffto further improve the delivery of education tothe students.Alice Lang, Rhodes Scholar for 2011.(Left) New Chinese book, “SiliconSolar Cells: Advanced Principlesand Practice” by M. Green will bereleased in early 2011.(Right) New Japanese book,“Fundamentals and Applicationsof Solar Cells” by M. Yamaguchi, M.Green, Y. Ohshita and N. Kojima.115

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORT5.6 Promotional ActivitiesThe Centre regularly participates in promotionalactivities organised by the Faculty of Engineeringand UNSW Student Recruitment. These events areimportant for increasing awareness and interest inthe Centre’s educational and research programs.5.6.1 The Shi’s Family CharitableFoundation and The SydneyTheatre CompanyThe School has been a major participant in theSydney Theatre Company’s high-profile “Greeningthe Wharf” project that is improving the waterefficiency and greenhouse gas impact of the WharfTheatre, at Walsh Bay. UNSW, in consultation withthe Shi’s Family Charitable Foundation has agreedthat the STC’s “Greening the Wharf” project is aproject suitable to showcase the Pluto technologyas an example of UNSW and Suntech’s researchexcellence. The School contributed $2m, donatedby the Shi’s Family Charitable Foundation, towardsthe establishment of a 380kW pgrid-connectedphotovoltaic array on the Wharf’s roof, completedand opened in December 2010. The balance offunding for the $4 million project was raised bySydney Theatre Company from The Commonwealthceremony with Federal Minister for Climate Changeand Water Penny Wong and Federal EnvironmentMinister Peter Garrett, Mrs Vivienne Shi, UNSWChancellor David Gonski, Scientia Professor MartinGreen and Dr Richard Corkish.5.6.2 Print and Broadcast MediaStudent recruitment advertisements werealso placed in the HSC Survival Magazine, theHSC Change of Preference Guide, and the HSCUniversities Advisory and Information Day Guide.Media mentions of the School and the Centrebecame more frequent in 2010 as communityinterest in climate change and renewable energyissues grew.5.6.3 UNSW Information DayLocal undergraduate students must apply foradmission to UNSW programs through theUniversities Admissions Centre (UAC), and the firstweek of January is the final opportunity for studentsto change their preferences for entry to universityprograms. Therefore, the university hosted aninformation day in January 2010 to assist studentsobtain information to finalise their preferences.School administrative and academic staff attendedthis event and talked to many prospective studentswho were unsure of their career direction. Studentsreceived information on our programs and hadthe opportunity to ask questions of staff. A finalround of interviews for the Faculty of EngineeringAdmissions Scheme (FEAS) was carried out on thesame day.Pictured at the launch of the WharfTheatre’s rooftop PV array are,(left-right):Dr. Richard Corkish,Head of School; Mr GeorgeMaltabarow, Managing Director,EnergyAustralia; The Hon. FrankSartor, NSW Minister for ClimateChange and the Environment;The Hon. Tony Burke MP, FederalMinister for Sustainability,Environment Water, Populationand Communities; Senator TheHon Don Farrell, ParliamentarySecretary for Sustainability andUrban Water; Cate Blanchett,Artistic co-Director, Sydney TheatreCompany; Dr Zhengrong Shi, CEOof Suntech Power Holdings; MrsVivienne Shi, Director, Shi’s FamilyCharity; Andrew Upton, Artistic co-Director, Sydney Theatre Company.(Image produced by and used withpermission of AAP).Government’s Green Precincts Fund, NSWGovernment’s Public Facilities Program, corporatesponsor Energy Australia and philanthropists.The project grew from two sources. Firstly, Dr ShiZhengrong, CEO of Suntech Power, expresseda wish to make a gift to UNSW and to Sydneyfor helping him start his successful career inphotovoltaics and to use it as a promotional tool.This led to a study of several prominent sites inSydney and a pre-feasibility study, carried out asa final year thesis project by Samantha Wong. Inparallel, Cate Blanchett and Andrew Upton, theArtistic Directors of the Sydney Theatre Company,were seeking opportunities to generate a significantpart of the Company’s energy from renewablesources. Bringing the two ideas together led tothe official launch in July 2009 of the project at a5.6.4 UNSW Courses and Careers DayUNSW Courses and Careers Day is the annualinformation day, held in the first week ofSeptember, for prospective students to obtaininformation about programs and student life atthe university. As part of this day students andacademic and administrative staff from the Schoolprovided advice and information to prospectivestudents from information desks in the Roundhouseand in a marquee in the UNSW Quad, nearby theSunsprint model solar car race finals. This year,the weather was, unfortunately, far from ideal forthe 2010 event. During the day Dr Alistair Sproulpresented two public lectures about the School’sand Centre’s programs as part of the lecture seriesorganised centrally by the university.5.6.5 Faculty of Engineering InformationDay for High School StudentsEach year the Faculty of Engineering organisesan information day to give high school studentsand their teachers an opportunity to learn aboutengineering and the programs offered at UNSW.As part of the day students visit three engineeringschools of their choice and engage in interactiveactivities aimed at demonstrating the relevantengineering area. Once again in 2010 the demandto attend this event was so great that two HighSchools Days were required, with the Schoolrunning several tours.116

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTSome of the students who participated in the UNSWSunSprint Model Solar Car Challenge.5.6.6 Honeywell EngineeringSummer SchoolThe Honeywell Engineering Summer School isan event held in December and is conducted byEngineers Australia. As part of the summer schoolhigh school students from across NSW and the ACTabout to enter their final year take part in a week ofactivities which involves industry visits and lecturesand demonstrations at a number of universities.In 2010 approximately 30 students visited theCentre to participate in hands-on activities andinformation provision.5.6.7 Model Solar Vehicle ChallengesSunSprint Model Solar Car ChallengeThe SunSprint Model Solar Car Challenge is aphotovoltaics project-based learning eventdesigned for high school students. In the event,solar cars are designed and built accordingto specifications provided by the Australian-International Model Solar Challenge (AIMSC), anational organization overseeing and linking all ofthe state events. These specifications provide designcriteria that, when followed, allow students all overAustralia to build model solar cars that are uniformlycompetitive with respect to cost and use ofmaterials. The students contribute ideas, inspiration,time and money. The average vehicle costs over$600 and takes the team over four months to designand build. Many schools throughout NSW haveallowed final year design and technology studentsto adopt SunSprint as their year-long major project.High school physics and science teachers havealso used SunSprint as a class-based teachingproject as SunSprint addresses several aspects ofthe curriculum.The UNSW Faculty of Engineering initiated theproject in 1996 with the Photovoltaics Centreof Excellence taking over its operation in 1999.In 2005 the UNSW Faculty of Science joinedthe Faculty of Engineering and the Centre injointly sponsoring the event.The 2010 SunSprint Model Solar Car Challengewas run on September 3 rd and 4 th , 2010 undertruly horrible conditions with extraordinaryamounts of rain and virtually no sunlight. Thewinning cars were able to actively competeusing only 10% of the normal amountof sunshine.Only 10% Sunshine and lots ofrain at the UNSW Model SolarVehicle Challenges.117

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTThe MiniSprint Competition.MiniSprint and Model Solar BoatsThe MiniSprint Model Solar Car Challengeand the Model Solar Boat Challengeare offshoot events allowing primaryschool students to design and build solarvehicles and then experience the thrill ofcompetition. The MiniSprint competitorsstart with a kit containing all of theimportant parts of the solar car and canbe built in an evening. The kit requires soldering,gluing and taping.The Model Solar Boat Competition requires muchmore student design than in the MiniSprint. Theboats can be made of any material and can usealmost any size or shape of solar cells and motors.Getting the boat to float right side up and tocompetitively move through the water is quitea feat.Preparations for the Model SolarBoat Competition.The Model Solar Boat Competition.5.7 EducationalCollaborationsA range of educational collaborations havebeen established between the Centreand other educational institutions andorganisations. These collaborations involvethe development and implementation ofeducational programs and courses, theprovision of support for student projects andtheses, and the exchange of students.5.7.1 “2+2” AgreementsIn 2010 the School continued collaborationswith four Chinese Universities for thirdand fourth year students from thoseuniversities to study Photovoltaics andSolar Energy Engineering at UNSW andfor Masters by coursework study with AsiaPacific Partnership sponsorship. “2+2”agreements were signed in 2007 with twoprestigious Chinese universities, NankaiUniversity in Tianjin and Sun Yat-Sen(Zhongshan) University in Guangzhou. Thefirst students under this scheme started atUNSW in Semester1 of 2008. During 2008similar agreements were signed with additionaltwo Chinese universities. The two new partneruniversities are Tianjin University (two agreements,each with a different school) and ZhejiangUniversity, Hangzhou.A new agreement has now been made in 2009with Nanchang University, in Nanchang, which hasformed a new School of Photovoltaic Engineeringand an additional “2+2” agreement was signed inearly 2010 with South China University of Scienceand Technology, in Guangzhou.5.7.2 Thyne Reid FoundationThe Thyne Reid Foundation concluded in 2008an agreement to sponsor the studies of twoundergraduate students for four years of theirstudies, one in Photovoltaics and Solar EnergyEngineering and one in Renewable EnergyEngineering. The students began their studies in thefirst semester of 2009 and finished their second yearin 2010.Trophies for primary schoolparticipants.118

