Nuno Miguel Fernandes Reis Novel Oscillatory Flow Reactors for ...

Escola de Engenharia 

Nuno Miguel Fernandes Reis 

Novel Oscillatory Flow Reactors for 

Biotechnological Applications 

Tese de Doutoramento 

Doutoramento em Engenharia Química e Biológica 

Trabalho efectuado sob a orientação dos 

Doutor António A. Vicente 

Professor Doutor José A. Teixeira 

Abril de 2006

Autor Nuno Miguel Fernandes Reis 

e-mail nunoreis@deb.uminho.pt 

Telf. +351 253604400 

BI 11382186 

Título da tese 

Novel oscillatory flow reactors for biotechnological applications 

Orientadores 

Doutor António A. Vicente 

Professor Doutor José A. Teixeira 

Ano de conclusão 2006 

Doutoramento em Engenharia Química e Biológica 

É AUTORIZADA A REPRODUÇÃO INTEGRAL DESTA TESE/TRABALHO APENAS 

PARA EFEITOS DE INVESTIGAÇÃO, MEDIANTE DECLARAÇÃO ESCRITA DO 

INTERESSADO, QUE A TAL SE COMPROMETE. 

Universidade do Minho, 10 de Abril de 2006 

ii 

PhD dissertation Novel Oscillatory Flow Reactors for Biotechnological Applications

Acknowledgements 

I would like to acknowledge the important role that my supervisor, Dr. António 

Vicente, has played in this thesis, somehow operating as my mentor and my 

godfather! I also would like to thank to my co-supervisor, Prof. José Teixeira, 

for his precise guidelines and support on publications. 

My thankfulness to Polymer Fluids Group, at the University of Cambridge, UK, 

for hosting me within the group during two research training periods at 

Cambridge, in particular to Dr. Adam P. Harvery (now at the University of 

Newcastle upon Tyne), to Mingzhi Zheng, and for last (but not the least) to 

Prof. Malcom R. Mackley, the group leader, for his supreme guidelines, for the 

pleasant collaboration kept throughout these years, and for the full support of 

my research. 

Many thanks to Cassilda, for being such an outstanding wife, my orientation 

on earth and my breath in the time I got down; I really fill this thesis also 

belongs to her! To all my family for such consideration of my work and for 

bringing a smile when it could not exist… I am also grateful to all my friends 

for the good leisure times we shared together throughout these years, in 

particular to Diana and Eduarda for seeding together the scientific research 

when we were just undergraduate students. 

After all, I am very grateful to my mother, whom too early has left me, for her 

fully support, love and unique energy, which I saw irreversibly fading 

throughout the first two and half years of this thesis… 

Thanks are also due to Fundação para a Ciência e a Tecnologia (FCT) for 

financial support by means of scholarship SFRH/BD/6954/2001. 

PhD dissertation Novel Oscillatory Flow Reactors for Biotechnological Applications 

iii

In memory of my mother, Georgina. 

“…acredito que nada do que é importante se perde verdadeiramente. Apenas 

nos iludimos, julgando ser donos das coisas, dos instantes e dos outros. 

Comigo caminham todos os mortos que amei, todos os amigos que se 

afastaram, todos os dias felizes que se apagaram. Não perdi nada, apenas a 

ilusão de que tudo podia ser meu para sempre.” 

Miguel Sousa Tavares 

iv 


Summary 

This thesis explores the biotechnological applications of two novel scale-down 

oscillatory flow reactors (OFRs). A micro-bioreactor (working mostly in batch) and a 

continuous meso-reactor systems were developed based on a 4.4 mm internal 

diameter tube with smooth periodic constrictions (SPC), both operating under 

oscillatory flow mixing (OFM). 

The first part is dedicated to the flow characterisation in the novel SPC geometry. Flow 

patterns within SPC geometry were experimentally studied using Particle Image 

Velocimetry (PIV) technique at different combinations of fluid oscillation frequency (x0) and amplitude centre-to-peak (x0), and afterwards used for validation of numerical 

simulations via Computational Fluid Dynamics (CFDs). CFD simulations were run with 

2-D axisymmetric and 3-D laminar models as wells as using a turbulent Large Eddy 

Simulation (LES) model using Fluent (New York, USA) software. 

Mixing times of the micro-bioreactor were determined for batch operation at f and x 0 of 

0 to 20 Hz and 0-3 mm, respectively, and correlated using a newly defined mixing 

coefficient (km). The control of fluid dispersion in the novel SPC geometry was studied for continuous 

operation of both the micro-bioreactor and the meso-reactor at different combinations 

of f, x0 and fluid net flow rates (v). Macroscopic flow patterns were studied through the 

residence time distribution (RTD) and the non-ideal tracer response was modelled by 

four single-phase flow models, allowing the prediction of conversion ( X ) in the novel 

SPC tube geometry. Further RTD experiments were performed in the presence of a 

steady, continuous flow rate (at various values of v) and their results were compared 

with those obtained from CFDs simulations. 

Flow patterns within this novel SPC geometry were found to be very dependent of both 

x 0 and f. In particular, k m, RTD and X have demonstrated to be manipulated by the 

OFM conditions, as a result of a controlled fluid convection and dispersion within the 

SPC tube through vortex rings detachment. It is possible to drive the macroscopic flow 

patterns within both the micro-bioreactor and the meso-reactor towards the ideal flow 

cases of plug flow reactor (PFR) or completely back-mixed reactor (or a continuous 

stirred tank reactor, CSTR), being the convection maximized in relation to fluid 

dispersion mainly at smooth OFM conditions (i.e. x0 ≤ 1 mm and f ≤ 10 Hz). A 2-D 

axisymmetric laminar model was found to match the flow patterns at small values of f 

and x0 (where flow has demonstrated to match the PFR) while a 3-D laminar model 

was required to simulate non-axisymmetric flow patterns (as those found in a CSTR). 

The 3-D laminar model was highly grid-dependent, but numerical simulations with 3-D 

LES were found to overcome such grid dependency. 


v

Amongst the four single-phase models used in the modelling of macroscopic flow 

patterns by means of the analysis of RTD results, the tanks-in-series model with 

backflow is highly recommended due to the physical analogy with the SPC geometry 

(several interconnected stages – the cavities) and for considering the existence of a 

backflow rate, G, between the cavities. 

The second part of this work is focused on exploring both the micro-bioreactor and the 

meso-reactor in three main biotechnological applications: i) aerobic and anaerobic 

growth of Saccharomyces cerevisiae in the micro-bioreactor; ii) biotechnological 

production/screening of γ-decalactone in the micro-bioreactor; iii) dilution refolding of 

lysozyme for batch (micro-bioreactor) and continuous (meso-reactor) operation. 

Beforehand, mass transfer within the micro-bioreactor was studied by assessing the 

oxygen mass transfer rates in a gas-liquid system. The effect of f and x0 on the oxygen 

mass transfer coefficient (kLa) and on the gas hold-up (ε) were studied at a fixed gas 

flow rate v gas of 0.28 mL/min. An empirical correlation was developed for k La and 

related with the flow patterns observed by PIV and numerically simulated with CFDs. 

Gas-liquid mass transfer in the micro-bioreactor was shown to be enhanced in relation 

to other scale-down systems, as values of kLa up to 0.05 s-1 were obtained through 

OFM (f = 0 - 20 Hz and x0 = 0 - 3 mm) at a small value of vgas = 0.28 mL/min. Such 

improved oxygen mass transfer was suggested to be responsible for an 83 % 

improvement of yield of biomass growth on glucose (YX/S), obtained in the aerobic 

growth of S. cerivisiae in comparison with the value of YX/S obtained for a stirred tank 

reactor (STR). Also the 50 % reduction of the time needed for maximum γ-decalactone 

production with the strictly aerobic yeast Yarrowia lipolytica suggested improved mass 

transfer rates in the four-phase system as result of an improved contact between the 

different phases. 

It has been shown that the reciprocating nature of OFM (backflow) enhances the 

interaction between fluid elements. This lead to the conclusion that both the microbioreactor 

and the meso-reactor present design limitations for lysozyme dilution 

refolding, mainly when applied to continuous refolding (with the meso-reactor). In fact, 

an intensive protein aggregation was observed, leading to the suggestion that the 

meso-reactor could be used as a scale-down system for production of bio-aggregates 

and nano-particles. In summary, the two novel scale-down platforms are ready to 

contribute to accelerate the bioprocess design, by allowing the running of highthroughput 

screening experiments at reproduced and well-controlled conditions. 

vi 


Resumo 

Este trabalho explora as aplicações biotecnológicas de dois novos reactores de fluxo 

oscilatório (RFO) de pequena-escala: um micro-reactor e um meso-reactor contínuo, 

compostos por um tubo (diâmetro interno = 4.4 mm) com constrições suaves na 

parede (CSP), sujeitos a mistura por fluxo oscilatório (MFO). 

A primeira parte visa a caracterização do fluxo na nova geometria CSP. Analisaram-se 

os padrões de fluxo recorrendo à Velocimetria por Imagem de Partícula (VIP), para 

diversos valores de frequência (f) e amplitude centro-ao-pico (x0) de oscilação, os quais 

foram posteriormente utilizados na validação de simulações numéricas por Dinâmica 

de Fluidos Computacional (DFC). As simulações foram realizadas com o software 

Fluent (Nova Iorque, EUA), com base em modelos do tipo 2-D com simetria axial, 3-D 

laminar e 3-D com Simulação directa de Grandes Vórtices (SGV). 

Determinaram-se tempos de mistura para um funcionamento do micro-bioreactor por 

partidas, a diversas combinações de f e x0 (0 - 20 Hz e 0 – 3 mm, respectivamente), 

tendo-se encontrado uma correlação empírica para os tempos de mistura com base 

num novo parâmetro: o coeficiente de mistura (km). O controlo da dispersão nos dois novos reactores foi analisado para várias 

combinações de f, x0 e caudais de líquido (v), tendo-se analisado os padrões de fluxo 

macroscópicos através da distribuição dos tempos de residência (DTR). A resposta 

não-ideal do traçador foi modelada por quatro modelos hidrodinâmicos e permitiu 

prever a conversão ( X ) na nova geometria SCP. Realizaram-se experiências 

complementares para a situação de um caudal contínuo e estacionário (a diversos 

valores de v), tendo-se comparado os resultados com os previstos pelas simulações 

por DFC. 

Concluiu-se que os padrões de fluxo na nova geometria SCP são bastante 

dependentes quer de f quer de x0. Os parâmetros km, kLa, ε, RTD e X são 

manipulados pelas condições de MFO graças a um controlo efectivo sobre a 

convecção e dispersão do fluído no interior da geometria CSP por geração de anéis de 

vórtices. Em concreto, os padrões de fluxo macroscópicos no interior do microbioreactor 

e do meso-reactor podem ser aproximados aos casos ideais de um reactor 

de fluxo pistão (RFP) ou de reactor perfeitamente agitado (RPA), sendo que a 

convecção é maximizada (relativamente à dispersão do fluido) essencialmente a 

baixos valores de MFO (p. ex., x0 ≤ 1 mm e f ≤ 10 Hz). O modelo 2-D laminar com 

simetria axial é capaz de prever os padrões de fluxo a baixos valores de f e x0 (p. ex., 

RFP), mas um modelo 3-D laminar é indispensável para prever a assimetria axial do 

fluxo (situação de um RPA). 

Tese de Doutoramento Novos Reactores Oscilatórios para Aplicações Biotecnológicas 

vii

As simulações numéricas efectuadas com o modelo 3-D laminar apresentaram-se 

bastante dependentes do espaçamento da grelha, enquanto que o modelo 3-D com 

SGV permitiu ultrapassar tal dependência. 

Entre os quatro modelos hidrodinâmicos utilizados para a modelação dos padrões de 

fluxo por análise de DTR, o modelo ‘tanques em série com retro-fluxo’ é altamente 

recomendado visto existir uma analogia física com a geometria CSP (diversas 

unidades perfeitamente agitadas e interligadas - as cavidades) e por considerar a 

existência de uma taxa de retrodispersão (G) entre as várias cavidades. 

A segunda parte do trabalho focou a aplicação do micro-bioreactor e do meso-reactor 

a três bioprocessos: i) crescimento aeróbio e anaeróbio da Saccharomyces cerevisiae 

no micro-bioreactor; ii) optimização da produção biotecnológica da γ-decalactona no 

micro-bioreactor; iii) renaturação da lisozima por diluição por partidas (no microbioreactor) 

ou em contínuo (no meso-reactor). Primeiramente, estudou-se a 

transferência de massa no micro-bioreactor por medição das taxas de transferência 

de oxigénio (kLa) num sistema gás-líquido. Averiguou-se o efeito de f e x0 sobre kLa e a 

fracção de gás (ε) para a um valor fixo de caudal volumétrico de gás vgas = 0.28 

mL/min, o que permitiu encontrar uma correlação empírica para o kLa e relacionar kLa com os padrões de fluxo quer experimentalmente observados (por VIP), quer 

numericamente simulados (por DFC). 

Os estudos de kLa demonstraram que a transferência de massa de um sistema gáslíquido 

no micro-bioreactor é superior à obtida em outros sistemas de pequena-escala: 

kLa = 0.05 s-1 (para f = 0 - 20 Hz e x0 = 0 – 3 mm) a vgas = 0.28 mL/min. O aumento 

de kLa foi apontado como responsável pelo aumento em 83 % do rendimento de 

crescimento de biomassa em glucose (YX/S ) para a situação de crescimento aeróbio 

de S. cerevisiae, em comparação com os valores de YX/S obtidos num RPA. De igual 

modo, a diminuição em 50 % do tempo necessário para máxima produção de γdecalactona 

com a levedura aeróbica restrita Yarrowia lipolytica sugere taxas de 

transferência de massa incrementadas neste sistema de quatro-fases, graças a um 

aumento da área de contacto entre as diversas fases. 

A natureza recíproca na MFO aumenta a interacção entre os elementos do fluido. Por 

isso, quer o micro-bioreactor quer o meso-reactor apresentam limitações na 

renaturação de lisozima por diluição, essencialmente quando em contínuo (com o 

meso-reactor). A intensa agregação proteica observada sugere que o meso-reactor 

poderá ser eficazmente utilizado como um sistema de pequena-escala para a 

produção contínua de bio-agregados e nano-partículas. Em suma, os dois novos 

sistemas de pequena-escala contribuirão, certamente, para acelerar o processo de 

projecto de bioprocessos, permitindo realizar experiências como elevada 

selectividade, reprodutibilidade e condições bem controladas. 

viii 

Tese de Doutoramento Novos Reactores Oscilatórios para Aplicações Biotecnológicas

Table of contents 

Acknowledgements.......................................................................................iii 

Summary ......................................................................................................v 

Resumo ...................................................................................................... vii 

Table of contents.......................................................................................... ix 

List of publications ..................................................................................... xiii 

List of abbreviations ....................................................................................xiv 

List of figures .............................................................................................. xv 

List of general nomenclature..................................................................... xxvii 

List of tables.............................................................................................. xxix 

Chapter 1 Introduction................................................................................ 1 

Chapter 2 Literature review......................................................................... 5 

2.1 Types and applications of oscillating devices............................... 6 

2.1.1 Types of oscillating devices .................................................... 6 

2.1.2 Industrial applications of oscillating reactors........................... 8 

2.2 The Oscillatory Flow Reactor (OFR) ........................................... 11 

2.3 The Oscillatory Flow Mixing (OFM) ............................................ 18 

2.3.1 Parameters governing the OFM............................................ 19 

2.3.2 The effect of geometrical parameters ................................... 24 

2.3.3 Effect of f and x 0 in the flow patterns..................................... 27 

2.3.4 Power input......................................................................... 27 

2.3.5 Numerical simulation........................................................... 28 

2.4 Further studies regarding oscillatory flow mixing ....................... 30 

2.5 Tools in reactor engineering ..................................................... 31 

2.5.1 Measuring techniques.......................................................... 31 

2.5.2 Flow visualisation by Particle Image Velocimetry................... 35 

2.5.3 Assessment of the non-ideal flow ......................................... 37 

2.5.4 Computational flow modelling .............................................. 40 

2.6 Biotechnological process engineering ....................................... 42 

2.6.1 Application areas................................................................. 42 


ix

x 

2.6.2 Bioreactors and bioprocesses...............................................43 

2.6.3 Bioreactor engineering .........................................................46 

2.6.4 Bioprocesses monitoring ......................................................48 

2.6.5 Continuous cultures .............................................................50 

2.6.6 Biotechnological applications of OFM....................................51 

2.7 Scale-down of bioprocesses ......................................................52 

2.8 Conclusions..............................................................................55 

2.9 References ...............................................................................56 

Chapter 3 The novel oscillatory flow reactor designs ..................................75 

3.1 The novel SPC tube geometry ...................................................76 

3.2 The novel micro-bioreactor........................................................76 

3.3 The novel continuous oscillatory flow meso-reactor....................77 

Chapter 4 Fluid mechanics and catalyst particle suspension within the novel 

micro-bioreactor .........................................................................................79 

4.1 Introduction..............................................................................80 

4.2 Materials and methods .............................................................81 

4.3 Results and analyses ................................................................89 

4.4 Discussion and conclusions ....................................................107 

4.5 Nomenclature.........................................................................109 

4.6 References .............................................................................110 

Chapter 5 Mixing times and residence time distribution of liquid phase within 

the SPC geometry.....................................................................................113 

5.1 Introduction............................................................................115 

5.2 Experimental ..........................................................................117 

5.2.1 RTD of liquid phase in the novel micro-bioreactor................117 

5.2.2 RTD of liquid phase in the novel meso-reactor ....................124 

5.2.3 Mixing times for batch operation of the novel micro-bioreactor 

126 

5.2.4 Numerical simulations of RTD for steady flow in the SPC tube 

geometry.........................................................................................127 

5.3 Results and discussion ...........................................................127 

5.3.1 Analysis of RTD of liquid phase in the micro-bioreactor .......127 

5.3.2 Analysis of RTD of liquid phase in the meso-reactor ............141 


5.3.3 Determination of mixing times fro batch operation of micro- 

bioreactor at different combinations of f and x 0................................. 145 

5.3.4 Matching of numerical simulations of RTD in the SPC tube for 

steady flow...................................................................................... 147 

5.4 Conclusions ........................................................................... 149 

5.5 Notation................................................................................. 150 

5.6 References............................................................................. 152 

Chapter 6 Correlating the macroscopic fluid mixing and axial dispersion with 

the fluid mechanics of the micro-bioreactor............................................... 155 

6.1 Introduction ........................................................................... 156 

6.2 Materials and Methods........................................................... 157 

6.3 Results and discussion........................................................... 160 

6.4 Conclusions ........................................................................... 175 

6.5 References............................................................................. 176 

Chapter 7 Oxygen mass transfer rates for gas-liquid flow in the micro- 

bioreactor................................................................................................. 179 

7.1 Introduction ........................................................................... 180 



7.4 Conclusions ........................................................................... 203 

7.5 Notation................................................................................. 204 

7.6 References............................................................................. 206 

Chapter 8 Aerobic and anaerobic growth on glucose of Saccharomyces 

cerevisiae in the micro-bioreactor.............................................................. 209 

8.1 Introduction ........................................................................... 210 

8.2 Materials and methods........................................................... 212 


8.4 Conclusions ........................................................................... 227 

8.5 References............................................................................. 228 

Chapter 9 Biotechnological production of γ-decalactone by the strict aerobic 

yeast Yarrowia lipolytica in the micro-bioreactor......................................... 233 

9.1 Introduction ........................................................................... 234 



xi

xii 


9.4 Conclusions............................................................................241 

9.5 References .............................................................................241 

Chapter 10 Lysozyme Dilution Refolding..................................................245 

10.1 Introduction............................................................................246 

10.2 Materials and methods ...........................................................249 


10.4 Conclusions............................................................................258 

10.5 References .............................................................................258 

Chapter 11 General conclusions and suggestions for future work.............261 


List of publications 

Reis N, Vicente AA, Teixeira JA, Mackley MR. 2004. Residence times and 

mixing of a novel continuous oscillatory flow screening reactor. Chemical 

Engineering Science 59(22-23):4967-4974. 

Reis N, Harvey AP, Vicente AA, Teixeira JA, Mackley MR. 2005. Fluid 

Mechanics and Design Aspects of a Novel Oscillatory Flow Meso-Reactor. 

Chemical Engineering Research & Design 83(A4):357-371. 

Reis N, Vicente AA, Teixeira JA. forthcoming. The Control of Liquid Axial 

Dispersion in a Small-Scale Tube through Oscillatory Flow Mixing. Aiche 

Journal. 

Reis N, Mackley MR, Harvey AP, Vicente AA, Teixeira JA. in progress. The 

correlation between the macroscopic flow patterns and the deviation from 

ideal flow for a Smooth, Periodically Constricted Tube. 

Reis N, Vicente AA, Teixeira JA. forthcoming. Enhanced mass transfer rates in 

a novel oscillatory flow screening reactor. Chemical Engineering Science. 

Reis N, Gonçalves CN, Teixeira JA, Vicente AA. forthcoming. Proof-of-concept 

of a Novel Micro-bioreactor for Fast Development of Industrial Bioprocesses. 

Bioengineering & Biotechnology. 

Reis N, Gonçalves CN, Águedo M, Gomes N, Teixeira JA, Vicente AA. 

forthcoming. Application of a novel oscillatory flow micro-bioreactor to the 

production of γ-decalactone in a two immiscible liquid phase medium. 

Biotechnology Letters. 


xiii

List of abbreviations 

OFM Oscillatory Flow Mixing 

OFR Oscillatory Flow reactor 

POF Pure Oscillatory Flow 

RTD Residence Time Distribution 

PIV Particle Image Velocimetry 

CFD Computational Fluid Dynamics 

HTP High throughput 

xiv 


List of figures 

Figure 1-1. Multidisciplinary nature of biotechnology (Moo-Young et al. 1985).1 

Figure 2-1. Technical drawing of a Van Dijck’s US Patent (1935). .................. 6 

Figure 2-2. Examples of oscillating vessels: reciprocating plates - (A) and (B) – 

and oscillating piston – (C) and (D). (A) from Prochazka and Rod (1974). (B) 

from Ni (2002)(2002). (C) from Prochazka and Rod (1974), (D) from 

Hounslow and Ni (2004). See references for numbering details..................... 7 

Figure 2-3. Number of publications out coming from a global search in ISI Web 

of Knowledge (http://portal.isiknowledge.com/portal.cgi) using keywords 

“oscillatory flow”. All citation databases, document types and languages were 

considered in the search. ............................................................................. 8 

Figure 2-4. Schematic representation of cross section in an OFR. d i – reactor 

internal diameter, L – baffles spacing, d 0 – orifice diameter, δ - baffle 

thickness.................................................................................................... 12 

Figure 2-5. Mechanism of oscillatory flow mixing (OFM) in an OFR, according 

to Fitch et al. (2005). (A) Start of Up Stroke. (B) Maximum velocity in Up 

stroke, i.e. flow reversal. (C) Start of Down stroke. (D) Maximum velocity in 

Down stroke............................................................................................... 18 

Figure 2-6. The net flow in a plain tube. ...................................................... 19 

Figure 2-7. Oscillatory motion superimposed onto a net flow........................ 21 

Figure 2-8. The oscillatory (baffled) flow. ..................................................... 22 

Figure 2-9. Exemplification of sinusoidal movement of a piston (displacement, 

x, velocity, v, and acceleration, a) for w = 0.62 rad/s (i.e., 0.1 Hz), and x 0 = 5 

mm............................................................................................................ 23 

Figure 2-10. Particle flow pattern in a batch OFR. Tracer = pollen particles of 

25 µm in diameter, bulk fluid = water, f = 2.5 Hz, x 0 = 6mm, d = 50 mm, L = 

1.5d, α = 36 %, δ = 3 mm (from Ni et al. 2002a)....................................... 29 

Figure 2-11. Overview of PIV technique. (A) Schematic representation of the 

flow field illumination in a PIV system. (B) PIV interrogation analysis. (C) 


xv

Evaluation of the image density. Only build up of 2-D velocity vector maps is 

exemplified (adapted from dantecdynamics 2002). .....................................36 

Figure 2-12. Factors that influence the performance of a bioprocess and the 

complexity of interactions between them. Only some interactions are shown 

for illustrative purposes. The factors are grouped under three system 

properties, namely, physical, chemical and biological (adapted from 

Vaidyanathan et al (1999))..........................................................................44 

Figure 2-13. A schematic of the approaches to measurement in bioprocesses 

(adapted from Vaidyanathan et al., (1999). .................................................49 

Figure 2-14. Main stages crossing the bioprocess development....................53 

Figure 2-15. Examples of commercially available HTP screening bioreactor 

systems. (A) Infors Profors – 16 x 400mL, sparged column reactors. (B) 

DasGIP Fedbatch-pro – 16 x 300mL stirred tank reactors. (C) Infors Sixfors – 

6 x 500mL, stirred tank reactors.................................................................54 

Figure 3-1. Novel SPC tube geometry. All dimensions are in mm. ................76 

Figure 3-2. Geometry of the SPC tube composing the novel micro-bioreactor. 

..................................................................................................................77 

Figure 3-3. Simplified scheme of the novel continuous oscillatory flow meso- 

reactor. ......................................................................................................78 

Figure 4-1. Experimental setup used in experimental PIV. A. Laser source. B. 

