Copyright is owned by the Author of the thesis. Permission is given for a copy to be downloaded by an individual for the purpose of research and private study only. The thesis may not be reproduced elsewhere without the permission of the Author. TOWER FERMENTATION OF WHEY PERMEATE �D SUCROSE-ENRICHED WHEY PERMEATE TO ETHANOL A thesis presented in partial fulfilment of the requirements for the Degree of Doctor of Philosophy in Biotechnology at Massey University CHATURONG BOONTANJAI 1983 ABSTRACT Tower fermentation of sulphuric acid whey permeate using Kluy ver omyces marxianus Y42 has been investigated . The tower fermenter used was 0 . 025 m in diameter and 2 . 69 m high . The straight section o f the tower was 2 . 37 m . The total tower volume was 2 . 9 litres and the separator section made up 1 . 6 litres of the total volume. The operating temperature was 30°C . The optimum medium feed rate was observed at a superficial liquid velocity of 0 . 24 mm/s . I t was found that a tower height of only 0 . 82 m was required , excluding the separator sec tion , and the corresponding residence time was 1 hour . An exit ethanol concentration of 1 6 g/1 was produced at a productivity o f 1 6 g/lh from 45 g/1 lac tose i n the whey permeate feed ( 94% uti lization ) . This was an ethanol yield of 71 % on lactose utilized . If the separator sec tion were considered , the ethanol productivity was 5 g/lh and the exit ethanol concentration was 1 9 g/1 , while the overall re tention time was 3 . 7 hours . The cell concentration inside the tower varied between 1 0 and 1 00 g/1 dried weight ( 54 and 350 g/1 wet weight ) being greatest a t the bot tom o f the tower . K . . mar xianus was found to be inhibited by a high level of ethanol in the growth medium and unable to ferment comple tely a high concentration of lac tose when tested in 1 0 litre-scale-batch fermentation . Further tests in the presence of sucrose and lac tose found that this yeast exhibited diauxic behaviour by utili zing suc rose before lac tose . This behaviour generally resulted in incomplete lac tose utili zation in the tower . In the screening for a flocculent lac tose-fermenting yeast , the yeast strain K. mar xianu swas found to be the only flocculent yeast , but it was only moderately flocculent . Further investigation found that it had good flocculence when grown in media which support good growth , and poor flocculence when grown in acidic media and in media which do not support good growth . A subcul ture of this yeast strain showed moderate flocculence when grown in whey permeate . Tower fermentation of whey permeate enriched with molasses by mixed culture of Sacaharomyces cer ev{s.iae CFCC39 and K. marxianus Y42 was found to be difficult . The difficulty arose because of incomplete lac tose utilization even at a very low feed rate ( up to 0 . 1 4 mm/s ) and incompatible flocculation properties of the two yeast species employed . Blockage of the separator and gas slug formation were caused by the very flocculent yeast mass of S. cerevisiae CFCC39 . This caused ii iii K. marxianusto be s lowly washed out of the tower fermenter . Sucrose was completely uti lized at the bot tom of the tower fermenter , while lac tose utilization was slow and incomplete . The incomplete lactose utilization has been attributed to the diauxic behaviour of K. marxianus� ethanol inhibition and molasses inhibition ( probably due to its reaction with whey permeate during autoclaving ) . Results of tower fermentation of cane molasses have also been given for characterization of the tower fermenter used . Experiments to isolate an ethanol tolerant K. marxianus using a serial subculture in a medium containing increasing ethanol concentrations were performed . The isolate obtained could tolerate up to 50 g/l ethanol . I t could produce ethanol at a faster rate lac tose-fermenting yeast tested . ferment lac tose in whey permeate to than the parent strain and other The isolate was found to be stable . It was not used in the tower fermenter as it was non-flocculent . An attempt was made to isolate a sucrose-negative K. marxianus . This was only partially successful . The mutant did not grow on sucrose agar but reverted to the wild type when grown in liquid medium containing both sucrose and lac tose . An experiment to isolate a diauxie-negative K. marxianus strain using D-glucosamine as a glucose analogue was also described . This was unsuccessful because K. marxianus was able to grow on lactose in presence of the analogue . TO MY KIWI AND THAI PARENTS ACKNOWLEDGEMENT During the course o f his attempt to complete this study , the author owed grat itude to many people and wishes to thank the following Professor R . L . Earle and Dr . Mary D . Earle for the kindness and help given to him and to the development of Food Technology and Biotechnology studies in Thailand and in particular at Khonkaen University . The DSIR for providing financial support during part of this study . His supervisors Drs . Vidar Friis Larsen and John D . Brooks for their guidance and advice . Dr . Ian S . Maddox for his abl e guidance , advice and forever readiness to help and answer the many questions from the author on general industrial microbiology . Dr . Graham J . Manderson for his forever readiness to give help and advice. on yeas t morphology and behaviour , and photomicroscopy . Dr . Noel W . Dunn of the School of Biological Science , University of New South Wales , Sydney , Australia , for his guidance, advice and planning of the experiments on the cul ture improvement o f Kluyveromyces marxianus. Assoc . Professor Anthony M .MacQuillan of the Dep t . of Microbiology , University of Maryland , College Park , Maryland , USA, for providing two lactose fermenting yeast cultures in which one of them was flocculent. Thus , enabl ing the completion of this study , and for his advice on Eluyveromyces species mutation . Dr . Marion Ewen of the Dairy Research Institute (DRI ) , Palmerston North , for her various help and advice , particularly for providing references on whey technology , giving som� first hand knowledge on tower fermentation and reading the manuscrip t . Dr . Roy J . Thornton of the Dept . of Microbiology , Massey University , for giving consultation and advice on yeast mutation. The staff of DRI , Palmerston North . The Whey produc t Section : Messrs . Peter Hobman , Mike O ' Connell and John Bligh for their assistance in providing the whey permeate . The DRI library staff for their able assis tance in particular Lorraine Tremain . The technical and secretarial staff of the Faculty o f Technology : John Alger , Paul Shaw , Derek Couling , Mike Stevens , Robyn Calder , Mark Lubbers , Astrid Ndzinge , Melvin Smith , Terry Gracie , Margaret Bewley , V vi Raewyn Cheer , Beverly Hawthorn and Carol C louston for their abl e assis tance . The s taff of Massey Univers ity Lib rary in particular the Serials and Int erloan Departments . Cla ire Mudford for her ass is tance during the experiments on tower fermentation, and thesis preparation. The boiler house s taff for the ir cooperation in providing the s te am outs ide office hours . Bruce Walker for his assis tance in the transportation of whey permeat e . His fellow s tudents : Tom Clark , Sheelagh Wilkinson , Moazzem Hozzain , Tipavana Ngarmsak , Pisanu Vichiensanth , Tony Ret ter, Richard Gapes , J ohn Mawson , Warren Hollaway , Wong Tze Sen and Pamela Palfreyman. Many thanks for their helpful and encouraging words , advice and sharing sadness during bad time s . Many thanks t o the many friends i n the Faculty of Technolo gy who have provided a very friendly atmosphere to work in. B rian Wilkinson, Dick Poll , Mr . & Mrs Meredith and their respective families for their kindness and encouraging words and for o f fering the warmth o f their homes . Mrs . Bunnak Wickham and (Aunty) Bertha Zurcher for finding and let ting a comfortable accommodation . S arah Sant for proof reading part of the final manuscript . The Computer Centre s taff for their cooperation and assis tance in the preparation of the manuscript . Vivienne Mair and Joane Charles for typing part of the manus cript . Mr . A. Eustace, Sunanta Juntakul and Priscilla Burton for their kind assis tance during photocopying . Chris t ine Samaniego and Linda Poll for their assis tance in the preparation o f the manuscript for b inding . Finally , his Kiwi parents (Mr . & Mrs . Moxon) and their family for all the love , kindness , warmth, care , encouragement and help that they have given which could not b e express ed in a few words. Without their as.s istance the completion of this s tudy would have been extremely difficult . TABLE OF CONTENTS ABSTRACT ACKNOWLEDGEMENTS TABLE OF CONTENTS LIST OF FIGURES LIST OF TABLES LIST OF ABBREVIATIONS 1 2 WHEY , ITS UTILIZATION AND DISPOSAL 1 . 1 Introduction 1 . 2 Types of whey 1 . 3 Whey composition 1 . 4 Whey production 1 . 5 Whey utilization and d isposal 1 . 5 . 1 Whey disposal 1 . 5 . 2 Protein production 1 . 5 . 3 Lactose production 1 . 5 . 4 Fermentation of whey (a ) Beverage product ion (b ) Lactic acid production ( c ) Citric acid production (d ) Acetic acid fermentation ( e ) Microbial protein production ( f ) Butanol product ion ( g) Production of o ther fermentation (h ) Ethanol product ion 1 . 6 Summary LITERATURE REVIEW 2 . 1 Introduction to e thanol fermentation 2 . 1 . 1 General biochemistry 2 . 1 . 2 General microbiology 2 . 2 Ethanol fermentation of whey 2 . 2 . 1 Microorganisms products 2 . 2 . 2 �etabolism of lactose to etnanol by yeast 2 . 2 . 3 Yeast environmental considerations vii page ii V vii xviii xxi xxiv 1 1 1 3 3 6 6 6 7 7 7 7 7 8 8 8 8 9 1 1 1 2 1 2 14 14 15 15 viii (a) Substrate utilization Concentrated whey (b) Ethanol production ( c ) Ethanol inhibition (d) Aeration (e ) Nutrient requirements ( f ) Temperature 1 5 1 6 1 7 1 9 20 20 20 (g) pH 2 1 2 . 3 Ethanol fermentation processes 2 1 2 . 3 . 1 Al ternative processes to batch fermentation 2 1 (a) Continuous stirred tank fermentation 2 1 (b) Vacuum fermentation 2 1 ( c ) Rapid batch fermentation 23 (d ) Fermentation by immobilized cells 23 ( e ) Tower fermentation 24 2 . 3 . 2 Industrial and pilot plant processes used for the. 25 production of ethanol from whey 2 . 4 Ethanol product ion by tower fermentation 26 2 . 4 . 1 History o f tower fermentation 26 2 . 4 . 2 The characteristics and operation o f the tower 26 fermenter (a) Organisms for tower fermentation 28 (b) The effect of tower height 28 (c ) Res idence times in tower fermentation 2 9 (d ) The effect of original wort specific gravity on 30 l imiting volumetric efficiency ( e ) Aeration 2 . 4 . 3 Appl ications of tower fermentation 3 1 3 1 2 . 4 . 4 Industrial ethanol tower fermentation 3 1 ( a ) Laboratory scale investigations 31 (b) Media used and the effect of sugar concentration 32 (c) Fermentation temperature (d ) Operating pH ( e ) Conclusions 2 . 5 Flocculation of yeasts 2 . 5 . 1 Flocculent yeast classification (a ) Non-flocculent (b) Flocculent-physically l imited ( c ) Flocculent-fermentation l imited 33 34 34 35 35 35 35 35 2 . 5 . 2 (a ) (b) (c) ix Quantitative measurement of yeast flocculation Burn ' s method as modified by Helm e t al Burn ' s method as modified by Stewart Sharp ' s method 35 35 35 36 (d ) Spectrophotometric method 37 2 . 5 . 3 Factors influencing yeast flocculence 37 2 . 5 . 3 . 1 Inherited flocculent properties in yeast 37 (a ) Flocculent genes 37 (b ) The yeast cell wall 37 ( c ) Comparison of the cell walls of flocculent and 38 non-flocculent yeasts (d) Chemical effects on the disulphide bridge 38 (e) The fimbria of yeast cell wall 39 (f) The cell wall ionic charge 39 2 . 5 . 3 . 2 Environmental effect on yeast flocculation 39 (a) Flocculation aids 39 (b) Mechanism of ionic induced flocculation 40 (c) Deflocculating agents 40 (d) Growth media 40 (e ) Temperature ( f ) pH (g) Coflocculation (h) Agitation 2 . 6 Mixed culture and mixed substrate fermentation 2 . 7 Conclusions 3 MATERIALS AND METHODS 3 . 1 Materials 3 . 1 . 1 Chemicals 3 . 1 . 2 Gases 3 . 1 . 3 Media (a) Whey permeate (b) Molasses ( c ) Tower fermentation (d) Tower fermentat ion start up media media (e ) Culture preservation media ( f ) Flocculation test media (g) Basic nutrient base (h) Whey broth and agar 4 1 4 1 4 1 4 1 4 1 4 3 4 5 4 5 4 5 4 5 45 45 4 5 4 6 4 6 4 6 4 6 4 6 X ( i) Lactose agar ( j ) Sucrose broth and agar (k) Total cell plate count agar ( 1) pH adjustment of media 3 . 1 . 4 Organisms 3 . 2 Equipment 3 . 2 . 1 Tower fermenter (a ) The separator (b ) Temperature control ( c ) Tower and medium aeration (d) Air supply and fil ter ( e ) Medium pump 3 . 2 . 2 10 l itres batch fermenter 3 . 2 . 3 UV lamp 3 . 2 . 4 Repl ica plating 3 . 2 . 5 Glassware 3 . 3 S terilization 3 . 3 . 1 Media and glassware 3 . 3 . 2 Tower fermenter 3 . 4 Analytical methods 3 . 4 . 1 Lactose 3 . 4 . 2 Sucrose 3 . 4 . 3 Ethanol 3 . 4 . 4 Cel l concentration (a ) Cell dried weight and wet weight (b) Plate count ( c ) Haemacytometer count 3 . 4 . 5 Yeast flocculence (a) Flocculation scale method (b) Sharp ' s modified Burn ' s number 3 . 4 . 6 pH 3 . 4 . 7 Specific gravity 3 . 5 Culture preservation and maintenance 3 . 6 Inoculum preparation 3 . 7 Fermentation conditions 3 . 8 Tower fermentation 3 . 8 . 1 S tart up (a) Initial s tart up (b ) Subsequent s tart up 4 7 4 7 47 47 47 48 48 5 1 5 1 5 3 5 3 53 5 3 5 6 5 6 56 5 6 5 6 56 5 7 5 7 5 7 5 7 5 7 5 7 5 8 58 58 58 59 59 59 5 9 6 0 6 0 6 1 6 1 6 1 6 1 xi 3.8.2 Sampl ing procedure 3.8 .3 Continuous operation 3.9 Flocculation tests 3.10 Culture improvements 62 63 63 64 3.10.1 Isolation of ethanol tolerant K. marxianus using 64 e thanol gradient agar 3.10.2 An attempt to isolate sucrose negative K. marxianus 65 s trains (a ) Determination of optimum irradiation time (b ) Replica plating and i solation 3.1 1 Calculation methods 3.1 1 .1 Tower fermentation 3.1 1 .2 Batch fermentation 4 TOWER FERHENTATION OF WHEY PERMEATE 4.1 The relationship between tower height and various fermentation parameters 4.1.1 Lactose concentration and utilization 4.1.2 Volumetric rate of lactose utilization 4.1 .3 Specific rate of lac tose utilization 4.1.4 Ethanol concentration 4.1.5 Ethanol yield 4.1 .6 Volumetric rate of ethanol production 4.1.7 Specific rate of ethanol production 4.1 .8 Cell concentration 4.1.9 Medium pH 65 66 67 67 69 7 1 7 1 74 75 7 7 79 80 8 1 83 87 4.2 The effect of the residence time on various fermentation 88 parameters 4.2.1 Lactose concentration 88 4.2.2 Ethanol concentration 89 4.2.3 The rates of lactose utilizat ion and ethanol 91 production 4.3 The effect of superficial liquid velocity on various 9 4 fermentation paramet ers 4.3.1 Lactose concentrat ion 94 4.3.2 Ethanol concentrat ion 96 4.3.3 Rates of lactose uti lizat ion and ethanol 98 production 4.3.4 Cell concentration 103 4 . 3 . 5 4 . 4 Tower 4 . 4 . 1 4 . 4 . 2 4 . 4 . 3 4 . 4 . 4 xii Specific growth rate fermenter performance Optimum superficial liquid velocity Res idence time and tower height Sugar utilization Yield o f ethanol 106 107 107 109 1 10 1 10 4 . 5 Optimum conditions for the tower fermentation o f whey 1 10 permeate 4 . 5 . 1 Comparison with other tower f ermentation investigations 4 . 5 . 2 Comparison with batch fermentation of whey 4 . 6 Continuous operation and difficulties 4 . 6 . 1 Organism 4 . 6 . 2 Continuous operat ion monitoring curves 4 . 6 . 3 Contamination 4 . 6 . 4 Feed lactose concentration 4 . 7 Conclusions 4 . 8 Summary 5 TOWER FERMENTATION OF WHEY PERMEATE ENRICHED WITH· MOLASSES 5 . 1 The relationship between tower height and various fermentation parameters 5 . 1 . 1 Sugar concentrations and utilizations ( a) Sucrose (b) Lactose (c) Total sugar 5 . 1 . 2 Volumetric rates of sugar utilization ( a) Sucrose (b) Lactose (c ) Total sugar 5 . 1 . 3 Specific rates of sugar utilization (a) Sucrose (b) Lactose (c) Total sugar 5 . 1 . 4 Ethanol concentration 5 . 1 . 5 Ethanol yield 5 . 1 . 6 Volumetric rate of ethanol production 5 . 1 . 7 Specific rate of ethanol production 5 . 1 . 8 Ce ll concentration 1 1 1 1 13 1 14 1 14 1 14 1 16 1 18 1 18 1 19 1 20 1 20 1 20 1 2 2 1 2 5 1 25 1 25 1 2 5 1 2 7 1 2 7 1 2 7 1 27 129 129 1 3 1 1 3 1 1 33 1 34 ( a) K. marxianus ( b ) S. cerevisiae xiii (c) Total cell concentration 5 . 1 . 9 Medium pH 5 . 2 Contamination of continuous tower fermentation culture 5 . 3 Tower ferment er performance 5 . 4 Comparisons between fermentation of whey permeate enriched with molasses and with sucrose . 5 . 5 Fermentation comparison using different ratios of mixed yeast culture in the inoculum . 5 . 6 Effect on floc morphology of species ratio in inoculum 5.7 Conclusions 5 . 8 Summary 6 FLOCCULATION TESTS 6 . 1 Introduction 6 . 2 Test media 6 . 2 . 1 Glossary of abbreviations used in the flocculation tes ts 6. 3 Flocculation test results 6 . 3 . 1 Modified Burn ' s number and flocculation testing methods used (a ) Flocculation test media (b) Technique for the determination of modified Burn ' s number (c) Alternative technique ( d ) Flocculent scale method 6 . 3 . 2 Flocculating ability of some lactose-fermenting yeasts s trains 6 . 3 . 3 Observation of f locculence during shakeflask fermentation ( a ) KM Y42 (b ) KM.Y42 (TS ) ( c ) s. cerevisiae FT 146 ( SC 146 ) (d) s. cerevisiae CFCC39 ( CCC39 ) ( e ) se 1 4 6 + KM Y42 1 34 1 37 1 3 7 1 4 1 1 4 2 1 4 2 144 146 148 1 50 1 50 1 5 2 1 5 2 1 5 2 1 5 6 1 5 6 1 56 1 58 1 58 1 58 1 5 8 - 159 1 5 9 1 60 1 60 160 160 xiv ( f ) CC 39 + KM Y42 6 . 3 . 4 Flocculation of KM Y42 grown in whey permeate with no additive 6 . 3 . 5 The effect of the initial medium pH on flocculation of KM Y42 grown in whey permeate 6 . 3 . 6 The effect of membrane filtration on flocculation 1 6 1 1 6 1 1 6 2 o f KM Y42 grown in whey permeate with additives 1 6 2 6 . 3 . 7 Flocculation of KM Y42 grown in whey permeate supplemented with organic nutrients 1 6 3 6 . 3 . 8 Flocculation of KM Y42 grown in whey permeate supplemented with inorganic nutrients 1 64 6 . 3.9 Flocculation of KM Y42 grown in media s�pplemented with flocculation aids 165 6 . 3 .10 Flocculation of KM Y42 grown in double sugar substrates 166 6 . 3 . 1 1 Flocculation of KM Y42 in different media 166 6 . 3 . 1 2 Flocculation of KM Y42 grown as mixed culture 1 6 8 with CC 39 or SC 146 in mixed substrate 6 . 3 . 1 3 Flocculation of strains CC 39 and SC 146 1 68 6 . 3 . 1 4 Flocculation curves 1 69 6 . 4 Discussion 1 7 1 6 . 4 . 1 Flocculation o f K. marxianus Y42 1 7 1 (a) Initial investigation 1 7 1 (b) Membrane filtration 1 7 1 ( c ) The additions o f yeast and malt extract broths to whey permeate 1 7 2 (d ) The addition of peptone , urea and diammonium hydrogen phosphate 1 7 2 ( e ) Lactose , glucose or maltose a s a carbon source 1 7 2 ( f ) Enriched whey permeate 1 7 3 ( g ) The addi tion of flocculation aids t o the growth media (h) Medium pH ( i ) Subculture of KM Y42 6 . 4 . 2 Flocculation of CC39 and SC 1 46 grown as pure or mixed cultures with KM Y42 (a) se 1 46 (b ) cc 39 (c ) The effect of the inoculum-growth medium 1 7 3 1 7 3 1 7 4 1 74 1 7 4 1 74 1 7 4 6 . 5 Conclusions 6 . 6 Summary XV 7 MEDIUM OPTIMIZATION AND CULTURE IMPROVEMENT 7 . 1 Introduction 7 . 2 Medium optimization 7 . 3 Isolation of an ethanol tolerant K. marxianus 4 . 3 . 1 Preliminary batch fermentations (a) Whey permeate containing 40 g/1 lactose (b) Whey permeate containing lOO g/ 1 lactose 1 7 5 1 7 5 1 7 7 1 7 7 1 7 8 1 7 8 1 7 8 1 78 ( c) Whey permeate enriched with molasses 1 80 7 . 3 . 2 Selection of ethanol tolerating isolate 185 7 . 3 . 3 The stability of ethanol-tolerant isolate 187 KM 10Dl0 7 . 3 . 4 Fermentation comparison of some lactose fermen- 188 ting yeasts 7 . 3 . 5 10 1 batch fermentations of ethanol-tolerant 189 isolate (a) Whey permeate containing 100 g/1 lactose 1 89 (b) Whey permeate enriched with molasses 190 7 . 3 . 6 Conclusions 191 7 . 4 An attempt to isolate diauxie-negative K. marxianus 19 3 s trains 7 . 4 . 1 Introduction 7 . 4 . 2 Isolation experiment (a) First attempt (b) Second attempt (c) Fermentation test 7 . 4 . 3 The effect of D-glucosamine on growth of K. marxianus 7 . 4 . 4 Conclusions 7 . 5 A mutation attempt to isolate sucrose negative K. marxianus 7 . 5 . 1 Introduction 7 . 5 . 2 First mutation attempt (a) Fermentation of whey permeate by mutant FSN 1 & 2 193 193 194 196 196 1 9 7 1 98 1 98 198 199 1 99 xvi (b ) Fermentation of mixed substrate of lactose and sucrose by mutants FSN 1 and 2 ( c ) Fermentation comparison o f mutants FSN 1 and 2 with parent strain 7 . 5 . 3 Second mutation experiment (a ) Isolation of mutant ( b ) Culture improvement of mutant FSN3 7 . 5 . 4 Conclusions 7 . 6 Summary 8 F INAL DISCUSSION AND CONCLUSIONS REFERENCES APPENDICES A . FEED MEDIUM PUMP CAPACITY AND SAMPLING DATA SHEETS B . EXPERIMENTAL DATA B . 1 Factorial experiment B . 2 Batch fermentation of whey permeate by K. marxianus Y42 B . 3 Tower fermentation o f whey permeate B . 4 Tower fermentation of whey permeate enriched with molasses B . 5 Tower fermentation of molasses C . TOWER FERMENTATION OF MOLASSES C . l The relationship between tower height and various 20 1 203 203 203 205 209 2 10 2 1 1 2 1 6 226 229 229 230 23 1 234 236 fermentation parameters 238 C . 1 . 1 Sucrose and e thanol concentrations 238 C . 1 . 2 Rates of sucrose utilization and e thanol production 240 (a ) Volumetric rates 240 (b) Specific rates 242 C . 1 . 3 Ethanol yield 242 C . 1 . 4 Cell concentration 244 C . 1 . 5 Medium pH 246 C . 2 The effect of the residence t ime on various fermentation parameters C . 2 . 1 Sucrose and ethanol concentration C . 2 . 2 Rates of sucrose utilization and e thanol production C . 3 The ef fect of the superficial liquid velocity on var ious fermentation parameters 248 248 248 252 xvii C . 3 . 1 Sucrose and e thanol concentration 252 C . 3 . 2 Rates of sucrose utilization and e thanol production 254 (a) Volumetric rates 254 ( b ) Specific rates 254 C . 3 . 3 Cell concentration 25 7 C . 3 . 4 Specific growth rate 259 C . 4 Tower performance 259 C . 5 Conclusions and summary 261 D . FLOCCULATION TEST . OBSERVATIONS AND DATA D . l Obs ervations of flocculent D . l . l K. marxianus Y42 D . l . 2 K. marxianus Y42 (TS ) D . l . 3 S. cerevisiae FT 1 4 6 D . l . 4 S. cerevisiae CFCC D.2 Flocculation test data E . ESTIMATION OF UNCERTAINTIES E . l Sugar concentrations E . 2 Ethanol concentration E . 3 Ethanol yield E . 4 Rate of sugar utilization E.4 . 1 Volumetric rate E . 4 . 2 Specific rate 39 E . 5 Rates of e thanol production E . 5 . 1 Volumetric rate E . 5 . 2 Specific rate E . 6 Cell concentration behaviour during E . 6 . 1 Haemacytometer cell count E . 6 . 2 Plate count fermentation E . 6 . 3 Cell dried weight and centrifuged wet weight E . 6 . 4 Estimation of the cell dried weight of K. marxianus Y42 from cell plate count number 262 262 262 264 265 265 266 2 7 2 27 2 2 7 3 2 73 274 274 274 2 75 2 75 275 276 276 276 276 2 77 LIST OF FIGURES Figure Title number 1 . 1 Milk utilizat ion • Typical annual whey production 1 . 2 page 2 4 1 . 3 Summary of processes for whey utilization and disposal 5 2 . 1 EMP pathway 1 3 2 . 2 S chematic diagram of the APV tower fermenter 2 7 2 . 3 Progres s ive reduction o f wort gravity in a tower fermenter 29 2.4 The relationship b etween wort specific gravity and apparent 30 fermentation time 2 . 5 The effect o f original wort specific gravity on limiting 3 1 volumetric e fficiency 2.6 The effect of fermentab le sugars concentration on dilution rate 33 and productivity 2.7 2.8 3. 1 3 . 2 3.3 3.4 3 . 5 3.6 3. 7 3 . . 8 4 . 1 4 . 2 4 . 3 4 . 4 4 . 5 4 . 6 4 . 7 4 . 8 Determination o f modified Burn ' s number � - S relationship of two organisms Tower fermenter set up S chematic d iagram of the tower fermenter Schematic diagram of the separator and draught tube Water heating and air f iltration systems Batch fermenter ( 10 litres working volume ) Replication equipment Subculturing s teps used in the isolat ion o f ethanol tolerating K. marxianus Survival of cells irradiated with UV light Lactose concentration vs tower height Volumetric rate o f lactose utilization vs mean tower height Specific rate o f lactose utilization vs mean tower height Ethanol concentration vs tower height Ethanol yield vs tower height Volumetric rate of ethanol production vs mean tower height Specific rate of ethanol production vs mean tower height Specific rate of ethanol production vs ethanol concentration 4 . 9 Cell concentration vs tower height 4 . 1 0 Tower fermenter during whey permeate fermentation 4 . 1 1 Medium pH vs tower height 4 . 1 2 Lactose concentration vs residence time xviii 36 43 49 50 52 54 55 55 64 66 72 74 76 78 79 80 82 83 84 86 8 7 89 xix 4 . 1 3 Ethanol concentration vs residence time 90 4 . 14 Volumetric rate of lactose utilization vs mean residence time 92 4 . 1 5 Volumetric rate of ethanol production vs mean residence time 4 . 16 Specific rate of lactose utilization vs mean residence t ime 4 . 1 7 Specific rate of ethanol production vs mean 4 . 18 Lactose concentration vs superficial liquid 4 . 19 Ethanol concentration vs superficial liquid 4 . 20 Volumetric rate of lactose utilization vs V s 4 . 2 1 Specific rate of lactose utilization vs V s 4 . 22 Volumetric rate of ethanol production vs V 4 . 2 3 Volumetric rate of ethanol production vs V 4 . 24 Cell concent ration vs V s 4 . 25 (a) Tower operation-monitoring curves (b ) Titration curves 4 . 26 Bacterial contamination s s residence time velocity (V ) s velocity (V ) s 92 93 94 95 9 7 9 9 100 1 0 1 1 0 1 104 ll5 ll7 5 . 1 Sugar concentrations vs tower height 12 1 5 . 2 Volumetric rates of sugar utilization vs mean tower height 126 5 . 3 Specific rates of sugar utilization vs mean tower height 1 28 5 .4 Ethanol concentration vs tower height 1 30 5 . 5 E thanol yield vs tower height 1 3 1 5 .6 Volumetric rate o f ethanol product ion v s mean tower height 5 . 7 Cell numbers vs tower height 5 . 8 Cell concentration vs tower height 5 . 9 Channelling inside the tower 5 . 10 Medium pH vs tower height 5 . 1 1 The effect of mixed culture ratio on floc morphology 6 . 1 Flocculation of KM Y42 (TS) , SC 146 and CC39 grown and tested in different media . 7 . 1 Batch fermentation of whey permeate (40 g/1 lactose) 1 32 1 35 1 38 140 14 1 149 1 70 1 7 9 7 . 2 Batch fermentation of whey permeate (lOO g/1 lactose) 1 7 9 7 . 3 Batch fermentation of whey permeate enriched with molasses 1 8 1 7 .4 Diauxic behaviour s tudy in the fermentation o f whey permeate 183 enriched with molasses 7 . 5 K. marxianus : total cell number ratio vs fermentation time 184 7 .6 10 l itre batch fermentation of whey permeate ( 100 g /1 lactos e) 190 by KMlODlO XX 7 . 7 10 litre b atch fermentation of whey permeate enriched with 191 molasses by KMlODlO 7 . 8 Growth of KMlODlO in lactose and glucose agars 1 95 7 . 9 Comparison of growth of possible sucrose negative mutants 200 7 . 10 Sequence of isolation of mutants 235C and 256 A 7 . 1 1 Sequence of plating and s treaking to check s tability of mutant 256 A 7 . 12 Streaking sequence to check s tability of 235 c 7 . 1 3 First culture improvement sequence of mutant FSN3 7 . 14 Second culture improvement sequence of mutant FSN3 A . l Pump capacity curves for 25mm� tower fermenter B . 1 K. mar.xianus Y42 cell plate count number vs cell dried weight B . 2 Cell centrifuged wet weight vs cell dried weight C . 1 (a) sucrose and (b) ethanol concentrations vs tower height C . 2 Volumetric rates of (a) sucrose utilization and (b) ethanol production vs mean tower height C . 3 Specific rates of (a) sucrose utilization and (b) ethanol production vs mean tower height C .4 Ethanol yield VS tower height C . 5 Cell concentration vs tower height C .6 Medium pH vs tower height C . 7 (a) sucrose and (b) ethanol concentrations vs residence t ime C . 8 Volumetric rates of (a) sucrose utilization and (b) ethanol production vs mean residence time C . 9 Specific rates of (a) sucrose utilization and (b) ethanol production vs mean residence time C . lO (a) sucrose and (b) ethanol concentrations vs super­ ficial liquid velocity (V ) s C . ll Volumetric rates of (a) sucrose utilization and (b ) e thanol production vs V s C . 12 Specific rates of (a) sucrose utilization and (b) e thanol production VS V s C . 13 Cell concentration vs V s 204 204 206 206 208 226 235 235 239 24 1 243 244 245 247 249 250 25 1 253 255 256 258 LIST OF TABLES Table Title page number 1 . 1 Typical c omposition of whey 2 1 . 2 Composition o f deproteinated whey 2 1 . 3 Est imated quantities o f whey production 4 1 . 4 Whey powder production 5 2 . 1 Fermentation o f concentrated whey 1 6 2 . 2 Ethanol concentration , product ivity , and yield in whey 1 8 fermentation 2 . 3 Inductrial e thanol tower fermentation s tudies or proce sses 27 2 . 4 The effect of sugar concentration on the limiting 32 volumetric e fficiency 2 . 5 Common terms for microbial interactions 3 . 1 Typical composition of sulphuric whey permeate 3 . 2 Yeas t cul tures used 3 . 3 Plat ing dilutions used to determined optimum UV irradiat ion t ime 4 . 1 4 . 2 4 . 3 4 . 4 4 . 5 4 . 6 4 . 7 5 . 1 5 . 2 5 . 3 5 . 4 5 . 5 5 . 6 6 . 1 6 . 2 6 . 3 6 . 4 Lactose utilization Mean specific growth rate Comparison of optimum superficial l iquid velocities Comparison o f the effective tower heights and residence t ime Exi t conditions at velocity of 0 . 24 mm/ s Comparison o f tower fermentation conditions Comparison with batch fermentat ion Percentage sugar utilization Comparison of lactose utilization Specific rate o f ethanol production The concentrations of K. marxianus and S. cerevisiae Comparison between fermentation o f whey permeate enriched with molasses and with sucrose Fermentation comparison using d ifferent ratios of mixed yeast culture in the inoculum Whey permeate as the base medium Molasses medium Lactose as the sole sugar source Maltose as the sole sugar source xxi 42 45 47 66 7 3 1 06 1 08 1 09 1 1 1 1 1 2 1 1 2 1 2 2 1 24 1 3 3 1 36 145 147 1 53 1 55 155 155 xxii Table Title 6 . 5 Glucose as the sole sugar source 6 . 6 Prepared media 6 . 7 Flocculence measurement media 6 . 8 Flocculating ability of some lactose-fermenting yeasts 6 . 9 Flocculation of KM Y42 grown in whey permeate with no additives 6 . 1 0 Flocculation of KM Y42 grown in whey permeate with additives at pH of 4 . 6 and 5 . 0 6 . 1 1 Flocculat ion of KM Y42 grown in whey permeate and add itive : with and without membrane f iltration page 1 5 5 1 56 1 5 6 1 5 9 1 6 1 1 62 1 63 6 . 1 2 Flocculation o f KM Y42 grown in whey permeate supplemented 1 64 with organic nutrients 6 . 1 3 Flocculat ion of KM Y42 grown in whey permeate supplemented 164 with inorganic nutrients 6 . 1 4 Flocculat ion of KM Y42 grown in media supplemented with 165 flocculation aids 6 . 1 5 Flocculation of KM Y4 2 6 . 1 6 Flocculation of KM Y42 6 . 1 7 Flocculation of KM Y42 s. cerevisiae in mixed 6 . 1 8 Flocculation of strains grown in double-sugar in dif ferent media grown as mixed cul ture subs trate CC39 and SC 146 substrates with 1 66 1 6 7 1 68 1 6 9 7 . 1 t-rat io of parameters 1 7 7 7 . 2 Comparison o f fermentation ability o f 4 ethanol tolerant 1 8 6 isolates of K . marxianus UCD FST 7 1 58 7 . 3 Tes t of the stability of KM1 0D10 1 8 7 7 . 4 Summary of fermentation comparison of lac tose fermenting 188 yeas t s 7 . 5 Comparison o f lactose utilization and e thanol product ion 190 of 10 l itre batch fermentation of whey permeate ( 100 g / 1 b y KM10D1 0 and parent strain 7 . 6 Observation of growth of KM1 0Dl0 in glucose and lac tose 1 94 agars 7 . 7 Observation of growth of KMlODlO in glucose and lac tose 1 9 6 agars ( second a t tempt ) 7 . 8 Fermentation test o f isolate obtained from lactose agar 1 9 7 containing 10 g / 1 DGA 7 . 9 Fermentation of whey permeate by two possible sucrose- 2 0 1 negat ive mutants of KMlODlO xxiii Table Title page 7 . 1 0 Fermentation of whey permeate enriched with sucrose by 202 FSN 1 and 2 7 . 1 1 Fermentation comparison of mutants FSN 1 and 2 with parent stra in KMlOD lO 7. 1 2 Fermentation of whey permeate ( l OO g/1 lac tose) by mutant FSN 3 ( cul ture no . 6 ) 7 . 1 3 Fermentation of whey permeate ( 1 00 g/1 lactose) by mutant FSN 3 ( cul ture no.l l ) B. l B . 2 B . 3 B . 4 B . 5 B . 6 Variables and their concentrations used at various RUNS Fermentation results o f fac torial experiment Batch fermentation of whey permeate by K. marxianus Y4 2 A summary of the dimensions o f the tower fermenter Tower fermentation of whey permeate data at various sampling points Tower fermentation of whey permeate data at various tower sec tions B . 7 Data for tower fermentation of whey permeate enriched with molasses at various sampling points B . 8 Data for tower fermentation of whey permeate enriched with molasses at various tower sec t ions 203 207 209 229 229 230 2 3 1 232 233 234 234 B.9 Tower fermentation of molasses dat� at various sampling 236 points B . l O Tower fermentation o f molasses dat� at various tower sec tions 237 C . l Percentage sucrose utilization at various heights 240 C . 2 Mean spec ific growth rate at various · superficial l iquid 259 velocities C . 3 Comparison of tower fermentation of 1 00 g/1 sucrose media 260 LIST OF ABBREVIATIONS PREFIX 6 change in concentration , g/1 or % SUBSCRIPTS a average E effec t ive i condition at a particular tower height or sec t ion 1 lactose o overall r residence t ime s sucrose or superficial t total sugar u substrate utilization NOTATIONS A aluminium sulphate , A12 (so4 ) 3 AFEB attached-film-expanded-bed fermenter B broth B 95% confidence interval uncertainty Ca calcium sulphate , Caso4 CB yeast cleaning buffer ( Caso4wash) CP Candida p seudot rop icaZis CC39 Saccharomyces cerevisiae CFCC39 CSTR continuous s tirred tank reactor D dilution rate DGA D-glucosamine DW cell dried weight , g/1 DW E ethanol concentrat ion , g/1 E ' volumetric rate o f ethanol product ion , g/lh EF extremely flocculent F membrane f il tration ( 0 . 45 �m) FM flocculating medium (acetate buffer) G glucose H , HE tower height , effec t ive tower height , mm or m H* average tower height KL KZuyv ero my ces Z actis �� KZ uyv ero my ces marxianus xxiv XXV KMY42 K.marxianus Y42 LVE l imiting volumetric efficiency M mal t extract broth M* mal t extract broth (Oxoid) Ma maltose MBN modified Burn ' s number MBN* non-standard modified Burn ' s number Me mal t extract powder medium MF moderately flocculent Mo molasses Ms mal t extract syrup (Maltexo ) P whey permeate P percentage uncertainty Pe peptone P4 . 6 whey permeate with no pH adjustment q specific rate of subs trate utilization , g/gh Q volumetric flow rate , ml r (Linear regression) correlation coefficient R rough S substrate concentrat ion , g/1 S ' volumetric rate of substrate utilization , g/lh SC Saccharomyces cerevisiae SC146 S.cerevisiae FT146 (AWRI 350) SG exit specific gravity e SM spent mal t extract broth S substrate utilization , % u T residence t ime , h r T* average residence t ime , h r T overall residence t ime , h ro TS subcultured from the tower fermenter VE effective tower volume , ml V. volume of a sec tion of the tower fermenter , ml � V superficial l iquid velocity , mm/s s VF very flocculent WF weakly flocculent X total cell number or cell weight , cell/ml or g/1 X average cell concentration , g/1 a � K.marxianus cell number or weight , cell/ml or g/ 1 xxvi Y yield cofficien t , yield of ethanol on substrate utilized , % yeast extract 10 100 g/1 whey permeate solution 44 , 46 ratio of lactose to sucrose of 40 : 40 g/1 and 40 : 60 g/1 5 pH 5 . 0 GREEK NOTATIONS � specific growth rate , g /gh V specific rate o f ethanol product ion , g/gh r 112�()� I r ACID CASEIN WHEY� standardization BUTTER ACID CASEIN l '---�-CULTURED PRODUCTS ium morzi 7 i fmmc _. GIBBERELLIC ACID Aspe�gi Z lua rziger - ciTRIC ACID Cnnduia spp. mixed bacterial - METHANE population Morchel la spp. K . marxianun Peniei llium -----. MICROBIAL PROTEIN cyclopium P. javanicum ACETIC ACID - Acetobacter aceti Fig . l . 3 Summary of processes for whey utilization and disposal . 1.-:SA, total of which for food · can:tda E EC of which France Netherlands TABLE 1 . 4 Whey po wder production 1966 1970 1 97 1 1 972 ('000 tons) 2 1 4 zs� 308 346 1 10 1 3 3 1 45 1 7 1 1 9 20 �4 25 93 2 10 274 325 26 80 1 1 5 148 2!1 S I 65 7 1 Germany, FR I S 43 48 56 United Kingdom I t 1 3 1 4 I S Belgium 6 7 1 0 9 Others 7 16 :!2 26 Austria I 7 8 9 Finl:.!nJ 6 1 2 1 4 1 7 Total 1 3 countries 333 53 1 628 722 1 973 338 1 78 24 389 1 70 99 66 1 5 9 30 1 1 1 7 779 1 ton • 1 . 016 tonne (Coton 1976) Whey drying as � � of total whey supply 1 973 40 - 36 30 39 53 26 I S s s 6 35 65 34 6 1. 5 . 1 Whey disposal Unutilized whey has generally been discarded e ither by spray irrigat ion or discharged into natural waterways or municipal sewers . Very few factories have effluent treatment plants ( Bisset t & Riddle 1976 ; Oborn 1968) because of the plant capital cost required to treat the high Biological oxygen demand ( BOD) of whey . Spray irrigat ion i s a common disposal method used in N . Z . About 20% of the effluent from all but ter , cheese and milk powder factories and SO% of the effluent from all casein fac tories are disposed i n this manner . Fac tories are normally s i tuated in rural areas and disposal of e ffluent to s ewage treatment plants is d ifficul t . The area required for spray irrigation is determined by the total waste volume , the applicat ion rate , and the return cycle . The re turn cycle is set by the infiltration rate recovery , which in turn is dependent upon the BOD application rate (Parkin & Marshal! 1 9 7 6 ) . Feeding whey to animals is another traditional disposal method . In the UK , 1 9 73 , 40% of the whey produced was s till used as pig feed ( Co ton 1976 ) . 1 . 5 . 2 Pro tein production One of the oldest means of recovering protein from acid casein whey is the lactalbumin process. The whey is heated to 9 8°C for 20 min and the prote in precipi tate is recovered and dried . The pro duct called "lac talbumin" is tan in colour and has a gri t ty texture which l imits its application in food stuffs ( Robinson et al 1 9 7 6 ) . Ultrafiltration of whey proteins is a new technology which is gaining wide popularity , as the recovered proteins have high solubility , whippability , heat gelat ion , and stabili ty in acid solution. They are used in the flour and sugar confectionery , soft drink and baby foods industries ( Co ton 1976 ) 1 . 5 . 3 Lactose production Lactose is almost exclusively p�oduced from whey and in par t icular swee t whey . It is a major component in the infants or invalids formulae , being added to cow' s milk to approximate the lac tose content of human milk . When added to food , i t contr ibutes to body and viscosity of the food without excessive sweetness; absorbs and stabilizes food and beverage flavours and aromas; promo tes and strengthens agglomerate structure . Refined grade l actose is used in the pharmaceutical industry as a 7 fermentation substrate for antibio tic production , a filler and inert carrier for capsules and tablets . It has also been used in production of S-galactosidase . Lactose can react with urea to form lactosylurea which is used as a nitrogen supplement in rum'inant feeds (Widell 1979) 0 Deproteinated whey can be concentrated by heating at 40-44 C to produce crude lactose through crystallization , ion exchange or electrodialysis . Crude lactose f inds application in formulated dairy foods (Webb & Whittler 1970) . 1 . 5 . 4 Fermentation .of whey Whey can be fermented , us�ng different microorganisms , to obtain many riifferent products . (a) Beverage production Non-alcoholic and alcoholic beverages have been producuced from whey . In the production of non-alcoholic beverages , deproteinated whey is used a s a base for a number o f carbonated ac id beverages fermented by lactic bacteria (Short 1978b) . Alcoholic beverages may be produced by a fermentation in which whey is utilized as the medium for ethanol production or a more readily fermentable sugar such as glucose may be added to the liquor ( Short 19 78b ) . (b) Lactic acid production Lactic acid is a syrupy liquid used in food and beverage manufacture , and is produced by Lactob ac il l us bulgaricus . The lactic acid can be recovered from the fermented l iquor by the addition of calcium lactate , calcium carbonate , and calcium hydroxide . After further treatment sulphuric acid is added to yield lactic acid and calcium sulphate crystals . The lactic acid obtained is cleaned using activated carbon (Campbell 195 3 ) . ( c ) Citric acid production Citric acid is commonly used as a food acidulant because of its ease of assimilation , palatability and low toxicity . Most of the world ' s supply of c itric acid is produced by fermentation of molasses or o ther carbohydrate sources using selected strains o f Aspergi llus niger or Candida sp • • S ince the optimum environment for citric acid production is , for most microorganisms , around pH 2 . 5-3 . 5 , acid whey should be a suitable medium. However , there is little published work on production of citric acid from whey . Generally 15-20% sugar concentration is required for fermentation and if this falls below 5% microbial cells only are produced (Hossain 1983 ) . (d) Acetic acid fermentation Acetic acid is produced by oxidation of ethanol using Acetobacter aceti . Ethanol can be produced from whey 8 and after the yeast has been separated from the broth , a suitable s tarter culture o f Acetobacter may then be added . Vinegar bacteria grow both in the liquid and on exposed surfaces to form a gelatinous film. Large scale fermentation vessels are usually packed columns , which provide the large surface area required for rapid oxygen transfer . The production of vinegar fron whey has already been commercialized in France , USA, and Switzerland (Short 1978b ) . (e ) Microbial protein production Lactose in whey can be converted to microbial protein by fermentation . The protein produced may be processed by cooking , extruding , texturising etc . into food grade products similar to textured vegetable protein or may be hydrolyzed to form the base for s tocks , gravies and sauces or used as animal feed s tuff for high quality meat producing animals such as cattle , pigs and chicken (Berns tein et al 1977 ) Yeasts of the Brettano myces, Candida & Kl uyveromyces spp . can grow on lactose under �aerobic condit ions (Meyrath & Bayer 1979 ; Wasserman et al 1960) . A yield of around 1 . 5 kg yeast mass per kg of whey so l id has been reported (Bernstein & Everson 1974 ) . A plant in France is reported to produce Penicil l ium cyclop ium from whey ultrafiltrate (Kosaric & Miyata 1981) . ( f) Butanol production Butanol is used as cosurfactant in tertiary oil recovery ; solvent for paints , varnish and cellulose esters and its derivatives are used in the pharmaceutical and perfume industries . The bacteria Clo stridium buty licum and Cl . ac et obuty licum are used for its )/ -product ion by fermentaion . Whey is an ideal substrate for fermentation to produce butanol because of its low sugar content , as butanol levels greater than 12 g/1 inhibit the fermentation . The proportion of butanol , acetone and ethanol produced varies according to the strain used but normally the ratio is 6 : 5 : 2 respectively (Gapes 1982 ) . (g) Production of other fermentation products A process for the produc tion of butylene glycol , a raw material for butadiene production , from whey has been described using Baci l l us polymyxa (Speckman & Collin 1982) . Fusarium mi niforme has been used to produce gibberellic acid from whey (Maddox & Richert 197 7 ) . Many digester designs have been described for anaerobic treatment of whey to produce methane (Boening & Larsen 1982 ; Archer , Larsen & McFarlane 1983) . (h) Ethanol production Whey is currently being used as the feed stock for potable and industrial ethanol production in many parts of the world , using Kl uyveromyces marxianus . Presently , there is a plant 9 producing potable ethanol from whey permeate by batch fermentation at Carbery, Southern Ireland ( Sandbach 1 98 1 ) . Microbial protein and ethanol are produced from whey at a Milbrew Inc . plant in Juneau , Wisconsin . This process could operate in a batch, semi-continuous or continuous mode (Anon 1977 ) . A continuous pilot plant has been successfully tested in Denmark (Reesen 1978 ) whereas in New Zealand the New Zealand Co-operative Dairy Co . commissioned its first full scale whey to ethanol fermentation plant in September , 1 980 at Reporoa . This plant utilises lactalbumin serum and a batch fermentation process . A second plant at Tirau , which uses a continuous stirred tank fermentation process and lac talbumin serum , was commiss ioned in September 1 982 . The total annual production will supply all of the 4 . 5 to 5 . 0 million litres market requirements in New Zealand with a surplus of 1 . 6 to 2 . 1 million l itres (Howell 198 1 ) . There are two more batch fermentation plants producing ethanol from whey at Edgecumbe (Mawson 1 983) and Temuka (Goading 1982) . It is worth mentioning here that if all lactose in acid whey produced in New Zealand were fermented to ethanol it would replace only 1% of all New Zealand motor spirit consumption (Marshal! 1 9 78 ) . Its impact will , however, be large in a region where the dairy industry is concentrated such as in the Waikato and Taranaki . If all the ethanol produced was used only within the regions , it could replace a large portion of the petrol used within the regions . Ethanol is used as a chemical intermediate in the synthesis of a large variety o f compounds by simple chemical reac tions , such as reaction with carboxyl ic acids to fats , waxes ; dehydration to produce olefins and ethers ; sulphonation and ethoxylation to form surfactants for use as detergents ; and other synthesis yielding compounds for use as plasticizers , emulsifiers , lub ricants , emollient and foaming agents . It may be dehydrated to ethylene , an important synthetic chemical starting point . It is second in importance only to water as a solvent for drugs , lacquers , perfumes , cosmetic s , detergents and plastics . Recent applications are in aerosol and mouth­ wash products , motor and rocket fuels ( Anon 1974 ) . 1 . 6 SUMMARY When all the fermentation processes described here for the utilization of whey are considered , only beverage , lactic acid and ethanol productions have been carried out on a commercial scale , while the o ther processes are still being investigated . The product ion of ethanol from whey provides a very attractive alternative through its simple fermentation process and . great demand for ethanol . The seasonal nature of whey production and the 1 0 cost o f ethanol from whey are o f considerable importance when considering the alternatives to be used for the treatment o f whey . Greater utilization of whey will help to reduce the need to dump it into natural waterways and reduce environmental pollut ion . The ethanol that can be produced will provide another product for export from New Zealand . CHAPTER 2 LITERATURE REVIEW 2 . 1 INTRODUCTION TO ETHANOL FERMENTATION Ethanol fermentation -. is one of man ' s oldest technologies . The Sumerians and Babylonians were known to produce beer before 6000 BC . Much later , about 4000 BC , the Egyptians discovered the use of yeast for baking and ancient Egyptians relief drawing dating from 2400 BC described the methods of baking and brewing . Kui , a Chinese rice beer , has been traced back to 2300 BC ( Demain & ·Solomon 1981 ) . Past eur , in 1851 , began a research proj ect which led him to the conclus ion that "alcoholic fermentat ion is an act correlated with the l ife and organization of yeast cells" . Subsequently he came to another conclusion that alcoholic fermentation was the result of life without oxygen . Later many other researchers have contributed to the knowledge of the fermentation processes and established that the yeast is responsible for the alcoholic fermentat ion . Today , fermentation technology is very important to man in the product ion o f wine ; beer ; bread ; cheese ; organic chemicals such as ethanol , citric acid , and antibiotics ; animal feed and waste treatment (Nord & Weiss 1958 Harrison & Graham 1970) . Before the advent of the petrochemical industry , all industrial ethanol was produced by fermentation of carbohydrate substrates . However , ethanol production from agricultural sources declined from the 1930 ' s as petrochemical-based ethanol was considerably cheaper to produce . The demand for fermentation-derived ethanol in the food industry has however been � �eady whereas the demand for ethanol as a fuel varied considerably . In Europe , immediately after the first World War , when the supply of petroleum was uncertain and its price was relatively high , there was intensive research into "alcohol fuels" . For a number of inter-war years from 1922 to 1935 an ethanol7petrol blend was the sole motor spirit in France (Titchenerl980) . In the Mackay district of Queensland , ethanol­ petrol blend was in continuous use from 1929 to 1 956 (Titchener l980) . Rac ing car and motor cycle engines have made use of alcoholic fuels for many years . During and after World War II circumstances revived interest in alcohol as a motor fuel in a number of Third World countries such as India . 1 1 12 In the 1970 ' s , the OPEC nations increased the price of oil dramatically and this led to renewed interest in ethanol as an alternative l iquid fuel . There was also a general conclusion that petrol suppl ies will dwindle in the coming decades and must in the near future be replaced by an alternative fuel . This has led many countries including New Zealand to search for alternative fuel sources , and Brazil mounted an ambitious ethanol programme based on sugar cane and other starchy root crops (Tichener 1980 ; Pimentel 1980) . Thus , ethanol may once again be used as a substitute for petroleum-based fuels . 2 . 1 . 1 General biochemis t�y The main substrates for ethanol fermentation are simpl e carbohydrates . These are the hexoses : glucose , fructose , galactose , and mannose ; disaccharides : sucrose , maltose , lactose and the trisaccharide : . maltotriose . More complex carbohydrates , such as s tarches and celluloses , must normally be hydrolysed by acids or enzymes to simple sugars before fermentation can proceed (Kosaric et al 1980) . The conversion by yeasts or bacteria of sugars to ethanol proceeds th�ough a series of enzymatic reactions that were elucidated in the 1930 ' s and which later become known as the Embden-Meyerhof-Parnas (EMP) glycolytic pathway ( fig . 2 . 1 ) . Under aerobic conditions , glycolysis provides pyruvate for entry into the tricarboxylic acid cycle and thence by oxidative phosphorylation ATP , carbon dioxide and water are produced . This gives the maximum quantity of energy per unit of sugar utilized . Under anaerobic conditions or high sugar concentration yeasts convert the pyruvate to ethanol and carbon dioxide . Less energy is produced per unit sugar utilized than under aerobic conditions , since ethanol has remaining elec tron pairs available for oxidation . The s toichiometric relationship for glucose oxidation to ethanol may be used as a basis from which ac tual yields may be evaluated . glucose ethanol Thus , theoretically the yield is 51 . 1 % (W/W) ethanol and 48 . 9 % (W/W) co2 from glucose . However , in practice some sugar is converted to cell mass and other minor products such as glycerol and succ inate , giving a practical ethanol yield o f approximately 90 % o f the above theoretical maximum . These minor compounds are produced only in trace quantities but are of great importance for the aroma and taste of alcoholic beverages (Harrison & Graham 1 9 70 ; Nord & Weiss 1958 ) . �CH 20H H OH H O HO H H OH 0 - GLUCOSE 3- phosphoglycerate CHzO ® I CHOH I coo- ! TH20H CHO® I coo- 2 -phosphoglycerate 13 � �CHzO® H C�Hz� �zOH --. H HO HO OH H OH H OH OH H glucose 6- phosphote fructose 6 - phosphate �ATP f'--ADP g l yceraldehyde CHzO ® C HzO ® 3 - phosphate l �CH20® CHOH -- H HO �HO H OH ! OH H d 1 hydro•y- )HzOH acetone C=O phosphate tHzO ® j.-- NAD+ p , - �NADH ""I\' �HzO ® ATP ADP 7HOH coo® fructose 1 ,6-diphosphate I ,3- d 1phosphoglycerate ·e H 11 2 �o® ADP ATP JJ_ coo- phosphoenol- pyruva te pyruvate C H3 I H - C-OH I H ETHANOL acetaldehyde Fig . 2 . 1 Embden-Meyerhof-Parnas pathway ( Rose 197 7 ) . 2 . 1 . 2 General microbiology There are a large number of yeas ts that can ferment carbohydrates to e thanol . The yeasts from the genus Saccha ro my ces are the mos t widely used and s tudied for potable and indus trial e thanol produc t ion . One limitation of the yeasts from this genus is that they cannot ferment such sugars as lactose , arabinose , cellobiose , rhamnose , sorbose and xylose (Ladder 1970 ; Harrison 19 71 ) . K. marxianus, on the o ther hand , can utilize lactose . I t has been used for s tudies of the produc tion o f microbial protein and ethanol from lactose( Meyrath & Bayer 197 9 ) . It has also recently been shown that some s trains of K.marxianus can ferment xylose to ethanol under aerobic conditions (Margaritis & Bajpai 1982) . Yeasts of the genus Schwanni o my ces have been inves tigated for use in the production of microbial protein or ethanol from starch without the 14 need of an enzyme hydrolysis step (Wilson et al 1982) , whereas Candida tr opi calis and Pachys ole n tannophi lis have been used to convert xylose to ethanol (Wong , Manderson , & Larsen 198 2 ) . Recently , there has been considerable interest in the use o f Zym omo nas mobi lis for ethanol fermentation because of its high specific rates of sugar uptake and ethanol production plus improved yields . However , it can not utilize starch or cellulose directly ( Rogers et al 1980) . A saprophytic filamentous fungus of the genus Mo nilia has been investigated for use in the direct conversion o f cellulose to ethanol (Gong et al 1981 ) . 2 . 2 ETHANOL FERMENTATION FROM WHEY 2 . 2 . 1 Microorganisms Early studies on lac tose fermentation found Candida p se udot ropi cali s (To rul a crem o li s) to be the best o f four yeast species owing to its rapid fermentation and high e thanol yield . It was able to utilize 46 g /1 lac tose in whey in 22 h at 30-34°C (Browne 1941) . In a s imilar study C. pse udot ropi calis required 55 h to utilize 50 g / 1 lactose at 30-34°C (Rogosa et a l 194 7 ) . Al though it appeared to be considerably s lower in this work than in the previously reported work (Browne 1941) it was selected as the most effic ient organism out o f the 11 yeasts s trains tes ted . Among the o ther yeasts tes ted were another s train of C. pse udotr opi cali s, K. l actis (S. lacti s) and K. marxi anus (S. fragi li s) . When the growth medium was concentrated whey , containing 200 g/1 lactose , C. pse udot ropi calis (ATCC 8619) was also selected out of nine yeast strains ( Izaguirre & Castillo 1982 ) . It produced 99 g / 1 e thanol in 192 h , at 30°C . The long fermentation t ime was a result of the completely anaerobic fermentation condition used in this s tudy . K.marxi anus (K.fragili s) was also found to be a rapid lactose fermenter . K.marxi anus CBS 397 fermented 150 g/1 lactose in whey , without any additive , in 72 h at 28 °C (Laham-Guillaume et al 1979 ) . I t was better than 8 other yeasts tested including one C.pse udot ropi cali s and three o ther K. marxi anus strains . A different K. marxianus strain (CBS 5 7 9 5 ) fermented a similar concentration of lactose in whey in 36 h yielding 92% ethanol ( Burgess & Kelly 19 79 ) . Two K. marxianus strains (Yl8 & Y42 ) were found to be the greatest ethanol producers from 100 g/1 lactose medium in comparison with 26 other lactose fermenting yeast strains (Yoo 1974 ) . 15 They produced up to 6 . 3% v/v (50 g/1 ethanol ) in ten days , under s tatic conditions and 28°C . Further information on lactose fermenting yeasts and their synonyms can be found in Ladder (1970 ) or Barnett et al ( 1 9 79 ) . 2 . 2 . 2 Metabol ism of lactose to ethanol by yeast The metabolism of lactose is essentially the same as for glucose ( f ig . 2 . 1 ) with the exception of the lactose hydrolysis and transport mechanism . The lactose is transported across the yeast cell membrane by an inducible specific enzyme system. This is followed by hydrolysis inside the yeast cell by an intracellular enzyme , 8-galactosidase , with the production of glucose and galac tose . Glucose is converted to ethanol via the EMP pathway . Galactose , on the other hand , is converted to D-glucose-6-phosphate through three intermediary react ion steps before it is converted to ethanol via the EMP patHway (Yoo 1974 ) . 2 . 2 . 3 Yeast environmental consideration (a) Substrate utilization Early investigation reported fermentation t ime of 22 h to ferment 46 g/1 of lactose in whey , by C. p seudot� opical is, at 30 °-34 ° C . The resultfng lac tose utilization rate was 2 . 1 g/lh (Browne 1941) . This utilization was considerably higher than reported for a different s train of C.pseudot�op ical is which required 55 h to ferment 50 g/1 lac tose in whey at the same temperature . This gave a lactose utilzation rate of 0 . 9 g/lh (Rogosa et al 1947 ) . These workers did not provide information on residual lactose concentration or indicate v1hether agitation was used . K. marxian us ( CBS 5795 ) has been reported to ferment 50 g/1 lactose in whey in 1 7 h in aerated shake f lasks . The reported lactose utilization was 100% and this gave a rate of util ization of 2 . 9 g/lh . C. p seudot rop icalis (NCYC 744 ) fermented similar medium, under the same conditions , in 12 h which gave utilization rate of 4 . 2 g / lh again with reported 100% lac tose utilization (Burgess & Kelly 1979 ) . Fermentation t ime of 12 h has been reported for different s trains of K. ma�xianus & C. pseudot�op ical is on whey ultrafiltrate containing 50 g/1 lac tose under s imilar conditions as above (Moulin et al 1980) . No data were given on the residual lactose but this gave an average utilization rate of 4 . 2 g/lh . These reported rates were very similar to the commercial scale fermentation t ime of 16 h using whey containing 4 1 g / 1 lactose . The utilization rate was 2 . 8 g/lh (Howell & Tichbon 1980) . 16 When the fermentation condition was completely anaerobic , K. marxian us (NCYC 151) & C. p seudot rop ical is (ATCC 8619 ) required 72 h to ferment 99% 0 of 51 g/1 lactose in whey , at pH 4 . 5 & 30 C giving a rate o f lactose utilization ·of 0 . 7 g/lh ( Izaguirre & �as tillo 1982 ) . This was between 4 to 6 t imes slower than the rate indicated previously when agitation and some air were available . / Concentrated whey Fermentations o f concentrated whey containing between 100 to 200 g/1 have been reported ( table 2 . 1) using both K. ma rxian us and C. p seudotrop ical is. Table 2 . 1 Fermentation of concentrated whey . yeast strains 51o 5 1u S ' 1 t ime T pH reference g/1 % g/lh h oc ( a) partially aerob ic KM CBS 5 795 100 >95 * 3 . 1 32 2 8 4 . 6 1 CP NCYC 744 100 >95 4 . 2 24 28 4 . 6 1 KM CBS 5795 150 >95 3 . 9 38 2 8 4 . 6 1 CP NCYC 744 150 >95 5 . 0 30 2 8 4 . 6 1 CP IP 513 200 >95 3 . 2 63 28 4 . 6 2 (b) anaerobic CP ATCC 8619 lOO >95 0 . 7 9 2 30 4 . 5 3 KM NCYC 151 150 88 0 . 7 192 30 4 . 5 3 CP ATCC 8619 150 9 3 0 . 7 192 30 4 . 5 3 KM NCYC 151 200 60 0 . 6 19 2 30 4 . 5 3 CP ATCC 8619 200 84 0 . 9 192 30 4 . 5 3 KM NCYC 151 250 46 0 . 6 216 30 4 . 5 3 CP ATCC 8619 250 48 0 . 6 192 30 4 . 