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTThe project team and Imakivillage volunteers preparingthe long trench for the powercables from the turbine to thevillage. (Image produced byand used with permission ofE. Tehan).5.7.3 Cundall, Infigen, NorcoastRefrigeration and OthersCundall, a sustainable design consulting company,and the Darcy and Patricia Wentworth Fund joinedforces to financially support a final year thesisproject to bring hydro power to a remote villageschool in Vanuatu, mentioned in Sec. 5.1.Thestudent undertaking this project, Eden Tehan, wasstrongly supported by an alumnus, Chris McGrath,who started the project in 2009. In 2010 the projectwas extended to include the secondary (boarding)school, medical clinic and church in an integratedpower system supplied by two hydro turbinesand including a new low-power electric vaccinerefigerator. Donors for the project extension in 2010included Cundall, Infigen, Norcoast Refrigeration,and several private donors: John Tehan, SusanMcParland, Sean Tehan, Don and Marg McGrath,Tyler Troy, Helena Colnaric, Leah Greengarten,Athron McCann, Peter Baker, Noam Field and DavidGazal. Eden and Chris also raised $4,500 for theproject at a Vanuatu Shakedown concert, thanksto donated performances by “The Bakery”, “TheLiberators” and Emma Davis. David Thompsonof Gwynneville Electrical Services was the team’selectrician during the first trip, paying his ownway and donating a week of work as well as manyelectrical components. Dave Selke of All-roundSupplies, Sydney, was a very helpful tradesman whoalso donated some vital electrical components.Samantha Paver, Sasha Siljanovic and Sean Tehan allcontributed their time to the project.5.7.4 Suntech Power CorporationSuntech, a partly Australian-owned companyoperating in China, has been actively involved inassisting UNSW with the development, testingand evaluation of educational material in thePV area such as in the development, testing andevaluation of a teaching software package calledthe Virtual Production Line. Significant numbersof UNSW undergraduate students participate inindustrial training or in “Taste of Research” projectsat Suntech and postgraduate students often travelthere to use vital research equipment. Additionally,Suntech sponsored an undergraduate CooperativeScholarship and two PhD scholarships and hasmade significant financial and in-kind contributionsto UNSW to support photovoltaics research.5.7.5 Singapore Polytechnic, Ngee AnnPolytechnic and Temasek PolytechnicThe Polytechnics in Singapore offer three-yeardiploma programs, including in engineering.During 2007 the School arranged for futurearticulation of graduates from the new clean energydiploma of Clean Energy program at SingaporePolytechnic into either of the School’s Bachelor ofEngineering programs. In 2008 the School set upa similar articulation arrangement with Ngee AnnPolytechnic and Temasek Polytechnic was includedin 2009. Significant numbers of students may flowin future years from these institutions, once the firstgroups of students make their way through thenew programs, fuelling Singapore’s drive to becomea major photovoltaics manufacturer. RichardCorkish has accepted in 2010 the role of ExternalExaminer for the Diploma in Clean Energy programSingapore Polytechnic.5.7.6 Darcy WenworthDarcy Wentworth donated his services through2010 as a visiting academic, supervisingundergraduate student projects and coordinatingtwo subjects.5.7.7 Kema ConsultingDr Hanzheng Duo, a practicing Engineer at KemaConsulting, in his role as visiting academic, donatedhis time and industrial project experience tosupervise and assess an undergraduate thesisproject through 2010.119

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORT6. ORGANISATIONALSTRUCTUREWhen the Key Centre for Photovoltaic Engineeringwas awarded in 1999, it was established asan autonomous Centre within the Faculty ofEngineering of UNSW, becoming independentfrom the School of Electrical Engineering whereUNSW photovoltaic activities had been previouslylocated. At that time, the Key Centre was giventhe same operational independence and rights asthe other Engineering schools, with the DirectorStuart Wenham being given the same status andauthority as other Heads of Schools. With theawarding of the Centre of Excellence in 2003,the University committed to the Key Centre forPhotovoltaic Engineering becoming the Schoolof Photovoltaic Engineering, within which theARC Photovoltaics Centre of Excellence would belocated. This transition occurred in January 2006.The corresponding organisational relationships areshown in the upper part of Fig. 6.1.Within the Centre of Excellence, Deputy Directorshave been appointed for the major strands ofresearch as shown. The Management Committeeof the Centre comprises the six Directors andDeputy Directors, along with the Head of Schooland the Business and Operations Manager. Thiscommittee meets fortnightly on the 1 st and 3 rdFridays of each month, with each Deputy Directorgiving a report on the activities in his or her areaover the preceding fortnight. This committeetakes responsibility for decision making withinthe Centre that affects the Centre as a whole,while the individual Deputy Directors receive theirown annual budgets to allow them to make andimplement decisions that impact only their ownlaboratory areas and research activities.The Advisory Committee for the Centre comprisesthe Centre Directors, the Head of School forPhotovoltaic Engineering, leading academics fromother institutions, industry leaders such as CEO’sof various companies involved in the field, andresearch leaders. This committee provides high leveladvice, feedback and recommendations in relationto the Centre’s activities and their relevance. TheAdvisory committee meets annually, taking intoaccount the geographical separation of its members(Australia, USA, China, Germany, Italy, Spain),although more frequent correspondence takesplace when necessary with individual committeemembers either by email or telephone. Themembership of the Advisory Committee during2010 has included:• Professor Stuart Wenham, Centre Director• Professor Martin Green, Centre ExecutiveResearch Director• Associate Professor Gavin Conibeer, DeputyDirector (3 rd Generation PV Research)• Associate Professor Thorsten Trupke, DeputyDirector (Silicon Photonics)• Dr Sergey Varlamov, Deputy Director (2 ndGeneration PV research)• Professor Allen Barnett: world leading academicin the field; recipient of many prestigiousinternational prizes such as the William CherryAward; former President of AstroPower, one of theworld’s largest solar cell manufacturers, beforebeing purchased by GE.• Prof. Andres Cuevas, ANU;• Dr Francesca Ferrazza, Chief Scientific Officer,Enitechnologie;• David Hogg: COO of Suntech Power; until recentlyCEO of CSG Solar AC, Germany; former CEO ofPacific Solar, Australia;• Mr. David Jordan, until recently, Manager of NewTechnology, BP Solar International;• Dr Larry Kazmerski, until recently Head,Photovoltaic Division, US National RenewableEnergy Laboratory;• Prof. Antonio Luque, Head, ESTI, PolytechnicalUniversity of Madrid; Leading European academicin the photovoltaic field; founder of Isofoton, oneof Europe’s most established manufacturers;• Dr Zhengrong Shi, CEO of Suntech-Power, theworld’s largest silicon cell manufacturer; recipientof various industry prizes; former Deputy ResearchDirector of Pacific Solar; major collaborator ofthe Centre;.• Prof. Peter Würfel, Institut fur Angewandte Physikder Universität Karlsruhe.120

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTCentre organisational structure.Figure 6.1Other distinguished researchers and industrialistshave been co-opted to the Committee as required.Also within the Management structure and amember of the Centre’s Management Committee isthe Centre’s Business, Technology and OperationsManager Mark Silver, given the extensiveinvolvement of Centre staff with industry, largenumber of collaborative research projects, and thehigh level of success of the Centre in generating,marketing and commercialising technology. TheHead of Administrative Support for the Centreduring 2010 was Joyce Ho. Not shown in theManagement Structure is the Financial OfficerJulie Kwan who looks after the Centre Accounts,interfaces with the University Financial System,advises the Centre Directors and Deputy Directorson budgeting and financial matters, and generallyassists with sourcing of funding, purchasing,reporting, writing grant applications, and advisingon business opportunities.purchasing of appropriate new infrastructureand equipment.The Director works closely with the Centre’sExecutive Research Director to assist in settinghigh level research direction and priorities forthe Centre’s programs. In the teaching area, withthe Key Centre for Photovoltaic Engineeringand its activities being incorporated into theCentre of Excellence, the Director also takessignificant responsibility in ensuring the successfuldevelopment and implementation of the Centre’seducational programs, both at undergraduate andpostgraduate levels. In this area, the Director worksclosely with Dr Richard Corkish, the Head of School.A primary responsibility of the Director in theCentre is to ensure the Management structure asshown functions efficiently and effectively, withappropriate support and resources to facilitatethe Centre’s achievement of its milestones andperformance targets. An important aspect of this ismanaging the Centre’s finances to ensure suitablelevels of funding are made available for supportingthe major research strands, while simultaneouslyensuring adequate funding is available for repairand maintenance of existing facilities and the121