Laser sheet. C. Optical box made of Perspex. D. CCD camera. E. captured 

image-pair. F. SPC tube. G. Oscillation unit. The optical box (C) and the jacket 

of SPC tube (F) were filled with glycerol to avoid optical distortions. .............82 

Figure 4-2. Mesh for 3-D numerical simulations (units are radii of the tube, R). 

A detail of mesh in zones a, b and c may be found in Table 4-2...................84 

Figure 4-3. Instantaneous velocity vector maps at Re o = 348, x 0 = 1.1 mm, f = 

11.1 Hz coloured by absolute velocity magnitude (mm/s) and different phase 

angles (black vortex rings and arrows added to aid visualization):.................91 

Figure 4-4. Instantaneous velocity vector maps at Re o = 1,350, x 0 = 4 mm, f = 

12.1 Hz and different phase angles (blue vortex rings and arrows added to aid 

visualization):..............................................................................................92 

xvi 


Figure 4-5. Influence of the grid size on the recirculation strength of steady 

state solution, w max - w wall, for continuous net flow, based on 2-D planar model 

results (Re = 100). ..................................................................................... 94 

Figure 4-6. Simulated flow patterns for Re o = 11, x 0 = 0.2 mm, f = 2 Hz, no net 

flow, using a 2-D axisymmetric laminar model, after 2 simulation cycles. 

Contours of stream functions (kg s -1 ) at: ...................................................... 95 

Figure 4-7. Simulated flow patterns for Re o = 348, x 0 = 1.1 mm, f = 11.1 Hz, 

using a 2-D axisymmetric laminar model, after 12 simulation cycles, no net 

flow. Velocity vectors coloured by velocity magnitude (m/s) at:.................... 96 

Figure 4-8. Simulated flow patterns for Re o = 348, x 0 = 1.1 mm, f = 11.1 Hz, 

using a 3-D laminar model, after 26 simulation cycles. Velocity vectors 

coloured by velocity magnitude (m/s), on plane z = 0, (black arrows added to 

aid visualization) at:.................................................................................... 97 

Figure 4-9. Comparison of the total areas occupied by the vortices () in 

different instants of the oscillation cycle (Re o = 348, x 0 = 1.1 mm, f = 11.1 Hz; 

no net flow, i.e. Re n = 0) for: ....................................................................... 98 

Figure 4-10. Average of axial velocities through the oscillation cycle at Re o = 

348, f = 12.1 Hz, x 0 = 1.2 mm, using a) experimental data from PIV, b) data 

from numerical modelling using a 2-D laminar axisymmetric model and c) data 

from numerical modelling using a 3-D laminar model. () global averaged 

axial velocity; () average of positive values of axial velocity; () average of 

negative values of axial velocities. ............................................................... 99 

Figure 4-11. Average of radial velocities through the oscillation cycle using a) 

experimental data from PIV, b) data from numerical modelling using a 2-D 

laminar axisymmetric model and c) data from numerical modelling using a 3-D 

laminar model using cells at plane z = 0. () global averaged radial velocity; 

() average of positive values of radial velocity; () average of negative 

values of radial velocities. Connection lines just intend to represent a 

tendency. Re o = 348; f = 12.1 Hz, x 0 = 1.2 mm......................................... 100 

Figure 4-12. a) maximum concentration of ion exchange resin particles 

completely suspended at different fluid oscillations frequencies and 

amplitudes for a vertically fixed SPC tube; b) minimum u(t) max (maximum 


xvii

oscillation velocities) for complete suspension of particles, at different fluid 

oscillation frequencies. .............................................................................103 

Figure 4-13. Instantaneous velocity vector maps of fluid phase at Re o = 990, x 0 

= 3 mm, f = 12.1 Hz and 45º of tube position in the presence of a small 

amount of ion exchange resin particles:.....................................................105 

Figure 4-14 Complete suspension of 40 % v/v of ion exchange resin particles 

at varying angles and similar oscillation conditions: (a) vertical position, f = 

12.1 Hz, x 0 = 4 mm; (b) 45º, f = 12.1 Hz, x 0 = 4 mm; (c) 10º, f = 12.1 Hz, x 0 = 

3 mm; (d) horizontal position, f = 12.1 Hz, x 0 = 3 mm. In b), c) and d), the 

right hand side corresponds to the bottom of the tube. ..............................106 

Figure 4-15. Proposed “in series” configuration for a single screening reactor 

unit. .........................................................................................................108 

Figure 5-1. Experimental setup. A: Peristaltic pump; B: Reservoir; C: Electric 

motor; D: Piston pump; E: 350-mm-long SPC tube; F: Micro transmission dip 

optical probe; G: Reflection optical probe; H: Aluminium foil; I: In-line cell; J: 

Tungsten halogen light source; K: 475 nm LED light source; L: Multi-channel 

fibre optic spectrometer; M: Personal computer; N: Tracer injection; O: Optical 

path of reflection probe; P: Optical path of transmission probe (2 mm); Q: 

section of dye injection; R: detail of SPC geometry (all units are in millimetres); 

S: inlet tube; T: outlet tube........................................................................118 

Figure 5-2. Relation between absorbance (A = log (P 0/P)) measured by optical 

micro-probes and the tracer concentration (x)............................................120 

Figure 5-3. (a) Response of micro-probes during the consecutive phases of a 

complete RTD experiment; (b) comparison of generated Laplace step-down 

function (‘generated x in ’) found by deconvolution of x measured by micro- 

probe 1 (located downstream the injection point) with a perfect Laplace input 

step function (‘perfect x in ’). Example is given for an experiment at steady, 

continuous flow (Re o = 0 and v = 1.94 ml/min). I: system cleaning; II: feeding 

of the system with the tracer; III: stabilisation of concentration in the system 

through the recirculation and oscillatory mixing; IV: RTD experiment running. ○ 

micro-probe 1 response, micro-probe 2 response. Line in (b) represents: 

‘generated x in ’ = ‘perfect x in ’. .................................................................121 

xviii 


Figure 5-4. Schematic of the SPC configuration for RTD experiments, as seen 

in Laplace’s domain. ................................................................................ 123 

Figure 5-5. Experimental apparatus used for RTD studies in the novel 

continuous meso-reactor. ......................................................................... 125 

Figure 5-6. Light absorbance (A = log (P 0/P)) measured by micro-probes 3, 4 

and 5 at dye tracer concentration (x) of 0 to 10 mg/L. .............................. 126 

Figure 5-7. (a) Reduced RTD curves for different superficial liquid tube 

velocities (u Ls) (in cm/s) and comparison with the pure-convective flow 

(Danckwerts 1953); (b) mean tracer residence time ( t ) as a function of inlet 

liquid flow rate (v)..................................................................................... 128 

Figure 5-8. Tracer response curves at the outlet of a 350-mm-long SPC tube at 

f = 20 Hz and v = 1.94 ml/min. (a) Repeatability of two different experiments 

(x 0 = 1 mm); (b) Experimental data for x 0 of 0, 0.5, 1.0, 2.0 and 3.0 mm... 129 

Figure 5-9. Average mean residence times of the tracer in a 350-mm-long SPC 

tube as a function of fluid oscillation frequency (f) and amplitude(x 0) for a net 

flow rate (v) of 1.94 ml/min. .................................................................... 131 

Figure 5-10. Details of best-fitting of cumulative dimensionless concentration of 

tracer (Fθ-diagram) and transfer function g(T) to single-flow models. (a) and (b) 

shows parameters N tis and D P estimated through direct comparison of Fθ- 

diagrams (Levenspiel 1972) and using best-fitting criteria of Equation (5-5); (c) 

and (d) shows model parameters estimated by best-fitting (with Equation (5-6) 

of transfer function g out(T) to transfer function g(T) derived by mass balance of 

single-phase models. Fluid oscillated at: (a) and (c) 3 Hz and 0.3 mm; (b) and 

(d) 20 Hz and 3 mm. Net flow rate of 1.94 ml/min. Note that some curves are 

coincident. The fitting range refers to the integration intervals in Equation (5- 

6)............................................................................................................. 133 

Figure 5-11. Effect of fluid oscillation frequency (f) on the dimensionless 

number I PD for constant fluid oscillation amplitudes (x 0), using values of D P 

estimated by different methods. ■ 0 mm, □ 0.5 mm, ▲ 1.0 mm, ♦ 2.0 mm, 

○ 3.0 mm; (a) I PD found by the moments method; (b) I PD found by fitting of 

g out(T) to g(T); (c) I PD found by direct nonlinear regression of the analytical 

equation of axial dispersion model presented by Levenspiel (1972) Vertical 


xix

error bars represent spread (standard deviation) of values for different 

experiments. Net flow rate of 1.94 ml/min. Area not shaded corresponds to 

the region where an improvement of RTD is achieved, i.e. where dispersion 

becomes less significant than convection. .................................................134 

Figure 5-12. Cross-correlation of dimensionless axial dispersion number (P D) 

with the backmixing (G), the number of tanks-in-series (N tis) and the volume- 

fraction of ideal PFR (V p/V), according the values of parameters estimated 

through different methods. (a) Linear plot of G vs. 1/P D; (b) linear plot N tis vs. 

P D; (c) log-plot of V P/V vs. P D; model parameters estimated through fitting of 

moments, fitting of g out(T) to g(T) of the model. × fitting of Fθ-diagrams 

(Levenspiel 1972). Line in (a) represents the theoretical relation G + 0.5 = 

N sw/P D (Mecklenburgh and Hartland 1976); continuous line in (b) represents 

the theoretical (Westerterp et al. 1963) relation: N tis = 0.5 P D + 1; dash line in 

(b) shows relation of N tis with the values of P D estimated by fitting of 

experimental Fθ-diagram to that given by the Levenspiel’s equation (Levenspiel 

1972).......................................................................................................137 

Figure 5-13. Predicted deviation on conversion ( X ) in a 350-mm-long SPC 

tube (micro-bioreactor) for a homogeneous, isothermal chemical reaction as 

determined directly from the Eθ-diagram in the SPC tube, at v = 1.94 ml/min. 

................................................................................................................140 

Figure 5-14. Reduced RTD curves at three axial distances of the meso-reactor, 

operated at eight combination of hydraulic mean residence times (τ) and 

oscillatory flow Reynolds number (Re o). M – micro-probe 3; S1 – microprobe 

4; S2 – microprobe 5; (a) τ = 60 min, steady flow (i.e. Re o = 0); (b) τ = 60 

min, x 0 = 1 mm, f = 10 Hz, Re o = 312; (c) τ = 60 min, x 0 = 2 mm, f = 10 Hz, 

Re o = 625; (d) τ = 60 min, x 0 = 3.5 mm, f = 6 Hz, Re o = 657; (e) τ = 10 min, 

Re o = 0; (f) τ = 10 min, x 0 = 1 mm, f = 10 Hz, Re o = 312; (g) τ = 10 min, x 0 = 

2 mm, f = 10 Hz, Re o = 625; (h) τ = 60 min, x 0 = 3.5 mm, f = 6 Hz, Re o = 

657. Note that θ = t/ t , where t was determined from tracer response in 

micro-probe 3 (i.e. located at the higher axial distance). ............................142 

Figure 5-15. Number of tanks-in-series (N tis) estimated by direct comparison of 

Eθ-curve of micro-probe 5 response for increasing values of net flow Reynolds 

number (Re n). ○ Steady flow, i.e. Re o = 0; ● x 0 = 3.5 mm, f = 6 Hz, Re o = 312; 

xx 


■ x 0 = 2 mm, f = 10 Hz, Re o = 625; ♦ x 0 = 1 mm, f = 10 Hz, Re o = 657. Error 

bars shows standard deviation of values extracted for different experiments. 

................................................................................................................ 144 

Figure 5-16. a) determination of mixing time t 90 parameter from experimental 

data at 20 Hz and 1 mm; b) comparison of experimental t 90 parameter with 

estimated values with Eq. (5-14)............................................................... 145 

Figure 5-17. Variation of the mean values of mixing coefficient k m with fluid 

oscillation (a) frequency and (b) amplitude at different oscillation conditions. 

................................................................................................................ 147 

Figure 5-18. Effect of net flow rate over a) tracer mean residence time and b) 

backmixing, g, assuming a perfect step input at steady flow (no fluid 

oscillations). A comparison is presented between experimental (■) and 

simulated values (●) using a 2D-axisymmetric model................................ 148 

Figure 6-1. Illustration of the procedure applied to the determination of 

instantaneous-average (Vradial, Vaxial) and cycle-average velocities ( V axial , V radial ). 

................................................................................................................ 160 

Figure 6-2. Maps of instantaneous-average radial velocity (V neg radial in left hand 

side of figures and V pos radial in right hand side) and of axial velocity (V axial), through 

three complete fluid oscillation cycles, when the fluid is oscillated in batch 

mode at: (a) 4.1 s -1 and 1 mm, Re o = 117; (b) 5.1 s -1 and 1 mm, Re o = 203; (c) 

10.1 s -1 and 1 mm, Re o =259; (d) 11.1 s -1 and 1 mm, Re o = 348; (e) 15.1 s -1 

and 1 mm, Re o = 430; (f) 20.1 s -1 and 1 mm, Re o = 630............................ 162 

Figure 6-3. Maps of standard deviation of radial velocities (σ Vx) and axial 

velocities (σ Vy) as obtained from PIV velocity vector maps for three complete 

fluid oscillation cycles, when the fluid is oscillated in batch mode at: (a) 4.1 s -1 

and 1 mm, Re o = 117; (b) 5.1 s -1 and 1 mm, Re o = 203; (c) 10.1 s -1 and 1 mm, 

Re o = 259; (d) 11.1 s -1 and 1 mm, Re o = 348; (e) 15.1 s -1 and 1 mm, Re o = 

430; (f) 20.1 s -1 and 1 mm, Re o = 630. White dots represents the cycle- 

average parameter R s = σ radial/σ axial, while the sloping-dashed line represents the 

relationship σVx/σVy = d L = 0.294........................................................... 164 

Figure 6-4. Cycle-average velocity vector maps as seen in PIV measurements. 

(a) Re o = 117, x 0 = 1 mm, f = 4.1 s -1 ; (b) Re o = 203, x 0 = 1.4 mm, f = 5.1 s -1 ; 


xxi

(c) Re o = 259, x 0 = 0.9 mm, f = 10.1 s -1 ; (d) Re o = 348, x 0 = 1.1 mm, f = 11.1 

s -1 ; (e) Re o = 430, x 0 = 1.0 mm, f = 15.1 s -1 ; (f) Re o = 630, x 0 = 1.1 mm, f = 

20.1 s -1 ; (g) Typical flow patterns in a stirred tank reactor (side and bottom 

view) when a propeller and wall baffles are used (adapted from J. H. Rushton 

and J. Y. Oldshue, Chem. Eng. Prog., 49, 161 (1953))..............................166 

Figure 6-5. Effect of f on the best-fitting of G, P D and N tis dispersion parameters 

at a constant fluid oscillation, x 0 = 1 mm. (a) Effect of f on R V, R S, P D and N tis 

(white dots in P D and N tis curves represent interpolated data); (b) Correlation of 

PD and G with the products RV f 

xxii 

. Axial dispersion data is from Reis et al. 

(2004), using a net flow rate of 1.94 ml min -1 ............................................168 

Figure 6-6. Illustration of the effect of R V on the RTD (net flow rate of 1.94 ml 

min -1 ) at increasing f (from 0 to 15 s -1 ) and x 0 = 1 mm. (a) Oscillatory velocity 

(in mm s -1 ) at input; (b) Cycle-average axial and radial velocities within the 

cavities of SPC tube; (c) Experimental F(θ)-diagram (comparison with ideal 

plug-flow and stirred tank).........................................................................170 

Figure 6-7. Effect of Re o on mixing coefficient k m and comparison with the effect 

of Reo on the product RS f 

. ....................................................................172 

Figure 6-8. Correlation between the product s RS f and the products RV f . 

................................................................................................................173 

Figure 6-9. Cycle-average V axial and V radial as a function of Reo at different 

combinations of f and x0, i.e. Reo. (a) cycle-average axial velocities, V axial . (b) 

cycle-average radial velocities, V radial . () experiments at constant x0 of ∼1 

mm, i.e. run a); () experiments at different combinations of f and x 0, i.e. run 

b). Vertical dashed line represents the critical Re o where break of flow 

symmetry was detected. ...........................................................................174 

Figure 7-1. Typical flow patterns within a SPC-tube’s geometry (Reis et al., 

2005).......................................................................................................183 

Figure 7-2. Geometry of a 350-mm-long SPC tube (SPC1 – micro-bioreactor) 

and a 75 mm length tube (SPC2); details of SPC geometry. All distances are 

in mm. .....................................................................................................184 


Figure 7-3. a) Illustration of the variation of the dynamic O 2 method used in this 

work; b) experimental time profiles of O 2 dissolved saturation level using the 

proposed modification of the dynamic method ( SPC1 tube, x 0 = 1 mm and f = 

3 to 20 s -1 )................................................................................................ 186 

Figure 7-4. Experimental setup used in k La studies. ................................... 187 

Figure 7-5. Effect of OFM on the mean value of O 2 saturation levels at the 

outlet of SPC2 tube: a) effect of f; b) effect of x 0......................................... 192 

Figure 7-6. Estimated k La values for the SPC2 tube. a) 3-D representation of 

the effect of f and x 0; b) plot of k La regimes................................................ 193 

Figure 7-7. Comparison of k La values obtained with SPC2 tube and with the 

work of Oliveira and Ni (2004) using a conventional 50 mm internal diameter 

OFR for similar fluid oscillation conditions. ................................................ 194 

Figure 7-8. Effect of f on ε G when operating the SPC1 tube under OFM and a) a 

continuous fluid net flow (v = 1.58 ml min -1 ) or b) in batch mode (i.e. v = 0 ml 

min -1 )........................................................................................................ 196 

Figure 7-9. Comparison of experimental k La values with estimated ones, using: 

a) the semi-empirical correlation shown in Eq. (7.12); b) the coarse correlation 

presented in Eq. (7.8). The solid line represents y = x. .............................. 199 

Figure 7-10. Correlation between the experimental k La values and of the best- 

fitted backmixing parameter (G) of liquid phase (from Reis et al., 2004) for f in 

regime II (7.5 to 15 s -1 ). ............................................................................ 200 

Figure 7-11. Variation of k La with ε G at different f. Dotted lines represents the 

general tendency. 0 to 7.5 s -1 : regime I; 7.5 to 15 s -1 : regime II; 15 to 20 s -1 : 

regime III. ................................................................................................ 201 

Figure 7-12. Schematic representation of bubbles behaviour in the three 

identified regimes, in the studied range of f............................................... 202 

Figure 7-13. Ten frames sequence showing the bubble breakage phenomenon 

under OFM at 12 s -1 and 4 mm................................................................. 203 

Figure 8-1. (A) Experimental setup used in batch fermentations of S. 

cerevisiae: A- rotary motor; B- piston pump; C- gas inlet; D- gas outlet; E- fluid 

heating inlet; F- fluid heating outlet; G- SPC tube; H- purging port; I- sampling 


xxiii

port. (B) Detail of SPC (Smooth Periodic Constricted) tube geometry, which 

composes the novel, designed oscillatory flow Micro-bioreactor. All dimensions 

are in mm. ...............................................................................................214 

Figure 8-2. Time course of glucose concentration (S), cell dry weight (X) and 

ethanol concentration (P) in batch aerobic-growth on glucose of S. cerevisiae 

(bioprocess IIIa and IIIb – see Table 8-1). Fermentations in the 5-L stirred tank 

bioreactor (A), with an aeration rate of 1.1 vvm and in the micro-bioreactor (B) 

with an aeration rate of 0.064 vvm............................................................219 

Figure 8-3. Time profiles of cell dry weight, X (log scale) in aerobic-batch 

glucose-growth of S. cerevisiae (bioprocesses I to IV). Fermentations in the 5-L 

stirred tank (ST) bioreactor (A) and in the micro-bioreactor (B) with an aeration 

rate of 1.1 vvm for the 5-L ST and 0.064 vvm for the micro-bioreactor. (C) 

Time profiles of dry cell weight in two replicates of S. cerevisiae growth in a 

shake flask (SF) starting with a glucose concentration of 20 g/L (bioprocess 

IVc – see also Table 8-1); yeast was cultivated at 27 ºC and agitated in an 

orbital shaker at 150 rpm (these experiments correspond to the seed culture’s 

growth).....................................................................................................220 

Figure 8-4. Time profiles of residual glucose concentrations (S) in the aerobic 

batch growth on glucose of S. cerevisiae in bioprocesses I to IV (see Table 

8-1). Fermentations running in the 5-L stirred tank (ST) bioreactor (A) and in 

the micro-bioreactor (B). ...........................................................................221 

Figure 8-5. Specific growth rates (µ) for batch growth on glucose of S. 

cerevisiae at 25 ºC and different initial glucose concentrations (S 0) in the 5-L 

stirred tank (ST) bioreactor and in the micro-bioreactor. The specific growth 

rate presented for the SF was the averaged µ found for the seed culture 

growth, incubated at 27 ºC and 150 rpm. .................................................222 

Figure 8-6. Increase in dry cell weight, ∆X = X – X 0, obtained until complete 

depletion of glucose in the aerobic batch growth of S. cerevisiae on glucose in 

bioprocesses I to IV, for initial glucose concentrations S 0 of ∼ 5 - 20 g/L....223 

Figure 8-7. Time course of anaerobic batch growth on glucose (expressed as a 

relative function of the OD) of S. cerevisiae in the 2-L stirred tank (ST) reactor 

and in the micro-bioreactor. Experiments were run in parallel and started with 

glucose concentrations of 5 g/L (A), 10 g/L (B), 15 g/L (C) and 20 g/L (D). 

xxiv 


No seed culture was prepared and fermentation temperature was controlled at 

25 ºC. Note that OD was turned dimensionless with the OD peak obtained at 

the end of growth phase (i.e. at the instant of glucose depletion, as indicated 

from pH measurements). ......................................................................... 226 

Figure 9-1. Experimental setup used in batch biotransformations. 1- rotary 

motor; 2- piston pump; 3- air inlet; 4- air outlet; 5- fluid heating inlet; 6- fluid 

heating outlet; 7- SPC tube; 8- purging port; 9- sampling port.................... 236 

Figure 9-2. Concentration of γ-decalactone experimentally obtained with a SPC 

tube in the four biotransformations carried out in this study (details given in 

................................................................................................................ 237 

Figure 9-3. Evolution of the number of cells (n) of Y. lipolytica in suspension 

within a SPC tube in the four biotransformations carried out in this study 

(details given in ........................................................................................ 238 

Figure 9-4. Evolution of the specific rate of production of γ-decalactone (υ) with 

the oscillatory mixing intensity (i.e. oscillatory Reynolds number, Re o). ....... 239 

Figure 10-1. Simplified kinetic scheme showing first-order refolding competing 

with higher-order aggregation, where k r is the refolding rate constant and k a is 

the aggregation rate constant (Hevehan and Clark 1997). ......................... 248 

Figure 10-2. Relation between lysozyme concentrations (before dilution with 

TFA) and slope of decrease on absorbance (450 nm) of a cell suspension 

(0.15 g/l Micrococcus lysodeikticus)......................................................... 251 

Figure 10-3. Refolding yield (Y ref) of lysozyme in a batch, unstirred Falcon tube 

(denatured-reduced lysozyme added with a sharp micropipette stroke and 

solution briefly mixed); ● this work; □ results from Buswell & Middelberg 

(2003). .................................................................................................... 253 

Figure 10-4. Refolding yield (Y ref) of lysozyme in a batch, small stirred beaker 

(refolding initiated with a sharp addition of denatured lysozyme); two parallel 

experiments are shown. Vertical bars represent standard deviation of Y ref.. . 254 

Figure 10-5. Refolding yield (Y ref) of lysozyme in a batch, small stirred beaker; 

refolding initialled through a slow addition (30 s) of denatured-reduced 

lysozyme solution (average injection rate = 0.7 ml/min)............................ 255 


xxv

Figure 10-6. Refolding yield (Y ref) of lysozyme in a batch, 350-mm-long SPC 

tube at a constant x 0 = 3 mm and varying f and injection procedures. ○ f = 10 

Hz, sharp injection; ♦ f = 10 Hz, injection time = 4 min; □ f = 3 Hz, injection 

time = 2 min; ▲ f = 10 Hz, injection time = 2 min and OFM stopped at t = 4 

min. .........................................................................................................256 

Figure 10-7. Comparison of refolding yields (Y ref) of lysozyme in the continuous 

meso-reactor (along the residence time, t) with the values of Y ref in batch 

dilution refolding. ● batch refolding in a small stirred beaker, with injection 

time >> 0 s; ○ continuous refolding in a meso-reactor at x 0 = 1 mm, f = 1 mm, 

Re o = 30; □ batch refolding in the SPC tube, x 0 = 3 mm, f = 3 Hz, injection 

time = 2 min. ...........................................................................................257 

xxvi 


List of general nomenclature 

Symbol 

A acceleration [L T -2 ] 

Cd orifice discharge coefficient dimensionless 

d tube diameter [L] 

D’ characteristic dimension of effective width of obstacle [L] 

d 0 orifice diameter [L] 

d i reactor internal diameter [L] 

f oscillation frequency [T -1 ] 

f v frequency of vortex shedding [T -1 ] 

H reactor or column height [L] 

h half channel width [L] 

h max maximum channel width [L] 

k La oxygen mass transfer coefficient [T -1 ] 

L baffles spacing [L] 

N number of baffles per unit length [L -1 ] 

P power input [M L 2 T -2 ] 

p pressure drop [M L -2 ] 

Re o oscillatory Reynolds number dimensionless 

Re n net-flow Reynolds number dimensionless 

Re ob obstacle Reynolds number dimensionless 

Re p pulsating Reynolds number dimensionless 

u mean superficial flow velocity [L T -1 ] 

u ∞ liquid free velocity [L T -1 ] 

u p pulsating velocity [L T -1 ] 

u peak peak velocity at the maximum channel width [L T-1] 


xxvii

v velocity [L T -1 ] 

S object to image scale factor dimensionless 

Sc Schimdt number dimensionless 

Sh Sherwood number dimensionless 

St Strouhal number dimensionless 

St f Strouhal number (by Sobey (1980)) dimensionless 

t time [T] 

V reactor volume [L 3 ] 

V r fluid velocity at cylindrical coordinate r [L T -1 ] 

v s superficial gas velocity [L T -1 ] 

V z fluid velocity at cylindrical coordinate z [L T -1 ] 

Vθ fluid velocity at cylindrical coordinate θ [L T -1 ] 

w angular velocity [T -1 ] 

x displacement [L] 

x 0 oscillation amplitude (centre-to peak) [L] 

Greek symbols 

α free baffle area dimensionless 

δ baffle thickness [L] 

δ’ Stokes layer thickness [L] 

µ viscosity [M L -1 T -1 ] 

µ 0 normal laminar viscosity [M L -1 T -1 ] 

µ t turbulent viscosity [M L -1 T -1 ] 

ρ density [M L -3 ] 

υ kinematic viscosity [L 2 T] 

xxviii 


List of tables 

Table 2-1: Examples and applications of oscillating devices since the 1970’s. 9 

Table 2-2: Summary of main USA patents related with oscillating systems. f 

and x 0 are the fluid oscillation frequency and amplitude, respectively ........... 10 

Table 2-3: Experimental studies and applications of oscillatory flow reactor 

(OFR) in the last 12-15 years. f and x 0 are the fluid oscillation frequency and 

amplitude, respectively............................................................................... 13 

Table 2-4: Summary of works concerning the fundamental study of OFM in 

OFR’s......................................................................................................... 20 

Table 2-5: Relevant studies concerning the research of OFM and the effect of 

constrictions............................................................................................... 32 

Table 2-6: Some of the applications of Biotechnology (Lee 1984) ................ 43 

Table 2-7: Summary of the main features of reactor classes (Cabral et al. 