5 3 KM - K. marxia nus . ,, CP - c. pseudot ropi calis s -lo initial lactose concentration 5!u - lactose utilization s I -1 rate o f lactose utilization t ime - fermentation time T - fermentation temperature . pH - initial pH of whey ' * - >95 indicates complete lactose utilizat ion 1. Burgess & Kelly 1979 2 . Moulin et al 1980 3 . Izaguirre & Castillo 1982 17 Complete lactose util ization was reported for fermentation of concentrated whey containing as high as 200 g/1 lactose under partially aerobic fermentation (Burgess & Kelly 1 9 79 ; Moulin et al 1980) . Fermentation of concentrated whey containing 250 & 300 g/ 1 , under the same fermentaton conditions , was reported to be incomplete but no data were given on residual lactose and fermentation t ime (Moulin e t al 1980) . K. marxi anus (NRRL Y 1109) util ized only 70% of lac tose in concentrated whey . This was reported to be a result of ethanol inhibition (Burgess & Kel ly 1 9 7 9 ) . Fermentation of concentrated whey . under completely anaerobic fermentation condition required long fermentation time ( tabl e 2 . 1 ) ( Izaguirre & Castillo 1982) . There was incomplete lactose utilization , when the lactose content was greater than 150 g/1 , and even after a prolonged fermentation time as long as 200 h . At all concentrations of lactose , C. p seudotropi cali s utilized lactose at a slightly faster rate than K. marxi anus. In summary , complete lactose util ization occurred between 50 and 200 g/1 lactose depending on the organisms . Lactose utilization decreased as lactose concentration increased . Increased lactose utilization and higher lactose utilization rate occurred when there was agitation and partially aerobic fermentation conditions . In all cases , the extent of utili zation of lactose and rate of utilization were also found to be dependent on the strain of yeast . ( b ) Ethanol production Two early independent investigators reported that in the fermentation of whey , e thanol yield based on lactose utilized was 80 & 91% when the lactose concentrations were 4 6 & 50 g/1 , respectively (Browne 1941 ; Rogosa et al 194 7 ) . These workers did not provide data on ethanol concentrations . Later investigators reported ethanol production from 50 g/1 lactose to be from 2 1 to 26 g/ 1 ( table 2 . 2 ) because dif ferent fermentation conditions were used . The corresponding e thanol productivity varied from 0 . 3 t o 2 . 0 g/lh . The lower productivity value of 0 . 3 was obtained under anaerobic fermentation conditions , while the productivity of 2 . 0 g/lh was obtained under partially aerobic conditions . The yield of ethanol on lactose used varied from 79 to 9 7% with no clear trends between different yeast strains and fermentation conditions . Some investigators used a term called " ethanol yield efficiency" which was calculated as the ratio of the amount o f ethanol produced to the amount of total substrate available . The results showed decreasing yield effic iency as lac tose concentration 18 Table 2 . 2 Ethanol concentration , product ivity and yield in whey fermentation . (a ) K. mar xi anus CBS 5 7 9 5 , 25°C , pH s l ' g/ 1 E u g /1 ' Time , h E ' ' g/lh y e / t s ' % (b ) K. marxianus CBS 39 7 , 28°C ' pH sl ' g/ 1 E u g/1 ' Time , g/ 1 E ' ' g/ 1 y e / t s ' % ( c ) K. marxianus NCYC 151 , 30°C . ' ph s l ' g/ 1 E u g/ 1 ' Time , h E ' ' g/lh y ' % lac tose g/1 so 100 150 200 250 4 . 6 , partially aerobic . 50 100 150 28 45 74 17 32 38 1 . 4· 1 . 4 1 . 9 87 84 9 2 na , partially aerobic . na na na na na 25 4 7 6 9 9 0 90 12 na na na na 2 . 0 na na na na 90 90 8 7 8 6 68 4 . 5 , anaerobic . so na 1 32 110 106 22 na 63 29 28 7 2 na 192 96 144 0 . 3 na 0 . 3 0 . 3 0 . 2 8 2 na 89 so 48 300 ref . na 7 9 na na so 1 2 3 (d ) c. pseudotr�picalis NCYC 744 , 28°C pH 4 . 6 , partially aerobic . 1 ' - slu ' g/1 E , g/ 1 Time , h E ' ' g/lh ye/ts ' % (e) C. pseudotropicalis s l ' g/ 1 E u g/ 1 ' Time , h E ' ' g/lh y e / t s ' % so 21 12 1 . 8 7 9 IP 513 , 0 28 C , pH na 25 12 2 . 0 90 100 150 44 70 24 30 1 . 9 2 . 3 8 3 86 na , partially aerobic . na na 200 na 46 70 90 69 na na 63 na na na 1 . 4 na 8 7 89 86 68 (f) C. pseudotropicalis ATCC 8619 , 30° C , pH 4 . 5 , anaerobic . sl ' g/ 1 so 89 140 178 1 14 E u g/1 26 5 1 75 98 32 ' Time , h 7 2 120 192 216 216 E ' ' g/lh 0 . 4 0 . 4 0 . 4 0 . 5 0 . 1 Y , % 9 7 100 100 100 52 E - ethanol concentration ; E' - rate of ethanol product ion Y - ethanol yield on substrate utilized Y I - e thanol yield on total subs trate available e ts 1 . Burges s & Kelly 1 9 7 9 ; 2 . Moulin e t al 1980 ; 3 . Izaguirre & Cas t illo 1982 . 2 na 8 7 na na 5 5 3 19 inc reased . However , the report did not provide all fermentation times or residual lactose concentrations (Moulin et al 1980 ; Burgess & Kelly 1 9 7 9 ) . At higher lac tose concentrations o f 100 g/l , the ethanol concentration var ied from 44 to 51 g/1 , product ivity from 0 . 4 to 1 . 9 g/lh , and yield from 83 to 100% . Again the lower productivity was from anaerobic fermentation . The high value of e thanol yield ( 1 00%) was obtained for C. p9 eudot rop ical is (ATCC 8619) under anaerobic conditions . Some reservations must be placed on these 100% yields because o ther workers found yields of be tween 80 to 91% only for the same yeast species but different strains ( Izaguirre & Castillo 1982 ) . S imilar product ivity and yield were found at lac tose concentrations of 1 50 & 200 g/1 except in the case of K. marxianus NCYC 151 , which produced only 29 g/ 1 ethanol from 200 g / 1 lac tose . When the lactose concentrations were 250 & 300 g/1 , there were poor e thanol production and the maximum e thanol concentration was 90 g/1 under partially aerobic fermentation of 200 g/1 lactose (Moulin et al 1 980) . The ethanol productivity was only 0 . 1 to 0 . 2 g/lh for the anaerobic condition . There were no data available for part ially aerobic conditions and the yield was very low (between 48 to 52%) for the anaerobic condition . Thus , the maximum rate of ethanol production reported was 2 g /lh . The e thanol concentrations and yields were virtually independent of lactose concentration up to 200 g /1 . At higher lactose concentrations , there was poorer ethanol production and much lower yield . Ethanol yield was found generally to be between 79 to 92% . Different yeast strains showed a different capability to produce ethanol . It was found that prolonged fermentation could reduce the ethanol yield because both yeast species changed from lactose to e thanol metabolism (Burgess & Kelly 1979 ) . K. marxi anus � however , grew poorly on ethanol as a carbon source ( Sarfacon et al 197 2 ) . ( c ) Ethanol inhibition Ethanol inhibition was found to reduce lactose consumpt ion and ethanol yield efficiency (Yoo & Mat tick 1969 ) . I t had been found that 3 5 g/1 e thanol could reduce the ac t ivity of B-galactosidase in K. marxianus by 7 7% (Wendorf et al 19 70a) . Adaptation to high ethanol concentration had b een reported to improve ethanol tolerance to up to 79 g /1 for some s trains of K . marxianus (Gawel & Kosikowski 19 78 ) . Di fferent strains of yeas ts had been found to have dif ferent e thanol tolerance . Ethanol added to a final concentration of 20 55 g/1 could kill more than 90% of the initial cell population of K. marxi anus in 48 h. Ethanol was added at the beginning of the f ermentation (Yoo 1974 ) . (d ) Aeration and agitation It has been found that the fermentation rate of lactose-fermenting yeasts were very slow in the complete absence of aeration and agitation (Burgess & Kelly 1979 ; Yoo 1974) . One K. marxi anus s train (Yl8 ) was able to ferment 150 g/1 lactose to completion in 10 days if it were aerated during the initial 24 h. Static cul ture could not u tilize lactose completely after 20 days (Yoo 1974 ) . K. marxi anus ( CBS 5795 ) was found to ferment the same lactose concentration in 38 h under constant ag�tation and p·artial aerobic conditions . (Burgess & Kelly 19 79 ) . ( e ) Nutrient requirements Most reported investigations on fermentations of whey to ethanol did not involve nutrient addition to the whey (O ' leary et al 1 9 7 7 a & b ; Moulin e t al 1980 ; Izaguirre & Castillo 1982) . Burgess & Kelly ( 1979 ) , however , added yeast extract and urea to their whey permeate at a rate of 1 and 0 . 5 g/1 , respectively . Owing to the different fermentation conditions used by various workers , i t was not possible to compare the resul tant fermentation rates . In the production of microbial protein from whey us ing K.marxi anus NRRL Yll09 , the addition of (NH4 ) 2 so4 , K2HP04 and yeast extract at rates ' of 5 , 5 , 1 g /1 , respectively , was required to improve the cell yield (Wasserman et al 1958) . NaCl and CaC12 at a level of 15 g/1 or higher were found to cause noticeable inhibition of fermentation by K.marxi anus (Gawel & Kosikowski 1978 ) . ( f ) Temperature It was reported that between 30 and 4 2°C , a strain of C.ps eudot ropi calis was able to ferment lac tose most rapidly at 3 7°C but the optimal t emperature was fou nd to be 3 3-34°C (Rogosa et al 1947 ) . The rate of lactose fermentation of another s train (NCYC 744) was found to increase with temperature within the range of 25 to 35°C . The time required to utilize completely 150 g /1 lactose was reduced from 22 to 20 h as the temperature was increased from 30 to 35°C (Burgess & Kelly 1 9 7 9 ) . I t was reported that when the fermentation temperature was between 22 to 32°C there was greater cell population at lower t emperatures . As the temperature increased , the t ime required to reach the maximum cell level decreased . The rate of ethanol product ion was more rapid at higher temperatures . The opt imal t emperature for ethanol production from lactose was 28°C (Yoo 1974 ) . 21 (g ) � I t was found that a pH between pH 4 to pH 6 had no effect on the fermentation of lactose to ethanol by e ither C. pseudot ropical is or K. marxianus. The optimal fermentation pH was reported to be between pH 4 . 5 to pH 5 . 0 (Rogosa et al 1947 ; Yoo 1974 ; Burgess & Kelly 1 9 7 9 ) . Viable yeast cells were destroyed faster at lower pH (Yoo 1 9 74 ) . In summary , there were considerable variations in the t ime reported for whey fermentation by lactose-ferment ing yeas t s . The t ime required was ·affected by the initial substrate level , fermentation temperature , inoculum siz e , aeration , e thanol produced , and the yeast s train used . The fermenta t ion t ime was reduced when a fast lac tose-ferment ing yeast strain , increased tempera�ure , inoculum size and sufficient aeration was used . 2 . 3 ETHANOL FERMENTATION PROCESSES . Ethanol fermentation can be carried out in batch or continuous reactors , with batch fermentation being the most widely used process for producing industrial ethanol . Three quarters of the estimated to tal world industrial e thanol production of 3 , 200 million l itres is produced using the batch fermentation process (Maiorella et al 1981 ) . The process is s imple and reliable , but slow. The overall productivity for a batch process is typically between 1 . 8 to 2 . 5 g /lh for fermentation of molasses containing 1 00 g/1 sucrose , with 2 8-32 h f ermentat ion at 30°C , while for whey , rates of up to 2 . 0 g/lh are reported . There are many e thanol fermentation processes in operation and under development which will improve the process over that of batch fermentation . S imple criteria for comparing various processes have been described as low operating and capital costs . Low operating cost is made possible through a continuous process with s imple operation , low energy input and near complete sugar utilization . Low capital cost is the resul t of high productivity through small reactor volume and a mechanically s imple reactor (Ma iorell et al 1981 ) . 2 . 3 . 1 Alternative processes to batch e thanol fermentat ion (a) Continuous st irred tank fermentation ( CSTR) Continuous f ermenters offer many advantages over batch fermenters . These include reduction in operating and maintenance costs , consistency of produc ts · and reduction in processing t ime for the same holding capacity . The process is however less flexible than the batch process and it is essential that the contamination by for��gn organism be 22 avoided , as this can affect the function of the yeast in the system (Hospodka 1966 ) . The theory of continuous fermentation is well reviewed (Hough & Wase 1966 ; Fredrickson & Tsuchiya 1977 ; Levenspiel 1980) . It has been reported that several plants in the USSR use CSTR processes for the production of industrial e thanol (Hospodka 1966 ) . The to tal productivity o f a CSTR system is l imited , however , by the low cell density ( 10-12 g/ lDW) achieved in the fermentation ( typically S-10% of the influent sugar concentration) and the cell maximum specific growth rate . The overall productivity was reported to be about 6 g/lh from 100 g / 1 sugar input which was 3 t imes that . of batch productivity (Maiorella et al 1981 ) . By adding a continuous cell recycle , cell densities as high as 83 g /1 DW could be maintained with ethanol productivity of 30-40 g/lh from 100 g /1 sugar feel (Maiorella et al 1981) . The CSTR systems , however , required constant energy input to provide agitation , to transfer l iquor from one vessel to the next and for centrifugation (Maiorella et al 1981) . (b ) Vacuum fermentation This process was developed to overcome ethanol inhibition in order that a high sugar concentration could be used . By drawing a vacuum of 7 3 kPa ( SS mmHg) , a t 3S°C , ethanol was continuously boiled off from the liquid and withdrawn from the fermenter . In continuous mode , 333 g/1 glucose feed was fermented to give total ethanol productivities of up to 82 & 40 g / lh with and without cell recycling , respectively . The product ivities obtained here were up to 20 to 40 t imes that o f batch productivity (Cysewski & Wilke 1 9 7 7 ) . There were some drawbacks to this process . I t was reported that the energy consumed in providing the vacuum led to a total requirement S% greater than for the conventional process . There was also a constant pure oxygen requirement to meet the yeast oxygen maintenance requirements under vacuum . This added O . S cents (US ) to the cost o f every litre of ethanol produced . The vacuum pumps required must operate at unusually low pressures and are extremely large and they would be difficult to control . There was also the extra capital required to provide the vacuum pumps . Vacuum operation could increase the l ikelihood of fermenter contamination and shutdown . Doubts were expressed as to whether the outstanding productivity of the vacuum fermenter could compensate for these potential difficul t ies (Maiorella et al 1981) . A process called "flash fermentation" modified the s imple vacuum fermentation process to overcome some of its operating difficul t ies . The fermentation was carried out in an atmospheric pressure fermenter so that the yeast oxygen requirement could be cheaply met with sparged 23 air . Ethanol was removed by rapidly recycl ing the beer , via a flow regulating valve , to a small auxillary flash vessel where it boiled off under vacuum . Ethanol was recovered as the f lash vessel overhead product and the ethanol depleted beer was pumped back to the fermenter . The contaminat ion problem w.as greatly reduced as only the small flash vessel was under vacuum . The energy requirement was s lightly lower than the vacuum fermenter but the flash fermentation plant was more complicated thap the vacuum fermenter because i t required an added vessel and associated beer cycling pumps . Again , there were doubts as to whether the high productivities possible would offset the operation and control difficulties and the l ikely high vacuum pump costs (Maiorella et al 1981 ) . (c ) Rapid batch fermentation Rapid batch fermentation is a process that used a large cell inoculum to achieve short fermentation t imes . A honey · solution containing 250 g/1 total sugar was fermented to 100 g/ 1 ethanol in 3 h , at 30°C . The inoculum was 109c ells/ml or SO g/1 DW and this gave a productivity of 3 3 g/lh . The drawback with this fermentation was loss of viability with only 2% viable yeast at the end of the fermentation . Continuous aeration to maintain the dissolved oxygen level at 13% of saturation and reduct ion of the temperature to 15°C increased cell viability to 94% . The fermentation t ime was however increased to 6 h, while · ethanol productivity reduced to 1 7 g/lh (Nagodawithana et al 1974 Nagodawithana & Steinkraus 1976 ) . At this low temperature , energy would be required to provide refrigerated cooling to remove the heat of the fermentation . As this process w as a batch process , the product ivity would be reduced when the down t ime was taken into account . The t ime and fermentation capacity required to prepare such a large inoculum would also be considerable . (d ) Fermentation by immobilized cells Cells can be immobilized by entrapment in a gel matrix such as alginate gel . This process employs the entrapped cells in a packed-bed column fermenter . There is no requirement for agitation or yeast recovery equipment . The medium is fed from the bot tom of the column . An investigation using immobilized Zy mo mo nas 0 mob il ia to ferment 150 g/1 glucose at 30 C , reported an ethanol productivity of 53 g/lh , ethanol concentration of 63 g/1 and a residence of 1 . 2 h. There was 30% reduction in the cell ac tivity after 33 days of continuous operation (Grote e t al 1980 ) . Immobilized S. cerevisiae were employed to ferment molasses solution containing 197 g/1 reducing sugars , with a reportea ethanol product ivity of 25 g/lh , ethanol concentration of 71 g/1 , residence time of 2 . 9 h, and cell half life of 1800 h 24 (Ghose & Bandyopadhyay 1980) . Continuous fermentation of whey using immobili zed yeas ts has been reported . One investigation described s table continuous fermentation operation for one month but the resul ts showed a steady reduction of the ethanol concentration with time (Linko et al 1981) . When the inlet lactose levels in demineralized whey were 50 , 100 , and 150 g/1 , the ethanol levels were initially 25 , 45 and 48 g/ 1 , but after one month , the?e levels decreased to 21 , 30 and 33 g/1 , respectively . Thus , ethanol productivit ies were reduced from 6 . 4 , 5 . 4 and 3 . 2 to 5 . 4 , 3 . 6 and 2 . 2 g/lh , respectively . The residence t ime were 3 . 9 , 8 . 3 and 1 5 h , respect ively . The initial pH and temperature were 4 . 5 and 25°C , respectively . K. marxi anus (NRRL Y ·1109) immobilized in acrylamide gel was used in a different s tudy to ferment whey containing 5 1 g/1 lactose ( Dillon 1980) . -1 The optimum dilution rate reported was . 15 h which corresponded to a residence t ime of 6 . 8 h . The productivity obtained was 3 . 1 g /lh which resulted in 21 g/ 1 ethanol at the exi t . There was 93% lactose utilization and 79% yield of ethanol on lactose utilized . The yeast cells were found to lose less than 10% o f the initial fermentative activity in 10 days and had a half l ife of 50 days . Finally , this process may give considerably higher e thanol productivity than batch fermentation but the cells have a short stable life . This means frequent replacement of immobilized cells in order to maintain continuous fermentation operat ion . I t had been estimated that gel cost contributed24% of the operating cost of a fermentation-distillation plant producing ethanol from whey as a substrate , using entrapped K.marxi anus with half l ife of 50 days (Dillon 1980) . A modification to the fixed bed reactor system used in the immobilized cell fermenter is the attached-film-expanded-bed fermenter (AFEB ) in which cellulose acetate was used as the yeast supporting medium .When ferment ing a 100 .g /1 lactose feed , the residual lactose increased gradually from 22 to 83 g/1 , while ethanol decreased from 35 to 7 . 6 g/1 as the dilution rate increased from . 07 to 0 . 9 h-1 . The e thanol productivity was 2 . 4 g/lh at 1 -1 dilution rate of . 07 h- whereas it was 6 . 9 g/ lh at D of 0 . 9 h . The corresponding values of sugar utilization were 78 & 17% , respect ively (Chen & Zall 1982) . ( e ) Tower fermentation The continuous tower fermentation process employs a s ingle tubular reactor with a conical bottom and an expanded top section to facilitate yeas t settling . The fermentor is filled with flocculent yeast and the cell concentration of 200 to 400 g/1 WW is maintained in the tower . Typical ly , an e thanol productivity of 10 g/lh 25 and concentration of 39 g /1 are ob tained at a residence time of 3 . 8 h from a feed beet molasses solution containing 100 g /1 reducing sugar . The advantages of tower fermentation l ie in the s implicity of construction and mode o f operat ion . No agitation is required to keep the yeast cells in suspension . Yeast separation and recycling are carried out inside the tower (Hough et al 1976 ; Coote 1974 ) . Thus , there is considerably less energy and capital requirements compared with fermentation syst ems described previously . Tower fermenters have been operated successfully on a commercial scale to produce different products such as beer and vinegar ( Greenshields & Smith 1974 ) . 2 . 3 . 2 Industrial and pilot plant processes used for the produc t ion of e thanol from whey At present , there are 5 plants producing e thanol from whey by batch fermentation . These plants are located at Reporoa , N . Z . (Howell 1981 ) ; Edgecumbe , N . Z . (Mawson 1983) ; Temuka , N . Z . ( Goading 198 2 ) ; Carbery , Ireland (Sandbach 1981) ; and Juneau , Wisconsin , USA (Anon 1 9 7 7 ) . The process at Reporoa which is s imilar to the process at Carbery produces 18 g/1 ethanol from 41 g/1 lactose in 16 h . This gives an e thanol yield of 90% and product ivity of 1 . 1 g/lh . There were no technical data available for other plants . A commercial continuous fermentation process employing three fermenter tanks in series (CSTR) is used at Tirau , N . Z . The plant processes a maximum of 1500 m3 of deproteinated whey per day to produce 32 , 000 l itres of 96% ethanol (Howell 1981) . A pilot plant scale inves tigation was carried out in Denmark . The process employed two fermenters in series to carry out continuous fermentation . The minimum residence t ime indicated was 12 h (Reesen 19 78 ) . Thus , among all the ethanol fermentation processes considered i9 these last two sections (sect . 2 . 3 . 1 & 2 . 3 . 2 ) , with the exception of the batch process , only the CSTR and the tower fermentation processes are in commercial use . The tower fermentation process f its all the s imple criteria of low operating and capital cos t required to improve the economy of the process . Its high productivity is surpassed only by the energy intensive vacuum fermentation process . I t has better productivity than the CST� processes and is c�nsiderably less and capital intensive . 26 2 . 4 ETHANOL PRODUCTION BY TOWER FERMENTATION 2 . 4 . 1 History of tower fermentation The concept o f tower fermentation was described as early as 1945 (Alzola 1945) . The process described employed a column divided into s ix parts . The feed medium was pumped in from the bottom of the column and the carbon dioxide produced from the fermentation provided agitation . A similar design was subsequently used for continuous fermentation but with the _ feed medium gravity fed from the top of the column ( Owen 1948) . A process which employed a series of tanks s tacked one on top o f another to form a tower was patented in the same year (Vic torero 1948 ) . The feed l iquor entered from the bottom of the lowest tank , passing through the upper tanks and out at the top of the tank . A continuous fermentation process whereby a s ingle vert ical tubular ferment er was employed , was patented in 1960 (Watson & Shore 1960) for . APV Co Ltd . The tower fermenter was filled with a high concentration o f microoganisms and had an expanded top sect ion which permit ted the microorganisms to settle and return to the tower ' s lower sec t ion . This process was intended for beer product ion but o ther applications have been described ( Smith & Greenshields 1974 ) . Other types o f tubular column fermenters have been described by various workers , such as the gradient-tube continuous fermenter (Portno 1967 ) , mult istage tower fermenter ( Paca & Gregr 1979 ) , tower fermenter with two fluid feed nozzles ( Go to et al 1981 ) , column fermenter packed with immobilized yeast (Linko & Linko 1981 ) . A summary of tower fermenters is given in table 2 . 3 . The following is concerned mainly with the APV tower fermenter type . 2 . 4 . 2 The characteristics and operation of the tower fermenter The tower fermenter behaves as a heterogeneoss system , with a progression from substrate solution to fully fermented l iquor within the s ingle vessel . The APV tower fermenter is made up of a vertical cylindrical tower with a conical bottom ( fig . 2 . 2 ) . The medium enters at the bot tom and passes up through a dense suspension of yeast cells . At the top , the vessel expands into a large settling zone in which yeast is separated from the effluent l iquor and returns to the main body of the tower , while the clear alcoholic l iquor overflows from the top of the tower . Fermentation proceeds as the l iquor rises while the yeast cells settle back and are retained . Thus , the superficial medium velocity mus t not exceed the settling velocity of the cell floes . The APV tower fermenter 27 Table 2 . 3 Industrial ethanol tower fermentation s tudies or processes . reference fermenter s izes & other details . Alzola 6-sections column ; mash enter from the bottom & agitated by co2 1945 produced (patent) . Victorero tanks stack one on top o f another 1948 at top (patent) . mash enter from bottom & exit Owen 73mm� x760mm high ; s ix-sec tions ( l08mm . long) -glass tower 1948 molasses feed from top to bottom . media na na molasses Wa tson & Shore APV tower , s ingle stage tower using flocculent yeast ; l . lm� x 10 various m high & 1 . 8m� x 12 . 5m high ( 6 . 5 & 7 . 2m straight section height 1960 respectively ; commercial s izes) ; lSOmm$ x 7m (p ilot s ize) . Coote 1974 Pyrex , 2 5 . 5mm� x 1 . 75m high (887ml) + 80mm� x . 54 m h igh (2 . 714 1) + expanded spherical section 60mm� x 60mm liquid depth ( 5 7ml) ; total heigh t 2 . 35m varying feed rate ; sugar concentrations 100-250 g/ 1 ; used flocculent S. ae�evisiae ; 20-28°C . lager beer beet molasses Henderson Perspex , lOOmm$ x four 1 . 5 1 sections ; 6 1 total volume ; 3 fodder beet & Smith perforated plates ; feed rate varied ; sugar concentrations sugar beet 1982 100-150 g/1 ; used flocculent S. ae�eviaiae ; 35°C. extracts Prince & Pyrex , 7 5mm� straight sect ion ; l SOmm� yeast settling section ; cane j uice Barford overall H/� 22 : 1 ; 10 . 1 1 total volume ; flocculent S. ae�evisiae, Whey 1982 S. diastatiaus, & Z . mobiZis with flocculent aid ; feed rate varied cane stillage ; sugar concentrations 100-250 g/ 1 ; 30°C . $ - internal diameter H - height Fig . 2 . 2 Schemat ic diagram of the APV tower fermenter ( Greenshields & Smith 1971 ) . 28 used for commercial beer product ion has an overall aspect rat io ( s traight section to internal diameter) between 7 : 1 to 10 : 1 and tower diameters from 0 . 9 to 2 . 0 m (Royston 1966 ) . Baffles are incorporated within the tubular sec tion of the fermenter to reduce gas channelling and prevent back mixing of the yeast and the beer (Klopper et al 1965) . At relatively low liquid flow rates , the tower fermenter can be considered as a fixed bed , catalytic reac tor and at relatively high flow rates as a fluidized bed catalyst reactor . S toke ' s law has been used to described the characteristics of yeast floes inside tower fermenters . However , i t was found that this could not adequately describe the characteris tics of the yeast floes (Greenshields & Smith 19 71 ) . A summary of the equation for estimat ing the hindered set t ling velocity of particles in a liquid has been described by Boening & Larsen (1983) but these require exact knowledge of particle s ize , density and spherici ty as well as the voidage of the fermenter "if reasonable result s are to be ob tained . (a) Organisms for tower fermentation A flocculent s train of organism is essential in tower fermentation to prevent wash out , as the density of the yeast cells differs only sl ightly from that of the l iquid . Flocculent yeast that have been used in tower fermentation are from the genus Saccharo my ce s. One study (Coote 1974 ) carried out extensive tests on flocculent Saccharo my ce s strains for tower fermentation of beer and molasses . More than 20 s trains of flocculent yeasts were l is ted . S. cerevi si ae NCYC 1257 was used in the s tudy on tower fermentation of cane juice ( Prince & Barford 1982 ) . S. cerevi si ae CFCC 39 was used in the tower fermentation of beet j uices (Henderson & Smi th 1982) . A flocculent s train of Z . mobi li s has also b een tes ted for use in ethanol production from cane j uice using a tower fermenter ( Prince & Barford 1982) . Further discussion on yeast behaviour in tower fermentation is given in sect . 2 . 5 . ( b ) The effec t of tower height In tower fermentation o f beer there is a gradual decrease in the wort specific gravity with increasing height from the tower inlet ( fig . 2 . 3 ) . The wort specific gravity decreases sharply from 1 . 036 to 1 . 010 for most flowrates over the first half of the tower , whereas the decrease is lower ( 1 . 010 to 1 . 005 ) over the second half of the tower . As the superficial l iquid velocity increases there is an increase in the specific gravity at various tower heights . The initial rapid fall of wort gravity is due to fermentation of glucose , fruc tos e , sucrose , and some maltose , whilst the slower fall over the middle and top o f the tower is due to fermentation o f the remainder o f the maltose u ... u w 0.. "' 29 and maltotriose in the wort . I t is also found that , at a particular superficial l iquid velocity , higher wort gravity resul ts in higher final specific gravity of the effluent beer . 1 03 0 0 2 1 6 V mm/ s s 0 . 3 6 X . 4 5 6. . 5 0 • . 5 7 0 . 6 4 • . 7 2 20 D I STANCE FR0"4 I N L E T l f t l 6 s m Fig . 2 . 3 Progressive reduc tion of wort gravity in a tower fermenter at various flow rates using physically l imited yeas t . ( Greenshields & Smith 1971 ) (c) Residence t imes in tower fermentation There were considerable differences between the residence t imes for the gas , liquid , and solid phases in a tower fermenter ( Smith & Greenshields 1973) . The rise velocity of the gas bubbles was reported to be dependent on the floc size and concentration and medium throughput . In a full scale (0 . 9 m diameter) tower , reported r ise velocity was 0 . 1 m/s and the residence t ime was 80 s . The average residence time for the yeast in tower fermentation of beer has been estimated to be approximately 100 h ( Smith & Greenshields 19 73 ) . At present there is n0 published informat ion on the effect of the medium specific gravity , feed rate , aeration and foaming on the residence t ime of yeast in tower fermenters . In beer fermentation , l iquid residence t imes between 4-6 h have been reported for wort with an initial specific gravity o f 1 . 035 (58 g/1 reducing sugars ) , with 81% sugar utilization , at 21°C . The sugars present in the beer wort were found to have different fermentation t imes inside the tower . No free sucrose was detec ted after 25 min , whereas glucose and fructose were both utilized in 40 min . Some mal tose was utilized during �his period but 2 h was required for complete utilization of mal tose . Mal totriose required 3 . 75 h to be completely utilized with 30 virtually no util ization in the first 1 . 5 h. Thus , the necessary 4-6 h residence t ime in beer tower fermentation was due to the t ime required to ferment maltotriose . This still compares favorably with 48 to 72 h required for batch fermentation (Ault et al 1969 ) . This difference in t imes is reflected when using a plo t of wort gravity vs apparent fermentation t ime ( f ig . 2 . 4 ) . It shows that the wort gravity is reduced from 1 . 035 to 1 . 01 0 in 1 h but further reduct ion from 1 . 010 to 1 . 005 required 19 h (Klopper et al 1965 ) . 1 0 3S r---�----�----�-----r----�----, I I I I ' > .... u , 030 , 025 � 1 0 1 5 � I � 1 0 1 0 \ • �X o x-.. V s mm/ s 0' . 071 X . 14 L::. . 21 0 . 28 • . 36 1 oo5 r- -"-l'lo----o I I - - - - - - o � e 1 2 1 6 20 2 1. A P P A R E N T F E R M E N TAT ION T IME l hou r s l Fig . 2 . 4 The relat ionship b etween feed liquor specific gravity and apparent fermentat ion t ime in tower fermentation of beer using physically l imited yeast . (Klopper et al 1965 ) In the fermentation of beet molasses (Coote 1974 ) , containing 100 g/1 reducing sugars , the limit ing volumetric eff iciency ( the maximum ratio , between the volume of beer produced in 24 h and the void volume of tower fermenter , at which there was 80% sugar utilization using feed l iquor of specific gravity 1 . 050) was found to be 4 . 7 or a d ilution rate of 1 . 1 h-1 . This was equivalent to an est imated T of 0 . 9 h. The exit r ethanol concentration was 39 g/1 . (d ) The effect of feed l iquor specific gravity on limiting volumetric efficiency I t was found that , for a particular original feed l iquor gravity , there was a limit ing volumetric efficiency a t which wash out occured . This i'Ilcra.ased rapidly as the original gravity was decrease d from 1 . 090 to 1 . 030 ( fig . 2 . 5 ) . When the original gravity was 1 . 030 , the volumetric efficiency could be increased without wash out ( Greenshields & Smith 1971 ) . This made the tower fermenter ideal for use in the fermentation of dilute feed medium if an appropriate flocculent organism were available . 31 1 1 00 r-�---r--�--r--.---r--;l---,--,--, > "' a: 1 0 80 "' 1 060 u u. u w Cl. "' .._ 1 OLO a: 0 � ..J "' � 1 020 Q: 0 o laboratory tower • commercial tower \ - \ . � - - 1 ooooL.......l---2.1-�l--��.L------:-s--�6;---::7---;----:-�,o VOLUt-IETR I C E F F IC I ENCY l d - 1 ) Fig . 2 . 5 The effect of original feed l iquor specific gravity on the l imiting volumetric efficiency for a physically limited yeast . (Greenshields & Smith 1 9 7 1 ) ( e ) Aeration Aeration was found t? be important in the operation of the tower fermenter (Ault et al 1969 ) . It allowed a greater degree of yeas t growth and improved yeast viability . During anaerobic fermentation , aeration of the feed medium to 70-75% saturation , was found to be sufficient t o maintain yeast viab ility . During the cell build up period , the tower was aerat ed directly . The superficial air velocity was l imited to 10 mm/s to avoid foaming and slug flow . The typical air to l iquid volumetric flow ratio used was 50 : 1 . 2 . 4 . 3 Applications of tower fermentation The APV tower fermenter has been used successfully in ale brewing . The t ower fermenter has also been reported to be used commercially for the product ion of vinegar ( Greenshields & Smith 1 9 7 4 ) . The product ion o f microbial protein and waste treatment using the tower fermenter have been investigated (Morris et al 1973 ; Ewen 1980 ; Burgess 1 9 7 7 ) . At present there is no commercial application of the tower fermenter for industrial ethanol product ion although there have been a few laboratory scale investigations . 2 . 4 . 4 Industrial ethanol tower fermentations (a ) Laboratory scale investigat ions It was indicated previously that there is no commercial scale use of the APV type tower fermenter for industrial ethanol product ion . There were three reported laboratory scale investigations . The details of the tower sizes used are given in table 2 . 3 ( Coote 1974 ; Henderson & Smith 1982 ; Prince & Barford 1982 ) . 32 (b) Medium used and the effect of sugar concentration Industrial ethanol production by tower fermenter reported to date has used only sucrose based substrates such as b eet molasses (Coote 1974 ) , cane j uice (Prince & Barford 1982 ) , fodder and sugar beet extracts (Henderson & Smith 1982 ) . The sugar concentrations used vary from 100 to 200 g/ 1 . In the tower fermentation of beet molasses (Coote 1974 ) , i t was found that , at a particular reducing sugar level , there was an optimum feed rat e or volumetric efficiency at which 80% sugar utilization occured and above which the utilization decreased and cell wash out occurred . The l imiting volumetric efficiency decreased with an increase in the reducing sugar concentration of the feed medium ( table 2 . 4 ) . The e thanol product ivity also decreased with an increase in the sugar concentration . The reduction in the productivity was due to LVE as percentage sugar ut ilization and percentage ethanol yield remained constant . The reduction in the LVE was a resul t of the reduct ion in the yeast set tl ing veloc ity due to increase in l iquid specific gravity with increase in sugar concentration . Table 2 . 4 The effec t o f sugar concentration on the l imiting volumetric efficiency (LVE) in the tower fermentation of molasses ( Coote 1974 ) reducing sugars , g/1 100 150 175 200 LVE , day -1 4 . 7 2 . 5 2 . 0 1 . 1 V s ' mm/ s . 5 3 . 29 . 2 3 . 13 -1 D , h 1 . 1 . 59 . 47 . 26 *ethanol product ivity , g/lh 10 8 . 0 8 . 2 4 . 9 * The productivity was calculated based on the overall residence t ime inside the tower including the residence time inside the separator . In the tower fermentation of cane j uice (Prince & Barford 1982) i t was found that the dilution rat e required t o achieve a given sugar utilization and product ivity was at an optimum when the sugar concentration was 123 g/1 ( fig . 2 . 6 ) . As the sugar concentration was increased further , it was necessary to reduce the dilution rate and hence the productivity in order to maintain the same level o f sugar utilization . For 95% sugar utilization , the optimum dilution rate and ethanol product ivity were . 2 7 h-l & 16 g /lh , respectively . ...-l I .c: .. • 1ii a: c 0 ] 0 33 0·7 / / 0·6 /" .c: 0·6 ...-l -tiO • .. 0·4 / > Converalon .. ·:; ·� u :J "0 0·3 0 � a. 0 c Cl .s:. .. w 0·1 35 30 25 20 15 10 5 ;; • I • Conver&ion i � ���:: ............ 0 I 1 0 0 1 2 0 1 4 0 1 6 0 1 8 0 1 0 0 1 2 0 1 4 0 1 6 0 1 8 0 FERMENTABLE SUGARS , g/1 Fig . 2 . 6 The effect of fermentable sugars concentration on the dilution rate and ethanol productivity of tower fermentation of cane juice (Prince Barford 1982 ) . In the tower fermentation of beet juices containing 100 g/1 sucrose (Henderson & Smith 1982) , it was found that there was 91 & 77% sugar -1 utilization at dilutions rates of . 20 & . 33 h , respectively . The ethanol concentrations and the productivities were 49 ,40 g/1 and 9 . 8 , 13 -1 g/lh , at D of . 20 & . 33 h , respectively . When the sucrose level was increased to 150 g/1 , it was found that there was no sugar utilization after the ethanol concentration reached 60 g/1 due to ethanol inhibition . This corresponded to 70% sugar utilization . In summary , the dilution rate decreased as the feed sugar concentration increase�� The ethanol productivity decreased with an increase in the sugar concentration . ( c ) Fermentation temperature The operating temperature of tower fermenter differed from one group of investigators to another . In the tower fermentation study of beet / molasses the temperature used varied 0 from 20 to 28 C . It was found that this temperature increase resulted in 6% increase in substrate utilization rate from 193 to 204 g/lh . There was 4 . 2% increase in the e thanol production rate but 1 . 5% reduction in e thanol yield efficiency . The concentration of the fermentable sugar used was 34 150 g/1 and the volumetric efficiency was 7 . 2 (V = . 82 mm/s) ( Coote s 1974 ) . A temperature o f 30°C was used during tower fermentation o f cane j uice ( Prince & Barford 1982) . In the tower fermentation of sugar beet extract , a higher t emperature of 35°C was used (Henderson & Smith 1982 ) . ( d ) Operating pH No data on the effec t of pH of operat ion has been reported in the l iterature . A pH of 4 . 5 was used during a tower fermentat ion of beet molasses ( Coote 1974 ) . This was the same pH used in a different investigation to ferment sugar beet solution (Henderson & Smith 1982) . A lower pH o f 4 . 0 was used in a tower fermentation o f cane juice ( Prince & Barford 1982 ) . ( e ) Conclusions Finally , it is evident that there have been very few investigations on industrial ethanol tower fermentation . The three groups of workers reviewed here used different approaches to present their data and so it was dif ficult to cbmpare their results . None of the three papers referenced indicated direc tly the straight section o f the tower . �coo te ( 1 9 74 ) used the parameter "volumetric efficiency" . This is a very difficul t parameter to use in comparison with other workers because this parameter is not very often quot ed by o ther workers . In order to estimate the superficial l iquid velocity and dilution rate , Coote ' s values were mult iplied by 208 . 3 t o give a medium feed rate in ml /h . This conversion factor was ob tained from the author ' s data which indicate that a volumetric efficiency of 4 . 6 7 was equivalent to 9 7 2 ml /h . The ratio of these f igures gave 208 . 3 ml /h per one unit of volumetric efficiency , assuming that the relationship was constant for all feed rates . The feed rate values in ml /h were then converted to superficial l iquid velocity and dilution rate using the quoted informat ion on the tower diameter ( 25 mm) and tower height ( 1 . 75 m) . The results reported by the three groups of investigators showed that tower fermenters gave high ethanol product ivities ranging from 5 to 16 g/lh . The residence time varied from 0 . 9 to 3 . 9 h for 80% sugar utilization upward . These values were affec ted by the concentration o f the feed sugar and temperature . 3 5 2 . 5 FLOCCULATION OF YEASTS 2 . 5 . 1 Flocculent yeast classification Yeast for ethanol fermentation in tower fermentation can be classified into three group s , based on their flocculent properties (Greenshields & Smith 1971 ) . (a) Non-flocculent These yeasts do not attain a concentration above 30 g/1 centrifuged wet weight (WW) and are rapidly washed out from the tower . (b) Flocculent-physically limited These yeas ts at tain a concentration in the tower of 200-300 g/1 WW and form fine floes up to 2 mm in diame ter . At any part icular specifi� gravity of the wort , there is a critical superficial l iquid velocity which , if excee ded , causes a complete wash out of this type of yeast . At superficial l iquid velocities up to this critical . wash out value the normal fermentation of the wort is achieved . (c ) Flocculent-fermentation l imited These yeasts attain high cell concentrations ( 250-400 g/1 WW) and form heavy " s ticky" floes up to 1 3 mm in diameter . They are retained in the tower at all wort gravit ies and up to relatively high l iquid velocities . At a cer tain critical superficial l iquid velocity (V ) , the complete fermentation of the wort can no longer s be achieved with this class of yeast s due to insufficient l iquid residence t ime . If the superficial liquid velocity is furthur increased complete wash out occurs . 2 . 5 . 2 Quantitative measurement of yea st flocculation Several methods have been used by various workers for the measurement of yeas t flocculation , of which some have undergone subsequent modification . (a) Burn ' s method as modified by Helm et al ( 1953) Burn ' s method is a relatively simple method of flocculation measurement . One gram WW of yeast cells was carefully weighed into 10 ml calc ium sulphate solution in ( 0 . 51 g/1) - a graduated tapering 15 ml centrifuged tube . Af ter mixing , the tube was allowed to s tand . The volume of yeast sediment after 10 min became the Burn ' s number . It was , however , found that the Burn ' s number could not be related to the performance of flocculent yeas ts in the tower fermentation . Different s trains of flocculent yeasts gave different l imiting volumetric efficiency in the tower fermenter whereas the Burn ' s number could not dis tinguish between them (Greenshields et al 19 71 ) . (b ) Burn ' s method as modified by Stewart ( 1975 ) This was another s imple method very similar to the one described previously (sect . 2 . 5 . 2a ) to determine quickly yeast flocculence . Yeas t cells cleaned in deionized water were suspended in 10 ml of deionized water containing 80 mg/ 1 36 calcium chloride at pH 4 . The degree of flocculence was determined after 10 min o f mixing by visual examinat ion using a subj ective scale ( 0-5 ) ranging from extremely flocculent ( 5 ) to non-flocculent (0 ) . This method has the same l imitation as the previous method but any amount of cells can be used . Thus , it can be used to quickly distinguish between flocculent and non-flocculent yeasts before the more tedious modified Burn ' s number is determined . (c) Sharp' s method (Greenshields e t al 1 9 7 2 ) A further modification to the method described in sect . 2 . 5 . 2 (a ) gave flocculence as modified Burn ' s number (MBN) . This value was obtained by plot ting the yeast sedimented volume against t ime at 1 min interval for 15 min on 1 / 10 inch graph paper ( fig . 2 . 7 ) . MBN V . is 1 V -V { ( 0 1 ) tl-tO � 10 � � 8 ...:l 0 :> H 6 z � 4 H Q J:il U) � 2 :> o - 1 1 I I I I I I I I 1 -� !1-.l on o - 1 � 5 10 15 t , TIME , m in v1-vr: + ( :) ) t5-tl + v5-vlO ( ) t1o....;t5 the sediment volume (ml ) at MBN 0 to 5 non-flocculent 90 120 flocculent 20 25 V -V + ( 10 15 ) } x l tl5-tl0 4 t ime t . (min) . 1 Fig . 2 . 7 Determinat ion of modified Burn ' s number (MBN) (Greenshields e t a l 1 9 7 2 ) . The slopes of the straight l ines between 0-1 , 1-5 , 5-lO , and 10-15 min were summated and divided by 0 . 4 to give the MBN value . The fac tor o f 0 . 4 was necessary because 2 ml o f yeast was plotted as 1 inch while 5 min was plotted as 1 inch ie . a slope of 1 was equivalent to 0 . 4 ml /min . The method is more tedious and t ime consuming tha� the previous two methods . I t was , however , found that the MBN value can be related proportionately to the performance of yeasts in the tower fermentation . This was because the method takes into account the yeast set tled volume during the first five min where very rapid settling of flocculent yeast 37 occurs . The previous two methods did not take this into account . Very f lo cculent yeast gave MBN values as high as 160 . For tower fermentation , the yeas t should have MBN greater than 70 (Coote 1 9 74 ) . (d) Spectrophotometric method A spectrophotometric method measured the absorbance of 0 . 1 g of washed yeas t suspended in 4 ml acetate buffer in a 10 mm spec trophotometer cuvette (Greenshields et al 1 9 7 2 ) . The yeast floes were loosened and dispersed by gentle agitation with a small spatula followed by slow inversion of the cuvet te 10 t imes . The cuvet t e was then inserted rapidly into an automatic recording spectropho tometer . Flocculent yeas ts showed rapid decrease in absorbance (at 670 nm) from 1 . 8 to almost zero in less than 40 s , whereas non-flocculent yeasts showed lit tle decrease in absorbance with time . The absorbance normally· remained at 1 . 9 after a time period o f 2 . 5 min . This method is s t ill in the development stage . 2 . 5 . 3 Factors influencing yeast flocculence The tower fermentation system can operate only with a highly flocculent yeast (Royston 1966) . There are two determining fac tors that influence the flocculence of yeas t , one being the genotype and the o ther external fac tors which refer to the growth medium and other environmental factors such as temperature , pH , chemicals and the presence of o ther organisms (Atkinson & Daoud 1976) . Microbial floes and flocculation in fermentation process es is well reviewed (Atkinson & Daoud 1976) while the flocculent behaviour of the Sacchar omyces spp is extensively studied particularly in brewer ' s yeasts ( S tewart 1975 ) . The flocculent behaviour o f yeas t described here applies mainly to Sacchamy ces spp unless s tated o therwise . 2 . 5 . 3 . 1 Inherited flocculent properties in yeas t ( a) Flocculent genes It is known that there are four genes regulating the flocculat ion in yeast . They have been labelled as FLO 1, FLO 2, fl o 3. and FLO 4. All are dominant genes except "f'l o 3" (Stewart & Russel 1 97 7 ) , and only one of these genes need to be present for flocculation t o occur . Spontaneous gene mutation or mitotic segregation rates from f locculent to non-f locculent are high and much higher than those rates in the reverse d irection (Lewis et al 1 9 76 ) . It is known that spontaneous yeast mutation can occur and is considered to be of importance in continuous f ermentation X' (Thorne 196 8 , 1970) . (b ) The yeas t cell wall The inheritable character is expressed in the yeast cell wall structure , which shows differences in various yeast 38 cul tures , grown under d ifferent conditions ( Jayatissa & Rose 1 9 7 6 ) . The yeast cell wall is thought to consist e ssentially o f two layers , the outer mannan�phosphate protein layer which is connected to the inner structure glucan layer . In general the relative composition of each component is 40% glucans , 4 0% mannans , 10% proteins and the remaining proportion are hexosamine , l ipids and inorganic materials ( Stewart 1975 Lyons & Hough 19 71 ) . · (c) Comparison of the cell walls of flocculent and non-flocculent yeasts There was evidence of a higher level of phosphorus in flocculent walls than in non-flocculent walls (Lyons & Hough 1970 , 1 9 71 ) . 7 0% of the phosphorus was incorporat�d as phosphodiester which bonds mannan and protein of the yeast cell wall together . Walls from flocculent yeas ts were found to bind on average twice as much calcium as did wall s from non-flocculent yeas ts . Removal of the phosphorylated manna-protein decreased the capacity of the cell walls to b ind calcium ions and rendered the wall non-flocculent . It was pos tulated that the phosphate of the glycoprotein complexes with calcium ions to form bridges between adjacent cells . In non-flocculent walls the level of the phosphate is presumably not sufficiently high to permit formation of s table sal t bridges (Lyons & Hough 1970 , 1 9 7 1 ) . Subsequent studies on yeast mannan structure of a number o f S.cerevisi ae mutants which have altered mannan structures found a mutant which had a lower phosphate content than its parent strain but exhibited a much higher degree of flocculence ( Ballou e t al 1973) . These workers also detected no difference between the phosphate level of flocculent and non-flocculent cell walls (Cawley & Ballou 1972 ) . It was found that there was a much higher content of t otal carbohydrate in the flocculent yeast cell walls than in non-flocculent walls . The increased carbohydrate content was found to be due to an elevated level of mannan while the glucan level appeared to be very s imilar ( S tewart 1975 ; Beavan et al 1979 ) . Hence , there were disagreements on the different phosphorus levels in flocculent and non-flocculent yeast cell walls . Later inves tigation found higher total carbohydrate content in the walls of flocculent than in non-flocculent yeast . (d ) Chemical effects on the disulphide bridge Chemical modification with reagents , known to act on disulphide bridge s , carboxyl and/or phosphate groups , phenol ic groups , amino groups and imidazole groups , was found to destroy the ability of yeast cells to flocculat e . This is 39 a s t rong indication that these functional groups o f amino acid residues of the protein are essential for the floc forming ablility of brewer ' s yeas t cells (Nishihara et al 1 9 7 7 ) . (e ) The fimbria of yeast cell wall Some flocculating yeast strains possess minut e fimbria ( 0 . 5 �m) on their cell wall . These fimbria can be easily removed when treated with a-amylase and treated cells lose the ability to flocculate . It was suggested that the fimbriae may be the surface mannan-protein complex known to be involved in flocculation (Day et al 1 9 75 ) . ( f ) The cell wall ionic charge The nature and density of ionic charges on the surface of cell envelope and interference from adsorbed material from the medium were thought to have an effect on flocculation ( Jayatissa & Rose 1976) but electrophoretic mobility studies on strains of top and bot tom ferment ing yeasts , at pH 3-7 , have shown that the mobilities were independent of the flocculation characteristics of yeasts . This indicates that flocculence and surface charge of the yeas t cells are not direc tly interrelated (Beaven et al 1 9 7 9 ) . 2 . 5 . 3 . 2 Environmental effects on yeast flocculation Environmental conditions exert a subsidary influence either at the cell surface or indirect ly on the cell metabol ism . These include the pH , temperature , and medium composition (Rainbow 1966 ) . (a) Flocculation aids Calcium , magnesium and manganese ions were found to induce flocculation . Maximum flocculation of yeas t occurred a t a Ca2+ concentration of 0 . 2 mM ( 22 mg/ 1 ) . Above this concentration there was no further increase in flocculation (Mill 1 946b) . The optimum concentration for magnesium and manganese were found to be 10 mg/ 1 but the flocculation induced was o f reduced int ensity when compared with calcium . A ten t imes increase in their concentration did not increase the flocculation intensity . Low concentrations ( 1-10 mg/ 1) of e ither sodium or potassium ions were found to induce flocculation of yeast strains displaying intense flocculation with calcium ions but high concentrations of either ions ( 50-100 mg/1 ) were found to antagonise floc format ion (Mill 1964b ; S tewart & Goring 1976 ) . Certain organic substances such as furfural , certain polysaccharides ( treberin) , ethanol , and colloidal wort components ( such as humic acid , metanoidins , and phlobaphene) are known to cause flocculation ( Rainbow 1966) . However , their op timum concentrations for maximum flocculation were not reported . 40 It is worthwhile not ing here that typical whey permeate was found to contain calcium, potassium and sodium in concentrations of 1 . 17 , 1 . 45 & 0 . 6 g/kg , respectively (Anon 1981) . It is evident from the data j us t described that whey permeate contain a greater concentration of calcium than that required for maximum flocculation . The concentrations of potassium and sodium , however , are high in the range which reduces flocculation . • Hence , whey contains a flocculating agent as well as deflocculating agents . (b) Mechanism of ionic induced flocculation It was suggested that divalent ions act by bridging cells through negative charges on the cell surface , whereas monovalent ions induce flocculation via a "counter ions" effect where the repellent forces of the negative charges on the cell surface are neutralized , thus allowing some floc formation due to hydrogen bonding· or other types of non-ionic bonding between cells . The antagonism of sodium or potassium ions may be due 'to neutralization o f all available cell surface charges by the monovalent ions and thus prevent cell to cell hydrogen bonding ( S tewart & Goring 1 9 7 6 ) . ( c ) Deflocculat ing agents Certain anions which can complex with calcium such as EDTA, potass ' um sal ts of phosphate , fluoride , b icarbonate , oxalate , citrat e , diglycollate and nitrilotriacetate can affect flocculation . The lowest inhibiting concentration of each salt was found to be in proportion to its complexing power . EDTA (at 10 mM) was effective but the inhibitory effect of EDTA was reversible after washing (Taylor & Orton 1973 ; Stewart 1975 ) . A similar effect was found for Schizosaccharomyces pombe ( Calleja 1970 ) . Some unspecified proteins have been reported to prevent flocculation ( Rainbow 1966) . There was evidence that some sugars fermentable by brewers ' yeas t , such as sucrose and mal tose , could prevent flocculation but the resul ts were inconclusive (Mill 1964b ) . (d ) Growth media The composition of the growth and flocculating medium exert considerable influence on yeast flocculation . Some ale yeast strains have been reported to be able to flocculate when cul tured in a defined medium of glucose , ammonium salts , vitamins and ions ( S tewart et al 1975 ) while some required the presence of nitrogen containing inducer in the growth medium . A peptide has been identified as the inducer of flocculation in wort ( St ewart et al 1975) . It was found to contain a high level of acidic amino acid residues . Mos t flocculent lager strains examined were able to flocculate after growing in a defined medium without the peptide inducer . A strain of brewers t yeas t grown in medium containing ammonium phosphate as a nitrogen source was 41 found to flocculate more weakly than cells grown in medium containing urea or ammonium sulphate and failed to flocculate when grown in medium deficient in magnesium . The minimum concentration required was found to be 20 ]JM (Nishihara et al 1 9 76a) . Thus , the growth media effect yeast flocculence by providing nutrients necessary for flocculation . ( e ) Temperature Yeast cell floes have been reported to be s table ' betw�en 20-50°C and at higher temperature rapid deflocculation occurred . If the temperature were not high enough to kill the yeast cells , the floes reappeared upon cooling (Mill 1964a ; Calleja 1970 ) . ( f ) � Flocculent yeast cells grown in medium o f initial pH less than 3 were found to flocculate poorly (Nishihara et al 1976a) . Another study observed that flocculation was low at pH 2 but increased with increasing pH and reached a maximum between pH 4 . 5 to 5 . 5 (Mill 1964b ) . ( g) Coflocculation Some strains of brewers ' yeast which are non-flocculating , were found to flocculate readily in presence of an appropriate partner ie . another strain ( Eddy 1958 ) . This type of flocculence has been termed "coflocculation" ( Stewart 197 5 ) . (h) Agitation Agitation of the growing medium exerts a constant ac t ion on microbial aggregates (Atkinson & Daoud 19 76 ) . I t has been indicated that in the pract ical operating range o f tower fermentation wheu superfic ial l iquid velocities are greater than . 14 rnm/s . The yeas t floes appeared to be affect ed by an increase in the liquid flow rate but the yeast bed expanded with this increase which eventually lead to wash out when a high enough velocity was reached ( Greenshields & Smith 19 71 ) . 2 . 6 MIXED CULTURE AND MIXED SUBSTRATE FERMENTATION The concept of growing a mixed cul ture on a mixed substrate is important in dairy fermentation , alcoholic fermentation and biological waste treatment processes . Mutual existence of species is termed symb iosis . There are a var iety of possible interactions and several o f these can take place s imultaneously (Table 2 . 5 ) (Bungay & Bungay 1968 ) . S tudies on mixed culture have been reviewed by many workers ( Bungay & Bungay 1968 ; Bungay & Krieg 1966 ; Veldkamp & Jannasch 1972 ; Jannasch & Mateles 1974 ) . A number of kinetic models have also been proposed by many workers ( Chiu et al 1972 ; Yoon et al 1 9 7 7 ; Fredricson & Tsuchiya 19 7 7 ; Yoshida et al 1979 ) . There have been very few investigat ions on the concept of improving ethanol product ivity by utilising two yeas t species with different abilities . 42 Table 2 . 