T H E U N I V E R S I T Y O F M I.C H I G A N COLLEGE OF ENGINEERING Department of Chemical and MetalLurgical Engineering Progress Report OXYGEN TRANSFER MECHANISMS IN THE GLUCONIC ACID FERMENTATION BY Pseudomonas ovalis Gary F. Bennett ORA Project 05766 under contract with% NATIONAL SCIENCE FOUNDATION GRANT NO. GP-1007 WASHINGTON, D.C. administered through: OFFICE OF RESEARCH ADMINISTRATION August 1963 ANN ARBOR

This report was also a dissertation submitted in partial fulfillment of the requirements for the dedegree of Doctor of Philosophy in The University of Michigan, 1963 o

TABLE OF CONTENTS Page LIST OF TABLES iv LIST OF FIGURES vii LIST OF APPENDICES x ABSTRACT xi Io INTRODUCTION 1 General Considerations 1 Effect of Agitation 10;Effect of Viscosity 12 Statement of the Problem 18 II. MATERIALS AND METHODS 21 The Fermentor 21 The Oxygen Electrode 26 The Rotational Viscometer 31 Bacterial Species 32 Media 32 Stock Cultures 33 General Experimental Conditions 34 Physical and Chemical Measurements 34 Procedure for a Run 38 Bacterial Densities 39 Alteration of Viscosity 40 Sulfite Oxidation 42 Measurement of Oxygen Transfer Rates Using the Oxygen Electrode 43 III, EXPERIMENTAL RESULTS 45 A. Preliminary Experiments 45 Physical and Chemical Measurements 45 Growth Curve 45 Thermal Death Rate 48 Products of the Fermentation of Glucose by Pseudomonas ovalis 56 Turbidity and the Estimation of Bacterial Population 59 ii

TABLE OF CONTENTS (Concluded) Page B. Effect of pH, Temperature, and Concentration of Cells on the Rate of Production of Gluconic Acid by Pseudomonas ovalis 67 Effect of pH 67.Effect of Temperature 71 Effect of Cell Concentration 73 C. Direct Use of the-Oxygen Electrode 78 Measurement of the Critical Oxygen Concentration 78 Measurement of the Rate of Oxygen Utilization by Pseudomonas ovalis 84 Effect of Oxygen Concentration on the Rate of Gluconic Acid Production 86 D. Effect of Viscosity 93,Sulfite System 93 Production of Gluconic Acid in Non-Newtonian Broths 102 Effect of Viscosity on the Rate of Oxygen Transfer in a Cell-Free System 111 IV. DISCUSSION 120 Mechanism of Cell Adsorption on Bubbles 120 Effect of Viscosity on the Oxygen Uptake Rate 128 V. SUMMARY 133 APPENDICES 135 BIBLIOGRAPHY 151 iii

LIST OF TABLES TABLE Page I, Critical Dissolved Oxygen Concentrations for Microorganisms in the Presence of Substrate. 8 II, Chemical Analysis of Pseudomonas ovalis..46 III. Growth of Pseudomonas ovalis NRRL B-8S in a 5% Glucose Medium at3 50.. *..... 50 IV. Effect of Time and Temperature on the Number of Viable Cells of a 24 Hour Culture of Pseudomonas ovalis in Gelatin-saline Broth. 53 V, Effect of Temperature on the Thermal Death Rate Constant for a 24 Hour Culture of Pseudomons ovalis in Gelatin-saline Broth. 54 VI. Temperature Coefficients for Thermal Death Rates of a 24 Hour Culture of Pseudomonas ovalis in Gelatin-saline Broth....... 55 VII. Relationship Between Turbidity and the Logarithm of the Multiple of Dilution of a Suspension of Pseudomonas ovalis with a Glucose-phosphate Medium......... 61 VIII. Relationship Between Turbidity and Dry Cell Weights for Pseudomonas oYall is..... 66 IX. Effect of Temperature on the Rate of Gluconic Acid Production by Pseudomonas ovalis in a Nitrogen-free Medium at pH 7.0, with an Agitator Speed of 300 RPM, an Air Sparging Rate of 1.16 WM and a Klett Reading of 100 74 X. Q10 Values for the Rate of Production of Gluconio Acid by Resting Cells of Pseudomonas ovalis at pH 7.0 with an Agitator Speed of 300 RPM and an Air Sparging Rate of 1.16 VVM 75 XI. Effect of Variations in the Agitation Speed on the Rate of Gluconic Acid Production by Resting Cells of Pseudomonas ovalis at pH 7.0 and 25~0. o 0 o...... o *... 87 XII. Effect of Dissolved Oxygen Concentration on the Rate of Production of Gluconic Acid by Resting Cells of Pseudomonas ovalis at pH 7.0 and 250.... 91 iv