2001) ........................................................................................................ 46 

Table 4-1: Experimental Conditions ............................................................. 83 

Table 4-2: Details of mesh used for 3-D numerical simulations presented in 

Figure 2 ..................................................................................................... 85 

Table 4-3: Minimum critical Re o observed for the screening reactor and 

comparison with some reported values for the conventional OFR................. 93 

Table 4-4: Comparison of cycle average axial, radial (and tangential) velocities 

measured by PIV with the results from numerical simulations, using 2-D 

axisymmetric and 3-D laminar models; Re o = 348, f = 11.1 Hz, x 0 = 1.1. Re n = 

0.............................................................................................................. 102 

Table 4-5:. Comparison of measured mixing intensity by PIV with the results 

from numerical simulations, using 2-D axisymmetric and 3-D laminar models; 

Re o = 348, f = 11.1 Hz, x 0 = 1.1 mm......................................................... 102 

Table 5-1: Conversion between mean flow rate (v), superficial liquid velocity 

(u LS), net flow Reynolds number (Re n) and hydraulic residence time (τ) for the 

meso-reactor experiments. ....................................................................... 125 


xxix

Table 5-2: Correspondence between superficial liquid velocity (u LS), liquid flow 

rate (v) and net flow Reynolds number (Re n). All values are based on d = 4.4 

mm..........................................................................................................127 

Table 5-3: Equations for cross-correlation of dimensionless axial dispersion 

number (P D) with the backmixing (G), the number of tanks-in-series (N tis) and 

the volume-fraction of PFR (V p/V), using the values of parameter estimated 

through various techniques. ......................................................................138 

Table 5-4: Summary of maximum and minimum values of dimensionless 

parameters I Ntis, I PD, I G and I Vp/V obtained with the introduction of OFM, towards 

the ideal convective or dispersive systems, respectively. * Ideal flow case for 

convective flow is a PFR; ideal case for dispersive flow is a CSTR...............139 

Table 5-5: Predicted conversions and respective deviations from conversion in 

a PFR, calculated from the experimental RTD (Eθ-diagrams) in the 350-mm- 

long SPC tube, for a continuous, homogeneous, isothermal first-order reaction, 

at v = 1.94 ml/min...................................................................................141 

Table 6-1: Experimental conditions (f, x 0 and Re o) used for PIV measurement of 

flow patterns in SPC tube (from Reis et al. 2005). Experiments run (a) are 

those performed at a constant value of x 0 ≈ 1 mm, while experimental run (b) 

comprises experiments performed at further values of x 0. u peak is the maximum 

theoretical axial velocity of the fluid (i.e. equal to 2 π f x0), while V axial, 

theo is 

the (theoretical) cycle-average axial velocity throughout a complete oscillation 

cycle (i.e. V = 2 f x cos( 

2π 

f t) 

1 2 f 

xxx 

axial, 

theo 

5 / 4 f 

∫ 

3 / 4 f 

π ); these values will be 

0 

used for comparison of experimental data in Figure 6-2 and Figure 6-9......157 

Table 7-1: Dimensions and constants used in the experiments with tubes 

SPC1 and SPC2 .......................................................................................184 

Table 7-2: Comparison of performance of the SPC geometry for O 2 mass 

transfer in a gas-liquid system with further reported works in literature.......198 

Table 8-1: Averaged yields of biomass on substrate (Y X/S) and specific substrate 

uptake rate (q s = µ/Y X/S) during the exponential phase of the aerobic growth of 

S. cerevisiae in bioprocesses I to IV and in three different small-scale vessels: 

5-L stirred tank (ST) bioreactor, micro-bioreactor and shake flask (SF). S 0 is the 


initial glucose concentration, as measured after inoculation with 10 % v/v of 

seed culture ............................................................................................. 224 

Table 9-1: Fluid oscillation conditions used in the four biotransformations 

carried out in this work, at different combinations of fluid oscillation frequency 

(f) and amplitude (x 0) (expressed as centre-to-peak); Re o is the oscillatory 

Reynolds number and is a measure of mixing intensity.............................. 235 


xxxi

Chapter 1 Introduction 

Biotechnology is a multidisciplinary field having its roots in the 

biological, chemical and engineering sciences (Figure 1-1) leading to a 

host of specialities, e.g. molecular genetics, microbial physiology, 

biochemical engineering (Moo-Young et al. 1985). 

Figure 1-1. Multidisciplinary nature of biotechnology (Moo-Young et al. 


1985). 

1

About 160 biopharmaceuticals have recently gained medical approval and several hundred are in the 

pipeline (Walsh 2005). Biopharmaceuticals (recombinant therapeutic proteins, monoclonal antibody-based 

products used for in vivo medical purposes and nucleic acid-based medicinal products) actually represent 

approximately one in every four genuinely new pharmaceuticals (Walsh 2003). But the successful 

commercialization of novel processes/products developed by pure scientists requires the development of 

large-scale processes which are both technologically viable and economically efficient. 

Biochemical engineering is focused in conducting biological processes to the industrial scale. The role of 

the biochemical engineers has become more important in recent years due to the dramatic developments 

of biotechnology (Lee 1992). They actually play an important function on the commercialisation of 

biotechnology, linking the biological sciences with the chemical engineering design. Nowadays, the 

challenge for the biochemical engineer is enhanced. To carry out a bioprocess at large scale the engineer 

needs: 

2 

a) to work together with biological scientists; 

b) to obtain the best biological catalyst (microorganism, animal cell, plant cell, or enzyme) for a 

desired process; 

c) to create the best possible environment for the catalyst, by designing the bioreactor and 

operating it in the most efficient way; 

d) to separate the desired products from the reaction mixture in the most economical way. 

The biological reactor (bioreactor) is of such importance in biological processes as the heart on a live 

body. A bioreactor can be understood as “a vessel where a biological reaction or change takes place, 

usually a fermentation or biotransformation, including tank bioreactors, immobilised cell bioreactors, 

hollow fibre and membrane bioreactors and digesters” (Bains 1998). The design of biological reactors is 

an integral part of biotechnology. Especially when designing bioreactors, integration of biological and 

engineering principles is essential (Cabral and Tramper 1993). 

Proteomics research as a result of the human genome project demanded many recombinant constructs 

with potentially beneficial therapeutic products to be designed and needing to be tested for efficacy of 

expression (Betts et al. 2005). This calls for the performance of a vast number of development 

fermentations. In order to speed up this process, the use of controlled high-performance parallel (scale- 

down) reactor systems is required. 

N. Reis Novel Oscillatory Flow Reactors for Biotechnological Applications

The preceding tasks involve process design and development including the bioprocessing at a small-scale, 

which is familiar to chemical engineers for the chemical processes. Techniques which have been applied 

successfully in chemical processes can be used in bioprocesses with small modifications (Lee 1992). 

Biochemical conversions with the aid of biological catalysts differ from purely chemical processes in a few 

numbers of ways (Atkinson 1974). In both cases, the best possible environment must be created by 

designing efficient reactors. 

A wide range of bioreactor classes may be identified, attending to their design, power source and number 

of phases (Crueger 1987). One particular design has gained increasing interest in the last decade: the 

oscillatory flow reactor (OFR). It is basically a column provided with periodic sharp constrictions (baffles) 

and operating under oscillatory flow mixing (OFM). The formation and dissipation of eddies has proved to 

result into significant enhancement in processes such as heat transfer (Mackley and Stonestreet 1995; 

Mackley et al. 1990), mass transfer (Hewgill et al. 1993; Ni et al. 1995a; Ni et al. 1995c), particle mixing 

and separation (Mackley et al. 1993), liquid-liquid reaction (Ni and Mackley 1993), polymerization (Ni et 

al. 1998b; Ni et al. 1999), flocculation (Gao et al. 1998) and crystallization. Research has been further 

extended to include: flow patterns (Brunold et al. 1989; Mackley and Ni 1991; Mackley and Ni 1993), 

local velocity profiles and shear rate distribution (Ni et al. 1995b), residence time distribution (Dickens et 

al. 1989; Mackley and Ni 1991; Mackley and Ni 1993; Ni 1994), dispersion (Howes 1988; Howes and 

Mackley 1990), velocity profiles (Liu et al. 1995) and scale-up correlations (Ni and Gao 1996). 

Unlike conventional tubular reactors, where a minimum Reynolds number must be maintained, mixing in 

an OFR is independent of the net flow, allowing long residence times to be achieved in a reactor of greatly 

reduced length-to-diameter ratio. For example, OFR is able to achieve a required product specification in a 

saponification process with a residence time one- eigth th that of a full-scale batch reactor (Harvey et al. 

2001). In this case, OFR comes in line with the process intensification that is redirecting the reactor 

engineering (Harvey et al. 2003; Mackley 2003). Overall, the Oscillatory Flow Mixing (OFM) is presented 

as a “technology ready to deliver” (Harvey and Stonestreet 2001). 

The aim of this thesis is to present and characterise innovative configurations of oscillatory flow reactors 

for biotechnological applications. Key areas of interest are the scale-down of OFRs for fast upstream 

development of biotechnological processes, from a single (liquid) phase to a multi- (four) phase (gas-liquid- 

liquid-solid) system. Such novel designs may be very useful in some stages of bioprocesses development, 


3

while selecting and optimising biotransformation media and operational conditions, as well in bioprocess 

design. 

The text is organised in eleven chapters. Chapters of experimental results (chapters 4 to 10) are provided 

with a specific introduction and a specific list of references. Chapter 2 presents an overview of reactor 

designs based on OFM, and more particularly the research on OFR, introducing some concepts and 

dimensionless groups very important in designing oscillatory reactors. The different designs and 

applications of OFRs in several previous studies are deeply reviewed. A state-of-the-art of reaction 

engineering tools is presented as well as a review of the main applications in biotechnology and of the 

main topics of bioprocess design. Conventional bioreactor designs are classified and examined and the 

main issues in reactor’s scale-down are explained. 

A novel tube geometry is introduced in Chapte 3 (Materials and Methods Section) and two scale-down 

reactor configurations (micro-bioreactor and meso-reactor) developed during the running time of this thesis 

are presented. In Chapter 4 the fluid mechanics generated within this particular tube geometry are 

investigated and consequently used in the validation of numerical simulations carried out with CFD’s 

technique. Chapter 5 assesses the deviation of both micro-bioreactor and meso-reactor from the ideal flow 

cases of ideal plug flow reactor and completely back-mixed reactor, while the steady flow was matched 

with numerically predicted flow backmixing using CFDs. Also, batch mixing in the micro-bioreactor is 

considered, thus mixing times results are presented. Chapter 6 shows a statistical correlation of deviations 

from ideal mixing/flow (summarised in Chapter 5) with the flow patterns observed in the tube geometry 

from (Chapter 4). The aeration capacity in the small-scale geometry was studied in Chapter 7. Chapters 8 

to 10 are dedicated to the biotechnological applications of both micro-bioreactor and meso-reactor. In 

particular, chapters 8 and 9 test the two scale-down platforms with two workhorse microorganisms: 

Saccharomyces cerevisiae and Yarrowia lipolytica¸respectively, while Chapter 10 assesses the dilution 

refolding of lysozyme. Overall conclusions and suggestions of future work are presented in Chapter 11. 

4 


Chapter 2 Literature review 

The application of external energy in the pulsing form (oscillatory flow 

mixing - OFM) has, for a long time, been a common practice to improve 

reaction performance, namely mass transfer rates in chemical 

engineering units. The general principles associated with the pulsing 

column were established by Van Dijck (1935), at the Royal Dutch/Shell 

Laboratory in Amsterdam, in the 1930’s (Figure 2-1). Since then, a 

number of techniques, based on several principles, have been 

developed and adapted for their applications to very different fields 

(Lema et al. 2001). 


5

6 

Figure 2-1. Technical drawing of a Van Dijck’s US Patent (1935). 

2.1 Types and applications of oscillating devices 

2.1.1 Types of oscillating devices 

In general, oscillating equipment may be classified in two main types (Lema et al. 2001): 

a) Alternating motion of some intrinsic elements of the column. It is worth mentioning the 

reciprocating plate columns (Figure 2-2A and 3B), in which the pulsation is generated by 


means of an upwards-downwards motion of plates (e.g. Baird and Rao (1988); Skala and 

Veljkovic,(1988b)) and the columns with oscillating piston (Figure 2-2C and 3D), where a plug 

is coupled to the bottom of the column (Harrison and Mackley 1992). 

b) Oscillation is generated by the hydraulic transmission of a perturbation to the liquid contained 

in the column. This perturbation is typically generated by e.g. systems using positive 

displacement pumps (plug or membrane) to introduce the feed into the column (Mak et al. 

1992) and the pneumatic oscillating systems. In the latter example, the oscillation is 

generated by means of a pressurized gas which propels the liquid contained in a parallel 

branch to the column (Murthy et al. 1987). The self-propelled oscillators are based on a 

different concept. In this case, fluid oscillating is the result of the liquid entering the columns, 

through a pulsation chamber. Once the pressure in the chamber is high enough, the 

membrane covering the feeding tube injects the liquid into the column; this membrane then 

closes the inlet tube again, which creates a cyclical feed system. In contrast with the previous 

pulsators, in this system, the motion of the liquid in the column is always generated in the 

upward direction (Baltar 1972). 

Figure 2-2. Examples of oscillating vessels: reciprocating plates - (A) and (B) – and oscillating piston – (C) 

and (D). (A) from Prochazka and Rod (1974). (B) from Ni (2002)(2002). (C) from Prochazka and Rod 

(1974), (D) from Hounslow and Ni (2004). See references for numbering details. 


7

2.1.2 Industrial applications of oscillating reactors 

The oscillating reactors were firstly used in separation processes in order to enhance the contact between 

the phases and, consequently, to improve mass transfer rates. Since then, they have been applied to a 

number of systems, either chemical or biochemical, under several configurations. Table 2-1 summarises 

some applications of oscillating vessels since the 1970’s. 

In the last three decades, the number of publications resulting from the study of OFM has increased 

several times, as seen in Figure 2-3, which demonstrates that this is a technology creating an increasing 

interest in the scientific community. Several patents are currently protecting novel oscillating devices’ 

designs and/or their commercial applications. 

Table 2-2 summarises the main registered patents in the US Patents Office. Very different types of 

oscillator systems were coupled to several types of unit operations and processes (Table 2-1 and Table 

2-2). 

8 

# of publications 

90 

80 

70 

60 

50 

40 

30 

20 

10 

0 

1970 1975 1980 1985 1990 1995 2000 2005 

Publication year 

Figure 2-3. Number of publications out coming from a global search in ISI Web of Knowledge 

(http://portal.isiknowledge.com/portal.cgi) using keywords “oscillatory flow”. All citation databases, 

document types and languages were considered in the search. 


Table 2-1: Examples and applications of oscillating devices since the 1970’s 

Oscillating reactor Designation Application Reference 

Plug oscillator Column of perforated 

plates 

Production of SCP Serieys et al. (1978) 

Pneumatic Packed-bed column with Anaerobic treatment of waste- Brauer and Sucker (1978) 

oscillator perforated plates water 

Pneumatic Packed-bed column with Alcoholic fermentation Navarro and Goma (1980) 

oscillator perforated plates 

Membrane Column of perforated L-L extraction Golding and Lee (1981) 

oscillator plates 

Alternating motion 

pumps 

Packed-bed column S-L extraction Goebel and Fortuin (1986) 

Reciprocating Column of perforate Absorption Skala and Veljkovic (1988a) 

plates column plates 

Oscillating pump Ultrafiltration unit Clarification of juices Finnigan and Howell (1989) 

Membrane Anaerobic Filter Anaerobic treatment of Etzold and Stadlbauer 

oscillator 

wastewater 

(1990) 

Oscillating piston Batch bioreactor Production of biodegradable Harrison and Mackley, 

plastic 

(1992) 

Pulsative Pumping High-Efficiency 

Animal trials Bellhouse et al.(1973) 

System 

Membrane Oxygenator 

Oscillatory Flow Continuous OFR Process intensification of Harvey et al. (2003) 

Reactor 

biodiesel production 

Oscillatory Flow Continuous OFR Continuous production of sterols Harvey et al. (2001) 

Reactor 

in an ester saponification 

reaction. 

Oscillatory Flow Oscillatory Baffled Batch Crystallization of Paracetamol Chew et al. (2004a) 

Reactor 

Crystallizer (OBBC) 

Oscillatory Flow Pulsed Baffled Tubular Treatment of wastewater Fabiyi and Skelton (1999; 

Reactor 

Photochemical Reactor (photocatalytic oxidation) 2000a; 2000b), Gao et al. 

(2003) 

Reciprocating Novel pilot scale gas– Counter-current gas–liquid Gomaa and Al Taweel 

plates column liquid reciprocating plate 

column 

contacting 

(2005) 

Oscillating piston Novel oscillatory flow 

reactor 

Protein refolding Lee et al. (2002; 2001) 

Pulsed reactor Pulsed reactor Pulse combustion: dehydration, 

decomposition reactions, 

oxidation 

Begand et al. (1998) 

Reciprocating plate Reciprocating plate 

agitator 

Fluid mixing Masiuk (1999) 

Oscillating pistons Vortex wave membrane Aeration of high density Millward et al. (1996) 

(moving against 

two pump bags) 

bioreactor 

mammalian cell culture 

Oscillating piston Oscillatory baffled reactor 

(OBR) 

Polymer production Ni et al. (2002c) 

Pneumatic Pressure swing operated Bioconversion in a solid Lee and Fan (1999) 

oscillator reactor 

immobilised system 


9

10 

Table 2-2: Summary of main USA patents related with oscillating systems. f and x 0 are the fluid oscillation frequency and amplitude, respectively 

Oscillating equipment Reactor designation Settings Patented application Reference 

Reciprocating plates column Column of perforated plates f, x 0, N and L are determined in each particular case Improve efficiency of (immiscible) 

liquid-liquid washing or extraction 

process 

Reciprocating plates column 

or oscillating piston 

Reciprocating piston or 

diaphragm 

Large-diameter vibrating 

or/and pulsating column 

Tubular continuously flow 

reactor 

Different kinds of perforated trays 

L defined as proportional to d i 

Vibratory trays or fluid pulsating 

Oscillating motion is superimposed on the linear 

(laminar) flow of reactants in order to maintain turbulent 

flow throughout; f and volume liquid displaced adjusted 

to particular reaction situation; peak instantaneous 

Reynolds number > 3,000 ; reciprocating piston or 

diaphragm 

Compressed air Pulsed fluidised bed f is determined by the rotation speed of a disc valve; 

1-50 Hz is effective; 1-15 Hz seems adequate; 8-10 is 

optimum in several cases 

Piston oscillator Continuous oscillatory 

baffled reactor; 

Reciprocating plates column Batch oscillatory baffled 

reactor (Premixer reactor) 

Tubular tube, may be operated vertically or horizontally, 

temperature controlled 

d = 0.1 – 5 m, L = 1.8d, f = 0-10 Hz, x 0 = 0-20 mm, α = 

21 % 

Further possible dimensions: L = 1.2 – 2.0d (preferably 

Apparatus for bringing fluid phases 

(including gases) into mutual contact 

Method to prevent the solids deposits 

in the walls of tubular reactors by 

pulsed flow 

Processing materials in a batchwise or 

continuous fluidised bed, such as a 

drier; improvement is higher for 

particulate solids of a non-uniform size 

Continuous phase-separated synthesis 

of particulates; 

Continuous polymerization 

Van Dijck (1935) 

Prochazka and Rod (1974) 

Soubrada and Galvez (1981) 

Kudra et al. (1999) 

Ni (2002)(2002) 

1.5d); α = 10 – 40 % (preferably 21 %), d = 0.1 – 5 m 

Vertical tube, f and x0 adjustable Premixer of monomer with an initiator Ni (2002)(2002)

2.2 The Oscillatory Flow Reactor (OFR) 

In the late 1980s, research aiming at generating unsteadiness in a laminar flow showed that when a 

periodically reversing flow exists in a tube fitted with orifice-type baffles mounted transverse to the flow and 

equally spaced, vortex rings are formed downstream of the baffles. On each flow reversal the vortices are 

swept into the central region of the tube and the cycle of vortex formation, growth and ejection results in a 

state of ‘chaotic’, advected mixing in each inter-baffle cavity (Brunold et al. 1989; Dickens et al. 1989). 

This marked the birth of the oscillatory flow reactor (OFR). 

The application of periodic fluid oscillations to a cylindrical column containing evenly spaced orifice baffles 

is the basic concept of OFR. A schematic representation of an OFR is shown in Figure 2-4. The OFR can 

be operated batchwise or continuously in horizontal or vertical tubes. The liquid or multiphase fluid is 

typically oscillated in the axial direction by means of diaphragms, bellows or pistons, at one or both ends 

of the tube (Ni et al. 2002a). The sharp edged baffles are fixed and distributed along the tube at a regular 

spacing (L). Another system for generation of flow oscillations is also common and has already been 

described (reciprocating plates column), which works by moving a set of baffles up and down from the top 

of the tube. 

The mixing within an OFR is an efficient mechanism, where fluid moves from the walls to the centre of the 

tube. The intensity of this movement is affected by the oscillation frequency, f, and amplitude, x 0. The flow 

becomes progressively more complex as the oscillation frequencies and amplitudes increase. These 

results are consistent for a 25-mm (Brunold et al. 1989; Dickens et al. 1989) and also for a 50-mm 

internal diameter tube (Ni et al. 1995c), indicating that the fluid mechanical conditions in an OFR can be 

linearly scaled up, as demonstrated later on by Ni et al. (1996). 


11

Figure 2-4. Schematic representation of cross section in an OFR. d i – reactor internal diameter, L – baffles 

12 

spacing, d 0 – orifice diameter, δ - baffle thickness. 

Table 2-3 summarises the most frequent OFR design and operational settings used in past experimental 

research works. 

One particularly advantageous application area of OFR is for performing 'long' (usually over 10 minutes) 

reactions in configurations which are substantially more compact than batch reactors, and which have 

substantially smaller length to diameter ratios than conventional tubular reactors. A novel methodology for 

design of continuous OFRs is based on mixing, as presented by Stonestreet and Harvey (2002)(2002). 