5 Common terms for microbial interactions interaction neutralism competition commensalism mutualism predation parasit ism amensalism inhibition synergism definition lack of interaction a race for nutrients and space one member benefits while the other is unaffec ted each member benefits from the other one member feeds on another one member steals from another one member adversely changes the environment for the o ther excretion of a factor harmful to the o ther combination synthesizes by cooperat ive metabolism One study on the application of binary mixtures of yeasts in continuous beer fermentation found that in certain. pairs of yeas ts , one of the s trains becomes dominant within a few days , indicating competition for substrate , although in o ther mixtures the original proportions were maintained indicating mutualism . The efficiency of a mixture was usually less than the average value for its individual components ( Rudin & Hough 1959) . Improved product ivity resul ted when the two yeasts were chosen because of their different inhibition properties a t high sugar and high ethanol concentration respectively (Jones & Greenfield 1981 ) . When there is more than one sugar source available , the sugar which is the easiest to utilize will be metabol ized f irs t . This growth pat tern has been termed "diauxic behaviour" (Monad 194 7 ) . Three terms are now used : catabolite repression describes the inhibition of specific enzyme synthesis by the preferred substrate catabolite inhibition describes the inactivation o f specific enzymes by the preferred medium ; and catabolite inactivation involves inactivat ion of already existing enzymes including their degradation (Holzer 1976 ) . D-glucose has been shown to exert catabolite inhibition on D-xylulose , D-xylose and D-xyt itol metabolism in some yeasts (Hsiao e t al 1982) . Growth of a mixed cul ture on mixed substrate in which one substrate is utilized at higher efficiency can lead to s teady state accompanied by an incomplete utilization of the o ther substrates present in the medium. However , if different growth l imiting substrates are used by different organisms , coexistence will occur ( Chain & Mateles 1968 ; Yoon & Blanch 197 7 ) . When two organisms are competing for the same growth-limit ing substrate and no other interactions between these organisms occur , their 43 behaviour can be predicted from the known relationships between substrate concentrations and growth rates . I f one organism has a greater growth rate and lower K ( subs trate s s a t urat ion ccnstant) than another , then for any substrate concentration , it will dominate and the o ther organism will be s electively excluded ( f ig . 2 . 8a) . If the saturation curves of both organisms cross ( fig . 2 . 8b ) . The dilution rate will determine the dominating organism. Fig . 2 . 8 �-S relationship of two organisms A and B . (a) KA < KB and A B A B A B s s � > � • (b ) K < K and � < � (Veldkamp & Jannasch 19 72 ) . max max ' s s max max There is one substrate concentration for which the corresponding growth rates of both organisms are equal . When this subs trate concentration is maintained · in the chemostat , the concentrations of both organisms will also be .maintained constant . Such behaviour has been observed experimentally (Veldkamp & Jannasch 1972 Chain & Mateles 1968 ) . 2 . 7 CONCLUSIONS The review has shown that K. marxianus and C.pseudot rop ical is are rapid lactose fermenting yeasts . K. marxianus is , however , more suitable for indus trial fermentation o f whey to ethanol becaus e , in contrast with C.pseudot rop ical is, it has not been reported to be pathogenic . It is also more widely used . Its optimum operating pH is between 4 . 5 to 5 . 0 , and optimum temperature o f 2 8 to 30°C . Tower fermentation i s a s imple and relatively cheap continuous fermentation process with high product ivity in comparison with the batch fermentation and o ther alternative ethanol fermentation processes . Its product ivity is affec ted greatly by the flocculating ability o f the yeas t used . This is because the yeast flocculence limits the medium and thus the medium feed rate . It is used in a number of industrial applications but has not yet been used for industrial ethanol produc tion . Thus , i t is a process that could be used to ferment whey rapidly to produce e thanol , as an alternative to the low productivity batch 44 fermentation , if a flocculent lactose fermenting yeast were available . This would enable the dairy industry to cope with the ever increasing volume of whey . Yeast flocculence is a genetically inherited behaviour which is influenced by various environmental factors . Yeasts exhibit diauxic b ehaviour when more than one sugar source is available , the sugars are util ized sequentially . Fermentation s tudies using mixed culture o f yeasts found no interact ion between the yeast s used . These topics have been included in the review because it was intended to investigate the possibility of using K.marxi anus and S . . cerevis iae together to ferment whey permeate enriched with sucrose to increase product ivity by using 100 g/1 total sugar and to improve the economic viability of the plant by removing the seasonal variability in the feedstock supply . CHAPTER 3 MATERIALS AND METHODS 3 . 1 MATERIALS 3 . 1 . 1 Chemicals Inorganic and organic chemicals (analytical reagent grade) were ob tained from BDH Chemicals (NZ) Ltd (Palmerston North , NZ) with the exception of D-glucosamine which was obtained from Sigma Chemical Co St Louis , Missouri , USA) . 3 . 1 . 2 Gases Carbon dioxide and hydrogen for the gas chromatograph were obtained from New Zealand Industrial Gas Ltd (Palmerston North , NZ) . 3 . 1 . 3 Media (a) Whey permeate Sulphuric acid whey permeate was supplied by the Dairy Research Inst i tute ( DRI) , Palmerston North , in 40 litre containers . It was s tored at -20°C unt il required , when it was thawed and autoclaved prior to use . The lac tose content of the whey permeate varied from 39 to 50 g/1 and the pH was 4 . 6 . There was one batch which had a pH of 4 . 2 but this was adj usted to 4 . 6 by addition of calcium hydroxide before autoclaving . Whey permeate was used without any additional nutrient . Table 3 . 1 gives typical composition of the whey permeate used . Table 3 . 1 Typical composition of sulphuric whey permeate (Anon 1981) . lactose TS ash 42 . 6 56 . 9 7 . 8 Ca Cl 1 . 17 . 09 K Na 1 . 45 0 . 6 P04 1 . 92 so4 1 . 51 g/kg (b ) Molasses Molasses was supplied by the Chelsea Sugar Refinery Ltd (Auckland , NZ) in cylindrical steel drums containing 292 kg o f molasses . The sucrose content varied from 500 to 600 g/kg of molasses . ( c ) Tower fermentation s tart up media During inoculum preparation and s tart up of tower fermentation of whey permeate , the growth medium used was whey permeate enriched with malt extract syrup ( 20 g / 1 ) (Maltexo , Wilson Mal t Extract Ltd , Dunedin) , Marmite ( 2 g / 1 ) / ( Sani tarium Heal th Food Co Ltd , Auckland) and . 51 g/1 of each of calcium chloride , ammonium sulphate and diarnmonium hydrogenphosphate ( (NH4 ) 2HP04 ) . 45 4 6 In the tower fermentation o f whey permeate enriched with molasses , there were further additions to the above medium . The addi tions were molasses ( 1 20 g molasses/1 of total medium volume) and urea (1 g /1 ) . The start up molasses medium for the tower fermentation o f molasses contained 100 g/1 sucrose . Urea and diammonium hydrogenphosphate were added using concentrations as indicated above . (d) Tower fermentation media In the tower fermentation of whey permeate no additive was used . In the tower fermentation o f whey permeate enriched with molasses , the medium used contained the ratio of lactose to sucrose o f 40 : 60 g / 1 . Urea ( 1 g/1 ) and diammonium hydrogenphosphate ( 0 . 5 g/1 ) were also added . Molasses (100 g/1 sucrose) was used as the fermentation medium in the tower fermentation of molasses . Urea and diammonium hydrogenphosphate were added using the same concentration as given above . ( e ) Cul ture preservation med ia YM agar ( Difco Laboratory , Detroit , Michigan , USA, supplied by Fort Richard Ltd , Wellington , NZ) was used for cul ture preservation at 4°C . Nutrient bro th (Difco) , containing 30% V/V gly�erol , was used to maintain the s tock cul ture at -20°C . ( f ) Flocculation test media The media used are listed in sect 6 . 2 for ease of cross reference during resul ts presentation because of the large numbe'r of media and variations used . ( g) Basic nutrient base For carbohydrate utilization tests . Ammonium chloride NH4Cl Dipotassium hydrogenphosphat e K2HP04 Nitrogen base (Difco) (Agar (Difco) pH 5 1 g / 1 0 . 5 2 15-30 ) If b roth medium were required only the appropriate sugar was added . If agar plates were required , agar was added . Normal agar plates contained 15 g/1 agar , replicating plates contained 30 g/1 agar . (h) Whey broth and agar This was used extensively in mutation and culture improvement experiments . Yeast extract ( Difco) Ammonium chloride Dipotassium hydrogenphosphate (Agar (95% ethanol ( . 7897 gm/ml ) was added as needed . ) pH 5 3 g / 1 1 0 . 5 15-30 ) 47 Whey permeate was used as the make up l iquid . High concentration o f agar was required because the agar would not se t at the normal concentration of 15 g/1 . Ethanol was added aseptically as needed after autoclaving . ( i ) Lactose agar Lactose was added to the nutrient base ( sect 3 . 1 . 3 g ) using concentrations of 1 . 7 and 40 g/1 for mutation and K. marxi anus plate counting , respect ivel y. Lower lactose concentration ( 1 .� g/1 ) was used during mutation in order to avoid interference from lactose . This concentration was the minimum required for growth . ( j ) Sucrose broth and agar Sucrose was added to the basic nutrient medium described in sec t . 3 . 1 . 3 (g) at concentrations of 1 . 7 g/1 when making mutation sucrose agar plates and 20 to 60 g/1 for fermentations tes ts . (k) Total cell plate count agar YM agar (Difco) was used for total cell plate counts . ( 1 ) pH adjustment of media The pH of all media was adjusted using 0 . 5 M sulphuric acid or 1 M sodium hydroxide before autoclaving . For tower fermentation of whey permeate and flocculation tes ts , calcium hydroxide was used , s ince sodium ions could interfere with the yeast cell flocculation . 3 . 1 . 4 Organisms Twelve lactose ferment ing yeasts and three Sacc haromyces cere visi ae s trains were used ( table 3 . 2 ) . They -were obtained as slant or freeze dried cul tures . Table 3 . 2 Yeast cultures used . no species strains sources forms 1 Candida pse udotr opic alis CBS 2234 BT s 2 Kl uy ver omyces l actis NCYC 416 DRI FD 3 1 1 1 1 NCYC 469 BT s 4 K. marxi anus ATCC 10022 DRI FD 5 1 1 CBS 397 BT s 6 1 1 NCYC 100 DRI FD 7 1 1 NCYC 587 BT s 8 1 1 NRRL Yll09 DRI FD 9 1 1 UCD-FST 7 1-58 BT s 10 1 1 X DRI BT s 11 1 1 y 18 (NRRL Y-610) UM D 12 1 1 y 42 UM D no 1 3 1 4 1 5 species S. cerevisiae " " 48 strains sources forms AWRI 350 ( FT146 ) AWRI S CFCC 39 UA S Y l 6 BT S AWRI BT Aus tralian Wine Research Institute , Adelaide , Australia . Department of Biotechnology , Massey University , Palmerston North , New Zealand . DRI Dairy Research Inst itute , Palmerston North , New Zealand . UA Department of Biological Sc ience , University of Aston in Birmingham, U . K . UM Department of Microbiology , University of Maryland , College Park , Maryland , U . S . A . 2074 2 . D Dried on sterilized filter paper . FD Freeze dried culture . S Agar slope cul ture . Yeas t no . l 2 is a maltose-utilizing hybrid of yeas t no . l l and S. dobzhanskii Y 1 9 76 (Wickerham and Burton 1 956 ) . 3 . 2 EQUIPMENT 3 . 2 . 1 Tower Fermenter The tower fermenter ( figure 3 . 1 and 3 . 2 ) used in this s tudy was constructed from 3 j acketed P yre.x glass pipes (Jobling Lab . Div . , S tratfordshire , England) with an enlarged sec tion situated a t the top to act as yeas t / l iquid/gas separator . The internal diameter o f the lower and separator sec tions were 25 and 1 00 mm respec t ively . The total length of the fermentation section was 2 . 37 m and overall height was 2 . 7 m giving a fermentation and to tal volume of 1 . 2 9 and 2 . 92 1 respect ively . The overall view of the tower fermenter in s itu is given in figure 3 . l (a) , while f igure 3 . l (b ) shows from left to right the medium and air inlets the separator sec t ion , and a sampling tube . The column was fitted with 4 sampling points at regular intervals along the column as given in figure 3 . 2 samples were drawn from the vertical centre l ine of the tower at each sample point . In addition samples were drawn at the inlet ( sample point 0) and exit (sample point 5 ) . A bottom inoculation port was fitted to the tower at the same level as sample port no . l . The bot tom o f the column was fitted with a rubber ( a ) ( b ) Fig . 3 . 1 (a ) To1�er fer1nenter s e t up f o r cont inuous fermentation o f whey permeate suowing aerated feed medium ( large vessel ) and e f fluent collector ( small vessel ) . (b) Separator sec t ion ( centre ) ; medium feed and air inlets ( l e f t ) · and a sampling port tube ( right ) . +-- 1.0 inocu l a t ion inl e t i no c u l a t i on i n l e t 5 0 HE (rmn) VE (ml ) - - - - - - - - - - - 2 69 3 2 9 2 1 e f f l uent ( exi t ) 2 368 1 2 8 7 top o f s t ra igh t s e c t i on 2 3 1 6 1 1 8 6 t o p o f t h i r d s e c t i on 1 5 7 3 80 5 u p p e r m i d - t owe r 8 1 7 4 1 8 l owe r mid-towe r 96 49 bo t tom o f t owe r 0 0 datum l- ig . 3 . 2 Schema t ic d iagram of the towe r f e rmen te r . In te rnal d iame t e r 2 5 . 5 mm, c ro s s s e c t i ona l area 5 . 1 2 HE - e f f e c t iv e he igh t ,mm ; V E - e f f ec t ive vo lume ,m l 2 cm , 5 1 s topper f itted with medium and air inlet pipes ( figure 3 . l (b ) ) . (a ) The separator The separator ( figure 3 . 3 ) consisted of a Pyrex glass housing ( supplier as in section 3 . 1 . 1 ) and a concentric s tainless s teel ( grade 18 : 8 ) draught tube (N . Z . S teel and Tube Ltd , Auckland ) . The Pyrex housing consisted o f a truncated inverted cone 25 mm high , a straight pipe section 400 mm high and a hemisphere o f 1 00 mm radius . The top and bottom radii of the inverted cone were 100 mm and 25 mm respec tively while the radius of the pipe sec tion was 100 mm ( figure 3 . l (a) ) . A liquid effluent port was posit ioned 250 mm from the bot tom o f the s traight section while gas exit and inoculum ports were positioned at the top of the hemisphere . The draught tube ( f igure 3 . 3 ( b ) and ( c ) ) was 38 mm in d iameter and 300 mm long . The draught tube wa� supported on three s tainless s teel pins 15 mm x 5 mm ( the bot tom end of the pins were tapered to follow the contour of the glass tube) , thus the bottom of the draught tube was 1 5 mm above the top of the fermenter straight section and the top o f the draught tube 1 5 mm above the l iquid effluent p ort . 0 Two rec tangular baffles , 90 apart were positioned on the out side of the draught tube . The baffles were 1 50 mm in height and o f sufficient width to reach the internal wall of the Pyrex glass housing . In between the baffles a 30 mm sec t ion of the draught tube was cut to a depth of 40 mm forming a weir for the liquid effluent . The weir was positioned d iametrically opposite to the liquid eff luent p ort . The flow pattern in the separator was therefore as follows : The liquid ef fluent from the fermenter flowed up the draught tube and over the weir into the zone between the baffles . At this point it was forced to flow down under the baffles and up into a quiescent zone where its super ficial veloc ity decreased . The cross-sec tional 2 area o f the quiescence zone was 50 cm . This decrease in superficial liquid velocity allowed the yeas t flo es to settle under gravity and the c lear l iquid ef fluent flowed out of the tower via the l iquid effluent port . The effective working volume o f the separator W3S 1 . 6 1 . (b ) Temperature Control The temperature of the fermenter was controlled by c irculating water 30°C from a ho t water bath ( Compenstat water c irculator , Gallenkamp , London , England , supplied by Smith-Biolab T 4 0 0 j 5 2 ( a ) separator gas exit ( b ) cro s s-sect ion o f s eparator l O O � + inocula t i on port ef f luen t exit-1 4 0 1 -r l 5 0 I 2 5 0 13 � 0 _j I 5 0 1 -+-- l 5 ( c ) d e tai ls o f draugh t tub e All d imens ions are in mm . � = internal d iame t er Fig . 3 . 3 S ch emat ic d iagram o f th e separator and the draught tub e . 5 3 Auckland , New Zealand ) through the 3 water j acke t s ( f igure 3 . 4 ) . The j acketed s e c t ions were connec ted in series , wi th the water enter ing the bo t t om of the l owe s t sect ion and returning to the water bath via the top exi t o f the uppermo s t j acketed sect ions . The t emperatures a t the four sample points on the tower were moni tored us ing copper-cons tantan thermocouple s connec t ed to a 1 2 input - 3 channel - chart recorder (Versaprint , Honeywel l , USA) . ( c ) Tower and medium aerat ion The tower was aerated f rom the bo t tom via a s tainle s s s teel cap il lary tube to keep air bubbl e s i z e very sma l l ( f i gure 3 . 2 (a ) and 3 . 4 ) a t a rat e o f 1 2 0 ml /min . , 5 0 kPa during the c e l l build up period . The med ium was aera ted for 1 h before b eing pumped into the tower . ( d ) Air supply and f il ter The air supply was p rovided b y the Univers i ty central service . The supply. to the tower was f i t ted wi th a vapour - o i l f il ter and pres sure regulator ( Norgren , type F04 , supplied by Kidd Garre t t , Auckland ) . The air supply for med ium reservoir and tower aerat ion was s teri l i zed by pas sage through an air f il ter cons t ruc ted o f s ta inless s teel tube 2 5 mm ID x 350 mm packed with f ibreglas s . A s imilar f i l ter was f i t t ed to the gas exi t l ine o f the towe r . The air f l ow r a te was moni tored us ing a 0- 1 000 ml /min variable area f lowmeter ( GAP , Basings toke , UK supplied by Homersham L td . , Chris tchurch , New Zealand ) . ( e ) Med ium pump The med ium was fed cont inuously in to the bot tom inle t of the tower ( f igure 3 . 2 (a ) and 3 . 4 ) us ing a Mas ter f lex Peris tal t ic pump , model no . HA IR 05 1 ( Co l e-Parme r , Chicago , USA s u p p l i e d by Smi th­ Bio lab , Auckland , New Zea land ) f i t t e d with pump head no . 7 0 1 3 for f eed rates up to 350 ml /hr and no . 7 0 1 4 for feed rates u p to 1 000 ml /hr . S i l ic o n e t ub ing was used throughout . Cal ibrat ion curves f or del ivery rate to the tower at var ious mo tor speeds are given in Append ix E 3 . 2 . 2 1 0 1 Batch f ermenter The fermenter vessel was a 15 l New Brunswick S c ien t i f ic Co Ltd , Mode l CMF 1 4 , glass vessel (Watson Victor L td , Wel l ington , New Zealand ) . The f e rmenter support and contro l unit was cons truc ted by the Depar tment o f B i o technology workshop , Massey Univers i ty . The uni t has facil i t ies for aerat ion , c ontrol o f t emp era ture , agitat ion speed , and foaming . The fermenter vessel was 1 5 1 total volume with a working volume o f 1 0 1 . a! c ·-< ,.. c.J 0 .... ._; •rl :1) ...... ,.. :lJ u :1) c CJ 6 ::l ::l ...... . .... "-' -o CH ili :]) 6 AF e f f l ue n t wa t er t e mp r e c o r d e r f e e d AF 5 4 TC TC 3 0 ° C w a t P r i n l e t � c ·rl ...... ,.. 0 u C1l ,.. ili (1j ,.. a! :3 0 Fig . 3 . 4 Wat er heat ing and a ir f i l t ra t ion s y s t ems for the tower fermen ter . AF - air fil t er ; TC - thermocoup l e inl et Fig . 3 . 5 Batch fermenter ( 1 0 l itre working volume ) . Fig . 3 . 6 Replication equipment . Wooden supporting base and replicat ing cloth . \Jl \Jl 5 6 The medium t emp era ture was maintained a t 3 0°C by recirculat ion o f hot wa ter . The agi ta t ion speed was 2 5 0 rpm ( f igure 3 . 5 ) . 3 . 2 . 3 UV lamp The UV lamp used for mut a t ion exper iment s was a CAMAG universal - UV - l amp ( 2 9 2 00 ) oper a t ing a t a wavelength o f 254 nm , a t 300 mm above the yeas t cell suspension . 3 . 2 . 4 Replica pla t ing Repl ica p la t ing was carried out us ing a s impl e round wooden base having a d iameter s l ightly sma l l er than presteri l ized p l as t ic agar p l a t e . The t ransfer c lo th ( f igure 3 . 6 ) was held in p lace b y an adj ustab l e Jub i l e e c l ip s teel band which can be removed easily from the base . The transfer cloths were s tacked in groups .o f 50 with paper tmvel s a s separa tors and autoc l aved a t 1 00 kPa , 1 2 0°C , 1 5 min be fore being dried overnight at 7 0° C . 3 . 2 . 5 Glas sware S t andard l aboratory glassware obtained from commerc ial sources wa s used throughout this s tudy . All glassware wa s rout inely washed in ho t \vater containing "pyroneg" (Diversey-Wa l lace Ltd , Auckland , New Z ea land ) , rinsed wi th d i s t i l l ed water and air dr ied a t ea . 5 0°C . 3 . 3 STERILIZATION 3 . 3 . 1 Med ium and glas sware s terilization Al l media were s t eri l ized at l O O kPa for 1 5 min . The media in large volume containers of 4 to 20 1 were s teril i zed under the s ame c ond i t ions for 2 5 min . Glas sware was s t er il i zed a t 1 60°C for 2 h . 3 . 3 . 2 Towe r fermenter The tower was thoroughly c leaned with water containing Pyroneg and s terili zed by free s teaming for a total of 24 h . The s teaming was c arried out for 1 2 h , s topped for 1 2 h , and then s t eaming was continued for 1 2 h . After the second s t eaming the tower f i t t ings were immediately ins ta l l ed . All tower f i t t ings were au toclaved a t l OO kPa for 1 5 min be fore ins tallation . During the commiss ioning run using mola s ses med ium , purity t e s t s us ing nutrient agar were carried out to t e s t the e f f ec t iveness o f 5 7 the s t e r i l iza tion . No contaminat ion was detec ted . 3 . 4 ANALYTICAL METHODS 3 . 4 . 1 Lac tose Lac to s e was analyzed u s ing a Y S I industrial sugar analyzer , model 2 7 (Yellow Spring Ins . Co , Ohio , USA) . The sugar analyzer was fit ted with a lactose membrane c ontaining galac to s e oxidase , immobil ized by glutaraldehyde and al lowed d irect measurement o f lac tose concentrat ion . 3 . 4 . 2 Sucrose The same sugar analyzer was used as in s e c t ion 3 . 4 . 1 but the analyzer was f i t ted with a gluco se membrane which c ontained the enzyme glucos e oxidase immobil ized by glutaral dehyde Suc rose was f ir s t hydrolysed to gluc o s e and fruc tose u sing yea s t inver�ase concentrate ( BDH) which was added at a rate of 250 �1 / g sucro s e in the samples . The sampl e solution was then incuba ted at 30°C for 1 h before inj e c t ion into the analyze r . The glucose reading ob tained was mul tiplied by 1 . 9 ( the ratio o f molecular weigh t o f sucrose t o gluco se ) t o give the suc rose concentration . 3 . 4 . 3 E thanol E thanol concentrat ions in the samples wer e determined using a Varian Aerograph gas chromatograph , model 600 D ( California , USA) . The c o lumn was packed with Porapak Q (Appl ied S c i ence Lab Inc . , S tate Co llege , Pennsylvania , USA) and operated at l 7 0°C . I sopropanol solution was add ed t o t he samples as internal standard . The external s tandard s used c ontained e qual concentration o f e thanol and i sopropano l and covered the concentrat ion range from 1 0 to 50 g / 1 in steps o f 1 0 g / 1 . An add i t ional s tandard o f 5 g / 1 were used for low e thano l concentrat ion . The inj e c t ion volume used was 1 w l in all cases and al l e thanol concentrat ions were expressed in g / 1 . 3 . 4 . 4 Cell conc entra t ion ( a ) Cell d ried weight and we t weight ( DW and WW) Samples o f 1 0 ml volume were c entrifuged and the supernatant f luid was decanted . The we t yeast c e l l pellets were weighed to give the centrifuged c e l l we t we ight (m�) ( g /1 ) . The cells were then d r ied at l 0 5°C for 2 4 h to determine the c e l l dried wei ght ( DW ) ( g / 1 ) . 5 8 ( b ) Plate count The to tal l ive cell c ount was obtained f rom YM agar incubated aerobically at 2 5°C while lactose agar ( s ec t ion 3 . 1 . 3 i ) was u sed f o r the K. marxianus coun t . Pep tone wa ter ( 5 g / 1 peptone) ( Difco ) was used as d i lut ion medium . For f locculent yeast s , the d ilution medium contained NaC1 ( 9 g / 1 ) and EDTA ( 4 g / 1 ) . NaCl provided an osmo t ic p ressure balance and EDTA induced de f locculat ion . During d ilut ion , each dilution was vigorously agitated b e fore the next dilut ion s tep was made . S erial d i lution was cont inued to give approximately 30-300 colonies when 1 ml al iquo t s were p lated . Dup l ica te plate s were made f o r each d ilut ion . The agar plates were incubated at 2 5°C for 3 days be f o re counting . ( c ) Haemacytometer cell count The to tal cell number was ob tained by us ing a haemacytometer (As s i s tent , Germany) . Samples to be counted tha t c ame f rom very c loudy media were �sual l y diluted 5- 1 0 t imes in o rder that cell coun t could be made . 3 . 4 . 5 Measurement o f yeas t f lo cculence Two me thods wer e used for the measurement of yeast f locculence . The f ir s t method was adapted f rom Helm e t al ( 1 9 5 3 ) and S t ewar t ( 1 9 7 5 ) . Thi s method used visual observat ion to give a f locculat ion scale . The second method was an adap tat ion of that desc ribed by Greenshields et al ( 1 9 72 ) using yea s t s e t t l ing volumes in rela t ion to t ime to g ive a numer ical floccul a t ion value . (a ) Flocculat ion scale method 1 . Yeast cel l s were p repared as desc ribed in sec t ion 3 . 9 . One gram o f centri fuged ce l l s was resuspended in 9 ml o f the t e s t ing medium and transferred t o a 1 5 ml capac i ty gradua ted tapered centrifuged tub e , giving a total vo lume of 1 0 ml . The tub e was al lowed t o reach thermal equilibrium with amb ient , then gently shaken f or 5 minutes to ensure all f loes were evenly b roken up . 2 . The yeas t s e t t led vo lume was recorded at 1 , 2 , 5 , 1 0 , 1 5 and 60 minute s . 3 . The f locculent scale was then de termined u s ing the fol lowing sub j ec t ive scale ( S t ewar t 1 9 7 5 ) . 4 . Each f locculat ion tes t wa s carried ou t in dupl i cates f rom s tep l to 3 . 59 5 extreme ly f loccul ent EF 4 very f locculent VF j moderately floccul en t MF 2 weakly f l o c culent WF 1 rough floc cul ent R 0 non- flocculent NF This me thod i s very sub j e c t ive . The p rocedure used here d i f f ered from that of Helm et al ( 1 9 5 3 ) in t hat the subj ec t ive scale was used ins tead of us ing the yeast s e t tl ed volume a f ter 1 0 min . The me thod also d i f f ered f rom that of S tewart ( 1 9 7 5 ) in that 1 gm we t yea s t was used in a t o tal t e s t med ium volume o f 10 ml . The c leaning and t e s t media used were a l so d i f ferent . The f locculation s cale used was the same . ( b ) Sharp ' s modi fied Burn ' s numb'er ( MBN) Yea s t flocculat ion was a l so measured by the mod i f ied Burns number ( Greenshields e t a l 1 9 7 2 ) . The yeas t volumes recorded ( a s in sec t ion 3 . 4 . 5 ( a) ) a f te r 0 , 1 , 5 , 1 0 and 1 5 min . were u s ed to calculate MBN o r MBN* for the yea s t t e s ted depend ing on the growth and the test medium used . I f the yeas t s were grown , prepared and t e s ted following the s tandard method described by Greenshields et al ( 1 9 7 2 ) , the value obtained was des ignated as "MBN" . However if the yeasts were no t gro\vn and tes t ed in the s t andard media , the value ob tained was des ig­ nated as "MEN'"" . 3 . 4 . 6 El! The pH o f the medium was measured u s ing a Triac DPH- l pH meter f i t ted with a comb inat ion e lec trode ( E . L . Kay Ltd , Auckland , New Zealand ) . 3 . 4 . 7 Spec ific gravitv Medium spe c i f ic gravity was measured u s ing a Zeal hydrome ter ( G . H . Zeal Ltd , London , England , suppl ied by Smi th-Bio lab Ltd , 0 Auckland , New Zealand) at 2 0 C . 3 . 5 CULTURE PRESERVATION AND MAINTENANCE The yeast cul tures were maintained on YM agar slopes a t 4° C . Two s lopes o f the culture were maintained . One slope was used for experiments whil e the o ther s lope was kep t as s t o ck cul ture . Cul tures were subcul tured onto new s lopes at s ix month interval s . Ano ther s e t 60 o f s tock cul tures were maintained a t -20°C in nutrient bro th containing 30% v/v glycerol . 3 . 6 INOCULUM PREPARATION Inocula for experiment s were prepared us ing the fol lowing s teps . The number o f s teps required increased with the volume o f ino culum required . The incubat ion per iod for each s t ep was 24 h and the volume trans ferred was 5 % v/v o f the f inal t o tal volume o f the next s tep . 1 . S tock cul ture + loop 2 . 1 0 ml mtd ium in 25 ml bo t t l e . 5 ml 3 . 9 5 m l mtd ium in 2 5 0 ml shake flask ( 1 00 ml ) 5 0 ml 4 . 4 5 0 m l �edium in 2 1 shake flask ( 500 ml ) 5 00 ml 5 . 9 . 5 1 mtd ium in 1 5 1 batch fermenter ( 1 0 1 ) ( concentrated before inocula t ion ) 6 . 2 x 4 5 0 ml medium in 2 1 shake f lasks ( 1 1 ) ( us ed inoculum prepared in s tep 3 ) 7 . 1 2 x 4 5 0 ml med ium in 2 1 shake flasks ( 6 1 ) ( used inoculum prepared i n s t ep 6 ) ( concen trated before inoculat ion) 8 . 30 1 medium in 4 5 1 batch f ermenter ( 3 0 1 ) ( u sed inoculum prepared i n s tep 6 ) ( concen t rated b efore inocula t ion) 3 . 7 FER}lliNTATION CONDITIONS All fermenta t ions were carried out a t 3 0°C . All shake flask fermen tat ions were agi tated a t 1 50 rpm . Fo r normal fermentat ion the inoculum volume was 5% v/v of the f inal t o tal med ium volume . For s ome mutat ion and cul t ure improvement s hake f l ask fermenta t ions an inoculum vo lume o f 1 0 % v/v of the f inal t o tal med ium volume was u sed . Ten l i tre batch fermen tat ion was ag i ta ted at 2 5 0 rpm and the inoculum wa s prepared as in s t eps 1 to 4 in sect ion 3 . 6 . Agar plates were incuba ted a t 2 5°C f o r 3 days . The lower incu­ bation tempera ture was used to reduce the drying o f t h e agar . 3 . 8 TOWER FERMENTATION 3 . 8 . 1 S tart up 6 1 ( a ) Ini t ial s tart up 1 . The inoculum was grown on the s tart up medium (whey permeate with Mal texo ) us ing s teps 1 to 5 , 6 & 8 in sec t ion 3 . 6 . 2 . After inoculat ion into the tower , the floes were allowed to settle and the spent liquid was decanted off at sample po int 2 ( . 8 2 m) . The tower was then filled with fresh med ium , and an aerated batch fermentat ion took place for 4 h af ter which the yeas ts were allowed to set tle for a period of 1 h . (Aerobic fermentat ion o f whey was reported to require 3 to 4 h for complete utilization of 4 0 g/1 lac tose on whey us ing ' 9 an ini t ial cell concentrat ion of 2 x 10 cells /ml . ) (Wasserman et al 1 9 5 8 ) . 3 . The procedure o f decanting o f supernatant l iquid , ref i l l ing the tower , and batch fermentat ion was repeated 3 t imes . In this way large cell concentrations were obtained in the tower . 4 . After the last batching , the supernatant l iquid was decanted . A cont inuous slow feed of 80 ml /h was commenced with aerat ion a t 1 2 0 ml /min for 10 days in order to increase the cell concentrat ion fur ther before sampl ing could begin . ( b ) Subsequent s tart up After one month of operation the experiment was terminated because of whey shortage . A s l ightly dif ferent inoculation procedure was employed in the subsequent start up for ease of operat ion and to reduce contaminat ion risk . The inoculum was prepared using s teps 1 to 4 , 6 & 7 in sec tion 3 . 6 . The inoculum was asep t ically concentrated and inoculated into the tower . Then the s tart up s teps 2 to 4 described in sec t ion 3 . 8 . 1 ( a ) were followed . The tower was considered ready for cont inuous operat ion when the set tled yeast volume occupied 2 / 3 of the tower s traight sec t ion . In the tower fermentat ion o f whey permeate enriched with molasses , the inocula o f the two yeast cul tures were prepared separately using similar s teps but the growth medium for S. cerevisiae CFCC 39 was molasses medium ( 100 g/1 sucrose) while K. marxianus Y 4 2 was grown in the whey permeate with Mal texo s tart up medium . They were then inoculated in to the tower together . The procedure described in sec t ion 3 . 8 . 1 (a) from steps 2 to 4 was followed using whey permeate enriched with molasses and Mal texo as the s tart up medium . Tower fermentat ion o f molasses was s tarted us ing the " subsequent s tart up" procedure and molasses solut ion ( 100 g/1 sucrose) as the s tar t up medium . 6 2 3 . 8 . 2 Sampl ing procedure Daily sampling of pH and cell cen t r ifuged wet weight for evaluat ion of the onset o f s teady s ta t e was perfo rmed a t sampling point 2 ( . 82 m) and the exi t ( sampl ing poin t 5 ) together with the speci f ic gravi ty and f low rate of the e f f luen t . Once s teady s ta t e was estab l ished as seen by two sub sequent analyses agreement to wi th in 5 % , complete s e t s o f samples were collec t ed f rom all samp l ing poin t s inc lud ing the exi t ( f i gure 3 . 1 ) . Approximately 1 5 ml samples were taken a t s ample po int s 0 to 4 . The e f f l uent was col l e c ted prior to samp l in g for approximately 1 h durat ion . A 10 ml samp l e f rom each sampl ing po in t \vas c entr i fuged immediately a f t e r c o l lec t ion . The superna tan t l i quids \vere ana lyzed for s ugar & e thanol . The centri fuged c e l l we t \vei gh t (vl\,.7) and the c el l d r ied we i ght (DW) were de termined . The remaining samp l e s were 0 S P rt f 0 r oH measurement . I t was pos s ible to monitor the SG o f the e f f luent onJ y , as the t ower fer111enter volume was too small to p ermi t the removal of l ar ge vo lumes at each of the sample points . An example of the data shee t i s given in Append ix E . Fur ther s e t s o f samples were c o l lec t ed a f ter minimum periods o f 2 res idence t imes . There were generally f ive sets o f samples a t each feed rat e . In the towe r fermentat ion o f whey permeate enriched wi th molas ses , a s imilar p rocedure as above was f o l lowed wi th some add i t ions . Af ter samp les were ob tained f rom various sampl ing po in t s , 1 ml o f each sample was asep t ical l y trans f erred to s t er i l i z ed 9 ml EDTA-saline solut ion f or u se in the de te rminat ion o f the c e l l numb ers b y p la te count sec t ion 3 . 4 . 4 b ) . The remainder o f each sample was proces sed us ing s imilar p ro c edure as for pure cul ture fermenta t ion . The parameters measured were pH ; glucose , suc ro s e , lac tose and ethanol c oncentrat ions ; t o tal c e l l count and K. marxianus cel l count . In the tower fermentation o f mol a s s e s s imilar sampl ing procedure as f or the tower f e rmenta t i on of whey permea t e was used . For the las t two t ower fermenta t ions only two complete s e t s o f samples were t aken for each feed ra t e . However , daily sampl ing a t sample p o in t 2 and the exi t was con t inued t o ensure s teady s ta t e . Thi s was neces sary b ecause o f the d i f f icul t i es i n car rying out ext ens ive samp l ing of mixed culture and mixed sub s trate f ermentat ions ( see chap ter 5 ) , and the l imi ted amount o f molasses ava ilab le . 6 3 3 . 8 . 3 Cont inuous operat ion When the s tart up operat ion was comple ted , the air was turned o f f and the medium was fed t o the tower a t the s lowest f eed rate . Af t er 2-3 days a t this f eed ra te , daily s ampl ing began , once s teady s ta te wa s es tabl ished . Once the required se t s o f samples were collected , the medium feed ra te was increased to the next feed rate and the s ame procedure was repeated . This was c arried out unt il t he wash out f low ra te wa s reached . 3 . 9 FLOCCULATION TESTS Flocculat ion tes t s involved two maj or s teps : l . Yea s t cell preparat ion was carried out by growing t he yea s t to be tes t ed in a medium of interest and observing floccul a t ing behaviour during f ermentat ion . 2 . Measurement o f yeas t flocculence to determine f locculent scale . The yea s t cells were grown a s follows : l . The yea s t t o be s tudied was grown in the medium o f interes t , with or without added nut rients . The ini t ial pH , f inal pH and formu­ la t ion of the medium were recorded . Ob servat ions during fermen­ ta t ion were also�recorded . 2 . The yea s t cells were harves ted a f ter 4 8 h ( to tal co l lected l g centrifuged wet we igh t ) f o r use in f locculat ion measurement s . I f the yea s t were t e s ted for f locculence in i ts growth medium , no cleaning was necessary . I f the yeas t were tes ted in a d i f ferent med ium , that med ium was used to c lean the yeas t twice . Mos t yeas t types were tes t ed for flocculence in the f loccul a t ion t e s t ing medium ( sec t ion 6 . 2 . 8 med ium 3 6 . FM) . The c leaning procedure involved twice washing in calc ium sulphate ( . 5 1 g/ 1 ) and centri­ fuging a t 2 500 rpm for 5 min . 3 . The cl eaned yeas t c e l l s were then tes t ed in an appropriate medium to determine the f loccula t ion scale or mod i f i ed Burns number as described in sec t ion 3 . 4 . 5 . 64 3 . 1 0 CULTURE IMPROVEMENT 3 . 1 0 . 1 Isolat ion o f ethanol to lerant K. marxianus using ethanol gradient -�gar IJ KMl O gradient a gar E 30 --t> 1-::-----;;:J --t> E40 E30 840 ?\45 --{> --l> --{> jjso--<> [is KM lOA {i n !140 845 E5� [5; 1 ':/1 � � KHlOB KMlOC (� --t> 0 . 1 ml inoculum for p lat ing . Number a f t er E ind icates added ethanol concen tration in g/ 1 Broth cul ture , 4 8 h , inoculum was 10% v/v o f the whey permeate broth volume ( 50 ml) in 250 ml flasks . E thanol was added at the beginning of the fermen tat ion . Fi g . 3 . 7 Subculturing s teps used in the isolat ion of ethanol tolerating K . marxianus . The subcul turing sequence used in this experiment is summari sed in figure 3 . 7 . In pouring the grad ient alcohol c oncent rat ion agar , normal whey agar was Doured f irst on a s lope . When the agar s o l id i f ied , the agar plates were placed on a l evel surface and whey agar with added e thanol was poured over the s loping agar . l . A 4 8 h cul ture o f K. marxianus UCD FST 7 1 5 8 , grown in whey permeate ( 1 00 g/ 1 lactose ) was us ed to inoculate 30 g/ 1 e thanol grad ient agar plates us ing . 1 ml inoculum and spread on the agar sur fac e . 65 2 . After incubation , single colonies growing at the end o f the agar plate with the highest ethanol concentration were used to inoculate a 30 g/1 ethanol whey broth ( 1 00 ml ) and fermented for 48 hours . 3 . Af ter 4 8 h , this culture was used to inoculate 4 0 g /1 e thanol gradient agar and whey broth . These were incubated as before . 4 . The s teps that followed are given in figure 3 . 7 . 5 . Once the sequence described above was complete, the cul tures in 55 and 60 g /1 ethanol whey broth were compared on their ability to ferment 100 g /1 lactose whey permeate . These cul tures were called KM l OA , KMl OB , KMlOC and KMl OD . Each cul ture was used to inoculate a separate flask containing 100 ml whey permeate ( 1 00 g / 1 lactose) and incubated . Samples o f 5 ml were collected at 0 , 24 , and 48 h measured . and pH , cell number , lactose and ethanol were 6 . The cul ture found to be the fastest lac tose fermenter was put through a further series o f subculturing in whey broth containing 35 and 55 g /1 ethanol to ensure that the ethanol tolerance was s table . The cul ture was subcul tured using a series of 1 00 ml whey broths with 35 and 5 5 g /1 ethanol added . The inoculum size used was 10% v/v in order to compensate for the cel l loss due to ethanol . 3 . 1 0 . 2 An at tempt to isolate sucrose negative K. marxianus strains The cul ture was irradiated with UV light to induce mutation . The procedure involved the es tablishment o f a yeas t-kill-curve by UV irradiation in order to determine the optimum irradiat ion time , fol lowed by a replica plating experiment to isolate the desired yeas t . ( a ) Determina tion of optimum irradiation time 1 . A 24 hour cul ture of alcohol tolerant K. ma�xianus (KM10D10 ) ( sect ion 3 . 1 0 . 1 ) was prepared in 1 0 ml whey permeate bro th in a 250 ml shake flask . 2 . The yeast suspension was washed by centrifuging and resuspending the cells in 1 0 ml of 9 g/1 saline solution . This s tep was repeated once more . 3 . The suspension was shaken and placed in a steril izedr petri ' dlsh and irradiated with 254 nm UV light , at a dis tance of 300 mm above the d ish for the following t imes : 1 , 1 . 5 , 2 , 3 , 4 , 5 , 7 , and 1 0 min . 4 . For each irradiation t ime , the number o f viable yeast cells was estimated by taking 0 . 1 ml o f suspension and preparing a d ilution '5 ' 66 series us ing pep tone water and pla ting appropriate dilu tions on YM agar . Two plates were poured for each concentrat ion . The plating d i lut ions used are given in table 3 . 3 . It should b e no ted here that s teps ( 3 ) and ( 4 ) should b e carried out i n darknes s t o prevent r epair by v i s ible l ight repair mechan isms . Tabl e 3 . 3 UV t ime ( s ) D ilut ions ( tens ) The viab l e The opt imum Plating d ilut ions used to determine op timum UV irradiation t ime 0 l 1 . 5 2 3 4 5 7 -8 -7 -7 -7 -7 -6 -6 -6 -9 -8 -8 -8 -8 - 7 - 7 -7 cell count ob ta ined was plot ted vs irradiation time was that which gave 9 9 . 9 9% kill ( f igure 1 0 - 5 - 6 t imes . 3 . 8 ) . l n X 5 IRRADIATION TIME, min 10 t 9 . 2 Fig . 3 . 8 Survival o f cel l s irrad iated with UV l ight . !OUODlO . ( 24 6 nm , d i s tance 300 mm , in 9 g / 1 sal ine solut ion ) (b ) Replica plat ing and isolat ion l . S teps l to 3 o f sect ion 3 . 1 0 . 3 ( a ) were followed excep t that the irrad iation t ime used was 9 min 1 5 s . 2 . The irrad iated c e l l s were c e n t r i f u ge d and r e s u s p e n d e d in 10 ml 0 whey broth and a l l owed to expres s for 6 hours at 30 C , 1 5 0 rpm . 3 . Then the number o f viable ce l l s were determined b y pla t ing onto whey agar and incub a ted at 2 5°C for 3 days in order to de termine the d il u t ion which gave a count of approximately l OO colonie s / plate . 6 7 4 . The d ilut ion which c ontained 1 03 c e l l s /ml ( i . e . l OO cell s / p la te ) was used t o p la te 300 pla te s o f whey permeate agar which were incuba ted a t 2 5°C for 3 days . 5 . The plates were used for replica t ion onto sucrose agar p lates ( see f i gure 3 . 5 ) . After incubation , the two plates , o riginal and replicated plate s , were compared and mis s ing colonies in sucrose agar pla t e s were noted . 6 . The colonies which were found to be miss ing on sucrose agar p lates were s treaked f rom the original whey permeate agar plates onto a pair o f sucrose and whey permeate agar plates to check for growth . Those colonies which showed growth in b o th plates were rej e c t ed , tho se only on whey were retained . 7 . Those pos s ible sucro se negative i solates were s treaked and p l a t ed onto a series o f pairs o f sucrQse and whey agar plates to check for s tabil ity , i . e . no growth in sucro se agar during subcul turing before a fermentat ion check was c arried out ini t ially in whey permeate bro th to build up populat ion and then in a mixed sub s trate medium of lactose and sucrose ( 40 : 4 0 and 20 : 2 0 g / 1 ) . 8 . For the f erment a t ion tes r , a number o f s ingle colonies were used to inoculate whey permeate ( 1 00 ml in 250 ml shake f lask) , one colony per f lask , and fermenta t ion was c arr ied out for 4 8 hours 0 at 30 C , 1 5 0 rpm . The cell populat ion , lac tose , pH , and e thanol were measured . Thes e cul tures were then used to inoculate lactose and sucrose media , correspondingly numbered and fermen­ tat ion was carried out . 3 . 1 1 CALCULATION METHODS 1 . 1 1 . 1 Tower ferment a t ion Vo lume t r i c f lowra te ( Q ) (ml /h ) Q = e f f luent vo lume (ml ) collec t ed over t ime t (h ) . Re s idence t ime ( T . ) ( h ) a t a sample point r1 T . = tower vo lume at this sample point (ml ) 7 Q (ml /h ) . r1 Res idence t ime ( T ) ( h ) for the to tal tower s t raight sec t ion was r c a lculated us ing the s t raight s e c t ion volume o f 1 2 8 7 ml . T 1 2 8 7 7 Q (ml / h ) y Dilut ion rate ( D ) (h-l ) D = l � T ( h ) r 68 Mean res idence t ime (T*) (h ) for a particular t ower s ect ion r T* T . + { ( T ' +l - T . ) � 2 } r r1 rl r1 where T . and T ' +l are res idence t imes at the bot tom and the t op of th e r 1 r 1 s ec t ion respectively . S uperfi c ial l iquid velocity (V ) (mm/s ) s V s Q � ( tower cross sect ional area) 2 { Q ( ml /h ) x 10} � { ( 5 . 119 5 ( cm ) x 3 600 ( s /h) } Q X 5 . 42 5 7 X 10-S The mean volume tric rate o f s ubs t ra t e utilization ( S ' ) w i th in a part icular tower s ec t ion was calculated f rom the amount o f s ubs trate u t i lized within th e s e c t ion during the t ime period th at med ium was in the section . g / lh where S i and S i+l are s ubs t rat e concent ra tions at the b ot tom and the top sample point s o f the sect ion respect ively ( g/1 ) . T . and T ' +l are the residence t imes at the b ot tom and the t o p r 1 rl s ampl e poin ts o f the s e ction respectively (h ) . The mean speci f i c rate o f s ub s t rate utili zat ion ( q) with in a part icular t ower s e ct ion g / lh where X i and Xi+l are cell c o n c en t r a t ions at the b o t tom and the t op samplin g point s o f the sect ions respectively . The cell concen t rat ion in the bot tom s e c t ion o f the tower ( up to 0 . 09 6 m) was as sumed to be the s ame throughout . Th is ass umpt ion was made because there was no s ampling at zero t ower he igh t for the cell concentrat ion . Th e mean volumetric rate o f e thanol produc tion ( E ' ) w i th in a par t i cular tower sect ion was ca lculated f rom the amount o f e thanol produced w i th in the sect ion during the t ime period that the medium was in the s e c t ion . E ' = ( E - E . ) · ( T T ). i+l 1 7 ri+l - r i g / lh where E i and E i+l are ethanol concentrations at the bo t t om and the top sample point s o f the sec t ion respectively ( g/ 1) . The mean speci f i c rat e o f e thanol product ion ( v) with in a part i cular tower sect ion was calculat ed f rom the mean vol ume t ric rat e o f e thanol produc tion ( E ' ) within the s e ction d ivided by the mean cel l concen trat ion in the sect ion . 69 g/ gh where X i and Xi+l are cel l concentrat ions at the bo ttom and the top sample points of the section respectively ( g/1) . E thanol productivity at a partdcular t ower he ight is = E -;- T . i rl g/ lh whe re E . is ethanol concen tration at that h e igh t ( g/1 ) . l The y ield o f e thanol (Y ) on s ub s t rate ut il i zed based on theoret ical y ie ld a t a par t i cular tower heigh t , y = (E . X lOO ) • [ ( S - S . ) X 0 . 5 3 8 ] l 0 l where S is 0 the feed sub s trate concentration ( g / 1 ) % and S . is the subs trat e concentration at this height ( g/ 1) . l The mean speci f i c growth rate ( � ) . for the tower excluding the separator , w = ( 1 /X) X ( dX/ dt) X • ( X T ) e a r g/ gh where X is the cell concent rat ion l eaving the tower and e X is the mean cell concentration f or the tower f rom the b o t t om a of th e tower up to 2 . 32 m heigh t . T is the res idence t ime a t th is h eigh t . r Growth inside the s eparator was negligib le . 3 . 1 1 . 2 B at ch fermen tation The sugar utilization rate and e thanol produc tion rate used were averaged over the f ermentation time period . v1here s I t ( S . 1_ 8 i+l) E ' ( E i+l E . ) t l qt s ' . X t at \) E ' • X t t at y [ ( E i+l - E . ) l t t i + [ ( t i+l x at ( Xi + x i+l ) . X ( t i+l ( t i+l 100 } . t . ) • 2 ] l 2 t . ) l g/ lh t . ) g/ lh l g/ gh g/ gh [ ( s . l - s i+l ) X 0 . 5 3 8 } % h g / lDW s is the sub s t rate concen tration ( g / 1) , S ' is the mean volumetric rate o f s ub s trate u t i li zat ion ( g/ lh) , E ' is the mean volumetric rate o f e thanol product ion ( g/ lh ) , q is the mean speci f ic r a t e of s ub s t rate ut i l i z a t i on ( g/ gh ) , 7 0 V is the mean s peci f ic rat e o f e th anol p roduction ( g/ gh ) ' y is the e th anol y ield bas ed on theore t i cal yield ( % ) ' t is the f ermentat ion t ime (h) ' X is the mean ce l l con cent rat ion ( g / 1 DW) . a Percentage s ub s trate ut il iza tion , S ( feed sugar concentration - res idual sugar concen trat ion) u x 100 � feed s ugar concentration . CHAPTER 4 TOWER FERMENTATION OF vlliEY PERMEATE In this s tudy , K. marxianus Y42 was used in the tower f ermenter t o ferment whey permea t e continuously · The cont inuous fermenta t ion operation was carried out a t d i f ferent s uperficial liquid velocities (V ) un ti J s cell wash out was observed . Such fermentat ion parameters as lactose , e thanol and c ell concentrations , and the medium pH were monitored a t d i f ferent locations o n the tower including the exit . These and o ther parameters were considered with respect to the height in the tower , the residence t ime and the s uperfic ial l iquid veloc i ty in order to s t udy the performance o f the to>ver fermenter when used to ferment whey permeate and to determine optimum operating condi t ions for this substrate and yeast cul ture . The averaged data for each superf ic ial l iquid veloci ty have been g iven in Appendix B . 3 . The following conversion data can b e used to convert the super ficial velocity (V ) to o ther related parameters as s required . Dilut ion rate (D ) based on the total tower he ight o f 2 . 3 7 m . D 1 . 44 V (mm/ s ) h- l s Volumetric f low rate ( Q ) Q Five 1 8 80 V (mm/ s ) s medium feed rates V s mm/ s 0 . 04 4 0 . 08 0 0 . 1 7 0 . 24 0 . 3 0 ml /h were used D h- 1 Q ml /h 0 . 063 80 0 . 1 2 1 5 0 0 . 25 3 1 0 0 . 34 440 0 . 4 3 5 5 0 4 . 1 THE RELATIONSHIP B ETWEEN TOWER HEIGHT AND VARIOUS FERMENTATION PARAMETERS 4 . 1 . 1 Lac tose concentration and u t i l i za t ion The lac tose concentra t ion of the ferment a t ion bro th decl ined with a n increase i n the height in the tower ( f ig . 4 . 1 ) . 7 1 .-I - 40 OD 30 w (/) 0 E--1 u < .....:l .-I (/) 20 10 72 2 0 . 5 1 . 0 1 . 5 TOWER HE IGHT , m 4 V mm/ s s • � - - 0 . 04 4 • -- 0 . 08 0 o -- 0 . 1 7 £> - --- 0 . 24 .... - -- 0 . 30 2 . 0 Q ml / h 80 1 5 0 3 1 0 4 4 0 5 50 2 . 5 5 Fig . 4 . 1 Lac tose concent ra t ion vs t ower height a t various superficial l iquid velo c i t ies . There was a very rap id decrease in lactose concentra t ion , f rom 40 t o 8 g /1 over t he f irs t 0 . 096 m o f t he tower t o sample p o in t 1 for a low superficial l iquid veloc i ty (V ) ( 0 . 04 4 mm/ s ) . The lac tose concen tra t ion s then decreased s lowly to 2 g / 1 , a t 0 . 82 m height ( s ampl e p o in t 2 ) , and reduced further to 1 g / 1 , a t 1 . 5 7 m. It remained at this concentra t ion over the remainder of the tower heigh t . This corresponded to 85 and 96% lac tose u t i l izat ion a t heigh t s o f 0 . 09 6 and 0 . 82 m , res pec t ively ( table 4 . 1 ) . The rate o f reduc t ion in lactose concentra t ion with heigh t was l ower at velocit ies b e tween 0 . 080 and 0 . 24 mm/ s . The concentrat ion was reduced from 40 g/1 at the inlet to 32 g / 1 at 0 . 096 m and to 4 g / 1 at 0 . 82 m when the veloc i ty was 0 . 080 mm/ s . Fur ther increase in height had only a marginal e f f e c t on further reduct ion o f lac tose concentrat ion which was 1 . 5 g / 1 a t the exit ( 2 . 6 9 m) . The corresponding lactose u t iliza t ion a t 0 . 096 and 0 . 8 2 m was 35 and 9 1 % , respe c t ively . S imilar pro files with only minor variations in lactose concentration at 0 . 096 and 0 . 8 2 m were observed for veloci t ies of 0 . 1 7 and 0 . 24 mm/s ( table 4 . 1 ) . 7 3 Table 4 . 1 Lactose utilization at various tower heights and super f ic ial l iquid veloc ities . locat ion/height slu ( % ) , a t various V (mm/s ) s m 0 . 044 0 . 080 0 . 1 7 0 0 . 24 0 . 30 1 0 . 096 85 35 3 2 2 7 1 4 2 0 . 82 9 6 9 1 9 1 9 0 5 7 3 l . 5 7 98 9 2 9 3 9 4 8 4 4 2 . 32 9 7 9 2 94 95 9 0 5 2 . 6 9 9 7 9 3 9 5 9 6 9 2 When the velocity was increased to 0 . 30 mm/ s the r educ tion in c onc entra tion with height wa s very much s lower . The concentrat ion rema ined high at 34 g / 1 , at 0 . 096 m , and 1 7 g / 1 at 0 . 82 m . At this velocity 2 . 3 2 m of height were required to reduce the concentrat ion to 4 g /1 . The corresponding lac tose ut il izat ion at 0 . 096 , 0 . 8 2 and 2 . 3 2 m was 1 4 , 5 7 and 9 0 % , respec t ively . I t was evident that at a cons tant veloc ity there was a decrease in the concentration o f lac tose with the increasing he ight . The height a t which lactose was reduced to a part icular concentrat ion increased when the velocity was increased . At a he ight o f 0 . 82 m , lac tose was reduced to l e s s than 4 g / 1 for super ficial veloc i t ies below 0 . 24 mm/ s corresponding to a lactose utilization of 90 to 94 % . This compared favorably wi th the batch fermentat ion util izat ion of 9 3 % (Appendix B . 2 ) . The resul ts presented here were averages o f f ive samples with at l eas t two res idence t imes ( T ) between each samplin g . This represented r more than 1 0 res idence t imes for each velocity used . The except ion to this was at the velocity of 0 . 30 mm/ s where only 2 sets of consecut ive s ampl ings were used . There was a s low reduc tion in the cell concentrat ion inside the tower ( s ec t . 4 . 6 ) and an increase in the exit effluent spec i f ic gravity ( SG ) ( fig . 4 . 26a) . This was an indication of cell wash out c e This super ficial veloc ity thus represents an approximat ion o f the c r i t ical veloc i ty above which the tower cannot be operated . Clearly this set o f readings canno t be regarded a s s teady s tate data o S imilar trends o f decreasing sugar concentration with increasing tower height have been observed for tower fermentation o f beer ( Greenshields & Smith 1 9 7 1 ) where the speci f ic gravity reduced rapidly f rom 1 . 03 5 , a t the inle t , t o 1 . 0 1 0 as the height increased to 2 m, when the veloci ty was 0 . 36 mm/ s . Further increase in height to 7 m caused a slow decrease in specific gravi ty to 1 . 006 . ..c ,....; - t;Q 74 The trend observed in the present work of rapid reduc tion in lactose concentrat ion with height at the low veloci ty of 0 . 044 mm/ s was very similar t o that observed in the tower fermentat ion o f fodder beet extrac t (Henderson & Smith 1 98 2 ) . Sucrose concentration showed a rapid reduc tion from 1 00 to 8 g / 1 as the height increased from 0 to - 1 0 . 09 5 m at a veloc ity o f 0 . 04 2 mm/ s (D = 0 . 2 h ) . Further increase in hei gh t to 0 . 4 8 m caused a slow decrease in sucrose concentrat ion to 3 g/1 . 4 . 1 . 2 Volumetric rate o f lac tose utilization ( S i ) The ra te o f lactose u t il i za t ion ( Si ) des cribed here was calculated as the mean for each o f the 5 tower sections ( f ig . 3 . 2 ) and plotted agains t the heigh t at the mid-point o f each section . The mean volumetric rate decreased rapidly with increasing height in the t ower ( fig . 4 . 2 ) . L::.. l 2 3 \ � 100 \ V mm/ s Q ml /h :z: 0 H f-; w U) 0 f-; u x-. 0 w f-; 22 u H � f-; w § ....:l 0 ::> 80 60 40 20 \ 0 \ \ \ \ \ \ \ \ \ \ \ \,\ \ I '\�>, ��- .. --- s • --- 0 . 044 80 • 0 . 08 0 1 50 0 - - 0 . 1 7 3 1 0 /::; - - -- 0 . 24 440 ,. - -- 0 . 0 5 5 0 - ,....; �' ... � 0 80 + , � 24 0 . 30 ---U) 0 . 044 -t �' � - ----... ... _ 0 . 5 1 . 0 1 . 5 2 . 0 2 . 5 MEAN TOWER HEIGHT , m Fig . 4 . 2 Volumetric rate o f lac tose ut il izat ion vs mean tower height at various superfic ial l iquid veloci ties . 7 5 I t was greatest in the f irs t tower sec t ion below 0 . 096 m and reduced t o near zero as the mean height increased to 1 . 20 m f or all superf ic ial l iquid veloci ties except the highest ( 0 . 30 rnrn/s ) . At the lowes t s uperf ic ial velocity (0 . 04 4 rnrn/ s ) , there was a rapid reduc t ion in the rate from 66 g / lh at 0 . 048 m to l g / lh at 0 . 4 6 m . I t then decreased to a rate b elow 1 g / lh over the r emaining height . At 0 . 080 rnrn/s , the rate reduced from 4 5 to 1 0 g /lh as the mean height increased from 0 . 048 to 0 . 46 m . It then reduced further to less than 1 g /lh at 1 . 2 0 m . S imilarly f o r veloc ities o f 0 . 1 7 and 0 . 24 rnrn/ s , the rate reduced to less than 1 g / lh above 1 . 2 0 m . At the crit ical velocity ( 0 . 30 mm/ s ) the mean rate over the in i t ia l sect ion o f the tower (up to 0 . 4 6 m) was low compared t o those at lower veloc i t ies , sugges t ing a lower yea st concentrat ion was present . Utilization occurring in the higher sect ion o f the tower ( from l o 2 0 t o 1 . 95 m) was greater than at veloci t ies less than 0 . 30 rnrn/s because o f the greater concentration o f yeas t and lactose in these s e c tions . The resul t showed that the greatest mean rate o f lactose utilization occurred at the lower sec tion o f the tower up to 0 . 82 m and for a l l veloci t ies up to 0 . 24 rnrn/s . There was l i ttle lactose utilization at heights greater than this . At a veloc ity o f 0 . 30 rnrn/s f ermentation was ob served over the full height o f the tower including the separa tor . The trend observed here was s imilar to the trend observed for the tower fermentation o f fodder beet extract (Henderson & Smith 1 9 82 ) . At a velocity o f 0 . 04 2 rnrn/s (D = 0 . 2 h- 1 ) , the volumetric rate o f sucros e utilization decreased rapidly from 7 2 to 3 g/lh a s the height increased from 0 . 09 5 to 0 . 29 m . A further reduct ion to l g / lh oc curred be tween 0 . 38 m and the exit at 0 . 7 6 m . 4 . 1 . 3 Speci f ic rate o f lactose u tiliza tion ( q 1 ) The mean spec i fic ra te o f l a c tose util ization ( q 1 ) decreased rapidly to a low value as the mean tower height increased ( f ig . 4 . 3 ) . The spec i f ic rate decreased f rom 0 . 68 to 0 . 02 g / gh as the mean height increased f rom 0 . 048 to 0 . 4 6 m at the lowest velocity ( 0 . 044 rnrn/s) . Beyond this mean height the spec ific rate was less than 0 . 02 g / gh . Thi s was because lactose was 9 6 % util ized when the height o f 0 . 82 m was reached so there was very l ittle lactose util izat ion after this height . At the higher velocities o f 0 . 080 , 0 . 1 7 and 0 . 24 rnrn/s , the specif i c ..c bJ) - bJ) z 0 H E--< .0:: N H ..-1 H E--< :::J � CIJ 0 E--< u .0:: ..-1 � 0 � E--< 2Z u H � H u w 0.. CIJ r-1 0"' 7 6 1 2 3 4 l:l. \ 1 . 6 V rnm/s Q ml /h s • ..;... _ _ 0 . 044 8 0 • - 0 . 080 1 5 0 o -- 0 . 1 7 3 1 0 1 . 4 6. - --- 0 . 24 4 4 0 ... - -- 0 . 30 5 5 0 1 . 2 ... 1 . 0 0 . 8 0 . 6 .. .. ....... " ' 0 . 4 ' 0� 1 7 "" ' ' 0 . 2 "" ' ' MEAN TOWER HEIGHT , m F ig . 4 . 3 Sp ec i f i c ra t e o f l ac t o s e u t i l i za t ion vs mean towe r h e i gh t a t various superf ic i a l l iq u i d veloc i t i e s . ra t e d e c r ea sed f rom b e tween 0 . 45 and 1 . 64 g /gh t o l es s t h a n 0 . 1 g /gh a s the mean h e igh t i nc r eas ed f rom 0 . 0 4 6 t o 1 . 20 m , a t this h e i gh t l a c tos e was 9 5 % u t i l i z e d ( ta b l e 4 . 1 ) . Beyond th i s h e i gh t , t he s pec i f i c ra t e wa s l owe r t han 0 . 1 g /gh s ince th ere wa s l i t t l e l a c to s e l ef t to b e u t i l i ze d . A t t h e h i gh e s t veloc i t y o f 0 . 30 mm/ s , t h e s pe r i f i c r a t e d e crea s ed 7 7 less rapidly than a t lower veloci t ies . The speci f ic rate fell from 1 . 07 to 0 . 1 0 g /gh as the mean height increased from 0 . 048 to 1 . 9 5 m . This slow decrease occurred because there was lactose utilization throughout the tower a s a resul t o f the decreased residence t ime . The cell concentrat ion at the bot tom of the tower was reduced whil e there was an increase in lactose concentration at greater heights as the veloc ity increased . Thus , lactose u t i l i zat ion occurred throughout the tower . I n the separator , the spec i f ic rate was zero f o r a l l veloc it ies . The resul t s showed that the specific rate o f lac tose utilization decreased as the height increased and the height at which the spec i f ic rate was zero depended on the residence t ime and the cell concentrat ion . I t was pos s ible that ethanol cencen trat ion would have an effect on the spec ific rate , s ince e thanol concen trat ion increased with height ( f ig . 