TABLE Page XIII. Effect of Agitation Speed on the Rate of Oxygen Utilization and on the Dissolved Oxygen Concentration: Resting Cells of Pseudomonas ovalis in a Glucose Medium at 25~0, pH 7.0 and Air Flow Rate 0.46 VVM in Run D-15................. 94 XIV. Effect of Air Flow Rate and Agitator Speed on the Dissolved Oxygen Concentration in a Glucose Medium Containing Resting Cells of Pseudomonas ovalis at 250C, pH 7.0 and Klett Reading 100............... 96 XV. Effect of Viscosity on the Rate of Oxygen Transfer to Sparged, Agitated Solutions of Sodium Sulfite: Air Flow Rate 1.0 VVM, Agitation Speed 300 RPM, Temperature 250C. 99 XVI. Sulfite Oxidation Rate in Sucrose Solutions Corrected for Reduced Solubility of Oxygen in Equilibrium with Air........ 103 XVII. Values of the Exponent n in the Power-Law Equation for Several Concentrations of Natrosol and Methocel......... 107 XVIII. Production of Gluconic Acid by Pseudomonas ovalis in Solutions Having Pseudoplastic Viscosity: Agitation Rate 570 RPM, Air Rate 1.85 WM, Temperature 25~0 110 XIX. Effect of Agitation Speed on the Mass Transfer Coefficient for Oxygen Diffusing through the Surface of a Glucose Solution. Measurements Made with the Oxygen Electrode. 115 XX. Effect of Increasing Concentrations of Methocel on the Oxygen Transfer Rate Measured by the Oxygen Probe: Agitation Rate, 300 RPM; Air Flow Rate to Sparger 0.2 WM, Temperature 25~0........... 116 XXI. Effect of the Air Flow Rate on the Oxygen Transfer Rate Measured by the Oxygen Electrode in a Fermentor Agitated at 300 RPM. The Methocel Concentration in the Liquid was 2.5 g/l................... 118 XXII. Calibration Data for Fischer and Porter Flowmeter No. 2F 1/4-20-5, Sapphire Float; Calibrated Using a Wet Test Meter at 78~F, and 730.3 mm Pressure.......... 136 v

TABLE Page XXIII. Change of Dissolved Oxygen Concentration in a Suspension of Resting Cells of Pseudomonas ovalis agitated at 114 RPM after the Air Plow to the Sparger was Stopped o o o o o. o.o o...... 144

LIST OF FIGURES FIGURE Page 1, Path of oxygen transfer in submerged aerobic fermentations o.0 0 o o 0 0 o 0 o o..* 2 20 Path of oxygen transfer with microorganisms adsorbed on the surface of a bubble.... 4 3. Flow curves for various types of timeindependent9 non-Newtonian fluids o o o o.. 13 4o Shear flow diagram for a pseudoplastic fluid showing the flow behavior index,9, and the consistency index, k o o o o o.... 15 5. Equipment used in the study of the rates of oxygen transfer o o o o. o..... 24 6. Schematic drawing of the oxygen electrode.. 27 7. Electron micrograph of Pseudomonas ovalis, NRRL B-8S, grown for 24 hours on an agar surface (X 109000) 0 0 0 0 0 0 0. 47 8o Growth of Pseudomonas ovalis9 NRRL B-8S in a 5% glucose medium at 30~ 0 o o o....... 49 9o Effect of temperature on survival of Pseudomonas ovalis in a gelatin-saline medium 0 0 52 10o Arrhenius plot for the effect of temperature on the survival of Pseudomonas ovals in a gelatin-saline medium 0000000000 0 57 11o Graphical determination of the pKa of the acid formed during the fermentation of glucose by Pseudomonas ovalis, NRRL B-8S in a resting cell suspension.0 0 0 0 0 0 60 12o Relationship between turbidity and cell concentration for a suspension of Pseudomonas ovalis in glucosesphosphate medium.o 0 o 62 13o Relationship between transmittance (D) or its metameters and the logarithm of the multiple of dilution of a suspension of Pseudomonas ovalis with glucose-phosphate medium. 65 140 Effect of pH on the rate of production of gluconic acid by Pseudomonas ovalis in a growth medium at 25~0 and Klett reading 300. 70 vii