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORT7. FINANCIALSUMMARYThe total income for 2010 for the AustralianResearch Council (ARC) Photovoltaics Centre ofExcellence was $6.8 million. This does not includeover $2 million from the Host Institution based onEFTSU and related income. This income has beengenerated by the Centre through its educationalactivities that have formed the basis of a Schoolwithin the Faculty of Engineering. This income hasbeen used for the development of new courses andteaching materials and to fund the salaries of mostof the academic staff associated with the Centre.The largest two components of Centre income interms of cash contributions were from the ARCwith $2.1 million for the Centre Grant and industrywith $2.0 million. The former was complementedby a further $0.49 million in ARC funding for othergrants. However, for the latter, in-kind contributionsfor collaborative research add a further almost$1 million, effectively making industry funds thelargest income source for the Centre for the fourthconsecutive year. With the booming photovoltaicindustry, the success of Centre PV technology andthe high demand for companies wishing to workwith the Centre, this trend is expected to continue.Figure 7.2 shows a more detailed breakdown ofincome derived from industry and related sources.The largest industry contributor was again Suntech-Power, but with the majority of the $0.63 millionbeing in-kind contributions resulting from mostof the collaborative research being carried out atthe premises of Suntech in China. The largest cashcontributor was the consortium of Toyota, GeneralElectric, Schlumberger and ExxonMobilfor the Global Climate & Energy Project,administered by the Stanford University,who contributed $0.61 million. The nextthree largest cash contributions were from CSGSolar from Australia, Shinsung Holdings from Koreaand Guodian Solar from China, each contributingabout $0.3 million. Other cash and in-kindcontributions have been made by a large numberof industry collaborators and contributors of whomthe most significant have been Hyundi HeavyIndustries, the European Commission, RenewableEnergy Corporation (REC), BP Solar, CEEG NanjingPV Tech, Toyota Motor Group, E-ton Solar, Silex, Roth& Rau, Global Sunrise Energy, Infigen Energy, theAustralian Academy of Science and Advent Solar. Inaddition, many other companies contribute to theCentre through the expertise and experience theyoffer in collaborative research areas and the in-kindsupport they provide through access to equipment,facilities and personnel not available at UNSW.The third largest cash component of income wasHost Institution support. UNSW contributed $1.0million not including additional Host Institutionsupport in the form of EFTSU and related income.The next largest cash component of income was$0.70 million from the State Government, primarilythrough the Department of State and RegionalDevelopment (DSRD). This funding has beenparticularly important for the purchase of newequipment and the development of facilities.Other Centre contracts, consulting work andtechnology transfers are conducted throughNewSouth Innovations, the commercial arm ofthe university. These are handled on behalf of122

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORTSources of Centre income.Figure 7.1the Centre of Excellence and its staff throughthe NewSouth Innovations accounts and notincluded in this financial report.The Centre also earns income through thesale of educational CDs, books and computersoftware and running short courses.Centre 2010 Expenditure of ARC Grant was $2.3million as shown in Figure 7.3, approximately$200k more than the corresponding income.By far the largest component of expenditurewas for salaries and scholarships with $1.4million, including for research appointments,technical support, administrative staff, financialmanagement, specialised consultants andfacility cleaning. Appointments include fulltime,part-time and casual.Equipment expenditure, mainly for small itemsor small contributions leveraged to purchaselarge items, was $101k from the ARC Grant. Withthe large amount of industry funded research,travel expenditure was quite significant at$145k, with several staff and students makingmany trips to industry partners for industrycollaborative research and technology transfers.Other travel included attendance and paperpresentation at international conferences.2010 industry income to the ARCPhotovoltaics Centre of Excellence.Figure 7.2Another major component of Centreexpenditure from the ARC grant is onconsumables and maintenance with $633K.These costs are strongly dominated bylaboratory consumables to support the deviceresearch such as high purity gases, chemicalsand general laboratory supplies.Expenditure of the ARC grant byexpenditure category.Figure 7.3123

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11ANNUAL REPORT8. PUBLICATIONS8.1 BOOKY. Yamaguchi, Martin A Green, Y. Ohshita and N. Kojima, “Fundamentals and Applications of Solar Cells”, ISBN 978-4-621-08211-9, 231pp (Maruzen, Japan, 2010)8.2 BOOK CHAPTERSG. Conibeer, “Up and Down Conversion forPhotovoltaics” in “Physics of Solar cells” Ed. Badescu, V& Paulescu, M, accepted after review April 2009 (NovaScience Publishers, July 2010).G. Conibeer, “Applications of Si Nanocrystals inPhotovoltaic Solar Cells” in “Silicon Nanocrystals:Physics and Applications”, Ed. Rasit Turan and LorenzoPavesi (Wiley VCH, 2010).R. Corkish, “Solar Energy”, in Mohd. Nordin Hassanand Sukanta Roy (eds.), “Sustainable Energy in Asiaand the Pacific”, Akademi Sains Malaysia, 2010 ISBN:978.983-9445-43-5 (invited).8.3 patent applicationsI. Brazil and A. Lennon, “Method for TexturingSurfaces”, International Patent Application PCT/AU2010/001208, filed on 16 September 2010.G. Conibeer, S. Shrestha, D. Konig and M.A. Green,“Hot Carrier Energy Conversion Structure andMethod of Fabricating the Same”, InternationalPatent Application PCT/AU2010/000848, filed on 2July 2010.R. Corkish, B. Jaya, Y. Augarten, C. Rizos and L. Ge,“Method to Detect Faults in Photovoltaic Arraysby Observation of Electroluminescence Signal”, IPNotification to New South Innovations, 1 July 2010.M.A. Green,“Improved Scheme for Light-Trappingin Organic Photovoltaic and Related Thin-FilmOptoelectronic Devices”, Australian ProvisionalApplication 2010900576, filed on 12 February 2010.M.A. Green, “Si Cells with III-V Multijunction Cells”,Australian Provisional Application 2010900574, filedon 12 February 2010.M.A. Green, A. Wenham and S.R. Wenham, “ImprovedMetal Adhesion”, International Patent ApplicationPCT/AU2010/000215, filed on 24 February 2010.P. Hamer and S.R. Wenham, “Method of Fabricatinga P-N Junction”, Australian Provisional Application2010905292, filed on 1 December 2010.D. König, “Acceptor Modulation Doping ofNanocrystals and Quantum Structures, InternationalPatent Application PCT/AU2010/001043, filed on 13August 2010.D. König, “Donor Modulation Doping of QuantumStructures”, International Patent Application PCT/AU2010/001042, filed on 13 August 2010.A. Lennon, D. Lu and Y. Chen, “Metal Contact Schemefor Solar Cell”, Australian Provisional Application2010902694, filed on 2 July 2010.A. Lennon, S.R. Wenham and A. Ho-Baillie, “Methodof Patterned Etching of Selected Material”, NationalPhase Entry PCT/AU2009/000098 filed on 1August 2010.A. Lennon and S.R. Wenham, “A Method for theSelective Delivery of Material to a Substrate”,International Patent Application PCT/AU2010/001001, filed on 6 August 2010.A. Lennon, Y. Yao, K. Chin and D. Coyle, “Method andApparatus for Light Induced Plating of Solar Cells”,Australian Provisional Application 2010903869, filedon 27 August 2010.D. König, “Photovoltaic Device Physics at theNanoscale”, Chapter 3 (pp. 73 - 146) in L. Tsakalakos(Ed.), Nanotechnology for Photovoltaics, Taylor &Francis, April 2010, ISBN-13: 9781420076745 [AmazonBooks bestsellers rank 680,938] (invited).L. Mai, A. Wenham and S.R. Wenham, “Solar CellMethods and Structures”, International PatentApplication PCT/AU2010/000036, filed on 15 January2010.L. Mai, A. Wenham and S.R. Wenham, “Photoplatingof Metal Electrodes for Solar Cells”, InternationalPatent Application PCT/AU2010/001460, filed on 3November 2010.C-Y. Tsao, X.J. Hao and M.A. Green, “A Methodof Forming a Germanium Layer on a SiliconSubstrate and a Photovoltaic Device Including aGermanium Layer”, Australian Provisional Application2010905488, filed on 15 December 2010.S.R. Wenham, B. Tjahjono and L. Mai, “PhotovoltaicDevice Structure and Method”, International PatentApplication PCT/AU2010/000145, filed on 11February 2010.S.R. Wenham, B. Tjahjono, A. Lennon and N. Kuepper,“Improved Metallization Method for Silicon SolarCells”, International Patent Application PCT/AU2010/001421, filed on 25 October 2010.S.R. Wenham, B. Tjahjono, A. Lennon and Y. Yao,“Method and Apparatus for Light Induced Platingof Solar Cells”, International Patent Application PCT/AU2010/001461, filed on 3 November 2010.124