4 . 4 ) , thus , e thanol inhibi t ion could increase with height . However , the maximum ethanol concentration r eached was 2 3 g/ 1 which was cons iderably l ower than a reported inhibi t ing concentrat ion o f 30 g /1 (Wendor f et al 1 9 70a) . The t rends observed were s imilar to those obtained in the tower fermentat ion of molasses ( f ig . C . 3 ) . However , the spec ific ra te of sugar utilizat ion was greater in the molasses and at a s imilar veloc ity , fermentation was completed at a mean height o f 1 , 20 m . The spec i f ic rate was greater due to the higher feed sucrose concentration used ( 1 00 g / 1 ) and the greater rate at which sucrose could b e utilized . 4 . 1 . 4 E thanol concen tration ( E ) Ethanol concentrat ion increased with increas ing height for a given superficial veloc ity ( f ig . 4 . 4 ) . The ethanol concentration rose rapidly with height at the lowest veloc ity ( 0 . 044 mm/ s ) . The concentration increased from 0 to 2 0 g /1 at 0 . 09 6 m and further increase in the height had minor e f fect on the concentrat ion , which was 2 2 g/1 at 2 . 3 2 m . There was a slower rate o f increase i n ethanol concentrat ion with height at greater veloc ities up to 0 . 24 mm/ s . When the veloc i ty was 0 . 080 mm/ s , the concentration increased to 5 . 8 g/1 at 0 . 09 6 m, to 1 7 g / 1 at 0 . 82 m , and t o 1 8 g / 1 a t 1 . 5 7 m . Further increase i n height had no s ignifican t effect on the concentration which remained at 1 8 g / 1 until the exit was reached . S imilar pro files with minor variat ions in concentrat ion were observed for velo c i t ies of 0 . 1 7 and 0 . 24 mm/ s . 20 r-1 � 1 5 / I 0 . 5 2 �: 1 . 0 7 8 4 5 ---- - ··�o---- ------ -----. � - - -0..- � -- ..._. - ------:r .,- ------== � 1 . 5 V mm/ s s · -- - 0 . 044 • 0 . 080 0 - - 0 . 1 7 6 - --- 0 . 24 T -- -- 0 0 2 . 0 Q ml /h 80 1 5 0 3 1 0 440 550 2 . 5 TOWER HEIGHT , m Fig . 4 , 4 Ethanol concentrat ion vs t ower height at various super f ic ial l iquid veloci t ies . There was a slower increase in the concentrat ion wi th height a t the great e s t veloc i ty ( 0 . 30 rnrn/ s ) . A small amount o f e thanol was produced in the init ial 0 . 096 m of the tower and at 0 . 82 m less than 2 / 3 of the f inal e thanol concentration was produced ( 1 1 g / 1 ) . At this velocity the full fermenter height including the s ep arator ( 2 . 69 m) was required to achieve an ethanol concentration of 18 g /1 . The re sul t s showed that for all veloci t ies up to 0 . 24 mm/ s , mos t o f ethanol was p roduced wi thin the f irst 0 . 8 2 m of height . Thi s was a reflect ion o f the trend obtained for the lac tose reduc t ion because for these velo c i t ies lactose was more than 90% u t ilized wi thin this height . At the greatest veloc i ty ( 0 . 30 rnrn/s ) , ethanol concentrat ion increased cont inuously as the height increased because lactose was b eing u t i l iz ed cont inuously over the ent ire tower he ight . The t rend observed here , at the lowes t velocity ( 0 . 04 4 mm/ s ) was similar to the increase in ethanol concentrat ion with height ob served in the tower fermentat ion of sugar beet extrac t (Henderson & Smi th 1 982 ) . They noted tha t the concentra t ion increased rapidly to 4 8 g / 1 '""' '"Cl rl -. rl (lj u ·rl 4-J =l '"""" � H >< '"-" 0 � � � � >< 7 9 a s the height increased t o 0 . 095 m when the velocity was 0 . 04 2 mm/ s . From 0 . 09 5 m to the exit , the concen tration increased by only 3 g / 1 to 5 1 g / 1 . 4 . 1 . 5 Ethanol yield (Y) For all super ficial velocities but one ( 0 . 24 mm/s ) , the ethanol yield coe f f i c ient did not vary greatly between 0 . 82 and 2 . 3 2 m ( f ig . 4 . 5 ) . 1 00 90 80 70 t 1 50 \% at 0 . 09 6 m \ \ \ ' .....__----�... 0 . 04j_- - ---- ... ./ -= - -,::- - -- ---�-'� 0 . 30 ..... "" U' 1 ..... ..... 0 . 5 / � ..... / 2 0 . 080 --- --- -- ----,.._.- - - - ,..- ,..­ ,.. ,.. ,.. 1 . 0 V -- - _.... 6. - 0 . 24 1 . 5 V mm/ s s • --- 0 . 044 • 0 . 080 0 - - 0 . 1 7 6 - --- 0 . 24 0 2 . 0 TOWER HEIGHT , m Q ml/h 'c. 80 1 5 0 3 1 0 4 4 0 5 5 0 2 . 5 5 Fig . 4 . 5 Ethanol y ield vs tower height at various super f ic ial l iquid veloci t ies . The y ield was between 82 and 9 3 % . At velocity o f 0 . 24 mm/ s , the yield increased f rom 7 1 to 80% as the height increased f rom 0 . 82 to 2 . 32 m . The y ield at 0 . 09 6 m varied considerably be tween 6 3 to 1 5 0 % for all veloci t ies . The probable explanation for this behaviour is that , because o f the high sugar and low ethanol concentrat ions at this point , small f luctuation in the measured sugar and ethanol concentrat ions resul ted in considerable variat ion in the calculated yield value . The 80 uncertainty in the yield value a t this height was estimated to be as high as 490% ( sec t . E . 3 ) . When the effluent left the tower ( 2 . 69 m) , the yield was generally lower than ins ide the tower (between 1 and 5% ) for all veloc i t ies except 0 . 30 mm/ s . This small reduc tion could have been caused by gas s tripping of ethanol when carbon d ioxide le f t the tower and consumpt ion of ethanol by K.marxianus ( Sarfacon et al 1 9 7 2 ) . However , this was no t s tu died further . The trend observed between 0 . 82 and 2 . 32 m was similar to the trend observed in the tower fermentation o f fodder beet extrac t (Henderson & Smith 1 98 2 ) . They observed that the yield was invariant with height for all velocitie s used . The yield was between 94 and 98% . Thus , for all veloc i ties , there was l it tle variat ion in the yield of ethanol within analyt ical uncertainty ( 1 2% from sect . E . 3 ) within the heights o f 0 . 82 to 2 . 32 m at a particular velocity . At 0 . 096 m , there was considerable variat ion in the yield due t o analytical error . The yield of e thanol in the e ffluent was sl ightly lower than that in the tower . 4 . 1 . 6 Volumetric rate o f ethanol product ion (E ' ) The volumetr ic rate o f e thanol product ion (E ' ) ( fig . 4 . 6 ) was plot ted in a s imilar way to that described for the volumetric rate o f lactose utilization ( sec t . 4 . 1 . 2 ) . 30 \ \ j 4 V mm/ s s · ..:.. -- 0 . 044 • -- 0 . 080 o -- 0 . 1 7 6 - --- 0 . 24 ... - -- 0 . 30 \ \ \� � ""-.... '- =-----,....._ ....... ... O . L t......_ ' , Y .. ---.£ . 30 � ' - - '-e:. _ _ J:2 . 24--- - - ... __ - - - - - 0 . 5 1 . 0 1 . 5 2 . 0 MEAN TOWER HEIGHT , m 5 Q ml /h 80 1 5 0 3 1 0 4 40 5 5 0 2 . 5 Fig . 4 . 6 Volurne tric ra te o f ethanol produc t ion vs mean tower height at various superf ic ial l iquid veloci t ie s . 8 1 At a constant velocity , the volumetric rate o f ethanol produc t ion decreased with increasing mean height in a similar way to the trend ob tained for the volumetric rate of lac tose u til ization ( f ig . 4 . 2 ) . At the lowes t veloc ity ( 0 . 044 mm/ s ) , the volumetric rate decreased rapidly from 33 g/lh to less than l g/lh as the mean height increased from 0 . 04 8 to 0 . 4 6 m and remained below 1 g /lh over the remaining sec t ions of the tower . The pro file was s imilar for veloc ity of 0 . 1 7 mm/ s . At velo c i t ies o f 0 . 24 and 0 . 30 mm/ s ini t ial rates remained high , but e thanol production occurred over more o f the to tal tower height . For all velocities ethanol produc t ion ceased inside the separator ( 2 . 5 1 m) . In all cases , the volumetric rate o f e thanol produc t ion was high within the f ir s t 0 . 82 m because o f the high lactose and cell concentrat ions presen t . At greater height s,the volumetr ic rate was very low because there was only a small amount of lac to se l e f t to be util ized . Thus , the volumetric rate o f e thanol production at a part icular tower height followed the observed trend for the volumetric rate o f lactose utilization . For all velocities under 0 . 30 mm/ s , the height greater than 0 . 82 m was no t essent ial . The trend observed here was s imilar to that observed in tower fermentation o f fodder beet extract (Henderson & Smith 1 9 82 ) . They found that the volumetric rate o f e thanol produc t ion decreased rapidly f rom 38 to 2 g/lh over the first 0 . 38 m of the tower he ight when the velocity was 0 . 042 mm/s (D = 0 . 2 h- 1 ) . It then reduced further to a low value ( 0 . 6 g / lh) and showed no signi f icant change be tween heights o f 0 . 38 and 0 . 7 6 m . The effluent left the tower a t 0 . 7 6 m . 4 . 1 . 7 Spec i f ic rate o f ethanol product ion (v) At a constant velocity , the spec ific rate of ethanol produc t ion (V ) showed a general trend o f decrease from high value (between 0 . 2 and 0 . 7 g /gh) at a mean height o f 0 . 048 m to zero with increasing height in the tower ( f ig . 4 . 7 ) . The mean height at which the specific ra te reduced to zero was a f fec ted by the velocity being , further up the tower at a grea ter velocity . The spec i f ic rate decreased with height because there was a decrease in lactose concentrat ion as the height increased but the cell concentration , at a part icular velo c i ty , be tween 0 . 8 2 and 2 . 32 m , s howed 82 l 2 3 ' V mm/s s Q ml /h 80 \ .c � >Lo -o eo 0 . 6 e ..:.. -- o . o44 • -- 0 . 080 o -- 0 . 1 7 to. - --- 0 . 24 ... - - - 0 . 30 1 5 0 3 1 0 4 4 0 5 5 0 0 . 5 1 . 0 1 . 5 2 . 0 2 . 5 MEAN TOWER HE IGHT , m Fig . 4 . 7 Spec i f ic rate o f ethanol produc tion vs mean tower he igh t a t various superficial l iquid veloc i t ies . l i ttle change as the height increased due to cell recycl ing ( fig . 4 . 9 ) . Thus , the amount o f e thanol produced by an almos t cons tant c e l l concen tration , decreased with heigh t . Henc e , the speci f ic ra te o f ethanol product ion decreased with hei ght . The trend o f decrease was s imilar to the trend ob served in the tower f erment a t ion of molasses ( s ec t . C . l . 2 ) at s imilar veloc i ty but the rate observed in the molasses f ermentat ion was higher ( be tween 2 " 8 and 7 . 0 g / g h , at a mean height o f 0 . 048 m) . There was a trend of decreasing spec ific rate o f e thanol product ion with an increase in the ethanol concentrat ion ( f ig . 4 . 8 ) . However , the evidence available was insuffic ient to conclude that this reduc t ion was due to e thanol concentrat ion alone because there was low lactose concentrat ion when e thanol concentrat ion was high ( f ig . 4 . l c ompared with f ig . 4 . 4 ) while the cell concentrat ion was kept h igh and changed l it tle b e tween 0 . 8 2 and 2 . 3 2 m ( f ig . 4 . 9 ) . Thus , the reduc t ion o f the speci f ic rate with an increase in e thanol concentrat ion was probably due to a combined effect of the av a i l ab i l i ty of lactose , e thanol concentration and the high b iomass concentrat ion . 83 ..c: bi) -bi) I I I I I I I I I z 0 V mm/ s Q ml /h 6 \ s -H E-< u ::::J � 0 . 6 0 � p., ....:l 0 � 0 . 4 ::r:: E-< w � 0 \ • 0 . 044 80 \ • 0 . 080 1 5 0 1- \ T 0 0 . 1 7 3 1 0 -\ 6 0 . 24 4 4 0 \ 1- \ T 0 . 30 550 \ f- 0 \ -\ '\ T ' 6 • - w E-< � 0 . 2 " ....... ' f- · • '-Q. T --. u - - - - - - -- -H � H r- • T 6 -u w p., Cl) I I I I I I I I � 4 8 1 2 1 6 2 0 ::> E , ETHANOL CONCENTRATION , g / 1 Fig . 4 . 8 Specific rate o f e thanol product ion v s e thanol concentration at various superf icial l iquid veloc i t ies . 4 . 1 . 8 Cell concentrat ion For all values o f superf i c ial l iquid velocities ( V ) , the cell s concen tration ( f ig . 4 . 9 ) was h ighest at height o f 0 . 09 6 m with values in the range of 5 6 to 100 g / 1 d ried weight (DW) and 220 to 350 g / 1 centrifuged wet weight (WW) . The exact values depended on the veloc ity and will be d iscussed in sect . 4 . 3 . 4 . As the height in the tower increased from 0 . 096 to 0 . 8 2 m , the cell concentration decreased rapidly . Then as height in the tower increased fur ther to 2 . 32 m, the concentration decreased at a s lower rate . The excep t ions to thi s t rend were for the velocities o f 0 . 24 and 0 . 30 mm/ s at which the cell mas s was more evenly dis tributed throughout the tower because of fluidization of the floes . In all cases , the concentrat ion of . c el l s leaving the fermenter from the separator was less than 0 . 5 g / 1 DW ( 1 8 g / 1 WW) . The resul t s showed that the cell concentrations inside the tower between 0 . 82 to 2 . 32 m were much lower than at height of 0 . 096 m . Highes t concentrat ions corresponded with the grea test fermentative activity ( u p to 0 . 8 2 m) . An increase in the superf ic ial velocity resulted in the gas carbon d ioxide being produced at a higher rate because of the increase in the ra te of lactose utilization with velocity ( sec t . 4 . 3 . 3 ) . 84 ..-l ........ bi) ( a) � 10 0 E-< t§ H i:Ll � Cl i:Ll H p:: Cl ,.-4 ,.-4 i:Ll u � � Cl ..-l ........ !)J) E-< :c v H '-'-1 3: ['--< w 3: Cl w v ;:::::, � H p:: ['--< z w u ..-4 ,.-4 w u � � 80 60 4 0 20 300 200 100 a�� ·�� .... � - A 0 . 2 4 ::1:330 ( b ) ... ...,. _ _ _ _ _ - --- - - - - - -� � � ------ - -....... --.._. -::. 1 -=---- --o-LlL '� -- - - - - e-.0....0!4_ _ -c.........:._:\ -1- 0 80 ----- � - V mm/ s s ·- -- 0 . 04 4 • 0 . 080 0 - - 0 . 1 7 6 - - -- 0 . 24 'Y -- - - 0 . 0 Q ml /h 8 0 1 5 0 3 1 0 4 4 0 5 5 0 A � � '� , "� l ' "'-� --A-0 . 24 - - - - - - · ::-....."\.. "--0 - - - -� • " 0 . 5 ��.:::::::::--=:..::. t . 3 0 0_iL_17_ ___ "- - - - -- - e-!:L.] 4 4 o , \ 2 3 1 :0 1 . 5 TOWER HEIGHT , m --- " \ 80 - -- ·-- �\ 2 . 0 2 . 5 Fig . 4 . 9 Cell concentration vs tower height at various super f i c ial liquid veloc it ies . Thus , the up f l ow movement o f this gas and the fermen t ing l iquor comb ined to l i f t part o f the cell population to a greater heigh t . At the highe s t vel o c i ty ( 0 . 3 0 mm/ s ) , this combined upl i f t ing act ion caused the devia t ion in the pro file from the general trend of decreas ing cell concentration with height between 0 . 8 2 and 2 . 3 2 m . There was a corres pon- d ing increase in the fermentative act ivity within these great er heights . In general , the cell concentrat ion reached inside the tower (between 0 . 8 2 and 2 . 3 2 m) was lower than 40 g / 1 DW ( 1 5 0 g / 1 WW) b ecause the yeast used was only modera tely flocculent in whey permeat e . 85 This was lower than the cell concent rat ion reached in other tower fermentat ions using S. cerevisiae . An average value o f 5 7 g / 1 DW ( 25 0 g / 1 WW) was repor ted f o r beer fermentation ins ide a tower fermenter ( Greenshields & Smith 1 9 7 1 ) . Prince & Barford ( 1 9 8 2 ) operated their tower fermentation o f cane j uice us ing cell c oncentrations of 54 to 83 g / 1 DW ( 2 2 0 to 230 g / 1 WW) . Henderson & Smith ( 1 982 ) operated their tower fermentat ion o f beet extract s us ing cell concentrations of 1 8 to 80 g/1 DW (90 to 320 g/1 \1W) . On the o ther hand , Chen & Zall ( 1 98 2 ) operated their AFEB fermenter to fermen t whey using a non-f locculent lactose-fermenting yeast at a concentrat ion o f 4 to 2 2 g /1 DW . The very flocculen t yeast used by the f ir s t two group s of workers permit ted a much greater med ium throughput than in this s tudy . The cell concentrat ion reached in the present s tudy of the tower fermentation of whey p ermeate was approximately hal f tha t reached in the tower fermentation using more flocculent S . cerevisiae s trains . Thus , the fermentation capabil ity o f the culture as a whole was l imited by the yeast s train used . The typical appearance o f K. marxianus Y42 floes in the tower a t the velocity o f 0 . 1 7 mm/ s is shown in f ig . 4 . 1 0 . In the inle t zone and sample p o int 1 ( 0 . 09 6 m) , the cell concentration was very high ( 1 00 g / 1 DW) ( f ig . 4 . 10 a ) . The floes were large ( 1 to 3 mm in d iameter ) . Some l arge floes were broken up by s treams o f whey permeate ent ering the tower . From sample point 1 upward (above 0 . 096 m) , there was considerable carbon dioxide produc t ion . The upward movement o f this gas and the medium provided l i f t ing action t o suspend and disperse the yeas t floes throughout the tower . Inside the separator ( f ig . 4 . 1 0 b ) , the beer and yeast floes entered the baffled sect ion of the separator f rom . the top o f the separa tor draught tube (upper left hand side ) . Carbon d ioxide separated f rom the beer and le f t the tower via the t o p exit port (no t shown) . The beer and the yeas t floes flowed down pa s t the bot tom o f the baffled section and then up toward the effluent exit (on the r ight hand s ide ) . The yeast f loes that were suf f ic ien t ly dense to resist the upward mo t ion of the fluids returned to the tower s t raight sect ion . There was a very clear separat ion between the eff luent and the f loc suspension . I t was observed that this zone of suspended yeast f loes ac ted as a f il ter that trapped free yeast cells to form larger floe s . When this zone of suspended yeast floes was not present (at the low velocities of 0 . 044 and 0 . 080 ( a ) ( b ) Fig . 4 . 1 0 Tower fermenter during whey permeate fermentation at superficial l iquid velocity of 0 . 1 7 mm/ s ( 3 1 0 ml /h) . ( a ) Tower bot tom showing large yeast floes ; feed and air inlet ( lower centre) ; sampling port 1 (mid-right ) , (b) Separator . CXl 0\ 8 7 mm/s) , there was a slightly greater cell loss with the effluent . 4 . 1 . 9 Medium pH The pH of the fermentation broth decreased with height over the initial 0 . 82 m ( fig . 4 . 1 1 ) . The influent pH of 4 . 7 to 5 . 4 was reduced to between pH 4 . 4 and 4 . 8 over the initial 0 . 096 m, and then decreased slowly to between pH 4 . 2 and 4 . 6 as the height increased to 0 . 82 m. Further height increase beyond this to the exit ( 2 . 69 m) , resulted in no s ignificant change in pH. It remained between pH 4 . 2 and 4 . 6 . pH 5 . 0 4 . 5 0 . 080 V mm/ s s • � - - 0 . 044 • - 0 . 080 o -- 0 . 1 7 t::. - --- 0 . 24 • --- 0 . 30 Q ml /h 80 1 50 3 1 0 440 550 "'-:._ - � 0.&]0 A -- ....... ---- - - A - 0:24 ---.::a=. --_.._----=----3f 0 . 5 2 3 4 5 1 . 0 1 . 5 2 . 0 2 . 5 TOWER HEIGHT , m Fig . 4 . 1 1 Medium pH vs tower height at various superficial l iquid velocities . The high initial pH of greater than pH 5 was due to the aeration of the medium. Aerated medium was found to have pH betwe�n 5 . 0 to 5 . 4 even though the pH before autoclaving was 4 . 6 . Once the medium was in the inlet zone at the bottom of the tower , there was considerable fermentative activity . Ethanol and carbon dioxide were produced . This caused the initial pH reduction . The slower reduction , b etween 0 . 096 and 0 . 82 m , was a result of a much l ess intense fermentative activity . Above 0 . 82 m , the pH remained constant because there was virtually no fermentation taking place . The lower effluent pH of 4 . 2 , when the velocities were 0 . 24 and 0 . 30 rnm/s , was caused by a new season whey permeate having a lower initial pH ranging from pH 4 . 2 to 4 . 4 . Even though the pH was corrected to pH 4 . 6 before autoclaving , the operating pH still decreased to between 4 . 2 and 4 . 4 inside the tower . 88 The trend observed was similar to that observed in the tower fermentation of molasses ( sect . C . l . 5) but the operating pH was higher at pH 5 . 0 . During the tower fermentation of whey permeat e , a sudden decrease in the pH to below pH 4 . 0 in association with a reduction of ethanol concentration ins ide the tower was found to be an indication o f bacterial contaminat ion ( sect . 4 . 6) . 4 . 2 THE EFFECT OF THE RESIDENCE TIME ON VARIOUS FERMENTATION PARAMETERS The liquid residence t imes (T ) used in this section were calculated r using the tower height at the respect ive sample point s . I t could be referred to as " apparent fermentation t ime " . The mean res idence time (T*) was , however , calculated in a different manner from the true r residence t ime (T ) . It was calculated using the mean height of each r tower section (sect . 3 . 1 1 . 1 and Appendix B . 3) . The mean re sidence t ime was plotted versus the mean rates in each tower section . 4 . 2 . 1 Lactose concentration ( S 1) Lactose concentration decreased rapidly from about 40 g / 1 to about 6 g / 1 as the residence time increased to 1 hour and then showed a small reduction to between 2 and 4 g /1 as the residence time increased further to 15 hours ( fig . 4 . 1 2 ) . The rapid reduction of lactose concentration during the first hour in the tower fermenter was a result of the high cell concentrat ion within the lower t ower section (up to 0 . 82 m) . Lactose WqS utilized without any delay normally required to build up the cell concentration in b atch fermentation . The fermentation t ime observed here was very much less than the 16 hours required to complete the batch fermentation (Appendix B . 2 ) . It was also very much shorter than the 1 2 hours reported for continuous fermentation of whey containing 5 8 g /1 lactose using two continuous st irred-tank fermenters in series (Reesen & Strube 1978 ) . Further comparison could be made with the longer residence time of 3 . 9 hours reported for 85% utilization of sugar in continuous fermentation of SO g/1 lactose in whey using an immobil ized yeast system (Linko et al 1 9 8 1 ) . The res idence t ime of 1 hour for 90% lactose util izat ion observed in the present work was only sl ightly longer than a re sidence t ime of 40 rl 'M 30 .. � Ul 0 H u < o-l � 20 Ul 10 I I • I I I I I I I I .... I I I I I el ..... 8 9 V mm/s s Q ml/h • 0 . 044 80 • 0 . 080 1 5 0 0 0 . 1 7 3 1 0 A 0 . 24 440 • 0 . 30 550 Q� .6. � .6. .o_- - - - _.JI! • . .- -- -- -- -- -- -- - -- - - - - - 5 1 0 T , RES IDENCE TIME , h r F ig . 4 . 1 2 Lactos e concentration vs r es idence time at various superficial l iquid velocity . �0 . 7 hour observed during tower fermentation of cane molasses in which 96% u tilization of 100 g/1 sucros e occurred ( f ig . C . 7a ) . The trend obs erved here was s imilar to that reported for tower fermentation of beer (Greenshield & Smi th . 19 7 1 ) . However , the beer wart c ontained many different sugars . Thus , as the r esidence t ime increas ed , there was a slower reduction in the sugar concentration fol lowing the initial rapid reduction . One hour was required to reduce the f ermenting l iquor specific gravity from 1 . 0 35 to 1 . 0 1 0 , but further reduc tion from 1 . 0 10 to 1 . 00 5 required 1 9 hours . 4 . 2 . 2 Ethanol concentration (E) Ethanol conc entration increas ed with increasing residence time ( f ig . 4 . 1 3 ) . The concentra tion increased rapidly from 0 to greater than 1 6 g/1 as the residence time increased from 0 to 1 hour . Further increase 90 in the res idence time resulted in a smal l increase in the concentration to greater than 1 7 g /1 . The final ethanol concentrations reached in these experiments were different for various superficial l iquid velocities because of the variation in the feed lac tose concentration (Appendix B . 3) . 20 � 15 - ClO .. o-l 0 � t5 w 1 0 .. w 5 • • 0 0 6. - · - - - - - - - - - · - - - - - - - - - ---- ·- - - � ..,.tf I I I p I I T V mm/s s • 0 . 044 • 0 . 080 0 0 . 1 7 6 0 . 24 • 0 . 30 5 10 T , RESIDENCE TIME , h r Q ml/h 80 150 3 1 0 440 550 Fig . 4 . 1 3 Ethanol concentration vs res idence time at various superf icial l iquid velocities . I t was shown previous ly that up to 90% lactose ut il ization occurred in 1 hour ( f ig . 4 . 1 2 ) . Thus , as would be expected , mos t ethanol was produced during this time . The trend observed here was different from that observed during the init ial period o f ba tch fermentation ( fig . 7 . 1 ) . During the early s tage of batch fermentation7 there was a short delay before ethanol was produced . This period was then followed by a moderately rapid increase in the e thanol concentration to the maximum concentration . A fermentation time of 1 6 hours was required to produce 1 8 g /1 ethanol compared with the 9 1 1 hour required t o produce 1 6 g / 1 ethanol in tower fermentation . I t is eviden.t tha t in the tower fermentation o f whey permeate , the maj ority of the lac to se was util ized with the product ion of ethanol during the f irst hour in the f ermenter . There was no delay in the p roduction of ethanol as ob served in batch f ermentation . 4 . 2 . 3 The rates of lac tose util i zation and ethanol production The volumetric rates of lac tose utili zation ( S i ) and ethanol production ( E ' ) reduced from high values ( 10 8 and 37 g/lh) to 1 g /lh as the mean res idence time increas ed to 2 hours and were less than 1 g/lh when the mean res idence time was greater than 2 hours ( f ig . 4 . 14 and 4 . 1 5 ) . Similar trends were observed for the specific rate of lac tose utilization q1 ( f ig . 4 . 1 6 ) and the specific ra te of ethanol production (v) ( f ig . 4 . 1 7 ) . For all veloci ties , the two speci fic rates ( q1 and v) were less than 0 . 0 3 g /gh at mean residence t imes greater than 2 hours . Th e resul t s showed that , for all velocities , there was no s ignificant lactose consumpt ion after 2 hours in the t ower and as a resul t , there was also no ethanol production af ter 2 hours . This rapid f ermentation was made pos s ible by the high yeas t cell concentration achieved by b iomass feedback from the s eparator . Thus , an increase in the residence time o f the f ermenting medium ins ide this tower beyond two hours would result in no further utilization of the lactose or further production of ethanol . The h igh b iomass concentration was greater than the concentration that could be supported in a stirred-tank fermenter with no f eedback utilizing the s ame f eedstock . A similar trend was obs erved in the tower fermentation o f molas ses ( s ect . C . 2 . 2 ) . The mean res idence time at which the two rates of lac tos e utilization ( Si and q1) were almos t zero was 1 . 5 hours , while for the two rates o f ethanol production (E ' and v ) , the t ime was 1 . 2 hours . This was shorter than the time ob tained in the t ower fermentation of whey permeate , even though the op timum velocity reached in the tower fermentat ion o f mol ass es was higher ( 0 . 3 3 mrn/s ) and the feed sucrose was als o higher ( 100 g /1 ) . In the tower fermentation o f fodder beet extract ( Henderson & Smith 1 9 82 ) , the mean res idence t ime required for the volumetric rate of sucrose utili za tion to reduce from 1 00 g /lh to less than 3 g/lh was 2 hours , and a shorter t ime o f 1 . 5 hours was required for the volumetric rate o f ethanol produc t ion to reduce from 50 g /lh to less than 3 g/lh . This was s imil ar ..c .-1 - bO �100 :z 0 H E-t � N H �� 80 H � J:il Cll 0 E-t j 60 � 0 J:il � 40 u H p:: E-t � ....:I 20 0 :> � - .-1 Cll I I I , \.6. \ \ T o \ \ T '• 2 hours , + T. 2 9 2 4 6 V mm/s s • 0 . 044 • 0 . 080 0 0 . 1 7 6. 0 . 24 ... 0. 30 10 T* MEAN RES IDENCE TIME , h r ' Q m1 /h 80 1 50 3 1 0 440 550 12 Fig . 4 . 1 4 Volumetric rate of 1actose utilization vs mean res idence time a t vari ous superficial liquid velocities . ..c V mm/ s Q ml /h s .-1 30 � - 0 bO J:il � E-t :z � 0 H E-t u u H p 20 p:: Q • 0 . 044 80 • 0 . 080 1 50 0 0 . 1 7 3 1 0 t::. 0 . 24 4 40 ... 0 . 30 550 E-t � J:il s p.., ....:I ....:I � 0 � 10 - E-t J:il J:il hours 2 4 6 8 10 1 2 T� , MEAN RESIDENCE TIME , h Fig . 4 . 1 5 Volumetric rate of ethanol produc tion vs mean residence time a t various superficial liquid veloci ties . 9 3 1 . 6 V nun/ s s Q ml/h • 0 . 044 80 • 0 . 080 150 1 . 4 0 0 . 1 7 3 1 0 "' 0 . 24 440 ...c:: ... 0 . 30 550 00 -00 � 1 . 2 z 0 H � N H � H 1 . 0 E-t ::::> � U) 0 E-t u j 0 . 8 � 0 � � 0 . 6 ... u \ H \ � H \ ... u � \ p.., U) \ � 0 . 4 0 \ .-i 0' \ \ \ 0 . 2 10 12 T� , MEAN RESIDENCE TIME , h Fig . 4 . 1 6 Specific rate of lactose utilization vs mean res idence time a t various superficial liquid veloci ties . .;, � "'00 0 . 6 0 0 . 4 0 . 2 \ � \ 1-2 hours \ 94 6 8 V mm/ s s • 0 . 044 • 0 . 080 0 0 . 17 6. 0 . 24 'Y 0 . 30 10 T� , MEAN RES IDENCE TIME , h Q ml /h 80 1 50 3 1 0 440 550 12 Fig . 4 . 1 7 Specific rate of ethanol production vs mean residence time a t various superficial liquid veloci ties . to the time of 2 hours obs erved in the present work for tower fermentation of whey permeate . Thus , in a tower fermenter a sucros e bas ed sub s trate would be used up at a fas ter rate than lactose based sub s trate such as whey and so ethanol production would b e completed in a shorter time . The fermentation sec tion of such a tower could therefore be shortened . 4 . 3 THE EFFECT OF SUPERFICIAL LIQUID VELOCITY ON VARIOUS FERMENTATION PARAMETERS 4 . 3 . 1 Lactose concentration ( S 1 ) The effec t of the superficial l iquid velocity (V ) on the lactose s concentration varied with the heigh t in the tower ( fig . 4 . 1 8) . At the bot tom of the tower ( 0 . 09 6 m , curve 1 ) , the concentration increased rapidly from 7 to 28 g/1 as veloci ty increased to 0 . 30 mm/ s , but did not reach the feed lac tose concentration . At heights from 0 . 82 to 2 . 32 m including the s eparator ( 2 . 69 m) (curves 2 to 5) , there was a cons tant low level of lactose a t approximately 5 g / 1 at all veloci ties up to 0 . 24 rnrn/s . When a velocity of 0 . 30 mm/s was reached , there was an increase in lactose concentration . The increase in lac tose concentration at 0 . 09 6 m ( curve 1 ) was M -eo .. M Cll 30 10 9 5 0 . 2 1 HE m • 1 . 0 . 096 0 2 .- - 0 . 82 /:}. 3 .- --- 1 . 5 7 • 4 .- - - 2 . 32 0 5 .--- 2 . 69 0 . 3 V , SUPERFICIAL LIQUID VELOCITY , mm/s s Fig . 4 . 1 8 Lactose concentration vs superficial liquid veloci ty at various tower heigh ts . b ecaus e this sampl ing location was very close t o the inlet . As the veloci ty increased there was an increas e in the feed liquor velocity and c arbon dioxide produc t ion . These two fac tors loosened and lifted up the yeas t floes a t the bo ttom of the tower . There was a cons tant lactose concentration b etween heights of 0 . 82 and the exi t ( 2 . 69 m) at vel oci ty up to 0 . 24 mm/s becaus e lactose was reduced to 5 g/1 in the f irs t 0 . 82 m of the column . The variations in lactose concentration a t heights greater than 0 . 09 6 m at velocities above 0 . 24 mm/s were due to the shifting of the yeas t b ed up the column ( fig . 4 . 25 ) and the reduction of res idence t ime in each tower sect ion . Above this velocity ( 0 . 24 mm/s ) cell wash out occurred . The resul ts showed that for all veloci ties up to 0 . 24 mm/s , 88% of available lactose was utilized in the firs t 0 . 82 m, and at the highes t velocity (0 . 30 mm/s ) wash out occurred . A similar trend was observed in the tower fermentation of molass es ( fig . C . l O a) . The sucrose concentration increased rapidly with velocity 96 increase at a heigh t of 0 . 09 6 m. At heigh ts b etween 0 . 82 and 2 . 3 2 m , the concentration was general ly lower than 4 g/1 and increas ed when the veloci ty was greater than 0 . 33 mm/ s . At veloci ties greater than this , there was cell wash out and incomplete sucrose utilization . Thus , the wash out velocity during growth on molas s es was greater than for whey permeate even though the feed molas s es had a greater dens ity than that of whey permeate , and that the yeas t s train (S. cerevisiae FT146 ) , used in the molasses fermentation , was only moderately flocculent . A s imilar trend , was ob served in the tower fermentation of fodder beet extrac t ( Henderson & Smith 1 982 ) . The sucrose concentration at a cons tant height increased with the veloci ty . Henderson & Smi th ( 1 982 ) showed an increase in the sucro se concentration even at a vel ocity of 0 . 0 7 mm/s ( their maximum veloci ty) which was much lower than the velocity o f 0 . 24 mm/s reached in the present s tudy . Th is could be due to the different tower fermenter internal cons truction ( their tower was baffled) s ince they us ed a more flocculen t yeas t s train and s imilar range of cell concentration ( 20 to 80 g / 1 DW) . Prince & Barford ( 1 9 82 ) observed that the exit sucros e concentration in tower fermentation of cane j uice was init ially constant as the velocity increased and then , after a certain velocity was reached , the concentration of sucrose increased with an increase in the velocity . Thi s l imiting velocity decreased as the feed sucros e concentration increased . I t is evident that , in tower fermentation , the subs trate concentration at the bottom of the tower increas es with veloci ty while the exi t subs trate is not affected by veloci ty increase until the wash out veloci ty is reached , at which point the subs trate concentration increases wi th vel oc i ty . 4 . 3 . 2 Ethanol concentrat ion (E ) The e thanol concentration showed a general trend of decreas ing concentration with an increase in the veloci ty ( fig . 4 . 1 9 ) . However , there were exceptions to this trend . Ethanol concentration at height of 0 . 09 6 m ( curve 1 ) decreased rapidly from 20 to 6 g /1 wi th a small increase in the velocity to 0 . 0 80 mm/s and then decreas ed slowly to 3 g/1 wi th further increas e to 0 . 30 mm/ s . At greater heights between 0 . 82 and 2 . 32 m ( curves 2 to 4 ) , e thanol concentration was relatively cons tant (between 1 7 and 1 8 g / 1 ) until a veloc i ty of 0 . 24 mm/s was reached at which there was a decreas e in the ethanol concentration with veloci ty . The high ethanol concentration (between 20 to 23 g / 1 ) a t the lowes t veloci ty o f 0 . 044 mm/s was due to 20 ,..., 15 9 7 � .\'\ � ' ", ��� -o-�-"" � -::::-- ... oc:.:::: ... - - h. - 4 --....a " _::::.==-__....... .......::: :--.... -: .... > - � """'o 3 ' 'h. - " 2" OD � ....:l 0 � ::r:: E-< � 10 HE m "\ • 1 . 0 . 096 0 0 2 . - - 0 . 82 � 6. 3 . -- -- 1 . 57 � 5 ... 4 . - -- 2 . 32 0 5 . - -- 2 . 69 0 . 1 0 . 2 V , SUPERFICIAL LIQUID VELOCITY , mm/s s 0 . 3 Fig . 4 . 1 9 Ethanol concentration vs superficial l iqui d vel oci ty a t various tower heights . the greater feed lactose concentration ( 4 7 g/1 , Appendix 8 . 3 ) than for o ther veloci ties . . At the exi t ( �urve 5 ) , the ethanol concentration was affected by the fluctuat ion of the feed lac tose concentra tion for each veloc i ty . The resul t s showed that as the vel ocity increased , the maj or ethanol produc t ion occurred at heights greater than 0 . 09 6 m . For velocities up to 0 . 24 rrnn/ s , mos t ethanol was. produced within the first 0 . 82 m. The trend observed was s imilar to that ob tained in the tower fermentation of cane molasses (fi g . C . 10 b) . Prince & Barford (1 9 8 2 ) -1 showed that at velocity greater than 0 . 28 rrnn/ s (D = 0 . 50 h :) , the exi t ethanol concentra t ion decreased wi th velocity when the feed sucrose c oncentra tion was 1 00 g /1 . Coote ( 19 74 ) showed tha t at a veloc i ty 98 greater than 0 . 1 7 mm/ s there was a reduction in ethanol concentration with velocity when the feed sucros e concentration in the beet molasses was 100 g / 1 . The trend was different in the tower fermentation of fodder b eet extract (Henderson & Smith 1 9 82 ) , the ethanol concentration decreas ed as the velocity increas ed at all heigh ts in the tower including the exi t . Thi s difference was probably becaus e their tower was much short er ( Q . 7 6 m maximum compared with 2 . 69 m in the present s tudy) . Thus , their tower corresponded with the bot tom of the pres ent apparatus . Thus , in tower fermentation , e thanol concentration a t the bot tom o f the tower decreased wi th an increase in the velocity, bu•t a t . . . the tower exit the concentration decreas ed with the velocity only when the ferrnenter was operated at wash out veloci ty . 4 . 3 . 3 Rates of lactos e utilization and ethanol production The volumetric ( Si ) and specific (q1 ) ra tes o f lac tose util ization ( fig . 4 . 20 and 4 . 2 1 ) , and the volumetric ( E ' ) and specific (\)) rates of ethanol production ( fig . 4 . 22 and 4 . 23 ) , al l showed s imilar changes wi th veloci ty . At a mean height o f 0 . 048 m (curve 1 ) , al l the rates were high and increas ed with veloci ty to 0 . 24 mm/ s and then decreased as the velocity increased furth .er to 0 . 30 mm/s . The exception to this trend occurred at a veloci ty of 0 . 0 80 mrn/s at which th.ere was a reduction in the rates . This reduction was probably a result o f lower feed lactose concentration (43 g / 1 ) than for the lower veloci ty Of 0 . 044 mm/s (4 7 g / 1 ) . Measured rates for two sample sets. ob tained at this mean heights and velocity , were low thus reducing the mean value ob tained . The reduction of the rates at 0 . 30 mm/ s and 0 . 04 8 m, was due to the reduction in the cell concentration and the residence t ime caus ed by the increas e in the velocity . Thi s caused a reduction in the overall lactose utiliza tion and a reduction of the rates compared with those at the lower vel ocity of 0 . 24 mm/ s . A s imil ar trend was obs erved at a mean heigh t of 0 . 46 m (curve 2 ) . However , because lac tos.e and cell concentrations were l ower than in the f irs t s ection of the tower , these rates were lower . At mean heights of 1 . 20 and 1 . 9 5 m ( curves 3 and 4 ) , there was no change in the rates until a velocity of 0 . 24 mm/s was reached , a t which point the rates increased with further velocity increas e . The increas e in the rates at velocities greater than 0 . 24 mm/s was a resul t o f the incomplete util i zation of lactos e in the lower sections . An increase 99 in the cell concentration in these sections was observed ( fig . 4 . 24 ) and this was consis tent with th.e greater availability of lactose as the velocity increased . In the separator ( curve 5 ) , the rates were effectively zero for al l velocities up to 0 . 30 mm/s . At the wash out velocity (0 . 30 mm/ s ) , there was some lactose utilization in the s eparator (0 . 2 g/lh) and so some ethanol production occurred (0 . 3 g/lh ) . Thus , at mos t operating velocities the separator was not involved in lactos e utilization , and its sole function was s eparation of the yeas t from the ef fluent bro th and return of the yeas t cells to the ac tive column . � -00 .. 100 z 0 H � � H � 80 H � :;::l rzl Ul 0 � u 60 :s � 0 rzl � 40 u H p::: � � � 20 0 :.> .. - .-l Ul H* m • 1 . - 0 . 048 0 2 . -- 0 . 46 6. 3 . - --- 1 . 20 • 4 . --- 1 . 95 0 5 .-- 2 . 5 1 0 . 1 0 . 2 0 . 3 V , SUPERFICIAL LIQUID VELOCITY , mm/s s Fig . 4 . 20 Volumetric rates of lactose util i zation vs superficial liquid velocity at various mean tower heigh ts . MASSEY UNIVERSITY LIBRARY • 1 . 6 1 . 4 fo -eo � 1 . 2 z 0 H E-< � H � H E-< 1 . 0 ::::> i:>l Cl) 0 E-< u < � 0 . 8 J:<.. 0 i:>l � u 0 . 6 H J:<.. H u i:>l p., Cl) � 0 . 4 0" 0 . 2 H* m • 1 . - 0 . 048 0 2 . - - 0 . 46 6. 3 . - --- 1 . 2 0 -. 4 . - -- 1 . 95 D 5 . - -- 2 . 5 1 1 00 / / / / P-- 2 / ----o I I I I I . I 3 / I f I � - .,.. --..:.-"'rr-- 5 0 . 1 0 . 2 0 . 3 V , SUPERFICIAL LIQUID VELOCITY , mm/s s Fig . 4 . 2 1 Specific rate o f lactose utilization vs superficial l iquid velocity at various mean tower heights . .. ..c:: � r-i 0 -00 � � .. z 0 1-1 u tj 1-1 p::: 2 E-< � � p.. 6 ...:I > 0 .. :ij - ::z:: � E-< � 30 20 10 1 01 .,...- H* m • 1 . - 0 . 048 0 2 . - - 0 . 46 !::. 3 . - --- 1 . 20 ., 4 . - - - 1 . 95 0 5 . - -- 2 . 5 1 ---- o-_ -l_ 0--- 0 0 . 1 0 . 2 0 . 3 V , SUPERFICIAL LIQUID VELOCITY, mm/s s Fig . 4 . 22 Volumetric rate of ethanol production vs superficial l iquid veloc i ty at various mean tower heights . ..c:: 00 � - . 0 00 0 . 6 0 . 4 0 . 2 H* m • 1 . - 0 . 048 0 2 . - - 0 . 46 !::. 3 . ---- 1 . 20 "' 4 . --- 1 . 95 0 5 .-- 2 . 5 1 2 _.....o- -� 0 . 1 0 . 2 0 . 3 V 8 , SUPERFICIAL LIQUID VELOCITY , mm/s Fig . 4 . 23 Specific rate of ethanol production vs superficial l iquid velocity at various tower heights . 102 The resul ts showed that as the veloci ty was increased , more o f the of the total tower was involved with lac tos e fermentation and lessv ethanol was produced in the bot tom of the towe r . The increas e in the s pecific rates with velocity implies that the cul ture at mean heights o f 0 . 48 to 1 . 95 m was limited by the c arbon feed rat e . The resul ts also showed that for al l velocities up to 0 . 24 mm/s , there was negligib le lactos e utiliza tion and ethanol production for mean heights greater than 0 . 4 8 m because there was l i t tle lactose remaining . Thus , tower s ec tions higher than 0 . 82 m were redundan t . The trend observed here was s imilar to that found in the tower fermentation of molas ses (.C . 3 . 3 ) . However , the optimum velocity ( 0 . 33 mm/s )was greater than in the �ower fermentation o f whey permeate in which a l ower concentration ( 40 g/1 lactos e ) was used . Thus , the overall productivity on sucrose was superior . The specific rates ob served in the whey fermentation were as low as 1 /5 of thos e observed in the tower fermentation of mol ass es ( fig . C . 1 1 and C . 1 2 ) at s imilar velocities . This showed that the utilization o f lac tos e and ethano l production from lactose by K . marxianus occurred at a much s l ower rate than the utiliza­ tion of sucrose and ethanol production from sucros e by S. cerevisiae . A s ligh tly different trend was observed in the tower fermentation of fodder beet extract (Henderson & Smith 19 82) . The maximum volumetric rates of sucrose utilization and ethanol production were obs erved at the lower p art of the tower (below 0 . 095 m) , and increas ed with velocity . However , at heights greater than this the two volumetric rates were not a ffected by velocity and over the small range of velocities s tudied (0 . 04 t o 0 . 07 mm/s ) , the cell floes were s till concentrated at the bot tom of the tower . The bed was not lifted up by the low velocity us ed . Thus , mos t sucrose was utilized with the production of ethanol in the lower tower s ec tion . Henderson & Smith ( 1 982) indicated that their fermenta­ tion was limited by the medium us ed . Further comparison could be made between the specific rates o f ethanol p roduction (_v) ob tained in batch and continuous fermentation (Appendix B . 2 ) . The highes t specific rate during b atch fermentation was found to b e 1 . 4 g/gh . This was twice the maximum observed at a velocity of 0 . 24 mm/s in the tower fermentation o f whey permeate . However , thes e two ob servations were not s trictly comparable , s ince in the continuous tower fermentation the cel l concentration was considerably higher than in the batch fermentation because of cell recycle ( fig . 4 . 23 ) . Thus , in 1 0 3 the batch fermentation , the specific rate of ethanol production was l imited by the cell physiology but in the continuous tower fermentation , the specific rate was limited by the feed rate , which in turn was limited by the floc s ettling veloci ty . I t was not pos s ib le to compare the s pecific rates obs e rved here with the results o f other tower fermentation inves tigations because ins uffi­ cient data on the cell concentration were provided . However , it was reported that for a continuous s t irred-tank ( CSTR) fermentation of glucose medium ( 9 . 4 g/1 ) by S .cerevisiae , the specific rate o f e thanol production ( v) increased from 0 . 23 to 1 . 83 g/gh with an increas e in the - 1 dilution rate from 0 . 0 72 t o 0 . 42 h (Bazua & Wilke 1 9 7 7) . The biomas s concentration , in contras t , decreas ed from 1 . 5 to 0 . 69 g/ 1 DW with the same increase in the dilut ion rate at cons tant glucose feed concentration . Thus , in the tower fermentation o f whey permeate , the s pe ci fic rate of e thanol product ion increased wi th an increas e in the superf icial l iquid veloci ty at heights up to 0 . 82 m and velocity below wash out (up to 0 . 24 mm/s ) . The calculated s pecific rate , however , was l ower than for b atch and continuous s tirred-tank fermentations due to the high b iomass concentration employed to p rovide rapid convers ion o f lactose t o ethanol . Finally , the conclus ions to be drawn from these measurements are that at heigh ts up to 0 . 82 m, there was an increas e in the various rates of lactose util i zation and ethanol production with an increas e in the velocity to 0 . 24 mm/s . At heights greater than this there was no change in the rates and the rates were generally zero . Therefore , thes e sections o f the t ower were redundant for this operating rang_e as shown previously ( s ect . 4 . 1 . 1 ) . The spe ci fic rates were lower than the specific rates obs erved for fermentation systems us ing lower biomass concentration . 4 . 3 . 4 Cell concentration The effect o f the superficial liquid velocity on the cell concentra­ tion was closely rel ated to height ( fig . 4 . 24 ) . At the bottom of the tower ( 0 . 09 6 m, curve 1 ) , the concentration was relatively unaffected by an increas.e in the velocity from 0 . 044 to 0 . 1 7 mm/s . The average concentration was 9 8 g/1 DW (344 g/1 WW) . However , as the velocity increased further to 0 . 30 mm/s , i t showed a rapid reduc tion to 56 g/1 DW (220 g/1 WW) . This was an indication that the cell plug at the b ot tom of the tower was slowly being l ifted by the liquid to greater heigh ts . .--i -()() .. E-4 G H � ;3:: � � H � � .....:1 .....:1 � u .. iS .--i - ()() .. E-4 ::X:: (.!) H � ;3:: E-4 � ;3:: � � (.!) � H � E-4 � u .....:1 .....:1 � u .. � 60 40 20 100 1 1 104 . l , ---- 02 . - -- Ll.3 . --- ­ T 4 . -- - ­ t;J 5 . - · - ·- V , SUPERFICIAL LIQUID VELOCITY , mm/s s HEl'l 0 . 096 0 . 82 1 . 5 7 2 . 32 2 . 6 9 Fig . 4 . 24 Cell concentrat ion vs superficial liquid velocity at various tower heights . 105 There would be l es s cell compaction als o . Thus , the cell concentration at this location decreased as the velocity increased . At a greater heigh t o f 0 . 82 m , the trend was different . The concen­ trat ion decreased from 2 1 to 10 g/1 DW ( 9 2 to 55 g/1 WW) as the velocity increased from 0 . 044 to 0 . 080 mm/s and then increas ed to a peak of 36 g/1 DW ( 1 40 g/1 DW) as the velocity increased furth er to 0 . 24 mm/s . There was further decrease in the concentration to 24 g/1 DW ( 106 g/1 WW) with a further velocity increas e to 0 . 30 mm/s . A s imilar trend was obs erved at heights of 1 . 5 7 and 2 . 32 m ( curves 3 and 4 ) . As the velocity increased , the concentration of ce lls in the upper parts o f the tower increas ed , indicating f luidization and washout o f the b iomass . At 0 . 24 mm/s , the cel l concentration was uniform between 0 . 82 and 2 . 32 m. At 0 . 30 mm/s , washout was apparen t . The reduction o f the cell concentration as the velocity increas ed from 0 . 044 to 0 . 080 mm/s , may be due to di fferent cell compaction . At the lowes t velocity , there could be a greater cell compact ion between 0 . 82 m and 2 . 32 m, b ecause the liquid velocity was not high enough . Thus , the cell concentration was greater than at the higher veloci ty (0 . 080 mm/s ) . The increas e in the cell concentration at veloci ties b etween 0 . 1 7 and 0 . 30 mm/s for heigh ts be tween 0 . 82 and 2 . 32 m, corres­ ponded with the reduction in the cell concentration at the bot tom of the tower ( 0 . 096 m) . Finally , at the exi t of the tower after the effluent left the s eparator , the concentration was practically cons tant at b etween 0 . 2 and 0 . 5 g/1 DW (10 and 1 8 g /1 WW) as the velocity increas ed from 0 . 04 to 0 . 08 mm/s . This in effect showed that th.e exit cel l concentration was unaffected by the velocity even at the washout velocity . The' res ul ts indicated that the separator was functioning very effectively for the yeas t KM Y42 . If the veloci ty were greater than 0 . 30 mm/ s , an increas e in the exit cell concentration woul d be expected . A different trend was ob tained in the tower fermentation of cane j uice ( Prince & B arford 1 982 ) , the concentration ( exact s ampling heigh t was not indicated) increased from 67 t o 83 g/1 DW with an increase in the velocity from 0 . 1 4 to 0 . 54 mm/ s ( D 0 . 25 to 0 . 9 5 h- 1 ) when the feed s ugar concentration was 100 g/1 . This difference was probab ly due to c arbon limitation at lower ve�ocities . However , s ince the sampl ing l ocation was not indicated , it is difficul t to make further speculation on the cause of this difference . In s ummary , in the tower fermentation o f whey permeate , the cell concentration at the b ot tom of the tower ( 0 . 096 m) was not affected by 106 the increase in the velocity until a velocity o f 0 . 1 7 mm/s was reached , af ter which the concentration decreased with an increas e in the velocity . At greater heigh ts b etween 0 . 82 and 2 . 32 m , the concentration increased , with an increase in the velocity , to a peak a t 0 . 24 mm/ s and then decreas ed with further velocity increas e and washout occurred . 4 . 3 . 5 Specific growth rate ( �KM) Estimates of specific growth rate are useful for purpos es o f comparison o f different proces ses . The mean specific growth rate was calculated using the exit cell concentration b ased on the average cell concentration in the tower up to 2 . 32 m and the residence t ime at this heigh t . � X ( X T ) g/gh e a r where X the exit cell concentration , g/1 DW . e X the mean cell concentration between 0 and 2 . 32 m . a T the res idence time at 2 . 32 m . r The speci fic growth rate increas ed from 3 . 9xl0-4to 6 . 3xl0-3 g/gh as the velocity increas ed from 0 . 044 to 0 . 3 3 rnrn/s ( table 4 . 2 ) . Table 4 . 2 Mean specific growth rate at various s uperfici al l iquid velocities . V , mm/ s 0 . 044 0 . 0 80 0 . 1 7 0 . 24 0 . 30 s � KM ' X a , g /gh 3 . 88xl0 -4 8 . 9 6xl0 -4 1 . 2l xl0 -3 . -3 3 . 2lxl0 6 . 25xl0 g/ 1 DW 32 . 4 25 . 7 4 0 . 4 4 1 . 6 34 . 3 X = the es t imated average cell concentration in the tower a calculated from concentrations at various heights . -3 The increase in specific growth rate with velocity was due to the reduction in the res idence time as the velocity increas ed to 0 . 1 7 mm/s s ince the exit cell concentration was cons tant at 0 . 2 g/1 DW . At greater veloc ities than thi s , there was an increase in the exit cell concentration with velocity (0 . 4 and 0 . 5 g/1 DW) . This increase would also contribute to the increas e in the specific rate with the velocity since there was little change in the average cel l concentrat ion through­ out this velocity increase ( s ect . 4 . 3 . 4 ) . The results showed that the overall s pe ci fic growth rate was low because o f the high cell concentration resulting from b iomass feedback . The values ob tained here were lower than in the tower fermentation of 1 0 7 molasses ( 4 . 4xl0-3 t o 0 . 059 g/gh f rom Appendix 13 . 5 ) but showed a s imilar trend of increas e with velocity . This was b e cause the exit cel l concen­ tration was higher in the molass es fermentation . Thus , there was a low mean: specific growth rate in the tower fermen­ tation o f whey permeate as a resul t o f the high yeas t cell concentration used, and the employment of the yeas t s eparator as an integral part of the fermenter to recycle the yeas t cell which res ulted in a low, exit cell concentration . In the tower fermentation of whey permeate the dilution rate was not equal to the specific growth rate . 4 . 4 TOWER FERMENTER PERFORMANCE It was shown that in the tower fermentation o f whey permeate , there was 90% l ac tose utilization and greater than 16 g/1 of ethanol was produced over the f irs t 0 . 82 m of the h eight in the tower for al l velocities up to 0 . 24 mm/s ( sect . 4 . 1 . 1 and 4 . 1 . 4 ) . This was s trongly reflected by high rates of lactose utilization and e thanol production within the f irs t 0 . 82 m of the tower ( s e ct . 4 . 1 . 2 , 4 . 1 . 3 , 4 . 1 . 6 , and 4 . 1 . 7) . The rates were negligibl e at greater heights . It was als o shown that for all velo cities us ed , there was greater than 90% lactose util ization in 1 hour ( s e ct . 4 . 2 . 1 ) . This was als o the t ime during which more than 1 6 g/1 of ethanol was produced ( s e ct . 4 . 2 . 2 ) . There were high rates of lactose utilization and e thanol production during the firs t 2 hours in the tower and th.e rates were negligible at res idence t imes greater than this (.s e ct . 4 . 2 . 3) . This reflected the rapid rates of lactose utilization and e thanol production in this tower fermenter . The res ulting yield of e thanol based on lac tose utilized was betweeen 70 and 96% at heights b e tween 0 . 82 and the exit for al l velocities ( se ct . 4 . 1 5 ) . The cell concentration which gave this rapid lacto s e utilization and ethanol production was between 66 and 100 g / 1 DW ( 244 and 348 g/1 WW) at a heigh t of 0 . 096 m for all velocities up to 0 . 24 mm/ s ( s ect . 4 . 1 . 8) . 4 . 4 . 1 Optimum superficial l iquid veloc ity (V ) . s The optimum velocity for the tower fermentation of whey permeate was 0 . 24 mm/s . This was lower than th.e optimum velocity reached by other tower fermentation inves tigations with one exception ( table 4 . 3 ) . 1 08 Table 4 . 3 Comparison o f optJmum superficial liquid veloci ties from various tower fermentation s tudies . medium 81o V D references s - 1 g/1 mm/s h l .Whey permeate 40 0 . 24 1 . 0 This s tudy 2 . Cane molas s es 100 0 . 33 1 . 4 Appendix C , s ec t . C . 4 3 . Beet molasses 100 0 . 53 1 . 1 Coote 1 9 74 4 . Cane j uice 100 0 . 34 0 . 60 Prince & Barford 1 982 S . Fodder b ee t extract 100 0 . 05 0 . 24 Henderson & Smi th 1 9 82 6 . Beer wart 58 0 . 57 - Aul t et al 1 969 S 10 - feed sugar concentration 3 . Coat ' s value was calculated from l imiting volumetric efficiency of 4 . 7 . In this s tudy , the optimum veloci ty was lower than that obs erved by mos t o ther workers b ecause the yeas ts used by the other workers were more flocculen t . I t was expected that i f a more flocculent K. marxianus were avail able the optimum veloci ty could be as high as the veloci ty obtained in the o ther inves tigations s ince whey permeate is less dens e than the more concentrated sugar solutions us ed by the o ther workers . I t was shown that the maximum operating superficial liquid velocity for a particul ar flocculent yeas t s train decreased as the medium concentra­ t ion increased (Prince & , Barford 1 982 ; Coote 1 9 74 ) . Thus , the lower density whey could permit operation at a higher velocity if the yeas t were more flocculent . Of course , the s pecific rate of ethanol production would remain lower . I t i s worth noting that the us e of the parameter "superfic ial l iquid veloci ty" (V ) t o define tower fermenter perfo rmance is more suitab le s than use of "dilution rate" (D ) . In table 4 . 3 , the values of D f or the firs t three tower fermenta tion s tudies were h igh and were 1 h- l or greater . This was because D was calculated based on the tower height up to the bot tom of the separator only . The superficial liquid velocity, on the o ther hand , is not affected by the change in the tower heigh t or the overall volume . It is rel ated direc tly to the fermenting medium input and output ra te . Thus , it is a more realis tic paramet er for use to describe the performance of a tower fermenter . 1 09 4 . 4 . 2 Res idence time and tower height The minimum tower heigh t of 0 . 82 m with respect to the veloci ty o f 0 . 24 mm/s giving a res idence time of 1 hour (excluding the t ime in the separator ) was comparable to o ther tower fennentation values ( t able 4 . 4 ) . Tab l e 4 . 4 Comparison of the effec tive tower heigh ts and residence t ime o f various tower fermentation s tudies . HE