FIGURE Page 15, Effect of pH on the rate of production of gluconic acid by Pseudomonas ovalis in a nitrogen-free medium at 25~0 and Klett reading 100................. 72 16. Effect of temperature on the rate of gluconic acid production by Pseudomonas ovalis in a nitrogen-free medium........ 76 17. Arrhenius plot for the effect of temperature on the rate of gluconic acid production by Pseudomonas ovalis in a nitrogen-free medium. 77 18. Effect of cell concentration on the rate of gluconic acid production and on the equilibrium oxygen concentration of the liquid during the fermentation of glucose to gluconic acid by resting cells of Pseudomonas ovalis..... 79 19. Measurement of the rate of oxygen utilization by resting cells of Pseudomonas ovalis using the oxygen probe with agitation at 114 RPM and no aeration............... 81 20. Effect of dissolved oxygen concentration on the rate of oxygen utilization by resting cells of Pseudomonas ovalis, with no aeration at pH 7.0 and 25u0. *. * * * * * * * * * * 83 21. Effect of oxygen concentration on the rate of gluconic acid production (speed and air flow rate varied) by resting cells of Pseudomonas ovalis at pH 7.0 and 2500.......... 89 22. Effect of oxygen concentration on the rate of gluconic acid production by resting cells of Pseudomonas ovalis at 2500, a constant air rate of 1.00 VVM and a pH of 7.0..... 92 23. Effect of agitation speed on the rate of oxygen utilization by resting cells of Pseudomonas ovalis in a glucose medium at 25~C and pH 7.0. Rates determined with aeration (gluconic acid) and without aeration (electrode). 95 24. Effect of stirring speed and air flow rate on the dissolved oxygen concentration in a suspension-of resting cells of Pseudomonas ovalis at 25~0, pH 7.0 and Klett reading 100.... 97 viii

FIGURE Page 25. Effect of increasing viscosity due to the addition of glycerin or sucrose on the rate of sodium sulfite oxidation. The liquid was sparged with air at a rate of lo0 VVM and agitated at 300 RPM at 25~0 o 0o o o o o o o 100 266 Logarithmic plot showing the effect of increasing viscosity due to the addition of glycerin or sucrose on the rate of sodium sulfite oxidationo The liquid was sparged with air at a rate of loO VVM and agitated at 300 RPM at 25 o.o. 101 27. Flow curve for a medium containing 15 grams of Natrosol per liter, The plot of shear rate versus shear stress was determined with a rotational viscometer 0 o o o o o o. 105 28. Relationship between shear stress and shear rate for Natrosol and Methocel...... 106 29, Relationship between the concentration of Natrosol and Methocel added to water to increase the viscosity and _n the slope of the logarithmic plot of the flow curve for that medium o o o o a. o o o o o o o.. 109 300 Oxygen transfer by diffusion through the surface of a glucose-phosphate medium stirred at 300 RPM at 2500 o. o o o o. o o o. 0. 113 31o Effect of air flow rate on the oxygen transfer rate measured by the oxygen electrode in a fermentor agitated at 300 RPMo The Methocel concentration in the liquid was 2.5 g/1 o.. 119 32. Relationship between the response of the oxygen electrode and the dissolved oxygen concentratio n distilled water measured at 2500 by the Winkler method o o. o. o. 138 33, Effect of temperature on the oxygen electrode response. The electrode was placed in distilled watr raer uted with oxygen from the air o o o o 0 0 0 0 o 0 0 0 0 0 0 0.. 140 ix