8.4 Papers in refereed Scientific andTechnical JournalsP. Aliberti, S.K. Shrestha, R. Teuscher, B. Zhang, M.A.Green and G.J. Conibeer, ”Study of Silicon QuantumDots in a SiO 2Matrix for Energy Selective ContactsApplications”, Solar Energy Materials & Solar Cells, Vol.94, pp. 1936-1941, 2010.P. Aliberti, Y. Feng, Y. Takeda, S.K. Shrestha, M.A.Green and G. Conibeer, “Investigation of TheoreticalEfficiency Limit of Hot Carriers Solar Cells with a BulkIndium Nitride Absorber”, Journal of Applied Physics,Vol. 108, 094507 (10 pp.), 2010.Y. Augarten, T. Trupke, M. Lenio, J. Bauer, J. Weber,M. Juhl and O. Breitenstein, “Calculation ofQuantitative Shunt Values Using PhotoluminescenceImaging”, Progress in Photovoltaics (re-submittedJanuary 2011).G. Barber, T. Mallouk, S. Zakeer, M. Graetzel, M.A.Green, A. Lee, P.G. Hoertz, N. Abrams, J. Mikulca and P.Liskat, “Tandem Hybrid Concentrator Using Dye Cellsand Silicon”, Journal of Physical Chemistry Letters, Vol.2, pp. 518-585. 2011.G. Barber, S-H. Lee, P. Liska, Paul, M. Graetzel, S.M.Zakeeruddin, M.A. Green, A. Ho-Baillie, J. Mikulca,T. Mallouk, P. Hoertz and N. Abrams, “Utilization ofDirect and Diffuse Sunlight in a Dye-Sensitized SolarCell – Silicon Photovoltaic Hybrid ConcentratorSystem”, Journal of Physical Chemistry Letters(submitted 9 February 2011).I. Brazil and M.A. Green, “Measuring Strain Changesduring Production of Thin Film Crystalline Siliconon Glass Photovoltaic Modules”, Journal of MaterialsScience: Materials in Electronics, Vol. 21, No. 11, pp.1207-1211, 2010.I. Brazil and M.A. Green, “Investigating PolysiliconThin Film Structural Changes during Rapid ThermalAnnealing of a Thin Film Crystalline Silicon on GlassSolar Cell”, Journal of Materials Science: Materials inElectronics, Vol. 21 (10), pp. 994-999, 2010.R. Clady, M.J.Y. Tayebjee, P. Aliberti, D. König, N.J.Ekins-Daukes, G.J. Conibeer, T.W. Schmidt and M.A.Green, “Femtosecond Investigation of High EnergyCarrier Relaxation Processes in GaAs and InP forHot Carrier Solar Cell Applications”, Progress inPhotovoltaics (accepted October 2010).R. Clady, T. Schmidt, M. Tayebjee, P. Aliberti, G.Conibeer, D. Konig, M.A. Green and N.J. Ekins-Daukes,“Interplay Between the Hot Phonon Effect andIntervalley Scattering on the Cooling Rate of HotCarrier in GaAs and InP”, Progress in Photovoltaics:Research and Applications (accepted 10 March 2011).G. Conibeer, M.A. Green, D. Konig, I. Perez-Wurfl, S.Huang, X. Hao, D. Di, L. Shi, S. Shrestha, B. Puthen-Veettil, Y. So, B. Zhang and Z. Wan, “Silicon QuantumDot Based Solar Cells: Addressing the Issues ofDoping, Voltage and Current Transport”, Progress inPhotovoltaics: Res. Appl. 18 (2010) In press, Publishedonline in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/pip.1045 (acceptedSeptember 2010).G. Conibeer and S. Huang, “International Journal ofNanoparticles”, ed Chang Hui Ye, Nanomaterials forThird Generation Photovoltaics (in press), 2010.G. Conibeer, R. Patterson, L. Huang, J-F. Guillemoles,D. König, S. Shrestha and M.A. Green, “Modellingof Hot Carrier Solar Cell Absorbers”, Solar EnergyMaterials and Solar Cells, Vol. 94, pp. 1516-1521, 2010(doi:10.1016/j.solmat.2010.01.018).Richard Corkish, Anna Bruce and Deo Prasad,“Integrated Solar Photovoltaics for Buildings”, Reviewof Architecture and Building Science (ISSN: 1225-1666), Vol. 54 (1), pp.86-92, 2010.D. Di, I. Perez-Wurfl, A. Gentle, D-H. Kim, X. Hao,L. Shi, G. Conibeer and M.A. Green, “Impacts ofPost-metallisation Processes on the Electrical andPhotovoltaic Properties of Si Quantum Dot SolarCells”, Nanoscale Research Letters, Vol. 5 (11), pp.1762-1767, 2010.D. Di, I. Perez-Wurfl, G. Conibeer and M. A. Green,“Formation and Photoluminescence of Si QuantumDots in SiO 2/Si 3N 4Hybrid Matrix for All-Si TandemSolar Cells”, Solar Energy Materials and Solar Cells, Vol.94 , pp. 2238-2243, December 2010.C. Flynn, D. König, I. Perez-Wurfl, M.A. Green and G.Conibeer, “Correlation between Fixed Charge andCapacitance Peaks in Silicon Nanocrystal Metal-Insulator-Semiconductor Devices”, SemiconductorScience and Technology, Vol. 25, 045011 (8pp), 2010.C. Flynn, D. König, M.A. Green and G. Conibeer,“Modelling of Metal-Insulator-SemiconductorDevices Featuring a Silicon Quantum Well”, Physica E,Vol. 42, pp. 2211-2217, 2010.M.A. Green and A. Ho, “43% Composite Split-Spectrum Concentrator Solar Cell Efficiency”, Progressin Photovoltaics: Research and Applications, Vol. 18,pp. 42-47, 2010.M.A. Green, K. Emery, Y. Hishikawa and W. Warta,“Solar Cell Efficiency Tables (Version 35), Progress inPhotovoltaics, Vol. 18, pp.144-150, 2010.M.A. Green, K. Emery, Y. Hishikawa and W. Warta,“Solar Cell Efficiency Tables (Version 36), Progress inPhotovoltaics, 2010, Vol. 18 (5), pp. 346-352, 2010.M.A. Green, “Enhanced Evanescent Mode LightTrapping in Organic Solar Cells and Other Low IndexOptoelectronic Devices”, Progress in Photovoltaics,Vol. 19, pp. 473-477, 2011.M.A. Green, “Learning Experience for Thin-Film SolarModules: First Solar, Inc. Case Study”, Progress inPhotovoltaics, Vol. 19, pp. 498-500, 2011.M.A. Green, “Is Sour Crude or Sour Gas a PotentialSource of Se and Te?”, Progress in Photovoltaics(accepted 1 February 2011).M.A. Green, “Analytical Treatment of Trivich-Flinnand Shockley-Queisser Photovoltaic Efficiency LimitsUsing Polylogarithms” Progress in Photovoltaics:Research and Applications (accepted 11 March 2011).P.J. Gress, P.I. Widenborg, S. Varlamov and A.G. Aberle,“Wire Bonding As a Cell Interconnection Techniquefor Polycrystalline Silicon Thin-film Solar Cells onGlass”, Progress in Photovoltaics: Research andApplications, Vol. 18, pp. 221-228, 2010.Z. Hameiri, L. Mai, A.B. Sproul and S.R. Wenham,“18.7% Efficient Laser Doped Solar Cell on CZ p-typeSilicon”, Applied Physics Letters, Vol. 97, pp. 222111-222113, 2010.Z. Hameiri, L. Mai and S.R. Wenham, “Advantages ofPhotoplating for Laser Doped Solar Cells”, Progressin Photovoltaics: Research and Applications (DOI:10.1002/pip.1072), 2010.Z. Hameiri, T. Puzzer, L Mai, A.B. Sproul and S.R.Wenham, “Laser Induced Defects in Laser DopedSolar Cells”, Progress in Photovoltaics: Research andApplications (DOI: 10.1002/pip.1043), 2010.Z. Hameiri, T. Puzzer, L Mai and S.R. Wenham, “SR.Influence of Laser Power on the Properties of LaserDoped Solar Cells”, Solar Energy Materials and SolarCells, 95, pp. 1085-1094, 2011.S. He, J. Jannsens, J. Wong and A. Sproul, “TheInfluence of Base Doping Density on the Performanceof Evaporated Poly-Si Thin-Film Solar Cells bySolid-Phase Epitaxy”, Thin Solid Films, Vol. 519, pp.475-478, 2010.S. He and A. Sproul, “Optical Properties of EvaporatedPoly-Si Thin-Films on Glass”, Thin Solid Films, Vol. 519,pp. 351-356, 2010.S. He, J. Wong, D. Inns, B. Hoex, A. Aberle and A.Sproul, “Influences of Oxygen Contamination onEvaporated Poly-Si Thin-Film Solar Cells by Solid-Phase Epitaxy”, Thin Solid Films, Vol. 518, pp. 4351-4355, 2010.G. Jin, P. I. Widenborg, P. Campbell and S. Varlamov,“Lambertian Matched Absorption Enhancementin PECVD Poly-Si Thin Film on AluminiumInduced Textured Glass Superstrates for Solar CellApplications”, Progress in Photovoltaics: Research andApplication, Vol. 18, pp. 582-589, 2010.D. König, K. Casalenuovo, Y. Takeda, G. Conibeer,J.F. Guillemoles, R. Patterson, L.M. Huang and M.A.Green, “Hot Carrier Solar Cells: Principles, Design andMaterials”, Physica E: Low-Dimensional Systems andNanostructures, Vol. 42, pp. 2862-2866, 2010.A. Lahreche, Y. Beggah and R. Corkish, “The Effectof Electron Range on Electron Beam InducedCurrent Collection and a Simple Method to Extractan Electron Range for Any Generation Function”,Ultramicroscopy (submitted July 2010).M. Lenio, A. J. Lennon, A. Ho-Baillie and S. R. Wenham.“Effect of Electroless Nickel on the Series Resistanceof High-Efficiency Inkjet Printed Passivated EmitterRear Contacted Solar Cells”. Solar Energy Materialsand Solar Cells, Vol. 94 (12), pp. 2102-2107,December 2010.D. Lin, L. Ma, G. Conibeer and I. Perez-Wurfl,“Study on electrical properties of Si quantum dotsbased materials”, Physica Status Solidi (acceptedSeptember 2010).L. Ma, D. Lin, G. Conibeer and I. Perez-Wurfl,“Introducing Dopants by Diffusion to Improve theConductivity of Silicon Quantum Dot Materials in 3rdGeneration Photovoltaic Device”, Physica Status Solidi(accepted September 2010).B. Mitchell, T. Trupke, J. Weber and J. Nyhus, “BulkMinority Carrier Lifetimes and Doping of SiliconBricks from Photoluminescence Intensity Ratios”,Journal of Applied Physics (accepted March 2011).E.J. Mitchell, S. Lindekugel, M. Künle, K. Schillinger,S. Janz and S. Reber, “c-Si Wafer-equivalent EpitaxialThin-film Solar Cells on Isolating Substrates”, SolarEnergy Materials and Solar Cells, Vol. 95, Issue 4, pp.1163-1167, 2011 (doi:10.1016/i.solmat.2010.12.046).E.J. Mitchell, N. Brinkmann and S. Reber, “Progresswith Epitaxy Wrap-through Crystalline Silicon ThinfilmSolar Cells”, Progress in Photovoltaics: Researchand Applications (accepted December 2010).Z. Ouyang, O. Kunz, A. B. Sproul and S. Varlamov,“Influence of the Absorber Doping for p-typeEvaporated Polycrystalline Silicon Thin-Film SolarCells On Glass, Journal of Applied Physics, Vol. 109,p.060104, 2011.Z. Ouyang, S. Pillai, F. Beck, O. Kunz, S. Varlamov, K.Catchpole, P. Campbell and M.A. Green, “EffectiveLight Trapping in Polycrystalline Silicon Thin-FilmSolar Cells by Means of Rear Localised SurfacePlasmons”, Applied Physics Letters, Vol. 96, p.261109, 2010.125