Table 2-3: Experimental studies and applications of oscillatory flow reactor (OFR) in the last 12-15 years. f and x 0 are the fluid oscillation frequency and amplitude, 

respectively 

Oscillation equipment Reactor designation Settings Characterisation/application Reference(s) 

Oscillating piston Baffled tubes di = 12 cm, length: 2 x 1,0 m, 55 orifice baffles Heat transfer measurements for pulsative flow Mackley et al. (1990) 

f = 3-14 Hz, x0 = 1-6 mm 

Operated Horizontally 

Operation fluid: lubrificating oil at 60 ºC 

Oscillating piston Baffled tube di = 12 cm, length = 1.0 m 

Heat transfer and associated energy dissipation Mackley and Stonestreet (1995); 

(2 x external) 

f = 0-10 Hz, x0= 1-7 mm 

measurements for oscillatory flow in baffled tubes, (same vessel as Baird and 

Operated horizontally 

using mineral oil 

Stonestreet (1995)) 

Oscillating piston Pulsed baffled tube di = 75.6 mm, length: 910 mm, L = 0.5d Photocatalysed mineralization of methylene blue in a Fabiyi and Skelton (1999; 2000a; 

photochemical reactor f =: 0-11 Hz, x0 = 0-4.5 mm 

continuous flow operation 

2000b) 

(PBTPR) 

UV lamp in its central axis 

Oscillating piston Pulsed baffled tube bundle di = 25 mm, length = 1.0 m, L = 1.5d 

Experimental flow pattern and associated residence Ni (1994) 

f = 0.5-9 Hz 

time distribution measurements 

Mackley and Ni (1993) – 

Operated vertically 

multitube arrangement 

Oscillating piston Pulsed baffled bioreactor di = 50 mm, length: 500 mm, L = 1.5d 

f = 1-12 Hz, x0 = 0-14 mm 

Mass transfer measurement in yeast culture Ni et al. (1995c) 

Oscillating piston Baffled tube di = 25 mm, length = 1 m, L = 1.5d 

Fluid dispersion and concentration profile 

Ni (1995) 

f = 0.5-9 Hz, x0 = 1-10 mm 


measurement 

Oscillating piston Batch pulsed baffled di = 50 mm, length = 500 mm, L = 1.5d 

Study of mass transfer of oxygen in yeast re- Ni et al. (1995a) 

bioreactor 

f = 0.5-9 Hz, x0 = 1-10 mm 

f = 1-12 Hz, x0 = 0-14 mm 


suspension and yeast culture 

Oscillating piston Double-pass tube Section 1 : di = 38 mm, length = 2 m 

Measurement of velocity of single particles for steady Liu et al. (1995) 

Section 2 : di = 43.5 mm, length = 2 m, L = 1.5d 

f = 0.5-10 Hz 


and oscillatory flows in plain and baffled tubes 

Oscillating piston (2x) Pulsed baffled reactors Reactor 1: di = 50 mm, length = 525 

Scale-up correlation for based on mass transfer Ni and Gao (1996) 

Reactor 2: di = 100 mm, length = 875 mm measurements in two pulsed baffled reactors, with 

f = 1-10 Hz, x0 = 1-12 mm 

different diameters but in which the water level is 


maintained constant 

13

14 

Table 2-3: (Continued) 


Oscillating piston Pulsed baffled reactor di = 50 mm, length: 800 mm 

Determination of degree of oil-water dispersion by two Zhang et al., (1996) 

f = 1-10 Hz, x0 = 0-15 mm 

methods: sampling technique and the visualization 


method; surfactants effects on dispersion 

Oscillating piston Pulsed baffled reactor di = 50 mm, height: 525 mm, L = 1.8D 

Effect of surfactants on mass transfer of oxygen into Ni et al (1997) 

f = 1-10 Hz, x0 = 0-12 mm 


water glycerol solutions; KLa measurements; 

Oscillating piston Modified pulsed baffled di = 50 mm, H = 1 m, L = 35-100 mm 

Experimental flow visualisation Gough et al, (1997) 

reactor 

f = 1-6 Hz, x0 = 5-25 mm 


Oscillating piston Batch oscillatory-baffled di = 50 mm, H = 750 mm 

Droplet size and size distribution in 

Ni et al. (1998b) 

column 

f = 1-10 Hz, x0 = 1-15 mm 

methylmethacrylate suspension 


Correlation of particle size with droplet size in 

suspension polymerisation of methylmethacrylate 

Ni et al. (1999) 

Oscillating piston Oscillatory-baffled column di = 50 mm, d = 950 mm 

The effect of gap size between baffle outer diameter Ni and Stevenson (1999) 

f = 1-10 Hz, x0 = 1-15 mm 


and tube inner diameter on the mixing characterists 

Oscillating piston Oscillatory baffled column di = 50 mm, H = 500 mm, L = 75 mm 

The measurement of stains rate using particle image Ni et al (2000a) 

f = 0.2-10 Hz 


velocimetry (PIV) 

Oscillating piston Novel continuous oscillatory di = 40 mm, total length = 25 m, L = 1.8d Study of parameters affecting fluid dispersion. Ni and Pereira (2000) 

baffled tube 

f = 1-4 Hz, x0 = 1-20 mm 

This new reactor consist on 14 glasses tubes 


vertically disposed and connected by a straight Ubends 

Oscillating piston Batch oscillatory baffled di = 50 mm, H = 500 mm 

Flocculation of bentonite and Alcaligenes eutrophus; Gao et al. (1998) 

flocculator 

f = 0.2-10 Hz, x0 = 1-12 mm 

the measurement of mean stains rates and their 


distribution using digital particle image velocimetry 

Oscillating piston Gassed oscillatory baffled di = 50 mm, H = 1.5 m, L =1.5d 

Gas hold-up and bubbles diameters Oliveira and Ni (2001) 

column 

f = 1-5 Hz, x0 = 2-8 mm 


Oscillating piston Continuous oscillatory di = 40 mm, total length = 25 m 

Droplet size distribution in the absence of surfactants Pereira and Ni (2001) 

baffled reactor (COBR) f = 0-5 Hz, x0 = 0-60 mm 


and coalescence inhibitors



Oscillating piston Oscillatory baffled reactor di = 50 mm, H = 1 m 

Polymer product engineering: particles production Ni et al. (2002c) 


with controlled size and morphology in batch and 

continuous mode 

Oscillating piston Batch oscillatory baffled di = 50 mm, H = 950 mm, L = 1.5f 

The effect of tracer density (tracer solution of Ni et al. (2002b) 

column 

f = 1-10 Hz, x0 = 1-15 mm 

potassium nitrite) on axial dispersion; comparison 


with both “Tank-in-series” and “Plug flow with axial 

dispersion” models; mechanical energy empirical 

correlations establishment 

Oscillating piston Oscillatory baffled column di = 50 mm, H = 500 mm 

Computation fluid dynamics (CFD) modelling of flow Ni et al. (2002a) 

f = 0.2-10 Hz 

patterns; 3-D numerical simulation of oscillatory flow 


in a baffled column 

Oscillating piston Baffled tube d = 26 mm, length = 1.08 m, L = 1.5d 

Gas-liquid (air/water) mass transfer enhancement Hewgill et al. (1993) 

f = 0-8 Hz, x0 = 0-6 mm 

determination and visualisation using oscillatory flow 


in a baffled tube 

Oscillating piston Periodic baffled tube arrays di = 26 mm, length = 5 x 1.0 m, L = 1.5d Flow pattern and associated residence time Mackley and Ni (1993) 

in two different 

f = 0.5-9 Hz 

distribution measurements 

configurations: serial and 

parallel (multitube) 


Oscillating piston Novel oscillatory flow d = 2.4 cm, d0 = 1.2 cm, H = 28 cm, baffled 

reactor 

w = 22 rad-1 and x0 = 3 mm, or 

w = 4.09 rad-1 Refolding of denatured-reduced lysozyme Lee et al. (2002; 2001) 

and x0 = 1 mm 

Oscillating piston Oscillatory baffled columns Column 1: d = 50 mm, H = 950 mm 

Study of the effects of geometrical parameters on Ni et al., (1998a) 

(column 1) 

Column 2: d = 50 mm, H = 990 mm 

mixing time; the effect of tracer concentration, baffle 

Reciprocating plates 

Column 3: d = 90 mm, H = 730 mm 

spacing and free baffle area 

(columns 2 and 3) 

f = 1-10 Hz, x0 = 1-20 mm 


Oscillating piston Glass baffled tube di = 23 mm, length = 1 m, L = 1.5d 

Mixing and separation of particle suspension using Mackley et al. (1993) 

(electromagnetic) 

w = 0-125 rad/s, x0 = 0-4 mm 


oscillatory flow in baffled tubes 

Oscillating piston Baffled tube di = 12 cm, length =1.0 m, 55 orifice baffles Determination of energy dissipation 

Baird and Stonestreet (1995)) 

(external) 

f = 3-14 Hz, x0 = 1-6 mm 


Operation fluid: light oil 

15

16 



Oscillating piston Baffled tube d = 25 mm, length = 1.08 m, 28 cylindrical baffles Observations on the dispersion of fluid; local profile Mackley and Ni (1991) 

f = 0.5-9 Hz 

Operated vertically and horizontally 

measurements 

Reciprocating plates Reciprocating baffled-plate Column 1: di = 19.4 cm, H = 90 cm 

Power dissipation and flow patterns determined with Bair and Rao (Baird and Rao 

column 

column 

f = 0.6-3.0 Hz, x0 = 5-20 mm 

water 

1995) 

(x2) 

Column 2: di = 15.0 cm (nominal), H = 3.0 m 

f = 0.6-3.0 Hz, x0 = 1-10 mm 

Different plates distance and configuration 

Reciprocating plates Reciprocating baffled-plate di = 15.0 cm (nominal), H = 3.96 m 

Time-average power dissipation rates and hold-up Column description: Hafez and 

column 

column 

f = 2.0-5.0 Hz, x0 = 1-10 mm 

determination 

Baird (1978); Work: Baird et al 

(1996) 

Reciprocating plates 0.38 m diameter oscillatory di = 0,38 m, H = 2 m, Only 2 baffles 

Flow patterns and oil-water dispersion in a 0.38 m Ni et al. (2000b) 

column 

baffled column 

f= 0-1 Hz, x0 = 60-200 mm 


diameter OBC 

Reciprocating plates Novel self-aerating pilot di = 19 cm, H = 0.9 m 

Mass transfer measurements of self-aerating system Mackley et al. (1998) 

columns 

scale oscillating baffle f = 0.25-2 Hz, x0 = 0-4.2 mm 

for oxygenation of water 

Same vessel as Baird and Rao 

column 


(1995) 

Oscillating piston Oscillatory Baffled Batch di = 30 cm, L = 1.5d mm, d0 = 15 mm, δ = 2 mm Crystallization of paracetamol Chew and Ristic, Chew et al 

(electromagnetic) Crystallizer (OBBC) f = 1-20 Hz, x0 = 1-4 mm 


(2005; 2004a) 

Oscillating piston Oscillatory baffled column di = 50 mm, H = 1.5 m, d0 = 28 mm, δ = 3 mm, L 

= 1.5d 

f = 0.2-10 Hz, x0 = 1-10 mm 


Oxygen mass transfer rates Oliveira and Ni (2004) 

Oscillating piston Pulsed sieve plate column di = 39.6 mm, Length = 800 mm, L = 25, 50 or Analysis of axial dispersion in an oscillatory-flow Palma and Giudici (2003) 

(PSPC) 

100 mm, α = 22.3 % 

f = 0-4.5 Hz, x0 = 5-25 mm 

continuous reactor 

Oscillating pistons U-tube di = 24 mm, d0 = 12 mm, L = 1.5d 

Heat transfer performance Stephens and Mackley (2002) 

2 vertical-interconnected operated tubes 

Same geometry as Mackley and 

Stonestreet (1995)



Oscillating piston Pulsed baflled tubular di = 75 mm, Total length: 1,500 mm, L = 70 mm, Photooxidation of a model pollutant (salicylic acid) Gao et al.(2003) 

photochemical reactor δ = 122 mm, ratio of L/di = 1.41 

(PBTPR), 

high f and x0 UV lamp in its central axis 

Oscillating piston Oscillatory baffled column di = 50 cm, H = 0.5 m 

f = 0.5-10 Hz, x0 = 2-6 mm 


Effect of fluid viscosity on mixing in an OFR Fitch et al. (2005) 

17

2.3 The Oscillatory Flow Mixing (OFM) 

The mechanism of oscillatory flow mixing (OFM) can be understood with the help of Figure 2-5. The 

essential feature is that sharp edges are presented perpendicular to a periodic and fully reversing flow. 

The flow patterns of OFM exhibit a complicated eddy mixing pattern due to the presence of wall baffles. 

Two half cycles can be identified, each containing flow acceleration and deceleration, corresponding to a 

sinusoidal velocity-time function. On each acceleration, vortex rings are formed downstream of the baffles. 

A peak velocity is reached and then as the flow decelerates, the vortices are swept into the bulk, and 

consequently unravel with bulk flow acceleration in the opposite (axial) direction. It is the radial velocities, 

arising from the repeating cycles of vortex formation, and of similar magnitude of the axial ones, which 

create a uniform mixing in each inter-baffle zone and cumulatively along the length of the column (Brunold 

et al. 1989; Mackley and Ni 1991; Mackley and Ni 1993). 

18 

Figure 2-5. Mechanism of oscillatory flow mixing (OFM) in an OFR, according to Fitch et al. (2005). (A) 

Start of Up Stroke. (B) Maximum velocity in Up stroke, i.e. flow reversal. (C) Start of Down stroke. (D) 

Maximum velocity in Down stroke. 

The study of OFM within OFRs has steadily grown in the last decade. The areas of research now include 

several aspects related to OFR characterisation and applications. In recent years, the science of the OFR 


has increasingly been applied to various industrial processes, such as suspension polymerisation, 

crystallisation, paint dispersion, flocculation and fermentation. Several scientific articles and industrial 

applications demonstrate that OFR is an exciting type of reactor and can be a process technology with 

major commercial applications (Ni and Gough 1997). Table 2-4 compiles the fundamental studies on of 

OFR’s in the last decade. Several science aspects will be discussed with more detail in the forthcoming 

sections. 

The concepts and key developments of OFM enhancement through pulsation and oscillation are reviewed 

by Ni et al. (2003). This configuration can generate high heat and mass transfer rates in both batch and 

continuous modes of operation, whereas potential applications may include pipes, mixers, (bio)reactors, 

filtration units and crystallizers (Mackley 1991). 

2.3.1 Parameters governing the OFM 

The dynamical nature of OFM may be presently characterised by a few fundamental dimensionless 

groups, namely: the classical Reynolds number, Re n, the oscillatory Reynolds number, Re o, and the 

Strouhal number, St. In addition, two dimensionless geometrical parameters contribute to describe the 

fluid mechanics within OFRs: the interbaffle spacing defined as L/d i, and the baffle free area, α, defined 

as: d 0/d i (Ni and Pereira 2000). A brief definition of each dimensionless group is presented below. 

a) Net-flow Reynolds number, Re n 

In flow in pipes the Reynolds number, Re n, is the dimensionless number used as the indicator of the type 

of flow in question and captures all the parameters shown in Figure 2-6. 

Figure 2-6. The net flow in a plain tube. 


19

20 

Table 2-4: Summary of works concerning the fundamental study of OFM in OFR’s 

Science aspect of OFR Reference(s) 

Axial Dispersion / RTD’s Howes (1988), Howes and Mackley (1990), Ni et al. (2002b), Palma 

and Giudici (2003), Takriff and Masyithah (2002), Dickens et al. 

(1989), Mackley and Ni (1991; 1993), Ni (1994) 

Bioprocessing Ni et al. (1995a), Lee et al. (2002; 2001), Fabiyi and Skelton (1999; 

2000a; 2000b), Gao et al. (2003; 1998), Lee et al. (2002; 2001) 

Chemical reaction Ni and Mackley (1993) 

Crystallisation Chew et al. (2004a), Chew and Ristic (2005) 

Dispersion Mackley and Ni (1991; 1993), Ni (1995), Ni , (2000)Ni et al. 

(2002a),(2000) Ni and Pereira (Ni and Pereira 2000); Palma and 

Giudici (2003), Fitch and Ni (2003), Ni and Stevenson (1999), Ni et 

al., (1998a) 

Fluid mechanics Baird and Rao (1995), Ni (1994), Liu et al.(1995), Fitch et al. 

(2005), Gao et al. (2003), Mackley and Ni (1991; 1993), Ni et al. 

(1995b; 2000b), Gough et al.(1997), Brunold et al. (1989), Ni et al. 

(2002a), Chew et al.(2004b), Mackley et al. (1996), Gao et al. 

(1998) 

Gas-liquid systems Oliveira and Ni (2001), Oliveira et al. (2003a; 2003b), Hewgill et al. 

(1993), Baird et al. (1996), Mackley et al.(1998) 

Heat transfer Mackley et al. (1990); Mackley and Stonestreet (1995), Stephens 

and Mackley (2002) 

Liquid-liquid systems Hounslow and Ni (2004), Ni et al (1998b); Ni et al. (1999); Ni et al. 

(2002c); Ni et al. (2002b); Harvey et al. (2003), Zhang and Ni 

(1996), Pereira and Ni (2001) 

Mass transfer Hegwill et al. (1993); Ni et al. (1995a); Ni and Gao (1996) 

(Ni et al. 1995a), Oliveira and Ni (2004), Lau et al. (2004) 

Numerical simulations Howes (1988), Howes et al.(1991), Jian and Ni (2003), Roberts and 

Mackley (1996), Mackay et al. (1991); Chew et al.(2004b); Ni et al 

(2002a) 

Particle suspension Mackley et al.(1993); Liu et al. (1995) 

Power input Mackley and Stonestreet (1995), Baird and Stonestreet (1995), Baird 

and Rao (1995), Baird et al. (1996) 

Scale-up Ni and Gao (1996), Ni (2001) 

Fluid viscosity Fitch et al. (2005) 

The Reynolds number is defined as follows 

Re n 

u d 

= (2.1) 

υ 


where d is the tube diameter, υ the kinematic viscosity of fluid and u the mean superficial flow velocity. 

b) The oscillatory Reynolds number, Re o 

When an oscillatory motion is superimposed onto the net flow (Figure 2-7) an additional dimensionless 

group is often needed to characterise such a motion, in conjunction with the above defined Re n. 

Figure 2-7. Oscillatory motion superimposed onto a net flow. 

The characterisation of such a pure oscillatory flow (POF) can be backdated in the 1940’s (e.g. Binnie 

(1945)). Since then, oscillatory flow was studied by several tube arrangements (see Ni and Gough (1997), 

for references). In all the published works, the characterisation of POF was achieved by using a 

dimensionless group called the pulsating Reynolds number, Re p, defined as: 

Re 

p 

u p d 

= (2.2) 

υ 

where u p is the pulsating velocity. In most cases, u p was taken as the product of x 0w, Re p describes the 

oscillatory motion applied to the system, and Re n (as defined in Eq. (1)) gives a measure of the state of 

flow in question. However, other authors used different definitions for u p. Sarpkaya (1966), for example, 

defined it as the amplitude of the periodic component of the cross-sectional mean velocity 

(= piston tube A / A x f 0 

π ), where A piston and A pipe are the cross-sectional areas of the piston and tube, 

respectively. No reason was given why ‘πf’’ was used instead of ‘2πf’. Sinada and Karim (1984a; 1984b), 

used a different approach: they replaced u p by u and d by the Stokes layer thickness defined 

as δ' = ( 2υ 

/ w ) in Eq. (2.2), when working with a special application, using a fixed stroke length. 

The situation is more complex when an oscillatory motion is imposed into a net flow in the presence of 

baffles (Figure 2-8). 


21

22 

Figure 2-8. The oscillatory (baffled) flow. 

Following previous studies, Brunold et al. (1989) defined the first of the two dimensionless groups 

controlling the fluid mechanics of OFR: the oscillatory Reynolds number, Re o: 

Re o 

w x0 

d 

= (2.3) 

υ 

Re p and Re o for both POF and OFR are basically identical. However, they describe different states of flow 

since, at certain oscillatory conditions, the fluid mechanics in Figure 2-7 will predominately be axial, while 

in Figure 2-8 will be complex and chaotic with similar magnitudes for both axial and radial velocity 

components. 

Since the oscillator normally operates sinusoidally, the variations in time of displacement, x, velocity, v, 

and acceleration, a, take the forms of (Ni and Gough 1997; Ni et al. 2002a): 

( w t) 

x = x0 

sin 

(2.4) 

( w t) 

v = x0 

w cos 

(2.5) 

2 

( w t) 

a = −x 

w sin 

(2.6) 

0 

where w is the angular piston velocity and x 0 is the oscillation amplitude, measured as centre-to-peak. The 

maximum velocity during the oscillation cycle is ‘x 0w’, as seen in Eq. (5) when ‘cos (w t) = 1’. An example 

is given in Figure 2-9. 


x [m], v [m/s], a [m/s 2 ] 

6.E-03 

4.E-03 

2.E-03 

0.E+00 

-2.E-03 

-4.E-03 

-6.E-03 

0 1 2 3 4 5 6 7 8 9 10 

x0 

v 

a 


TIME [s] 

Figure 2-9. Exemplification of sinusoidal movement of a piston (displacement, x, velocity, v, and 

acceleration, a) for w = 0.62 rad/s (i.e., 0.1 Hz), and x 0 = 5 mm. 

From extensive studies, there is now a solid understanding of the mixing nature in an OFR. At low Re os of 

100-300, it exhibits plug flow characteristics: the vortices are axisymmetrically generated within each 

baffled cavity (plug flow mode). When Re o increases further, the symmetry is broken and flow becomes 

intensely mixed and chaotic; flow achieves the mixing mode, as defined by Ni et al. (1999; 2002b). 

c) The Strouhal number, St 

The description of POF develops further when tube inserts or varying tube shapes are incorporated. Sobey 

(1980) introduced another dimensionless number, apart from Re p, when working in a flow through a 

furrowed channel to account for the additional parameters involved. This was named the Strouhal number, 

St f: 

St f 

f h 

u peak 

= (2.7) 

23

where h is the half channel width and u peak the peak velocity at the maximum channel width, h max. The 

physical meaning of such dimensionless group was just given in Sobey’s later work of flow past an 

indentation in a channel (Sobey 1985) as the ratio of the channel length scale to the scale of the fluid 

particle displacement. Since then the characterisation of various structures in oscillatory flows has 

followed a similar line (e.g., Nishimura et al. (1985)). 

At the end of the 1980’s, Brunold et al. (1989) followed Sobey’s examples and definitions and reported 

the second dimensionless group to define the fluid mechanics in OFR’s, referring to it as the Strouhal 

number St: it represents a measure of the effective eddy propagation and is defined as the ratio of column 

diameter to the stroke length: 

d 

St = (2.8) 

4 π x 

24 

0 

This re-definition of St is actually the most used. In a simplified form, St represents the ratio of orifice 

diameter to oscillation amplitude (Ni and Gough 1997). 

2.3.2 The effect of geometrical parameters 

The recent advances in the OFR research have suggested the introduction of a term that involves either 

the orifice diameter (d 0) or the baffle spacing (L), since they play an important role in OFR and do not 

participate either in Re o or St numbers. For example, L influences the shape of eddies while d 0 controls the 

width of the vortices within each baffled cavity, either of which affects the onset of fluid mixing within OFR 

(Ni and Gough 1997). 

In the presence of sharp edges, Knott and Mackley (1980) and Brunold et al. (1989) have reported that 

eddies’ interaction is optimal for a baffle spacing of 60 % in a tube with a d i of 25 mm. Since these two 

studies, the effect of the geometrical parameters in OFRs was intensively explored, mainly weighted by the 

residence time and liquid-liquid dispersion characteristics, in single horizontal (e.g. Dickens et al. 1989), 

vertical (e.g. Mackley and Ni 1993) or an array of tubes (e.g. Pereira and Ni 2001). Reproducible and 

consistent results have shown that the introduction of OFM, coupled with periodically spaced baffles, 

greatly enhances fluid mixing even at laminar flow conditions. Each baffle cavity acts as a continuously 

stirred tank, in which the radial velocity components are comparable to the axial ones. The events at the 


walls are similar to the events at the centre. This resulted, for example, in a six-fold increase in the mass 

transfer of oxygen into water was reported for oscillatory flow in a baffled tube with an air-water system (Ni 

et al. 1995a). 

a) The effect of free baffle area, α 

Several studies considered the effect of α = d 0 2 /d 2 on mixing time or axial dispersion. Ni et al. (1998b) 

studied the effect of α (11 to 51 %), δ (1 to 48 mm) and L (d to 2.5d). The lowest value of α tested (11 %) 

exhibited the best mixing and, consequently, required shorter mixing times, presumably due to a higher 

power input, or by the increased mixing efficiency due to the higher dispersion rate. Gough et al. (1997) 

found L = 0.57d, α = 0.63 as the optimized sizes to achieve efficient mixing of a polymerization 

suspension (for d = 50 mm, L = 0.7 – 3.3d and d 0 = 0.51 – 0.69d tested). For the smallest orifice 

diameter tested (with a corresponding α = 0.26) small symmetrical eddies were formed at the sharp 

edges of the baffles and the vortex rings did not encompass the entire column cross-section, nor the 

complete length of the entire-baffle region. Thus stagnant regions between eddies were identified. For a 

higher α = 0.32, eddies extended to the reactor walls covering a greater area of the section. Vortex rings 

were still symmetrical along the centre line (axisymmetric) and displaying small interaction. At α = 0.40, 

the axisymmetry was lost and the intense interaction between eddies led to the disappearance of the 

stagnant regions within the baffled cavity, inducing characteristics of plug flow when in continuous 

operation. For the maximum α tested (0.47), a high degree of channelling through the baffle orifice was 

observed and the formation of eddies was destroyed by the predominant axial movement, thus low mixing 

could take place (Gough et al. 1997). 