U) � E-< 0 E-< � � U) 60 40 20 40 20 1 00 80 60 40 20 (a) S 8 , SUCROSE (c) S t ' TOTAL SUGAR \ � ' 1 2 1 V mm/s Q ml/h s • 0 . 087 1 6 0 ·-- - 0 . 1 2 220 ·--- 0 . 14 260 � 1 ,, , �J.:!2 �-;::--_ t. 2 0 . 5 0 . 14 �-- -- -- - - .,__ . -- - -....a_ . 087 ..- - -· 3 1 . 0 1 . 5 2 . 0 2 . 5 TOWER HEIGHT , m Fig . 5 . 1 Sugar concentration vs tower. height at various superficial liquid velocities . (a) Sucrose , (b) Lactose , ( c ) Total sugar . 1 22 furthe r to 2 . 32 m . The decrease of sucrose concentra tion wi th height was slower for the two greater veloci ties ( 0 . 1 2 and 0 . 1 4 mm/s ) over the ini tial 0 . 096 m , where the concentrations were 22 and 23 g/1 at the respec tive veloci ties . Table 5 . 1 Percentage sugar utilization at various tower heights and superficial liquid velocities V8 , mm/s 0 . 087 0 . 1 2 0 . 1 4 location , m ssu slu stu ssu slu stu s su slu stu 1 0 . 096 78 1 2 5 1 62 1 37 62 4 37 2 0 . 82 99 29 7 1 96 3 58 97 1 6 62 3 1 . 57 99 68 87 98 1 7 64 98 3 1 70 4 2 . 32 99 77 90 98 30 70 98 47 76 5 2 . 69 99 80 9 1 98 33 7 1 98 49 77 Csu - Sucrose utili zation , % SJ..u - Lactose utili zation , % stu - Total sugar utilization , % However , at 0 . 82 m , the sucrose concentration was reduced to a s imilar value as for the lowest veloci ty and showed no further reduction with increasing height . This difference in sucrose concentration at 0 . 096 m was attributed to an increase in the liquid veloci ty , wi th the resulting decrease in the residence time at a particular tower height , as well as the reduction in the total cell weight ( fig . 5 . 8 ) . A similar rapid reduction of the sugar concentration was obtained in the tower fermentations of whey permeate ( fig . 4 . 1 ) and molasses ( fig . C . 1 a) at the same low superficial veloci ties . In all cases , the sugars were completely utili zed in the firs t 0 . 82 m height of the towe r . ( b ) Lactose Cs.t ) Although there were minor variations from one super ficial velocity to another , there was generally only a small decrease in lac tose concentration up t o a height of 0 . 82 m , after which a mode rate decrease occurred until a height of 2 . 32 m was reached ( fig . 5 . 1 b ) . At the lowest velocity , there was a moderate lactose utili zation 1 23 over the first 0 . 82 m , with a reduction in concentration from 40 to 34 g/1 . A significant reduction occurred over the next 0 . 75 m with lac tose concentration being 1 1 g/ 1 at 1 . 57 m. A near linear decrease occurred over the remaining height of the tower . The lac tose concentration profiles wi th heights were nearly identical for the two higher velocities ( 0 . 1 2 and 0 . 1 4 mm/s ) wi th one exception . For both velocities , there were small decreases in concent ration over the firs t 0 . 82 m height where the concentration was 39 and 40 g/ 1 , respec tively . The small reduction in lac tose concentration below 0 . 82 m , was att ributed to the diauxic growth behaviour of K. marxianus Y42 . The presence of sucrose in the medium up to 0 . 82 m would inhibit lac tose utilization through glucose repression of the enzyme 6 -galactosidase . The smal l amount of lac tose utilized could possibly be attributed to a diffusion effect visual observation of the tower ind icated that for all three veloci ties the tower did not behave as a fluidi zed bed but as a packed bed and channel ling of the liquid medium was observed ( fig . 5 . 9 ) . Sucrose could have been rapidly utilized by the yeast cells close to the liquid channels and as the medium diffused into the inner layers of the packed bed , only lactose remained in the medium . The reduction within the ini tial tower sec tions was slightly greater at the lowest velocity because there was a greater cell concentration ( table 5 . 2 ) and the fermentation time was longer as a result of the lower veloc i ty . Sucrose was completely utili zed below 0 . 82 m at this low veloc i ty . Thus , lactose utilization could occur below 0 . 82 m at the lowest velocity . This would explain the greater reduction in lac tose concentration at 0 . 087 mm/s than at higher veloci ties . Comparison of the estimated cell weight of K. mar xianus ( table 5 . 2 ) within the tower up to 0 . 82 m with those in the tower fermentation of whey permeate ( without enrichment ) at similar veloci ties , showed that the cell d ried weight ( 57 g/1 ) at 0 . 087 mm/s was more than double the cel l weight ( 25 g/1 ) at a similar velocity ( 0 . 080 mm/s ) in the whey permeat e fermentation . For the two higher veloci ties , the cell concentration was comparable with those in the whey permeate fermen tation , but this did not result in a comparable lac tose utilization , since the fermentation of lactose in the mixed sugar fermentation was dependent on the absence of sucrose and probably sugar d iffusion to the inside of the yeast floes . 1 24 Table 5 . 2 Comparison of lac tose utilization with that of the tower fermentation of whey permeate with respect to the cell dried weight K. marxian uswithin tower sections above and below 0 . 82 m . Whey permeate + molasses Whey permeate V , mm/s s 0 . 087 0 . 1 2 0 . 1 4 0 . 080 0 . 1 7 1 . Tower heights between 0 and 0 . 82 m X KM , g DW I 57 24 35 25 28 s , % 29 3 1 6 91 90 lu 2 . Tower heights between 0 . 82 and 2 . 32 m X KM' g DW 77 59 47 s , % 48 27 31 lu X KM - Estimated K. m arxianus cell dried weight from table 5. 4 . Slu - Percentage o f lac tose utilized from table 5 . 1 . The slow reduction in lac tose concentration at heights greater than 0 . 82 m and the incomplete lac tose consumption at the exi t for all veloci t ies were probably a result of ethanol inhibition of the cell growth . I t has been reported that an ethanol concentration greater than 30 g/1 could reduce the activity of the enzyme s -galactosidase obtained from K. marxian us ( NRRL Y 1 1 09 ) by 70% ( Wendorf et al 1 97 1 b) and similar inhibition was observed for another strain of K. marxianus (UCD FST 71 58 ) in batch culture ( sec t . 7 . 3 ) . As the ethanol concentration in the tower was greater than 25 g/1 for all velocities ( fig . 5 . 3 ) i t was likely that ethanol inhibition took place . This was supported by the estimated concentration of K. marxian us present within these heights the estimated total K. marxian us cell weight was between 2 - 3 times the concentration that was required to .achieve 90% lac tose utili zation in the firs t 0 . 82 m in the tower fermentation of whey permeate at similar velocities ( table 5 . 2 ) . These observations should be interpreted wi th caution , since the cell weights of K. marxian us estimated in mixed culture were subj ect to considerable uncertainty ( 1 2 - 460% ) see sect . E . 6 . 4 . As a result o f the d iauxic behaviour o f the yeast and probably ethanol inhibition , the profiles observed were different from those in 1 25 the tower fermentation of whey permeate in which lactose concentration was reduced to 4 g/l wi thin the firs t 0 . 82 m and showed l it tle change with further height increas e . ( c ) Total sugar ( S t )The total sugar curve i s the sum of the sucrose and lactose curves ( fig . 5 . 1 c ) . Thus , there was an incomplete sugar util i zation due to incomplete lactose utili zation . The trend observed here was similar to that observed for tower fermentation of beer wort which contained mixed sugars . Such sugars as sucrose , g lucose and fructose were all uti l ized in the lowe r regions (0 to 1 . 4 m ) of the tower while maltose and maltotriose required a longer time to be uti lized . Thus , they remained in the tower up to a greater height ( between 1 . 4 and 8 . 4 m ) in a similar manner to lac tose ( Aul t et al 1 969 ) . In summary , in the mixed sugar fermentation of lac tose and sucrose , sucrose was utilized within the first 0 . 82 m of the tower , whi le lac tose uptake was repressed due to the diauxic behaviour o f K . marxianus and a small amount o f lactose was utilized . A t heights greater than 0 . 82 m, there was only slow uptake of lactose , due probably to ethanol inhibition . 5 . 1 . 2 Volumetric rate of sugar uti l i zation (S ' ) All volumetric and specific rates given here were calculated as mean values between subsequent sampling points , and the mean heights are the midpoint between these sampling points ( table B . 4 ) . (a ) Sucrose ( s; ) The volumetric rate of sucrose utilization decreased rapidly from a high value between 1 48 and 208 g/lh to less than 1 g/ lh as the mean height increased from 0 . 048 to 1 . 20 m for all velocities ( fig . 5 . 2 a ) . It was less than 1 g/lh over the remaining mean heights . Thi s showed that sucrose utilization occurred rapidly up to 0 . 46 m as a result of high cell and sucrose concentrat ions , and both yeast species consumed suc rose . The trend of change in the rate as the mean height increased was similar to those observed in the tower fermentation of whey permeate ( fig. 4 . 2 ) and molasses ( fig . C . 2 a ) using a similar velocity . The rate reduced to a negligible value as the height increased to 1 . 20 m . (b ) Lactose (S l ) The rate was low for a l l velocities studied and showed some fluc tuations with heights ( fig . 5 . 2 b ) . The fluc tuations 200 1 80 160 20 1 5 ..c: 10 .-t -bO .. 5 z 0 H E-< < N H t-l H 1 5 E-< p � 10 p Cl.l li. 5 0 t: .. 1 80 - Cl.l 160 20 15 10 5 (a ) S ' s ' SUCROSE \ \ \ �\ ',-1 0 . 14 \ \ \ s� , ' \ '� .... ' o .?2'� LACTOSE TOTAL SUGAR ' �4 . 5 ' . 1 2 " .. ... ' 1 . 0 1 2 6 V mm/s s • 0 . 087 ·--- 0 . 1 2 • --- 0 . 14 MEAN TOWER HEIGHT , m Q ml/h 160 220 260 Fig . 5 . 2 Volumetric rates of sugar utilization vs mean tower height at various superficial liquid velocities . (a) Sucrose , (b ) Lactose , ( c ) Total sugar . 1 27 were due to diauxic behaviour and hence the p resence of sucrose in the lower region , and possibly to ethanol inhibition of 6 -galactosidase in the upper regions . Some redistribution of the biomass as velocity increase may also have had some effect upon the rates of uti li zation . The trends obtained were different from those obtained in the tower fermentation of whey permeate , whe re the init ial rate was much higher and fel l to less than 1 g/lh above 0 . 46 m when sugar was almost exhausted ( fig . 4 . 2 ) . ( c ) Total sugar ( St ) The volumetric rate of total sugar uti lization at a mean height of 0 . 048 m was between 1 64 and 2 1 7 g/lh . Sugar utilization occurred throughout the tower , but the rates were low at mean heights above 0 . 46 m . The greatest rate achieved ( 2 1 7 g/lh) was lower than that obtained for molasses fermentation ( 350 g/lh at 0 . 1 7 mm/s ) using the same concentration of sugar ( 1 00 g/1 ) in the feed . 5 . 1 . 3 Specific rates of sugar utilization ( q ) ( a ) Sucrose ( q ) The specific rate o f sucrose utilization was s calculated using the total cell weight since both yeast species utilize sucrose . The specific rat e decreased wi th an increase in the mean tower height ( fig . 5 . 3 a ) . At the lowest velocity ( 0 . 087 mm/s ) , the specific rate decreased rapidly from 0 . 43 to 0 . 03 g/gh as the mean height increased from 0 . 048 to 0 . 46 m and then decreased to zero as the mean height increased further to 1 . 20 m . It remained at this level over the remaining tower sections . Similar profiles with higher specific rates were observed at the two higher veloci ties ( 0 . 1 2 and 0 . 1 4 mm/ s ) . The results showed that the yeast 'cel ls were very active in utilizing sucrose . At the mean height of 1 . 20 m upward , there was no sucrose utilization since all sucrose was used up below 0 . 82 m ( sect . 5 . 1 . 1 a ) . The trend observed was similar to those in the tower fermentation of molasses ( sec t . C . 1 . 2 b ) at s imilar velocities . ( b ) Lac tose (� ) The specific rate of lactose utilization was low for all velocities s tudied and showed some fluctuations with height ( fig . 5 . 3 b) . The irregular t rend observed was considered to be due to the different concentrations of K. marxianus at various heights in the tower ( table 5 . 4 ) . At all veloci ties , there was the e ffect o f ethanol � 00 -00 .. z 0 H � N H ....:1 H E-< :::::> � :::::> tf) J:1< 0 � E-< � u H J:1< H u � p.. tf) .. 0' 1 . 6 q , SUCROSE s 1 . 2 \ , . 14 0 . 8 \ ' 1\ \ \ 0 . 4 0 .. 1 2\\ \ \ 0 . 4 (b ) ql ' LACTOSE 0 . 3 0 . 2 0 . 1 1 . 6 TOTAL SUGAR 1 . 4 1 . 2 1 . 0 0 . 2 0 . 1 1 28 4 V mm./s s • 0 . 08 7 ·- -- 0 . 12 ·--- 0 . 14 ,_...A, , " � 0 . 1 2 ' MEAN TOWER HEIGHT , m Q ml/h 160 220 2 60 Fig . 5 . 3 Specific rate of sugar utilization vs mean tower height at various superficial liquid velocities . 1 29 concent ration on the specific rate of lac to se uti li zation between 0 . 82 and 2 . 32 m . Within these regions , the ethanol concentration was high (between 26 and 43 g/1 ) ( fig . 5 . 4 ) , and would contribute to lowering the specific rate of lac tose utilization ( Wendorf et al 1 971 a ) . Thu s , K. marxian us cells were active throughout the towe r . However , d iffering mean cell concentrations at each mean height led to different specific rates throughout the tower . The results obtained here were different to the trend observed in the tower fermentation of whey permeate , which showed a reduction of the spec ific rate as the height increased ( sec t . 4 . 1 . 3 ) . Here the rates generally decreased with height , but there was significant lac tose utilization throughout the ent ire length of the tower . ( c ) To tal sugar ( q ) The resulting mean specific rate o f total t sugar utilization ( fig . 5 . 3 c ) showed rapid decrease from between 1 . 03 and 1 . 75 g/gh to between 0 . 05 and 0 . 1 4 g/gh as the mean height increased from 0 . 048 to 0 . 46 m . The speci fic rate was between 0 . 0 1 and 0 . 05 g/gh over the remaining mean height increase to 1 . 95 m and was zero in the separator . These trends were similar to those of the volumetric rate of total sugar utilization described previously ( se ct . 5 . 1 . 2 c ) . However , the high cel l concentration used in this mixed culture fermentation gave a very low specific rate between 0 . 46 and 1 . 95 m and only K. marxian us was invo lved in the fermentation . Thus , the specific rates observed in this fermentation were lower than those observed in the tower fermentation of molasses ( as high as 1 0 . 8 g/gh) ( sect . C . 1 . 2 b ) . 5 . 1 . 4 Ethanol concentration ( E ) Ethanol concentration ( fig . 5 . 4 ) increased rapidly a s the height increased to 0 . 82 m and as the height increased further to 2 . 32 m , the concentration increased slowly . The concentration increased to 36 g/1 over the first 0 . 82 m of the tower height at the lowest velocity ( 0 . 087 mm/s ) . It then increased slowly to 43 g/1 as the height increased further to 2 . 32 m . The concentration was 4 1 g/1 as the effluent liquor left the tower at the exit . A similar profile was observed for the two higher velocities ( 0 . 1 2 and 0 . 1 4 mm/s ) but the concentration was lower at all heights . The effluent ethanol concentrations reached were 33 and 34 g/1 , respect ively . 40 r-l ';;o 30 / � , V 1 0 0 . 5 1 30 ___ .. 1-- - t ... -- - ­ ..__ --::-: - ---� 0 . 1 2 , 0 . 1 4 V mm/s s Q ml /h • 0 . 087 1 60 ·-- 0 . 1 2 220 ·-- o ; 1 4 260 2 3 4 5 1 . 0 1 . 5 2 . 0 2 . 5 TOWER HEIGHT , m Fig . 5 . 4 Ethanol concentration vs tower height at various superfic ial l iquid veloci t ie s . The ini tial increase in the concent ration wi th increase in height up to 0 . 82 m was mainly a result of suc rose utili zation . As the tower height increase further to the tower exi t there was a slow utilization of lac tose , resulting in the slow increase in the e thanol concent ration . The smaller increases in the ethanol concentration wi th height at the two higher velocities was a result of incomplete lac tose utilization . This was discussed previously ( sect . 5 . 1 . 1 ) . The trend observed was similar to those observed a t s imilar veloci t ies in the tower fermentation of whey permeate ( fig . 4 . 4 ) and molasses ( fig . C . 1 b ) , in that there was a rapid initial increase in ethanol concentration . The re was a difference at heights between 0 . 82 and 2 . 32 m in which there was a s low increase in the e thanol concentration in the mixed sugar fermentation , but this did not occur in the two pure sugar fermentations , since most ethanol was produced in the lower regions of the tower . 1 3 1 5 . 1 . 5 Ethanol yield ( Y ) Ethanol yield increased wi th increasing height i n the tower ( fig . 5 . 5 ) . The yield was lower at greater superficial liquid In the lower 0 . 096 m of the velocities and ranged between 74 and 97% . tower , the yield at all veloci ties was low . ,....._ "0 r-i Cl) "M >-. r-i Cl! (.) "M +.J Cl) H 0 Cl) ,.c:: +.J � 0 � ..._, � ...:I 0 � ::r:: E-< I'Ll � ;>; 90 80 / 70 / V mm/s Q ml /h s 60 2 3 0 . 5 1 . 0 1 . 5 TOWER HEIGHT , m • 0 . 087 1 60 • --- 0 . 1 2 2 2 0 • --- 0. 14 260 2 . 0 2 . 5 5 Fig . 5 . 5 Ethanol yield vs tower height at various superficial l iquid velocities . The results observed were similar to those in the tower fermentation of whey permeate ( fig . 4 . 5 ) in that the yield was lowe r at 0 . 096 m and then increased , at 0 . 82 m , to · a level where it showed little change within analytical error as the height increased to 2 . 32 m and at the exi t . It was explained previously that the lower yield at 0 . 096 m was affec ted by analytical unce rtainty more than at the o ther heights , because of high sugar concentration and concentration ( sect . 4 . 1 . 5 ) . 5 . 1 . 6 Volumetric rate of ethanol production ( E ' ) low ethanol The volumetric ra te of ethanol production ( E ' ) ( fig . 5 . 6 ) fo llowed a similar trend to that observed for the volumetric rate of total sugar uti lization ( fig . 5 . 2 c ) . The rate decreased rapidly from 57 to 2 g/ lh ,..c:: M - OD � z 0 H H u ;::::> p 0 p:: P-o ....:l 0 � ::X:: H � � 0 � H � u H p:: H � ....:l 0 ::> - � 1 32 • V rnm/s Q ml /h � s • 0 . 087 1 60 65 ·- -- 0 . 12 220 ·---0 . 14 260 ,\ \\ 60 \ \ ' \ A " 55 \ \ 10 5 ::.:=-.:-� ...... .. 0 . 5 1 . 0 1 . 5 2 . 0 2 . 5 MEAN TOWER HEIGHT , m Fig . 5 . 6 Volumetric rate of ethanol production vs mean tower height at various superficial liquid velocities . as the mean height increased from 0 . 048 to 0 . 46 m at the lowest veloci ty ( 0 . 087 mm/s ) . I t then decreased slowly to 0 . 7 · g/lh as the mean height reached 1 . 95 m . A similar trend was observed for the two higher veloc ities . Inside the separator ( 2 . 5 1 m ) , the rate was zero for all velocities . This showed that there was ethanol production throughout the tower , though the volumetric rate of ethanol production was highest in the lower section of the tower ( less than 0 . 82 m) due to high sucrose utilization . Most of the ethanol was produced within this section . The rate was lower in the upper sec tion of the tower ( greater than 0 . 82 m ) because of the low rate of lac tose utilization . The ini t ial trend ( up to 0 . 82 m) was similar to the trend observed for similar veloci t ies in the tower fermentation of whey permeate ( fig . 4 . 6 ) and molasses ( fig . C . 2 b ) in which there was a rapid reduction in the rate 1 33 to less than 1 g/ lh within the same tower sec tions . 5 . 1 . 7 Specific rate of ethanol production (v ) The spec ific rate of ethanol production decreased to zero as the mean height increased ( table 5 . 3 ) . Table 5 . 3 Specific rate of ethanol production (v ) ( g/gh ) at various mean tower height and superficial liquid velocities . V \) (g/gh) at s various mean height (m ) mm/s 0 . 048 0 . 46 1 . 20 1 . 95 2 . 5 1 0 .087 0. 4 0 . 1 0 0 0 0 . 1 2 0 . 5 0 0 0 0 0 . 1 4 0 . 5 0 . 1 0 0 0 At the lowest velocity , the specific rate decreased from 0 . 4 to 0 . 1 g/gh as the mean height increased from 0 . 048 to 0 . 46 m and then decreased to zero as the mean height increased further to 1 . 20 m and remained at this level up to the exit . Similar trends were observed at the two higher veloci ties . ( 2 . 5 1 m ) for all velocit ies . The rate was zero inside the separator This showed that there was a high specific rate of e thanol production at the bo ttom of the tower for all velocities studied . It was shown previously that ethanol was produced throughout the tower ( sect . 5 . 1 . 5 ) , however , the specific rate was effectively zero at mean heights greater than 0 . 46 m because the amount of ethanol produced wi thin these heights ( between 1 and 5 g/l ) was very smal l compared to the total cell weight ( between 58 and 1 85 g/l DW ) . It should be noted that the specific rate was calculated based on the total cell weight , but at heights greater than 0 . 82 m only lac tose was being utilized by K. marxianus since sucrose was all used up ( sect . 5 . 3 . 1 a ) . Thus S. c eP evisiae cells were no t involved in ethanol production but were includ ed in the calculation of the specific rate . 1 34 5 . 1 . 8 Cell conc entration (a) K. m arxianus concentration (XKM) The number of K. marxianus cells in the tower between 0 . 096 to 2 . 32 m was high ( between 3 - 5x 1 08and 2 . 4x1 09cells/ml ) and was very low ( between 7 . 2x1 05and 1 . 1 x 1 0 7 cells/ml ) in the effluent liquor ( fig . 5 . 7 a ) . Similar profiles were obtained at all veloc i ties studied . At the lowest velocity , the cell number decreased from 2 . 4x 1 09to 6 . 5x1 08cells/ml as the height increased from 0 . 096 to 2 . 32 m . This resulted in a high log mean cell number of 1 . 2x1 09cells/ml and gave an overall lactose utilization of 80% . The corresponding cel l dried weight at this velocity was estimated from fig . B . 1 relating plate count to dried weigh t , and was found to be higher than the cell concentration ( at 0 . 080 mm/s ) measured in the tower fermentation of whey permeate ( table 5 - 4 ) . The estimated cell dried weights ( table 5 - 4 ) showed some discrepancies when compared with the cel l number ( table B . 7 ) . I t was , however , the only method available to determine the K. marxianus cell dried weight from the cell plate count number and these estimated weights were essential for comparis on with the cell dried weight fermentation of whey permeate . obtained in the tower Considering the next veloci ty of 0 . 1 2 mm/s , the cell number changed little as the tower height increased from 0 . 096 to 2 . 32 m and the log mean cell number within these heights was 7 . 6x1 08cells/ml . At the highest velocity ( 0 . 1 4 mm/s ) , the cell number was relatively constant throughout the tower with the exception of the cell number at 0 . 82 m ( 2 . 3x 1 09 cells/ml ) . The log mean cell number was 5 . 3x 1 08 cells/ml . In the separator , the yeast cells . were mainly K. m arxianus numbering between 7 . 2x 1 05and 1 . 1 x 1 0 7cells/ml . There was between 2 . 2x 1 05 and 2 . 6x 1 06 cells/ml of S. c erevis iae . This predominance of K. marxianus was confirmed by observations during continuous operation that in the separator the yeast floes were small and suspended i e . typical of K . marxianus floes . They were different to the glutinous and very flocculent yeast cells inside the tower . It was also observed , during operation at the two higher velocities , that the concentrated yeast plug had expanded to the bo ttom of the separator . � § ·� N (::l E: � � � :><; .--i s - .--i .--i Q) tJ � � w � H H w u ,....... 00 0 .--i ..._, � (::l ·� CO ·� :::> � N � tl � � u tl.l >< 2 (a ) V mrn/s Q ml /h s 11 0 . 087 1 60 .&---- 0 . 1 2 220 ·--- 0 . 14 260 1 35 3 0 . 08 7 ""­.... .. - - -- - - -A- - -_-_:..:::::--- •�:. -- 0 . 11_� - - - . � 0 . 1 2 ... ___ (b ) xsc \ \\ 0 . 087 \\, \ 2 3 0 . 5 1 . 0 1 . 5 2 . 0 2 . 5 TOWER HEIGHT , m 5 \ • 5 Fig . 5 . 7 Cell numbers vs tower height at various superficial liquid velocities . ( a) K. marxianus , (b) S. cerevisiae . The corresponding estimated cell concentrations are given in table 5 . 4 . 1 36 Table 5 . 4 The concentration of K. mar xianusY42 and S. cer evisia e CFCC39 at various heights in the tower and superficial liquid velocities . whey permeate + molasses whey permeate* V , mm/s 0 . 087 0. 1 2 0 . 1 4 0 . 080 0 . 1 7 s sample height cell concentration , g/ l DW point m XKM x sc XKM x sc XK.M x sc X K.M x s c 1 0 .096 1 70 0 66 70 23 99 1 0 1 96 2 0 . 82 98 66 47 95 1 62 1 7 1 0 30 3 1 . 57 81 94 62 1 1 2 28 93 6 26 4 2 .32 5 1 7 65 9 29 85 4 22 5 2 .69 1 0 . 1 0 . 1 o . 1 o . 5 0 . 1 0 . 2 0 . 2 � - K. marxianus concentration was estimated by using fig . B . 1 to convert the K. mar xianusplate count number ( Appendix B . 4 ) to cell dried weight . �C - S. cer evisiae concentration was the difference between the measured total cell dried weight ( Appendix B . 4 ) and XKM above . * - The K. mar xianuscel l dried weights obtained in the tower ferment­ ation of whey permeate ( Appendix B . 3 ) . These are for comparison . The results showed that a high concentration of K. marxianus throughout the tower was essential for improved lac tose utilization in mixed sucrose and lac tose medium by a mixed cul ture of K. marxianus and S. cer evisiae. This was because at the lowest velocity ( 0 . 087 mm/s ) , the K. marxianus concentration was high throughout the tower and this resulted in lac tose being uti lized throughout the tower ( table 5 . 4 ) . However , at a higher veloc ity of 0 . 1 2 mm/s , the concentration was moderate throughout the tower but was not high enough in the first 0 .82 m of the tower to give a noticeable util ization of lac tose within these heights . So there was an uneven reduction of the lac tose concentration and when the effluent liquor left the tower only 33% of lac tose was utili zed ( table 5 . 4 ) . Finally , at the highest velocity (0 . 1 4 mm/s ) , there was a moderate concentration of K. mar xianus inside the tower and an extra high concentration at 0 . 82 m . This high concentration had enough impact on lac tose utilization to give 49% lactose uti l ization at the exit even though the cell concentration at 1 37 o ther heights was lower than at the lower velocity of 0 . 1 2 mm/s . I f the cell concentration were highe r , i t is l ikely that more lac tose could be utilized . The results also showed that K . marx�anus occupied the separato r and formed the majority of the biomass leaving the tower . (b ) S . a erevisiae (X5� The number of S. cerevisiae cells inside the tower was high ( be tween 6 . 5x 1 08and 6 . 2x 1 09 cells/ml ) and was very low in the effluent liquor ( between 2 . 2x 1 0S and 2 . 6x1 06 cel ls/ml ) ( fig. 5 . 7 b ) . Wi thin the tower , the concentration was generally greater than that of K. marxianus at corresponding heights . Thus , these cells which were very flocculent caused the very high yeast cell concentrations reached in the tower ( fig . 5 . 8 ) and contributed to the rapid consumption of sucrose . Some of the values of S. a erevisiae cell dried weight listed in table 5 . 4 d o no t give a true representation of the cell concentration because these values were obtained as the difference between the measured total cell dried weight and the estimated K. marxianus cell dried weigh t . The uncertainty of the estimated K. marxianus weight was high ( between 280 and 460% of the values reported ( sect . E . 6 . 4 ) ) . ( c ) Total cell concentration At a cons tant veloci ty of 0 . 087 and 0. 1 2 mm/ s , the total cell concentration ( fig . 5 . 8 ) increased to a high value as the height in the tower increased from 0 . 096 to 1 . 57 m and then decreased rapidly as the height increased to 2 . 32 m . There was a different profi le a t the highest veloc ity ( 0 . 1 4 mm/s ) . The cel l concentration showed a general reduction as the height increased from 0 . 096 to 2 . 32 m . The concentration a t 2 . 32 m was greater than for the two lower veloci ties . The resul ts showed a high cell concentration wi thin the first 1 . 57 m o f the tower . Wi thin these sec tions , there was a greater concentration of the more flocculent yeast S . cerevisiae CFCC39 . Thus , the cell concentration was high . I t was considered that the chemicals present in whey and the slow velocities at which the tower was operated contributed to this high flocculence . The ini tial increase in the cell concentration as the height increased to 1 . 57 m for the two lower veloci ties ( 0 . 087 and 0 . 1 2 mm/s ) was considered to be the effect of gas production in the yeast floes . More gas was produced in the lower tower sec tions up to 0 . 82 m and as a result the void volume increased . At 1 . 57 m , there was less gas 1 38 2 3 5 ....... 1 60 -Ill) .. H o .a: --......... rJ .... H "--fJ:l 120 :3 A fJ:l H p:: A ...:I ...:I 80 fJ:l u � H 0 V mm/ s Q ml/h H s .. 40 0 . 087 160 • � 0 . 1 2 220 ---· 0 . 14 260 --- · (b ) 600 ..-1 -Ill) .. 500 H ::r: 0 H i:il :3 H i:il :3 A i:il 0 � H p:: H z i:il u ...:I ...:I i:il u � H 0 H .. � 400 300 200 0 . 5 1 . 0 1 . 5 2 . 0 2 . 5 TOWER HEIGHT • m Fig . 5 . 8 Cell concentration vs tower height at various superficial liquid velocities . (a) Cell dried weight , (b) Cell centrifuged wet weight . 1 39 production and , thus , the cell concentration was greater . The decrease in the cell concentration at 2 . 32 m was due to the highly flocculent nature of the yeas t . The floes were large and heavy and were able to settle to the bot tom of the tower because the liquid veloci ty was low and was not able to suspend the floes . The different pro file observed at the highest veloci ty of 0 . 1 4 mm/s was a result of the expansion of the yeast bed with an inc rease in the velocity , dec reasing the cell concentration at heights between 0 . 82 and 2 . 32 m . I t was observed that the concentrated cell mass had expanded to the bo ttom of the separator and blocked the return path of the floes . Thus , the concentration was lower at 1 . 57 m , and was greater at 2 . 32 m , than for the other two veloci ties . The total cell concentration obtained in this experiment was considerably greater than those reached in the tower fermentation of whey permeate ( fig . 4 . 9 ) and molasses ( fig . C . 5 ) . This difference was due to the increased flocculence of the yeast strain CFCC39 which was used here . The high cell concentration and glutinous yeast cell mass was found to cause considerable channelling inside the tower ( fig . 5 . 9 ) . This was a result of the low velocity used and the extremely flocculent nature of the yea�t strain CFCC39 . A very high veloci ty would be required to suspend the yeast floes . The high concentration of calcium ions in the whey permeate probably contributed to the good flocculation of the yeast cells inside the tower . The modified Burn ' s number ( MEN* ) for the tower yeast culture was 1 66 indicating that the yeasts were very flocculent . The continuous cell mass did not permit free movement of carbon dioxide and the medium up the tower but induced coalescence and thus channelling occurred . This in effect reduced the cell to sugar contact and created gas slugs which slowly washed out the less flocculent K. mar xianus by creating turbulence in the separato r . As the flow rate was increased the volume occupied by the yeast cells expanded . When the expansion reached the bot tom of the separato r , the yeast recycle path was blocked . This also created turbulence in the separator and cell washout . Thus , it is possible that the use of a very highly flocculent yeast together wi th a moderately flocculent yeast as in this study may be undesirable . It may be more appropriate to use the less flocculent S. cerevisiae strain SC 1 46 , which was used in the tower fermentation of 140 Fig , 5 . 9 Channell ing inside the tower during mixed cul ture and mixed subs trate tower fermentation . 1 4 1 molass e s ( Appendix 6 ) , e i ther as mixed culture w i th KM Y42 , o r to use only KM Y42 . The re was , however , insuffic i en t t ime to inves tigate this in the t ower fermente r . The natu re o f the c e l l mass also caused problems during sampl ing . The cell mas s would not flow out through the sampling tubes eas i ly . Each tube had to be squee zed l ike a too thpas te tube to obtain suffic ient sample f o r analysis . An uncertainty a l lowance o f 5% should be added to the cell concentra t ion values fo r h eights up to 1 . 57 m ( sect . E . 5 . 3 ) . The sampl ing problems inc reased analysi s errors by inc reas i ng t ime between sampling and separa t ion o f yeast cel l s from the liquor , which was pa rticularly c ri t i cal at the bo t tom sampl ing po int ( 0 . 096 m) because o f the rapid red uc t ion in the sugar concen t ra t ion in thi s reg ion . Chemical agen t could no t be added to k i l l t he yeas t cel l s because the sampl e was requi red fo r a plate count . In add i t i on , the chemical agent if add ed could in t erfere with the enzyme membranes used fo r sugar ana lysis . 5 . 1 . g f'l ed ium pH The pH ( fig . 5 . 1 0 ) o f the fermen t ing med i um d e creased from a n ini tial value between pH 4 . 0 and 5 . 0 t o between pH 4 . 6 and 4 . 8 ove r t h e fi rst 0 . 096 m o f the tower heigh t . I t then increased to between pH s . o and 5 . 1 as the heigh t increased to O . b2 m and remalned l n this range up to the ex i t fo r all veloc i ties used . pH 5 . 4 . 0 . 5 ._ 0 . 1 4 __.__Q.J� ----A... I. --- -- -� 2 --_..:::--:::. --- .-'::.:::.... 0 . 08 7 I V mm/ s Q ml/h s • 0 . 08 7 1 60 ·--- 0 . 1 2 2 2 0 · -- - 0 . 1 4 260 4 1 . 0 1 . 5 2 . 0 2 . 5 TOWER HEIGHT , m Fig . 5 . 1 0 Medium pH vs tower height a t various superf i c ial l i quid veloc i ties . 5 1 42 The trends obse rved here were similar to those in the tower fermentation of molasses ( fig . C . 6 ) in that there was an ini tial pH reduction ove r the fi rst 0 . 096 m of tower height and the pH returned to between pH 4 . 8 and 5 . 0 as the height increased . The major pH change occurred in the tower sections where ethanol was produced and thus there was lit t le change in pH at heights greater than 0 . 82 m . A s imilar pattern was observed during fermentation of whey permeate . 5 . 2 C ONTAMINATION OF CONTINUOUS TOWER FERMENTATION CULTURE Great care was requi red to ensure that the tower was not contaminated by al lowing unfiltered air to enter through the sample ports and the feed inlet . Infec tion problems were compounded in that the high res idual lac tose concentration in the upper sec tion of the tower permit ted growth of bacteria . This occurred toward the end of the continuous operation at the velocity of 0 . 1 4 mm/s . Ins tead of abandoning the operation , it was decided that penicillin and streptomycin should be added to the medium to suppress the infect ion and permit the continuation of the fermentation . 5 . 3 TOWER FERMENTER PERFORMANCE I t was evident from the results described previously ( sec t . 5 . 1 ) that the tower operation at the velocity of 0 . 087 mm/s was the only velocity at which there was 90% to tal sugar utilization at a tower height of 2 . 32 m . Thus , only the performance at this veloc ity could be used as a comparison with other tower fermentations . This veloci ty was lower than the optimum velocities of 0 . 24 and 0 . 33 mm/s observed in the tower fermentation of whey permeate ( sect . 4 . 4 ) and molasses ( sec t . C . 4 ) , respectively . It was shown that this fermentation was limi ted by the diauxic behaviour of K. marxianus due to the presence of sucrose and lac tose together and sucrose was util i zed before lac tose in the tower sec tion below 0 . 82 m . At greater heights , the rate of fermentation of lac tose was probably reduced by the presence of ethanol . Thus , it was shown that in order to achieve the same rapid rate of fermentation as for whey permeate fermentation , the K. marxianus cell concentration should be considerably higher than for whey permeate fermentation . S. cerev is iae cells did not take part 1 43 in the fermentat ion o f lactose . This could be regarded as an inefficient use of the yeast cells and it would be more efficient to use only K. marxianus in the towe r , provided that a very flocculent strain were available ( sec t . 5 . 5 ) . The height of 2 . 32 m , required for 90% total was greater than the 0 . 82 m required for the sugar other uti l i zation , two towe r fermentations to allow longer residence time for lac tose to be uti lized . The residence time of 7 . 4 hours was long when compared with the time of 1 . 0 and 0 . 7 hour for the tower fermentations of whey permeate ( sect . 4 . 4 ) and molasses ( sec t . C . 4 ) , respectively . The increased time was a d i rect result of the d iauxic behaviour of the yeast and probably also ethanol inhibition . TI1e resul ting ethanol productivity from this was 5 . 8 g/ lh at 2 . 32 m . The productivity was lower than 1 6 and 67 g/lh obtained ( exclud ing the separator) in the tower fermentat ion of whey permeate and molasses , respec tively . Batch fermentat ion of a s imilar mixed sugar med ium required 48 hours to produce 39 g/1 ethanol ( fig . 7 . 3 a ) . This gave a productivity o f 0 . 8 g/lh which was approximately one seventh of the rate observed in the present work . Thus , towe r fermentation of mixed sugar did no t achieve a very high produc tivity , but was better than batch fermentation of a similar substrata . The fac tors which limited the performance o f the tower fermenter in these expe riments could be listed as follows : 1 . The presence of sucrose and lac tose together caused K. marxianus to exhibit diauxic behaviour . Lactose uti lization was repressed by glucose produced from suc rose , so a longer res idence time was requi red to achieve lactose fermentation and this fermentation occurred at heights within which the re was high ethanol concentration . Thus , the maximum superfic ial veloc i ty was limi ted and the height required increased in order to achieve a long fermentation time . 2 . In the upper sec tions o f the tower above 0 . 82 m , there was probably ethanol inhibi tion , because the rate of lac tose uti lization was low even though the cell concentration was comparable with that used in whey permeate fermentation . K. marxianus reduced considerably Thus , the fermentative ac tivity of within these heights . This effect contributed to the incomplete K. marxianus concentration than lactose utilization . A greater that achieved in this fermentation 1 44 would be required to achieve the same degree of lac tose utili zation as rapidly as in the whey permeate fermentation . 3 . There was a difficulty in maintaining a sufficiently high concentration of K. marxianus in the tower fermenter to compensate for the reduction in the lac tose fermentation rate throughout the tower . The difficul ty was caused by a combined effec t o f the moderately flocculent nature of this yeast strain compared wi th that of S. cerevisiae CFCC39 and the incompatability of 'the two yeasts in the tower. The highly flocculent yeast mass caused gas slug formation and channell ing which washout of the velocities used . c reated turbulence in the separator . less flocculent yeast occurred even Thus , the slow at the low It was desirable , therefore , to investigate whether the incomplete lactose uti l i zation was caused by an insufficient concentration of K. marxianus as well as by the diauxic behaviour , or as a result of the interac t ion of molasses and whey permeate . This is d escribed in sect . 5 . 4 . An investigation of the effect of the concentration of each cul ture in the inoculum on the final cell ratio was also considered desirable , because it could clarify whether the observed change in the K. marxianus concentration was caused by washout as described earlier . This investigation is described in sect . 5 . 5 . 5 . 4 COMPARISON BETWEEN FERMENTATION OF WHEY PERMEATE ENRICHED WITH MOLASSES AND WITH SUCROSE In various fermenta tions of whey permeate enriched with molasses ( sect . 5 . 1 and 7 . 3 ) it was observed that there was always a greater amount of res idual lac tose than sucrose remaining at the end of the fermentatio n . An experiment was conducted to investigate the effect of the addi tion of molasses to whey permeate . Molasses and pure sucrose were added to separate quantities of whey pe rmeate to give a final sucrose concentration of 60 g/1 . The fermen tation was carried using 450 ml of media volume in 2 litre shake flasks . The media were inoculated with 50 ml of inoculum and incubated at 30°C with shaking at 1 50 rpm . I t was found that the sucrose concentration ( table 5 . 5 ) was reduced rapidly in both media and was completely utilized in 24 hours . The residual concentrations were 0 . 4 and 0 . 1 g/l in molasses and 1 45 sucrose enriched media , respectively . In contrast , lactose was reduced , after 48 hours , to 6 . 7 and 2 . 1 g/ 1 ( 83 and 94% lac tose utili zation , respectively ) . After 72 hours , there was a very small decrease in the lac tose concentration . Tabl e 5 . 5 Compar ison between fermenta tion of whey permea te enriched with molass es and with s ucro se by K. marxianus Y42 and S. cerevisiae CFCC39 . PM - Permeate + molas s es ; P S - Permeate + sucros e media time s sl st E XKM/X pH s h g /1 g / 1 g / 1 g / 1 % PM 0 5 5 3 9 94 - 43 4 . 9 P S 0 58 37 9 5 - 44 5 . 1 PM 24 0 . 4 2 2 2 2 2 5 46 4 . 5 PS 24 0 . l l l l l 29 55 3 . 8 PM 48 0 . 4 6 . 7 7 . l 32 58 4 . 4 P S 48 0 . l 2 . l 2 . 2 34 66 4 . 2 PM 72 0 . 4 3 . 2 3 . 6 2 8 56 4 . 4 PS 7 2 0 . l 1 . 7 1 . 8 3 2 6 3 4 . 3 s s Sucrose concentration St - To tal s ugar concentration s1 Lactos e concentration E - E thano l concentr at ion XKM/X - Ratio of K. marxianus to to tal ce ll number The to tal residual sugar concent ration was higher in molasses-enriched whey permeate than for sucrose-enriched permeate , throughout the fermentation . After 48 hours , the whey the concentrations were 7 . 1 and 2 . 2 g/ 1 in molasses and sucrose-enriched whey ( 92 and 98% utilization , respec tively ) . The major component of the residual to tal sugar was lac tose . There was less ethano l in the molasses-enriched whey throughout the fermentation . Maximum ethanol concentrations were reached after 48 hours and were 32 and 34 g/1 , respec tively . This was an ind ication of a slower fermentation in molasses-enriched whey . The ratio of K . marxianus Y42 to the to tal cell count ( XKM/X ) remained close to 50% throughout the fermentation , showing a smal l increase as the fermentation progressed . The pH of whey enriched with sucrose decreased more than the pH of 1 46 whey pe rmeate enriched wi th molasses . This was probably due to a grea ter rate o f ca rbon dioxide produc tion in whey permeate enri ched wi th suc rose . The resul ts ind icated that molasses had an inhibitory effec t on fermentative ac t iv ity of K. marxianus permea t e . Y42 when used to enrich whey It has been reported tha t lactose can be thermal ly rearranged to lactulose upon steril ization ( Thayani thy et al 1 982 ) . The lac tulose fo rmed was no t metaboli z ed by a bac t erial 8 -amylase enzyme . The amount o f lac tose ( lactulose ) remaining was reported to be be tween 20 and 36% . However , a shake-flask fermentation test using 20 g/ 1 lac tulose demonst rated that K . marxianus Y42 grew on lac tu lose . There fo re , the unmetabo l i z ed lactose detec ted in this s tudy was unlikely to be due to the formation of lactulose . 5 . 5 FERMENTATION COMPARISON USING DIFFERENT RATIOS OF MIXED YEAST CULTURE IN THE INOCULUM I t was observed that during the tower fermentation by K. marxianus Y42 and S . cerevisiae CFCC39 of whey permeate enriched with molasses ( sec t . 5 . 1 . 8 ) , there was a dec rease in the number of K. marxianus Y42 cel l s as the fe rmentation progressed . The expe riment described here inves t igated the effec t o f the ratio of the two yeas t s t ra ins in the inoculum on the cell ratio during fermentat ion of whey permeate enriched with mo lasses . Three cell ratios in the inoculum were used , 9 : 1 , 1 : 1 and 1 : 9 (K. marxianus Y42 : S . cerevisiae CFCC39 ) The fe rmenta tion was carried out using 450. m l medium volume in 2 l i t re shake flasks . The flasks were inoculated with 50 ml o f inoculum and were incubated at 30°C with shaking at 1 50 rpm . The resu l ts o f the fermentat ion are shown i n table 5 . 6 . The total cell counts fo r the three inoculum ratios were similar at inoculation and afte r 24 hours . However , after 48 hours , the 9 : 1 inoculum ra tio had the lowest to tal ce l l number whi l e the medium inoculated wi th 1 : 9 inoculum ratio contained the greatest cel l concentration . For the 1 : 9 inoculum ra tio , lactose uptake required longer time because of the lower initial K. marxianus Y42 concentration . The ratio o f K . marxianus Y42 to the total cell count ( XKM/X ) was found to be c lose to the ra tio at inoculation time fo r 9 : 1 and 1 : 1 1 47 Table 5 . 6 Fermenta tion comparis on us ing different ratios o f yeas t cul ture in the inoculum . KM: SC - Ratio of K . marxianus Y42 to S. cerevisiae CFCC39 in the inoculum . KM: se X XKM �/X s sl 8l u s t E pH s ratio cells /ml cells /ml % g / 1 g / 1 % g / 1 g / 1 0 hour 8 8 9 1 4 7 . 5 3 8 . 0 85 . 5 4 . 7 9 : 1 1 . 2 x10 8 1 . 1 x 1 0 7 - - 1 : 1 1 . 0x 108 6 . 0x 10 7 5 9 4 7 . 0 3 7 . 1 - 84 . 1 - 4 . 7 1 : 9 1 . 1 x 1 0 1 . 2x 10 1 1 4 6 . 6 3 8 . 0 - 84 . 6 - 4 . 7 24 hours 8 8 84 0 . 5 1 2 . 3 6 8 1 2 . 8 3 8 . 6 4 . 4 9 : 1 8 . 7x 1 09 7 . 3x 1 08 1 : 1 1 . 1 x 10 8 5 . 2x 10 8 4 8 0 . 4 1 9 . 8 47 20 . 2 29 . 3 4 . 3 1 : 9 7 . 3x 1 0 9 . 0x 10 1 2 0 . 4 30 . 5 20 30 . 9 25 . 7 4 . 4 4 8 hours 9 : 1 8 8 94 0 . 5 7 . 0 82 7 . 5 3 9 . 3 4 . 4 3 . 2x 1 0 9 3 . 0x 108 1 : 1 1 . 1 x1 09 5 . 1 x1 08 4 4 0 . 4 6 . 9 8 1 7 . 4 29 . 5 4 . 4 1 : 9 2 . 2x 1 0 6 . 0x 10 2 8 0 . 4 9 . 9 74 10 . 3 28 . 5 4 . 4 X - To tal cel l number XKM - K. marxianus cel l numb er slu - percentage lactose utilizat ion flasks . However , in the case of the 1 : 9 ratio flask the c e l l ratio inc reased more than 2 fold after 48 hours . Suc rose ( S s ) was found to be completely u t i l i zed after 24 hours for all inocu lum rat ios . The grea test uti l i zation o f lac tose (82% ) occurred in the 9 : 1 ratio flask . The lowest u t i l i za tion ( 74% ) occurred at the 1 : 9 inoculum ra tio . Thus , lactose uti l i zation was affec ted by the conc entra tion of K. marxianus Y42 in the inoculum fo r a ll ino culum ratios whereas suc rose consumption was not affec ted . The re was more rapid total sugar uti l i zation when the ratio of K. marxianus Y42 in the inoculum was higher . The inocu lum ratio o f 9 : 1 gave the lowest resi dua l to tal sugar level o f 7 . 3 g/ l after 48 hours ( 92% u t i l i zation ) . After 24 hours , the med ium inocula ted with the 9 : 1 ratio contained the highest concentration o f ethanol ( 39 g/ 1 ) . The lowest concent ra tion o f 26 g/ 1 was in the 1 : 9 inoculum ra tio . After 48 hours , the order was the same . 1 48 5 . 6 EFFECT ON FLOC MORPHOLOGY OF SPECIES RATIO IN INOCULUM The mo rphology of the yea s t floes present in the med ium at the end of the fermentation is shown in fig . 5 . 1 1 . In ( a , 9 : 1 ) , the re were mainly small floes of K. marxianus Y42 , as it accounted for 90% o f the total ce l l concentration . In ( b , 1 : 1 ) , there were more large floes o f diame t e r up t o 1 . 5 mm . The two yeast strains had apparently formed mixed f loes because the flo c size increase corresponded with an inc rease in S. cerevisiae CFCC39 . In ( c , 1 : 9 ) , there were many larger floes up to 4 mm in diameter . This was a resu l t o f the higher ra tio o f S . cerevisiae CFCC39 . I t i s ev ident tha t the higher concent ration o f K . marxianus Y42 in the ino c ulum resu l ted in fas ter sugar u ti l ization and e thanol produc tion during the early stage o f the fermentation . This also resul te d in grea ter ethano l p roduc tion over the same fermentation time when compared with the fe rmentation using a lower initial concentrat ion of K. marxianus Y42 . There was an inc rease in the ratio o f K . marxianus Y42 when its concentration i n the inoculum was low . This occurred afte r sucrose was used up and lac tose was being u t i l i zed . Thus , the cell ra tio changed from the level in the inoculum since S . cerevisiae was unable to uti li ze the remaining substrate . Th is , howev e r , occurred when the e thanol concentration was high and the ac tivity of K. marxianus was therefore red uced . This resul ted in slow sugar uptake and e thanol produc tion . Hence , there was evidence that it may be more d esirable to use a concent ration of K. marxianus Y42 greater than a 50 : 50 ratio in the towe r . Accord ing to the resul ts observed here , it should be the dominant cell population throughout the tower fermente r . There would then be a more rapid lac tose S . cerevisiaeCFCC39 resul ted uptake . in la rger desirab le in the tower fermenter . Higher floc concentration o f fo rmation and this is The composi t ion of the mixed cul ture is therefore determined by a compromise between maximum sugar uti l i zation and floc re tention at high superficial liquid veloc ity . 149 (a ) 9 1 ( b ) l (c ) 1 9 ac tua l s i z e F ig . S . l l Th e e f f ec t of mixed c u l t u r e ra t i o o n f l o c nw r p h o l o g y i n s h a k e f l a s k c u l tu r e . 1 50 5 . 7 C ONCLUSIONS 1 . The re was incomplete lac tose u t i l iza tion in the tower fermenta tion of whey permeate enriched with molasses by m ixed cul ture of K. marxianus Y42 and S. cerevisiae CFCC39 . Th is was found to be a result o f many fac tors : Diauxic behaviour of K . marxianus Y42 which caused sucrose to be u t i l ized befo re lac tose , Ethanol and molasses inhibi t ion of K. marxianus Y42 fermentative ac tivity , The difficul ty in main taining a suffic iently high K. marxianus Y42 population in the tower due to cell washout caused by channel l ing , gas s lug formation and the only moderately flocculent nature o f K . marxianus Y42 , 2 . Channell ing and gas slugging occurred because the yeas t bed was too dense ( no t fluid i zed ) at the low superfic ial liquid veloc i t i es used . 3 . The two yeast s t rains employed were no t compatible for use in this fermentation because the mixed cul ture floes that were formed could be broken down easily by gas bubbles , and the less flocculent K. marxianus was s lowly washed out of the tower . 4 . An increased concentra t ion o f K. marxianus in the mixed cul tu re improved both the rate and yield of e thanol when grown o n the mixed subt ra te . 5 . I t may be desi rable to use a pure culture of K. marxianus for fe rmentation of this subs trate . This area of the work was not taken further , but it is clear tha t more experiments could be des igned to s tudy the behaviour o f mixed cul tures o f various s trains in completely flui d i z ed bed s , examining the effec ts of popula t ion ratios on floc size and ope ra ting at greater superfic ial veloc i ty . 5 . 8 SUMMARY There was incomplete lac tose uti l i zation i n the towe r fermentation o f whey permeate enriched with molasses by mixed cul ture o f K. marxianus Y42 and S. cerevisiae CFCC39 . This was shown to be the resu l t of many fac tors : The diauxic behaviour of K. marxianus Y42 in the presence of suc ros e and lac tose toge the r caused sucrose to be 1 5 1 u t i li z ed first ; e thanol inhibition s lowed down lac tose uptake fermentat ion was inhibited by molasses and there was d ifficulty i n maintaining a high populat ion of K. marxianus Y42 because o f c ell washout , caused by channel ling , gas s lug fo rmation and the only mod erat e ly flocculent nature o f K. marx&anus Y42 . CHAPTER 6 FLOCCULATION TESTS 6 . 1 INTRODUCTION Tower fermenta tion of whey permeate to e thanol requi res a flocc ulent lac tose-fe rmenting yeast . flocculent lac tose-fe rmenting yeas t There are no published reports o f species . There was a need , therefo re , to select for such an o rganism . Thus , the experiments d es c ribed in the fo llowing sections were pe rformed . Once the desi red yeas t was obtained , i ts flocculent behaviour in various media was s tudied , inc lud ing its behaviour in the presence o f S. cerevisiae in the same medium . It was intended t o util i z e both yeast species together to ferment a mixed subs trate cons is t ing of whey permeate enri ched wi th molasses . 6 . 2 TEST MEDIA The composi t i ons of the various med ia used in these experiments have been inc luded here ( table 6 . 1 to 6 . 6 ) for convenience during the presentation and d is cuss ion of results . All media were steri li zed as described in sec t . 3 . 3 . 1 . The pH value was adj us ted by the add i tion o f sulphuric acid or calcium hyd roxid e . Sodium hydroxide was not used for pH ad jus tment because sod ium ions may interfere wi th certain sites on yeast cell walls whi ch normally complex wi th cal c ium ions to form floes with other cells ( Stewart & Goring 1 976 ) . 6 . 2 . 1 A B C a CB CC39 EF F FM G Glossary of abb reviations used in flocculation tes ts Aluminium sulphate , Al2 ( S04 )3 Broth Calcium sulphate , Ca SC4 Yeast c leaning buffer ( Ca S04 wash ) S. cerevisiae CFCC39 Ext remely floccul en t Medium was membrane fi l te red ( 0 . 45� ) Flocculation med ium ( acetate buffer ) Glucose 1 5 2 KM Y42 M M* Ma MBN MBN* Me MF Mo Ms na NF p Pe P4 . 6 R SC 1 46 SM TS VF WF y YM 1 0 44 46 5 1 . p 1 F . PF 1 5 3 K . marxLanus Y 42 Ma l t exrac t bro th ( Di fco ) Ma l t ex t ract bro th ( Oxoid ) Mal tose Mod ified Burn ' s numbe r Non-s tandard mod if ied Burn ' s number Mal t ex t ra c t powder medium Moderately flocculent Molasses Mal t ex t ract syrup ( Maltexo ) Data not available Non-flocculent Whey permea te Pep tone Whey permeate wi th no pH correc tion Rough S . cerevisiae FT 1 46 Spent mal t ex t ra c t bro th Subcultured from the bot tom of the tower fe rmente r Very flocculen t Weakly flocculent Yeas t ext ract powde r Yeas t-mal t ext ract b ro th 1 00g/ l whey extra c t solution Rat i o of lactose to suc rose of 40 : 40 g/ 1 Rat i o o f lac tose to sucrose o f 40 : 60 g/ 1 pH 5 . 0 Table 6 . 1 Whey permeat e as the basic medium . Whey permea te containing approximately 40 g/ 1 lactose ini t ial pH was 4 . 6 ; adjusted to 5 . 0 . Whey permeate , membrane fil tered after autoclaving t o ex t ract precipi tate ; pH adj us ted to 5 . 0 before autoclaving . 2 . P4 . 6 Whey permea t e a t pH 4 . 6 ; no pH adjus tment o r precipi tate removal aft e r aut o c laving . 2F . P4 . 6F As fo r 2 . P4 . 6 bu t membrane fi l tered aft er aut oclaving . 1 54 3 . PCa Whey permeate plus 0 . 5 g/1 calcium chloride whi ch was added after autoclaving pH adjusted to 5 . 0 . 3F . PCaF As for 3 . PCa but membrane filtered after autoclaving and before calcium chloride was added . 4 . py Whey permeate plus yeast extract powder ( 1 g/1 ) pH ad justed to 5 . 0 . 5 . PYM Whey permeate plus yeast extract powder ( 1 g/ 1 ) and mal t extract powder ( 1 g/1 ) ; pH was 4 . 7 not adjusted . 6 . PYMCa Whey permeate plus yeast extract powder ( 1 g/1 ) malt extract powder ( 1 g/1 ) ; and calcium chloride ( 0 . 5 g/1 ) ; pH 4 . 6 ; no adjustment . 1 . PXM Whey permeate plus mal t extract broth powder ( 1 5 g/1 ) no pH adjustment . 8 . PXM5 As for 7 . PXM but pH adjusted to 5 . 0 . 8F . PXM5F As for 8 . PXM5 but membrane fi ltered after autoclaving . 9 . PMs Whey permeate plus mal t extrac t syrup ( 20 g/1 ) ( Maltexo ) , Vegemite ( 20 g/1 ) ; and 0 . 5 g/1 of each of calcium chloride ; diammonium sulphate adjusted to 5 . 0 • diammonium hydrogenphosphate pH 1 0 . PPe Whey permeate plus peptone powder ( 3 g/1 ) pH not adjusted membrane filtered . 1 1 • PUAm Whey permeate plus urea ( 1 g/1 ) and diammonium hyd rogenphosphate ( 0 . 5 g/1 ) filtered . pH adjusted to 5 . 0 membrane 1 2 . P 1 0 Whey permeate with added lac tose to 1 00 g/1 pH adjusted to 5 . 0 . 1 3 . PMo46 Whey permeate plus molasses ; lactose to sucrose ratio of 40 to 60 g/1 ; g/ 1 urea ; 0 . 5 g/1 diammonium hyd rogenphosphate pH adjusted to 5 . 0 . 1 3F . PMo46F As for 1 3 . PMo46 but membrane fil tered after autoclaving . 1 4 . PMo44 As for 1 3 . PMo46 but the sucrose concentration was 4 . 0 g/1 . 1 5 . PB Whey permeate containing 40 g/1 sucrose ; 1 g/1 mal t extract powder ; 1 g/1 yeast extract ; 0 . 5 g/1 of each of diammonium sulphate ; diammonium hydrogenphosphate ; & calcium chloride ; no pH adj us tmen t . 1 6 . PB5 As for 1 5 . PB but pH adjus ted to 5 . 0 . 1 6F . PB5F As for 1 6 . PB but membrane fil tered after autoclaving . 1 7 . PMoYMCa As for 6 . PYMCa but molasses was added to give 60 g/1 1 5 5 sucrose ; p H was not adjus ted . 1 7F . PMoYMCaF As for 1 7 . PMoYMCa but membrane fi ltered afted autoclaving . 1 8 . Mo Table 6 . 2 Molasses medium . Mo lasses med ium containing 1 00 g/1 sucrose ; 1 g/ 1 urea g/ 1 diammoni um hydrogenphosphate ; pH adjusted to 5 . 0 . Table 6 . 3 Lac tose as the sole sugar source . 0 . 5 1 9 . LYA Lac tose ( 40 g/ 1 ) plus 0 . 5 g/1 aluminium sul phate 0 . 5 g/ 1 d ihydrogen sulphate ; 0 . 5 g/1 dipotass ium hydrogenphosphate ; 1 g/1 yeast extra c t powder ; 2 g/ 1 yeast ni trogen base ; pH adjusted to 5 . 0 . 20. LYCa As for 1 9 . LYA with calcium chloride ( 0 . 5 g/1 ) instead o f aluminium sul phate ; p H 4 . 8 no adjustment . 2 1 . LSM Lac tose ( 40 g/ 1 ) was add ed to spent mal t extract bro th pH adjusted to 5 . 0 . Table 6 . 4 Ma ltose as the sole sugar source . 22 . Ma Mal tose ( 1 0 g/1 ) plus yeas t nitrogen base ( 2 g/ l ) ammonium chloride ( 1 g/ 1 ) and dipo tassium hyd rogenphosphate ( 0 . 5 g/ 1 ) ; pH adjusted to 5 . 0 . 23 . MaCa As for 22 . Ma but calcium chloride ( 0 . 5 g/ 1 ) was added . 24 . MaCa4 As for 23 . Ma Ca but mal tose con tent was inc reased to 40 g/ 1 . Table 6 . 5 Glucose as the sole sugar source . 25 . G As for 24 . MaCa4 but glucose ( 40 s/1 ) replaced mal tose . 26 . GCa As for 25 . G but with calcium chloride ( 0 . 5 g/ 1 ) . 156 Table 6 . 6 P repared media . 27 . YM Yeas t-ma l t ex t ract bro th ; pH adjusted to 5 . 0 . 28 . YMCa As for 27 . YM but with cal c ium chloride ( 0 . 5 g/ 1 ) . 29 . YMA As for 27 . YM but wi th aluminium sulphate ( 0 . 5 g/1 ) . 29F . YMCaF As for 29 . YMA but membrane fi l tered aft er auto claving . 30 . M Mal t extract bro th ( Difco ) ; pH adjusted to 5 . 0 . 3 1 . M* Mal t ex t rac t bro th ( Oxoid Ltd . , Basings toke , UK ) ( 40 g/1 ) pH adj usted to 5 . 0 . 32 . MCa As for 30 .M but with cal c ium chl oride (0 . 5 g/1 ) . 33 . Ms I1al t ext ract syrup ( Mal t exo) ( 50 g/ 1 ) ; pH ad justed to 5 . 0 membrane fi l t e red aft e r aut o c laving . 34 . Me 35 . CB Mal t ext rac t powder ( 1 5 g/1 ) ; pH adjusted to 5 . 0 . Table 6 . 7 Flocculence measurement med ia . Yea s t cleaning buffer ( calc ium sulphate wash ) solution o f calcium sulphate ( 0 . 5 g/1 ) . aqueous 36 . FM Flocculation med ium ( acetate buffer ) con tained cal c ium sulphate ( 0 . 5 g/ 1 ) ; sodium acetate ( 6 . 8 g/1 ) ; glacial ace t i c a c i d ( 4 . 05 g/1 ) i n deioni zed wa ter ; p H 4 · 9 · 6 . 3 FLOCCULATION TEST RESULTS These expe riments were intended to select flo cculent sui table for tower fermen ta tion of whey permeat e . yeas ts 6 . 3 . 1 Mod ified Burn ' s number and flocculation tes ting methods used ( a ) Flocculation test media The s tandard method for measuring mod i fied Burn ' s number ( MBN ) ( G reenshields et al 1 972 ) requires the yeas t to be grown in a very concentrated mal t ext ract solution ( 200 g/ l ) which has been treated wi th pepsin and trypsin to make t he solut i on pro tein free and then fi l tered to extrac t pre cipitates which could interfere wi th the flocculation tes t . Thi s medium was found to be unsui table for lac t o se fermenting yeasts which general ly cannot uti li ze the mal tose present in the mal t extrac t . K . marxianus Y42 is an except ion to thi s , since i t can ferment both mal tose and lac t ose . 1 5 7 Malt extrac t solution was found t o give poo r growth and t o be difficul� to prepare . Normal malt extract bro th ( 1 5 g/ 1 ) , using 1 00 ml volume in a 250 ml f lask , yielded approximately gm o f wet yeas t which was sufficient for the flo ccu lation tes t . This medium was reasonably c lear after aut o c laving . If more yeast cells were required , they could be prepared easi ly by increasing the number o f shake flasks . The mod ifled Burn ' s numbers determined using normal malt extract bro th were of the same o rd e r as those attained using mal t extract solution . This s tudy obtained MBN 1 72 for s train CC39 whi le C oote ' s ( 1 974 ) value for a s t rain o f S . cer>evisiae ( CFCC 54 ) , the mos t flocculating yeast tested by him , was 1 70 . The main advantage o f mal t ex trac t i s that i t i s easier t o prepare and t o s tandard i z e since i t i s a c omme r c i a l ly prepared med ium . In o rd e r to carry out f locculation tes ts on lac tose ferment ing yeasts , i t was found to be more appropria te to grow them in whey permeat e rather than in mal t ex tract bro th . The disadvantage o f whey permeate was i ts tendency to precipitate upon auto c laving and i ts residual pro tein content . I t also has a high content o f calcium i ons . It was found during the test that membrane fil tration extracted components important for good growth and floc culence o f Kl-'I Y42 . However , floc culence tests using whey permeate were considered to be appropriate to the aim o f this work , that is identifying flocculent lac tose fermenting yeasts for growth in the tower fermenter . The standard method required MBN measurement to b e carried out using the flocculation med ium ( ac etate buffer or 36 . FM ) . Thi s med ium was found to reduce flocculence of KM Y42 grown in mal t ex t rac t b ro th . In o rd e r to have a be t ter ind i cation o f the yeas t set t l ing abi l i ty , i t may be bes t to measu re the flocculence in the medium tha t wi l l b e used for fermentation . The standard med ium could be retained o r used for comparing the flo cculating abi l i ty of d i fferent yeas t s � rains . I t was with thi s concept in mind tha t many tests were carried out using the growth med ia or med ia of po tential interest as the flo c culatlon test ing med ium . The summation of the slope of defined as MBN* ( sect . 3 . 4 - 5 b ) . the yeast set t l ing This , with an v o l ume was appropria t e mul tiplication fac to r , should more c o r r e c t ly b e cal led " the rat e o f yeast s e t t lement " hav ing a uni t o f "ml/minute " . 1 58 ( b ) Technique for the determination of modified Burn ' s number Ano ther shortcoming of the Sharp ' s modified Burn ' s method ( Greenshields et al 1 972 ) is the summation method for the slope of the settling volume curve . This involves the addition of the average slopes between 0- 1 , 1 -5 , 5- 1 0 , and 1 0- 1 5 minutes which were taken from a curve plotted to the scale of inch/2 ml and 1 inch/5 minutes . Another time zone which was found to be important was between 1 -2 minutes . During this period some very flocculent yeasts set tled at a much greater rate than between 2-5 minutes ( see curves for strains CC39 in comparison wi th strains SC 1 46 and KM Y42 in fig . 6 . 1 ) . In this situation it would be difficult to obtain a reasonably accurate average s lope from the curve between 1 -5 minutes . Consideration of this time zone would help to refine this method further . ( c ) Alternative technique The calculation of the average slopes between 0- 1 , 1 -5 , 5 - 1 0 , and 1 0- 1 5 minutes using the yeast settled volume measured at 0 , 1 , 5 , and 1 5 minutes , summation and then mul tiplication by 25 to convert the sum to Sharp ' s graphical scale (Greenshields et al 1 972 ) will give the same MBN value wi thout having to plot and estimate the slope as described for the original method . ( d ) Flocculent scale method The method used to determine the flocculent scale from 0-5 is not an accurate method as it is subjective . However , it would be difficult to determine MBN value for each of tests carried out in this study , since The Sharp ' s modified Burn ' s number method is not sensi tive enough for weakly flocculent yeasts . There were many sets of tes t conditions in which the strain KM Y42 was weakly flocculent . Thus , this subj ective scale was used to grade quickly the flocculence of yeast after growth in various media , while the MBN values were determined only when the cells were very flocculent . 