Z. Ouyang, X. Zhao, S. Varlamov, Y. Tao, J. Wongand S. Pillai, “Nanoparticles Enhanced Light-Trapping in Thin-Film Silicon Solar Cells”, Progress inPhotovoltaics (submitted November 2010).R. Patterson, M. Kirkengen, B. Puthen-Veettil, D.Konig, M.A. Green and G. Conibeer, “Phonon Lifetimesin Model Quantum Dot Superlattice Systems withApplications to the Hot Carrier Solar Cell”, SolarEnergy Materials & Solar Cells, Vol. 94, pp. 1931-1935, 2010.S. Pillai and M.A. Green, “Plasmonics for PhotovoltaicApplications”, Solar Energy Materials and Solar Cells,Vol. 94, pp. 1481-1486, 2010.S. Pillai, I. Perez-Wurfl, G.J. Conibeer and M.A. Green,“Surface Plasmons for Improving the Performance ofQuantum Dot Structures for Third Generation SolarCell Applications”, Physica Status Solidi, Vol. 8, pp.181-184, 2011.B. Puthen-Veettil, D. König, G. Conibeer and M.A.Green, “Impact of Disorder in Double Barrier QDStructures on Energy Selectivity Investigated by TwoDimensional Effective Mass Approximation”, EnergyProcedia, Vol. 2, p. 213, 2010.Santosh K. Shrestha, Pasquale Aliberti and Gavin J.Conibeer, “Energy Selective Contacts for Hot CarrierSolar Cells”, Solar Energy Materials and Solar Cells, Vol.94, pp. 1546-1550, September 2010.R.A Sinton and T. Trupke, “Limitations on DynamicExcess-Carrier Lifetime Calibration Methods”, Progressin Photovoltaics (accepted March 2011).Y-H. So, M.A. Green, G. Conibeer, S. Huang and A.Gentle, “Size Dependent Optical Properties of SiQuantum Dots in Si-rich Nitride/Si 3N 4SuperlatticeSynthesized by Magnetron Sputtering”, Journal ofApplied Physics (accepted February 2011).Yong-Heng So, Shujuan Huang, Gavin Conibeer andMartin A Green, “Formation and Photoluminescenceof Si Nanocrystals in Controlled Multilayer StructureComprising of Si-rich Nitride and Ultrathin SiliconNitride Barrier Layers”, Thin Solid Films (acceptedFebruary 2011).A. Sugianto, L. Mai, M.B. Edwards, S.R. Wenham etal., “Investigation of Al-Doped Emitter on N-TypeRear Junction Solar Cells”, IEEE Transactions onElectron Devices, Volume: 57, Issue: 2, pp. 525-529,February 2010.Y. Takeda, T. Motohiro, D. König , P. Aliberti, Y. Feng, S.Shrestha and G. Conibeer, “Practical Factors LoweringConversion Efficiency of Hot Carrier Solar Cells,Applied Physics Express”, Applied Physics Express 3,p. 104301, 2010.C.Y. Tsao; J. W. Weber; P. Campbell; G. Conibeer; D.Song and M.A. Green, “In Situ Low TemperatureGrowth of Poly-Crystalline Germanium Thin Filmon Glass by RF Magnetron Sputtering”, Solar EnergyMaterials and Solar Cells, Vol. 94, pp. 1501-1505, 2010.C.Y. Tsao, J. Wong, J-L. Huang; P. Campbell, D. Songand M.A. Green, “Structural Dependence of ElectricalProperties of Ge Films Prepared by RF MagnetronSputtering”, Applied Physics A, Vol. 102, pp. 689-694, 2010.C.Y. Tsao, P. Campbell, D. Song and M.A. Green,“Influence of Hydrogen on Structural and OpticalProperties of Low Temperature Polycrystalline GeFilms Deposited by RF Magnetron Sputtering”,Journal of Crystal Growth, Vol. 312 (9), pp. 2647-2655, 2010.C.Y. Tsao, J.L. Huang, X.J. Hao, P. Campbell andM.A. Green, “Heavily Boron-Doped HydrogenatedPolycrystalline Ge Thin Films Prepared byCosputtering”, Electrochemical and Solid-StateLetters, Vol. 13 (10), pp. H354-H356, 2010.T. Trupke, J, Nyhus and J.Haunschild, “LuminescenceImaging for Inline Characterisation in SiliconPhotovoltaics”, Phys.Stat.Sol. (RRL) (published online2 February 2011).A. Uddin, C. B. Lee and T. G. Anderson, “Initial Rise ofTransient Electroluminescence in Doped Alq 3Films”,Physica Status Solidi A, Vol. 207, No. 10, pp. 2334-2338, 2010A. Uddin and C. B. Lee, “Exciton Behaviour in DopedTris (8-hydroxyquinoline) Aluminum (Alq 3) Films),Physica Status Solidi C, Vol. 8 (1), pp. 80-83, 2011.A. Uddin, C. B. Lee and J. Wong, “Emission Propertiesof Dopants Rubrene and Coumarin 6 in Alq 3Films”,Journal of Luminescence, Vol. 131, pp. 1037-1041, 2011.Z. Wan, S. Huang, M.A. Green and G. Conibeer,“Residual Stress Study of Silicon Quantum Dot inSilicon Carbide Matrix by Raman Measurement”,Physica Status Solidi C, Vol. 8 (1), pp. 185-188, 2011.Z. Wan, S. Huang, M.A. Green and G. Conibeer, “RapidThermal Annealing and Crystallization MechanismsStudy of Silicon Nanocrystal in Silicon CarbideMatrix”, Nanoscale Research Letters, 6:129, 2011.S. Wang, A. Lennon, B. Tjahjono and S. Wenham,“Overcoming Overplating Problems for PECVD SiNxPassivated, Laser-doped p-type MulticrystallineSilicon Solar Cells, Solar Energy Materials and SolarCells (submitted March 2011).K. S. Wang, B. S. Tjahjono, J. Wong, A. Uddin and S.R. Wenham, “Sheet Resistance Charaacterizationof Laser-dope on Crystalline Silicon Wafers forPhotovoltaic Applications”, Solar Energy materialsand Solar Cells, Vol. 95, pp. 974-980, 2011.K. S. Wang, B. S. Tjahjono, A. Uddin and S. R. Wenham,“A Method to Characterize the Sheet Resistance ofa Laser Doped Line on Crystalline Silicon Wafers forPhotovoltaic Applications”, Applied Physics Letters,Vol. 98, pp. 094105 (1-3), 2011.J. Wong, J. L. Huang, S. Varlamov, M. A. Green,R. Evans, M. Keevers and R. Egan, “StructuralInhomogeneities in Polycrystalline Silicon OnGlass Solar Cells and Their Effects on DeviceCharacteristics”, Progress in Photovoltaics: Researchand Applications, DOI: 10. 1002/pip.1089, 2011.J. Wong, J.L. Huang, B. Eggleston, M.A. Green, O. Kunz,R. Evans, M. Keevers and R.J. Egan, “Lifetime LimitingRecombination Pathway in Thin-Film PolycrystallineSilicon on Glass Solar Cells” Journal of AppliedPhysics, Vol. 107, p.123705, 2010.Y. Yao, A. Sugianto, A. Lennon, B. Tjahjono and S.Wenham, “Use of ICP-MS and OES Measurementsto Characterise Light-induced Plating for SiliconSolar Cells”, Solar Energy Materials and Solar Cells(submitted January 2011).G. Zatryb, A. Podhorodecki, X.J. Hao, J. Misiewicz,Y.S. Shen and M.A. Green, “Quantitative Evaluationof Boron-Induced Disorder in Multilayers ContainingSilicon Nanocrystals in an Oxide Matrix Designed forPhotovoltaic Applications”, Optics Express, Vol. 18, No.21, pp. 22004-22009, 2010.B. Zhang, S. Shrestha, M.A. Green and G. Conibeer,“Surface States Induced High p-type Conductivityin Nanostructured Thin Film Composed of Genanocrystals in SiO 2Matrix”, Applied Physics Letters,Vol. 97, p. 132109, 2010.B. Zhang, S.K. Shrestha, M.A. Green and G.J. Conibeer,“Size Controlled Synthesis of Ge Nanocrystals inSiO 2at Temperatures Below 400 degree C UsingMagnetron Sputtering”, Applied Physics Letters, Vol.96, p. 26901, 2010.B. Zhang, S. Shrestha, P. Aliberti, M.A. Green and G.Conibeer, “Characterisation of Size-Controlled andRed Luminescent Ge Nanocrystals in MultilayeredSuperlattice Structure”, Thin Solid Films, Vol. 518, pp.5483-5487, 2010.B. Zhang, Y. Yao, R. Patterson, S. Shrestha, M. A. Greenand G. Conibeer, “Electrical Properties of ConductiveGe Nanocrystal Thin Films Fabricated by LowTemperature in Situ Growth”, Nanotechnology, Vol.22, p. 125204, 2011.8.5 UNREFEREED Papers in Scientific andTechnical JournalsRichard Corkish, “Our Time to Shine: Australia’sBurgeoning Solar Energy Industry”, Ecogeneration,Issue 57, pp. 24, 26, March/April 2010.http://ecogeneration.com.au/news/our_time_to_shine_australias_burgeoning_solar_energy_industry/012453/ (invited).T. Trupke and W. McMillan, “PhotoluminescenceImaging Speeds Solar Cell inspection”, Laser FocusWorld, 12/2010.T. Trupke, “Photolumineszenz Bildgebung für dieSilizium-Photovoltaik”, Photonik 1/2011, p.38, 2011.126