Research on liquid-liquid dispersions by Zhang et al. (1996) was consistent with Gough’s observations. A 

minimum value of α tested (0.19) showed to be the most appropriated value for dispersion of liquid-liquid 

solutions, leading to the use of the lowest minimum f (on average) to achieve the complete dispersion. 

The rate of increase of the degree of oil-water dispersion with the oscillatory component of velocity is 

greater for lower values of d 0 than for higher ones defined before, as reported by Ni et al. (2000b). 

b) The effect of baffles spacing, L 

Several authors have suggested that different values of L may result into different flow behaviours. The 

baffle spacing is a key design parameter in an OFR as it influences the shape and length of eddies within 


25

each baffle cavity, for a given x 0 (e.g. Brunold et al. 1989; Knott and Mackley 1980). However, it is usually 

not included in the dimensionless groups in OFRs. Some authors are of the opinion that it should 

participate in the equations characterising the mixing in OFRs. Since in the usual flow regimes both L and 

d 0 are close enough to pipe diameter (L = 1 - 2d; (1-α 2 )/α 2 = 26 - 35 %), both parameters are not 

independent in the dimensionless groups (Ni and Gough 1997). Mackley et al. (1993) used a new 

dimensionless group called the Stroke ratio, intending to classify the flow in terms of the relation between 

x 0 and L. 

The optimal L should ensure a full expansion of vortex rings generated behind baffles so that vortices will 

spread effectively throughout the entire inter-baffle zone. At a small value of L, the generation of vortices is 

strongly suppressed. This effectively restrains the growth of the vortices and reduces the required radial 

motion within each baffled cell. Conversely, if the baffles are spaced too far apart, the opposite effect 

occurs. The vortices formed behind baffles cannot effectively cover the entire inter-baffle regions. In this 

case, it is most likely that stagnant plugs will be created, into which the vortices disperse and diminish. 

This demonstrates that vortex rings generation is not independent of L. Brunold et al. (1989) reported the 

optimal L as 1.5d for flow visualisation studies. However, L = 1.8d was suggested by Ni and Gao (1996) in 

their mass transfer studies. Ni et al. (1998a) reported that the maximum L/d ratio tested (2 to 2.5) is the 

one minimizing the mixing time. But L seems to have little effect on oil-water dispersions, as the relation 

L/d is linearly scaled-up, as repeated by Ni et al. (2000b). 

26 

c) The effect of baffle thickness, δ 

The generation of vortices in each baffle of an OFR is similar to that of vortices formed in a fluid flowing 

around an object. Each eddy needs an edge to cling on for and has an optimal time of processing of 

shedding (Ni and Gough 1997). As there should be an optimal δ, Ni et al. (1998b) also investigated the 

effect of δ on the mixing time, for top and bottom injection locations. Six values of δ (between 1 and 48 

mm, for a d of 50 mm) were tested. Mixing time decreased with the increase of f or x 0. Overall, the results 

suggest that the thinner baffles (i.e. low δ) favoured the generation of vortices. If vortices attach to baffle 

edges for too long prior to shedding, their shape can distort somewhat, thereby affecting mixing time. The 

higher values of δ resulted in higher mixing times, in the order of five-fold greater than those of the 

thinnest baffles. 


2.3.3 Effect of f and x 0 in the flow patterns 

The oscillation frequency (f) and amplitude (x 0) are the most important operational parameters in OFR. At a 

given L and d 0 changing the combination of f and x 0 allows control of the generation of eddies and 

produces a range of fluid mechanical conditions as broad as required, as reported by e.g. Gough et al 

(1997) from their work in application of polymerisation suspensions. Research reported by Zhang et al. 

(1996) on oil-water dispersions demonstrated that both x 0 and f have a significant effect on the minimum 

frequency for complete dispersion in liquid-liquid extraction processes; a 50 % reduction occurred when x 0 

was increased from 6 to 12 mm. Ni et al. (1998b) found that the mixing time decreased as f or x 0 

increased (valid for L = 1 - 2.5d) 

A similar work of Ni et al. (2000b) on oil-water dispersions in a scaled-up OFR (d i = 380 mm) illustrated 

that f and x 0 affect the nature of mixing much more than design parameters, such as d 0 and L. The degree 

of dispersion increased linearly with the oscillatory velocity until a complete dispersion is achieved. The 

oscillatory f and x 0 were also found to affect the mass transfer measurements (for wall baffles) in a yeast 

cell suspension (Ni et al. 1995c). The oxygen mass transfer coefficient, k La, increased with the increasing 

of f (from 3 to 12 Hz) for all the tested values of x 0 (4 to 14 mm), in a 25 mm internal diameter OFR. 

Changes in x 0 affected k La more than changes in the f, meaning that x 0 controls the length of eddy 

generated in the column. 

For some applications an optimum f or x0 may be identified. For example, Dickens et al. (1989) identified 

x0 = 1 mm as the minimum value for full axial dispersion in a pulsed packed bed. 

2.3.4 Power input 

There are essentially two models for estimation of the power consumption in an OFR: i) the quasi-steady 

flow model (Jealous and Johnson 1955), and ii) the eddy acoustic model (Baird and Stonestreet 1995). 

The quasi-steady flow model was originally derived for packed columns and subsequently used by Baird 

and Garstang (1967) for pulsed columns. This method is based on a quasi-steady assumption to calculate 

the pressure drop and power density for oscillating flow. By applying Bernoulli’s equation between two 

planes adjacent to a baffle, the pressure drop across the orifice plate can be obtained and an 

instantaneous power density can be calculated (Hewgill et al. 1993; Ni and Mackley 1993). By integrating 


27

this over a cycle and allowing for a number of orifice plates, it gives a time-averaged power density defined 

as (Ni et al., 1998b): 

28 

2 2 

D α 

2 

P 2 ρ N 1 − α 2 3 

= x0 

w 

(2.10) 

V 3π 

C 

where N is the number of baffles per unit length, ρ is the density of fluid and C D is the orifice discharge 

coefficient (usually equal to 0.7). For small values of α the term (1-α 2 )/α 2 increases, and Eq. (2.10) 

predicts high mixing intensity and a reduced mixing time. This suggests the existence of a threshold in the 

uniformity of mixing in OFRs. The decrease in α would have a similar effect on mixing time as the product 

‘f x 0’ is increased (Ni et al., 1998b). But the power input for this model is valid for high x 0 and low f, i.e. 5 - 

30 mm and 0.5 - 2 Hz (Baird and Stonestreet 1995). The eddy acoustic model (Baird and Stonestreet 

1995) is based on acoustic principles and uses the concept of eddy viscosity with reasonable accuracy. 

The power input for this model appears to be justified for conditions of low x 0 and high f, i.e. 1 - 5 mm, 3 - 

14 Hz, where the quasi-steady model was shown to be inappropriate for predicting the power dissipation 

of oscillatory flow (Baird and Stonestreet 1995). The eddy acoustic model relates the frictional resistance 

to the acoustic resistance of a single orifice in a thin plate and assumes that the eddy kinematic viscosity 

is a function of f and of a mixing length corresponding to the average distance travelled by turbulent 

eddies. Through several experiments, the mixing length was shown to be equal to the orifice diameter d 0. 

2.3.5 Numerical simulation 

Although technological applications of oscillatory flow to pursuit enhancements in unit operations have 

been reported since the early 1930s (Van Dick 1935), numerical simulations for oscillatory flow in baffled 

geometry were not cited until 1980. Sobey was perhaps the first one reporting his extensive 2-D numerical 

studies (Sobey 1980; 1983) followed by Ralph (1986), while in different situations. These studies revealed 

that the vortex mixing mechanism was the key factor responsible for high mixing efficiency of the system. 

Based on those works, Howes (1988) developed a numerical code for studying dispersion of unsteady flow 

in baffled tubes. Following on, Roberts (1992) extended Howes’ work to 2-D baffled channel flows. A solver 

based on finite difference axisymmetrical, time dependent Navier-Strokes equation plus a stream function 

and vorticity was used. Even with some assumptions such as a flow spatial periodicity, with flow in each 


cell being identical and the formation of axisymmetric vortices (Howes 1988),(Mackay et al. 1991); 

Roberts (1992), these models were successfully applied to many fields, namely to predict the onset of 

chaotic motions, and they evaluated concentration gradients by incorporating transport such as heat and 

mass transfer (Roberts and Mackley 1995) and provided fluid particle motion simulations (Neves-Saraiva 

1998) up to a critical Re o. 

The first numerical study taking d 0 into consideration into POF was done by Jones and Bajura (1991), by 

carrying out a numerical analysis on a pulsating laminar flow through a pipe orifice while considering two 

Reynolds numbers: the numerical Reynolds number, Re n, and the orifice Reynolds number, Re op. 

The flow characteristics of a POF are dominated by the axial velocity components, but thanks to the 

contribution of numerical studies, there is nowadays a good understanding of the nature of OFM. It is 

known that at low values of Re o of 100 - 300, the OFR exhibits plug flow characteristics, where the vortices 

are axisymmetrically generated within each baffled cavity (referred to as the plug flow mode). On the other 

hand, for high values of Re o, the symmetry condition is no more valid and flow becomes intensely mixed 

and chaotic (referred to as the mixing mode). Depending upon the column geometry and the viscosity of 

the fluid, these critical values of the oscillatory Reynolds number may vary. When the Re o number 

increases beyond such critical values, the generation of vortices is no longer axisymmetrical, as show in 

Figure 2-10. 

Figure 2-10. Particle flow pattern in a batch OFR. Tracer = pollen particles of 25 µm in diameter, bulk fluid 

= water, f = 2.5 Hz, x 0 = 6mm, d = 50 mm, L = 1.5d, α = 36 %, δ = 3 mm (from Ni et al. 2002a). 


29

Recently, Chew et al. (2004b) used the Computational Fluid Dynamics (CFD) technique to model spatial 

and temporal behaviour of flow patterns in an OFR (L = 48 mm, d i = 30 mm, d 0 = 15 mm). Large eddy 

simulation (LES) was found suitable for simulations of OFM at two combinations of f and x 0, respectively: 

10 Hz – 3 mm and 10 Hz – 5 mm. The volume-averaged shear rate was found to be of one order of 

magnitude larger than that of an impeller-driven stirred tank and a marked distinction between the 

temporal shear rate distributions was observed. The modelling also showed that particles in an OFR spend 

most of their residence time in high shear regions. 

The effect of fluid viscosity on OFM was qualitatively assed by numerical simulations (further validated with 

experimental measurements) by Fitch et al. (2005). A ratio of the plane-averaged axial over the radial 

velocity was defined to quantify such viscosity effects. For the given geometry the velocity ratio approached 

to 2 very quickly at increased the values of Re o, regardless of Newtonian and non-Newtonian fluids. An 

empirical critical value of velocity ratio equal to 3.5 was identified, below which the system mixed 

sufficiently. 

Jian and Ni (2003) tested the modelling of turbulence with the traditional Reynolds Averaged Navier-Stokes 

(RANS) model. Results are sufficiently good for simulating flows in stirred tank reactors but the RANS 

turbulence models showed a poor prediction of turbulence in periodic flows in an OFR as the methodology 

of averaging in time in RANS has effectively removed the turbulence. As in OFR eddies of various sizes are 

the main ingredient for mixing, the large-eddy simulation (LES) is particularly suitable for such type of 

flows (Jian and Ni 2003). 

Outside the OFR field, Komoda et al. (2001) carried out CFD simulations in a reciprocating disk cylindrical 

vessel. Simulations were experimental validated (with laser Doppler anemometry velocity measurements) 

and represented well the flow patterns and the force acting on the disk during the oscillation cycle. 

2.4 Further studies regarding oscillatory flow mixing 

In complement to many studies regarding the industrial application of OFM, many further studies were 

carried out in relation to the science of OFM and the effect of tube constrictions. A survey is presented in 

Table 2-5. Apart from the works in OFR, fundamental studies (fluid mechanics and numerical simulations) 

govern the major part of publications of OFM (e.g. Bolzon et al. 2003). Several authors also seek the 

30 


understanding of control of mixing/dispersion (e.g. Crittenden et al. 2005) or the science behind the 

enhancement of mass/heat transfer rates (Nishimura et al. 2000). More recently (since 2004), oscillatory 

flow was scaled-down to microfluidics applications (e.g. Morris and Forster 2004). 

2.5 Tools in reactor engineering 

Reactor engineering activity is related to the engineering of (chemical or biochemical) transformations. 

Such transformations can occur only if the reactant molecules are brought into short contact (mixed) 

under the appropriate environment (temperature and concentration fields, catalysts/biocatalysts) for and 

adequate time. The process vessel (reactor) must provide the necessary conditions to favour the desired 

reaction and allow for removal of products. To describe a reactor’s behaviour it is necessary to 

characterise it in terms of flow patterns and mixing, eventually for the different phases in presence. 

Recently, CFD tools appear to make a substantial contribution in establishing the best way to carry out a 

desired transformation, as on accelerating the reactor engineering tasks (Ranade 2002). 

2.5.1 Measuring techniques 

The description and design of multiphase (gas–liquid, gas–liquid–solid and gas-liquid-liquid-solid) reactors 

still relies to a large extent on empirical rules and correlations, which in turn are based on measurements 

made under conditions as relevant as possible to industrial practice. This is true for the classical chemical 

engineering approach, where such quantities as liquid hold-up (fraction) or pressure drop are predicted via 

empirical correlations based on data as numerous and precise as possible. Nevertheless, more modern 

approaches appeared in the last years to help in the design of multiphase reactors, such as CFD. Even in 

this case, the physical models used require information on local and transient flow characteristics (e.g. 

turbulence characteristics, wake coefficients, etc.), since ab initio calculations are up to now impossible. 

Reliable measuring techniques are therefore needed for the rational description and the design of 

multiphase reactors. Different types of measurements are required depending on the aim of the analysis. 

Measurement techniques can be classified according to different criteria. A first classification distinguishes 

between ‘time-averaged’ and ‘transient’ measurements and between ‘local’ and ‘global’ measurements. 


31

32 

Table 2-5: Relevant studies concerning the research of OFM and the effect of constrictions 

Main research subject Study description Reference 

Bioprocesses Oscillatory flow in a cone-and-plate bioreactor Chung et al. (Chung et 

al. 2005) 

Bioprocesses Differential responses of the Nrf2-Keap1 Hosoya et al.(Hosoya et 

system to laminar and oscillatory shear 

stresses in endothelial cells 

al. 2005) 

Bioprocesses Tissue factor activity is upregulated in human 

endothelial cells exposed to oscillatory shear 

stress 

Mazzolai et al.(2002) 

Bioprocesses An harmonic analysis of arterial blood 

pressure and flow pulses 

Voltairas et al.(2005) 

Dispersion/simulations Simulation of concentration dispersion in 

unsteady deflected flows 

Hwu et al. (1997) 

Dispersion Oscillatory flow and axial dispersion in packed 

beds of spheres 

Crittenden et al.(2005) 

Dispersion Effect of turbulence on Taylor dispersion for 

oscillatory flows 

Ye and Zhang (2002) 

Dispersion/diffusion Augmented longitudinal diffusion in grooved 

tubes for oscillatory flow 

Ye and Shimizu (2001) 

Fluid mechanics Linear stability analysis of flow in a periodically Adachi and Uehara 

grooved channel 

(2003) 

Fluid mechanics Birth of three-dimensionality in a pulsed jet 

through a circular orifice 

Bolzon et al.(2003) 

Fluid mechanics Asymmetric Flows and Instabilities in 

Symmetric Ducts with Sudden Expansions 

Cherdron et al (1978) 

Fluid mechanics Characterisation of impeller driven and OFM Chew et al.(2004b) 

Fluid mechanics Bifurcation phenomena in incompressible 

sudden expansion flows 

Drikakis (1997) 

Fluid mechanics Nonlinear Flow Phenomena in a Symmetric 

Sudden Expansion 

Fearn et al. (1990) 

Fluid mechanics Characteristics of laminar flow induced by 

reciprocating disk in cylindrical vessel 

Komoda et al.(2001) 

Fluid mechanics Instability in three-dimensional, unsteady, Mallinger and Drikakis 

stenotic flows 

(2002) 

Fluid mechanics 3-D analysis of the unidirectional oscillatory 

flow around a circular cylinder 

Nehari et al. (2004) 

Fluid mechanics Three-dimensionality of grooved channel flows Nishimura and 

at intermediate Reynolds numbers 

Kunitsugu (2001) 

Fluid mechanics Flow around a short horizontal bottom cylinder 

under steady and OFM 

Testik et al.(2005) 

Heat transfer Cooling of micro spots by OFM Chou et al. (2004) 

Heat transfer Convective heat transfer enhancement in a Herman and Kang 

grooved channel using cylindrical eddy 

promoters 

(2001) 

Heat transfer Effect of oscillating interface on heat transfer Chen et al. (1997) 



Main research subject Study description Reference 

Heat transfer Local heat transfer in the presence of a single 

baffle within a channel 

Chen and Chen (1998) 

Heat transfer The effects of gas-liquid interfacial movement 

on heat transfer using oscillations 

Chen et al. (1997) 

Heat transfer Effect of the distance between a single baffle 

and the solid wall on the local heat transfer in 

a rectangular channel due to an oscillatory 

flow 

Chen and Chen (1998) 

Mass transfer Enhancement of liquid phase adsorption 

column performance by means of oscillatory 

flow 

Lau et al. (2004) 

Mass transfer Oscillatory flow of droplets in straight capillary Graham and Higdon 

tubes 

(2000a) 

Mass transfer Oscillatory flow of droplets constricted in Graham and Higdon 

capillary tubes 

(2000b) 

Mass transfer A comparison between the enhanced mass Thomas and Narayanan 

transfer in boundary and pressure driven 

oscillatory flow 

(2002a) 

Mass transfer Influence of x0 and f on mass transfer 

enhancement of grooved channels 

Nishimura et al. (2000) 

Microfluidics DNA molecules in microfluidic oscillatory flow Chen et al. (2005) 

Microfluidics Oscillatory flow in microchannels Morris and Forster 

(2004) 

Microfluidics Numerical simulation of micromixing by 

pulsative micropump 

Kim et al (2003) 

Mixing Mixing performance by reciprocating disk in 

cylindrical vessel 

Komoda et al. (2000) 

Mixing/dispersion Interstage backmixing in oscillatory flow in a Takriff and Masyithah 

baffled column 

(2002) 

Mixing/Microfluidics Chaotic mixing in cross-channel micromixers Tabeling et al. (2004) 

Mixing/Numerical simulations Simulation of mixing in unsteady flow trough a Howes and Shardlow 

periodically square obstructed channel (1997) 

Particle suspension Influence of wall proximity on the lift and drag Fischer et al. (Fischer et 

of a particle in an oscillatory flow 

al. 2005) 

Rheology Vibrational flow of non-Newtonian fluids Deshpande and Barigou 

(2001) 

Rheology Viscous dissipation of a power law fluid in an Herrera-Velarde et al. 

oscillatory pipe flow 

(2001) 

Suspension Response of concentrated suspensions under 

large x0 oscillatory shear flow 

Narumi et al. (2005) 

Suspension The use of pulsative flow to separate species Thomas and Narayanan 

(2002b) 

Suspension/fluidisation Using pulsed flow to overcome defluidization Wang and Rhodes 

(2005) 


33

Since the classification between local and global measurements is not always possible other classification 

has been preferred by Boyer (2002), relying on the physical basis of the measurement, thus distinguishing 

between ‘invasive’ and ‘non-invasive’ measuring techniques as follows: 

34 

a) Non-invasive techniques 

(a) Global techniques 

i. Time-averaged pressure drop 

ii. Measurement and analysis of signal fluctuations 

iii. Dynamic gas disengagement technique (DGD) 

iv. Tracing techniques 

1. Tracing of the liquid 

2. Tracing of the gas-phase 

3. Tracing of the solid (coloured tracers, magnetic tracers, fluorescent 

tracers) 

v. Conductimetry 

vi. Radiation attenuation techniques 

1. X-ray, γ-ray or neutron absorption radiography 

2. Light attenuation 

3. Ultrasound techniques 

(b) Techniques yielding local characteristics 

b) Invasive techniques 

i. Visualisation techniques 

1. Photographic techniques 

2. Radiographic techniques 

3. Particle image velocimetry 

4. NMR imaging 

ii. Laser Doppler anemometry and derived techniques 

iii. Polarographic technique 

iv. Radioactive tracking of particles 

v. Tomographic techniques 

1. Tomography by photon attenuation measurement 

2. Electrical tomographic system 

3. Ultrasonic tomography 


(a) The so-called ‘needle probes’ (optical probes, resistive or conductive probes, or 

‘impedance probes’) 

(b) Heat transfer probes 

(c) Ultrasound probes 

vi. Ultra-sound transmittance technique (UTT) 

vii. Pulse echo technique 

(d) Pitot tubes. 

A detailed analysis of time and space resolution as well some examples of the use of measuring 

techniques with industrial constraints in the petrochemical and refinery industry is also presented by Boyer 

(2002). 

2.5.2 Flow visualisation by Particle Image Velocimetry 

The Particle Image Velocimetry (PIV) has become quite classical for the determination of velocity fields 

essentially in single-phase flow (e.g. Boyer et al. 2002). While large-scale turbulence structures have been 

recognised historically by fluid dynamicists as significant phenomena, most of today’s fluid dynamics 

measurements are made with point-based techniques. The PIV system, on the other hand, provides 

practical quantitative whole-field turbulence information and thus has the potential to give a new 

perspective on flow phenomena. The PIV measurement process usually involves (dantecdynamics 2002): 

a) Seeding the flow: seed particles are suspended in the fluid to trace the motion and give a 

visible reflection for the cameras. 

b) Flow field illumination: when a thin slice of the flow field is illuminated by a light-sheet (of laser 

light), the illuminated seeding scatters the light. This is detected by a camera placed at right 

angles to the light-sheet. The light-sheet is pulsed (switched on and off very quickly) twice at a 

known intervals (∆t) (Figure 2-11A). 

c) Image acquisition: the first pulse of the laser freezes images of the initial positions of the 

seeding particles (at time t) onto the first frame of the camera. The camera frame is advanced 

and the second frame of the camera is exposed to the light scattered by the particles from the 

second pulse of laser light (at time t + ∆t). There are thus two camera images, the first 

showing the initial positions of the seeding particles and the second their final positions after 

an interval of time equal to ∆t due to the movement of the flow field (Figure 2-11B). 


35

36 

d) Vector processing: the two camera frames are then processed to find the velocity vector map 

of the flow field. This involves dividing the camera frames into small areas called interrogation 

regions. In each interrogation region, the displacement of groups of particles between frame 1 

and frame 2 (∆x) is measured using correlation techniques. The velocity vector, v, of this area 

in the flow field is then calculated using the equation 

∆x 

v = S 

(2.11) 

∆t 

where S is the object to image scale factor between the camera’s CCD chip and the measurement 

area (Figure 2-11C). 

This is repeated for each interrogation region to build up the complete (2-D) velocity vector map. 

Figure 2-11. Overview of PIV technique. (A) Schematic representation of the flow field illumination in a PIV 

system. (B) PIV interrogation analysis. (C) Evaluation of the image density. Only build up of 2-D velocity 

vector maps is exemplified (adapted from dantecdynamics 2002). 


The PIV technique has been successfully used in the study of fluid mechanics within an OFR. The first 

study of this kind was performed by Ni et al. (1995b), thus demonstrating that it is possible to directly 

measure velocity vector fields and strain-rate distributions in an OFR using time-resolved PIV. It also 

allowed finding a correlation between the strain rate and the power dissipation generated within OFRs, as 

seen in Ni et al. (2000a). More recently, Fitch et al (2005) used the PIV technique to validate CFD 

simulations and concerning the effect of fluid viscosity on mixing in a OFR: Gao et al. (2003) used PIV 

measurements to assist in obtaining the design optimum oscillatory flow conditions for catalyst dispersion 

(in photochemical oxidation of organic compounds) whilst avoiding the possible side effects of strong 

scattering or reduction of quantum yield. The PIV studies showed that uniform mixing can be readily 

achieved at low Re o (i.e. at Re o above 2,000). 

2.5.3 Assessment of the non-ideal flow 

The most extensively used concept in reactor engineering is that of an ‘ideal’ reactor. The simplest 

reactor, whose performance is governed by the so-called ‘zero dimensional’ equation, is the ‘completely 

mixed reactor’. The key assumption is that mixing in the reactor is complete, so that the properties of the 

reaction mixture are uniform in all parts of the reactor and are, therefore, the same as those of the ‘exit’ 

stream. The other ideal reactor concept, known as ‘plug flow reactor’ is based on a ‘one dimensional’ 

approximation of the material and energy balance equations. In an ideal plug flow reactor, unidirectional 

flow through the reactor is assumed (similar to the flow through a pipe) (Ranade 2002). 