6 . 3 . 2 Flocculating ability of some lac tose-fermenting yeast strains Thirteen lac tose-fermenting yeast strains , grown in whey permeate (medium 2 . P4 . 6 ) ) , were tes ted for flocculation in the flocculation medium (medium 36 . FM ) . It was found that all strains except KM Y42 were non-flocculent ( table 6 . 8 ) . The strain KM Y42 was found to be weakly flocculent , forming small floes whi ch did not settle very rapidly . Thus , the 1 5 9 flocculence of KM Y42 in various media was sele c ted for further investigation . Table 6 . 8 Flocculating abi l i ty o f some lac tose -fermenting yeas ts . The yeasts were grown in whey permeate ( medium 2 . P4 . 6 ) They were tes ted in flocculation med ium ( med ium 36 . FM ) yeas t flocculent yeast floccul ent strains scale s t rains scale 1 . CP 2234 NF 0 8 . KM Y 1 1 09 NF 0 I 2 . KL 41 6 N:F 0 9 · KM 7 1 58 NF 0 3 . KL 469 NF 0 1 0 . 1G·I XDRI NF 0 4 . KJI1 1 0022 NF 0 1 1 • llo46 1 . p 1 1 na 1 9 . LYA 1 2 na 1 . P ( TS ) 5 4 na 1 9 . LYA ( TS ) 4 4 na 2 . P4 . 6 1 1 1 20 . LYCa 3 3 2 2 . P4 . 6 ( TS ) 5 , 4 5 , 4 na 2 1 . LSM 0 0 na 3 · PC a 1 1 1 22 . f.ia 0 0 na 3 . PCa ( TS ) 5 5 na 23 . MaCa 0 0 na 4 . py 3 2 0 24 . MaCa 0 0 na 5 . PYM 3 2 0 25 . G 3 3 na 6 . PYMCa 3 3 0 26 . GCa 2 2 na 7 . PX!1 4 4 2 27 . YM 1 1 na 8. PXM5 4 0 na 28 . YMCa 1 1 na g . PMs 4 , 5 3 na 29 . YMA 5 5 na 1 0 . PPe 2 2 na 30 . M 2 1 na 30. M ( TS ) 3 2 na 1 3 . PMo46 4 na 1 32 . MCa 4 1 0 1 6 . PB5 5 5 na 33 . Ms 2 2 na 30 . Me 1 0 na The cel ls grown in media 1 9 . LYA , 21 . LSM , 23 . MaCa and MaCa4 showed poor flocculation in the test ing media . 168 Thus , the f lo cculation medium (36 . FM) was the mos t suitable medium for comparison o f f lo cculence o f KM Y42 cells grown on various media . 6 . 3 . 1 2 Floc cula tion of KM Y42 grown as mixed cul ture wi th either C C39 o r SC 1 46 in mixed substrata Thes e experiments were int ended to s tu dy the flocculat ion o f mixed cul tures of KM Y42 + CC39 , and KM Y42 + SC 1 46 when grown in whey permea t e enriched wi th molasses ( 1 3 . PMo46 ) . The mixed culture of KM Y42 + CC39 was more flocculent than the mixed culture of KM Y42 + SC 1 46 , al though bo th miied cul tures were very flocculent ( table 6 . 1 7 ) . Table 6 . 1 7 Flocc ulation of KM Y42 grown as mixed cul ture w i th S. cePeVisiae in mixed substra t a growth KH Y42 36 . FM 1 3 . Pli'Io46 med ium and sca l e MBN* scale MBN* 1 3 . PJI!o46 CC39 5 1 34 5 1 74 " CC39 4 33 4 43 1 1 SC 1 46 5 50 5 na 1 1 SC 1 46 4 29 4 30 1 5 . PB SC 1 46 5 na na na The observed flocculence of both mixed cultures , as measu red in the MBN* , was d iffe rent in the growth med ium compared with the flocculation med ium , whi l e the flocculen t scale values were the same . Us ing strain CC39 , large d iffe rences i n MBN* values were obs e rved between tests carried out unde r apparent ly iden t ical cond i t i ons . No explana t ion could be found for this behaviour . 6 . 3 . 1 3 Flocculat ion o f s t rains CC39 and SC 1 46 Yeas t s t rain CC39 was found to be more flocculen t than yeas t s train S C 1 46 ( table 6 . 1 8 ) . The observed flo cculence o f both yeas t s t rains , as measured in the MBN* , was d ifferent ln the g rowth medium compared wi th the flocculation med ium . I t should be noted here tha t 169 the MBN* values for cells grown in mal t extract broth ( 30 . M ) that were tes ted in the flocculation medium are standard MBN values since the tests were carried out under standard conditions . Both yeasts generally displayed good flocculence in molasses ( 1 8 . Mo ) and in mal t extract broth ( 30 . M and 31 . M* ) wi th some variations observed when different batches of cells were used . Yeast strain CC39 showed bet ter flocculation if the inoculum were grown in either mal t extract bro th o r yeast-malt extract broth before inoculation into molasses ( 1 8 . Mo ) . Thus , strain SC 1 46 was only moderately flocculent while strain CC39 was extremely flocculent . Table 6 . 1 8 Flocculation of strains CC39 and SC 1 46 CC39 se 1 46 growth pH . flocculence in pH . flocculence in 1. 1. medium 36 . FM growth medium 36 . FM growth scale MBN* scale MBN* scale MBN* scale 1 8 . Mo ( M ) 5 . 0 4 46 5 1 20 1 8 . Mo ( M ) 5 . 0 4 4 1 4 43 1 8 . Mo ( Mo ) 5 . 0 4 44 1 0 4 . 9 1 8 . Mo ( Mo ) 5 . 0 1 8 . Mo ( YM ) 5 . 0 5 1 54 4 40 30 . M 5 . 0 5 1 72 5 1 87 4 . 5 3 1 . M 5 0 1 4 58 4 53 5 . 2 4 · 9 5 . 1 ( M ) inoculum grown in mal t extract bro th 3 30 3 25 4 6 1 4 na 3 52 2 3 1 (Mo ) inoculum grown i n molasses solution ( 1 00 g/1 sucrose ) (YM) inoculum grown in yeast malt extract broth 6 . 3 . 1 4 Flocculation curves 4 3 na 3 4 2 medium MBN* 56 3 1 na na 47 1 6 Sample flocculation curves are given in fig . 6 . 1 . Yeas t s train CC39 settled to a small volume (2 ml ) in 5 minutes ( fig . 6 . 1 a ) . The moderately flocculent strain SC 1 46 showed a slower reduction in volume ...-1 15 8 "' � ;L; :;::l 6 .....:1 0 :> 0 4 z H .....:1 f-< � 2 Vl rg "' 8 � � 6 ....:1 0 :> 0 4 z H .....:1 f-< � 2 Vl 1 70 ( a) P ure culture , g rown & t es t ed in s tandard media . Grown in malt ext ract b roth ( 30 . M) T es ted in a cetate buffer ( 36 .FM) �-=-==s_:: 146 , MEN = �2 -=--- 10 ') (\ 4- V C C 3 9 , MEN 172 3 0 40 50 ( b ) Pure culture , grown & test ed in their growth med ia . KM Y42 (TS ) in wh ey permeate ( 2 . P 4 . 6 ) CC 39 & S C 146 in molas s es ( 100 g/ 1 s ucrose) ( l8 . Mo) -- _ s e 146 , MEN>"= 56 - - - - - -- -- -- -- -- -- -- MEN*= 1 20 ; KF Y42 ( TS ) , MEN*= 8 8 ( c) Mixed culture , grown & t es t ed in whey permeat e + molas s es medium no . l 3 . PMo46 20 30 40 50 SETTLING TIME , minut es Fig . 6 . 1 F locculat ion of �f Y42 ( TS ) , SC146 , and CC39 grown and tes ted in d i fferent media . 1 7 1 to 5 . 1 ml in ? minutes . SC 1 46 , respect ively , when The MEN values we re 1 72 and 52 fo r CC39 and grown and tested in s tanda rd med ium ( A modified Burn ' s number method i s given i n descript ion of Sharp ' s sec tion 2 . 5 . 1 ( c ) ) . Strain CC39 showed less floc c ulence ( MEN 1 20 ) when i t was grown and tes ted for flocculation in molasses ( fi g . 6 . 1 b ) . The volume reduced to 3 . 1 ml after 5 minutes . St rain KM Y42 ( TS ) gave an MEN value of 88 when grown and tes ted in whey permeate . The yeas t s e t t led volume was 4 . 2 ml after 5 minute s . St rain SC 1 46 showed less flocculence i n molasses giving an MEN* value o f 56 . Its volume reduced to 7 . 0 ml after 5 minutes . A mixed culture of s t rains , CC39 and KM Y42 , when grown and tes ted in whey permeate enriched wi th molasses gave a MEN* value o f 1 74 . The yeas t vo lume red uced to 3 ml in 5 minutes . In comparison , a m1xed culture o f KM Y42 and SC 1 46 , when grown and tes t ed in a similar med ium gave a MEN* value of only 30 . The yeast volume reduced to 9 . 5 ml after 5 minutes . 6 . 4 DISCUSSION 6 . 4 . 1 Flocculati on of K. marxianus Y42 ( a ) Ini tial inves t igation During ini t ia l inves t igation , this yeas t was not flo c cu lent when grown in whey permeate . Further investigations showed that it floc cu lated strongly when grown in malt extract broth but cel l s from thi s med ium flocculated rather poorly when tested in the flocculation medium . The add i t ion of calcium chloride to mal t ex trac t bro th improved flocculence s light ly . The same behaviour was observed for cel l s grown in malt extra c t syrup . Mal t extrac t powder by i tself was not a very good growth medium fo r thi s yeas t , and i t flocculated poorly . I f lac tose were added to the spent mal t extrac t bro th , growth and yeas t floccu lence were found to be poor . These observations would tend to ind icate that good flocculation of KM Y42 is related to rich med ia which contain many nut ri ents impo rtant fo r good growth . ( b ) Membrane fi l t ration Membrane fi l t rat ion cont ributed toward the unders tanding o f the flo cculent behaviour of this yeas t because the yeas t grew and flo c c ulated poorly in those med ia which had been 1 7 2 membrane fil t ered afte r auto c laving . This ope ration may extrac t some nutrients which are impo rtant for good growth and floc c ula tion . Al t ernatively , the suspended parti c les , which would be removed by fi l tration could provide sites to trap cel l s during fe rmentation and thus eventual ly form floes . Thus , the flocculence of KM Y42 was affe c t ed by fil tra tion of the med ia in which it original ly showed good flo c cu l ence . This effec t was probably a result of the removal of nut rien t s and precipi ta ted particles tha t may be required for flocculation . ( c ) The add i ti on of yeast and mal t ex t ra c t b ro ths to whey permeat e The add i t ion of yeast and mal t ex t ra c t to whey permeat e res u l t ed in good growth and moderate floccu lation o f KM Y42 in a number o f media inc luding the flocculation medium . The add i tion of calc ium chloride did no t resu l t in bet t e r flocculation than tha t of cel l s grown in whey permeate enriched with yeas t -mal t ext rac t bro th . Flocculen c e o f cells grown in whey permeate supplemented wi th either mal t ex t rac t broth o r mal t extract syrup exhibi ted a reasonably s table flocculence when tes ted in the flo cculat ion med ium compared with those grown and flocculated in mal t ex trac t bro th . The lat t e r cel l s flocculated poorly in the flocculat ion med ium . (d ) The addition o f peptone , urea and diammonium hyd rogen phosphat e I t was consi dered tha t pep tone , which is general ly added to the formulation of mal t extra c t bro th , may have made some contribution to the flocculen ce of this yeas t . I ts add i t ion to the permeat e resul ted in weak flo c cu lat ion bo th during fermentation and when tes ted in the flo c cu lation medium . The add i t ion of urea and d iammonium hyd rogenphospha te to whey permeat e produced cel ls of KM Y42 wi th similar flocculence to tho se grown in whey permea te w i th add ed peptone . ( e ) Lac tose , glucose or mal tose as a carbon source The growth and the yield of cel l s in these media were found to be poo r due to the low buffering capac i ty of these media . The final pH ( approximately 3 ) was probably too ac id i c for good growth and flocculence of K. marxianus (Helm et a l 1 953 ) . In general , c e l ls cul t ivated in ei ther glucose or lac tose as the so le sugar source were found to have be t ter flocculence than tho se grown in mal tose . Thus , glucose and lac tose are probably bet ter substrates for produc t ion of flocculent yeas t cells than mal tose . I t should be not ed , however , tha t acidity of the med ia could 1 73 have exerted a great influenc e on the flocculence . ( f ) Enriched whey permeat e Flocculence of KN Y42 grown i n whey permeate with added sucrose , mal t ex t rac t , yeas t ex tra c t and a few other nut rients was very good . When these cel l s were tes t ed 1n flocculation med ium and spent whey permeate , flocculence remained good . These c e l l s formed more s table floes than did the cel ls grown in mal t extrac t bro th and YM bro th . Thi s i s an advantage since these c e l ls would be abl e t o tolerate changes in med ium conditions . Growth and floc cu lence in whey permeate supplemented with molasses was good . I t should be no ted , howeve r , tha t these were rich med ia containing nutrients whi ch contributed to good yeas t growth . Thus , a lthough flocculation was improved in these media , the effec t s of the presence o f mal tose and sucrose in whey permeate , on yeas t floc culation could no t be c learly isolated because of the presence o f some o ther nut ri ents . In cont ras t , when whey permeate was suppl ement ed with yeas t-mal t ex tract and calc ium a s well a s molasses floccu lence decreased . Th1s resul t could not be explained , although it could be an erroneous resul t in that the dark colour of the medium made flocculence diffi cu l t t o observe . ( g ) The add i tion of flocculation aids to the growth med ia The addition of a luminium sul phat e or calcium chloride was found to improve floc culation in some med ia but this effe c t was not as pro found as the influence of the med ium composi tion . That is , the effect of the carbon source and non-specific growth fac tors source were mo re significant . The flocculation improvement observed aft er the add i tion o f calc ium t o yeast or mal t ext ract was probably because these media were deficient i n c a l c i um ini tially . Whey permeat e con tains suffic i ent quant i ty of calc ium such that further add it ion of calcium caused no improvement to flocculation . Thus , the add i t ion o f cal cium as a floccula t ion aid provided no improvement on the floc culation of KM Y42 grown i n whey permeate b u t improved flocculation when the bas e med ia were yeas t or mal t ext rac t . Aluminium , o n the o ther hand , improved the fl o cculation in yeast-mal t extract med ium . ( h ) Med ium p H The pH did not affect flocculence wi thin the range used for most med ia ( pH 4 . 6 5 . 0 ) but acidic pH was found to be associated wi th poor flocculence . The acidic pH ( 3 ) occurred because 1 7 4 of the poor buffering capacity o f these media . The poor flocculence a t low med ium pH probably occurred a s a res u l t o f the effe c t o f the low pH on yea s t growth . ( i ) Subculture of KM Y42 The o riginal s train KM Y42 was not as floc culent as the tower subcul ture ( KM Y42 ( TS ) ) when grown in whey permea t e or whey permea te supplemented with lac tose ( 1 00 g/1 ) . Finally , whey permeat e supplemen ted wi th malt ext rac t syrup was the bes t med ium for producing flocculent KM Y42 . The towe r fermenter subcul ture KM Y42 ( TS ) was sufficien t ly flocculen t to be used i n the tower fermenter even though this abi l i ty varied considerably . Since none of the other lac tose ferment ing yeasts tested was flocculen t , �1 Y42 was the only choice of flocculent yeas t available fo r use in the tower fermen tation of whey permea t e . 6 . 4 . 2 Flocculat ion o f CC39 and SC 1 46 grown as pure or mixed cul tures wi th KM Y42 ( a ) SC 1 46 The S. cerevisiae s t rain SC 1 46 was found to be modera te ly floc culent . The mixed culture o f this yeas t s t raln and KM Y42 could be used in the tower fermen ta t i on s ince the MBN* obtained was in the same range as that obtained for SC 1 46 alone , and this yeas t had been used successfully in the towe r fermenter ( Append i x C ) . The f l o e s of the mixed cul t ure were no t spherical , but were serra ted and sma l l . ( b ) CC39 The yeast stra in CC39 was found to be an ext remely flocculent yeast in mos t med ia tes ted inc lud ing mal t extract bro th . The floes were s table when tes t ed in the flocculat ion medium and in the growth medium . It was also very flocculen t when grown in a mi xed culture with KM Y42 in whey permeate supplemented wi th molasses . The flocculence of the mixed cul ture was also stable in the flocculat ion med ium . When CC39 was grown as a mixed cul ture wi th K M Y42 in whey permeate enriched with molas ses , the flo cculence was bet t e r than that of the m ixed cul ture of KM Y42 + SC 1 46 . I t was cons idered tha t mixed cul tu re of KM Y42 + CC39 would be more sui table for tower fermenta t i on o f whey permeat e/molasses m ixtures since thi s mixed cul t ure was very floccu lent . Thi s mixed c u l ture would be able to remain in the towe r fermen t e r a t a higher med ium feed ra t e than a mixed c u l ture o f KM Y42 + SC 1 46 . ( c ) The effec t o f the ino culum-growth med ium In some cases d u r i n g flocculation tes t s , c e l ls were prepared from the same med ium o n 1 75 differen t occasions and the resul ting floccul ence behaviour obse rved was variable . It was observ ed tha t the flocculence o f CC39 was affected by the med ium in which the inoculum was prepared . F locculence observed for cel l s grown in molasses med ium ( 1 8 . Mo ) was bet t e r i f t h e inoculum were grown up in ei ther mal t extrac t bro th ( 30 . M ) o r yeas t-malt ex t ract bro th ( 27 . YM ) rather than in mo lasses solution ( 1 8 . Mo ) . This was a good indication o f the influence that the growth environment , in this case substrate and nutrients , has o n yeas t flocculation . 6 . 5 CONCLUSIONS Ko marxianus Y42 was the only flo cculent lac tose-fermenting yeas t iden t i fied . I t was found to be moderately flocculent when grown in media which support good growth but showed poor flocculence when grown in ac id i c med ia or med ia which did no t support good growth . I t showed poo r flocculence when grown in whey permea te and showed modera te flocculence when grown in whey permeate enriched wi th molasses . However , i ts subcu l tu re showed good flocculence in whey permeate . 6 . 6 SUMMARY 1 . The hybrid yeast K. marx1/anus Y42 ( KM Y42 ) was tr1e only flocculent yeast found amongs t the lactose-fermenting yeasts t es ted . 2 . Flocculence o f KM Y42 grown in aci d i c media o r med ia whic h d i d not support good growth was poor bu t the yeast was mode rate ly flocculent in many med ia which suppo rted good growth . 3 . The flo c culence o f KM Y42 ce l l s grown in whey permeat e supplemented with yeast ex tract o r mal t ext rac t o r mal t ex t rac t syrup , to which ammonium sulphate , d iammonium hydrogensulpha te , and calcium chloride were added , was more stable than tha t o f cells grown i n mal t extrac t or yeast-ma l t extract bro th , even though the cel l s grown i n th� las t two med ia flocculated read i ly . 4 . The c e l l s o f KM Y42 grown i n whey permeate we re weakly flocculent . The subcul ture o f KM Y42 taken from the tower fermen t er was more flocculent than the parent s trai n . Thi s subcul ture was ini tially prepared in whey permeate supplement ed wi th mal t ex trac t syrup . 1 7 6 5 . S . cerevisiae CC39 was an extremely flocculen t yeas t . I t was more flocculent than the s t rains SC 1 46 and KM Y42 . 6 . The mixed cul ture of KM Y42 + enriched with molasses ( lac tose to C C 39 grown in whey permea te sucrose ratio of 40 : 60 ) was mo re flocculen t than the mixed culture of K M Y42 + SC 1 46 , but not as flocculen t as the s train CC39 grown in pure cul t u re . CHAPTER 7 MEDIUM OPTIMIZATION AND CULTURE IMPROVEMENT 7 . 1 INTRODUCTION Med ium op timizat ion and batch fermentat ions were carried out using a d i f ferent s train of K. marxianus from the one used for tower fermenta tion o f whey permeat e . Thi s resul ted in further inves t igations to improve e thanol tolerance of K. marxianus� to selec t for an isolate which showed no diauxic behaviour and to mutate K. marxianus using UV radiat ion in order to isolate a mutant which could no t util ize sucrose . The resul t ing improved s t rain could then be used to provide gene tic material s to produce a floc culent K. marxianus mutant which could tolerate high concentrat ions of e thanol , and was d iauxiE: -negat ive or sucrose nega t ive . The s t r a in KH Y4 2 was no t used in the mutat ion s tudies becaus e i t was no t ava i l ab l e a t t h e t ime that thes e experiments were performed . 7 . 2 MEDIUM OPTIMI ZATION A factorial experiment was performed using a 3 3 des ign consis ting o f 1 4 runs (Webb 1 9 7 1 ) . Whey permeate (40g/l lacto se) of volume 2 5 0ml was used in 500ml shake flasks (agitated at 5 rpm) and the yeast used was K. marxianus UCD FST 7 1 5 8 . Three variables were invest igated : ( NH4 ) 2 so4 , K2HP04 and yeast extrac t . A summary o f fermentation resul t s is given in appendix B . l . There was very l i t t le not iceable d i f ference b e tween each nutrient condi t ion . S tatistical analysis o f the resul t s gave the following correlat ion : E = 1 . 7 8 - . 0 3N - . 02 1K - . 024Y + . 03 7 5NK + . 035KY + . 02 5NKY Where E i s e thanol concentrat ion . The t-ratio o f each nutr ient i s given in tab le 7 . 1 . Tab l e 7 . 1 : t Rat io o f p ar ame t e r s N= (NH4 ) 2 so4 , K=K2HP04 , Y =Yeast extrac t . The null hypo thesis was to consider whether N , K , and Y have any ef fec t on e thanol produc t ion . Parameter t-ra t io N K - . 64 y - . 7 4 NK l . 03 KY . 9 6 t-ratios a t 90 & 9 5 % confidence l evel for 7 d e g r e e s o f freedom a r e 1 . 9 and 2 . 3 7 respec t ively . Regressed again s t e thanol concentrat ions a t 3 0 h . Fur ther regre s s ion by d ropping progres sively the pa rameter with t h e lowes t t-ratio did not yield a t-rat io greater than the 90% confidence level . 1 77 1 7 8 At zero nutr ient addi t ion level , the e thanol concentration �as 1 9 g/ J and the corresponding yield was 89% , a f ter 30 hours which in this f ermenta t ion tes t was s l igh t ly greater than for o ther nutrien t addi tion l evels . The shake f lasks were agitated b ecause it was found that s t ill cul ture required too l ong a fermentat ion t ime and par t ially aerobic f ermentation resul t ed in rapid ethanol produc tion ( Burges s & Kelly 1 9 7 4 ) . Thus , nu t rient supplementation o f whey permeate was no t neces sary for e thanol fermentation by this K. marxianus s t rain . 7 . 3 ISOLATION OF AN ETHANOL-TOLERANT K. marxianus 7 . 3 . 1 Prel iminary ba tch fermentation ( a ) Whey permeate containing 40 g/1 lactose I t was found that K. marxianus UCD FST 7 1 5 8 was abl e to ferment 4 0 g / 1 lactose in whey permeate (using 1 0 l i t res med ium vo lume ) to complet ion in 1 6 hours ( f ig . 7 o l ) . The res idual lactose was 3 g / 1 ( 9 3% lac tose u til izat ion ) and 1 8 g / 1 ethanol was produced in that t ime giving a yield o f 9 2% e thano l on lactose utilized . The resul t s ob tained were in good agreement with tho se obtained by o ther workers . Burges s & Kel ly ( 1 9 7 9 ) reported complete u ti li zation o f 5 0 g / 1 lactose i n cheddar cheese whey permeate b y two s trains o f K. marxianus ( NRRL-Y- 1 1 0 9 and CBS 5 7 95 ) in 1 8 and 1 2 hours , respec t ively ( 2 8°C , in 2 50 ml shake f lask) . Industrial batch fermentat ion t ime o f 1 6 hours was repor ted for the fermentation o f depro teinated whey containing 44 g / 1 lac tose (Howell & Tichbon 1 9 8 0 ) . ( b ) Whey permeate conta ining 100 g /1 lac tose As a resul t o f the previous fermentation , a further 1 0 l i tres batch fermentat ion was carried out us ing the same yeast strain but the whey p e rme a t e c o n t a ined a grea t e r lac tose concentrat ion o f l OO g / 1 . K. marxianus had not u til ized all lactose available a f ter 4 8 hours ( f ig . 7 . 2 ) Lactose was reduced from l OO to 1 5 g /1 during this time but onl y a small amount of lac tose ( 1 . 6 g / 1 ) was util i zed between 3 5 and 4 8 hour s . At 3 5 hours , 2 9 g / ] o f e thanol had been p roduced and then increased t o 3 3 g / l a f ter � 8 hours . The f inal concen t rat i on o f e thanol was produc ed a f t e r 6 9 % o f lac tose was u t il ized result ing in 9 2 % e thanol yield based o n lac tosP u til i zed . The resul t above showed that when the e thano l concentra t ion reached 30 g / 1 , the fermentat ion activity of K. marxianus UCD FST 7 1 5 8 reduced considerab ly . This was considered to have b een a resu l t o f ethano l inhib it ion . I t has b een repor ted tha t 32 g / 1 ethanol could reduce the 0 rl X ::r: 0.. 30 ,....., ....___ 01) ,...:1 0 � ;:r:: H � 1 79 Whey permea te ( 4 0 g/ 1 lactose) 'a.. � K . marxianus UCD FST 7 1 5 8 'o....o-o-�o.-0 --o--o--o- -{)- -- -- - -O o--D·_o_·....o- ·o-0-·-u· -o-· o ·- o-· -· - ·--· -.o ):]' 0 . .. .cr to tal cell -+ • �·-.-. -.-·-- - --.-- - - - -· / '� � ethanol " . �' // \ lactose -+ ........ \ _......- .... .. ......__ -·- ......__ - ·- 10 20 30 FERMENTATION TIME , h 40 so Fig . 7 . 1 Batch f ermenta t ion o f whey permea te ( 40 g / 1 lac tose ) . rl ........_ Oil 30 10 10 20 ' ' ............... _ _ - - --4 Whey p ermea te ( lOO g/ J lac t o s e ) K. marxianus UCD FST 7 1 5 8 30 40 so FERMENTAT ION T IME , h Fig . 7 . 2 Batch fermentation o f whe y permea t e ( 1 00 g / 1 lacto se ) . ,_, ......._ 60 01) � CfJ 0 H u < ,...:1 40 80 ---.. 60 01) Ul 0 r-1 6 1 0 9 3 "' r_) r-' 1 0 8 � ::t: '---< z w c_;. ;z 0 '� 1 80 ac t ivity o f B-galactos idase in K. marxianus NRRL Y- 1 1 09 by 7 7 % . The inab ili ty o f some K. marxianus s trains to ferment concentrated lactose solut ions o f greater than 100 g/1 has been reported by a numb er of workers (Yoo 1 9 74 ; O ' Leary et al 1 9 7 7 ; Gawel & Kosikowski 1 9 7 8 ) . There have been reports o f K. marxianus st rains which withstand greater ethanol concentrat ions than 3 0 g / 1 during fermentation , but the f e rmenta t ion condi t ions used were partially aerob ic ( Burges s & Kelly 1 9 7 9 ; Moul in et al 1 980 ) . ( c ) Whey permeate enriched with molasses I t was intended that whey permeate enriched with molasses would be used as a feed medium for tower fermenta t ion . A batch f ermentation s tudy was necessary to observe the behaviour of K. marxianus in mixed cul ture with S . cerevisiae when the total sugar concentrat ion was 1 0 0 g / 1 and to s tudy whether e thanol inhibition would occur as found in the previous fermentat ion ( sect . 7 . 3 . 1 b ) . One 1 0 1 batch fermentation was carried out us ing a lac tose to sucrose ratio of 2 0 : 80 g / 1 ( fig . 7 . 3 a ) . I t was found that l i t t le lactose was u t i l ized . Af ter 24 h , lactose had been reduced from 20 to 16 g / 1 ( 2 1 % lactose u tilization) whereas sucro se was reduced f rom 7 5 t o 1 g / 1 ( 99% utilization ) . At this t ime , 3 5 g / 1 of e thanol had been produced . As the fermentat ion t ime increased to 50 h , lactose was r educed further to 1 1 g / 1 ( 4 5% lactose u t il ization ) . The resul t ing total sugar concentra t ion reduced from 95 to 1 . 2 g/1 ( 8 7 % sugar utilization) . The res idual sugar was mainly lactose ( 9 3% lactose ) . By this t ime 38 g/1 e thanol had been produced ( 8 7 % ethanol yield on sugar utilized ) . S imilar results were obtained for ano ther fermentat ion us ing a lac tose to sucrose ra t io o f 1 0 : 90 g / 1 ( f ig . 7 . 3 ) . In this fermenta t ion , lactose dec reased from 1 0 to 7 . 3 g / 1 ( 30% lac tose utilization) whi l e sucrose reduced f rom 84 to 0 . 9 g / 1 ( 99 % sucro se ut i l ization) to produce 38 g / 1 e thanol , a f ter 24 h . As the fermentat ion t ime increased to 4 8 h , lac tose was reduced further to 5 . 5 g/ 1 ( 4 7 % lactose ut il ization) and the total sugar reduced f rom 94 to 6 . 3 g / 1 ( 9 3 % sugar u t i l i zation) . By this t ime , 3 9 g / 1 e thanol had been produced . This was 8 3 % yield on sugar utilized . The resul ts presented here indicated that when sucrose and lac tose were availabl e s imultaneously , sucrose would be util ized f irst at a very much faster rate than lactose . Lac tose util izat ion did not occur until a l l sucrose 1vas utilized , but by this t ime the ethano l concentrat ion was greater than 30 g / 1 which could reduce S-galac tosidase ac t ivity cons iderably (Wendorf et al 1 9 70a) . Henc e , even though lac tose utilizat ion occurred , 50 ( a ) o-()0---a... - e:. ... e:. _,_ pH o 'O-o_Q__o -0- ' t:::.. ' 1 8 1 100 o--o-o- - -- --o ;::; 4 0 'f.otal s ugar -+ ', +- e thanol _ - -• 80 X :::r:: CL rl .._,_ bf) � ..-4 0 z � f--j !:.:) 0 rl X ;:c CL r-1 .._,_ 01) ....:< 0 z - ---0- - - -() ( lactos e : s ucros e 10 : 90 g/ 1) / �\ - - -4---.t. � _.__ .... _ _ �-�::...4.- 7 ... - �actose -+- ,-- - - -� -=- --=� � \7 . . 10 20 30 40 FERMENTATION TIME , h '50 60 40 20 !:.:) [fJ 0 � u j l O O � (.fl ,__l �- � 7.,.__.,_ .... .._..... - -11'--- _.., _ __,_ �- __,..._ - -� -rr- - - -\' � ...., _ _ -:..... -� '\ .. �eros e ;, - - ··y 100 ·-e ....._ 80 109 rl rl CV u . ;z; 0 >--" 60 r-1 ....._ 10 8 I:-· ;:2 bO 1:--< z � P::: '-' < z C) 0 ;::J "--' (f) 40 10 7 :--1 """' w u """ bO 0 rl ..__., I / 10 20 10 ° _ /+-ethanol - 10 20 30 40 FERMENTATION TINE , h Fig . 7 . 7 1 0 l. i tre batch fermantat ion of whey permeate enriched with mo lasses by KM10D l0 ( lactose : sucrose 40 : 60 g / 1 ) . Thus , only 6 2% o f the total sugar was u til i zed and mos t o f this was due to sucrose u tilization . E thano l produced dur ing this t ime was 3 3 g / 1 , a yield on sugar u t i lized o f only 9 3% . The resul t s ind icates that KM1 0D 1 0 reta ined the d iauxic behaviour o f the parent yeas t , u t il iz ing sucrose before lac tose . I t was not ab l e t o ful ly u t i l i ze 1 0 0 g / 1 o f mixed sucrose and lac tose wi thin 48 hours . I t could utilize either one o f these sugars in a s ingle sub s trate fermentation us ing the same sugar concentration wi thin 48 hour s . 7 . 3 . 6 Conc lus ions K. marxianus UCD FST 7 1 58 could ferment whey permeate without any add i t ional nutrient in batch fermentat ion . In a 1 0 l itre batch fermentation o f whey permeate i t required 16 hours to u t i l ize 9 3% of 4 0 g/1 lac tose produc ing 1 8 g / 1 e thano l . Thi s was 9 2 % yield o f e thanol on lac tose util i zed . I t was , however , inhibi ted b y e thanol c oncentrat ion greater than 30 g / 1 and exhibi ted d iauxic behaviour when it was g iven a choice o f both lactose and sucrose at the same t ime . Sucrose was always 1 9 2 uti l i zed f ir s t . Thi s was fo l l owed by poor l a c tose up take . A cul ture improvement programme was success ful in iso la t ing a more e thanol -tolerant s train , KMl OD l O . Compar ison o f the improved yeas t cul ture with i t s parent and o ther lactose- fermen t ing yeas t s showed i t to be a rapid lactose fermenter and more rapid e thanol producer . The isolate yeas t s tr a in , KMlOD l O , was found to have a s table e thanol tolerance . However , in the presence of two sugars , in this case l ac tose and sucrose , i t retained the d iauxic behaviour o f i ts parent , u t i l i z ing sucro se before lac tose . This was an undes irable charac teristic , because in order for l actose and sucrose to be rapidly converted to e thano l in in the tower fermen ter , the two sugars mus t b e met abol ized s imul taneously . There was a need t o a t t empt t o isolate a mutant which showed no preference for sugar ( d iauxie-negat ive) or a mut an t which u t i l ized only lactose ( sucro se-negat ive ) , s o that it could be used in mixed sub s t ra t e , mixed cul t ure , for produc t ion o f e thanol in economic concentrations f rom supplemented whey permeate . 1 9 3 7 . 4 AN ATTEMPT TO ISOLATE DIAUXIE-NEGATIVE K. mar•xianus s trains 7 . 4 . 1 Introduc t ion In the batch fermentation o f mixed substrate of sucrose and lactose by K. marxianus UCD FST 7 1 58 and KM1 0D 1 0 ( s ec t . 7 . 3 . 1 & 7 . J . 5 ) , sucrose was f ound to r epress lactose utilization . This is an undesirabl e charac teris t ic in tower f ermentat ion . Thus , the aim o f this experiment was to carry out an isolat ion exper iment for diauxie-negat ive K. marxianus s trains , us ing D-glucosamine ( DGA) as a gratuitous cataboli te repressor . DGA has b een used by a number o f inves tigators as a glucose analogue in carbohydrate me tabol i sm o f yeas t s ( Fur st & Michel s 1 9 7 7 ; Michel s & Romanowski 1 980 ) . DGA at a concentration o f 0 . 5 g / 1 was found to r epress comple t ely the growth o f a s train o f S. cerevisiae ( El l io t & Bal l 1 9 7 3 ) . Fur ther s tudy found that DGA repressed the respirat ion rate more rapidly than glucose , could repress the level of cytochrome oxidase to the same l evel as glucos e and repressed f ermenta t ion of mal tose and galac tose , but no t sucrose . In all , DGA produced a repre s sed s tate very s imilar to glucose in many aspec t s , but d id not affect growth on gluco se ( Furs t & Miche ls 1 9 7 7 ) . DGA has no general ized growth r epressive e f f e c t s a s does 2-deoxygluco se , ano ther glucose analogue that has been used f o r iso lat ion o f mutants i n S. cerevisiae insensi t ive t o glucose rep ression ( Zimmerman & S cheel 1 9 7 7 ) . Thus , the presence o f DGA in a med ium wh ich contained l ac to s e as t he so le sugar source , would inhib i t growt h o f K. marxianus only i f D GA could repress the synthesis of the enzyme ( s ) required for lactose utiliza t i on . The growth o f K. marxianus in the lac tose med ium with added DGA would be af fec ted b y the concentrat ion o f DGA . When the l evel o f DGA was greater than a minimum l imit ing level , there would b e reduced growth . Tho s e colonies that could grow a t this l evel o f DGA or greater should be c e l l s that could u t il iz e b o t h glucose and lac tose s imul taneousl y (Dunn 1 9 8 1 ) . They are no t repressed by the analogue , DGA , hence , the s e are the requi red d iauxie-nega t ive K. marxianus strains . 7 . 4 . 2 Isolat ion experiment In this expe rimen t , two sets o f lactose and glucose agar w i t h DGA added at d i f ferent concentra t i ons were prepared . One set each of lactose and glucose agar was plated with 0 . 1 ml of I01 10D l 0 yea s t suspens ion , \..rhile the remaining se ts were s t reaked with KMlOD l O . Af ter incuba t ion , 1 94 the growth in the lac tose agar plates was compared with the control glucose agar plates . KM10D 10 cells which had diauxic behaviour should no t grow in the presence of DGA in lactose agar plates , but grow in glucose agar plates . KM10D10 cells which had no diauxic behaviour should grow in both agars . Colonies with this behaviour were isolated and inves t igated further . Two isolation attempts were carried out . In the first attempt DGA level s used were . 1 , . 5 , 1 , 2 , and 5 g/1 based on the finding that 0 . 1 5 g/1 DGA could cause inhibi t ion and 1 . 5 g / 1 DGA had an e f fect equiva­ lent to 5 g/1 glucose on the growth o f S. cerevisiae ( Furs t and Michel s 1 9 7 7 ) . The KM10D10 yeas t culture used was grown in whey permeate for 24 hours and received no further treatment prior to inoculation . (a ) Firs t attempt The re sults from the firs t attempt are given in table 7 . 6 while f ig. 7 . 8 provides a visual representation o f the e f fect of DGA on growth . Table 7 . 6 Observation of growth o f KM1 0D 10 in glucose and lactose agars with added DGA ( First a t tempt , c f . fig . 7 . 8 ) DGA growth after 48 h ( 2 5 °C) in g/1 glucose agar lactose agar 0 . 1 Moderate growth , large Moderate growth , large colonies ( - . 5rnm) colonies ( . 5- . 6mm) 0 . 5 Moderate growth , colony s ize Moderate growth , co lony s ize . 2- . 4rnm - . 2- . 3rnm 1 Moderate growth , colony size Moderate growth , colony size . 2- . 3rnm - . 5- . 8rnm 2 Poorer growth , smaller Poorer growth , smaller colonies - . l - . 2mm colonies - . lmm 5 Moderate growth , colony size Poorer growth , smaller - . 5- . 6mm colonies - . 1- . 2mm 1 9 5 Fig . 7 . 8 Growth o f KMlOD lO in lactose and glucose agars which contained dif ferent levels of D-glucosamine . (cf . table 7 . 5 ) 1 9 6 KM1 0D l 0 grew on l ac tose agar which contained DGA u p t o 5 g/ l . Bo th plat ed and s treaked agar plates showed s imilar growth patterns to those ob tained for the c orresponding control g lucose agar plates . I t was concluded that DGA c oncentrat ions up to 5 g / 1 did no t repres s growth o f KM1 0D 1 0 . ( b ) Second a t t empt The concentrat ion o f DGA was inc reased to 1 0 and 2 0 g / 1 . The resul ts are g iven in tab le 7 . 7 . The growth in agar plates containing DGA up to 5 g/ 1 was s imilar to tho se descr ibed previously ( tabl e 7 . 6 and fig . 7 . 8 ) . There was very l i t t l e growth in p l ates with 1 0 & 20 g / 1 DGA both on glucose and lactose agar plates . Table 7 . 7 Observat ion o f growth o f KM1 0D 1 0 in glucose and lac tose agars which contained up to 20 g/1 DCA ( Second a t t emp t ) DCA growth a f ter 48 h ( 25 ° C ) in g / 1 glucose agar lac tose agar 1 0 Very l i t t le growth Very l it tl e growth , few I isolated colonies . I ! 2 0 Very l i tt l e growth Very l it t l e growth , very few I co l onies At this p o int , i t was considered that a l evel o f DCA a t which the growth of KMlOD l O was repres sed had been reached . The control glucose agar plate s showed the same level of poor growth . Thi s may be an ind icat ion of toxic i ty , ( rather than repres s ion o f enzyme synthesis by DGA) . In order to reso lve this problem , a fermentat ion test was carried ou t , u s ing whey permeate enriched with sucrose . ( c ) Fermentat ion test A c o l ony growing on the 1 0 g /1 DGA lac tose agar plate was used in the t es t . The whey permeate used contained equal concent rat ion of s uc ro s e and l ac to se ( 4 0 : 40 g / 1 ) t o check whether o r no t the yea s t was d iauxie­ nega t ive . The re sul ts from this fermentat ion are g iven in table 7 . 8 , wh ich shows tha t sucrose was u t i l ized f i r s t dur ing the in i tia l 2 4 hours f e rmen­ tat ion " A sma l l amount o f lac t o se was util ized during thi s ini t ial period . The amoun t o f sucrose and lac tose u t i l i zed were 4 3 and 1 2 � / 1 co r re s pond ing t o 96 and 2 4 % u t il i za t ion respe c t ive ly . 1 9 7 Table 7 " 8 Fermentation test o f isolate obtained from lactose agar containing 10 g / 1 DGA . The medium used was whey permeate enriched with 4 0 g/ 1 sucrose ( 1 00 ml med ium) t ime pH X h Cells /ml 0 4 . 9 5 . 0x 1 0 1 0 4 . 9 3 . 0x 1 0 1 7 4 . 5 2 . 9x 1 0 2 4 4 . 2 3 . 7 x 1 0 4 8 4 . 6 S . Ox l O l'IS sucro se used s 4 7 8 8 8 sucro s e g / 1 4 5 3 7 1 3 1 . 9 1 . 1 Ms l actose 6S1 e thanol g / 1 g / 1 g / 1 g / 1 - 50 - - 8 . 7 4 8 2 - 32 4 7 3 . 2 5 . 8 4 3 38 1 2 2 2 4 4 6 . 6 4 4 3 3 Af ter 4 8 h , the utilizations were 9 8 and 8 7 % respec t ively . I n s ec tion 7 . 3 . 5 (a ) , KM10D1 0 was shown to be able to util ize 9 7% of I 1 00 g / 1 l ac tose present in whey p ermeate in 48 hours . In this fermenta tion , there was 9 2 % total sugar utilization a f ter 4 8 hours . This indicates that the presence o f two sugars in the same med ium had delayed the rapid ut ilizat ion of lactose . Comparison with the parent s train c an b e made fur ther . From Table 7 . 11 , in a s imilar fermentation , KM10D l 0 reduced sucrose and lac tose from 4 1 and 4 0 g / 1 to . 8 and 1 . 8 g / 1 in 24 hours . Thi s corresponded to 9 8 and 96% utilization respec t ivel y . This shows tha t the isola t e ob tained here has shown no improvement over the parent s train . Thus , the isolate was no t d iauxie-negat ive . The poor growth on 10 and 20 g / 1 DCA agar plates was a resul t o f DCA inhib i t ion on growt h , no t repress ion o f specific enzyme synthesis . 7 . 4 . 3 The e f fect o f D-glucosamine on growth of K. marxianus I t was apparent from the resul t s of the two iso lat ion at tempt s that DCA did not repress the utilizat ion of lac tose by KM10D 1 0 at l eve l s up to 5 g / l . However , at 10 and 20 g / 1 , growth was inhib ited . The fermentation tes t o f a colony f rom the 1 0 g / 1 DGA lac tose agar plate showed tha t the diauxic behaviour was s till present when the two sugars , suc rose and lactose , were availab le . Sucrose was ut il ized before lac tose . Thus , the cells which grew on lactose agar plates which contained 1 0 and 2 0 g/1 DGA were those that could tolerate DCA rather than d iauxie- nega tive s t rains " 1 98 Other workers , had found that DGA repressed mal to se and galactose fermentation but not sucrose , while glucose repressed fermentation of all three sugars ( Furs t and Michels 1 9 7 7 ) . I t is generally known that sucrose is hydrolyzed a t the cell wall whereas mal tose and lactose are hydrolyzed inside the cel l s ( Sutton & Lampen 1 962 ; Yoo 1 9 74 ) . From these studies , one would expect lac tose to b e repressed in a s imilar way to maltose by DGA since they are both hydrolyzed inside the yeast cells . In this study , lactose was found to be affec ted in a s imilar manner to sucrose in that i t was util ized by KM10D10 in the presence of DGA up to 5 g/1 . I t was pointed out earlier that only 5 g / 1 of DGA was needed to completely repress the growth of a s train of S . cerevisiae . Thus , the inhibit ion of growth on the 1 0 and 20 g/1 DGA lactose and glucose agar plates was probably a result o f toxicity of DGA (due to its very high concentration) and not enzyme synthesis repression . 7 . 4 . 4 Conclusions K. marxianus util ized lac tose in the presence of concentrations of D-glucosamine up to 5 g/ 1 . At higher concentrations (up to 10 and 20 g/1 ) , the growth of K. marxianus was poor . This was considered to be due to the toxicity of D-glucosamine . Diauxie-negative strains were not isolated . 7 . 5 A MUTATION ATTEMPT TO ISOLATE SUCROSE-NEGATIVE K. marxianus 7 . 5 . 1 Introduc tion In the previous sec t ions , it was es tablished that K. marxianus and its mutant KM10D l0 ( sec t . 7 . 3 . 1 , 7 . 3 . 5 , 7 . 4 . 2 ) showed diauxic behaviour in the presence of sucrose and lactose . Sucrose was ut ilized before lactose , prolonging the fermentation t ime of a mixture of these sugars . There was a need to isolate a yeast strain which could utilize both sugars s imultaneously , or utilize lac tose only so that it could be used in conj unc tion with S. cerevisiae to ferment a mixture of sucrose and lactose . An attempt to isolate a diauxie-negative mutant was unsuccessful ( sect . 7 . 4 ) . This experiment investigated the second alternative , which was to isolate a lac tose-fermenting yeas t incapable of utilizing sucrose . The mutant could be used in conjunction with S. cerevisiae to ferment a � � mixture o f lactose and suc rose in the tower fermenter . 1 9 9 7 . 5 . 2 First muta t ion a t t empt In an at tempt to produce a sucrose-negat ive K. marxianus from KMlOD l O , two experiments were carried out , using UV light as mutagen . The f irst at tempt yielded 2 mutants which showed very l i t t le growth in sucro se agar but had a very much better growth in whey agar . Typic al growth compar i son i s given in f i g . 7 . 9 . These were des ignated as FSN l and FSN 2 . FSN l were found to be l arge round cells which tended to collapse a f ter 24 to 48 hours fermentation in l iquid cul ture . F SN 2 had the normal oblong shape o f K. marxianus in a s imilar medium . ( a ) Fermentat ion o f whey permeate by mutant FSN l & 2 A fermentat ion test in whey permeate was carried out to study the abi l i t ies of the mutants t o fermen t lactose and to select the fas tes t lac tose fermenter for tests in the presence o f sucrose and lactos e . The fermenta t ion used lOO ml medium in a shake f lask . Table 7 . 9 shows r esul t s o f the fermentat ion by 1 0 isolate colonies after 48 hours . I solates no . l-4 ( f rom FSN 1 ) , util ized on average 60% of the 7 6 g / 1 lactose available . They gave an average e thanol production of 1 4 g / l . Isolates no . 5- 1 0 ( from FSN 2 ) , util ized on average 2 6% of the lactose available to produce 6 . 8 g / 1 ethanol . For all isolates , The f inal pH was 4 . 5 , a normal value for K. marxianus , while the f inal cell number was about hal f that expec ted from th e parental s t rain . The resul t s showed that FSN 1 isolates could ferment more than twice the amount o f lactose util ized by isolates of FSN 2 , but that this was a poor f ermentat ion performance . I t was d ec ided that the behaviour o f these isolates i n the presence o f both sucro se and lactose should b e invest igated . 200 Fig . 7 . 9 Comparison of growth of possible sucrose-negative mutants in whey and sucrose agars . 201 Table 7 . 9 Fermentation of whey permeate by two possible sucrose­ negative mutants of KM1 0D1 0 (after 48 hours ) . (No . 1-4 were colonies taken from FSN 1 agar plate no . 5- 10 were colonies taken from FSN 2 agar plate) Ini tial pH = 4 . 9 , cell concentration 2 . 5x 1 04cells/ml , lactose colony X t�s1 8lu ethanol number cell s /ml g/1 % g/1 FSN 1 1 8 45 58 13 1 . 4x 107 2 9 . 2x 107 4 3 54 13 3 6 . 3x108 54 70 14 4 1 . 8x10 45 60 15 FSN 2 5 7 16 2 2 6 . 5 9 . 8x 108 6 1 . 2x 10 7 1 9 2 6 7 . 2 7 9 . 4x 108 1 6 2 1 7 . 5 8 1 . 4x108 20 26 6 . 7 9 1 . 4x108 20 2 7 4 . 8 10 1 . 2x10 24 30 5 . 8 6S 1 - lac tose consumed . ( b ) Fermentation of mixed subs trate of lactose and sucrose by 76 g/1 . mutants FSN 1 and 2 . The inocula for these fermentation tes ts were obtained from the previously described fermentation of whey permeate by FSN 1 and 2 ( table 7 . 9 ) . The tests were carried out to study the behaviour of the mutants in mixed substrate of lactose and sucrose . Nomenclature of isolates and fermentation conditions were as shown in table 7 . 8 . The resul ts of the fermentation at 0 and 48 hours are shown in table 7 . 1 0 , Bo th sets of isolates showed poor lactose and sucrose utilization and ethanol product ion . More sucrose than lac tose was utilized by isolates ( 1 , 2 , 3 , 5 , 6 , 7 ) while the remaining isolates u tilized more lac tose than sucrose . Isolates no . 2 and 3 produced the greatest amount of e thanol ( 3 . 1 g / 1 ) , but this was a much lower yield than that expec ted of the parent strain . There was also very lit tle increase in cell number from inocula t ion . This was another indication of poor growth . Isolates no . 4 , 9 , & 10 were selec ted for fermentation comparison with the parent yeast , KM10Dl0 , on the basis of their abil ity to util ize more lac to se than sucrose . 202 Table 7 . 1 0 Fermenta t ion of whey permeate enriched with sucro se b y isola te s FSN 1 & 2 . (Nomenclature as for table 7 . 9 ) 0 hour 48 hour s no pH X s l s s t pH X s l ss s t E 65 1 9 c e l l / m! g / 1 g/ 1 g/ 1 cell /m1 g/ 1 g/ 1 g/ 1 g /1 g/ 1 FSN I I 5 . 4 7 . 7 x ! 0 5 4 2 36 7 8 2 . 7 4 . 4xl 07 4 0 2 8 6 7 0 . 8 22 2 5 . 5 8 . 3x 105 41, 39 83 2 . 7 5 . 2x l 07 4 1 2 1 6 1 3 . 1 3 . 0 J 5 . 5 1 . 2 x 1 06 4 2 3 9 82 2 . 7 S . Ox l 07 4 0 2 0 6 0 3 . 1 2 . 2 !, 5 . 7 l . 2 x 1 0° 4 6 2 2 4 8 3 . 0 3 . 0x 1 07 1 3 1 1 2 3 1 . 1 1 3 FSN 2 5 5 . 7 4 . Ox I 0 '> 21. 2 2 46 2 . 7 4 . S x ! 07 2 2 1 5 3 7 0 0 6 5 . 7 1 . 1 x [ (1 6 2 5 2 9 53 2 . 7 6 . 7x l 07 2 1 2 2 4 3 1 . 1 3 . 4 7 5 . 8 6 . 8x ! 05 2 5 2 2 4 7 2 . 7 6 . ! x i 0 7 2 4 1 9 4 3 0 . 6 ! . 2 8 5 . 5 'l . Ox ! O 'i 4 3 34 82 2 . 7 4 . 4 x l 0 7 38 3 7 7 5 0 4 . 6 9 5 . 5 3 . 0x ! 05 4 3 3 7 80 2 . 7 S . Ox ! 07 36 3 5 7 2 0 . 6 6 . 6 10 5 . 8 3 . 8x i 05 2 4 1 8 4 2 2 . 7 6 . 0x ! 0 7 2 1 l 7 38 0 . 4 3 . 2 �st- total sugnr u t i l i za t ion ; s tu- percentngc to t a l sugar u t il i za t i on 580- pe rcent age sucrose u t il i z a t ion . LlS liS t 5 l u s 9 su g/1 g/ 1 :r. % 8 . 0 1 0 5 22 1 9 2 2 7 48 1 9 2 2 5 4 9 1 2 25 5 1 5 3 7 . 2 5 . 6 0 3 3 6 . 8 1 0 1 5 24 J . 0 4 . 6 5 l J 2 . 3 6 . 9 1 1 6 2 . 2 8 . 8 I S 6 0 . 9 4 . I 1 3 5 5 t u ); 3 I 26 26 52 I 9 1 9 9 9 1 1 1 0 ( c ) Fermentat ion compar ison o fmu tants FSN 1 and 2 with parent s train KM1 0D l 0 These fermentat ions were carried out to compare the mutan t s with the parent stra in , KM1 0D l 0 . S imilar cond i t ions to the previous fermentation were used . The resul t s are shown in t able 7 . 1 1 . At 24 hour s , KM1 0D l 0 had reduced both s ugars to res idual level and produced 3 1 g/1 e thanol , while the three mutants had u t il ized only sucro se . By 4 8 hours , they u t i l ized l ac t o s e and sucro se to res idual leve l s o f 2 and 0 . 8 g/ l , respec t ively . Final l y , i t i s evident tha t all mutant s were sucro se-posi t ive in l iquid med ium and s t i l l retained the d iauxi c behaviour . They showed no improvemen t over the abil ity o f the parent s train . I t was dec ided that fur ther mutat ion should be carried ou t . 203 Table 7 . 1 1 Fermentat ion comparison o f mutants FSN 1 and 2 with paren r s train KM1 0D 1 0 in whey b ro th enriched with sucrose ( 40 : 4 0 g / 1 lactose : sucrose . ( S imilar fermentat ion condi t ions t o those used in table 7 . 1 0 ) t ime organisms pH X lactose sucrose e thanol h c e l l s /ml g/ 1 g / 1 g / 1 0 KM1 0D 1 0 5 . 2 l . Ox 1 0 5 4 1 4 1 4 ( FSN 1 ) 5 . 2 1 . 1 x 1 0 6 4 1 3 9 1 l ( FSN 2 ) 5 . 2 2 . 4x l 0 6 4 0 4 1 1 2 ( FSN 2 ) 5 . 2 3 . 0x l 0 6 39 39 7 KMlOD 1 0 5 . 2 1 . 4x l 0 6 39 40 i 24 KH1 0D1 0 3 . 4 8 . 0x l 0 8 1 . 8 0 . 8 3 1 4 4 . 2 l . 4 x l 0 8 4 1 1 . 5 6 . 7 I l l 4 . l 8 4 0 0 . 8 I 9 . 0 l . 4x l 0 I I 8 1 ') 3 . 9 2 . 7x l 0 39 0 . 4 1 1 .l L I I 48 KMl ODlO 4 . 2 8 0 . 8 2 7 -1 7 . 8x l 0 1 . 8 I 8 4 3 . 9 3 . 3x l 0 2 . 4 0 . 3 2 2 I l l 4 . 3 8 i 2 . 0 0 . 8 2 8 2 . 4 x 1 0 I I 1 2 4 . 4 8 2 . l 0 . 8 2 9 I 2 . 4x l 0 7 . 5 . 3 S econd muta tion experiment ( a ) Isolat ion o f mutant In this at temp t , 300 whey agar p la tes wer e inocula ted with UV irradiated KM1 0D 1 0 cul ture . Aft er replication onto sucrose agar , 16 isolates were found to be unable to grow on sucrose . These were checked us ing a p a ir o f sucrose and whey agar p lates wh ich had b een d ivided into 8 sect ions ( the 3 rd & 4 th p a i r o f p lates i n f i g . 7 . 1 0 ) . Each i s o l a t e was s tr eaked onto one of these sec t ions on both whey and sucro se agar p l ates . The isolates that grew on both p la tes wer e rej ec ted , while those isolates that grew only on whey agar were re tained . The pho tograph shows that isolates no . 2 3 5C and 2 5 6A showed very l i t t le growth in sucrose agar but good growth in whey agar . Isolate 2 5 6A was ( th -checked fur t her by s treaking onto whey and sucrose agars the 5 pai r o t plates i n f i g . 7 . 1 0 ) . There were f ew l arge colonies growing i n sucrose agar ( agar p lates no . S ) . A colony f rom lac tose agar was used to inocu la t � the f i r s t s e t o f plates ( set no . l i n f ig . 7 . 1 1 ) . From this s e t , a col ony was taken f rom whey agar to inoculate the next set of p la tes (no . 2 ) in 204 1 2 3 4 5 Fig . 7 . 10 Sequence of isolation of mutant 235C (FSN 3 ) and 256A . From original replicated plate� and two subcul turing checks . Fig . 7 . 1 1 S equence of plating and s treaking to check the s tability o f mutant 256A. Growth i n sucrose agars indicates ins tabili ty . 20 5 in order to check for purity and s tability o f the isolate . A few colonies grew on the sucrose agar plate . The final subculture onto set number 3 showed more colonies on sucrose agar . At this poin t , this isolate was rej ec ted because of _ the growth on sucrose agar . Isolate no . 2 35C from the plate shown in fig . 7 . 1 0 ( the middle pair , 3rd set from right) was used to inoculate the first set of plates ( from left ) in fig . 7 . 1 2 . There was no growth on the sucrose agar plate . An isolated colony from whey agar was used to inoculate the next set of plates (no . 2 from left ) . There were 3 large colonies growing on the sucrose agar plate . At this point , there was a need to isolate a mutant which grew in lac tose only because there was contaminat ion by sucrose utilizing s trains . This was done by selecting a few isolated colonies from the whey agar plate and using them to inoculate a separate whey agar plate . Af ter incubation they were replicated onto sucrose agar plates . These plates are shown in fig . 7 . 1 2 ( no . 1 , 2 , 3 , in red , 3 sets of plates in the far right ) . There was no growth in the sucrose agar plates . The markings on these plates were a resul t of replication pressing on the agar . Thus , it appeared that , the isolate 235C was pure and s table ( I t did no t grow on sucrose agar ) . I t was assigned a new name of FSN 3 . ( b ) Culture improvement o f mutant FSN 3 I t was shown b y the previous experiment that FSN 3 was a stable mutant when tested for growth in whey and sucrose agar . In addition the mutants FSN 1 & 2 were slow fermenters which reverted to sucrose utilization in liquid medium . In order to improve the FSN 3 cul ture , it was decided that FSN 3 should be repeatedly subcultured into whey permeate bro th wi th added ethanol , the ethanol concentration being increased after each subcul ture so that the isolate would have an improved fermentation rate similar to the resul t obtained in the isolation o f the ethanol tolerant KM10D10 ( sec t . 7 . 2 ) . The inoculum size used was 10% v/v of the final medium volume to allow for cell loss due to ethanol toxicity . Fig . 7 . 1 3 illustrates the first subcul turing sequence . Six subcul tures were mad e . There was progressively poorer growth a s the ethanol level increased ( 5 g/1 per subcul ture ) to 20 g/ 1 . The f i f th and sixth flasks both contained 20 g/1 ethanol . (No te the clearer medium in these two flasks . Each f lask was thoroughly shaken before the pho tograph was taken in order to show the extent o f yeas t growth} . 206 Fig . 7 . l 2 S treaking sequence to check stability o f 235C ( FSN 3 ) and replication t o isolate pure culture . Fig . 7 . l 3 First culture improvement sequence of mutant FSN 3 to improve growth rate and ethanol-tolerance by growing in whey broth ( l OO g/1 lactose) with added ethanol ( s t epwise increase) . 20 7 This run was unsuccessful because the isolate was unable to grow in presence o f ethanol up to 20 g / 1 . Hence , the subcul turing was res tarted from the initial cul ture . In the second run , 5 g/1 ethanol increase was used for each subcult­ uring s tep and the resul ts are i l lus trated in fig . 7 . 14 . The extent of growth was also monitored on lac tose and sucrose agar plates ( fig . 7 . 1 4 b ) . The ethanol level o f 35 g / 1 was reached after 6 subcul tures . There was good growth in all o f these f lasks (up to no . 6 ) . However , as shown by fig . 7 . 1 4 (b ) , the growth on sucrose agar plates increased progressively as the e thanol concentration increased, while the growth in the whey agar plates was better and improved very sl ightly as ethanol increased . The growth improvement on the sucrose agar did no t approach that of the parent s train . There would be more growth also if there was a complete revers ion to sucrose utilization . At this point , cul ture from flask no . 6 ( fig . 7 . 14 a ) was used to ferment whey permeate ( l OO g /1 lac tose ) . This fermentation was carried out to s tudy the ability of the isolate to ferment lactose . Table 7 . 1 2 shows the fermentation results . The initial cell concentration of lx l 06 cells /ml was of normal inoculat ion level but the 7 . cell levels at 24 and 48 hours of l x l O cells/ml were low and lactose utilization was poor . Table 7 . 12 Fermentation o f whey permeate ( l OO g /1 lactose) by mutant FSN 3 ( lOO ml volume) . Af ter a series of 6 subcul tures in whey broth ( l OO g/1 lactose) with up to 35 g/ 1 added ethanol (no . 6-3 . 5E/ l ) ( cf . fig . 7 . 14 ) t ime pH X lactose h cells /ml g/1 0 5 . l 4 . l x l O 6 1 1 0 24 4 . 8 l . 5x l 0 7 1 03 48 4 . 5 l . 4x l 0 7 55 From these resul ts it was decided that cul ture improvemen t should be carried out further . The added ethanol level of 35 g / 1 was maintained . This was considered to be a suf ficiently high concentration because it was found that a s t rain of K. marxianus was inhib ited by ethanol at a concentration of 32 g/1 (Wendorf et al 1 9 70 ) . The resul ts of this las t subcul turing series were also illus trated in fig . 7 . 14 ( f rom no . 7- 3 . 5E/2 208 (a) (b ) Fig . 7 . 1 4 Second culture improvement sequence of mutant FSN 3 . (a) In whey broth , (b) Corresponding lactose and sucrose agar plates to monitor mutant s tability . No te increasing growth in sucrose agar as the e thanol content increased . 209 to 1 1-E/6) . There was good growth in all flasks ( fig . 7 . 14 a ) . The growth in both whey and sucrose agars improved s lightly ( f ig . 7 . 1 4 b ) . Finally , a fermentation test for the last culture was carried out using whey permeate with added sucrose ( lactose : sucrose 40 : 40) (in 100 ml medium) . Table 7 . 1 3 shows the fermentation results at 0 and 48 hours . Tabl e 7 . 1 3 Fermentation of whey permeate enriched with sucrose by mutano FSN 3 which was passed through a series o f 1 1 subcultures in whey broth ( 100 g/ 1 lactose ) with added ethanol ( 35 g/ 1 ) . (This was from the sixth subcul ture which was grown in presence of 35 g/ 1 ethanol . No . 1 1-3 . 5E/ 6 ) ( c f . fig . 7 . 14 a) . t ime pH sucrose lac tose ethanol h g/1 g/1 g/1 0 4 . 6 40 4 3 0 48 4 . 4 1 . 4 1 2 2 2 Sucrose was all utilized but only 3 1 g/1 of lactose was utilized . There was 1 2 g/1 remaining . From table 7 . 1 1 , the parent s train was able to completly utilize a similar amount of mixed sugars in .24 hours . 7 . 5 . 4 Conclusions The resul ts showed that even though the fermentation capabil ity o f the mutant FSN 3 was improved over i t s initial culture , it s till reverted to utilizing sucrose before lactose when both sugars were available s imultaneously . The growth of the mutant on sucrose agar increased with improvement , al though growth on whey agar was more profuse than on sucros e agar . The final fermentation test also showed that the improved ability of the mutant was not as good as that of the parent strain . Thus , the three mutants isolated were ' sucrose-negative when they were grown on sucrose and whey agars . However , all mutants reverted to utilizing sucrose before lactose when the two sugars were available s imul taneously in l iquid medium . The last mutant ( FSN 3 ) was passed through a series of subculturing in the presence of increasing concentrations of ethanol to improve its fermentation ability . This improved the culture but abilty to ut ilize sucrose also improved . Sucrose was also ut ilized before lactose when tes ted for growth in mixed substrate of lactose and sucrose . 2 10 7 . 6 SUMMARY 1 . Fermentation of whey permeate by K. marxianus UCD FST 7 1 58 could be carried out without nutrien t addition . This yeast strain was able to utilize 40 g/ 1 lactose in whey permeate in 16 hours but was not able to utilize completely 100 g/1 lactose . This was shown to be due to inhibition o f growth by e thanol . In the fermentation of whey permeate enriched with molasses , the culture exhibi ted diauxic b ehaviour , utilizing sucrose before lactose . The overall fermentation time was increased compared with that for unsupplemented whey fermentation . 2 . A s imple culture improvement technique was used to improve ethanol tolerance o f K. marxianus . An improved yeast s train , KM10D 1 0 , was isolated . I t was a rapid l ac tose fermenter and bet ter producer of ethanol from lactose than its parent , K. marxianus UCD FST 7 1 5 8 and o ther lac tose ferment ing yeasts under the same fermentation conditions and medium used . I t was found to be stable . 3 . An investigation was carried out to isolate a diauxie-negative K. marxianus using D-glucosamine as the glucose analogue . K. marxianus was found to be able to grow on lactose in the presence o f D-glucosamine up to 5 g/1 . At higher D-glucosamine concentrations ( 10 and 20 g/1 ) , the growth was inhibited . Fermentation tes ts of cells that could grow a t these concentrations showed them to be diauxie-positive . 