8.6 Conference PAPERS ANDOTHER PRESENTATIONSP. Aliberti, Y. Feng, R. Clady, M.J.Y. Tayebjee, T.W.Schmidt, S.K. Shrestha, M.A. Green, G.J. Conibeer, “OnEfficiency of Hot Carriers Solar Cells with a IndiumNitride Absorber Layer”, 5 th World Photovoltaic SolarEnergy Conference, Oral Presentation, Proceedings ofEuropean Photovoltaic Conference, Valencia, Spain,6-10 September 2010.P. Aliberti, B. Zhang, M.A. Green. G. Conibeer, S.Shrestha, A. Hsieh, R. Li and B. Puthen-Veettil,“Investigation of Single Layers of Silicon QuantumDots in SiO 2Matrix for Energy Selective Contactsin Hot Carriers Solar Cells”, Australian Solar EnergySociety Conference, Solar2010, Canberra, 1-2December 2010.X. Bai, A. Sugianto, B. J. Hallam and S. R. Wenham,“Improved Open Circuit Voltage for CommercialGrade Wafers by High Quality Hydrogen Passivation”,Australian Solar Energy Society (AuSeS) Conference,Solar2010, Canberra, 1-3 December 2010.G. Conibeer, M.A. Green, D. König, I. Perez-Wurfl, S.Huang, X. Hao, D. Di, L. Shi, S. Shrestha, B. Puthen-Veetil, Y.H. So, B. Zhang and Z. Wan, “Silicon QuantumDot Based Solar Cells: Addressing the Issues ofDoping, Voltage and Current Transport”, 5 th WorldPhotovoltaic Solar Energy Conference, Valencia,Spain, 6-10 September 2010.G. Conibeer, “Third Generation Photovoltaics”,Australian Solar Energy Society Conference,Solar2010, Canberra, December 2010.Richard Corkish, Stephen Bremner, Anna Bruce, EvattHawkes, Merlinde Kay, Alison Lennon, Alistair Sproul,Ted Spooner, Santosh Shrestha, Geoff Stapleton,Ashraf Uddin, Muriel Watt and Darcy Wentworth,“Renewable Energy Education at UNSW”, Solar2010, 48th Annual Conference of the AustralianSolar Energy Society (AuSES), Canberra, 1-3December 2010.Richard Corkish, “Renewable energy engineeringeducation at UNSW”, World Green Energy Forum,Gyeongju, Korea 17-18 November 2010 (invited,presentation slides published in proceedings).Richard Corkish, “Photovoltaics technologydevelopment at UNSW”, World Green Energy Forum,Gyeongju, Korea 17-18 November 2010 (invited,presentation slides published in proceedings).Richard Corkish, “Advanced silicon photovoltaicstechnology and education at UNSW, Australia”,International Photovoltaic Solar Energy Conferenceand Exhibition (IPVSEE 2010) 27-29th September2010, Beijing (invited).Richard Corkish, “Australia-based Research forPhotovoltaics Industry Expansion in the Asia-Pacific Region”, Ecogen 2010, 5-8 September 2010,Sydney (invited).Richard Corkish, “Australia-based Education forRenewable Energy Industry Expansion in the Asia-Pacific Region”, Ecogen 2010, 5-8 September 2010,Sydney (invited).Richard Corkish, “Advanced PV Cell ResearchDirections & Education at UNSW”, SOLARCON India2010, 28-30 July 2010, Hyderabad, India (invitedkeynote, slides distributed to participants).Richard Corkish, “Photovoltaics R&D at UNSW”,in KIER’s PV R&D Outcome Exchange 2010, pp.129-151. Daejeon, Korea 28 April 2010 (invited,opening session).Richard Corkish, “Photovoltaics Education andTechnology at UNSW”, 2 nd PV Summit Asia, 13-14 April2010, Beijing. Day 2, Paper #1. http://www.merisisasia.com/pv2010/Speakers.html(invited).Richard Corkish, “Antenna collection of solar energy”,The First International Symposium on Solar EnergyTechnology, Yeungnam University, Korea, 16November 2010 (invited).Richard Corkish, “Renewable Energy in Australia”,Australian Conservation Foundation, Central CoastBranch, Gosford, 18 September 2010 (invited).Richard Corkish, “Photovoltaics Technology andEducation at UNSW”, Cutting Edge Conference, MLC,Burwood, 15 September 2010 (invited).Richard Corkish, “Supporting Activities for RenewableEnergy Growth”, University of Delaware, 23 April2010 (invited).Richard Corkish, “Supporting Activities for RenewableEnergy Expansion in the Asia-Pacific Region”, Indo-Australia Solar Energy Workshop, Amity Institute ofRenewable & Alternative Energy, Amity University-Uttar Pradesh, India, 9-10 February 2010 (invited).T. Curtis, R. Corkish and R.J. Fuller, “Solar ThermalDrying of Canarium Indicum Nuts”, Solar 2010, 48thAnnual Conference of the Australian Solar EnergySociety (AuSES), Canberra, 1-3 December 2010.R. Dumbrell, P. Hamer, A. Lennon and B. Tjahjono,“Emitter etch-back application for selective emittersilicon solar cells”, Proceedings of SOLA 2010(AuSES), Australian Solar Energy Society Conference,Canberra, 1-3 December 2010.H. Cui, M.A. Green, P. Campbell, M. Wolf and G. Jin,“Enhanced Absorption of SPC Poly-Si Thin Filmson Aluminium Induced Textured (AIT) Glass withSubmicron Texture Feature Size”, 25th EuropeanPhotovoltaic Solar Energy Conference and Exhibition(25th EU PVSEC) / 5th World Conference onPhotovoltaic Energy Conversion (WCPEC-5), Session:3BV.3.59, Abstract No. [2008], Valencia, Spain, 6-10September, 2010 (poster presentation).Dawei Di, Ivan Perez-Wurfl, G. Conibeer andM.A. Green, “Fabrication and Characterisation ofSilicon Quantum Dots in SiO 2/Si 3N 4Hybrid Matrix”,Proceedings of SPIE – Optics and Photonics – NextGeneration (Nano) Photonic & Cell Technologies forSolar Energy Conversion, San Diego, August 2010.Hua Fan, Richard Corkish, Mary-Louise McLawsand Jian Gang Zhang, “Solar Electrolysis Systemfor Remote Emergency Production of DisinfectantSolutions”, Solar 2010, 48th Annual Conference of theAustralian Solar Energy Society (AuSES), Canberra,1-3 December 2010.M.A. Green, G. Conibeer, D. König, S. Shrestha, S.Huang, P. Aliberti, L. Treiber, R. Patterson, B. Puthen-Veettil, A. Hsieh, Y. Feng, A. Luque, A. Marti, P.G.Linares, E. Cánovas, E. Antolin, D. Fuertes Marrón, C.Tablero, E.Hernandez, J-F. Guillemoles, L. Huang, A. LeBris, T. Schmidt, R. Clady and M. Tayebjee, “Hot CarrierSolar Cells: Challenges and Recent Progress”, 35 thIEEE Photovoltaic Specialists Conference, Honolulu,Hawaii, 20-25 June, 2010.M. A. Green, G. Conibeer, D. König, S. Shrestha, S.Huang, P. Aliberti, L. Treiber, R. Patterson, B. PuthenVeettil, A. Hsieh, A. Luque, A. Marti, P.G. Linares, E.Cánovas, E. Antolín, D. Fuertes Marrón, C. Tablero, E.Hernández, J-F. Guillemoles , L. Huang, T. Schmidt,R. Clady and M. Tayebjee, “Recent Progress withHot Carrier Solar Cells”, 25th European PhotovoltaicSolar Energy Conference and Exhibition (25th EUPVSEC) / 5th World Conference on PhotovoltaicEnergy Conversion (WCPEC-5), Session: 1AP.1.1,Abstract No. [1191], Valencia, Spain, 6-10 September,2010 (plenary).Martin A. Green, “Price and Supply Constraint onTe and In Photovoltaics”, 35 th IEEE PhotovoltaicSpecialists Conference, Honolulu, Hawaii, 20-25June, 2010.Martin A Green, “The International Marketing Trendof Silicon Solar Cells”, World PV Forum, Chengdu, 28-30 September 2010 (invited).Martin A Green, “Possible Evolutionary Pathsfor Photovoltaics”, Symposium C, E-MRS 2010Fall Meeting, Warsaw, Poland, 13-17 September2010 (invited).Martin A. Green, “The Global Evolution of PVtechnology”, 2 nd Chinese Renewable EnergyConference & Solar Power Exhibition, Wuxi, 17-19September 2010 (invited).Martin A. Green, “Principles and Perspectives ofPhotovoltaics”, Walter Schottky Lecture 2010, RWTHAachen, Germany, 21 October 2010 (invited).Martin A. Green, “Technologies Shaping thePhotovoltaic Industry in the Coming 10 Years”,Lecture, Solland Solar, Aachen, Germany, 22 October2010 (invited).Martin A. Green, “High Efficiency Silicon and ThirdGeneration Solar Cells”, Colorado NanofabricationLaboratory Symposium, Colorado, USA, 24-26October 2010 (plenary).Martin A. Green, “Solar PV – The Real Answer toFuture Energy?”, AIE National Conference, Adelaide,14-16 November 2010 (invited).Martin A. Green, “Photovoltaic Physics: LightEmission, Ultimate Efficiencies, Light Trapping”,Australian Institute of Physics (AIP) 19 th NationalConference, Melbourne, 5-9 December 2010 (invited).Martin A. Green, “Photovoltaics Solar Conversionfor Sustainable Energy Supply”, First InternationalConference on Sustainable Urbanization (ICSU 2010),Hong Kong, 15-17 December 2010 (invited).P. Gress, P. Widenborg, G. Jin and S. Varlamov,“Incorporation of Interconnection Technique intothe Cell Design of Thin-Film Polycrystalline SiliconSolar Cells On Glass”, 25 th EUPVSEC, pp. 3580-3583,Valencia, Spain, 2010.B. Hallam, N. Western, S. Wenham, “Large Increase inBulk Lifetime of Commercial Grade CZ Silicon WafersThrough Hydrogenation”, Australian Solar EnergySociety (AuSeS) Conference, Solar2010, Canberra, 1-3December 2010.B. Hallam, A. Sugianto, B. Tjahjono, N. Kuepper,N. Western, G. Xu, S. Wenham “Deep JunctionFormation for Silicon Solar Cells Through the Useof Laser Doping”, Australian Solar Energy Society(AuSeS) Conference, Solar2010, Canberra, 1-3December 2010.P. Hamer, S. Wenham, Xue Bai, B. Hallam, “Applicationof Boron Depletion Techniques to Low ResistivitySilicon for Photovoltaics”, Australian Solar EnergySociety (AuSeS) Conference, Solar2010, Canberra, 1-3December 2010.A. Hsieh, S. Shrestha, M. A. Green and G. Conibeer,“Investigation of Si/SiO 2Quantum Well Structures forEnergy Selective Contacts in Hot Carrier Solar Cells”,5 th World Photovoltaic Solar Energy Conference,Valencia, Spain, 6-10 September 2010.A. Hsieh, S. Shrestha, M.A. Green and G. Conibeer,“Investigation of Si/SiO 2Quantum Well Structuresfor Selective Energy Contacts in Hot Carrier SolarCells”, Australian Solar Energy Society Conference,Solar2010, Canberra, December 2010.127