It is of extreme importance to evaluate the consequences of the assumptions involved in the concepts of 

ideal reactors to estimate the behaviour of an actual reactor, as the mixing may deviate significantly from 

the ideal flow cases. This deviation can be caused e.g. by channelling of fluid, by recycling of fluid or by 

the formation of stagnant regions within the reactor (e.g. Levenspiel 1972). The mixing of a phase may be 

experimentally characterised by tracing techniques (e.g. Boyer et al. 2002). 

The residence time distribution (RTD) is an important concept used for analysis of reaction engineering 

with idealised models. RTD, as the name suggests, indicates the spread of residence time experienced by 

different fluid elements while flowing through the reactor. The response data or measurements of the 

variation of reactor outlet concentration of a substance for the known change of inlet concentration of that 

substance can be used to estimate the RTD of a given reactor. The completely segregated (assuming no 


37

mixing between fluid elements of different ages) and completely mixed fluid elements constitute the two 

limiting solutions. Obtaining the RTD of an actual reactor and applying these two limiting assumptions to 

obtain the bounds of the performance of the reactor is a practical method for reaction engineering analysis 

(Ranade 2002). RTD affects heat transfer rates, interphase mass transfer rates and the conversion and 

selectivity of chemical and biochemical reactions (Briens et al. 1995). 

Several sophisticated techniques and data analysis methodologies have been developed to measure the 

RTD of reactors. Measuring the RTD of a tracer dissolved in the liquid phase is a well-known technique to 

evaluate the mixing of the liquid phase. This technique is easy to apply but may present some pitfalls, as 

demonstrated by (Briens et al. 1995). Main tracer types are (Boyer et al. 2002): 

38 

a) Tracer dissolved in the liquid phase, e.g.: 

(a) Salt tracer 

(b) Coloured tracers 

(c) Radioactive isotope tracer 

b) Particle tracking technique, i.e. neutrally buoyant solid particles followed by electromagnetic 

means. 

Various different types of models have been developed to interpret RTD data (tracer concentration versus 

time) and to use it further to predict the influence of non-ideal behaviour on reactor performances. Most of 

these models use ideal reactors as building blocks. In simple case, a two-parameter model (the mean 

residence time and the axial dispersion coefficient) may be sufficient to yield an adequate description of 

the global flow behaviour of a reactor: In more complex cases, models with more parameters have to be 

used (Levenspiel 1972). A flow model representing the actual flow patterns and mixing within a reactor is 

necessary for the realistic description of reactor behaviour (Ranade 2002). 

Another important issue in RTD studies is the physical boundaries of the reactor in study: closed or open 

type. When the flow patterns are disturbed across a boundary (e.g., a measurement point), such boundary 

is classified as being open. If flow patterns are not disturbed along the boundary it is classified as closed. 

Most academic and practically all industrial tracer studies are conducted with open boundary conditions, 

using the "imperfect pulse method". A pulse of tracer is injected upstream of the reactor and the resulting 

tracer concentration peaks are detected at two different locations in the reactor. Then, the residence time 

distribution between these locations is obtained by deconvolution (Briens et al. 1995). 


Care must be taken when measuring the RTDs in reactors with one or two open boundaries. In such cases 

tracer measurements do not provide RTD but a ‘transient response function’ from which RTD may only be 

obtained if separate experiments provide more information (Nauman and Buffham 1983). The 

measurement of the tracer concentration can be performed by three different techniques (Briens et al. 

1995): 

a) mixing-cup 

b) local concentration (e.g. as the measured by an effective fibre-optical probe or a conductivity 

probe) 

c) through-the-wall, along a diameter of a cross-section (e.g. conductivity meters or scintillation 

systems). 

Only the mixing-cup concentration provides the true RTD (Nauman and Buffham 1983). 

A second important concept in reactor engineering analysis, mainly for batch operating vessels, is the 

‘mixing time’. This is briefly the time required to reach a specified degree of uniformity the system being 

then said to be ‘mixed’. Practical mixing times can be measured by a variety of experimental tracing 

techniques, similarly to those applied to obtain RTDs (Harnby 1992): 

a) acid/base/indicator reactions 

b) electrical conductivity variations 

c) temperature variations 

d) refractive index variations 

e) light-absorption techniques. 

In each case it is necessary to specify the manner of tracer addition, the position and number of recording 

points, the sample volume of the detection system, and the criterion for deciding the cut-off point of the 

end of the experiment (Harnby 1992). An example of how to determine the mixing times in a process 

vessel used in biopharmaceutical manufacturing is presented by Ram et al. (2000). In such case an acid 

reaction was monitored by pH probes. 

Studies in OFRs have shown that OFM coupled to a net flow (of the correct magnitude) gives high fluid 

mixing and narrow residence time distribution (e.g. Dickens et al. 1989; Howes and Mackley 1990; 

Mackley and Ni 1991; Mackley and Ni 1993). The baffle edges promote the formation of eddies, which 

increase the radial mixing in the tube (Ni and Pereira 2000). 


39

2.5.4 Computational flow modelling 

Computational Fluid Dynamics (CFD) is an engineering-numerical tool which has gained large popularity 

during the last years. As opposed to the semi-empirical models (e.g. those use for modelling of RTDs), 

CFD aims at solving the (complete or simplified) fundamental physical equations that describe a flow 

phenomenon. The most general form of these equations has been given by Navier and Stokes more than 

150 years ago, therefore the set of equations that has been applied are named Navier-Stokes equations. 

These equations encompass mass, momentum and energy balances; they have to be adapted to the 

specific problem under consideration by additional closure laws. Also the subsidiary sets of reaction 

equations can be used in case of having reacting species. 

While CFD has been very popular among car manufacturers and in the air and space industry, chemical 

engineers have only recently become aware of the large potential it bears for the development and 

improvement of process equipment. This is mainly due to the fact that with modelling flow around a car 

body or an airplane wing, only single-phase flow has to be considered while in most applications in 

chemical reactors two- and three-phase flows are common. This poses a wealth of new questions and 

brings about serious difficulties in modelling and numerics. 

CFD simulations did bring some advances to move forward in the numerical simulations of OFRs in 

comparison to previous works. The stream function approach (Howes et al. 1991) was abandoned and the 

3-D Navier–Stokes equations are solved directly, as described below. 

40 

a) Model equations 

In most of CFD packages (e.g. Fluent 5 – Fluent Inc., Paris, France) the governing equations are solved in 

cylindrical coordinates, as follows (Ni et al. 2002a): 

Momentum equations: 

⎛ 

⎜ ∂ 

ρ 

⎜ 

⎝ 

∂t 

∂V 

∂r 

V ∂ 

+ θ Vr 

V 

− θ + V 

r ∂θ 

r 

∂V 

⎞ 

r ⎟ ∂p 

⎡1 

∂ 1 ∂τ 

∂ ⎤ 

= − − ⎢ ( r ) + r τ τ 

τ 

θ − θθ + 

∂ ⎟ 

rr 

z ∂ 

⎥ 

⎠ 

r ⎣r 

∂r 

r ∂ r ∂z 

⎦ 

2 

Vr + V r 

rz 

r 

z 

⎛ ∂Vθ 

∂V 

V ∂V 

V V ∂V 

⎞ ∂p 

⎡ ∂ 

∂ ∂ ⎤ 

⎜ + V θ + θ θ − r θ + V θ 1 1 2 1 τ 

⎟ = − − ⎢ ( r ) + θθ τ 

ρ 

− z 

r 

z 

τ 

θ 

rθ 

∂ 

⎥ 

⎝ ∂t 

∂r 

r ∂θ 

r ∂z 

⎠ r θ 2 

⎣r 

∂r 

r ∂θ 

∂z 

⎦ 

N. Reis Novel Oscillatory Flow Reactors for Biotechnological Applications 

(2.12) 

(2.13)

⎛ ∂ 

ρ⎜ 

⎝ ∂t 

∂V 

∂r 

V ∂ 

+ θ Vz 

+ V 

r ∂θ 

∂Vz 

⎞ ∂p 

⎡1 

∂ 1 ∂τ 

∂ ⎤ 

⎟ = − − ⎢ ( r ) + z τ 

τ 

θ 

rz − 

∂z 

∂ 

⎥ 

⎠ z ⎣r 

∂r 

r ∂θ 

∂z 

⎦ 

Vz + V z 

zz 

r 

z 

Continuity equations: 

1 ∂ 

r ∂r 

where 

( rV ) + + z = 0 

r 


(2.14) 

1 ∂Vθ 

∂V 

(2.15) 

r ∂θ 

∂z 

⎡ ∂V 

⎤ 

= − r 2 

τ rr µ ⎢2 

− ( ∇V 

) ⎥⎦ 

(2.16) 

⎣ ∂r 

3 

⎡ ⎡1 

∂V 

V ⎤ ⎤ 

= − ⎢ ⎢ + r 2 

τ µ 2 θ 

θθ ⎥ − ( ∇V 

) ⎥⎦ 

(2.17) 

⎣ ⎣r 

∂θ 

r ⎦ 3 

⎡ ∂V 

⎤ 

= − z 2 

τ zz µ ⎢2 

− ( ∇V 

) ⎥⎦ 

(2.18) 

⎣ ∂z 

3 

⎡1 

∂V 

∂ ⎛V 

⎞⎤ 

τ = = − r 

⎢ + r θ 

rθ 

τθr 

µ 

⎜ ⎟⎥ 

⎣r 

∂θ 

∂r 

⎝ r ⎠⎦ 

⎡1 

∂V 

∂V 

⎤ 

τ = = − z 

⎢ + θ 

zθ 

τθz 

µ 

⎥ 

⎣r 

∂θ 

∂z 

⎦ 

(2.19) 

(2.20) 

⎡1 

∂V 

∂V 

⎤ 

τ = = − z 

⎢ + r 

zr τ rz µ 

⎥ 

(2.21) 

⎣r 

∂r 

∂z 

⎦ 

where 

1 ∂ 

r ∂r 

( ∇V 

) = ( rV ) 

r 

1 ∂V 

∂V 

+ θ + 

r ∂θ 

∂z 

z 

(2.22) 

where V r, Vθ and V z are the fluid velocities (m/s) at r, θ and z coordinates respectively, p is the pressure 

drop (Pa). The viscous term in Eq.s (2.16) - (2.21) takes the form of µ = µ 0 + µ t, where µ 0 is the nominal 

laminar viscosity (kg m -1 s -1 ) and µ t the turbulent viscosity (kg m -1 s -1 ). For laminar flow simulation, µ t, = 0, 

and V r, Vθ and V z are the laminar velocity components. For turbulence simulation, µ t, is included and V r, Vθ 

and V z are averaged velocities. For 2-D simulations, all variables in the third direction (θ) are treated as 

constants, thus simplifying the above equations accordingly. 

41

Fluent (Fluent Inc., Paris, France) is one of the CFD software products commercially available in the 

market. It solves numerically the Navier-Stokes equations to find the flow pattern in the reactor. Three 

main steps are involved in numerical simulations with Fluent: 

42 

a) designing the geometry & meshing (descritisation of domain into finite elements) 

b) defining the fluid properties 

c) boundary conditions. 

For complex geometries, its designing and meshing are usually performed in a Fluent’s CAD tool, the 

‘Gambit’ software package. 

2.6 Biotechnological process engineering 

In 1989, the European Federation of Biotechnology proposed, in General Assembly, the following 

definition: “Biotechnology is the integration of natural and engineering sciences in order to achieve the 

application of organisms, cells, parts thereof and molecular analogues for products and services” (EFB 

General Assembly, 1989). 

Contrary to its name, Biotechnology is not a simple technology. Rather, it is a group of technologies that 

share two things in common: they manipulate living cells and their molecules, and they have a wide range 

of practical uses that can improve our lives. Simply defined, then, Biotechnology is a collection of scientific 

techniques that use living cells and their molecules to make products or solve problems (ncbiotech 2002). 

2.6.1 Application areas 

The applications of biotechnology are so broad, and the advantages so compelling, that virtually every 

industry is using the technology. Several examples are listed in Table 2-6. Biotechnology is enabling these 

industries to make new or better products, often with greater speed, efficiency and flexibility. The 

consumers are beginning to see the benefits in the foods they eat, the clothes they wear, the medicines 

they take, and the environment they live in, etc (ncbiotech 2002). 


Table 2-6: Some of the applications of Biotechnology (Lee 1984) 

Application field Main products 

Pharmaceuticals Antibiotics, antigens, diagnostics, endorphin, gamma globulin, human 

growth hormone, human serum albumin, immune regulators, insulin, 

interferon, interleukins, lymphokines, monoclonal antibody, neuroactive 

peptides, tissue plasminogen activator, vaccines, etc. 

Animal agriculture Products similar to those being developed in the pharmaceutical industry; 

development of disease-free seed stocks and healthier, higher-yielding food 

animals 

Plant agriculture Transfer of stress-, herbicide-, and pest-resistance traits to important crop 

species; development of plants with the increased abilities of 

photosynthesis or nitrogen fixation; development of biological insecticides 

and nonice nucleating bacterium 

Specialty chemicals Amino acids, enzymes, vitamins, lipids, hydroxylated aromatics, and 

biopolymers 

Agricultural chemicals Pesticides, fungicides, herbicides 

Environmental applications Mineral leaching, metal concentration, pollution control, toxic waste 

degradation, and enhanced oil recovery 

Foods and beverages Alcoholic beverages, sweeteners, single-cell protein 

Commodity chemicals Acetic acid, acetone, butanol, ethanol, and many other products from 

biomass processes 

Bioelectronics Biosensors, biochips 

2.6.2 Bioreactors and bioprocesses 

The commercialisation of biotechnology developments requires the scale-up of biological processes. To 

successfully design biological reactors (bioreactors) it is demanding to understand the bioprocesses 

mechanism/kinetics. There are several factors affecting the performance of a bioprocess and, in 

consequence, the operation of a biological reactor. They can be grouped in three systems, such as 

physical, chemical and biological properties, as listed in Figure 2-12 (Vaidyanathan et al. 1999). A 

complex network of interactions might exist in a bioprocess. 


43

44 

Figure 2-12. Factors that influence the performance of a bioprocess and the complexity of interactions 

between them. Only some interactions are shown for illustrative purposes. The factors are grouped under 

three system properties, namely, physical, chemical and biological (adapted from Vaidyanathan et al 

(1999)). 

The bioreactor is the ‘heart’ of biological processes and basically must display the following settings 

(Blenke 1985): 

a) a well-defined spatial distribution of all components (i.e., a good mixing, no concentration 

gradients) 

b) a good dispersion of all phases (gaseous, liquid and solid) 

c) avoid cell damage 

d) a high heat transfer rate 


e) an easy design and construction of high dimension bioreactors (volumes up to 100 m 3 ) at low 

construction cost 

f) easy operation: good sterility and possibility of keeping sterile conditions, low mechanical 

management, low power requirements and possibility to operate on high volume reactors 

g) easy set of operation conditions on a high range of temperature, concentration, viscosity, etc., 

(batch mode), or flexibility in production, (for continuous mode) 

h) design and operation performance must be easily appointed and proper to scale-up. 

As mentioned earlier in this text, mixing is of paramount importance in the bio/chemical process industry 

as it is determinant on heat/mass transfer, reaction performance and product uniformity. Engineers often 

require reactors with well defined residence times and good fluid mixing, while also searching for devices 

that exhibit near plug-flow behaviours, in some cases. 

Despite of all these issues, bioreactor scale-up may be indeed the biggest challenge in biotechnology due 

to mixing, oxygen transfer and shear stress restrictions. These parameters are often interrelated. Aeration 

should be as low as possible to avoid excessive shear stress, but must also ensure adequate oxygenation 

of the cells. Cells are delicate and their culture and processing invariably exposes them to intense 

hydrodynamic forces at some stage. A sufficiently intense force will destroy cells outright, while forces of 

lower magnitude may induce various physiological responses, without necessarily causing any obvious 

physical damage (Chisti 2001). Nowadays, little is known about shear fields in bioreactors but it is 

definitely desirable to know the maximum shear rates (usually near the walls) rather than averaged values, 

as this can be critical for the application of a reactor to a biotechnological process. 

An attempt to classify biological reactors in two main groups according to the source of power and the 

degree of homogeneity was made by the Working Party of Bioreactor Performance of the European 

Federation of Biotechnology (Crueger 1987). While doing that, authors noticed that many of the bioreactor 

designs attempt to keep the whole of their volume homogenous. Nevertheless, in the most stirred 

volumes, for example, total homogeneity becomes progressively more difficult to archive as the scale 

increases (Cabral and Tramper 1993). 

Biological reactor types may be summarised in a few number of classes (see Table 2-7). Some innovative 

designs appeared in the last decade. Most innovations addressed either oxygen transfer, shear induced by 

stirring, control of water activity in organic phase systems or waste biotreatment. An extensive review is 

presented by Deshussest et al. (1997). 


45

46 

Table 2-7: Summary of the main features of reactor classes (Cabral et al. 2001) 

Biological reactor type Main features 

Stirred tank, ST Cylindrical vessel, equipped with a stirrer, baffles 

and aeration 

Continuous flow stirred-tank reactor, CSTR A refined design of the stirred tank (provided with 

ports for inlet and outlet) 

Packed-bed Vertical mounted settled bed of particles, 

continuous, upward or downwards feeding 

Fluidized reactor Upward fluid feeding; flow rate must assure 

fluidisation of bed of particles 

Bubble column reactor, BC Attractive alternative to stirred reactor for aerobic 

processes, continuous or batch operation 

Air-lift loop reactor, ALR Similar to BC, but where the hydrodynamic flow 

pattern is well described and controllable 

Novel reactor designs Mainly at laboratory scale, e.g.: membrane and 

liquid-impelled loop reactor; liable for scale-up; 

integrated in downstream process 

For a long time, stirred tank (ST) was the most used reactor for chemical applications and for aerobic 

fermentations (Chisti 1989). While it still being the most used reactor in industrial applications, the 

growing attention on processes development at industrial scale brought into evidence that ST is not the 

most suitable for microorganisms’ culture. Several reasons exist for this statement, namely: cell damage is 

very intense, essentially due the high shear stress caused by the stirrer; sterility is very difficult to assure; 

they present low energetic efficiency, often involving heat removal by temperature control; usually high 

construction cost (Chisti 1989). 

The knowledge of such disadvantages of STs, namely the issue of excess shear stress and low energetic 

efficiency has fomented the investigation of other types of reactors, namely oscillating bioreactors. 

Recently, new reactor designs have been developed, but further development of innovative bioreactors 

remains a high priority, as a single bioreactor configuration will never provide a universal solution 

(Deshusses et al. 1997). 

2.6.3 Bioreactor engineering 

Producing more, faster, with higher yields and more reliably have been the main driving forces behind the 

evolution in bioreactor designs. It has been pointed out that mixing, oxygen transfer and shear stress 


emain the biggest challenges as far as the scale-up to industrial size bioreactors is concerned. These 

parameters are generally linked, and compromises need to be made, for instance, in aeration to avoid 

excessive shear stress. The latest developments in bioreactors for better mixing, oxygen transfer and lower 

shear stress are reviewed by Deshussest et al. (1997) 

One of the particularities of biotransformations is their polyphasic composition (gas-liquid-solid or gas- 

liquid-liquid-solid). Consequently, the mass transfer of nutrients (carbon and energy sources, organic 

nitrogen and oxygen) is more complex than for chemical processes thus controlling the performance of 

bioreactors (Galaction et al. 2004). 

a) Oxygen mass transfer rates 

The oxygen supply constitutes one of the decisive factors in submerged microbial cultures and can play an 

important role in the scale-up and economy of aerobic biosynthesis systems. The aeration efficiency 

depends on oxygen solubilisation and diffusion rate into the liquid-phase. The amount of dissolved oxygen 

in a culture is limited by its solubility and mass transfer rate, as well as by its consumption rate by cells’ 

metabolic pathways (Galaction et al. 2004). 

The oxygen mass transfer can be described and analyzed by means of the mass transfer coefficient, k La. It 

represents the most important parameter implied on the design and operation of mixing–sparging 

equipment of bioreactors. The correct measurement and estimation of k La is a crucial step in the design 

procedure of the bioreactors (Puthli et al. 2005). The k La values are affected by several factors, such as 

geometrical and operational characteristics of the vessels, media composition, type, concentration and 

microorganisms morphology, biocatalyst properties (particle size, porosity, etc.) (e.g. Chisti and Jauregui- 

Haza 2002). 

Numerous mathematical correlations have been proposed for k La, either as functions of adimensional 

groups, such as 

( Re, Sc,... 

) 

Sh = f 

(2.23) 

or using specific mechanical power input and superficial air velocity: 

⎛ P ⎞ 

kL a = f ⎜ ⎟ 

⎜ 

, vs 

,... 

⎟ 

(2.24) 

⎝ V ⎠ 


47

The second relation is preferred, being more useful in practical applications or for fermentation scale-up 

using oxygen mass transfer efficiency criteria. 

A survey of measurement techniques to assess k La in bioreactors is presented by Gogate and Pandit 

(1999). Main techniques may be summarised as follows: 

48 

a) Dynamic methods 

(a) Dynamic oxygen electrode method 

(b) Start-Up method 

b) Steady state sulphite method 

c) Dynamic pressure method (DPM) 

d) Peroxide method 

e) Response methods 

Several drawbacks and errors (up to 100 %) may be associated with any of these methods. Depending on 

the range of the variables i.e. (P/V) and v g (superficial gas velocity), the most appropriate method needs to 

be chosen (Gogate and Pandit 1999). 

2.6.4 Bioprocesses monitoring 

Bioprocesses monitoring is crucial in bioreactors operation. Preferably, on-line and real-time information of 

bioprocesses kinetics should be obtained with the final aim of process full control. Measurements of 

bioprocesses may occur at different levels, as presented in Figure 2-13. For many years, the approach to 

the measurement of non-physical variables (as shown in Figure 2-12) has been to perform those 

measurements outside and away from the reactor (‘off-line’), principally due to a lack of appropriate 

technologies with which to obtain the values directly from the reactor. However, for some years now 

measurement techniques have been applied increasingly ‘by’ (at-line) and where possible ‘on’ (on-line), 

and even ‘in’ (in situ) the bioreactor vessel or flow stream (Vaidyanathan et al. 1999). Ideally, in situ 

approaches are desirable. 


Figure 2-13. A schematic of the approaches to measurement in bioprocesses (adapted from Vaidyanathan 

et al., (1999). 

Measurement of physical variables, such as temperature, pressure, agitation speed, and flow rates, are 

not considered here as they can today be reliably made in situ, provided that appropriate maintenance 

schedules are maintained. Basically, the chemical and biological variables can be the measured using 

(Vaidyanathan et al. 1999): 

a) Optical Sensors 

(a) Light Absorption/Scattering Measurements 

(b) Fluorescence Measurements 

(c) Vibrational Spectroscopy 

(d) Image Analysis 

(e) Flow Cytometry 

b) Flow-Injection Analysis — Biosensor Systems 

(a) Flow-Injection Analysis (FIA) 

(b) Biosensors 

c) Chromatography 

d) Mass Spectrometry 

e) Dielectric Spectroscopy 

f) Nuclear Magnetic Resonance (NMR) Spectroscopy 

g) Calorimetry 

Further techniques (essentially ‘off-line’) have found specific utility: steric sedimentation field-flow 

fractionation, electrophoresis, etc. 


49

2.6.5 Continuous cultures 

Microbiological processes have been largely developed through batch-processing methods, i.e. one batch 

of material is completely processed in a given vessel before the next batch is started. This situation arose 

partly because of the operational problems involved (e.g. aseptic operation and the maintenance of a 

particular strain of microorganisms). The complex relationships between substrate consumption, microbial 

growth and product formation are also significant factors. One of the advantages of batchwise operation is 

that the capital cost is less then for a continuous process and, for this reason, it is frequently favoured for 

new and untried processes, which may be converted to continuous operation at a more advanced stage of 

development (Atkinson 1974). 

These factors are slowly being overcome as the tendency to large-scale continuous process inevitably 

continues. Examples of applications are illustrated by developments in beer production (Branyik et al. 

2002) and in the production of cellular material for use as protein (Jung 2006), or enzyme recovery 

(Papamichael and Hustedt 1994). The reasons why continuous processes are eventually adopted in 

almost all large-scale operations are (Atkinson 1974) 

50 

a) diminished labour costs (repeated filling and emptying operations of batch vessels are 

eliminated; in situ medium sterilisation) 

b) ease of application of automatic control to continuous processes (also leading to the reduction 

of labour costs) 

c) more stable bioreactor conditions, and hence greater steadiness in the quality of product 

(product recovery is also facilitated) 

d) homogeneous environmental conditions, leading to a reduced range of by-products 

e) products produced only during a very brief transient growth phase can only be produced in 

quantity in continuous mode under well-defined conditions 

f) a steady load on the services required by the process, e.g. air and steam. 

In fact, the continuous culture experiments offer a number of advantages over the conventional batch 

method. Batch cultures are traditionally used in biological experiments because of easy handling. But 

many often the interpretation of the results in batch culture is difficult because concentrations of 

substrates and products change constantly, pH varies, and osmotic pressure and redox potentials change 

(Nielsen and Villadsen 1992). In continuous culture, under steady-state conditions, the environment is 

well-controlled and defined and the results obtained (e.g. kinetic parameters and yield coefficients) may be 


more reliable and reproducible (Sipkema et al. 1998). Therefore, the cause/effect relationships are more 

easily determined in a continuous culture than in batch cultures. Continuous fermentation experiments 

can provide details and valuable information about a biological system and certainly it is the option for 

determining specific characteristics that are difficult to observe with non-continuous culture techniques. 