4 . UV irradiation and agar plate replication techniques were used to produce a sucrose-negative s train of K. marxianus . Three mutants were obtained and all were sucrose-negative in sucrose agar but always reverted to sucrose utilization when grown in a broth medium that contained both lac tose and sucrose . CHAPTER 8 FINAL DISCUS S ION AND CONCLUS ION The feasibi l i ty of fermenting whey permeate to ethano l us ing a tower fermente r has been investigated . The commercial advantage in using such a process is considerable , since the tower fermentation process provides continuous treatment of whey permeate to ethanol requi ring low energy input , plant cost , and maintainance but can give a high rate of ethanol production . Ini tial screening and investigations to obtain a lac tose-fermenting and flocculating yeast found that there was only one strain of K. m arxianus available that was flocculent . This yeast s train , K. marxianus Y42 , showed lac tose-fermenting ability comparable with o ther K. marxianus strains . Further investigations of the behaviour of this yeast species showed it to be unable to ferment comple tely a high concentration of lactose in whey permeate ( 1 00 g/ 1 lactose ) . This was found to be due to inhibition by ethanol . The optimum operating cond itions for tower fermentation of whey permeate by K. ma rxianus Y42 were found at a superficial liquid velocity of 0 . 24 mm/s , and pH be tween 4 . 2 and 4 . 6 when the fermentation temperature was 30 C . The minimum tower height was 0 . 82 m ( excluding separator section ) which corresponded to a residence time of hour . The concentration of e thanol produced under these conditions was 1 6 g/ 1 at productivity of 1 6 g/ lh from 45 g/ 1 o f lac tose ( 94% lac tose utilization ) and 7 1 % yield on lac tose utilized . If the separator section were included in the consideration , the productivity reduced to 5 g/ lh , and the effluent liquor contained 1 8 g/1 o f ethanol . The overall re tention time was 3 - 7 hours . The cell concentration inside the tower varied between 1 0 and 1 00 g/ 1 . dried weight ( 54 and 350 g/ 1 wet weight ) being greatest at the bo ttom of the tower . The fermentation was l imited by the moderately flocculent nature of K. marxi anus Y42 . The productivity was found to be lower than that observed for tower fermentation using sucrose based media , but was found to be more than 13 times that of current commercial fermentation of whey to ethanol . During continuous tower operation , bac terial contamination occurred . This lowered lac tose utilization , e thanol production and the med ium pH . When this occurred , the fermenter was emptied , c leaned and 211 2 1 2 s t erilized before res tarting the fermentation . Tower fermentation using mixed culture of K. marxia nu s and S. cerevisia e to ferment whey permeate enriched with molasses was difficult . The re was incomplete lactose util ization even at the low feed rates used (up to 0 . 1 4 mm/s ) . Sucrose was completely utilized at the bot tom of the tower fermenter . The incomplete lac tose utilization was found to be a result of the diauxic behaviour of K. marxianus in the simultaneous presence of sucrose and lactose . Suc rose was utili zed first then lac tose was uti l i zed , but by this time the ethanol concentration was high . The high concentration of ethanol inhibited lactose uptake in the tower . Thus , lactose utilization took place at a low rate . It was also found that molasses , when mixed with whey permeate , could slow down the rate of fermentation and contributed to the incomplete sugar utilization . The re was also some incompatability of the two yeast strains used . S. cerevi sia e CFCC39 was very flocculent in whey permeate enriched wi th molasses . The medium feed rate to the tower was very slow . This resulted in a very dense cell population causing the blockage of the separator and the formation of gas s lugs . The less flocculent K . marxianu s Y42 was siowly washed out of the tower and lactose uti l ization was therefore reduced . Investigation of the flocculent nature ind icated that this yeast s train exhibited behaviour when grown in different media . It was of K. ma rxianus Y42 different flocculating found to have good flocculence when grown in media prepared from yeast and mal t extract powder , mal t extract broth powder , mal t extract syrup or in whey permeate enriched with these nutrients . I t showed poor flocculence when it was grown in acidic media or media which did not support good growth . During the start up of the tower for fermentation of whey permeate and whey permeate enriched with molasses , the whey permeate feed was enriched further wi th mal t extract syrup in order to enhance the flocculence of K. marxianu s for rapid build up of cell concentration . An attempt to improve the ethanol tolerance of a strain of K. ma rxianus was successful . The isolate , called KM 1 0D 1 0 , fermented lactose in whey permeate rapidly and grew in presence of ethanol up to 50 g/1 . The isolate was found to be able to produce ethanol at a faster rate than all of the K. marxianus strains tested . The isolate was found to be stable upon repeated subculture . However , i t could not 2 1 3 be used in the tower fermenter as it was not flocculent . A similar culture improvement technique could be used to improve the e thanol tolerance of a flocculent parent strain and this could reasonably be expected to produce a suitable culture for use in the tower fermentation process . An attempt to isolate a diauxie-negative strain of K. marxianus , using D-glucosamine as a glucose analogue , was unsuccessful . This yeast utilized lac tose in the presence of the analogue in the growth medium . An at tempt to isolate a strain of K. marxianus , which was unable to utilize sucrose , using UV radiation as a mutagen was only partial ly successful . Mutants that showed no growth in sucrose agar plates were isolated . These mutants , when tes ted for growth in liquid medium containing mixed substrate of sucrose and lactose , reverted to utilizing sucrose before lactose . Further attempts were carried out , followed by a cul ture improvement subculturing sequence similar to that used in the isolation of an e thanol toleratant strain . The mutant obtained from this later attempt reverted to sucrose util ization in liqui d medium as in the first attempt . In summary , i t has been shown that the fermentation of whey permeate to ethanol using a tower fermenter is feasible and requiring no nut rient addition to the whey permeate except for the requirement of malt extract syrup during start up and periodically thereafter to maintain high yeast flocculence . The use of a mixed yeast cul ture to ferment whey permeate enriched with molasses in a tower fermenter was difficul t . The re remained some problems to be solved . These are the diauxic behaviour of K. marxianus and the incompatibility of the flocculence of the two yeast strains used . The flocculent behaviour of K. marxianus Y42 needs to be inves tigated further . Suggestions for future investigations are listed below . 1 . Further investigation of KM Y42 flocculence properties It was found that this yeast showed variable flocculence in different media but it was not possible to isolate the fac tors tha t caused this behaviour . KM Y42 is the only lactose fermenting yeast that is known to flocculate . Thus , further investigation to establish the cause of its flocculence would be of great benefit to the study of tower fermentation of whey to ethanol . 2 1 4 2 . Mutation to isolate flocculent lactose-fermenting yeast Flocculation is a genetically controlled behaviour. Transformation mutation could be employed to transfer gene information for lac tose transport and fermentation to a highly flocculent yeast such as S. cer evisiae. The transformation of S -galactosidase information from K. lact is to S. cer evisiae has bee n suc cessfully carried out but the mutant could not utilize lactose because lac tose transport information was lacking (Dickson 1 980 ) . Recombinants of S. cer evisiae and K. lact is through spheroplast fusion were found to be unstable ( Stewart 1 98 1 ) . Thus , presently fusion is an unsuitable approach because of the incompatability of the two yeasts . It is generally accepted that flocculation is a cell wall phenomenon ( Stewart 1 975 ) . Thus , an attempt to transfer flocculation genes into K. marxianus spheroplasts may not be successful because there may be up to four genes controlling flocculence and the existing cell wall may not be suitable for the exhibition of flocculence . 3 . Tower fermentation of whey containing 1 00 to 200 g/ i lac tose In this study , the concentration of lactose in the whey permeate was 40 g/ 1 . Fermentation of higher concentrations of lactose in whey ( up to 200 g/1 ) has been reported using either K. marxianus or C. p seudotropical is in batch fermentation (Moulin et al 1 980 Burgess & Kelly 1 979 ) . A high lactose concentration means greater ethanol productivi ty and improved economy on the utili zation of fermentation and distillation equipment . 4 . Mixed culture fermentation of whey permeate enriched wi th molasses It was found that the yeast strains CFCC39 and KM Y42 did not have compatible flocculence for use together in the tower fermenter . The less flocculent KM Y42 was washed out of the tower slowly because the highly flocculent CFCC 39 caused blockage of the yeast recycling path and the formation of gas slugs . The yeast strain SC 1 46 may be more compatible with KM Y42 since it is moderately flocculent . It may also be more desirable to use only K. mar xianus in the tower since it was found that the mixture of sucrose and lac tose was uti l i zed more rapidly when the ratio of K. mar xianus concentration was greater than 90% . to total cel l 2 1 5 5 . Fermentation o f whey permeate containing high concentration of lactose with added nutrient that induces ethanol tolerance It was found that high ethanol concentration exerted considerable inhibition on the fermentative activity of K . marxianus, causing incomplete lac tose utilization when lactose and sucrose were available together . I t has been shown for a mumber of S. cer evisiae strains that the addition of proteolipid at the early s tage of the fermentation can improve ethanol tolerance , (Hayashida & Ohta 1 98 1 ) . The addition of such a nutrient to whey permeate containing a high sugar concentration and studying its effect on ethanol tolerance could contribute to the understanding of the behaviour of K. marxia nus. REFERENCES Alzola , F . ( 1 9 4 5 ) Equipment and procedure for carrying out continuous fermentations , US patent 2 , 37 1 , 208 . Anon ( 1 9 7 4 ) Ethanol , Encyclopaedia Britannica , Macropaedia , 1 : 4 50-459 . Anon ( 1 9 7 7 ) Dairy news diges t /Adding yeast to whey starts process which produces alcohol , Modern Dairy , 5 6 ( 6 ) : 4 . Anon ( 1 98 1 ) Fermentation in N . Z . with lactose substrate , DRI internal report , Palmerston North , New Zealand , 9 March . Archer , R . H . , Larsen , V . F . and McFarlane , P . N . ( 1983 ) Comparison o f three high rate anaerobic diges ter designs for the treatment of whey : preliminary results . IPENZ conference , Hamilton , New Zealand , February . Atkinson , B . & Daoud , I . S . ( 1 9 7 6 ) Microbial floes and flocculation in fermentation process engineering , Adv . Biochem . Eng . , 4 : 4 1- 1 24 . Aul t , R . G . , Hampton , A . N . , Newton , R. & Roberts , R . H . ( 1 969 ) B iological and biochemical aspects of tower fermentation , J. Ins t . Brew . , 7 5 : 2 60-2 7 7 . Ballou , C . E . , Kern , K .A . & Raschke , W . C ( 1 9 7 3 ) Genetic control o f yeast mannan s tructure , J . Biol . Chem . , 248 ( 1 3 ) : 4 667-467 3 . Barnet t , J . A. , Payne , R . W . & Yarrow , D . ( 1 9 7 9 ) A guide to identifying and classifying yeasts , Cambridge Univ . Press , p . 56-59 , 283 . Bazua , C . D . & Wilke , C . R. ( 1 9 7 7 ) Ethanol effects on the kinetics o f a continuous fermentation with Saccharomyces cerevisiae , Biotech . B io eng . Symp . 7 : 1 05- 1 1 8 . Beavan , M . J . , Belk , D .M . , Stewar t , G . G . & Rose , A . H . ( 1 9 7 9 ) Changes in elec trophoretic mobility and lytic enzyme ac tivity associated with development of flocculating ability in Saccharomyces cerevisiae , Can . J . Microbia l . , 25 (8 ) : 8 88-895 . Berns tein , S . & Everson , T . C . ( 19 74 ) Protein production from acid whey via fermentation , Env . Pro t . Agency , Report no . EPA-660/2-74-025 , May 1 9 7 4 . Berns tein , S . , Tzeng , C . H . & Sisson , D . ( 1 9 7 7 ) The commercial fermentation of cheese whey for the production of pro tein and/or alcohol , Biotech . Bioeng . Symp . , 7 : 1-9 . Bisse t t , D . W . & Riddle ,M . J . ( 1 9 7 6 ) S tatus o f ef fluent treatment in the Canadian dairy industry , Int . Dairy Federation seminar on dairy e f fluent , Warsaw , Oc tober , 1 976 . 2 1 6 2 1 7 Boening , P . H . & Larsen , V . F . ( 1 9 8 2 ) Anaerobic fluidized bed whey treatment , B io tech . B ioeng . , 24 : 2 539-2556 . Browne , H . H . ( 1 94 1 ) Ethyl alcohol fermentation o f lactose in whey , Ind . Eng . Chem . , News ed . , 1 9 ( 22 ) : 1 2 7 2 & 1 2 7 6 . Bungay , H . R . & Bungay ,M. L . ( 1 968 ) Microbial interactions in continuous cul ture , Adv . Applied Micro . , 1 0 : 269-29 0 . Bungay , H . R . & Krieg , N . R . ( 1 966 ) Growth in mixed culture processes , Chem. Eng . Pro g . Symp . Ser . , 62 (69 ) : 68-72 . Burgess , K . J . ( 1 9 7 7 ) Produc tion of yeast pro tein from lactose permeate in a tower fermenter , Irish J . Fd . Sci . Technol . , 1 : 107- 1 1 5 . Burges s , K . J . & Kelly , J . ( 1 9 7 9 ) Alcohol production by yeast in concentrated ultrafiltration permeate from cheddar cheese whey , Irish J . Fd . Sci . Technol . , 3 ( 1 ) : 1 -9 . Callej a , G . B . ( 1 9 70 ) Flocculation in Schizosaccharomyces pombe , J . Gen . Microbial . , 64 : 247-250 . Campbell , L . A . ( 1953 ) Production o f calcium lactate and lac t ic acid from cheese whey , Can . Dairy Ice Cream J . , 3 2 : 2 9-3 1 , 7 7-7 8 . Cawley , T . N . & Ballou , C . E . ( 1 9 7 2 ) Identification o f two Saccharomyces cerevisiae cell wall mannan chemo types , J . Bacterial . ·, 1 1 1 : 690-695 . Chen , H . C . & Zall , R . R . ( 1982) Continuous fermentation o f whey into alcohol using an attached-film-expanded-bed reactor , Process Biochem . , 1 7 ( 1 ) : 20-25 . Chain , S . K . & Mateles , R . I . ( 1 9 68 ) Growth o f mixed cul ture on mixed sub s trates , I . Continuous cul ture , Appl . Microbial . , 16 : 1 3 3 7 - 1 34 2 . Chiu , S . Y . , Erickson , L . E . , Fan , L . T . & Kao , I . C . ( 1 9 7 2 ) Kinetic model iden tification in mixed populations using continuous cul ture data , Bio tech . Bioeng . , 14 : 207-23 1 . Cleland ,A . C . ( 1 98 3 ) Es t imat ion of data unc.ertainties , Dept . o f Bio techno­ logy , Massey University , Palmerston North , New Zealand , 46 pp . Coote , S . D . J . ( 19 74 ) Studies in the continuous tower fermentation of alcohol , PhD thesis , Univers ity o f As ton , Birmingham , UK . Co ton , S . G . ( 1 9 7 6 ) Recovery o f d airy waste , In Food from was t e , Birch , G . G . , Parker ,K . J . & Worgan , J . T . ( ed . ) , Appl ied Sci . Pub . , London , p . 22 1 -23 1 . Cysewski , G . R . & Wilke , C . R . ( 1 9 7 7 ) Rapid ethanol fermentations using vacuum and cell recycle , B io tech . B ioeng . , 1 9 : 1 1 25- 1 143 . Day , A . W . , Poon , N . H . & Stewart , G . G . ( 1 9 7 5 ) Fungal f imbriae Ill . The ef fect on flocculation in Saccharomyces , Can . J . Microb ial . , 2 1 : 558-564 . 2 1 8 Demain ,A .L . & Solomon , N .A . ( 1 9 8 1 ) Indus trial microbiology , Scientific American , 24 5 ( 3 ) : 42-5 1 . Dickson , R . C . ( 1 980) Express ion of a f oreign eukaryotic gene in Saccharomyces cerevisiae : S-galactos idase from KZuyveromyces Zactis , Gene , 1 0 : 34 7-356 . Dillon , J . A . ( 1 980) The immobilization o f KZuyveromyces fragi Zis and Saccharomyces cerevisiae in polyacrilamide gel , Mas terate thesis , Massey University , Palmers ton North , New Zealand . Dunn , N .W . ( 1 98 1 ) School o f Biotechnology , University o f New South Wales , Sydney , Australia , Personal communication . Eddy , A . A. ( 1 958 ) Composite nature o f the flocculation process o f top and bottom strains of Saccharomyces , J . Ins t . Brew . , 64 : 1 4 3- 1 5 1 . Ellio t , J . J . & Ball , A . J . S . ( 1 9 7 3 ) Glucosamine-resis tant mutants o f Saccharomyces cerevisiae � Genetics , 74 : 7 1 . Ewen , M . ( 1 980) Fermentation of milk wastes , PhD thesis , University of As ton , B irmingham , UK . Ewen ,M . & Greenshields , R . N . ( 1 982 ) Single cell protein produc t ion from Cheshire cheese whey . Brief communication , 1 2 thint . Dairy Congr . , 1 ( 2 ) : 5 29 . Fredrickson , A . G . & Tsuchiya , H . M . ( 1 9 7 7 ) Microbial kinetics and dynamics , In Chemical reac tor theory : A review , Lapidus , L . & Amundson , N . R . ( ed . ) Prent ice Hall , p . 405-483 . Furs t , A . & Michels , � . ( 1 9 7 7 ) An evaluation of D-glucosamine as a gratuitous catabolite repres sor of Saccharomyces carZsher.gensis , Molec . gen . Gene t . , 1 55 : 309-3 14 . Gapes , J . R . ( 1982 ) The acetone-butanol-ethanol fermentation : Preliminary s tudies on some pract ical aspec ts , Masterate thesis , Massey University , Palmerston Nor th , New Zealand . Gawel , J . & Kosikowski , F . V . ( 1 9 7 8 ) Improving alcohol fermentation in concentrated ul trafiltration permeates o f cottage cheese whey , J . Food Sc i . , 4 3 : 1 7 1 7- 1 7 1 9 . Ghose , T . K . & Bandyopadhyay , K . K . ( 1 980) Rapid ethanol fermentation in immobilized yeast cell reacto r , Biotech . Bioeng . , 22 ( 7 ) : 1 489-1496 . Gong , C . S . , McCracken , L . D . & Tsao , G . T . ( 1 98 1 ) Direct fermentation of D-xylose to e thanol by a xylo se-ferment ing yeast mutant , Candida sp . XF 2 1 7 , Bio tech . Let ters , 3 ( 5 ) : 2 1 3-2 1 8 . Gooding , B . ( 1 982 ) Whey was te recovery inspires exportable technology , N . Z . Farmer , 1 03 ( 1 4 ) : 94-95 . Goto , S . , Kuwaj ima , T . , Okabe ,M . , Okamoto , R . & Takamatsu , A . ( 1 9 7 9 ) Appl icat ion o f tower type fermenter with two-fluid nozzle t o biomass production from methanol , J. Ferment . Technol . , 5 7 (4 ) : 34 1- 348 . 2 1 9 Greenshields , R . N . & Smith , E . L . ( 1 9 7 1 ) Tower fermentation sys tems and their applications , The Chem . Eng . , 249 : 1 8 2- 1 90 . Greenshields , R . N . & Smith , E . L . ( 1 9 74 ) The tubular reactor in fermenta t ion , Process Biochem . , 9 ( 3 ) : 1 1 - 1 7 & 2 8 . Greenshields , R . N . , Yates , J . , Sharp , P . & Davies , T . M. C . ( 1 9 7 2 ) Studies in tower continuous fermentation . I . Methods for the characterization of flocculent yeas ts , J . Ins t . Brew . , 78 ( 3 ) : 2 36-240 . Gro te ,W . , Lee , K . J . & Rogers , P . L . ( 1 980) Cont inuous ethanol produc tion by immobilized cells o f Zymomonas mobi lis � Biotech . Let ters , 2 ( 1 1 ) : 4 8 1 -486 . Harrison , J . S . ( 1 9 7 1 ) Yeast production , Progress Ind . Micro . , 1 0 : 1 2 9- 1 7 7 . Harrison , J . S . & Graham, J . C . J . ( 1 9 70 ) Yeasts in d istillery prac tice , In The yeast s , Vol . 3 , Rose , A . H . & Harrison , J . S . (ed . ) , Academic Press , p . 2 8 1 -348 . Hayashida , S . & Ohta , K . ( 1 9 8 1 ) Formation o f high concentrations of alcohol by various yeasts , J . Inst . Brew . , 8 7 ( 1 ) : 42-44 . Helm , E . , Nohr , B . & Thorne , R . S . W . ( 1 953 ) The measurement o f yeast f locculence and its significance in brewing , Wallerstein Lab . Comm . , 1 6 ( 55 ) : 3 1 5-325 . Henderson , C . F . ( 1 9 8 1 ) Continuous fermentation o f ethanol from sugar and fodder beet using a tower f ermenter , Chemeca 8 1 - 9 th Aust . Conf . on Chem. Eng . , Chr,istchurch , New Zealand , p . 467-475 . Henderson , C . F . & Smith , J . M . ( 1 98 2 ) Tower and s tirred tank fermentation o f sugar bee t and fodder beet , Proc . 5 Int . Symp . on Alcohol Fuel Tech . , 1 : 1 9 7-205 . Hobman , P . G . ( 1 982 ) Prospec ts for whey util ization . Summary of Proc . 5 0 Annual Conf . o f Dairy Managers , Palmerston North , New Zealand , p . 9 - 1 0 . Holzer , H . ( 1 9 7 6 ) Catabolite inact ivat ion in yeast , Trends Biochem . , 1 : 1 7 8- 1 8 1 . Hodpodka , J . ( 1 966 ) Industrial appl ications of continuous fermentations , In Theoretical and methodological basis o f cont inuous cul ture of microorganisms , Malek , I . & Fencl , Z . ( ed . ) , p . 493-645 . Hossain ,M . ( 1 983 ) Submerged citric acid fermentation o f whey permeate by Aspergil lus niger , PhD thesis , Massey University , Palmers ton North , New Zealand . Hough , J . S . , Keevil , C .W . Maric , V . , Philliskirk , G . & Young , T .W . ( 1 9 76 ) Continuous culture in brewing , In Continuous culture 6 : Applications and new fields , Dean , A . C . R . et al (ed . ) , Ell i s Horwood Pub . , p . 226-23 7 . 2 20 Hough , J . S . & Wase , D . A . J . ( 1 966 ) Continuous culture , Theory and appl ications , Process Biochem . , 1 ( 2 ) : 7 7-83 . Howell , M . J . ( 1 98 1 ) High grade ethanol from whey , Food Technol . in N . Z . , 1 6 ( 8 ) : 1 3 , 1 5 & 1 7 . Howell , M. J . & Tichbon , C . R . ( 1 980 ) Ethanol from whey , N . Z . Ins t . o f Eng . Conf . , Dunedin , New Zealand , February . Hsiao , H . Y . , Lin Chang Chiang , Ueng , P . P . & Tsao , G . T . ( 1 98 2 ) Sequential utilization of mixed monosaccharides by yeasts , App . Env . Microbial . 4 3 (4 ) : 840-845 . Izaguirre ,M . E . & Castillo , F . J . ( 1 982 ) Selection of lactose-fermenting yeas t for e thanol produc t ion from whey , Bio tech . Let ters , 4 ( 4 ) : 25 7 - 262 . Jannasch , H .W . & Mateles ,R . I . ( 1 9 74 ) Experimental bacterial ecology s tudied in continuous culture , Adv . Microbial Physiol . , 1 1 : 1 65-2 1 2 . Jayatissa , P . M . & Rose , A . H . ( 1 9 7 6 ) Role o f wall phosphomannan in flocculation o f Saccharomyces cerevisiae , J . Gen . Microbial . , 9 6 : 1 65- 1 74 . Jones , R . P . & Greenfield , P . F . ( 1 9 8 1 ) Batch ethanol produc tion with dual organisms , B io tech . Let ters , 3 ( 5 ) : 2 25-2 3 0 . Klopper ,W . J . , Robert s , R . H . , Royston , M . G . & Aul t , R . G . ( 1 965 ) Continuous fer�entation in a tower fermenter , Proc . Eur . Brew . Convention , S tockholm, p . 2 38-259 . Kosaric , N . & Miyata , N . ( 1 9 8 1 ) Growth of morel mushroom mycelium in cheese whey , J . Dairy Research , 48 : 149- 162 . Kosaric , N . , Ng , D . C .M . , Russell , I . & Stewar t , G . C . ( 1980) Ethanol produc t ion by fermentation : An alternat ive l iquid fuel , Adv . Applied Microbial . , 2 6 : 1 4 7-22 7 . Led erberg , J . & Lederberg , E .M . ( 1 9 52 ) Replica plating and indirect selection o f bacterial mutants , J . Bacterial . , 63 : 399-406 . Lewis , C .W . , Johnston , J . R . & Martin , P . A . ( 1 9 7 6 ) The genet ics o f yeast flocculation , J . Inst . Brew . , 82 ( 3 ) : 1 5 8- 1 60 . Linko ,Y . , Jalanka , H . & Linko , P . ( 1 9 8 1 ) Ethanol product ion from whey with immobilized living yeast , B io tech . Letter s , 3 ( 6 ) : 26 3-268 . Linko , Y . & Linko , P . ( 1 98 1 ) Cont inuous ethanol product ion by immobilized yeas t reacto r , B iotech . Let ters , 3 ( 1 ) : 2 1-26 . Lodder , J . ( 1 9 7 0 ) The yeasts : A taxonomic study , North Holland Pub . , Amsterdam , Holland . 22 1 Lyons , T . P . & Hough , J . S . ( 19 70) Flocculation o f brewer ' s yeas t , J . Inst . Brew . , 7 6 : 564-5 7 1 . Lyons , T . P . & Hough , J . S . ( 1 9 7 1 ) Further evidence for the cross bridging hypothesis for flocculation o f brewer ' s yeast , J . Ins t . Brew . , 7 7 : 300-305 . Maddox , I . S . & Richert , S . H . ( 1 9 7 7 ) Produc tion of gibberell ic acid using a dairy waste as the basal medium , Appl . Env . Microbia l . , 33 ( 1 ) : 20 1-202 . Maiorell a , B . , Wilke , C . R . & Blanch , H . W . ( 1 98 1 ) Alcohol product ion and recovery , Adv . Biochem . Eng . , 20 : 43-92 . Margarit i s , A . & Baj pai , P . ( 1 982 ) Direc t fermentation o f D-xylo se to ethano l by Kluveromyces marxianus s t rains , Appl . Env . Microbial . 44 ( 5 ) : 1039- 1 04 1 . Marshall , K . R . ( 1 9 7 8 ) Prospect for ethanol produc tion from whey , Proc . 1 s t Biotechnology Conf . on Biomass and energy , Massey University , Palmerston Nor th , New Zealand , May . p . 280-284 . Marshall , K . R . ( 1 98 1 ) Dairy processing 1 . Compositions and properties of milk f luids , Proc . Biochemical Engineering - Applications and po tentials in N . Z . , Auckland University, New Zealand . Mawson , J . ( 1 983 ) N . Z . Distillery Co . Ltd . , Edgecumbe , N . Z . , Pr ivate communication . Meyra th , J . & Bayer , K . ( 1 9 7 9 ) Biomass from whey , Economic Microbial . , 4 : 207-269 . Michels , C . A . & Romanowski , A . ( 1 980 ) Pleio tropic glucose repression­ resistant mutation in Saccharomyces carlsbergensis , J . Bacterial . 1 4 3 : 6 7 4-6 7 9 . Mill , P . J . ( 1 964a) The effec t o f nitrogenous substances on the time of flocculat ion o f Saccharomyces cerevisiae , J . Gen . Microbia l . , 35 : 5 3-60 . Mill , P . J . ( 1 964b) The nature o f the interactions between flocculent cells in the flocculation of Saccharomyces cerevisiae , J. Gen . Microbial . 35 : 6 1-68 . Monod , J . ( 19 4 7 ) The phenomenon o f enzymatic adaptation , Growth , 1 1 : 223-289 . Morris , G . G . , Greenshields , R . N . & Smith , E . L . ( 1 9 7 3 ) Aeration in tower fermenters containing microorganisms , B iotech . Bioeng . Symp . , 4 : 5 35-545 . Moul in , G . , Guillaume , M. & Galzy , P . ( 1 980) Alcohol product ion by yeast in whey ultrafiltrate , Biotech . B ioeng . , 2 2 : 1 2 7 7 - 1 28 1 . Nagodawithana , T . W . , Castellano , C . & Steinkraus , K . H . ( 1 9 7 4 ) Effect o f dissolved oxygen , temperature , initial cell count and sugar concentration on the viabil ity of Saccharomyces cerevisiae in rapid 222 fermentation , Appl . Microbial . , 2 8 ( 3 ) : 383-39 1 . Nagodawithana , T . W . & S teinkraus , K . H . ( 1 9 76 ) Influence o f the rate of e thanol production and accumulation on the viability o f S accharo my ces cerevisiae in rapid fermentation , Appl . Microbial . , 3 1 ( 2 ) : 1 58- 1 62 . Nishihara , H . , Toraya , T . & Fukui , S . ( 1 9 7 6a) Fac tors af fec ting flocculation of brewer ' s yeast , J. Ferment . Technol . , 54 ( 6 ) : 35 1 -355 . Nishihara , H . , Toraya , T . & Fukui , S . ( 1 9 7 6b ) Induc tion o f floc-forming abili ty in brewer ' s yeast , J . Ferment . Technol . , 5 4 ( 6 ) : 356-360 . Nishihara , H . , Toraya , T . & Fukui , S . ( 1 9 7 7 ) Effect of chemical modifications of cell surface components o f a brewer ' s yeast on the floc-forming ability , Arch . Microbial . , 1 1 5 : 1 9-2 3 . Nord , F . F . & Weiss , S . ( 1 958) Fermentation and respiration , In The chemistry and b iology of yeasts , Cook , A . H . ( ed . ) , p . 323-368 . Oborn , J . ( 1 968) A review o f methods available for whey util ization in Australia , Australian J . Dairy Technol . , 23 : 1 3 1- 1 36 . O ' leary , V . S . , Green , R . , Sullivan , B . C . & Holsinger , V . H . ( 1 9 7 7a) Alcohol produc tion by selected yeast strains in lactase-hydrolysed acid whey , B iotech . Bioeng . , 1 9 : 1 0 1 9- 1035 . O ' l eary , V . S . , Sutton , C . , Bencivengo ,M . , Sullivan , B . & Holsinger , V . H . ( 19 7 7b ) Influence o f lactose hydrolysis and solids concentration on alcohol production by yeast in acid whey ultrafiltrate , Biotech . B ioeng . , 1 9 : 1 689- 1 702 . Owen , L . W . ( 1 948 ) Continuous f ermentation , Sugar , 4 3 ( 2 ) : 36-38 . Paca , J . & Gregr , V . ( 1 9 7 9 ) Effec t o f dilution rate on growth characteris­ tics o f Candida ut iZis in a multistage tower fermenter , J . Ferment . Technol . , 5 7 ( 3 ) : 2 27-2 32 . Parkin , M . F . & Marshall , K . R . ( 1 9 7 6 ) Spray irrigation o f dairy fac tory e ffluent - A review of current practice in N . Z . , N . Z . J . Dairy Sc i . Technol . , 1 1 : 1 96-205 . Pimentel , L . S . ( 1 980 ) The Brazil ian ethanol program , Biotech . Bioeng . 2 2 ( 1 0 ) : 1 989-20 12 . Portno , A . D . ( 1 9 6 7 ) New system o f continuous fermentation by yeast , J . Inst . Brew . , 7 3 ( 1 ) : 43-5 0 . Prince , I . G.& Barford , J . P . ( 1 98 2 ) Tower fermentation of sugar cane juice , B iotech . Let ters , 4 ( 7 ) : 469-4 74 . 223 Prince , I . G . , Callander , I . J . , Barford , J . P . & Prince , R . G . H . ( 1 98 2 ) Power ethanol production using flocculated microorganisms in �ontinuous upf low fermenters , Proc . V Int . Symp . on Alcohol Fuel Tech . , 1 : 245-25 2 . Rainbow , C . ( 1 966 ) Flocculation o f brewer ' s yeast , Process Biochem . 1 ( 1 2 ) : 4 89-492 . Reesen , L . ( 1 9 7 8 ) Whey and permeate as raw material for alcohol production , Dairy Sci . Abs t . , 4 0 ( 6 ) : 322 . Robinson , B . P . , Short , J . L . & Marshall , K . R . ( 1 9 7 6 ) Traditional lactalbumin­ manufacture , properties and uses , N . Z . J . Dairy Sci . Technol . 1 1 : 1 1 4 - 1 26 . Roger s , P . L . , Lee , K . J . & Tribe , D . E . ( 1 980 ) High product ivity e thanol fermentations with Zymomonas mobi lis , Process Biochem . , 1 5 ( 6 ) : 7- 1 1 . Rogosa ,M . , Browne , H . H . & Whittier , E . O . ( 1 9 4 7 ) Ethyl alcohol from whey , J . Dairy Sci . , 30 : 263-269 . Rose , A . H . ( 1 9 7 7 ) History and scientific basis of alcoholic beverage product ion , Economic Microb iology , 1 : 1-4 1 . Royston ,M . G . ( 1 966 ) Tower fermentation o f beer , Process Biochem . 1 ( 4 ) : 2 15-22 1 . Rudin , A . D . & Hough , J . S . ( 1 959 ) Influence o f yeas t strairt on production of beer by continuous fermentation , J . Ins t . Brew. , 6 5 : 4 1 0-4 1 4 . Sandbach , D .M . L . ( 1 98 1 ) P roduct ion o f potable grade alcohol from whey , Cul tures Dairy Products J . , 1 6 : 1 7-22 . Sarfacon , R. , Rouillard , R. Goupil ,M . & Heick , H .M . C . ( 1 9 7 2 ) Alcohol dehydrogenase ac t ivity in the yeast Saccharomyces fragi lis , Can . J . Microbial . , 1 8 : 35-44 . Short , J . L . ( 1 9 7 8a) Whey process ing , N . Z . J . Dairy Sci . Technol . , 1 3 ( 4 ) :A98-A99 . Short , J . L . ( 1 9 7 8b ) Prospec ts for the util:i,zat.ion o f deproteinated whey in New Zealand - A review , N . Z . J . Dairy Sci . Technol . , 1 3 : 1 8 1- 1 94 . Short , J . L . & Doughty , R . K . ( 19 7 7 ) Demineralization of deproteinated wheys by electrodialysi s , N . Z . J . Dairy S c i . Technol . , 1 2 : 1 5 6- 1 5 9 . Smith , E . L . & Greenshields , R . N . ( 19 7 3 ) Tower fermentation systems : Their application in continuous brewing , Biotech . Bioeng . Symp . , 4 : 5 1 9-533 . Smith , G . M . ( 1 9 7 9 ) The technology o f whey product ion , Pro c . Int . Food Sc i . (UK) , 1 2 : 1 9 1-202 . Speckman , R . A . & Collins , E . B . ( 1 98 2 ) Microbial product ion of 2 , 3-butylene glycol from cheese whey , Appl . Env . Microbial . , 43 ( 5 ) : 1 2 1 6- 1 2 1 8 . 224 Stewart , G . G . ( 1 9 7 5 ) Yeas t f locculation : Practical implications and experimental f indings , Brewers Diges t , 50 ( 3 ) : 4 2-56 , 62 . Stewart , G . G . ( 1 98 1 ) The genetic manipulation of indus trial yeas t s trains , Canadian Journal of Microbiology , 2 7 ( 1 0) : 9 73- 1 0 10 . Stewart , G . G . & Goring , T . E . ( 1 9 7 6 ) Effect o f some monovalent and d ivalent metal ions on the flocculation of brewer ' s yeast strains , J . Inst . Brew . , 82 : 34 1 -34 2 . Stewart , G . G . , Russel , I . & Garrison , I . F . ( 1 9 7 5 ) Some consideration o f the flocculation characteristics o f ale and lager yeast s trains , J . Inst . Brew . , 8 1 : 248-25 7 . Sutton , D . D . & Lampen , J . D . ( 1 962 ) Localization o f sucrose and maltose fermenting systems in Saccharomyces cerevisiae , Biochimica et B iophysica Acta , 56 : 303-3 1 2 . Taylor , N . W . & Orton , W . L . ( 1 9 7 3 ) E ffect o f alkaline-earth metal salts on flocculation in Saccharomyces cerevisiae , J. Ins t . Brew . , 7 9 : 2 94-29 7 . Thayanithy , K . , Harding , G . & Wase , D . A . J . ( 1 982 ) Rearrangement of lactose on sterilization , Biotech . Let ters , 4 ( 7 ) : 423-424 . Thorne , R . S .W . ( 1 968 ) Some observations on yeast mutation during continuous fermentation , J . Inst . Brew . , 7 4 : 5 1 6-524 . Thorne , R . S .W . ( 1 9 7 0 ) Further observations on yeast mutat ion during continuous fermentation , J. Inst . Brew. , 7 6 : 555-563 . Titchener , A . L . ( 1 980 ) The Brazil ian alcohol programme and some inplications for N . Z . , Paper no . 25 , N . Z . Inst . Eng . Conf . , Dunedin , February , 30 pp . Veldkamp , H . & Jannasch , H . W . ( 1 9 7 2 ) Mixed culture studies with the chemo stat , J . Appl . Chem . Bio technol . , 2 2 ( 1 ) : 1 0 1 - 1 2 3 . Vic torero , F . A . ( 1 948 ) Continuous fermentation apparatus , US patent 2 , 450 , 2 1 8 . Wasserman , A . E . ( 1 960 ) Whey util ization IV . Availability o f whey nitrogen for the growth o f Saccharomyces fragi lis , J . Dairy Sci . , 4 3 : 1 2 3 1 - 1 234 . Wasserman ,A . E . , Hopkins , W . J . & Porges , N . ( 1 958 ) Whey utilization - growth conditions for S. fragilis , Sewage & Ind . Was tes , 30 ( 7 ) : 9 1 3-920 . Watson , E . G . & Shore , D . T . ( 1 960) British patent 9 38 , 1 7 3 . Webb , B . H . & Whittier , E . O . ( 1 9 7 0 ) Eyproducts from milk , 2nd ed . , AVI Pub . � Connecticut , USA . Webb , S . R . ( 1 9 7 1 ) Small incomplet e fac torial experiment designs for two - and three - level factors , Technometric s , 1 3 ( 2 ) : 243-256 . Wendor f , W . L . , Amundson , C . H . & OLson , N . F . ( 1 9 7 1a ) Use o f yeast S-galac to­ sidase in milk and milk products , J . Milk Fd . Technol . , 34 : 294-29 9 . 225 Wendorf , W . L . , Amundson , C . H . & Olsen , N . F . ( 1 9 7 1b ) Characterization o f S-galac tosidase from Saccharomyces fragi lis , J . Milk Fd . Technol . 34 : 300-306 . Widel l , S . ( 1 9 7 9 ) A lactosylurea whey product for feeding to ruminants , Proc . Whey products Conf . , 1 9 74 , Philadelphia , US . Dept . o f Agri . p . 5 3-6 1 . Wong , T . S . , Manderson , G . T . & Larsen , V . F . ( 1 982 ) E thanol production from wood pentose sugars , Pro c . 1 4 N . Z . Biotech . Conf . , Palmerston North , New Zealand , p . 5 9-75 . Yoo , B . W . & Mat t ick , J . F . ( 1 969 ) Utilization o f acid and sweet cheese whey in wine product ion , J . Dairy Sci . , 5 2 ( 6 ) : 900 . Yoo , B .W . ( 1 9 74 ) Effect o f certain fac tors on cell mass and ethanol product ion by lactose fermenting yeas t s , PhD thesis , University of Maryland , Maryland , USA. Yoon , H . & Blanch , H .W . ( 1 9 7 7 ) Compet ition for double growth-limiting nutrients in continuous culture , J . Appl . Chem. Bio technol . , 2 7 : 2 60-268 . Zall , R . R . , Kuipers , A . , Muller , L . L . & Marshall , K . R . ( 1 9 7 9 ) Trends in whey Processing , N . Z . J . Dairy Sci . Techno l . , 1 4 ( 3 ) : 79-85 . Zimmermann , F . K . & Scheel , I . ( 1 9 7 7 ) Mutants of Saccharomyces cerevisiae resistant to carbon cataboli te repression , Molec . gen . Gene t . , 1 54 : 75-82 . APPENDIX A FEED MEDIUM PUMP CAPACITY AND SAMPLING DATA SHEETS A . l Capacity o f the feed medium pump for the tower fermenter Fig . A . l provides pumping capacity data for the Cole Parmer Masterflex pump used for pump ing medium into the tower . Two pump head sizes were used , they were sizes 13 & 14 . 1 2 00 � 800 H u <( � u 0 400 z H � p... COLE- PARI-fER MASTERFLEX pump head si ze 1 3 head size 1 4 Model no .WZ1R051 2 4 6 SPEED CONTROLLER SETTINGS F i g . A . l Pump capa c i t y c u rves f o r 25 mm:, t O\,,( ' r f e rrnl ' T" l t l:.> r . A . 2 Sampling data sheets for a The fo llowing data sheets were used to record data collected set o f samp les at a particular medium f low rate . 226 Tower F ermenta tion C . Boontanja i Run . . . . . . . . . . Date . . . . . · · · · · · · · Med ium . . . • . . . . . . Run duration fran . . . . . . . . . . . . . . . . . . . . to . . . . . . . . . . . . . . • . . . . . Sampl ing time . . . . . . . . . . . . . . . . . . . . . . . . Tower aera t ion • . . . . . . . • . . . . . . ml /m i n . L i ne pressure . . . . . . . . . . kPa . F l ow rate . . . • . • • . . . . . . . ml /hr. Pump head no . /motor control l er setti ng . . . . . . . Duration of sampl ing . . . . . . . . . . hr . Start . . . • . . . . . • F i nish . . . . . . . . . . Vohne beer col lected . . . . • • . . • . . . . . . ml . Ev . • . . . . . . . . Feed wort aerati on . . . . • . . . . . hr . a t . . . . . . . . . . ml /hr . . . . • . . . . . • kPa . Res idence time, T R . . . . • . . • . . hr . L iqu id superficial velocity, Vs . . • . . . . • . . cm/s (fl ow (Ml /hr) x 5 . 4257 x 10- s ) Dil ution rate, o_ . . . . . • • • . • hr-1 ( 1 /TR) Col lect sampl es from the top of column f i st . HA HE VE 0 - Feed wort 0 0 1 - Bottan of tower 95 96 2 - Lower mid-tower 817 817 3 - Upper m id-tower 1 573 1 573 4 - Top of 3rd section 2316 2316 5 - Effluent beer 2693 2693 S w/v g/100 •1 . tower cross sec:tiona 1 ar• • 5. 1 1 9 1 cm2 HA - Actual hei ght, .. Ha - Effecti ve height, • VE - Effecti ve vol ume, •1 49 418 805 1186 2921 RUN NO . . . . . . . . . . . Parameters Vol ume of sampl e , ml . Temperature, °C pH SG Total cel l weight Centri fuge tube wt. ( CT) , g . C T + wet wei ght , g . CT + dried wei ght , g . Centri fuged wet wt . , % w/v Cel l dried wei ght, % w/v K. [ragi lis O i l uti on Cel l /ml K . F . cel l dried �t. , % w/v Sucrose O i l uti on Sugar analyzer S W/V Lactose Oi l uti on Sugar analyzer % W/V Ethanol Internal s td . % w/v Peaks % W/V Ethanol para l lel std. % w/v 0 1 2 feed bottom m i d- l ower 3 4 mi d - upper top 5 effl uent N N -...) Tower Fennentat ion C . Boontanja i Run . . . . . . . . . . . . . . Date . . . . . . . . . . . Med i um . • . . . . . . . . . . . . . . . . . . . . Run durat i on . . . . . . . . • . . . . . . . . days . . . . . . . . . . . . . . . . . hrs . Tower aeration . . . . • . . • . . • . • . . ml /mi n . • . . . . . . . . . . . . kPa . Feed wort aeration . . . . • • . . . . . hr . . . . • . . . . . . . . . . . . . ml /mi n . . . . . . . . . . kPa . F low rate • • . • . . . . . . . . • • . . . . . . ml /hr . Ev . . . . . . . . . . . Residence time, TR . . . . . . . . . . . . hr . Superf ic ia l l i quid veloci ty , Vs . . . . . . . . • . . . . . . . • . . cm/s . Di l ution rate ,D • • • • . . . • • . . . . hr: 1 sample pH SG total cel l �. fragi lis TR sucrose ww DW ce 1 1 /111 DW '!ow/v '!ow/v '!ow/v �r · lw/v . 0 1 2 � 3 4 �. 5 : .. -. - 0 - Feed wort 3 - Upper mid tower l actose '!ow/v 1 - Bott0111 of tower 4 - Top of 3rd section 2 - lower mid-tower 5 - Effluent beer WW - Centri fuged wet weight , g/ 100 ml . DW - Dried weight,g/100 ml . ethanol '!ow/v r. I' :'o-··· 1 ower I ennen ld l 1 on C . !loon ta nja i Run . . . . . . . . . . . . . . Da te . . . . . . . . . . . Med ium . . . . . . . . . . . . . . . . . . . . . . Run dura t ion . . . . . . . . . . . . . . . . . days . . . . . . . . . . . . . . . . . hrs . Tower aerat ion . . . . . . . . . . . . . . . m1 /mi n . . . . . . . . . . . . . . I: Pa . Feed wor t aera t ion . . . . . . . . . . . hr . . . . . . . . . . . . . . . . . . m1 /mi n . . . . . . . . . . kPa . F l ow ra te . . . . . . . . . . . . . . . . . . . . ml /hr . Ev . . . . . . . . . . . Hes i dence t i me , TR · · · · . . . . . . . . hr . Superf ic ial l i qu i d veloc i ty , Vs . . . . . . . . . . . . . . . . . . . cm/s . D i lution rate , D . . . . . . . . . . . . . hr: 1 sample pH SG tota 1 cel l IK . [raEi l i s TR sucrose l actose ww DW cel l /ml DW 'l.w/v 'l.w/v '!ow/v hr lw/v '!ow/v 0 1 ! 2 ! - i I 3 , 4 . 5 ... 0 - Feed wort 3 - Upper mid tower 1 - Bottom of tower 4 - Top of 3rd section 2 - Lower mid- tower 5 - Effl uent beer WW - Centri fuged wet weight ,g/100 ml . DW - Dri ed wei g ht , g/100 ml . ethanol lW/V I ! N N 00 APPENDIX B FERMENTATION DATA B . l MEDIUM OPTIMIZATION DATA Variables and levels Three variables were used at three different levels of concentrations ( 33 design , 14 runs ) . Table B . l Variables and their concentrations used at various RUNS . level 0 1 2 (NH4 ) 2S04 N 0 0 . 5 1 . 0 K2HP04 K 0 0 . 5 1 . 0 Yeast extract y 0 0 . 1 0 . 2 All concentrations in g/1 RUNS 000 002 0 1 1 020 022 1 0 1 1 10 1 1 2 1 2 1 200 202 2 1 1 220 222 Table B . 2 Fermentation results o f factorial experiment on nutrient requirement of K. marxianus UCD FST 7 15 8 growing on whey permeate . lactose concentration ( g /1 ) after ethanol concentration ( g /1 ) after runs 7 . 5 1 9 . 5 30 48 66 . 5 (h) 7 . 5 1 9 . 5 30 48 66 . 5 000 33 14 4 . 0 3 . 6 2 . 7 3 . 3 1 1 19 1 9 18 002 3 3 1 7 4 . 3 3 . 3 2 . 6 3 . 0 8 . 7 19 19 1 8 0 1 1 36 1 2 3 . 9 3 . 6 2 . 7 3 . 3 1 1 1 8 1 8 1 8 020 3 3 1 4 3 . 9 3 . 7 2 . 8 3 . 3 1 2 1 8 1 8 1 7 022 34 1 7 4 .4 . 3 . 5 2 . 7 3 . 1 1 1 1 8 1 8 1 7 1 0 1 2 9 1 9 3 . 6 3 . 4 2 . 6 2 . 8 8 . 3 1 8 1 7 1 9 1 10 33 1 3 4 . 0 3 . 8 2 . 9 3 . 2 1 1 1 8 1 8 1 8 1 1 2 34 1 3 3 . 7 3 . 5 2 . 9 3 . 7 1 1 1 7 1 7 1 9 1 2 1 3 3 1 5 3 . 9 3 . 6 2 . 9 3 . 1 1 1 1 8 1 9 1 8 200 34 14 4 . 2 3 . 7 2 . 9 3 . 3 9 . 8 1 9 1 8 1 9 202 35 16 4 . 3 3 . 5 2 . 8 3 . 3 9 . 5 1 7 1 8 1 9 2 1 1 33 18 4 .8 3 .6 2 . 8 3 . 2 9 . 1 1 5 1 8 1 8 220 35 15 4 . 2 3 . 0 2 . 0 2 . 8 1 0 1 8 1 9 1 8 222 32 1 5 3 . 9 3 . 7 2 . 0 2 . 6 10 19 19 19 The initial lactose concentration was 43 g/ 1 . 2 29 230 B . 2 10 1 BATCH FERMENTATION OF WHEY PERMEATE (40 g/ 1 lactose) by K. marxia:nus Y42 E e thanol concentration , g/1 sl lactose concentration , g / 1 L!.Sl lactose utilization , g / 1 81u lactose utilization , % T f ermentation time , h X total cell number , cell/ml y yield of ethanol on lactose utilized , % Table B . 3 Batch fermentation o f whey permeate by K. marxianus Y42 T pH X sl L!.Sl 81u E y h cell /ml g / 1 g/ 1 % g /1 % 0 5 . 0 7 40 0 0 1 . 1x10 - - 3 5 . 0 2 . 4x 10 7 3 7 3 . 1 7 . 8 1 . 6 96 6 4 . 9 4 . 8x10 7 34 6 . 2 1 6 3 . 8 9 5 9 4 . 8 8 . 0x10 7 2 3 1 7 4 1 7 . 8 ' 89 1 2 4 . 7 1 . 1x 10 8 1 4 26 65 14 98 1 6 4 . 7 1 . 1x 10 8 4 . 1 36 90 18 94 24 4 . 6 1 . 1x10 8 2 . 0 38 95 18 90 Fermentation was completed a fter 16 h. There was 90 % lac tose utilization to produce 18 g/1 ethanol . This was 94 % yield of ethanol . The average rate o f lac tose utilization was 2 . 2 g /lh and ethanol productivity was 1 . 1 g /lh . The maximum specific rate of e thanol produc tion occurred between 9 and 12 h and was 1 . 4 g /gh ( cell dried weight = 1 . 4 g / 1 from f igure B . 1 ) . 23 1 B . 3 TOWER FERMENTATION OF WHEY PERMEATE ( 40 g/1 lactose) Data for the superfic ial liquid velocity (V ) of 0 . 044 , 0 . 080 , s 0 . 1 7 , & 0 . 24 mm/s were averaged from 5 separate sets of samples . For the superficial liquid velocity of 0 . 30 mm/s , the data were averaged from only 2 sets o f samples . - 1 D Dilution rate based on tower straigh t section , h H Tower height , m H* Mean tower height at the mid-point of each tower sec tion , m Q Feed medium volumetric f low rate , ml/h SG Exit specific gravity e T r T* r Residence time at various tower heigh t , h Mean res idence time at height of H* T . Residence time at height H , h r1 X Mean a cell concentrat ion in a tower sec tion , g/1 DW H* , T� and Xa are used in conj unct ion with Si , q1 , E ' , and v . Table B . 4 A summary of the d imens ions o f the tower fermenter . samp le point VE HE average H* f:..V ml m between m ml 0 ( inlet) 0 0 1 49 0 . 09 6 0-1 0 . 048 49 2 4 18 0 . 82 1-2 0 . 46 369 3 805 1 . 5 7 2-3 1 . 20 387 4 1 1 86 2 . 32 3-4 1 . 9 5 381 5 ( exit) 2921 2 . 69 4-5 2 . 51 1 7 35 In tables B . 5 and B . 6 , the s tandard deviations ( SD) for the data are given after a s lash following each datum . 232 Table B . 5 Tower fermentation of whey permeate, ,data at various sampling points H m T . r1 h pH ww g/ 1 DH g/ 1 s su % V = 0 . 044 , Q = 8 1 , D = 0 . 063 , T = 1 5 . 9 , SG = 1 . 003 s r e 0 0 . 096 0 . 82 1 . 5 7 2 . 32 2 . 69 0 0 . 6 5 . 2 1 0 . 0 1 4 . 7 36 . 3 5 . 4 / 0 . 04 4 . 6 / 0 . 1 3 4 . 6 / 0 . 1 3 4 . 6 / 0 . 1 5 4 . 6 / 0 . 1 6 4 . 6 / 0 . 1 2 2 5 / 2 . 1 348/ 6 . 6 9 2 / 14 . 1 7 8 / 14 . 8 54 / 1 1 . 6 1 8 / 2 . 0 2 / 0 . 1 9 7 / 20 . 1 2 1 / 4 . 5 1 6 / 4 . 0 1 0 / 3 . 0 0 . 2 / 0 . 1 4 7 / 0 . 5 7 . 0/ 3 . 0 1 . 9 / 1 . 3 1 . 3 /0 . 7 1 . 3 / 0 . 6 1 . 3 / 0 . 4 0 85 96 97 97 9 7 V = 0 . 080 , Q = 1 5 0 , D = 0 . 1 2 , T = R . 7 , S G = 1 . 004 s r e 0 0 . 096 0 . 82 1 . 5 7 2 . 32 2 . 69 0 0 . 3 2 . 8 5 . 5 8 . 1 1 9 . 8 5 . 1 / 0 . 1 8 4 . 7 / 0 . 06 4 . 5 / 0 . 02 4 . 5 /0 . 06 4 . 5 / 0 . 04 4 . 6 /0 . 06 1 9 / 4 . 0 344 / 22 . 0 5 5 / 8 . 5 39/ 5 . 0 2 8 / 2 . 8 1 6 / 0 1 / 0 . 2 1 0 1 / 1 1 . 8 1 0 / 2 . 3 6 / 1 . 2 4 / 0 . 8 0 . 2 / 0 4 3 / 2 . 3 28 / 2 . 9 3 . 9 / 1 . 1 3 . 4 / 0 . 8 3 . 3 /0 . 9 3 . 2 / 0 . 9 0 3 5 9 1 9 2 9 2 9 3 V = 0 . 1 7 , Q = 3 1 0 , D = 0 . 25 , T = 4 . 1 , SG = 1 . 004 s r e 0 0 . 096 0 . 82 1 . 57 2 . 32 2 . 69 0 0 . 096 0 . 82 1 . 5 7 2 . 32 2 . 69 0 0 . 096 0 . 82 1 . 57 2 . 32 2 . 69 0 4 . 9/ 0 . 1 5 0 . 2 4 . 8 /0 . 1 1 1 . 3 4 . 6 / 0 . 05 2 . 6 4 . 6/ 0 . 07 3 . 8 4 . 5 / 0 . 07 9 . 3 4 . 6 / 0 . 09 1 0 / 2 . 0 0 . 6 / 0 . 1 34 1 / 7 1 . 7 9 6 / 2 1 . 8 1 20 / 8 . 4 30/ 1 . 6 1 08 / 1 0 . 6 2 6 / 2 . 5 9 2 / 9 . 1 2 2 / 2 . 4 1 0 / 2 . 6 0 . 2 / 0 . 1 4 6 / 7 . 3 3 2 /5 . 3 4 . 8 / 0 . 9 3 . 1 /0 . 6 2 . 7 / 0 . 3 2 . 5 / 0 . 4 0 32 90 93 94 95 V = 0 . 24 , s Q = 440 , D = 0 . 34 , T = 3 . 0 , r SG = 1 . 004 e 0 0 . 1 1 . 0 1 . 8 2 . 7 6 . 7 4 . 9 / 0 . 05 4 . 5 / 0 . 04 4 . 2 /0 . 06 4 . 2 / 0 . 06 4 . 2 / 0 . 04 4 . 3/ 0 . 04 8 / 2 . 6 0 . 5 / 0 . 2 2 44 / 4 . 1 6 6/ 1 1 . 4 1 4 0 / 10 . 0 36 / 2 . 3 1 36/ 7 . 6 35/ 1 . 6 1 4 0/ 1 1 . 6 36 / 2 . 5 1 0 / 4 . 0 0 . 4 / 0 . 2 4 5 / 1 . 4 33 / 1 . 9 2 . 8 /0 . 1 2 . 8 / 0 . 6 · 2 . 2 /0 . 9 1 . 9 / 0 . 9 0 2 7 9 4 94 95 9 7 V = 0 . 30 , s Q = 550 , D = 0 . 4 3 , T = 2 . 3 , r SG = 1 . 005 e 0 4 . 7 / 0 . 02 0 . 1 4 . 4 /0 . 03 0 . 8 4 . 3/ 0 . 0 1 1 . 5 4 . 2 /0 . 0 1 2 . 2 4 . 2 / 0 . 0 1 5 . 3 4 . 2 / 0 . 02 8 / 1 . 9 0 . 4 / 0 . 1 2 1 9/ 1 2 . 8 5 6 / 6 . 0 1 06 / 3 . 1 24 / 0 . 4 1 34 / 6 . 4 3 2 / 0 . 2 1 40/ 5 . 9 3 3 / 0 . 2 1 1 / 3 . 2 0 . 5 / 0 38 / 1 . 3 3 3 /4 . 3 1 6 / 3 . 1 6 . 1 / 0 . 4 3 . 8 / 0 . 5 3 . 2 / 0 . 9 0 1 4 57 84 90 92 E g/1 0 2 0 / 2 . 7 2 2 / 0 . 9 2 3 / 0 . 6 23 I 1 . 3 2 2 / 0 . 8 0 5 . 6 / 1 . 0 1 7 / 0 . 6 1 8 / 1 . 3 1 R I 1 . 4 1 7 / 1 . 5 0 6 / 1 . 4 1 8 / 0 . 9 1 8 I 1 . 7 1 9 / 2 . 0 1 9 / 1 . 5 0 4 . 1 / 1 . 3 1 6 I 1 . 6 1 8 / 0 . 3 1 7 / 0 . 6 1 9 / 1 . 0 0 3 . 2 / 0 . 2 1 1 / 0 . 4 1 6 I 1 . 1 1 7 / 1 . 1 1 8 / 0 . 7 y % 9 7 9 3 9 2 9 5 9 0 7 5 8 2 84 83 82 7 7 7 7 80 84 82 63 7 1 78 80 75 1 50 9 3 9 1 9 1 9 6 233 Table B . 6 Tower fermentation of whey permeate , data at various tower sections . H* T* X s I q l E ' \1 r a 1 m h g / 1 DW g/lh g / gh g/lh g/lh V = 0 . 044 s 0 . 04 8 0 . 3 9 7 6 6 0 . 68 3 3 0 . 3 0 . 46 2 . 9 59 1 . 1 0 . 02 0 . 4 0 l . 20 7 . 5 19 0 . 1 0 . 0 1 0 . 1 0 1 . 9 5 12 . 3 1 3 0 0 0 0 2 . 5 1 20 . 5 5 0 0 0 0 V = 0 . 080 s 0 . 048 0 . 1 7 10 1 4 5 0 . 45 1 8 0 . 2 0 . 46 1 . 6 56 9 . 6 0 . 1 7 4 . 6 0 . 1 1 . 20 4 . 1 8 0 . 2 0 . 0 3 0 . 2 0 1 . 9 5 6 . 8 5 0 0 0 0 2 . 5 1 1 4 . 0 2 0 0 0 0 V = 0 . 1 7 s 0 . 048 0 . 08 96 9 1 0 . 9 5 38 0 . 4 0 . 46 0 . 75 6 3 2 3 0 . 37 1 1 0 . 2 1 . 20 2 . 0 28 1 . 3 0 . 0 5 0 0 1 . 9 5 3 . 2 24 0 . 4 0 . 0 2 1 0 2 . 5 1 6 . 6 1 1 0 0 0 0 V = 0 . 24 s 0 . 048 0 . 06 66 1 08 1 . 64 3 7 0 . 7 0 . 46 0 . 5 5 1 35 0 . 69 14 0 . 3 1 . 20 1 . 4 36 0 0 1 . 9 0 . 1 1 . 9 5 2 . 3 36 1 . 0 0 . 0 3 0 0 2 . 5 1 4 . 7 18 0 · o 0 . 3 0 V = 0 . 30 s 0 . 04 8 0 . 04 56 60 1 . 0 7 36 0 . 6 0 . 46 0 . 4 40 2 5 0 . 6 3 1 2 0 . 3 1 . 20 1 . 1 2 8 14 0 . 50 6 . 8 0 . 2 1 . 9 5 1 . 8 3 3 3 . 3 0 . 10 1 . 7 0 . 1 2 . 5 1 3 . 8 1 7 0 . 2 0 0 . 3 0 234 B . 4 TOWER FERMENTATION OF WHEY PERMEATE ENRICHED WITH MOLASSES (40 : 60 g/1 lac tose : sucrose) The resul ts for each V 8 were averaged from two sets o f samples . e - The power of 10 of the cell concentrat ion S ' - Volume t r ic rate of total sugar util izat ion q . t - Specific rate of to tal sugar util ization ; � - K.maPxianus cell concentrat ion S t - Total sugar concentration ; XsKM - Average K. mapxianu� cell concen trat ion Table 8 . 7 Data for tower fermentat ion o f whey p e rmea te enriched w i th molasses at various samp l ing points . H Tri pH ww DW X m h g/ 1 g/ 1 cell /ml V • 0 . 08 7 , Q - 160, D • 0 . 1 3 , Tr• 8 . 0 , SG • 1 . 0 1 4 s e 0 0 5 . 0/ 0 . 05 2 3 / 9 9/ 1 - 0 . 096 0 . 3 4 . 8/ 0 . 05 530/26 160/ 1 ( 5 . 1 / 3 . 9 ) e9 0 . 82 2 . 6 5 . 0/0 . 05 550 / 2 3 164/ 8 (4 . 1 / 2 . 5 ) e9 1 . 57 5 . 0 5 . 0/ 0 . 05 634 / 7 7 185/ 2 1 ( 2 . 1 /0 . 3)e9 2 . 32 7 . 4 5 . 0/ 0 . 1 0 209/ 1 1 58/ 4 ( 1 . 3 / 0 . 3 )e9 2 . 69 1 8 . 2 5 . 0/ 0 . 20 40/ 1 8 2 / 1 ( 1 . 3/0 ) e 7 Vs· 0 . 1 2 , Q - 220, 0 • 0 . 1 7 , T • S . B , SGe • I . 020 r 0 0 5 . 0/ 0 . 05 1 8/ 1 2 / 0 - 0 . 096 0 . 2 4 . 7 /0 . 03 466/50 1 36/ 1 1 ( 3 . 8/ 0 . 4 ) e9 0 . 82 1 . 9 S . 0/ 0 . 08 4 7 4 / 2 1 4 2 / 1 (4 . 0/ 1 . 6 ) e9 1 . 57 3 . 6 5 . 1 /0 . 1 5 S68 / 7 8 1 74 /42 ( 7 . 0/ 5 . 0)e9 2 . 32 5 . 3 5 . 1 /0 . 08 2S8/47 74/ l S ( 2 . 8 / 1 . 6 )e 9 2 . 69 1 3 . 2 5 . 0/0 . 08 1 7 / 1 1 / 0 ( 9 . 4 / 0 . 7 ) e5 vs- 0 . 1 4 , Q - 260, D • 0 . 20 , Tr• 4 . 9 , SGe• 1 . 020 0 0 4 . 8 / 0 . 0S 1 8/ 1 3/ 0 - 0 . 096 0 . 2 4 . 6/ 0 . 03 4 50/ 64 1 2 4 / 2 2 ( 3 . 7 / l . 4 ) e9 0 . 82 1 . 6 5 . 1 / 0 . 05 484 / 2 2 1 4 5 / 8 ( 8 . 7 / 2 . 3 ) e9 1 . 57 3 . 1 5 . 0 /0 4 16/ 8 1 2 1 / 2 ( 7 . 2 / l . 2 ) e9 2 . 32 4 . 5 5 . 0/ 0 . 08 400/ 1 8 1 14 / 5 ( 4 . 8/4 . 0) e9 2 . 69 1 1 . 1 5 . 0 /0 . 05 1 7 I 2 2 / 1 ( 6 . 8/ 2 . 8 ) e6 � c e l l /m! - ( 2 . 4 / 2 . 0 ) e9 ( 1 . 3 / 0 . 7 ) e9 ( l . l / O . l ) e 9 ( 6 . 5/ 3 . 5 ) e8 ( l . ! /O . l ) e7 - (8 . S / 7 . S ) e8 ( 5 . 9/ S . l ) e8 ( !l . 0/7 . O) e8 ( 8 . 4 / 7 . 6 ) e8 ( 7 . 2 / l . O) e5 - ( 2 . 7 / 0 . 3 ) e8 ( 2 . 3/ 1 . 7 ) e9 ( 3 . 4 / 2 . 2 ) e8 ( 3 . 8 / 1 . 8 ) e8 (4 . 2 / 2 . 2 ) e6 s l s s t s lu s s tu E y s su g/ 1 g/1 g / 1 % i. % g / 1 % 40/ 1 . 1 58/ 1 . 0 98 - - - - - 35/4 . 9 1 3 / 1 . 3 4!l 12 78 51 1 7 / 1 . 3 64 28/0 . 8 0 . 5/0 29 2 9 9 9 7 I 36/5 . 3 9 7 1 3/2 . 0 0 . 4 / 0 l ) 68 99 !i7 4 1 / 1 . 2 90 9 . 3 /0 . 7 0 . 4 / 0 9 . 7 7 7 99 90 4 3 /0 . 3 90 8 . 1 /0 0 . 4 /0 8 . 5 80 99 9 1 4 1 / 0 . 9 86 4 1 /0 . 9 S 7 / 7 . 8 98 - - - - - 40/0 . 8 22/2 . 8 6 2 I 62 37 I S / 0 . 4 7 S 39/ 1 . 4 2 . 3 /0 . 7 4 1 3 ')U 58 2 S / 1 . 5 83 34/ 1 . 6 1 . 3/ 0 . 2 3 S 1 7 98 64 30/ 0 . 3 87 29/ l . 6 1 . 3/ 0 . 1 30 30 98 70 34 / 1 . 3 92 27 / 1 . 4 1 . 3/0 . 1 28 33 98 7 1 33 /0 . S 88 46/6 . 8 62/ 3 . S l OB· - - - - - 4 4 / 5 . 3 2 3 / 0 . 8 6 7 4 62 3 7 1 3 /0 . 7 58 39/3 . 6 2 . 2 / 0 . 3 4 1 1 6 9 7 6 2 28/ 1 . 1 76 3 2/ 1 . 1 1 . 3/0 3 3 3 1 98 70 30/ 2 . 5 74 24/2 . 1 1 . 2 / 0 . 1 2 S 4 7 98 76 33 / 3 . 5 74 2 3 / 3 . 7 1 . 1 /0 24 49 98 7 7 34 /4 . 3 7 6 Table B . B Data for tower fermentat ion o f whey permeate enriched w i t h molasses at various t ower sect ions . H* T* xaKM X S ' ql S ' q s S ' q t E ' V r a 1 s t m h g/ 1 � g/1 DW g/1h g/gh g/1h g/gh g/lh g/gh g/lh g/1h V • 0 . 087 s 0 . 048 0 . 2 1 70 160 1 6 0 . 09 148 0 . 4 3 1 64 1 . 03 57 0 . 4 0 . 4 6 1 . 5 1 34 1 62 2 . 9 0 . 02 5,. 4 0 . 03 8 . 3 0 . 05 8 . J 0 . 1 1 . 20 3 . 8 90 1 74 6 . 9 0 . 07 0 0 6 . 6 0 . 04 2 . 0 0 1 . 95 6 . 2 66 1 2 0 1 . 4 0 . 0 1 0 0 1 . 4 0 . 0 1 0 . 7 0 2 . 5 1 1 1 . 6 26 30 0. 1 0 0 0 0 . 1 0 0 0 V • 0 . 1 2 s 0 . 048 0 . 1 66 1 36 1 . 8 0 . 03 1 6 2 1 . 1 9 1 64 1 . 2 1 66 0 . 5 0 . 46 1 . 1 57 1 39 0 . 5 0 . 0 1 1 2 0 . 09 1 2 0 . 09 6 . 4 0 1 . 20 2 . 8 ss 1 58 3 . 2 0 . 06 0 . 5 0 3 . 7 0 . 02 2 . 5 0 l . 95 4 . S 64 1 24 3 . 1 0 . 05 0 . 1 0 3 . 1 0 . 03 2 . 5 0 2 . 5 1 9 . 3 33 38 0. 1 0 0 0 0 . 1 0 0 0 V • 0 . 14 s 0 . 048 0 . 1 2 3 124 8 . 6 0 . 37 208 l . 68 2 1 7 I . 7 5 68 0 . 5 0 . 4 6 0 . 9 9 3 1 3 5 3 . 9 0 . 04 I S 0 . 1 1 1 9 0 . 14 1 1 0 . 1 l . 20 2 . 3 95 1 3 3 4 . 8 0 . 03 0 . 6 0 S . 4 0 . 05 I . 6 0 1 . 9S 3 . 8 29 1 1 8 5 . 0 0 . 1 8 0 . 1 0 S . I 0 . 04 1 . 9 0 2 . 5 1 7 . 8 I S 58 0 . 1 0 0 0 0 . 1 0 0 . 2 0 235 log � 6 . 96 + 1 . 08 log DW , r 80% • • 1 07�._��������_.�--�._��������_.�--�._�� 1 10 100 ( log) DW , CELL DRIED WEIGHT , g/1 Fig . B . l K. marxianus Y42 cell plate count number vs cell dried weight . Cells were obtained during the tower fermentation of whey permeate . .-1 - 00 400 E-t 300 ILl ;3: A ILl 0 ::::> � H 200 � E-t z ILl u ....l ....l ILl 100 u � � 3 . 83 DW + 1 8 . 66 , g / 1 , r DW , CELL 99% • KM Y42 6. KM lODlO 0 se cc39 o se Yl6 • se cc39 & KM Y42 • se Yl6 & KM lODlO Fig . B . 2 Cell centrifuged wet weight vs cell dried weight . 236 B . 5 TOWER FERMENTATION OF MOLASSES ( 1 00 g/ 1 sucrose ) Data for each superficial lquid velocity were averaged from 2 separat e sets of samples except for the last two velocities a t which there was only one set of samples . Table B . 9 Tower fermentation o f molasses , �ata at various sampling points . H T . pH ww DW s s E y r1 s su m h g/ 1 g/ 1 g/ 1 % g/1 % V = 0 . 084 , Q = 160 , D = 0 . 1 2 , T = 8 . 3 , SG = l . 020 s r e 0 0 5 . 1 / 0 . 10 1 4 / 0 . 4 2 /0 . 4 9 7 / 1 2 . 5 0 0 - 0 . 096 0 . 3 5 . 0/0 . 0 1 1 7 1 / 7 . 7 5 7 / 1 . 4 3 . 4 / 0 . 3 9 7 5 1 /0 . 8 1 0 1 0 . 82 2 . 7 5 . 0/0 . 0 1 1 6 6 / 5 . 3 5 6 /0 . 8 1 . 5 / 0 . 1 99 49 / 1 . 3 9 5 l . 5 7 5 . 2 5 . 0/0 1 64 / 2 . 2 5 5 /0 . 2 1 . 3 / 0 . 1 99 4 9 / 1 . 5 96 2 . 32 7 . 7 5 . 0/0 148/ 1 0 . 1 4 9 / 4 . 7 1 . 3 / 0 . 1 99 5 1 / 2 . 6 98 2 . 69 1 8 . 9 5 . 0/0 1 7 / 2 . 2 2 / 0 . 1 1 . 3 / 0 . 1 99 49 /0 95 V = 0 . 1 7 , Q = 3 1 0 , D = 0 . 24 , Tr= 4 . 2 , SG = 1 . 020 s e 0 0 5 . 2 /0 . 10 19/ 7 . 1 4 / 2 . 4 1 02 / 0 . 7 0 0 - 0 . 09 6 0 . 2 5 . 0/ 0 . 04 1 6 6 / 2 1 . 1 5 1 /0 . 3 4 6 / 1 4 . 6 55 25 /0 . 4 82 0 . 82 1 . 4 4 . 9 /0 . 02 225 / 1 1 . 2 7 2 / 3 . 8 2 . 5 / 0 . 5 98 4 3 / 1 . 4 80 l . 57 2 . 6 4 . 9 / 0 . 0 1 2 1 8 / 7 . 7 70 /2 . 5 l . 5 / 0 99 50 / 6 . 5 9 2 2 . 32 3 . 8 5 . 0/ 0 . 0 1 1 9 5 / 6 . 3 62 / 2 . 2 1 . 4 / 0 99 4 7 / 1 . 0 8 7 2 . 69 9 . 4 4 . 9/ 0 . 0 1 1 2 / 0 . 6 2 / 0 . 1 1 . 4 / 0 99 48 /0 . 2 89 V = 0 . 20 , Q = 360 , D = 0 . 28 , Tr= 3 . 5 , SG = 1 . 020 s e 0 0 5 . 2 /0 . 02 1 4 / 0 . 1 3 / 0 1 04 I 2 . 1 0 0 - 0 . 096 0 . 1 5 . 0/0 . 0 1 1 6 3/ 1 7 . 5 4 9 / 6 . 4 5 7 / 9 . 9 45 23 / 3 . 5 89 0 . 82 1 . 1 4 . 8 /0 . 03 224 / 1 . 7 7 4 /0 . 5 3 . 6 / 0 . 7 97 5 1 /0 . 4 94 l . 57 2 . 2 4 . 9 /0 . 0 1 228/ l . 4 7 6 /0 . 6 1 . 7 / 0 . 1 98 53 / 0 . 3 9 7 2 . 32 3 . 3 5 . 0 /0 . 02 2 1 5 / 0 . 7 7 2 /0 . 4 1 . 8 / 0 . 1 98 5 1 / 2 . 3 94 2 . 69 8 . 0 5 . 0/0 . 03 1 6 / 0 . 2 3 /0 1 . 6 / 0 . 1 99 50 /0 . 2 9 1 V = 0 . 33 , Q = 600 , D = 0 . 4 7 , T = 2 . 4 , SG = l . 020 s r e 0 0 5 . 2 / 0 . 02 1 2 / 0 . 5 2 / 0 . 1 1 02 / 3 . 2 0 0 - 0 . 096 0 . 1 5 . 1 /0 . 05 68/ 9 . 5 1 6 / 5 . 2 88/ 4 . 4 1 4 6 . 2 /3 . 6 8 2 0 . 82 0 . 7 4 . 8 / 0 . 0 1 203/ 3 . 2 60/0 . 7 4 . 4 / 0 . 6 96 4 7 / 1 . 0 89 l . 5 7 1 . 3 4 . 8 / 0 . 0 1 1 9 7 / 2 . 0 5 7 / 0 . 6 1 . 8 / 0 . 5 98 46 /0 . 4 86 2 . 32 2 . 0 4 . 9 / 0 . 0 1 1 84/ 4 . 1 5 4 /0 . 4 1 . 7 / 0 . 1 98 50 / 1 . 5 9 2 2 . 69 4 . 9 4 . 9 / 0 . 02 3 1 / 1 . 0 7 / 0 . 2 1 . 6 / 0 . 1 98 4 7 /0 . 2 8 7 V = 0 . 4 6 , Q = 850 , D = 0 . 66 , T = 1 . 5 , SG = 1 . 020 s r e 0 0 4 . 9 1 5 3 98 0 0 - 0 . 096 0 . 1 4 . 9 24 4 97 0 . 4 1 . 6 744 0 . 82 0 . 5 4 . 7 106 28 54 45 18 78 l . 5 7 0 . 9 4 . 6 1 35 38 1 6 84 38 8 7 2 . 32 1 . 4 4 . 5 1 24 35 4 . 7 95 39 7 8 2 . 69 3 . 4 4 . 6 5 1 1 2 1 . 4 9 9 45 8 7 V = 0 . 56 , Q = 1 040 , D = 0 . 80 , T = 1 . 2 , SG = 1 . 030 s r e 0 0 5 . 0 1 7 3 97 0 0 - 0 . 09 6 0 . 1 4 . 9 1 6 3 96 0 . 4 0 . 6 2 7 9 0 . 82 0 . 4 4 . 9 28 5 96 0 . 7 3 6 3 7 1 . 5 7 0 . 8 4 . 7 86 22 43 56 1 2 4 0 2 . 32 1 . 2 4 . 6 103 26 2 7 72 25 66 2 . 69 2 . 8 4 . 5 48 1 1 10 90 39 83 2 3 7 Table B . lO Tower fermentation of molasses ,data a t various tower sections . H* T* X S ' qs E ' \) r a s m h g/1 DW g/lh g/gh g/lh g/lh V = 0 . 084 s 0 . 048 0 . 2 5 7 2 96 5 . 1 9 1 60 2 . 8 0 . 4 6 1 . 5 5 7 0 . 8 0 . 0 1 0 0 1 . 20 3 . 9 56 0 . 1 0 0 . 2 0 1 . 95 6 . 4 52 0 0 0 . 5 0 2 . 5 1 1 3 . 2 2 6 0 0 0 0 V = 0 . 1 7 s 0 . 048 0 . 08 5 1 3 5 1 6 . 88 1 5 5 3 . 0 0 . 46 0 . 7 5 62 37 0 . 60 1 5 0 . 2 1 . 20 2 . 0 7 1 0 . 8 0 . 0 1 5 . 4 0 . 1 1 . 95 3 . 2 66 0 . 1 0 0 0 2 . 5 1 6 . 6 32 0 0 0 0 V = 0 . 20 s 0 . 048 0 . 07 4 9 349 7 . 1 2 1 6 7 3 . 4 0 . 4 6 0 . 6 62 53 0 . 86 28 0 . 5 1 . 2 0 1 . 7 7 5 1 . 8 0 . 02 2 .. 7 0 1 . 95 2 . 7 74 0 0 0 0 2 . 5 1 5 . 6 38 0 0 0 0 V = 0 . 33 s 0 . 048 0 . 04 1 6 1 7 2 1 0 . 7 5 7 6 4 . 8 0 . 46 0 . 4 38 1 36 3 . 58 66 1 . 7 1 . 20 1 . 0 5 9 4 . 0 0 . 07 0 0 1 . 95 1 . 7 5 6 0 . 2 0 5 . 0 0 . 1 2 . 5 1 3 . 4 3 1 0 0 0 0 V = 0 . 46 s 0 . 04 8 0 . 03 4 7 . 0 1 . 75 28 7 . 0 0 . 46 0 . 3 1 6 1 00 6 . 25 39 2 . 4 1 . 20 0 . 7 33 85 2 . 58 44 1 . 3 1 . 95 1 . 2 3 7 24 0 . 66 1 . 3 0 2 . 5 1 2 . 4 24 1 . 6 0 . 07 3 . 0 0 . 1 V = 0 . 56 s 0 . 048 0 . 02 3 8 . 4 2 . 80 1 3 4 . 3 0 . 46 0 . 2 4 0 . 8 0 . 1 6 5 . 9 1 . 5 1 . 20 0 . 6 1 4 142 1 0 . 5 1 24 1 . 8 1 . 95 1 . 0 24 4 3 1 . 7 9 34 1 . 4 2 . 5 1 2 . 0 1 9 1 0 0 . 54 8 . 2 0 . 4 APPENDIX C TOWER FERMENTATION OF MOLASSES Tower fermentation of molasses was performed using S . cerevisiae AWRI 350 ( FT 1 46 or SC 1 46 ) . The results of the tower fermentation of molasses have been included here in order to provide a comparison for the tower fermentation of whey permeate ( Chapter 4 ) and whey permeate enriched with molasses ( Chapter 5 ) . The data used for plo tting the graphs are given in Appendix B . 5 . C . 1 THE RELATIONSHIP BETWEEN TOWER HEIGHT AND VARIOUS FERMENTATION PARAMETERS C . 1 . 1 Sucrose and ethanol concentrations ( S , E ) s Sucrose concentration ( S ) ( fig . C . 1 a ) showed a general decrease s as height increased at a constant superficial l iquid velocity . The rate at which the concentration reduced with an inc rease in height decreased as the velocity increased . The height at which sucrose reduced to less than 1 0 g/1 ( 90% utilization) was higher at greater velocity ( table C . 1 ) . Sucro se concentration reduced to less than 4 g/1 ( 96% utilization ) at 0 . 82 m for all velocities up to 0 . 33 mm/s . Sucrose concentration decreased to less than 1 0 g/1 at the height of 2 . 32 m and the exit when the velocities were 0 . 46 and 0 . 56 mm/s , respectively . Cell wash out was observed at these two velocities ( sect . c . 1 . 4 and C . 3 . 3 ) . In contrast ethanol concentration ( E ) increased as the height increased , with some exceptions , at a constant velocity ( fig . C . 1 b ) . The concentration increased rapidly to 5 1 g/1 as the height inc reased to 0 . 096 m at the lowest velocity and then remained relatively constant over the remaining heights in the tower . For veloci ties of 0 . 1 7 and 0. 20 mm/s , the concentration increased rapidly to a value greater than 46 g/1 as the height increased to 0 . 82 m and then remained relatively constant over the remaining heights . For velocity of 0 . 33 mm/s , the concentration inc reased to 46 g/1 when the height was 1 . 57 m and was relatively constant over the remaining tower sec tions . Ethanol concentration inc reased slowly and peaked at 45 and 39 g/1 at the tower exit when the velocities were 0 . 46 and 0 . 56 mm/s , respectively . 238 1 00 S , SUCROSE s ' - - - � ., ,� \ 80 \\ \ l I l I \ � �\ -00 \ � 60 ? , en I I \ 0 � f \ u ::::> en \ �(1)40 \ \ en <\' \ \ \ \ \ � - 33 \� . 20 20 0 . 1 7 """' ', ', ,,� 0 . 084 ................. � ( b ) E , ETHANOL 50 40 30 10 239 3 V nnn/s D h- 1 Q ml/h s • 0 . 084 0 . 1 2 155 ' o --0. 1 7 0 . 2 4 3 1 0 "' l:t. ---- 0 . 20 0 . 2 8 365 'f' --- 0 . 33 0 . 47 600 \ 0 -·-0 . 46 0 . 66 850 "!· --- 0 . 56 0 . 80 1035 \ \56 \. • ""' " ' '\. "........... "' • ' ' "" "-&.:6 ............. . ........_ . - - o . sg/ 1 . 5 TOWER HEIGHT , m - ·--.o.._ . / Fig . C . 1 (a) Sucrose and (b ) ethanol concentrations vs tower height at various superficial liquid velocities . 240 Table C . 1 Percentage sucrose uti li zation at various heights in the tower and supe rfic ial liquid velocities . location/height S ( % ) , at various V (rrun/s ) so s m 0 . 084 0. 1 7 0 . 20 0 . 33 0. 46 0 . 56 1 0 . 096 97 55 45 1 4 0 . 4 0 . 4 2 0 . 82 99 98 97 96 45 0 . 7 3 1 . 57 99 99 98 98 84 56 4 2 . 32 . 99 99 98 98 9 5 72 5 2 . 69 99 99 99 99 99 90 The results showed that for veloci ties up to 0 . 33 mm/s the high cel l concentration used reduced 1 00 g/1 sucrose to less than 5 g/1 wi thin the first 0 . 82 m of the height in the tower . Most ethanol was also produced up to this height ( 0 . 