J. Huang, J. Wong, M. Keevers and M. Green,“Cell performance limitation investigation ofpolycrystalline silicon thin-film solar cells on glass bytransmission electron microscopy”, 25 th EUPVSEC, pp.3565-3567, Valencia, Spain, 2010.G. Jin, P. Gress, S. Varlamov and P. Widenborg, “RecentProgress in High Rate PECVD SPC Poly-Si Thin FilmSolar Cells at UNSW”, 25 th EUPVSEC, pp. 3577-3579,Valencia, Spain, 2010.D. König, J. Rudd, G. Conibeer and M. Green,“Introduction of Majority Carriers into Si and Ge NanoCrystals Embedded in SiO 2or Si 3N 4, Symposium J,E-MRS Spring Meeting, Strasbourg, 7 - 11 June 2010,Oral Presentation (invited).D. König, “Introduction of Majority Carriers into Siand Ge Nano Crystals by Modulation Doping ofEmbedding Dielectric”, Seminar at the ChemnitzUniversity of Technology, Germany, 4 June2010 (invited).D. König, “New Results on Si Nano Crystals: Interfaces,Doping, Energy Selective Contacts and Hot Carriers”,Seminar at the Institute of Microtechnology(IMTEK), University of Freiburg, Germany, 14 June2010 (invited).Eunjoo Lee, Hyunwoo Lee, Ilhwan Kim, JunyoungChoi, Dongjun Oh, Jimyung Shim, Kyungyeun Cho,Jisun Kim, Hae-Seok Lee, Soohong Lee, B. Hallam andS.R. Wenham, “The Potential Efficiency of Laser DopedSolar Cells Using Photoluminescence Imaging”, 35 thIEEE Photovoltaic Specialists Conference, Honolulu,Hawaii, 20-25 June, 2010.P.H. Lu, Y. Chen and A. Lennon, “Innovative Rear Point-Contact Scheme for Silicon Solar Cells, Proceedings ofSOLA 2010 (AuSES), Australian Solar Energy SocietyConference, Canberra, 1-3 December 2010.W. McMillan, T. Trupke, J.W. Weber, M. Wagner, U.Mareck, Y.C. Chou and J. Wong, “In-line Monitoring ofElectrical Wafer Quality Using Photoluminescence”,25 th European Photovoltaic Solar Energy Conference,Valencia, Spain, September 2010.W. McMillan, T. Trupke, J.W. Weber, S. Ha, B.S. You, W.Y.Park, K.Y. Lee and J. Nyhus, “Slip-line Detection onMono Ingots and Wafers at Production ThroughputUsing Photoluminescence”, 25 th EuropeanPhotovoltaic Solar Energy Conference, Valencia,Spain, September 2010.G.A. Parnis, A.B. Sproul, “Fast Thermal ModellingUsing Micro-Cap”, The Australian Solar EnergySociety (AuSES) Solar 2010 Conference, AustralianSolar Energy Society Conference, Canberra, 1-3December 2010.Rajesh S. Patel, J. Bovatsek, A. Sugianto, B. Hallam,B. Tjahjono, “Laser Doped Selective Emitters:Comparing Mode-locked UV 355nm and CW532nm Laser Process”, 25 th EU PVSEC, Valencia, 6-10September 2010.R. Patterson, Z. Wan, B. Puthen-Veettil, D. Königand G. Conibeer, “Phonon Decay in Naostructure:Computational Study”, Proc. of 35th IEEE PVSC,Honolulu, Hawaii, 20 - 25 June 2010, proc. pp. 1843(art.-no. 5615953).R. Patterson, B. Puthen Veettil, M. Kirkengen, D.Koenig, M.A. Green and G.J. Conibeer, PosterPresentation 1DV.3.77, 5 th World PhotovoltaicSolar Energy Conference, Valencia, Spain, 6-10September 2010.B. Puthen Veettil, R. Patterson, D. König, G. Conibeerand M. A. Green, “Modelling the Effects of Doping inQuantum Dot Structures for Tandem Cells”, PosterPresentation 1DV.3.64, 5th World PVSEC, Valencia,Spain, 6-10 September 2010.B. Puthen-Veettill, D. König, M.A. Green and G.Conibeer, “Analysis of QDs in Energy SelectiveContacts Using a 2-Dimensional Model”, Proc. E-MRSSpring meeting, Symposium B, June 2009, EnergyProcedia, 2(1), 2010. http://www.elsevier.com/wps/find/journaldescription.cws_home/718157/description#descriptionF. Qi, G. Conibeer, S. Shrestha and S. Huang, “TheStudy of Potential Application of Group IV Alloys inPhotovoltaic Field”, Australian Solar Energy SocietyConference, Solar2010, Australian Solar EnergySociety Conference, Canberra, 1-3 December 2010.Y. H. So, I. Perez-Wurfl, L. Shi, S. Huang, G. J. Conibeerand M. A. Green, “Silicon Nitride As Alternative Matrixfor All-Si Tandem Solar Cell”, Proceedings of 25thEuropean Photovoltaic Solar Energy Conference andExhibition, Valencia, Spain, 6-10 September 2010.A. Sugianto, J. Bovatsek, S. Wenham, B. Tjahjono, G.Xu, Y. Yao, B. Hallam, X. Bai, N. Kuepper, C.M. Chongand R. Patel, “18.5% Laser-Doped Solar Cell on CzP-Type Silicon”, 35 th IEEE Photovoltaic SpecialistsConference, Honolulu, Hawaii, 20-25 June, 2010.A. Sugianto, A. Lennon, B. Tjahjono and S. Wenham,“Impact of Laser-Doped Line Conductivity on NiLight-Induced Plating, Proceedings of the 25 thEuropean Photovoltaic Solar Energy Conference,Valencia, Spain, 6-10 September 2010.Y. Tao, S. Varlamov, J. Wong, O. Kunz and R. Egan,“Effects of SPC Temperature on Properties ofEvaporated Poly-Si Thin Films and Solar Cells”, 35 thIEEE PVSC, Hawaii, USA, 2010 (oral).L.S. To, “Social and Institutional Aspects ofPhotovoltaic Projects in Western China”, Climate:Science + Humanities Symposium, The Faculty Club,Harvard University, 4 March 2010.L.S. To, “Solar Photovoltaic Projects in Remote Areasof China”, Asia Pacific Week, Australian NationalUniversity, 11 February 2010.T. Trupke, B. Mitchell, J.W. Weber and J. Nyhus.“Bulk Minority Carrier Lifetime from LuminescenceIntensity Ratios Measured on Silicon Bricks”, 25 thEuropean Photovoltaic Solar Energy Conference,Valencia, Spain, September 2010.T Trupke, “Luminescence Imaging: A PowerfulCharacterization Tool for Photovoltaic Applications”,SPIE, Optics & Photonics, San Diego, August 1-5,2010 (invited).T Trupke, “Luminescence Imaging for InlineApplications in Photovoltaics”, 10th InternationalWorkshop on Beam Injection Assessment ofMicrostructures in Semiconductors, Halle, Germany,July 2010. (invited).A. Uddin and C. C. Teo, “High Efficiency Self-assemblyCdSe/ZnS Quantum dots Light-Emitting Devices inOrganic Matrix”, IEEE International Nano-ElectronicsConference, 3–8 January 2010, City University ofHong Kong.A. Uddin and C.B. Lee, “Exciton Behaviors in DopedTris (8-hydroxyquinoline) Aluminum (Alq 3) Films”, 9 thInternational Conference on Excitonic and PhotonicProcesses in Condensed and Nano Materials,(EXCON’10), 11–16 July 2010, Brisbane, Australia.S. Varlamov, Y. Tao, J. Wong, O. Kunz and R. Egan,“Crystallisation Kinetics of E-beam Evaporated SiFilms and Its Effects on Poly-Si Film and Solar CellProperties”, 25 th EUPVSEC, pp. 2823-2827, Valencia,Spain, 2010 (oral).Z. Wan, S. Huang, M.A. Green and G. Conibeer,“Raman Study of Stress Effect in Silicon Rich Carbide(SiC x) Film by Furnace and Rapid Thermal Annealingfor Photovoltaic Application”, 35 th IEEE PhotovoltaicSpecialists Conference, Honolulu, Hawaii, 20-25June, 2010.J. Wong, J.L. Huang, M. Keevers and M.A. Green, “Voc-Limiting Recombination Mechanisms in Thin FilmSilicon on Glass Solar Cells”, 35 th IEEE PhotovoltaicSpecialists Conference, Honolulu, Hawaii, 20-25June, 2010.D. Wang, B. Tjahjono, A. Uddin and S. Wenham,“Theoretical Analysis of Laser-Doped SemiconductorFingers Silicon Solar Cell Design Using a Shadingand Resistive Power Loss Model” 25 th EuropeanPhotovoltaic Solar Energy Conference, 6-10September 2010, Valencia, Spain.E. Wang, S. Mokkapati, F. J. Beck, Z. Ouyang,S. Varlamov and K. Catchpole “Periodic silvernanoparticles fabricated with soft-stampnanoimprint on silicon solar cells”, 25 th EUPVSEC,pp.77-81, Valencia, Spain, 2010 (oral).Y. Yao, B. Zhang, S. Shrestha, G. Conibeer and M.A.Green, “Photovoltaic Effect in Ge Nanocrystals/csiliconHeterojunctions Devices”, 35 th IEEEPhotovoltaic Specialists Conference, Honolulu,Hawaii, 20-25 June, 2010.Y. Yao, A. Lennon, B. Tjahjono and S. Wenham, “Useof Inductively-Coupled-Plasma Measurements toMonitor Light-Induced Plating of Silicon Solar Cells”,Proceedings of the 25 th European PhotovoltaicSolar Energy Conference, Valencia, Spain, 6-10September 2010.H. Zhang, S. Huang and G. Conibeer, “TandemPhotoelectrochemical Cell for Direct Water Splitting”,Australian Solar Energy Society Conference,Solar2010, Canberra, December 2010.B. Zhang, S. Shrestha, S.J. Huang, P. Aliberti, M.A.Green and G. Conibeer, “Structure Studies ofMultilayered Ge Nanocrystals Embedded in SiO 2Matrix Fabricating Using Magnetron Sputtering”, Proc.E-MRS Spring meeting, Symposium B, June 2009,Energy Procedia, 2(1), 2010. http://www.elsevier.com/wps/find/journaldescription.cws_home/718157/description#description8.7 OTHER REPORTSR. Armstrong, M.A. Green, L. Vant-Hull, E. Palominoand T. Plant, “2010 Solar Program Peer ReviewReport: An Independent Evaluation of ProgramActivities for FY2009 and FY2010”, Solar EnergyTechnologies Program, US Department of Energy,July 2010 (782pp.).Muriel Watt and Richard Corkish, “Assessment of theCurrent Status and Future Development Prospectsof Photovoltaics”, Report to Rio Tinto by New SouthGlobal and IT Power, March 2010.128

This report is available from:ARC Photovoltaics Centre of Excellence ARCPHOTOVOLTAICSUniversity of New South WalesCENTRE OFSydney NSW 2052, AustraliaEXCELLENCEPhone:2010/11ANNUAL REPORT+61 2 9385 4018Fax:+61 2 9662 4240Email:s.wenham@unsw.edu.auWeb:http://www.pv.unsw.edu.auDirector:Stuart WenhamExecutive Research Director:Martin GreenReport Coordinators:Martin Green, Stuart Wenham, Joyce Ho andRobert LargentGraphic Design:Peter Haywood (pfhaywood@gmail.com)Roman Balla (romanballa@bigpond.com)Photos, figures and graphs:Courtesy of Centre Staff and Others

ARCPHOTOVOLTAICSCENTRE OFEXCELLENCE2010/11The University of New South WalesANNUAL UNSW REPORT Sydney NSW 2052 AustraliaTel +61 2 9385 4018Fax +61 2 9662 4240Web www.pv.unsw.edu.auARC PHOTOVOLTAICSCENTRE OF EXCELLENCE