For example, pulse and medium shift experiments (Fiechter et al. 1981; Sipkema et al. 1998; Zhang and 

Greasham 1999) and accelerostat operation (Paalme et al. 1995) can be used in continuous fermentation 

for the preparation and the optimization of the chemical and physical environment to which an organism is 

exposed. 

2.6.6 Biotechnological applications of OFM 

Good aeration, mixing and mass transfer are important for biotechnological processes which require an 

efficient and adequate supply of oxygen to aerobic microorganisms. There are many different designs and 

methods to obtain gas dispersion. Some devices are quiescent, such as bubble columns and trickle beds. 

Others employ dynamic (mechanical) agitation, such as gas sparged stirred tanks widely used at industrial 

scale (Linek et al. 1991; Schugerl 1982), multiple impeller vessels (Linek et al. 1996), cascade reactors 

with rotating or axially moving mixing elements, and mechanical surface aerators. 

All such devices use constant “steady” mixing, such as superficial velocities or fixed agitation speed. The 

use of OFM is an alternative, with the relative periodic motion of fluid (usually, sinusoidal). OFM is found to 

significantly enhance mass transfer rates in bubble columns (Hewgill et al. 1993) and more efficient than 

mechanical (stirred) agitation with respect to gas hold-up (Baird and Garstang 1967). 

The development of high efficiency bioreactors has been an important research objective in the field of 

bioprocesses. Appropriate selection and design could greatly improve the efficiency of the overall process. 

Several bioreactor configurations (fixed/fluidized-bed, gas-lift, membrane fermentors, reciprocating 

bioreactors) have been considered (Chamy et al. 1990; Chisti 1989; Mehaia and Cheryan 1984). In many 

cases, gas-lift, fluidized and reciprocating bioreactors is better suited to particular applications (Brauer 

1991; Gilson and Thomas 1993). 

Many reports concerning the successful application of OFM to bioengineering can be found in the 

literature. Several fermentations and enzymatic processes where improved with fluid oscillations (gas or 


51

liquid) either by preventing operational problems or by facilitating the improvement of efficiency and 

control of multiphase bioreactors. Enhancement of mass transfer rate using pulsation has been achieved 

by Baird and Garstang (1967), applying f from 1.09 to 1.35 Hz and x 0 up to 9.4 mm, to a 76 mm 

diameter column packed with random rings of 12.5 mm. The introduction of pulsation gave a three-fold 

increase in gas hold-up. Bellhouse et al. (1973) used oscillations in furrowed channels to enhance blood 

oxygenation. Serieys et al. (1978) also reported that in a reciprocating column with perforated plates the 

gas hold-up was slightly higher than a turbine agitator, but lower than with airlift bioreactors. However, the 

k La values obtained were much higher than those published for any other technology. Beeton et al. (1991) 

applied fluid oscillations to a membrane and achieved at least a five-fold enhancement in mass transfer 

over flat membranes. Mass transfer of oxygen into water was reported for OFM in a baffled tube (Hewgill 

et al. 1993), and a six-fold increase in k La was measured as compared with those for a bubble column. 

Measurement of k La into yeast re-suspension and yeast cultures (Ni et al. 1995a) revealed on average a 

75 % increase in k La values in a OFR over those obtained for a ST bioreactor, explained by the better shear 

rate distribution inside the vessel, leading to averaged thinner liquid films (hence increasing the k L term). 

Another application of OFM is of consideration. The potentially lethal bubble break-up at the gas-liquid 

interface was minimized by the development of a vortex wave membrane bioreactor by Millward et al 

(1996). The vortex wave generates a very effective mixing under laminar flow conditions by generating, 

expanding and transporting vortices in an oscillatory flow field (Millward et al. 1996). Significant mass 

transfer enhancement has been achieved under laminar flow conditions, without a major increase in 

power dissipation. The low shear rate indicated that such vortex wave design may be an effective 

alternative to conventional bioreactors for shear-sensitive systems. 

2.7 Scale-down of bioprocesses 

During the development of a microbial cell cultivation process there are four key stages (Steven et al. 

2004), as represented in Figure 2-14. Throughout a development process, many native and modified cell- 

lines are created and many operating conditions are considered and therefore large numbers (>100) of 

experiments are usually performed (Chartrain et al. 2000). Since development time is precious to 

commercial success, approaches that increase the rate at which these experiments can be carried out are 

52 


of great value and therefore high throughput (HTP) screening methodologies are of increasing interest (Lye 

et al. 2003). The elements in an ideal high throughput approach are: 

a) experiments can be performed in parallel 

b) experiments can be operated at a small-scale 

c) experiments can be automated (or online monitored). 

Strain 

selection 

Figure 2-14. Main stages crossing the bioprocess development. 

The application of small-scale reactors to the early stages of the bioprocesses can effectively contribute to 

the integration of biocatalyst, medium and bioprocess designs (Weuster-Botz et al. 2005). For such 

reasons, the micro-scale processing techniques are rapidly emerging as a means to increase the speed of 

bioprocess design and reduce material requirements (Lye et al. 2003). 

Biocatalysis is also a key technology in the synthesis of optically pure fine chemicals and pharmaceuticals 

(Schmid et al. 2001). Drugs developed with the incorporation of biocatalytic steps in their syntheses are 

now involved in the treatment of diseases such as HIV, heart disease, cancer, diabetes, flu and bacterial 

infections including tuberculosis (McCoy 1999). More than 150 industrial bioconversion processes are 

currently in operation or have been used for the manufacture of kilogram quantities of materials (Liese et 

al. 2000). The implementation of a new bioconversion process requires careful consideration of several 

competing biocatalyst and process options. Biocatalyst and process decisions has historically been 

collected at the 1.5 – 2.0 l scale, which can be time consuming and often requires significant quantities of 

expensive synthetic substrates. 

Strain 

enhancement 

The most commonly used cultivation vessel in process development is the shaken flask (Buchs 2001; 

Maier and Buchs 2001). Erlenmeyer flasks (100 – 2000 ml), filled with low volumes of medium (10 – 25 

% of the total capacity) are shaken to promote mixing and mass transfer via surface aeration. 

Unfortunately shaken flasks cannot be easily automated and the number of simultaneous experiments is 


Process 

optimization 

Scale-up 

53

limited to several tens. Thus, recently several authors presented alternative designs to the shaken flask. A 

critical review is presented by Lye (2003). Also some HTP screening systems are commercially available, 

but limited to a small number (10 - 20) of bio-transformations in parallel (Figure 2-15). 

54 

Figure 2-15. Examples of commercially available HTP screening bioreactor systems. (A) Infors Profors – 

16 x 400mL, sparged column reactors. (B) DasGIP Fedbatch-pro – 16 x 300mL stirred tank reactors. (C) 

Infors Sixfors – 6 x 500mL, stirred tank reactors. 


Characterisation of the engineering environment in a small-scale system may be complex. The generation 

of quantitative process design data at the micro-litre scale first requires an understanding of the underlying 

mixing and mass transfer phenomena. The establishment of key parameters, such as k La, is necessary to 

enable scale-up, which also implies the study of the influence of well design and methods of agitation or 

aeration. On the other side, the lower scale requires the development of miniaturised techniques (e.g. 

Gernot T. John 2003). 

2.8 Conclusions 

The OFM is increasingly finding more applications and is now introduced as ‘a technology ready to deliver’ 

(Harvey and Stonestreet 2001). Since the 1930’s many authors were reporting the enhancement of 

chemical processes by operating under OFM conditions. The nuclear industry was the first one beneficing 

with the OFM following Van Dijck’s work (1935). 

The several studies with OFRs (essentially from the 1990’s) brought a deep knowledge of the OFM nature. 

CFD simulation tools developed in the last years offered the change to successfully predict the fluid 

mechanics within OFRs. The linear scale-up of lab-scale OFRs was successfully demonstrated as well as 

its enhanced capacity to deal with multiphase systems. 

Previous numerical simulations of OFM had some problems of validation (e.g. Howes 1988) due to the 

inexistence of appropriate experimental techniques. Thus, the simulation work in OFR has been at 

standstill since the mid 1990s. In recent years, novel high-resolution techniques appear as a validation 

tool, such as the digital Particle Image Velocimetry (PIV) technique. It is now possible to continue and 

extend the previous numerical research in this field with quantitative validation (Ni et al. 2002a). 

Although several micro-bioreactor designs are found in literature, few of them support a continuous 

operation. This is a gap that needs to be repaired. One exception is the work of Akgun (2004). It is 

basically a 250 ml shake flask provided with two inlet ports (one for gas supply and another for medium 

inlet) and one combined outlet on the side of the flask for the exhaust of gas and culture liquid, thus 

supporting the continuous growth. 

Continuous culture experiments with conventional fermentation technology (e.g. ST bioreactors) are very 

time- and material-consuming, and laboratory setup is complex. It is clear the challenge opportunity for a 


55

iochemical engineer in the design of scale-down platforms supporting the application of HTP screening in 

a continuous operation mode. 

The last decade brought several studies on multiphase systems, with promising results achieved in terms 

of mass transfer rates and particle suspension in OFRs, suggesting that the OFR could be a successful 

biological vessel. On the other hand, a scale-down of OFR (to less than 10 millilitres scale) was never tried 

before. The high demand of bioprocesses for reactor engineering asks for a systematic study and 

characterisation of OFRs for application to biotechnological processes. 

Several areas for novel reactor designs based on OFM technology may be identified in early stages of 

biotechnological processes development. The enhanced fluid mixing, heat and mass transfer rates 

highlight an opportunity for applications of such novel oscillatory reactor designs e.g. at the strain selection 

stage, allowing the parallel screening of strains and fermentation media. Batch HTP screening is allowed, 

assuring essentially small operation volumes and reduced reagent costs and waste generation. But the 

narrow RTDs found in OFRs forecast a chance to develop novel OFR configurations suiting the continuous 

process optimization, allowing keeping the same environment conditions (essentially the fluid mechanics). 

Anticipating the scale-up of OFRs, such novel scale-down reactor designs may complement the OFR and, 

together with the metabolic and genetic engineering work, concretise the two novel concepts in bioprocess 

development: integration and intensification. 

2.9 References 

Adachi T, Uehara H. 2003. Linear stability analysis of flow in a periodically grooved channel. International 

56 

Journal for Numerical Methods in Fluids 41(6):601-613. 

Akgun A, Maier B, Preis D, Roth B, Klingelhofer R, Buchs J. 2004. A novel parallel shaken bioreactor 

system for continuous operation. Biotechnol Prog 20(6):1718-24. 

Atkinson B. 1974. Biochemical reactors. London: Pion. 267 p. 

Bains W. 1998. Biotechnology from A to Z. Oxford: Oxford University Press. 411 p. 

Baird MHI, Garstang JH. 1967. Power Consumption and Gas Hold-up in a Pulsed Column. Chemical 

Engineering Science 22(12):1663-&. 


Baird MHI, Rao NVR. 1988. Characteristics of a Countercurrent Reciprocating Plate Bubble Column. 2. 

Axial Mixing and Mass-Transfer. Canadian Journal of Chemical Engineering 66(2):222-231. 

Baird MHI, Rao NVR. 1995. Power Dissipation and Flow Patterns in Reciprocating Baffle-Plate Columns. 

Canadian Journal of Chemical Engineering 73(4):417-425. 

Baird MHI, Rao NVR, Stonestreet P. 1996. Power dissipation and holdup in a gassed reciprocating baffle- 

plate column. Chemical Engineering Research & Design 74(A4):463-470. 

Baird MHI, Stonestreet P. 1995. Energy-Dissipation in Oscillatory Flow within a Baffled Tube. Chemical 

Engineering Research & Design 73(A5):503-511. 

Baltar C; 1972. Device for the generation of pulsing phenomena in fluids. Spain patent 44. 

Beeton S, Millward HR, Bellhouse BJ, Nicholson AM, Jenkins N, Knowles CJ. 1991. Gas Transfer 

Characteristics of a Novel Membrane Bioreactor. Biotechnology and Bioengineering 38(10):1233- 

1238. 

Begand S, Dahm B, Ambrosius S. 1998. The pulsed reactor for processing in the chemical industry. 

Chemical Engineering & Technology 21(5):420-424. 

Bellhouse BJ; National Research Development Corporation, London England, assignee. 1978 Feb 21. 

Method for Effecting Heat or Mass Transfer. USA patent 4, 075,091. 

Bellhouse BJ, Bellhouse FH, Curl CM, Macmilla TI, Gunning AJ, Spratt EH, Macmurra SB, Nelems JM. 

1973. High-Efficiency Membrane Oxygenator and Pulsatile Pumping System, and Its Application to 

Animal Trials. Transactions American Society for Artificial Internal Organs 19:72-79. 

Betts JI, Doig SD, Baganz F. The characterisation and application of a novel miniature stirred bioreactor 

for the scale down of industrial relevant microbial fermentations. In: Meerman HJ, editor. 

Proceeding of the Biochemical Engineering XIV; 2005; Harrison Hot Springs, BC, Canada. 

Engineering Conferences International. p 170. 

Binnie AM. 1945. A double-refraction method of detecting turbulence in liquids. Proceedings of the 

Physical Society 57(5):390. 


57

Blenke EH. 1985. Biochemical loop reactors. In: Reed G, editor. Biotechnology. Weinheim: VCH 

58 

Verlagsgesellschaft. p 465-517. 

Bolzon G, Zovatto L, Pedrizzetti G. 2003. Birth of three-dimensionality in a pulsed jet through a circular 

orifice. Journal of Fluid Mechanics 493:209-218. 

Boyer C, Duquenne AM, Wild G. 2002. Measuring techniques in gas-liquid and gas-liquid-solid reactors. 

Chemical Engineering Science 57(16):3185-3215. 

Branyik T, Vicente A, Cruz JM, Teixeira J. 2002. Continuous primary beer fermentation with brewing yeast 

immobilized on spent grains. Journal of the Institute of Brewing 108(4):410-415. 

Brauer H. 1991. Growth of Fungi and Bacteria in the Reciprocating Jet Bioreactor. Bioprocess Engineering 

6(1-2):1-15. 

Brauer H, Sucker D. 1978. Waste-Water Purification in a High-Performance Reactor. Chemie Ingenieur 

Technik 50(11):876-877. 

Briens CL, Margaritis A, Wild G. 1995. A New Stochastic-Model and Measurement Errors in Residence 

Time Distributions of Multiphase Reactors. Chemical Engineering Science 50(2):279-287. 

Brunold CR, Hunns JCB, Mackley MR, Thompson JW. 1989. Experimental-Observations on Flow Patterns 

and Energy-Losses for Oscillatory Flow in Ducts Containing Sharp Edges. Chemical Engineering 

Science 44(5):1227-1244. 

Buchs J. 2001. Introduction to advantages and problems of shaken cultures. Biochemical Engineering 

Journal 7(2):91-98. 

Cabral JMS, Mota M, Tramper J, editors. 2001. Multiphase Bioreactor Design. London: Taylor & Francis. 

528 p. 

Cabral JMS, Tramper J, editors. 1993. Bioreactor design: Harwood Academics Publishers, GmbH. 

Chamy R, Nunez MJ, Lema JM. 1990. Optimization of the Hardening Treatment of Saccharomyces- 

Cerevisiae Bioparticles. Enzyme and Microbial Technology 12(10):749-754. 


Chartrain M, Salmon PM, Robinson DK, Buckland BC. 2000. Metabolic engineering and directed evolution 

for the production of pharmaceuticals. Current Opinion in Biotechnology 11(2):209-214. 

Chen YL, Graham MD, de Pablo JJ, Jo K, Schwartz DC. 2005. DNA molecules in microfluidic oscillatory 

flow. Macromolecules 38(15):6680-6687. 

Chen ZD, Chen JJJ. 1998. Local heat transfer for oscillatory flow in the presence of a single baffle within a 

channel. Chemical Engineering Science 53(17):3177-3180. 

Chen ZD, Chen XD, Chen JJJ. 1997. Effects of an oscillating interface on heat transfer. Chemical 

Engineering Science 52(19):3265-3275. 

Cherdron W, Durst F, Whitelaw JH. 1978. Asymmetric Flows and Instabilities in Symmetric Ducts with 

Sudden Expansions. Journal of Fluid Mechanics 84(JAN):13-&. 

Chew CM, Ristic RI. 2005. Crystallization by oscillatory and conventional mixing at constant power density. 

Aiche Journal 51(5):1576-1579. 

Chew CM, Ristic RI, Dennehy RD, De Yoreo JJ. 2004a. Crystallization of paracetamol under oscillatory flow 

mixing conditions. Crystal Growth & Design 4(5):1045-1052. 

Chew CM, Ristic RI, Reynolds GK, Ooi RC. 2004b. Characterisation of impeller driven and oscillatory 

mixing by spatial and temporal shear rate distributions. Chemical Engineering Science 59(7):1557- 

1568. 

Chisti Y. 1989. Airlift Reactors - Design and Diversity. Chemical Engineer-London(457):41-43. 

Chisti Y. 2001. Hydrodynamic damage to animal cells. Critical Reviews in Biotechnology 21(2):67-110. 

Chisti Y, Jauregui-Haza UJ. 2002. Oxygen transfer and mixing in mechanically agitated airlift bioreactors. 

Biochemical Engineering Journal 10(2):143-153. 

Chou FC, Weng JG, Tien CL. 2004. Cooling of micro spots by oscillatory flows. Journal of Mechanics 

20(4):267-271. 

Chung CA, Tzou MR, Ho RW. 2005. Oscillatory flow in a cone-and-plate bioreactor. Journal of 

Biomechanical Engineering-Transactions of the Asme 127(4):601-610. 


59

Crittenden BD, Lau A, Brinkmann T, Field RW. 2005. Oscillatory flow and axial dispersion in packed beds 

60 

of spheres. Chemical Engineering Science 60(1):111-122. 

Crueger W. 1987. Physical Aspects of Bioreactor Performance. Frankfurt: DECHEMA on behalf of the 

European Federation of Biotechnology. 159 p. 

dantecdynamics. 2002. Measurement Principles of PIV. 

http://www.dantecdynamics.com/piv/Princip/Index.htm. 

Deshpande NS, Barigou M. 2001. Vibrational flow of non-Newtonian fluids. Chemical Engineering Science 

56(12):3845-3853. 

Deshusses MA, Chen W, Mulchandani A, Dunn IJ. 1997. Innovative bioreactors. Current Opinion in 

Biotechnology 8(2):165-168. 

Dickens AW, Mackley MR, Williams HR. 1989. Experimental Residence Time Distribution Measurements 

for Unsteady-Flow in Baffled Tubes. Chemical Engineering Science 44(7):1471-1479. 

Drikakis D. 1997. Bifurcation phenomena in incompressible sudden expansion flows. Physics of Fluids 

9(1):76-87. 

Etzold M, Stadlbauer EA. 1990. Design and Operation of Pulsed Anaerobic Digesters. Bioprocess 

Engineering 5(1):7-12. 

Fabiyi ME, Skelton RL. 1999. The application of oscillatory flow mixing to photocatalytic wet oxidation. 

Journal of Photochemistry and Photobiology a-Chemistry 129(1-2):17-24. 

Fabiyi ME, Skelton RL. 2000a. On the effect of pulsing conditions on reaction rates in a pulsed baffled 

tube photoreactor (PBTPR). Process Safety and Environmental Protection 78(B5):399-404. 

Fabiyi ME, Skelton RL. 2000b. Photocatalytic mineralisation of methylene blue using buoyant TiO2-coated 

polystyrene beads. Journal of Photochemistry and Photobiology a-Chemistry 132(1-2):121-128. 

Fearn RM, Mullin T, Cliffe KA. 1990. Nonlinear Flow Phenomena in a Symmetric Sudden Expansion. 

Journal of Fluid Mechanics 211:595-608. 


Fiechter A, Fuhrmann GF, Kappeli O. 1981. Regulation of Glucose-Metabolism in Growing Yeast-Cells. 

Advances in Microbial Physiology 22:123-183. 

Finnigan SM, Howell JA. 1989. The Effect of Pulsatile Flow on Ultrafiltration Fluxes in a Baffled Tubular 

Membrane System. Chemical Engineering Research & Design 67(3):278-282. 

Fischer PF, Leaf GK, Restrepo JM. 2005. Influence of wall proximity on the lift and drag of a particle in an 

oscillatory flow. Journal of Fluids Engineering-Transactions of the Asme 127(3):583-594. 

Fitch AW, Jian HB, Ni XW. 2005. An investigation of the effect of viscosity on mixing in an oscillatory 

baffled column using digital particle image velocimetry and computational fluid dynamics 

simulation. Chemical Engineering Journal 112(1-3):197-210. 

Fitch AW, Ni X. 2003. On the determination of axial dispersion coefficient in a batch oscillatory baffled 

column using laser induced fluorescence. Chemical Engineering Journal 92(1-3):243-253. 

Galaction AI, Cascaval D, Oniscu C, Turnea M. 2004. Prediction of oxygen mass transfer coefficients in 

stirred bioreactors for bacteria, yeasts and fungus broths. Biochemical Engineering Journal 

20(1):85-94. 

Gao P, Ching WH, Herrmann M, Chan CK, Yue PL. 2003. Photooxidation of a model pollutant in an 

oscillatory flow reactor with baffles. Chemical Engineering Science 58(3-6):1013-1020. 

Gao S, Ni X, Cumming RH, Greated CA, Norman P. 1998. Experimental investigation of bentonite 

flocculation in a batch oscillatory baffled column. Separation Science and Technology 33(14):2143- 

2157. 

Gernot T. John IK, Christoph Wittmann, Elmar Heinzle,. 2003. Integrated optical sensing of dissolved 

oxygen in microtiter plates: A novel tool for microbial cultivation. Biotechnology and Bioengineering 

81(7):829-836. 

Gilson CD, Thomas A. 1993. A Novel Fluidized-Bed Bioreactor for Fermentation of Glucose to Ethanol 

Using Alginate Immobilized Yeast. Biotechnology Techniques 7(5):397-400. 

Goebel JC, Fortuin JMH. 1986. Solids Holdup in a Pulsed Packed-Column Used for Solid Liquid 

Contacting. Chemical Engineering Science 41(10):2531-2539. 


61

Gogate PR, Pandit AB. 1999. Survey of measurement techniques for gas-liquid mass transfer coefficient in 

62 

bioreactors. Biochemical Engineering Journal 4(1):7-15. 

Golding JA, Lee J. 1981. Recovery and Separation of Cobalt and Nickel in a Pulsed Sieve-Plate Extraction 

Column. Industrial & Engineering Chemistry Process Design and Development 20(2):256-261. 

Gomaa HG, Al Taweel AM. 2005. Axial mixing in a novel pilot scale gas-liquid reciprocating plate column. 

Chemical Engineering and Processing 44(12):1285-1295. 

Gough P, Ni XW, Symes KC. 1997. Experimental flow visualisation in a modified pulsed baffled reactor. 

Journal of Chemical Technology and Biotechnology 69(3):321-328. 

Graham DR, Higdon JJL. 2000a. Oscillatory flow of droplets in capillary tubes. Part 1. Straight tubes. 


Graham DR, Higdon JJL. 2000b. Oscillatory flow of droplets in capillary tubes. Part 2. Constricted tubes. 


Hafez MM, Baird MHI. 1978. Power-Consumption in a Reciprocating Plate Extraction Column. 

Transactions of the Institution of Chemical Engineers 56(4):229-238. 

Harnby N. 1992. Mixing in the Process Industries. Ltd B-H, editor. Oxford, UK: Butterworth-Heinemann 

Ltd. 392 p. 

Harrison STL, Mackley MR. 1992. A Pulsatile Flow Bioreactor. Chemical Engineering Science 47(2):490- 

493. 

Harvey A, Stonestreet P. 2001. Oscillatory flow: a technology ready to deliver. Chemical Engineer- 

London(720):41-41. 

Harvey AP, Mackley MR, Seliger T. 2003. Process intensification of biodiesel production using a 

continuous oscillatory flow reactor. Journal of Chemical Technology and Biotechnology 78(2-3):338- 

341. 

Harvey AP, Mackley MR, Stonestreet P. 2001. Operation and optimization of an oscillatory flow continuous 

reactor. Industrial & Engineering Chemistry Research 40(23):5371-5377. 


Herman C, Kang E. 2001. An experimental study of convective heat transfer enhancement in a grooved 

channel using cylindrical eddy promoters. Journal of Enhanced Heat Transfer 8(6):353-371. 

Herrera-Velarde JR, Zenit R, Mena B. 2001. Viscous dissipation of a power law fluid in an oscillatory pipe 

flow. Revista Mexicana De Fisica 47(4):351-356. 

Hewgill MR, Mackley MR, Pandit AB, Pannu SS. 1993. Enhancement of Gas-Liquid Mass-Transfer Using 

Oscillatory Flow in a Baffled Tube. Chemical Engineering Science 48(4):799-809. 

Hosoya T, Maruyama A, Kang MI, Kawatani Y, Shibata T, Uchida K, Itoh K, Yamamoto M. 2005. 

Differential responses of the Nrf2-Keap1 system to laminar and oscillatory shear stresses in 

endothelial cells. Journal of Biological Chemistry 280(29):27244-27250. 