82 m ) . At heights greater than this there was little fermentation taking place . At velocities greater than this ( 0 . 46 and 0 . 56 mm/s ) fermentation occurred throughout the tower . The highest ethanol concentration was measured at the exit . At these two velocities ce ll wash out was observed and thus the data were not at steady state ( sect . C . 1 . 4 ) . C . 1 . 2 Rates of sucrose utilization and ethanol production ( a ) Volumetric rates There was a generally rapid reduction of the volumetric rates of sucrose utilization ( S� ) and ethanol production (E ' ) ( fig . C . 2 ) from high values , at 0 . 048 m, to less than 5 g/lh as the mean height increased for all velocities up to 0 . 33 mm/s . The mean height at which this occurred was 0 . 048 m for the lowest velocity (0 . 084 rrun/s ) and was 0 . 46 m for all other velocit �es up to 0 . 33 mm/s . The volumetric rates were less than 5 g/lh at heights greater than these . For velocities greater than 0 . 33 mm/s ( 0 . 46 and 0 . 56 mm/ s ) , the volumetric rates increased with the mean height to a peak and then decreased with further height increase . The maximum volumetric rate of sucrose utilization was 1 00 and 1 42 g/lh and occurred at mean heights of 0 . 46 and 1 . 20 m for the two respective velocities . The maximum volumet ric rate of ethanol production of 44 and 34 g/lh for these respective veloci ties was observed at greater mean height of 1 . 20 and 340 .. � 300 H � � 2 80 .....:1 H E-< ;::l 1 60 � U) � 1 40 u ;::l U) � 1 20 0 � E-< 1 00 � u H ex: E-< � .....:1 � 80 60 .. 40 - CIJ U) .. 20 � 1 60 H E-< u ;::l 0 � p.. s � � � 0 - � 140 1 20 1 00 80 60 40 20 24 1 (a) S ' s ' SUCROSE UTILIZATION \ V mm/ s s • 0 . 084 0 -- 0. 1 7 6 ---- 0 . 20 �\ � --- 0 . 33 0 -·- 0 . 46 • --- 0 . 56 � { \\ ' " '• �'\ �' \ I , \' \ , ' \ 1 �-\- I \jJ� S6 .�\\ \j' - o . ""' '\, � ·\ · � '\ h'J 'i) . 33 . • 5 Q ml/h 155 3 1 0 365 600 850 1035 /' ' , 0 . 2� " ' "' ........... ' ' o . ...... 0 . � � "· �· �----��=-=---.-- ETHANOL PRODUCTION \ �\ \ \ \ \ \ -- �-� \\\ �33 \\ ', 0 . 46 · D ,\ - · - · - "' ' . -· .4J 20 "" �- ----- 'Q 56 . 0 . 0 84 b '-..... -�· - � " · ' - -- �i' ""' ... 0 . 5 . 0 MEAN TOWER HEIGHT , m Fig . C . 2 Volumetric rates o f ( a) sucrose utilization and (b) ethanol production vs mean tower height at various superficial liquid velocities . 24 2 1 . 95 m respectively . This showed that there was high fermentative activity inside the tower between 0 . 096 and 1 . 57 m as a resul t of the yeast mass being lifted up by the medium from lower heights . The results showed that , for all velocities up to 0 . 33 mm/ s , mos t sucrose uti lization and most ethanol production occurred rapidly within heights up to 0 . 82 m . There was little sucrose uti lization and ethanol production at heights greater than this . For velocities greater than 0. 33 mm/s , there was sucrose uti lization and ethanol production throughout the entire tower but the maximum volumetric rates were lower than those observed at lower veloci ties because of lower biomass concentration ( sec t . C . 1 . 4 ) . (b) Specific rates The specific rates of sucrose utilization (� ) and ethanol production ( v ) ( fig . C . 3 ) dec reased rapidly with mean height inc rease for all velocities up to 0 . 33 mm/s . The trend was similar to the corresponding volumetric rates ( sec t . C . 1 . 2 a ) . The specific rates decreased from high values to less than 0 . 1 g/gh with mean height increase . A slightly different profile was observed for the specific rate of ethanol production when the velocity was 0 . 46 mm/ s . At this velocity , the specific rate of ethanol production was not zero until the mean height was 1 . 95 m . This was different from the profile observed for the volumetric rate at this velocity ( fig . C . 2 b ) . Thus , there was a general trend of decreasing volumetric and specific rates with mean height increase for al l velocities up to 0 . 33 mm/s and up to 0 . 82 m . Above 0 . 82 m , all rates were very slow . c . 1 . 3 Ethanol yield ( Y ) For all superficial liquid veloci ties except the two highest velocities (0 . 46 and 0 . 56 mm/s ) , the ethanol yield ( fig . C . 4 ) showed only minor change between 0 . 096 and 2 . 32 m . It varied between 78 and The yield at the lowest sampling point ( 0 . 096 m) was very high (between 280 and 740% ) for the two velocities of 0 . 46 and 0 . 56 mm/s . For the veloci ty of 0 . 56 mm/s , the yield decreased to a low value of 40% at height of 1 . 57 m and then increased to 83% when the effluent left the tower (2 . 69 m) . The fluctuations observed were due to back-mixing of ethanol into the feed medium which contained high concentration of sucrose , at the bot tom of the towe r , and analytical � 0 � � u H � H u � p... (/) .. ;::> .. z 0 H � N H .....:l H � u H � H u � p... (/) .. Cl) 0' ..s:: Cl() - Cl() .. z 0 H E-< u ::::> j::l 0 p:: p... .....:l 0 � E-< � 243 UfiLIZATION 4 V rrrm/s s • 0 . 084 o-- 0 . 1 7 /:,. --- 0 . 20 · - - - 0 . 33 D-·- 0 . 46 ·--- 0 . 56 MEAN TOV7ER HEIGHT , m Q ml/h 155 3 1 0 365 600 850 1035 2 . 5 Fig . C . 3 Specific rates o f (a) sucrose utilization and (b ) ethanol production vs mean tower height at various superficial liquid velocities . 244 uncertainty ( between 1 2 and 85% of the yield value , sec t . E . 3 ) at high sugar and low ethanol concentrations . ,.-..... '"Cl M Q) ·� :>. M (\j (.) ·� .j.J Q) H 0 Q) .c .j.J 4-4 0 � � � H :>-< � 0 z � H � � :>-< l OO 80 60 40 20 2 3 V rnm/s Q ml /h s • 0 . 084 o - - 0 . 1 7 Ll --- - 0 . 20 � -- - - 0 . 33 o- ·- 0 . 46 ·---0 . 56 0 . 5 155 3 10 365 600 850 1 0 35 1 . 0 1 . 5 TOWER HEIGHT , m 4 5 0 . 084 2 . 0 2 . 5 Fig . C . 4 Ethanol yield (% of theoret ical yield based of sucrose utilized) vs t ower height at various s uperficial liquid velocities . The results showed that the yield changed lit t le as the height increased for all veloci ties below wash out . The variations that occurred were due to experimental uncertainties . C . 1 . 4 Cell concentration The cell concentration ( fig . C . 5 ) increased as the height in the tower increased from 0 . 096 to 0 . 82 m for all velocities except the mwest velocity ( 0 . 084 mm/s ) . As the height inc reased further to 2 . 32 m , the concentration was relatively constant at 74 , 74 and 57 g/1 DW (2 1 0 , 220 and 200 g/1 WW) for velocities o f 0 . 1 7 , 0 . 20 and 0 . 33 mm/s , respectively . For a higher velocity of 0 . 46 mm/s , the concentration increased to a maximum of 38 g/1 DW ( 1 35 g/1 WW) at height of 1 . 57 m and then reduced to 35 g/1 DW ( 1 24 g/1 WW) as the height increased further to 2 . 32 m . The trend was slightly different for the highes t .-t -bO � � 100 c.!l H f:i � � H p::: � .....:l .....:l � u � ::;,: � � c.!l H 80 60 40 20 � 250 E-< f:i 200 � � g 1 50 .,.. H p::: � 1 00 � u .....:l 50 .....:l � u 1 (a) 245 4 V mm/ s s Q ml/ s V mm/s s Q ml / s • 0 . 084 155 • 0 . 33 600 o- - 0 . 1 7 3 10 0 - · - 0 . 46 850 t::. --- - - 0 . 20 365 · - - - 0 . 56 1035 -b--0 . 20 �-=-=--=- --<>--:---=-- - - -A ..-.:::::::- -to . 30 +o . r 7 - --�'-� - - - -- - - - -- -� ' ..... ' \ / 0 . 084 · - ·-- :"\. \ ---· ·-----o-ra . 46 · -.c..._ "'&� __.o-- --- -�,� _......- ·-· ..--- - ---o.-s 6 .... - _...-- - - - - - - �------ -�0�0 -:;:::::..::::- - - - A .,...._;:::..:::::; __._-- - -�- 0 . 1 7 _ _ - � - �?' , .., - 1·0':30 - -- -,� 7 • • _+0 . 084._. ____ ---- '�� .-----·- """'{] +o . 46 ·- -a....' � �·- �- - ·� � �..... � - · 0 . 56 �· - � _ ___..- � 0 . 5 1 . 0 1 . 5 2 . 0 2 . 5 TOWER HE ICHT , m F i g . C . 5 Ce l l concent rat i on v s t owe r he i gh t a t va r i ou s supe r f i c i a l l iquid ve l oc i t ies . 246 velocity of 0 . 56 mm/s , the concentration increased to the highest value of 26 g/ 1 DW ( 1 03 g/1 WW) at a height of 2 . 32 m . As the height inc reased further to the exit , there was a rapid reduction of the concentration to between 2 and 1 2 g/ 1 DW ( 1 6 to 5 1 g/ 1 WW) for all veloci ties . The cel l concentration was greater within heights between 0 . 096 and 2 . 32 m than within the moderately flocculent nature SC 1 46 ) . The cel l floes were heights of the lifted up to 0 . 096 m because of the yeast S. cer ev isiae FT1 46 ( or up to greater tower heights by carbon dioxide and the upward movement of the medium . For these two higher velocit ies of 0 . 46 and 0 . 56 mm/s , a considerable number of the yeast floes were lifted up from the bo t tom of the tower by the medium . This had resul ted in the cell concentration being high between 1 . 57 and 2 . 32 m but lower than the greatest concentration at the lower velocity . There was a s low cell wash out . This was shown by the increase in the cell concentra tion at the exit to 1 2 g/ 1 DW ( 5 1 g/1 WW) . There was an almost linear decrease in the cell concentration from 0 . 096 to 0 . 82 m at the lowest velocity of 0 . 084 mm/s . The cell concentration fol lowed the general trend by decreasing rapidly to 2 g/ 1 DW ( 1 7 g/ 1 WW) as the medium left the tower at the exi t ( 2 . 69 m ) . This showed that the medium flow and the gas production had not lifted the yeast plug off the bot tom of the tower at this velocity but the upflow movement was abl e to suspend a large number of the yeast floes inside the tower as shown by the almost constant cell concentration between 0 . 096 and 2 . 32 m . Thus , for a l l veloci ties up to 0 . 33 mm/s , the cell concentration between 0 . 82 and 2 . 32 m was constant and was greater than at the lowest height of 0 . 096 m . At velocities greater than this , the yeast bed was lifted by the medium and carbon dioxide to the upper sec tion of the tower and cells were slowly washed out of the tower. C . 1 . 5 Medium pH The fermenta tion broth pH ( fig . C . 6 ) showed a small decrease from between pH 4 . 9 and pH 5 . 1 as the height in the tower increased to 0 . 096 m for all velocities and then was relatively constant ( between pH 4 . 8 and 5 . 0 ) as the height increased further to 2 . 32 m and the exit for all velocities except 0 . 46 and 0 . 56 mm/s . 247 pH 2 3 4 5 -- ·- ·- 0 . 5 1 . 0 1 . 5 2 . 0 2 . 5 TOWER HEIGHT , m V mm/s Q ml /h V mm/s Q ml /h s s • 0 . 084 1 5 5 'f" ---0 . 33 600 o - - 0. 1 7 3 1 0 o - · - 0 . 46 850 t:. - - - - - 0 . 20 36 5 ·- - - 0 . 56 1 035 Fig . C . 6 Medium pH vs tower height at various superficial liquid velocities The initial pH decrease indicated that there was high fermentative act ivity within these sec tions . At greater heights , there was l ittle activity since most sucrose was uti lized when the height of 0 . 82 m was reached so that there was no change in the pH . At the two higher velocities of 0 . 46 and 0 . 56 mm/s , sucrose was utilized throughout the entire tower . Hence , there was a slow but continuous pH decrease to between pH 4 . 5 and 4 . 6 as the height increased to 2 . 32 m and the exit ( 2 . 69 m ) . Thus , for all velocities below wash out , there was lit t le change in the medium pH as the height increased except for the ini t ial reduction over the first 0 . 096 m o f the tower due to the high fermentative ac tivity within this section . There was a slow pH dec rease throughout the tower at the wash out velocities ( 0 . 46 and 0 . 56 mm/s ) because there was ac tive fermentation taking place over the entire tower. 248 C . 2 THE EFFECT OF THE RESIDENCE TIME ON VARIOUS FERMENTATION PARAMETERS C . 2 . 1 Sucrose and ethanol concentration Sucrose concentration ( Ss ) ( fig . C . 7 a) dec reased rapidly as the residence time inc reased . The sucrose concentration was reduced to less than 4 g/1 ( greater than 96% utilization ) in 1 . 5 hours for all velocities used . There was very small reduction after this time . In contrast , ethanol concentration ( E ) increased rapidly from 0 g/1 to greater than 49 . g/1 as the residence time inc reased from 0 to 2 hours ( fig . C . 7 b ) . The concentration increased slightly as the residence time inc reased further beyond this time . The results showed that sucrose was almost all uti l i zed in 1 . 5 hours and most ethanol was produced in 2 hours for all velocities used . This occurred as a result of the biomass concentration used . c . 2 . 2 Rates of sucrose utilization and e thanol production The volumetric rates of sucrose utilization (S � ) and ethanol production ( E ' ) ( fig . C . 8) reduced very rapidly from high values ( 350 and 1 67 g/lh respectively) to less than 1 g/ lh as the mean residence time increased from 0 . 1 to to 1 . 5 h ours for the volumetric rate of sucrose uti lization ( fig . C . 8 a ) and 1 . 2 hours for the volumetric rate of ethanol production ( fig . C . 8 b ) . The volumetric rates were less than g/lh when the mean residence time was greater than 1 . 5 hours . A similar trend was observed for the specific rates ( qs and v ) ( fig . C . 9 ) . Both specific rates decreased rapidly from high values ( 1 0 . 8 and 7 . 0 g/gh respectively ) to less than 0 . 1 g/gh during the same time period . The results showed that most suc rose utili zation and ethanol production occurred during the initial 1 . 5 hours in the tower . There was lit tle sucrose uti li zation after 1 . 5 hours and little ethanol production afte r 1 . 2 hours for all veloc ities used . The very high cel l concentration used had resulted in very high volumetric rates of sucrose utilization and ethanol production . The specific rates decreased with an increase in the mean residence time because the biomass was recycled and changed lit tle while sucrose concentration decreased as the mean residence time inc reased . 100 80 ...... -bO .. rz:l 60 tl.l � u :::> tl.l .. tl.lrtJ 40 20 50 40 .. ....:l 0 30 � � 20 1 0 (a) S , SUCROSE s • 249 V m.m/ s s • 0 . 084 0 0 . 1 7 6 o . 20 ... 0 . 33 0 0 . 46 • 0 . 56 Q ml/h 1 55 3 10 365 600 850 1035 � � . _, - � - - - - - -. - - - - - -• • I t 0 o' 2 hours I 0 0 I I I I I e • 1 ETHANOL 2 3 4 5 T , RESIDENCE TIME , h r 6 7 Fig .C . 7 (a) Sucrose and (b) ethanol concentrations vs residence time at various superficial liquid velocities . 340 -'= ..-! "bii320 .. z S 3oo � N H 280 ...:I H H p 160 rz:l Cf.l 0 fj 140 p Cf.l � 120 rz:l H � 100 (.) � 80 H � 60 ...:I � � .. Ill 40 Cf.l 20 .. � 160 H H g 140 � � � 120 ...:I 0 � 100 � rz:l rz.. 80 0 rz:l � 60 � 40 r:.:: H � 20 ...:I � 250 (a) S ' SUCROSE UTILIZATION s ' ' ' \ •• \ \ \ \ ' CtJ \ \ � \ \ 0 � b \ 1 . 5 hours ,+ (b) E ' , ETHANOL PRODUCTION \ \ \ \ \ \ \ � \ 0 0 ' • � \ 0 • 1 1 . 2 hours 2 3 4 V mm/s s • 0 . 084 0 0 . 1 7 !:. 0 , 20 • 0 . 33 [] 0 . 46 • 0 . 56 5 T� , MEAN RESIDENCE TIME , h 155 3 1 0 365 600 850 1 035 6 Fig . C . 8 Volumetric rates of (a) sucrose utilization and (b) ethanol production vs mean residence time at various superficial liquid velocitiee ..c Cl() - Cl() .. z 0 H E-t < N H ...:l H E-t ::::> J>;:l tr.l � u ::::> tr.l � 0 J>;:l � c..:> H � H u J>;:l p.. tr.l .. tll 0" ..c Cl() - 1 0 8 6 4 • \ \ 2 25 1 (a) q , SUCROSE UTILIZATION s V nnn/s Q ml/h s • o. 084 155 0 0 . 1 7 3 10 t:,. 0. 20 365 ... 0 . 33 600 0 0 . 46 850 • 0 . 56 1035 0 \ '• \ \ (b) v , ETHANOL PRODUCTION � Cl() 6 0 1 1 . 2 HOURS 2 3 4 T* , MEAN RESIDENCE TIME, h r 5 6 Fig . C . 9 Specific rates of (a) sucrose utilization and (b ) ethanol production vs mean residence time at various superficial liquid velocities . 252 C . 3 THE EFFECT OF THE SUPERFICIAL LIQUID VELOCITY ON VARIOUS FERMENTATION PARAMETERS c . 3 . 1 Sucrose and ethanol concentrations Sucrose concentration showed a general increase with velocity increase with some exceptions ( fig . C . 1 0 a ) . At a height of 0 . 096 m , the suc rose concentration increased from 3 to 96 g/1 as the velocity inc reased from 0 . 084 to 0 . 56 rnrn/s ( curve 1 ) . At heights greater than this the concentration was less than 5 g/1 as the velocity increased to 0. 33 rnrn/s and then increased with further increase in the velocity to 0. 56 rnrn/s ( curve 2 , 3 and 4 ) . At the exit ( curve 5 ) , there was an increase in the concentration at velocity greater than 0 . 46 rnrn/s . In contrast , ethanol concentration ( fig . C . 1 0 b ) a t a constant tower height , decreased with an increase in the velocity with some exceptions . At a height of 0 . 096 m , ethanol concentra tion decreased rapidly from 51 to 0 . 6 g/1 as tpe velocity increased from 0 . 084 to 0 . 56 rnrn/s as a result of the reduction of sucrose utilization with velocity increase . At heights greater than this ( curves 2 , 3 and 4 ) , ethanol concentration was high ( 50 g/1 ) and was not affected by an increase in the velocity to 0 . 20 rnrn/s . As the velocity increased further to 0 . 56 rnrn/s , the ethanol concentration decreased with an increase in the velocity. The ethanol concentration was not affected by the veloci ty below 0 . 20 mm/s because at these velocities sucrose was almost all utilized below 0 . 82 m . Thus , ethanol concentration was at its highest value wi thin the heights above 0 . 82 m. Then as the velocity increased further beyond 0 . 33 mm/s , ethanol was produced throughout the enti re tower because sucrose was being utili zed throughtout the tower . At the exit ( curve 5 ) , the ethanol concent ration did not dec rease with velocity until a high velocity of 0 . 46 rnrn/s was reached . The results showed that the reduction in the fermentation time and cel l concentration with velocity increase ( fig . C . 1 3 ) had resulted in an increase in the sucrose concentration at the height of 0 . 096 m . The trend at greater ' heights followed that for 0 . 096 m when the velocity was greater than 0 . 33 mm/s because there was a reduction in cell concentration at points below 0 . 82 ( fig . C . 1 3 ) . Hence , sucrose was not all uti li zed at the bot tom of the tower and thus the sucrose concentration increased with velocity . As a result of this less ...-l - 80 00 60 ...-l - 20 50 40 00 30 � s � � 20 1 0 HE m • 1 0 . 096 0 2 - - 0 . 82 1:. 3 ---- 1 . 5 7 . ... 4 -- - 2 . 32 0 5 - ·- 2 . 69 0 . 1 0 . 2 253 0 . 3 0 . 4 I I I 13 /' / � / / / < ..c . 0 \ 5 "t \ 0 . 5 \, V , SUPERFICIAL LIQUID VELOCITY , mm/s s Fig . C . lO (a) Sucrose and (b ) ethanol concentrations vs superficial liquid velocity at various tower heights . 254 ethanol was produced in the lower tower sec tion as the velocity increased . Thus , the velocity of 0 . 33 mm/s was the limiting velocity for complete uti lization of sucrose ( even though sucrose was reduced to less than 5 g/ 1 at the height of 2 . 32 m for veloci ty of 0 . 46 mm/ s ) since cell wash out occurred at this velocity . The highest concentration of ethanol reached for velocities up to 0 . 33 mm/s and height of 0 . 82 m was 46 g/ 1 . Ethanol concentration was less than this at greater heights and velocities . C . 3 . 2 Rates of sucrose utili zation and ethanol production ( a ) Volumetric rates ( S� , E ' ) The volumetric rates of sucrose utilization ( S� ) and ethanol production (E ' ) ( fig . C . 1 1 ) , at heights below 0 . 82 m ( curves 1 and 2 ) , increased with an increase in the velocity to a peak and then decreased with further velocity increase . Maximum rates of sucrose util ization of 35 1 and 1 36 g/lh were observed at veloci ties of 0 . 1 7 and 0 . 33 mm/s for mean heights of 0 . 048 and 0 . 46 m respectively Similarly maximum rates of ethanol production of 1 67 and 67 g/lh were observed at velocities of 0 . 20 and 0 . 33 mm/s for these two mean heights respectively . For greater heights ( curves 3 and 4 ) , the volumetric rates were not affected until the velocity was greater than 0. 33 mm/s . Above this velocity , the volumetric rates increased with velocity . At the exit ( curve 5 ) , this did not occur until a velocity of 0 . 46 mm/s was reached . ( b ) Specific rates ( q8 , v ) Similar trends were observed for the specific rates ( fig . C . 1 2 ) but the maximum specific rate of suc rose utilization occurred at higher velocities of 0 . 33 and 0 . 46 mm/s for mean heights of 0 . 048 and 0 . 46 m , respectively , and were 1 0 . 8 and 6 . 3 g/gh respectively ( fig . C . 1 2 a , curves 1 and 2 ) . The maximum specific rate of ethanol produc tion occurred at 0 . 46 m for bo th mean heights of 0 . 048 and 0 . 46 m ( fig . C . 1 2 b , curves 1 and 2 ) and were 1 . 0 and 2 . 4 g/gh respectively . The results showed that volumetric and specific rates inc reased with an increase in the velocity within the tower sec tions whe re there was sucrose util i zation and ethanol production . The velocity at which the peak rates occurred was related to the effect the velocity has on the cel l concentration ( sect . c . 3 . 3 ) and fermentation time . The ..c:: ..... -llO � � H � N H � H � ril t/) � u ::::::> t/) � 0 ril � u H � E-< � � 0 ::> � � tl) t/) ..c:: ..... -llO � z 0 H E-1 u 2 0 ::::::> p.. � 0 � ::r: E-< ril � 0 ril E-< � u H � E-< � � 0 ::> 28 160 140 1 20 1 00 80 60 40 20 255 HE m • 1 0 . 09 6 0 2- - 0 . 82 A 3 - --- 1 . 57 � 4 --- 2 . 32 0 5 -·- 2 . 69 E , ETHANOL PRODUCTION V , SUPERFICIAL LIQUID VELOpTY , rmn/ s s Fig . C . ll Volumetric rates of (a) sucrose utilization .and (b) ethanol production vs superficial liquid velocity at various mean tower heights . � 10 -00 .. z 0 H � N H 8 ...:l H � J;>:l Cl) � 6 u :::::> Cl) � 0 J;>:l � 4 u H � H u J;>:l 2 p... Cl) "' Ill 0' ..c: 00 6 � -0 00 J;>:l "' � z 0 H u b 4 H :::::> � 0 H � u J;>:l p... p... Cl) ....:l 0 .. � ;> ::X:: 2 H J;>:l 256 I 3 I I I 0 2/ / 0 / / / ..cyO (b ) V , ETHANOL PRODUCTION HE m • I 0 . 096 0 2 - - 0 . 82 6 3 ---- T 4 --- o s --- 1 . 5 7 2 . 32 2 . 69 __ _ o,� .., .... o --- L).- >.� ...-' / / .... / lt.' � / .�0 V , SUPERFICIAL LIQUID VELOCITY , mm/ s s Fig . C . l2 Specific rates of (a) sucrose utilization and (b ) ethanol production vs superficial liquid velocity at various mean tower heights . 257 specific rates , for mean heights up to 0 . 46 m , increased with velocity as a result of the reduction in the cell concentration and fermentation time caused by an increase in the velocity . Further dec rease in the specific rates with further increase in the velocity , after the peak rates were reached , was due to the yeast cells being lifted to greater heights . Thus , the rates of sucrose utilization and ethanol production , at mean heights up to 0 . 46 m , increased to a maximum wi th an increase in the velocity and then decreased with further velocity inc rease . At greater mean heights , · the rates were ini t ially unaffec ted by the velocity increase until velocity reached 0 . 33 mm/s , above which the rates also increased with velocity . This occurred as a result of the effect the medium velocity had on the cell concentration and fermentation time at a particular height in the tower . c . 3 . 3 Cell concentration The effect of the superficial liquid velocity on the cell concentration was closely associated with the height in the tower (fig . C . 1 3 ) . At height of 0 . 096 m , the concentration decreased from 57 to 3 g/l DW ( 1 7 1 to 1 6 g/1 WW) as the velocity increased from 0 . 084 to 0 . 56 wn/ s . The increase in the velocity had resulted in yeast bed expansion and the cell floes were lifted to greater heights . Hence , the reduction in the cel l concentration . Between 0 . 82 and 2 . 32 m ( curves 2 , 3 and 4 ) , the cel l concentration increased to a peak as the velocity increased from 0 . 084 to 0 . 20 mm/s . Then the concentration decreased wi th further increase in the velocity to 0 . 56 mm/s . The initial increase in the concentration was a result of the yeast cel l bed expansion from the lower heights . The concentration decrease after the peak concentra tion was a result of the cell floes being lifted to greater heights . At the exit , the cell concentration was low ( 3 g/l DW ) and changed l i t t le as the velocity increased from 0 . 084 to 0 . 20 mm/s . It then increased to 1 1 g/1 DW ( 48 g/1 WW) with further velocity increase to 0 . 56 mm/s . reflec ting a gradual increase in the amount of the cel l floes being washed out of the tower . Thus , for all sec tions of the tower , the cell concentra tion decreased with an increase in the velocity from 0 . 20 mm/s . (a) 80 ..-! -1:10 .. H s 60 H f%l � 0 f%l H p::: 0 40 � � f%l u .. � 20 (b) ..-! - 1:10 250 .. H s H f%l � H 200 f%l � 0 f%l (.!) � H 1 50 p::: H z f%l u � � f%l 100 u .. � 258 HE m • 1 0 . 096 0 2 - - 0 . 82 /��� 6 3 ---- 1 . 57 .. 4 2 . 32 - - - o s 2 . 69 -· - / J \ � / _- ���� � ��- 0 \ � \. .. __o-. �0 A� , (,// ;t '4� I I ,'\:: I // '� �� '� � '� 0 \2� \ . -o- · � 0 0 . 1 0 . 2 0 . 3 0 . 4 0 . 5 V 8 , SUPERFICIAL LIQUID VELOCITY , mm/ s Fig . C . l3 Cell concentration vs superficial l iquid velocity at various tower heights . 259 Specific growth rate ( � ) s e The mean specific growth rate increased from 0 . 0044 to 0 . 67 g/gh as the velocity increased from 0 . 084 to 0 . 56 mm/s ( table C . 2 ) . Table C . 2 Mean specific growth rate at various superficial liquid velocitie s . V , mm/s 0 . 084 0 . 1 7 0 . 20 0 . 33 0 . 46 0 . 56 s �se , g/gh 0 . 0044 0 . 0072 0 . 0 1 2 0 . 059 0 . 29 0 . 67 X , g/1 a 54 . 8 65 . 7 69 . 5 49 - 4 27 . 6 1 3 . 8 The inc rease in the specific rate as the velocity increased to 0. 20 mm/s was due mainly to the change in residence time , since the mean cell concentration in the tower was relatively constant during this increase . However , at greater velocities than this , the mean cel l concentration was reduced and the exit cell concentration increased . Thus , the re was an increase in the mean specific growth rate as the velocity increased due to the reduction in the residence time with velocity . For velocities increase greater than 0 . 20 mm/s , the increase was also caused by a reduction in biomass in these sec tions . C . 4 TOWER PERFORMANCE The results showed that up to 99% sucrose utilization occurred . The residual sucrose concentration at the exit varied from to 1 0 g/1 depend ing on the velocity . Most sucrose was utili zed in the lower tower sec tions up to 0 . 82 m for all velocities up to 0 . 33 mm/s . For all velocities , sucrose was more than 95% utilized in 1 . 5 hours . More than 46 g/1 of ethanol was produced within the first 0 . 82 m at a velocity of 0 . 33 mm/s . For all velocities , ethanol concentrations greater than 49 g/1 were produced in 2 hours . The yield of ethanol was between 80 and 97% for heights between 0 . 82 and 2 . 32 m and for velocities up to 0 . 33 mm/s . The cell concentration attained inside the tower ranged from 20 to 80 g/1 DW ( depending of the velocity and height ) for velocities up to 0 . 33 mm/s . The continuous operation was free from contamination . It is evident from the resul ts that the optimum velocity for this 260 fermentation was 0 . 33 mm/s . If the total tower height were considered , 47 g/1 of ethanol was produced from 1 02 g/1 of sucrose in the molasses solution ( 98% utilization) and ethanol yield was 87% . The overall ethanol productivity was 1 0 g/lh . I t was , however , shown that the greatest fermentative activity occurred in the first 0 . 82 m of the tower . The refore , the upper sections of the tower ( except for the separator ) were not essential . At 0 . 82 m and 0 . 33 mm/s , there was 96% sucrose uti l ization . Sucrose concentration reduced to 4 g/1 in 0 . 7 hour and 47 g/ 1 of ethanol was produced . The corresponding ethanol yield was 89% based on sucrose uti l i zed and the theoretical yield . The mean cell concentration was 60 g/1 DW ( 200 g/1 WW ) . The overall hold up time ( including 2 . 7 hours ethanol productivity ( E ' ) was 67 ( E ' ) including the separator 0 in the separator) was 3 . 4 hours . The g/lh while the overall productivity was 1 4 g/lh . The hold up time in the separator could increase sucrose util ization by a small amount to 98% ( table C . 1 ) and ethanol concentration could increase to 50 g/1 ( fig . C . 7 b ) . The ethanol productivity ( E ' ) was approximately 33 times the productivity reported for batch fermentation o f molasses ( sect . 2 . 3 ) . The results obtained were comparable to those obtained for other tower fermentations using sucrose based media ( table c . 3 ) . Table C . 3 Comparison of tower fermentation of 1 00 g/1 sucrose media . ssu E y E ' E ' V 0 D T T s r ro % g/1 % g/lh g/lh mm/s h- 1 h h 1 . Cane molasses 96 50 92 67 1 4 0 . 33 1 . 4 0 . 7 3 . 4 2 . Bee t molasses 90 39 81 4 1 1 0 o . 53 1 . 1 0 . 9 3 . 8 3 . Cane juice 90 43 89 1 5 1 3 0 . 26 0 . 45 2 . 2 2 . 5 4 . Fodder beet 93 49 97 1 2 1 1 0 . 05 0 . 24 4 . 2 na ext rac t 1 . This study 2 . Coote 1 974 3. Prince & Barford 1 982 4 . Hende rson & Smith 1 982 26 1 The productivity was better than other inves tigations ( no . 2 to 4 in table C . 3 ) because these investigators did not optimise their tower heights , even though they used more flocculent yeasts . C . 5 CONCLUSIONS AND SUMMARY In tower fermentation of molasses using 1 00 g/1 sucrose in the feed , sucrose was more than 96% uti lized within the first 0 . 82 m of the tower height and at superficial liquid velocities up to 0 . 33 mm/s . Thus , the optimum operating conditions using a moderately flocculent S. ce rev is iae were at a superficial liquid velocity of 0 . 33 mm/s , tower straight sec tion height of 0 . 82 m and the resultant residence time was 0 . 7 hour . The resulting ethanol concentration was 47 g/1 from 96% sucrose util ization and the corresponding ethanol yield on the sucrose utilized was 89% . Optimization o f the tower height has resulted in a productivity comparable to o ther investigations using more flocculent yeasts in tower fermenters . A maj or limiting fac tor of the performance of this tower fermentation was the flocculating ability of the yeast used . APPENDIX D FLOCCULATION TEST OBSERVATIONS AND DATA D . l Observation of flocculating behaviour during fermentation Observat ions of the flocculating behaviour of the yeasts during fermentation . The abbreviat ions and the medium numbers used are the same as those used in section 6 . 2 . D . l . l K.marxi anus Y42 ( or KMY42 ) Flocculating behaviour of this yeast s train observed during fermentation in some media was different from the behaviour displayed in the flocculat ing medium . The following is a brief description of the flocculating behaviour of this yeast during fermentation in various media . l . P In whey permeate at pH 5 . 0 , the growth was good but flocculation was poor . Even after fermentation was completed . l F . PF 2 . P4 . 6 Whey permeate which was membrane filtered after autoclaving to remove the precipitate did not encourage flocculation of KMY42 . Flocculation was poor than in l . P . There was no difference between flocculation in this medium and in l . P . 2 F . P4 . 6F Growth was moderate , poorly flocculating as very few floes were 3 . PCa seen . In whey permeate with Cac12 the yeast showed generous growth but s t ill flocculated poorly . 3F . PCaF In whey permeate with CaC12 added after autoclaving and membrane f il tration , the growth was poor and flocculat ion was poor . 4 . PY 5 . PYM The addi tion of yeast extract powder resulted in improved floccu­ lation of KMY42 . Few floes could be s een when fermentation was completed . Yeast-malt extract powder in whey permeate improved flocculation of KMY42 . Large floes ( 0 . 2 mm diameter) could be seen when fermentation was completed . 6 . PYMCa The addition o f CaC12 to the above medium (5 . PYM) d id not result 7 . PXM in marked difference in the flocculation of KMY42 . Growth on this medium was good . When the cells were allowed to settle they tended to s t ick to the bottom of the flasks and not all the cells flocculated readily . Flocculation was good . 262 8 . PXM5 263 The growth in this medium was good . Floc formation was similar to 7 . PXM. 8.F . PXMSF Membrane filtration had removed some nutrient from this medium which encourage good flocculation . There was poor f locculation when fermentation was completed . 9 . PMS lO . PPe KMY42 grown in whey permeate with added malt extract syrup f locculated readily when fermentation was complete . The floc size was large ( 0 . 2-0 . 3mm) . The floc once formed was reasonably s table . Part o f the cell population did not flocculate readily . The addition of peptone powder to whey permeate did not result in good flocculation o f KMY42 in this medium . The growth was moderate and some cells were st icky . l l . PuAm The addition of urea and (NH4 ) 2 HP04 to whey permeate did not result in good flocculation o f KMY42 in this medium . Only poor 1 2 . P l0 f locculation was observed . KMY42 did no t flocculate in whey permeat e with added lactose ( to l OOg/1 ) even after the completion of fermentation . Very few small flo cs(O . 1 mm) formed . 1 3 . PM04 6 Generous growth and moderate flocculation , some small floes could be seen . 1 3F . PMo46F The colour of whey permeate enriched with molasses was rather dark and it was difficul t to observe flocculat ion . Few large floes ( 0 . 2mm) could be seen but most o f the cells 1 5 . PB 1 6 . PB5 remained in suspension . This was a very rich medium . KMY42 f locculated well , growth was generous and cells did not s t ick to the bottom of the flask . Sucrose , malt extract and yeast extract were added to this medium . Growth was moderate , moderately flocculating small floes could be seen . 1 6F . PB5F This was as for the above medium ( 15 . PB5 ) but was membrane filtered . KMY42 flocculated moderately and part o f cell population remained in suspension . 1 7 . PMoYMCa Growth was good , moderate flocculation few small floes . 1 7 F . PMoYmCaF The growth in this medium was generous but flocculation was difficult to observe because of the dark colour . Flocculation was similar to 1 3 . PMo46 , i . e . moderate flocculation . 1 9 . LYA Growth in this medium was very poor , the medium buffering capacity was not good and the pH dropped to below 3 . KMY42 showed poor f locculation . 20 . LYCa Behaviour of KMY42 in this medium was s imilar to that in 1 9 . LYA . 2 l . LSM 22 . Ma 264 Growth in spent mal t extract broth with added lac tose was poor but better than in 1 9 . LYA. There was no flocculation in the medium . Growth in mal tose ( lOg/1 ) was poor , the medium pH dropped to 2 . 9 and no f locculation . 23 . MaCa S imilar results as for 2 2 . Ma . 24 . MaCa4 S imilar result s to those for 22 . Ma and 23 . MaCa . 25 . G 26 . GCa 27 . YM 2 8 . YMCa 29 . YMA Growth in glucose ( 40g/l ) , was better than in mal tose but the poor buffering capacity of the medium resulted in rather acidic medium after fermentation . Flocculation was very poor . S imilar result s as for 25 . G . Growth in yeast mal t extract broth was generous and KMY42 flocculated rapidly by 24 hours . Very flocculant and cells did not s t ick to the bottom of the flask when the flask was shaken , i . e . could be loesened eas ily . The behaviour was s imilar to that in 2 7 . YM . The behaviour was s imilar t o that in 27 . YM and 28 . YMCa . 2 9F . YMAF Membrane fil tration has taken some nutrients away , growth was poor and poor flocculation . 30 . M 32 . MCa 33 . Ms 34 .Me KMY42 showed generous growth and . very good flocculation which occurred rapidly by 24 hours . Cells loosen from the bot tom of the flask easily after having settled to the bottom when flask was allowed to stand . S imilar to the behaviour in 30 .M . Growth in malt extract syrup was generous even though no o ther nutrient was added . The yeast , KMY42 , was very flocculant and did not st ick to flask bottom when the f lask was shaken af ter having been left standing . Growth of KMY42 in mal t extract po.wder alone was poor and the yeast flocculated poorly . D . l . 2 K. marxianus .Y42 (TS) (or KMY42 (TS ) ) The subcul ture of KMY42 , which was originally grown in whey permeate enriched with mal t extract syrup and then used in the tower fermenter , was found to be very flocculent in whey permeate ( l . P and 2 . P4 . 6 ) and also in whey permeate with added calcium chloride ( 3 . PCa) . Cell floes ( 0 . 1 to 0 . 3 mm in diameter) formed after 24 hours fermentation . In later tests the s train was found to be only moderately flocculent in these media . When these media were membrane fil tered , the strain showed weak flocculence in these media ( lF . PF , 2 F . P4 . 6F , and 3F . PCaF) . 265 D . 1 . 3 S. cerevisiae FT146 ( or SC146) This yeast s train was used in the commissioning of the tower fermenter using cane molasses (Appendix C) . 1 3 . PMo46 The strain SC146 was grown as mixed culture with the strain KMY42 , in this medium . Moderately flocculating flocswere formed rapidly af ter 1 2 hours fermentation . The floc s ize was approximately 0 . 2-0 . 3mm in diameter . 1 8 . Mo The s train SC 146 was very flocculent in this molasses medium . The cells flocculated well in fresh whey permeate ( 2 . P4 . 6 ) , whey permeate supplement ed with molasses ( 1 3 . PMo46 ) and the molasses medium ( 18 . Mo ) . 30 . M The stra in SC146 was very flocculent i n mal t extract bro th . The f loes were observed when the fermentation times was 1 2 hours . D . 1 . 4 S. cerevisiae CFCC39 ( or CC39) 1 3 . PMo46 This strain was also grown as mixed culture with the strain KMY42 , in this medium. The cells flocculated very rapidly in 1 2 hours . The f loc s ize varied from 0 . 1 to 0 . 5 mm in diameter . Later tests found the cells to be only moderately flocculent but this was later found to be due to the medium used to prepare the inoculum . 1 8 . Mo The strain showed good growth and very good flocculation in the medium . The cells flocculated rapidly after 1 2 hours fermentation . Af ter 48 hours , the floes were spherical about 0 . 5mm in diameter . Cells stuck to the bottom of the flask after the flask was left standing for a while . Later tests found the cells to be only moderately flocculent but this was found to be due to the inoculum medium used . 30 . M The s train showed good growth and was extremely flocculent in mal t extract broth . The behaviour was the same as described above for 1 8 . Mo . 266 D . 2 FLOCCULATION TEST DATA The flocculation test data of various yeast s trains in different media and at different pH have been tabulated . The data which are given in each column are arranged as follow;s column t i tle 1 - yeast 2 - grown in 3 - pHi 4 - pHf data - The abbreviated name of the yeas t being tes ted . The growth medium in which the yeas t was grown . - Initial pH of the growth mediuN . - Final pH o f the growth medium . 5 - test medium - The medium in which the yeas t was tes ted for flocculat ion . 6 - pHt - The pH of the flocculation test medium . 7 - scale - The degree of flocculence of the yeast measured in the test medium ( after Stewart 1 9 7 5 ) . NF 0 - Non-flocculent R 1 - Rough WF 2 - Weakly flocculent MF 3 - Moderately flocculent VF 4 - Very flocculent EF 5 - Extremely flocculent 8 vol . ml . 10 The settled volume of the yeas t cells in ml . after 1 0 min . 9 vol . ml . 60 The settled volume of the yeas t cells in ml . after 60 min . 1 0 - MBN - Modified Burn ' s number , * indicates non-standard MBN value because the cells were no t grown and tes ted in standard media . 1 2 3 4 5 6 7 8 9 1 0 yeast growth medium test medium scale vol (ml /min) MBN grown in pHi pHf tes t in pHt 1 0 60 Y42 l . P 5 . 0 5 . 1 l . P 5 . 1 R 1 0 . 4 1 . 1 na 36 . FM 4 . 9 R 1 0 . 4 1 . 0 na Y42 (TS ) l . P 5 . 0 6 . 2 l . P 6 . 2 VF 4 4 . 9 3 . 0 50* 36 . FM 5 . 0 EF 5 3 . 8 2 . 3 9 1 * Y42 1 F . PF 5 . 0 5 . 8 36 . FM 4 . 9 R 1 0 . 1 1 . 0 na Y42 (TS ) 1 F . PF 5 . 0 5 . 9 36 . FM 4 . 9 R 1 0 . 1 1 . 1 na Y42 2 . P4 . 6 4 . 6 4 . 7 2 . P4 . 6 4 . 7 R 1 9 . 6 9 . 2 na 36 . FM 4 . 6 R 1 9 . 5 9 . 3 na 1 3 . PMo4 6 4 . 7 R 1 10 . 0 9 . 6 na Y42 (TS ) 2 . P4 . 6 4 . 6 4 . 6 2 . P4 . 6 4 . 6 EF 5 3 . 9 2 . 6 88">'< 36 . FM 4 . 9 EF 5 4 . 1 2 . 8 83* 267 1 2 3 4 5 6 7 8 9 1 0 yeast growth medium test medium scale vol (ml /min) MBN grown in pHi pHf tes t in pHt 1 0 6 0 Y42 2F . P4 . 6F 4 . 6 5 . 8 2F . P4 . 6F 5 . 8 R 1 9 . 5 9 . 1 na 36 . FM 4 . 8 R 1 9 . 5 9 . 0 na Y42 (TS ) 2F . P4 . 6F 4 . 6 4 . 7 2F . P4 . 6F 4 . 7 R 1 9 . 3 8 . 9 na 36 . FM 4 . 9 R 1 9 . 4 9 . 0 na Y42 3 . PCa 5 . 3 5 . 9 3 . PCa 5 . 9 R 1 9 . 4 8 . 9 na 36 . FM 4 . 8 R 1 9 . 4 9 . 0 na 1 3 . PMo46 4 . 7 R 1 9 . 6 9 . 3 na Y42 (TS) 3 . PCa 5 . 0 5 . 5 3 . PCa 5 . 5 EF 5 2 . 9 2 . 2 na 36 . FM 4 . 9 EF 5 3 . 3 2 . 5 na Y42 3F . PCaF 5 . 0 5 . 8 3F . PCaF 5 . 8 R 1 9 . 7 9 . 4 na 36 . FM 4 . 8 R 1 9 . 6 9 . 1 na Y42 (TS) 3F . PCaF 5 . 3 4 . 5 3F . PCaF 4 . 5 WF 2 1 0 . 0 8 . 1 na 36 . FM 4 . 9 R 1 1 0 . 0 9 . 3 na Y42 4 . PY 5 . 0 na 2 . P4 . 6 4 . 7 WF 2 1 . 8 1 . 7 na 36 . FM 4 . 9 MF 3 3 . 2 2 . 6 na 1 3 . PMo46 4 . 7 NF 0 0 . 1 0 . 2 na 32 . MCa 4 . 3 MF 3 1 . 8 2 . 2 na Y42 5 . PYM 4 . 7 na 2 . P4 . 6 4 . 7 WF 2 1 . 1 1 . 4 na 36 . FM 4 . 9 MF 3 2 . 5 2 . 2 na 1 3 . PMo46 4 . 7 NF 0 0 . 1 0 . 2 na Y42 6 . PYMCa 4 . 6 na 2 . P4 . 6 4 . 7 WF 2 1 . 6 1 . 7 na 36 . FM 4 . 9 MF 3 2 . 6 2 . 2 na 1 3 . PMo46 4 . 7 NF 0 0 0 . 2 na 32 . MCa 4 . 3 MF 3 1 . 9 2 . 4 na Y42 7 . PXM 4 . 5 4 . 4 2 . P4 . 6 4 . 0 VF 4 1 . 9 1 . 3 na 4 . 9 VF 4 1 . 9 1 . 4 na 5 . 3 VF 4 1 . 9 1 . 3 na 36 . FM 4 . 9 VF 4 1 . 9 1 . 5 na 32 . MCa 4 . 9 EF 5 3 . 3 2 . 6 na 5 . 1 na 2 . P4 . 6 4 . 7 VF 4 2 . 4 1 . 7 na ( fresh) ** 1 3 . PMo46 5 . 0 WF 2 0 . 1 1 . 6 na 18 . Mo 5 . 3 VF 4 1 . 9 1 . 3 na Y42 8 . PXMS 5 . 0 5 . 3 2 . P4 . 6 4 . 2 NF 0 0 0 . 6 na 36 . FM 4 . 9 VF 4 1 . 3 1 . 2 na 2 7 . YM 5 . 2 VF 4 2 . 0 2 . 1 na Y42 8F . PXM5F 5 . 0 4 . 7 2 . P4 . 6 4 . 6 WF 2 0 1 . 9 na 36 . FM 4 . 9 WF 2 0 1 . 4 na 8F . PXM5F 4 . 7 WF 2 1 . 4 1 . 2 na 28 . YMCa 5 . 3 VF 4 2 . 9 2 . 1 na ** ( fresh) - freshly prepared medium instead o f the spent medium . 268 1 2 3 4 5 6 7 8 9 10 yeast growth medium test medium scale vol (ml /min) MBN grown in pHi pHf test in pHt 1 0 6 0 Y42 9 . PMs 5 . 0 4 . 6 2 . P4 . 6 4 . 2 MF 3 1 . 3 1 . 9 na 36 . FM 4 . 9 VF 4 1 . 7 1 . 9 na 2 7 . YM 5 . 2 VF 4 0 . 2 2 . 1 na Y42 (TS ) 9 . PMs 5 . 0 4 . 6 36 . FM 4 . 9 EF 5 3 . 8 2 . 6 1 1 8* Y42 1 0 . PPe 4 . 5 4 . 6 2 . P4 . 6 4 . 0 NF 0 0 . 1 0 . 2 na 4 . 9 WF 2 0 . 7 0 . 5 na 5 . 3 WF 2 0 . 6 0 . 5 na 36 . FM 4 . 9 WF 2 0 . 6 0 . 6 na 32 . MCa 4 . 9 VF 4 2 . 8 2 . 5 na Y42 1 1 . PUAm 5 . 0 na 36 . FM 4 . 9 WF 2 2 . 2 1 . 9 na Y42 1 2 . P 10 4 . 9 4 . 5 36 . FM 4 . 9 R 1 9 . 7 9 . 3 na Y42 (TS) 1 2 . P 10 4 . 7 4 . 3 1 2 . P 1 0 4 . 3 VF 4 3 . 7 2 . 6 4.4 * . 36 . FM 4 . 9 VF 4 3 . 8 2 . 6 49* Y42 1 3 . PMo46 5 . 0 4 . 7 1 3 . PMo46 4 . 7 R 1 9 . 8 9 . 5 na 36 . FM 4 . 7 VF 4 2 . 8 1 . 7 45* CC39 & 1 3 . PMo46 5 . 0 5 . 3 1 3 . PMo46 5 . 3 EF 5 1 . 9 1 . 4 1 74* Y42 (TS ) 36 . FM 4 . 9 EF 5 . 2 . 7 2 . 2 1 34* CC39 & 1 3 . PMo4 6 4 . 9 5 . 5 1 3 . PMo46 5 . 5 VF 4 3 . 8 2 . 4 43* Y42 36 . FM 4 . 9 VF 4 6 . 1 3 . 4 33* SC146 & 1 3 . PMo4 6 5 . 0 na 36: FM 4 . 9 EF 5 4 . 5 3 . 0 50* Y42 2 . P4 . 6 4 . 7 EF 5 4 . 5 3 . 3 na ( fresh) 1 3 . PMo46 5 . 0 EF 5 5 . 8 2 . 0 na ( fresh) 1 8 . Mo 5 . 3 EF 5 6 . 0 3 . 1 na ( fresh) SC146 & 1 3 . PMo46 5 . 0 5 . 4 1 3 . PMo46 5 . 4 VF 4 7 . 7 3 . 6 30* Y4 2 36 . FM 4 . 9 VF 4 7 . 5 3 . 9 29* Y42 1 3F . PMo46F 5 . 0 na 36 . FM 4 . 9 MF 3 2 . 5 2 . 1 na SC146 & 1 5 . PB 4 . 7 4 . 1 2 . P4 . 6 4 . 2 EF 5 6 . 0 2 . 9 na Y42 36 . FM 4 . 9 EF 5 5 . 3 2 . 7 na 27 . YM 5 . 2 EF 5 5 . 3 3 . 1 na Y42 1 6 . PB5 5 . 3 3 . 2 2 . P4 . 6 4 . 9 EF 5 7 . 7 3 . 5 na 36 . FM 4 . 9 EF 5 3 . 9 2 . 5 na 1 6 . PB5 3 . 2 EF 5 4 . 3 2 . 7 na 28 . YMCa 5 . 3 EF 5 5 . 2 3 . 0 na Y42 1 6F . PB5F 5 . 3 3 . 0 1 6F . PB5F 3 . 0 R 1 9 . 8 9 . 4 na 36 . FM 4 . 8 R 1 9 . 6 9 . 2 na 269 1 2 3 4 5 6 7 8 9 yeast growth medium test medium scale vol (ml /min) grown in pHi pHf test in pHt 1 0 60 Y42 1 7 . PMoYMCa 5 . 6 5 . 2 1 7 . PMoYMCa 5 . 2 R 1 9 . 7 9 . 4 3 6 . FM 4 . 9 WF 2 9 . 6 9 . 4 Y42 1 7 F . PMoYMCaF 4 . 9 na 2 . P4 . 6 4 . 7 R 1 0 . 1 0 . 2 3 6 . FM 4 . 9 WF 2 1 . 6 1 . 7 1 3 . PMo4 6 4 . 7 NF 0 0 0 . 1 SC 146 1 8 . Mo 5 . 0 na 3 6 . FM 4 . 9 EF 5 8 . 1 3 . 4 36 . FM 4 . 9 EF 5 5 . 1 2 . 9 2 . P4 . 6 4 . 7 EF 5 5 . 5 2 . 6 ( fresh) 1 3 . PMo4 6 5 . 0 VF 4 4 . 3 1 . 1 ( fresh) 1 8 . Mo 5 . 3 VF 4 4 . 1 0 . 8 ( fresh) CC39 1 8 . Mo 5 . 0 4 . 8 1 8 . Mo 4 . 8 EF 5 2 . 9 2 . 5 36 . FM 4 . 9 EF 5 5 . 1 3 . 0 1 8 . Mo (M) *** 5 . 0 4 . 9 1 8 . Mo 4 . 8 VF 4 4 . 1 3 . 0 36 . FM 4 . 8 VF 4 4 . 6 2 . 7 1 8 . Mo (Mo ) *** 5 . 0 5 . 6 1 8 . Mo 5 . 6 R 1 9 . 5 9 . 0 36 . FM 4 . 8 VF 4 4 . 0 2 . 5 1 8 . Mo (YM) *** 5 . 0 4 . 8 1 8 . Mo 4 . 8 VF 4 4 . 8 2 . 9 36 . FM 4 . 8 EF 5 2 . 6 1 . 8 SC146 18 . Mo (Mo) *** 4 . 9 4 . 9 1 8 . Mo 4 . 9 VF 4 5 . 9 3 . 3 36 . FM 4 . 8 MF 3 7 . 3 3 . 7 1 8 . Mo (Mo ) *** 5 . 0 4 . 9 1 8 . Mo 4 . 9 MF 3 7 . 6 3 . 8 36., FM 4 . 8 MF 3 7 . 6 4 . 1 Y42 (TS ) 1 9 . LYA 5 . 0 3 . 6 2 . P4 . 6 4 . 2 VF 4 1 . 9 1 . 3 36 . FM 4 . 9 VF 4 1 . 6 1 . 3 2 7 . YM 5 . 2 VF 4 1 . 6 1 . 3 Y42 1 9 . LYA 5 . 1 3 . 5 2 . P4 . 6 4 . 7 WF 2 0 1 . 9 36 . FM 4 . 9 R 1 0 0 . 7 28 . YMCa 5 . 3 R 1 0 0 . 5 Y42 20 . LYCa 4 . 8 na 2 . P4 . 6 4 . 7 MF 3 1 . 9 2 . 4 36 . FM • 4 . 9 MF 3 2 . 7 2 . 9 1 3 . PMo46 4 . 7 WF 2 0 . 1 2 . 2 32 . MCa 4 . 3 MF 3 2 . 3 2 . 6 Y42 2 1 . LSM 5 . 0 3 . 9 2 . P4 . 6 4 . 2 NF 0 0 0 . 1 36 . FM 4 . 9 NF 0 0 0 . 2 2 7 . YM 5 . 2 R 1 0 . 2 0 . 2 Y42 22 . Ma 4 . 8 2 . 9 2 . P4 . 6 4 . 0 NF 0 0 0 . 2 4 . 9 NF 0 0 0 . 2 5 . 3 NF 0 0 0 . 1 36 . FM 4 . 9 NF 0 0 0 . 2 32 . MCa 4 . 9 NF 0 0 0 . 2 *** The medium in the bracket indicates the inoculum growth medium . 10 MBN na na na na na 2 1 * 46* na na na 120* 46* 4 3* 4 1 * 0 44* 40* 154* 56* 30* 25* 3 1 * na na na na na na na na na na na na na na na na na na 270 1 2 3 4 5 6 7 8 9 1 0 yeast growth medium test medium scale vol (ml /min) MBN grown in pHi pHf test in pHt 1 0 60 Y42 23 . MaCa 4 . 7 2 . 9 2 . P4 . 6 4 . 0 NF 0 0 0 . 1 na 4 . 9 NF 0 0 0 . 1 na 5 . 3 NF 0 0 0 . 1 na 36 . FM 4 . 9 NF 0 0 0 . 1 na 32 . MCa 4 . 9 NF 0 0 . 1 0 . 2 na Y42 24 . MaCa4 4 . 4 2 . 9 2 . P4 . 6 4 . 0 NF 0 0 0 . 1 na 4 . 9 NF 0 0 0 . 1 na 5 . 3 NF 0 0 0 . 1 na 36 . FM 4 . 9 NF 0 0 0 . 1 na 32 . MCa 4 . 9 NF 0 0 . 1 0 . 2 ne: Y42 25 . G 4 . 6 2 . 6 2 . P4 . 6 4 . 0 MF 3 1 . 0 1 . 1 n:J 4 . 9 MF 3 0 . 9 1 . 0 na 5 . 3 MF 3 1 . 0 1 . 1 na 36 . FM 4 . 9 MF 3 1 . 0 1 . 1 na 32 . MCa 4 . 9 MF 3 1 . 2 1 . 2 na Y42 26 . GC a 5 . 1 2 . 6 2 . P4 . 6 4 . 0 WF 2 0 . 1 0 . 9 na 4 . 9 WF 2 0 . 4 0 . 9 na 5 . 3 WF 2 0 . 4 0 . 9 na 36 . FM 4 . 9 WF 2 0 . 4 0 . 9 na 32 . MCa 4 . 9 MF 3 1 . 0 1 . 0 na Y42 2 7 . YM 5 . 0 5 . 2 2 . P4 . 6 4 . 2 R 1 0 0 . 6 na 36 . FM 4 . 9 R 1 0 0 . 7 na 27 . YM 5 . 2 WF 2 0 1 . 1 na Y42 28 . YMCa 4 . 9 5 . 1 2 . P4 . 6 4 . 2 R 1 0 0 . 6 na 36 . FM 4 . 9 R 1 0 . 1 1 . 7 na 27 . YM 5 . 2 MF 3 1 . 0 1 . 6 na Y42 2 9 . YMA 5 . 1 5 . 4 2 . P4 . 6 4 . 2 EF 5 4 . 5 2 . 1 na 36 . FM 4 . 9 EF 5 4 . 3 2 . 5 na 2 7 . YM 5 . 2 MF 3 2 . 0 1 . 7 na Y42 2 9F . YMAF 5 . 0 5 . 6 2 . P4 . 6 4 . 9 WF 2 0 1 . 2 na 36 . FM 4 . 9 WF 2 0 1 . 6 na 28 . YMCa 5 . 3 WF 2 0 1 . 9 na 2 9 F . YMAF 5 . 6 WF 2 0 1 . 4 na Y42 30 . M 4 . 9 5 . 4 2 . P4 . 6 4 . 0 R 1 0 . 1 0 . 4 na 4 . 9 WF 2 0 . 6 0 . 6 na 5 . 3 WF 2 0 . 4 0 . 4 na 36 . FM 4 . 9 WF 2 0 . 6 0 . 6 na 32 .MCa 4 . 9 VF 4 2 . 6 2 . 2 na Y42 (TS) 30 . M 5 . 2 5 . 2 3 0 . M 5 . 2 MF 3 3 . 0 2 . 6 na 3 6 . FM 4 . 9 WF 2 9 . 5 3 . 5 4 CC39 30 . M 5 . 0 4 . 3 30 . M 4 . 3 EF 5 2 . 1 1 . 5 187* 36 . FM 4 . 9 EF 5 2 . 3 1 . 6 1 72 SC146 30 . M 4 . 5 na 36 . FM 4 . 9 VF 4 4 . 6 2 . 7 6 1 2 7 1 1 2 3 4 5 6 7 8 9 1 0 yeast growth medium test medium scale vol (ml/min) MBN grown in pHi pHf test in pHt 1 0 60 CC39 3 1 . M* 5 . 1 4 . 7 3 1 . M* 4 . 7 VF 4 4 . 2 2 . 9 53* 36 . FM 4 . 8 VF 4 3 . 9 1 . 7 58 SC146 3 1 . M* 5 . 2 3 . 9 3 1 . M* 3 . 9 MF 3 1 . 5 1 . 1 na 36 . FM 4 . 8 MF 3 2 . 6 1 . 7 na 3 1 . M* 4 . 9 4 . 7 3 1 . M* 4 . 7 VF 4 4 . 0 2 . 3 4 7 * 36 . FM 4 . 9 VF 4 3 . 9 2 . 3 52 3 1 . M* 5 . 1 4 . 7 3 1 . M* 4 . 7 WF 2 7 . 5 3 . 0 16* 36 . FM 4 . 8 WF 2 6 . 5 3 . 4 3 1 Y42 32 . MCa 4 . 7 na 2 . P4 . 6 4 . 7 R 1 0 . 1 0 . 3 na 36 . FM 4 . 9 VF 4 2 . 2 1 . 9 na 1 3 . PMo46 4 . 7 R 1 0 . 1 0 . 2 na 32 . MCa 4 . 3 R 1 0 . 1 0 . 4 na Y42 33 . Ms 5 . 0 4 . 8 2 . P4 . 6 4 . 2 WF 2 0 1 . 1 na 36 . FM 4 . 9 WF 2 0 . 3 1 . 3 na 2 7 . YM 5 . 2 MF 2 1 . 0 1 . 7 na Y42 34 . Me 5 . 0 3 . 8 2 . P4 . 6 4 . 2 NF 0 0 0 . 3 na 36 . FM 4 . 9 R 1 0 . 1 0 . 2 na 2 7 . YM 5 . 2 R 1 0 . 2 0 . 4 na APPENDIX E ESTIMATION OF DATA UNCERTAINTIES E . 1 SUGAR CONCENTRATIONS The lac tose concentration readings were obtained for a set of replicate dilutions ( 50x ) of a whey solution . The readings were 1 07 . 2 , 1 07 . 5 , 1 1 0 . 5 , 1 1 1 . 8 , 1 1 4 . 0 , 1 1 4 . 8 , 1 1 6 . 5 and 1 1 6 . 5 gl l . The mean (x ) and the standard deviation ( SD ) were 1 1 2 . 4 and 3 - 7 gl l , respectively . The 95% confidence interval uncertainty in each reading was B = ± t x SD Where t is a constant multiplication fac tor for various sample size ( n ) at the 95% level of confidence . SD = I [ I (x-x) 2 -;- ( n- 1 ) ] and x is an individual reading ( Cleland 1 983 ) . Thus , the unce rtainty for each lactose reading is ( t 8 ) 2 . 37 X 3 . 7 s . 77 I 1 1 2 . 4 pe rcentage uncertainty ) 8 . 77 = 7 . 8 9 8 2 . 37 for n gl l % Thus , the re was approximately 95% certainty that one lac tose value was ± 8% of the true value . The glucose membrane of the sugar analyzer was found to give more consistent readings of sucrose concentration than the lac tose membrane . Thus , the uncertainty in the sucrose measurement would be less than 8% . In the tower fermentation of whey permeate , each datum point was generally averaged from 5 readings . The uncertainty in each mean value was = 8 I I 5 = 3 . 58 % . When the lactose concentration was lower the uncertainty would be lower since less dilution was required . The high lac tose concentration of 1 00 gl l was chosen to estimate the uncertainty because the uncertainty at this concentration would be greater than at lower concentrations since more dilution was required for the higher concentration . 272 273 E . 2 ETHANOL CONCENTRATION Replicate ethanol analyses were made using solutions containing equal concentrations o f ethanol and isopropanol ( 20 gll ) . The ratios of the peaks were 0 . 9 1 30 . 0 . 96 1 6 , 0 . 9677 , 0 . 9902 , 1 . 01 62 , 1 , 0544 and 1 . 1 1 80 . The mean and the standard deviation were 1 . 0030 and 0 . 0674 , respectively . The value of t for 7 replicate readings is 2 . 45 for 95% confidence level . Thus , the uncertainty in each injec tion was 2 . 45 X 0 . 0674 o . 1 65 1 I 1 . 003o = 0 . 1 65 1 = 1 6 . 46 1 7 % Each ethanol measu rement was obtained from an average of generally two sets of peaks and multiplied by the peak ratio of the external standard . Thus , the uncertainty of the ethanol measurement was to 1 . PE I [ ( 1 6 . 5 )2 X 2 I 12 ) + ( 1 6 . 5 )2 J 25 . 6 = 26 % Assuming the mean ratio of the two peaks was approximately equal Thus , the uncertainty of the average of 5 samples = 26 I 15 = 1 1 • 6 = 1 2 The percentage uncertainty was the same for high or low ethano l concentration since no dilution was used and the standards used were designed to give a ratio as close to 1 : 1 as possible . E . 3 ETHANOL YIELD The yield was calculated based on the rat io of ethanol produced to the amount of sugar utilized . Thus , the uncertainty in the ethanol yield is p = I [ (P )2 + ( P ) 2 J % Y E s where P ( 1 2% ) is the percentage uncertainty of the average E ethanol concentration while P5 ( 4% ) is that of the average sugar utilized . The value of P varied from 4 to 5 50% of the sugar util ized s for 99 and % uti lization of 1 00 gll sugar , respectively . This was because P was calculated as follows s p = [I ( � + B2 • ) J X 1 00 I /:, s % s so S1 where B and B . are the uncertainties of the ini tial and so 51 instantaneous sugar concentrations , respectively . t:,s is the change in the sugar concentration . Thus , for these two extremes of 4 and 550% uncertainty in the 274 value of the sugar utilized ( using B5 = 4 g/1 for 1 00 g/1 sugar and PE of 1 2% for average ethanol concentration) the magnitude of the unce rtainty of the ethanol yield was between 1 2 and 490% for 99 and 1 % sugar util ization , respectively . Thus , the re was considerable uncertainty in the yield value when there was low sugar consumption at high concentration of sugar . E . 4 RATES OF SUGAR UTILIZATION E . 4 . 1 Volumetric rate ( S ' ) The rate of sugar util ization ( S ' ) was calculated from the amount of sugar util i zed within a particular tower sec tion during the time that the medium was within that tower section . The uncertainty is p s I = [I ( B; i + B; i + 1 ) J X 1 00 I 6 s % Where Bsi and Bsi+l are the uncertainties ( 4% ) of the sugar concentrations at the bot tom and the top of the section , respectively . Assuming that there was negligible uncertainty in the value of the residence time . Thus , the value of Ps i varied in proportion to the mean sugar concentrations present being approximately 6% for a very high rate of sugar utili zation ( 350 g/ lh ) and as high as 80% for a very low rate of sugar utili zation ( 0. 1 g/ lh) . E . 4 . 2 Specific rate ( q ) The specific ra te was calculated by dividing the volumetric rate ( S ' ) in each tower section by the mean cell concentra tion ( Xa ) in that particular section . Thus , its uncertainty incorporated the uncertainty of the cell concentration . The percentage uncertainty of the specific rate ( q ) is 1 1 p = ( [ ( B2 + B2 )� X 1 00 .;- X y + P2 I l � % q DWi DWi+l a s Where B and B are the uncertainties of the cell DWi DWi+l concentration ( DW ) ( 4% ) at the bo ttom and the top of the tower section , respectively , and X a is the mean cell concentration in this tower sec tion . P 1 was shown previously to be between 6 and 80% . s High cell concentration ( 200 g/1 DW ) , generally occurred at the bo ttom of the tower where the volumetric rate ( S ' ) was high . The uncertainty of the volumetric rate ( P5 1 ) was 6% . Thus , the unce rtainty of the specific rate (Pq ) at this cell concentration was approximately 275 7% . However , when the cell concentration was low ( between 0 . 2 and 1 0 g/1 DW ) , the volumetric rate ( S ' ) was low because this occurred in the upper sections of the tower . The uncertainty associated with this low volumetric rate ( P8 y was 80% . Thus , the uncertainty of the specific rate ( Pq ) was approximately 80% . This showed that the uncertainties of the two rates of sugar util ization were of similar magnitude ( be tween 7 and 80% ) . E . 5 RATE OF ETHANOL PRODUCTION E . 5 . 1 Volumetric rate ( E ' ) The rate of ethanol production was calculated from the amount of ethanol produced within a particular tower sec tion during the time that the medium was wi thin that tower section . Thus , the unce rtainty is PE , = [I ( B�i+l + B�i ) ] X 1 00 I 6E % Where �i and �i+l are the uncertainties ( 1 2% ) due to the mean ethanol concentration at the bot tom and the top of the tower section , respec tively . Thus , the value of PE ' varied in proportion to the ethanol concentration present assuming that there was negligible uncertainty in the value of the residence time . The uncertainty was approximately 1 2% of the value of the volumetric rate calculated in the tower sec tions where there was ethanol production . I t is clear that in the upper sections of the towe r , where ethanol concentration was high , but ethanol production was negligible , the resul ting confidence limits will be very wide . E . 5 . 2 Specific rate (v ) The speci fic rate was calculated by dividing the volumetric rate ( E ' ) in each tower sec tion by the mean cell concentration ( Xa ) in that particular tower section . Thus , i ts unce rtainty incorpo rated the unce rtainty of the cell concentration . The percentage uncertainty of the specific rate ( v ) is p = ( [ ( B2 + B2 )� X 1 00 7 X ] 2 + P2E, l � V DWi DWi+l a B and B 1 are both 4% . DWi DWi+ For high cell concentration ( 200 g/1 DW ) , the volumetric rate ( E ' ) was high at the bot tom of the tower . The uncertainty i n the volumetric rate (E ' ) was shown to be 1 2% . Thus , using the above equation , the 276 uncertainty of the specific rate (v ) at this cell concentration was approximately 1 3% . When the cell concentration was low ( between 0 . 2 and 1 0 g/1 DW ) , the volumetric rate ( E ' ) was low because this occurred in the upper section of the tower . The confidence limits will be wide as discussed above ( sec t . E . 5 . 1 ) . This showed that the uncertainties of the two rate of sugar uti lization were of similar magnitude . E . 6 CELL CONCENTRATION E . 6 . 1 Haemacytometer cel l count A set of replicate readings of a yeast cell suspension using the Haemacytometer was determined . The cell number readings obtained were ( 1 . 6 , 1 . 6 , 1 . 6 , 1 . 7 , 1 . 9 , 1 . 9 , 2 . 0 , 2 . 0 , 2 . 1 , 2 . 3 ) x 1 08 cells/ml . The 8 7 mean and the standard deviation were 1 . 87x1 0 and 2 . 4 1 x 1 0 cells/ml , respectively . Thus , the 95% confidence inte rval uncertainty in each Haemacytometer cell count was 7 Ex 2 . 26 x 2 . 4 1 x 1 0 PX 5 . 4x1 0 7 / 1 . 87x1 08 ( t = 2 . 26 for n = 1 0 ) E . 6 . 2 Plate count 7 5 . 4x1 0 29 cel ls/ml % Ten replicate read ings of the cell plate count number of a cell suspension of K. marxianus Y42 (24 hours culture ) were determined . The readings were ( 1 . 0 , 2 . 5 ) x 1 09 ce lls/ml . The 1 . 1 , 1 . 1 , 1 . 2 , 1 . 4 , 1 . 4 , 1 . 5 , 1 . 5 , 1 . 7 , 9 mean and the standard deviation were 1 . 44x1 0 and 4 . 33x 1 08 cells/ml , respectively . Thus , the 95% confidence interval uncertainty in each cel l plate count number was Ex 2 . 26 x 4 . 33x 1 08 9 . 78x1 08 cells/ml PX 9 . 78x1 08 I 1 . 44x1 09 68 % E . 6 . 3 Cell dried weight and centrifuged wet weight Ten replicate readings of the cell concentration of a cell suspension of K. marx ianus Y42 obtained during the tower fermentation of whey permeate were determined . The readings were 274 . 0 , 278 . 7 , 281 . 0 , 283 . 2 , 283 . 5 , 283 . 5 , 284 . 1 , 284 . 6 , 285 . 8 , 292 . 0 g/ 1 WW for the centrifuged wet weight . The mean and the standard deviation were 283 . 1 and 4 . 7 g/ 1 WW , respec tively . 2 . 26 X 4 . 7 1 o . 62 I 283 . 1 277 1 0 . 62 4 gl l ww % The corresponding cell dried weight readings were 69 . 0 , 69 . 3 , 69 . 9 , 70 . 1 , 7 1 . 3 , 7 1 . 4 , 7 1 . 5 , 7 1 . 5 , 7 1 . 7 , 73 . 0 gl l DW . The mean and the standard deviation were 70 . 9 and 1 . 2 gl l DW , respectively . Bow = 2 • 2 6 x 1 • 2 PDW = 2 . 7 1 I 70 . 9 2 . 7 1 = 4 gll DW % This showed that the uncertainty in the cell weight determination was small . However , at sample point ( 0 . 096 m ) , there was considerable carbon dioxide production for all tower fermentation runs . This introduced more uncertatinty into the measurement than usual . The sampling difficulties encountered during the tower fermentation of whey permeate enriched with molasses were already discussed ( sec t . 5 . 1 . 7 ) . For the cell concentration measurements obtained under these conditions , the uncertainty should be higher at approximately 5% . E . 6 . 4 Es t imation of the cell dried weight of K. marxianus Y42 from cell plate count number The linear regression line used is from fig . B . 1 in · Appendix B . 4 . log KM = 6 . 96 + 1 . 08 logDW The standard deviation (SD ) o f the intercept is 0 . 5864 . Bintercept = 2 . 45 x 0 . 5864 t = 2 . 45 for n = 7 The SD o f the s lope is 0 . 4 1 06 , Bslope 2 . 45 X 0 . 41 06 Now logDW = ( logXKM - 6 . 96 ) I 1 . 08 The uncertainty in X is 68% KM = 1 . 436 = 2 1 % = 1 . 003 93 % ( sect . E . 6 . 2 ) and for X KM 1 x 1 06 and 2x 1 09 cellslml , the log of the . uncertainty in the value of the cell dried weight is Bnw [ l (4 . 0& + 1 . 4362 )� x 1 00 I ( 6 - 6 . 96 ) F + 932 ] � 460 % for XKM = 1 0 6 c ellslml 1 l Bnw = [ ! ( 6 . 1 22 + 1 . 4362 )� x 1 00 I ( 9 . 03 - 6 . 96 ) l 2 + 932 ]� 280 % for XKM = 2x1 09 cel lslml be tween es timated Thus , the unce rtainty in the estimation of the cell dried weight of K. marxianus Y42 from fig . B . 1 is between 280 and 460% for cell dried weight between 0 . 1 3 and 1 47 gll DW .