LIST OF APPENDICES APPENDIX A. Calibration of Rotameter.... Bo Calibration of Oxygen Electrode.... C, Effect of Temperature on the Oxygen Electrode o o...... *. Do Calculation of the Correction for the Rate of Production of Undissooiated Gluconio Acid E. Comparison of Oxygen Transfer Rate Measured by the Rate of Production of Gluconio Acid to the Rate Measured with the Oxygen Electrode. 0 0 0 0 0. 0 0. * * * F, Calculation of the Parameters in the PowerLaw Equation for non-Newtonian Viscosity G. Calculation of Reynolds Number for Agitation Ho Nomenclature. 0 0 0 o. 0 * * * *.. Page 135 137 139 142 145 146 148 x

ABSTRACT Submerged aerobic fermentations are important in the industrial production of antibiotics, yeast cells and vitamins as well as in waste treatment. In these fermentations, oxygen is an essential nutrient that must be continually supplied to the microorganisms, Many workers have evaluated the oxygenating capacity of fermentors.under conditions of aeration and agitation. Others have studied oxygen uptake by cells in the quiescent chambers of polarographs.Little data, however, have been published on the separate effects of agitation and aeration in an active fermentation, because it is difficult to design experiments to separate these two variables, The fermentation of glucose to gluconic acid by cells of Pseudomonas ovalis suspended in a nitrogen-free medium was first used to measure the rate of oxygen transfer in a sparged and agitated system, Then, using the same system, measurements were made of the rate of transfer of dissolved oxygen to cells suspended in a liquid that was being stirred but not aerated, The effect of aeration on.the rate of oxygen transfer was the difference between the rates found by the two methods, It was found that the oxygen transfer rate measured by the electrode was less than the rate measured by the rate of gluconic acid production in a sparged system. Agitation rate had no effect on the oxygen transfer rate in the absence of aeration. However, in the sparged and aerated system, an increase in either the air rate or the agitation rate increased the rate of oxygen uptake by the xi

cells, even though the dissolved oxygen concentration in the medium was high, These data suggested that an additional mechanism for oxygen transfer was present in addition to the direct supply from dissolved oxygen in the liquid. This second pathway was encountered when cells were adsorbed on the surface of bubbles. Adsorption merged the liquid films which surrounded the bubbles and the cells, resulting in a shorter path for the oxygen transfer, The effect of viscosity has been investigated. It was found that longchain carboxymethyl cellulose derivatives, which increase the viscosity when added to water in small amounts, did not affect the rate of oxygen transfer in the Pseudomonas ovalis system, However, these same compounds, as well as glycerol and sucrose, did reduce the oxygen transfer rates when these rates were measured by a chemical technique or with the oxygen electrodeo The effect of environmental conditions on the rate of gluconic acid production by resting cells of Pseudomonas ovalis was investigated. The maximum rate was found to occur at pH 7.35 and 37 C. An activation energy of 9,600 cal/mole was calculated for this fermentationO The critical dissolved oxygen concentration for resting cells of Pseudomonas ovalis was 1o1 mg/l, xii

I INTRODUCTION 1o General.onsiderations Submerged aerobic fermentations are used industrially to produce antibiotics chemicals9 yeast cells and vitamins, and are also used in waste treatment These ferme ntations require that large amounts of oxygen be supplied to microorganisms submerged in liquids in order to support growth of the cultures o The microorganisms must be continually supplied with all essential nutrients at rates sufficiently high to prevent the limitation of their metabolic activity. At the same time, metabolic products must be swept away to prevent inhibition of growth processes of the cells, In aerobic fermentations9 oxygen becomes an essential nutriento Not only does it have a long, difficult path to travel from the air to reaction sites in the cell, but it also has a very low solubility in the medi um, Bartholomew ej a o (3) have separated the oxygen transfer process into a number of steps each wih i its own resistance. The ese reistances shown diagramatioally in Figure 1, are as foll ws (1) Gas-film resistance9s between the gas and the gasliquid interface (2) Interfacial resistance9 between the gas and liquid films (3) Liquid-film resistance between the gas!liquidL' interface and the bulk of the liquid 1

Film Liquid I// 1 Cell Liquid Film Liquid Film Film Figure 1. Path of oxygen transfer in submerged aerobic fermentations.