Hounslow MJ, Ni X. 2004. Population balance modelling of droplet coalescence and break-up in an 

oscillatory baffled reactor. Chemical Engineering Science 59(4):819-828. 

Howes T. 1988. On the Dispersion of Unsteady Flow in Baffled Tubes [PhD Dissertation]. Cambridge, UK: 

Cambridge University. 

Howes T, Mackley MR. 1990. Experimental Axial-Dispersion for Oscillatory Flow through a Baffled Tube. 


Howes T, Mackley MR, Roberts EPL. 1991. The Simulation of Chaotic Mixing and Dispersion for Periodic 

Flows in Baffled Channels. Chemical Engineering Science 46(7):1669-1677. 

Howes T, Shardlow PJ. 1997. Simulation of mixing in unsteady flow through a periodically obstructed 

channel. Chemical Engineering Science 52(7):1215-1225. 

Hwu TH, Sobey IJ, Bellhouse BJ. 1997. Simulation of concentration dispersion in unsteady deflected 

flows. Chemical Engineering Science 52(19):3233-3242. 

Jealous AC, Johnson HF. 1955. Power Requirements for Pulse Generation in Pulse Columns. Industrial 

and Engineering Chemistry 47(6):1168-1169. 

Jian HB, Ni XW. 2003. On modelling turbulent flow in an oscillatory baffled column - RANS model or large- 

eddy simulation? Journal of Chemical Technology and Biotechnology 78(2-3):321-325. 


63

Jones EH, Bajura RA. 1991. A Numerical-Analysis of Pulsating Laminar-Flow through a Pipe Orifice. 

64 

Journal of Fluids Engineering-Transactions of the Asme 113(2):199-205. 

Jung KH. 2006. Continuous production of recombinant interferon-alpha in Escherichia coli via the 

derepression of trp promoter using casamino acid. Process Biochemistry 41(4):809-814. 

Kim MC, Kim S, Park JS, Park HD. 2003. Numerical simulation of micromixing by pulsatile micropump. 

Journal of Industrial and Engineering Chemistry 9(5):602-606. 

Knott GF, Mackley MR. 1980. Eddy Motions near Plates and Ducts, Induced by Water-Waves and Periodic 

Flows. Philosophical Transactions of the Royal Society of London Series a-Mathematical Physical 

and Engineering Sciences 294(1412):599-&. 

Komoda Y, Inoue Y, Hirata Y. 2000. Mixing performance by reciprocating disk in cylindrical vessel. Journal 

of Chemical Engineering of Japan 33(6):879-885. 

Komoda Y, Inoue Y, Hirata Y. 2001. Characteristics of laminar flow induced by reciprocating disk in 

cylindrical vessel. Journal of Chemical Engineering of Japan 34(7):919-928. 

Kudra T, Gawrzynski Z, Glaser R; Her Majesty the Queen in right of Canada, as represented by the 

Minister, assignee. 1999. Pulsed fluidised bed. USA patent 5,918,569. 

Lau A, Crittenden BD, Field RW. 2004. Enhancement of liquid phase adsorption column performance by 

means of oscillatory flow: an experimental study. Separation and Purification Technology 35(2):113- 

124. 

Lee C-K, Fan C-H. 1999. Pressure swuing reactor for converting suspension of solid substrate p- 

hydroxyphenylhydantoin by immobilized D-hydantoinase. Bioprocess Engineering 21:341-347. 

Lee CT, Buswell AM, Middelberg APJ. 2002. The influence of mixing on lysozyme renaturation during 

refolding in an oscillatory flow and a stirred-tank reactor. Chemical Engineering Science 

57(10):1679-1684. 

Lee CT, Mackley MR, Stonestreet P, Middelberg APJ. 2001. Protein refolding in an oscillatory flow reactor. 

Biotechnology Letters 23(22):1899-1901. 

Lee JM. 1984. Biochemical Engineering. Oxford: Prentice Hall. xxv,1136p p. 


Lee JM. 1992. Biochemical Engineering: Prentice Hall International, NJ. 

Lema JM, Roca E, Sanromán A, Nunez MJ, Moreira MT, Feijoo G. 2001. Pulsing bioreactors. In: Tramper 

J, editor. Multiphase Bioreactor Design. First ed. London: Taylor & Francis. p 309-330. 

Levenspiel O. 1972. Chemical reaction engineering. New York: John Wiley & Sons, Inc. 

Liese A, Seelbach K, Wandrey C, editors. 2000. Weinheim: Wiley-VCH. 423p. p. 

Linek V, Moucha T, Sinkule J. 1996. Gas-liquid mass transfer in vessels stirred with multiple impellers. 1. 

Gas-liquid mass transfer characteristics in individual stages. Chemical Engineering Science 

51(12):3203-3212. 

Linek V, Sinkule J, Benes P. 1991. Critical-Assessment of Gassing-in Methods for Measuring KLa in 

Fermenters. Biotechnology and Bioengineering 38(4):323-330. 

Liu S, Ni X, Greated CA, Fryer PJ. 1995. Measurements of Velocities of Single Particles for Steady and 

Oscillatory Flows in Plain and Baffled Tubes. Chemical Engineering Research & Design 73(A6):727- 

732. 

Lye GJ, Ayazi-Shamlou P, Baganz F, Dalby PA, Woodley JM. 2003. Accelerated design of bioconversion 

processes using automated microscale processing techniques. Trends Biotechnol 21(1):29-37. 

Mackay ME, Mackley MR, Wang Y. 1991. Oscillatory Flow within Tubes Containing Wall or Central Baffles. 


Mackley M. 2003. Process and product innovation. Journal of Chemical Technology and Biotechnology 

78(2-3):94-97. 

Mackley MR. 1991. Process Innovation Using Oscillatory Flow within Baffled Tubes. Chemical Engineering 

Research & Design 69(3):197-199. 

Mackley MR, Ni X. 1991. Mixing and Dispersion in a Baffled Tube for Steady Laminar and Pulsatile Flow. 


Mackley MR, Ni X. 1993. Experimental Fluid Dispersion Measurements in Periodic Baffled Tube Arrays. 



65

Mackley MR, Smith KB, Wise NP. 1993. The Mixing and Separation of Particle Suspensions Using 

66 

Oscillatory Flow in Baffled Tubes. Chemical Engineering Research & Design 71(A6):649-656. 

Mackley MR, Stonestreet P. 1995. Heat-Transfer and Associated Energy-Dissipation for Oscillatory Flow in 

Baffled Tubes. Chemical Engineering Science 50(14):2211-2224. 

Mackley MR, Stonestreet P, Roberts EPL, Ni X. 1996. Residence time distribution enhancement in 

reactors using oscillatory flow. Chemical Engineering Research & Design 74(A5):541-545. 

Mackley MR, Stonestreet P, Thurston NC, Wiseman JS. 1998. Evaluation of a novel self-aerating, 

oscillating baffle column. Canadian Journal of Chemical Engineering 76(1):5-10. 

Mackley MR, Tweddle GM, Wyatt ID. 1990. Experimental Heat-Transfer Measurements for Pulsatile Flow in 

Baffled Tubes. Chemical Engineering Science 45(5):1237-1242. 

Maier U, Buchs J. 2001. Characterisation of the gas-liquid mass transfer in shaking bioreactors. 

Biochemical Engineering Journal 7(2):99-106. 

Mak ANS, Hamersma PJ, Fortuin JMH. 1992. Solids Holdup and Axial-Dispersion During Countercurrent 

Solids Liquid Contacting in a Pulsed Packed-Column Containing Structured Packing. Chemical 


Mallinger F, Drikakis D. 2002. Instability in three-dimensional, unsteady, stenotic flows. International 

Journal of Heat and Fluid Flow 23(5):657-663. 

Mannal GJ, Tiitchener-Hooker NJ, Chase HA, Dalby PA. 2006. A critical assessment of the impact of 

mixing on dilution refolding. Biotechnology and Bioengineering 93(5):955-963. 

Masiuk S. 1999. Mixing time for a reciprocating perforated plate agitator. Inzynieria Chemiczna I 

Procesowa 20(4):601-612. 

Mazzolai L, Silacci P, Bouzourene K, Daniel F, Brunner H, Hayoz D. 2002. Tissue factor activity is 

upregulated in human endothelial cells exposed to oscillatory shear stress. Thrombosis and 

Haemostasis 87(6):1062-1068. 

McCoy M. 1999. Biocatalysis grows for drug synthesis. Chemical & Engineering News 77(1):10-14. 


Mehaia MA, Cheryan M. 1984. Ethanol-Production in a Hollow Fiber Bioreactor Using Saccharomyces- 

Cerevisiae. Applied Microbiology and Biotechnology 20(2):100-104. 

Millward HR, Bellhouse BJ, Sobey IJ. 1996. The vortex wave membrane bioreactor: Hydrodynamics and 

mass transfer. Chemical Engineering Journal and the Biochemical Engineering Journal 62(3):175- 

181. 

Moo-Young M, Blanch HW, Drew S, Wang DIC. 1985. Comprehensive biotechnology : the principles, 

applications and regulations of biotechnology in industry, agriculture and medicine. Oxford: 

Pergamon. xxv,1136p p. 

Morris CJ, Forster FK. 2004. Oscillatory flow in microchannels - Comparison of exact and approximate 

impedance models with experiments. Experiments in Fluids 36(6):928-937. 

Murthy V, Ramachandran KB, Ghose TK. 1987. Gas Hold-up and Power-Consumption in an Air-Pulsed 

Column Bioreactor. Process Biochemistry 22(1):3-8. 

Narumi T, See H, Suzuki A, Hasegawa T. 2005. Response of concentrated suspensions under large 

amplitude oscillatory shear flow. Journal of Rheology 49(1):71-85. 

Nauman E, Buffham B. 1983. Dispersion, diffusion and open systems, in Mixing Continuous Flow 

Systems. Toronto: Wiley. 

Navarro JM, Goma G; Institut National de la Propieté Industrielle, Paris, assignee. 1980. Nouveau 

dispositif de mise en oeuvre des microorganisms. France patent 78 28572. 

ncbiotech. 2002. About Biotech. http://www.ncbiotech.org/aboutbt/main.cfm ed: North Carolina 

Biotechnology Center. 

Nehari D, Armenio V, Ballio F. 2004. Three-dimensional analysis of the unidirectional oscillatory flow 

around a circular cylinder at low Keulegan-Carpenter and beta numbers. Journal of Fluid Mechanics 

520:157-186. 

Neves-Saraiva RMC. 1998. The characterisation of mixing for oscillatory flow within baffled tubes [PhD]. 

Cambridge: University of Cambridge. 


67

Ni X; Heriot-Watt University, assignee. 2002. Method and Apparatus for Phase Separated Synthesis. USA 

68 

patent 6,429,268. 

Ni X, Brogan G, Struthers A, Bennett DC, Wilson SF. 1998a. A systematic study of the effect of 

geometrical parameters on mixing time in oscillatory baffled columns. Chemical Engineering 

Research & Design 76(A5):635-642. 

Ni X, Cosgrove JA, Arnott AD, Greated CA, Cumming RH. 2000a. On the measurement of strain rate in an 

oscillatory baffled column using particle image velocimetry. Chemical Engineering Science 

55(16):3195-3208. 

Ni X, Gao S. 1996. Scale-up correlation for mass transfer coefficients in pulsed baffled reactors. Chemical 

Engineering Journal 63(3):157-166. 

Ni X, Gao S, Cumming RH, Pritchard DW. 1995a. A Comparative-Study of Mass-Transfer in Yeast for a 

Batch Pulsed Baffled Bioreactor and a Stirred-Tank Fermenter. Chemical Engineering Science 

50(13):2127-2136. 

Ni X, Gough P. 1997. On the discussion of the dimensionless groups governing oscillatory flow in a baffled 

tube. Chemical Engineering Science 52(18):3209-3212. 

Ni X, Jian H, Fitch AW. 2002a. Computational fluid dynamic modelling of flow patterns in an oscillatory 

baffled column. Chemical Engineering Science 57(14):2849-2862. 

Ni X, Liu S, Grewal PS, Greated CA. 1995b. A study of velocity Vector Profile and Strain Rate Distribution 

for Laminar and Oscillatory Flows in a Baffled Tube Using Particle Imaging Velocimetry. Jornal of 

Flow Visualisation and Image Processing 2:135-147. 

Ni X, Mackley MR. 1993. Chemical-Reaction in Batch Pulsatile Flow and Stirred-Tank Reactors. Chemical 

Engineering Journal and the Biochemical Engineering Journal 52(3):107-114. 

Ni X, Mackley MR, Harvey AP, Stonestreet P, Baird MHI, Rao NVR. 2003. Mixing through oscillations and 

pulsations - A guide to achieving process enhancements in the chemical and process industries. 



Ni X, Zhang Y, Mustafa I. 1998b. An investigation of droplet size and size distribution in 

methylmethacrylate suspensions in a batch oscillatory-baffled reactor. Chemical Engineering 

Science 53(16):2903-2919. 

Ni X, Zhang Y, Mustafa I. 1999. Correlation of polymer particle size with droplet size in suspension 

polymerisation of methylmethacrylate in a batch oscillatory-baffled reactor. Chemical Engineering 

Science 54(6):841-850. 

Ni XW. 1994. Residence Time Distribution Measurements in a Pulsed Baffled Tube Bundle. Journal of 

Chemical Technology and Biotechnology 59(3):213-221. 

Ni XW. 1995. A Study of Fluid Dispersion in Oscillatory Flow-through a Baffled Tube. Journal of Chemical 

Technology and Biotechnology 64(2):165-174. 

Ni XW, de Gelicourt YS, Baird MHI, Rao NVR. 2001. Scale-up of single phase axial dispersion coefficients 

in batch and continuous oscillatory baffled tubes. Canadian Journal of Chemical Engineering 

79(3):444-448. 

Ni XW, de Gelicourt YS, Neil J, Howes T. 2002b. On the effect of tracer density on axial dispersion in a 

batch oscillatory baffled column. Chemical Engineering Journal 85(1):17-25. 

Ni XW, Gao SW, Pritchard DW. 1995c. Study of Mass-Transfer in Yeast in a Pulsed Baffled Bioreactor. 

Biotechnology and Bioengineering 45(2):165-175. 

Ni XW, Gao SW, Santangeli L. 1997. On the effect of surfactant on mass transfer to water-glycerol 

solutions in a pulsed baffled reactor. Journal of Chemical Technology and Biotechnology 69(2):247- 

253. 

Ni XW, Murray KR, Zhang YM, Bennett D, Howes T. 2002c. Polymer product engineering utilising 

oscillatory baffled reactors. Powder Technology 124(3):281-286. 

Ni XW, Nelson G, Mustafa I. 2000b. Flow patterns and oil-water dispersion in a 0.38 m diameter 

oscillatory baffled column. Canadian Journal of Chemical Engineering 78(1):211-220. 

Ni XW, Pereira NE. 2000. Parameters affecting fluid dispersion in a continuous oscillatory baffled tube. 

Aiche Journal 46(1):37-45. 


69

Ni XW, Stevenson CC. 1999. On the effect of gap size between baffle outer diameter and tube inner 

70 

diameter on the mixing characteristics in an oscillatory-baffled column. Journal of Chemical 


Nielsen J, Villadsen J. 1992. Modeling of Microbial Kinetics. Chemical Engineering Science 47(17- 

18):4225-4270. 

Nishimura T, Kunitsugu K. 2001. Three-dimensionality of grooved channel flows at intermediate Reynolds 

numbers. Experiments in Fluids 31(1):34-44. 

Nishimura T, Ohori Y, Kajimoto Y, Kawamura Y. 1985. Mass-Transfer Characteristics in a Channel with 

Symmetrical Wavy Wall for Steady Flow. Journal of Chemical Engineering of Japan 18(6):550-555. 

Nishimura T, Oka N, Yoshinaka Y, Kunitsugu K. 2000. Influence of imposed oscillatory frequency on mass 

transfer enhancement of grooved channels for pulsatile flow. International Journal of Heat and Mass 

Transfer 43(13):2365-2374. 

Oliveira MSN, Fitch AW, Ni X. 2003a. A study of bubble velocity and bubble residence time in a gassed 

oscillatory baffled column - Effect of oscillation frequency. Chemical Engineering Research & Design 

81(A2):233-242. 

Oliveira MSN, Fitch AW, Ni XW. 2003b. A study of velocity and residence time of single bubbles in a 

gassed oscillatory baffled column: effect of oscillation amplitude. Journal of Chemical Technology 

and Biotechnology 78(2-3):220-226. 

Oliveira MSN, Ni X. 2001. Gas hold-up and bubble diameters in a gassed oscillatory baffled column. 

Chemical Engineering Science 56(21-22):6143-6148. 

Oliveira MSN, Ni XW. 2004. Effect of hydrodynamics on mass transfer in a gas-liquid oscillatory baffled 

column. Chemical Engineering Journal 99(1):59-68. 

Paalme T, Kahru A, Elken R, Vanatalu K, Tiisma K, Vilu R. 1995. The computer-controlled continuous 

culture of Escherichia coli with smooth change of dilution rate (A-stat). Journal of Microbiological 

Methods 24(2):145-153. 


Palma M, Giudici R. 2003. Analysis of axial dispersion in an oscillatory-flow continuous reactor. Chemical 

Engineering Journal 94(3):189-198. 

Papamichael N, Hustedt H. 1994. Enzyme Recovery by Continuous Crosscurrent Extraction. Aqueous 

Two-Phase Systems. p 573-584. 

Pereira NE, Ni XW. 2001. Droplet size distribution in a continuous oscillatory baffled reactor. Chemical 


Prochazka J, Rod V; Ceskoslosvenska Academie Ved, assignee. 1974. Apparatus for Bringing Fluid Phases 

Into Mutual Contact. USA patent 3,855,368. 

Puthli MS, Rathod VK, Pandit AB. 2005. Gas-liquid mass transfer studies with triple impeller system on a 

laboratory scale bioreactor. Biochemical Engineering Journal 23(1):25-30. 

Ralph ME. 1986. Oscillatory Flows in Wavy-Walled Tubes. Journal of Fluid Mechanics 168:515-540. 

Ram K, Vickroy TB, Lamb KA, Slater NKH, Dennis JS, Duffy LE. 2000. Mixing in process vessels used in 

biopharmaceutical manufacturing. Biotechnology Progress 16(2):244-247. 

Ranade VV. 2002. Computational flow modeling for chemical reactor engineering: Academic Press, 

c2002. 452 p. p. 

Roberts EPL. 1992. Unsteady flow and mixing in baffled channels [PhD]. Cambridge: University of 

Cambridge. 

Roberts EPL, Mackley MR. 1995. The Simulation of Stretch Rates for the Quantitative Prediction and 

Mapping of Mixing within a Channel Flow. Chemical Engineering Science 50(23):3727-3746. 

Roberts EPL, Mackley MR. 1996. The development of asymmetry and period doubling for oscillatory flow 

in baffled channels. Journal of Fluid Mechanics 328:19-48. 

Sarpkaya T. 1966. Experimental Determination of the Critical Reynolds Number for Pulsating Poiseuille 

Flow. Journal of Basic Engineering, ASME Vol. 88:pp. 589-598. 

Schmid A, Dordick JS, Hauer B, Kiener A, Wubbolts M, Witholt B. 2001. Industrial biocatalysis today and 

tomorrow. Nature 409(6817):258-268. 


71

Schugerl K. 1982. Some Basic Principles for the Layout of Tower Bioreactors. Journal of Chemical 

72 


Serieys M, Goma G, Durand G. 1978. Design and Oxygen-Transfer Potential of a Pulsed Continuous 

Tubular Fermentor. Biotechnology and Bioengineering 20(9):1393-1406. 

Sinada F, Karim AGA. 1984a. Physical and Chemical Characteristics of the Blue Nile and the White Nile at 

Khartoum. Hydrobiologia 110(MAR):21-32. 

Sinada F, Karim AGA. 1984b. A Quantitative Study of the Phytoplankton in the Blue and White Niles at 

Khartoum. Hydrobiologia 110(MAR):47-55. 

Sipkema EM, de Koning W, Ganzeveld KJ, Janssen DB, Beenackers A. 1998. Experimental pulse 

technique for the study of microbial kinetics in continuous culture. Journal of Biotechnology 64(2- 

3):159-176. 

Skala D, Veljkovic V. 1988a. Mass-Transfer Characteristics in a Gas-Liquid Reciprocating Plate Column .1. 

Liquid-Phase Volumetric Mass-Transfer Coefficient. Canadian Journal of Chemical Engineering 

66(2):192-199. 

Skala D, Veljkovic V. 1988b. Mass-Transfer Characteristics in a Gas-Liquid Reciprocating Plate Column. 1. 

Liquid-Phase Volumetric Mass-Transfer Coefficient. Canadian Journal of Chemical Engineering 

66(2):192-199. 

Sobey IJ. 1980. Flow through Furrowed Channels. 1. Calculated Flow Patterns. Journal of Fluid Mechanics 

96(JAN):1-26. 

Sobey IJ. 1983. The Occurrence of Separation in Oscillatory Flow. Journal of Fluid Mechanics 

134(SEP):247-257. 

Sobey IJ. 1985. Observation of Waves During Oscillatory Channel Flow. Journal of Fluid Mechanics 

151(FEB):395-426. 

Souhrada F, Galvez BB; Gulf Canada Limited, Toronto, Canada, assignee. 1981 1981. Method and 

apparatus for the prevention of solid deposits in a tubular reactor by pulsed flow. USA patent 

4,271,007. 


Stephens GG, Mackley MR. 2002. Heat transfer performance for batch oscillatory flow mixing. 

Experimental Thermal and Fluid Science 25(8):583-594. 

Steven DD, Frank B, Lye GJ. 2004. High throughput screening and process optimisation. In: Kristiansen 

B, editor. Basic biotechnology. Cambridge: Cambridge University Press. 

Stonestreet P, Harvey AP. 2002. A mixing-based design methodology for continuous oscillatory flow 

reactors. Chemical Engineering Research & Design 80(A1):31-44. 

Tabeling P, Chabert M, Dodge A, Jullien C, Okkels F. 2004. Chaotic mixing in cross-channel micromixers. 

Philosophical Transactions of the Royal Society of London Series a-Mathematical Physical and 

Engineering Sciences 362(1818):987-1000. 

Takriff MS, Masyithah Z. 2002. Interstage backmixing in oscillatory flow in a baffled column. Chemical 

Engineering Communications 189(12):1640-1652. 

Testik FY, Voropayev SI, Fernando HJS. 2005. Flow around a short horizontal bottom cylinder under 

steady and oscillatory flows. Physics of Fluids 17(4). 

Thomas AM, Narayanan R. 2002a. A comparison between the enhanced mass transfer in boundary and 

pressure driven oscillatory flow. International Journal of Heat and Mass Transfer 45(19):4057-4062. 

Thomas AM, Narayanan R. 2002b. The use of pulsatile flow to separate species. Microgravity Transport 

Processes in Fluid, Thermal, Biological, and Materials Sciences. p 42-56. 

Vaidyanathan S, Macaloney G, Vaughn J, McNeil B, Harvey LM. 1999. Monitoring of submerged 

bioprocesses. Critical Reviews in Biotechnology 19(4):277-316. 

Van Dick WJD; 1935. Process and apparatus for intimately contacting fluids. USA patent 2,011,186. 

Voltairas PA, Fotiadis DI, Massalas CV, Michalis LK. 2005. Anharmonic analysis of arterial blood pressure 

and flow pulses. Journal of Biomechanics 38(7):1423-1431. 

Walsh G. 2003. Biopharmaceutical benchmarks. Nature Biotechnology 21(11):1396-1396. 

Walsh G. 2005. Biopharmaceuticals: recent approvals and likely directions. Trends in Biotechnology 

23(11):553-558. 


73

Wang XS, Rhodes MJ. 2005. Using pulsed flow to overcome defluidization. Chemical Engineering Science 

74 

60(18):5177-5181. 

Weuster-Botz D, Puskeiler R, Kusterer A, Kaufmann K, John GT, Arnold M. 2005. Methods and milliliter 

scale devices for high-throughput bioprocess design. Bioprocess Biosyst Eng:1-11. 

Ye XF, Shimizu M. 2001. Augmented longitudinal diffusion in grooved tubes for oscillatory flow. 

International Journal of Heat and Mass Transfer 44(3):633-644. 

Ye XF, Zhang JW. 2002. Effect of turbulence on Taylor dispersion for oscillatory flows. International 

Journal of Heat and Mass Transfer 45(21):4373-4380. 

Zhang DQ, Ni ZM. 1996. Separation and determination of trace inorganic germanium in beta- 

carboxyethylgermanium sesquioxide by filtration chromatography and hydride generation graphite 

furnace atomic absorption spectrometry. Analytica Chimica Acta 330(1):53-58. 

Zhang J, Greasham R. 1999. Chemically defined media for commercial fermentations. Applied 

Microbiology and Biotechnology 51(4):407-421. 

Zhang YM, Ni XW, Mustafa I. 1996. A study of oil-water dispersion in a pulsed baffled reactor. Journal of 

Chemical Technology and Biotechnology 66(3):305-311.

Nuno Miguel Fernandes Reis Novel Oscillatory Flow Reactors for ...

You also want an ePaper? Increase the reach of your titles

Delete template?

Save as template?