3 (4) Bulk liquid resistance (5) Liquid film resistance, around the cell (6) Internal resistance, associated with reactions of oxygen and respiratory enzymes of the cell In addition to the path represented by the preceding resistances, Bartholomew and his co-workers presented evidence suggesting a shorter, but parallel path between the same concentration limits, by direct contact between the cells and the gas-liquid interface. This path, shown diagramatically in Figure 2, eliminates the bulk and liquid film resistances, since the liquid film surrounding the bubble is the same liquid film as that surrounding the cell. This is an important concept that appears to have received little experimental attention. The total amount of oxygen transferred by means of this alternate path depends on the number of bubbles contacting with the gas-liquid interface. The amount of contact will depend on the area of gas available to the cells. The area of the air surface available per unit time for oxygen transfer, in turn, is a function of the bubble size, the length of the path of bubbles in the liquid, and the relative velocity between the bubbles and cells. This path for direct transfer is favored by a large number of small bubbles and by a long retention time. Since Bartholomew et al. (3) lacked quantitative data, they were only able to evaluate this new path on a qualitative basis. They observed that Stregtomyces griseus and

4 Liquid Film -, / ----Gas Film (f/Gas \ I ~-Cl ~ I ubble' I Figure 2. Path of oxygen transfer with microorganisms adsorbed on the surface of a bubble.

5 Penlcillium chrysogenum adhered to and had mobility on the surface of bubbles. It is known (73) that microorganisms can form an intimate adsorptive contact with air bubbles, much as that occurring with froth flotation of minerals. In fact, froth flotation has been used (36) to concentrate microorganisms. The relative importance of each resistance is different in any given transfer system. In the first step, the oxygen encounters the resistances of the gas film, the interface and the liquid film. Between the liquid films surrounding the bubble and the cell, the oxygen must overcome diffusion resistance of the bulk liquid. At the cell, it meets the resistance of the liquid film surrounding the cell and finally, it encounters the reaction resistance due to the respiratory enzymes of the cell. For the transfer of oxygen, one may write: = KA (0 - 0) (1) dt L g In this equation, the first three resistances have been combined into an overall mass transfer coefficient, KLA, where A is the interfacial area. The driving force is the difference between the concentration of oxygen in the liquid in equilibrium with the existing partial pressure of oxygen in the air and the instantaneous dissolved oxygen concentration in the liquid. The other resitanoes apply, according to Richards (66), not only to oxygen transfer, but also to the transfer of

6 other substances such as nutrients0 The resistance to passage of such substances through the bulk liquid should be minor if sufficient mixing is provided to maintain homogeneity,.There is a liquid film around the cells which could be affected by agitationo In the production of glueonic acid by suspended cells of Pseudoonas ovalis' Tsao and Kempe (94) proposed that the rate-limiting step occurred at this interfaceo Finn (32) howev how r has eeil allate theortially allated the magnitude of this resistance and found it too small to be significanto He stated that agitatio in excess of that required to suspend single cells uniformly will not markedly improve oxygen transfer from the liquid to the cell wall, because the cells move with the liquids improvement can only be achieved by increasing relative movement between the liquid and the ello Even t gh c s th though d mov e with th liquid there is still an increasing resistance to their movement as the rotational speed increases0 Barker and Treybal (2) have shown that the mass transfer coefficient for solids9 suspended in agitated liquids, could be correlated with a Reynolds number calculated for each tank:b The internal resistance of the microorganism to oxygen uptake is independent of stirring unless it is insufficient: to remove metabolic prodtaets o The activity of these respiratory enzymes depends upon the method used- for growing the cells among other factors, Working with Peni cillium chrvsnm Rolinson (67) found

7" that the oxygen demand rate was considerably higher in stirred fermentors than in shaken flaskso He suggested that res piration was less active in systems where aeration was less intensiveo He believed that aeration intensity could affect the kind and quantity of enzymes present in the cells. He reported that mycelium, grown under the conditions of intensive aeration found in a stirred fermentor, maintained its high oxygen utilization rate when transferred to the less active system found in a shaken flask, Terui et al, (92) found a distinct difference in physiological properties between mycelia grown as surface and myoelia grown in submerged cultureso Expecially notable was the behavior of these two different cultures toward oxygen provided as gaseous or dissolved oxygen, They attributed this difference, in part9 to differences in enzyme systems. A2pergillus or zae grown in submerged cultures contained much more eytochrome oxidase and cytochrome oa but less catalase and flavins (FMN and FAD) than cells grown on a surfaceo Winzler (100) has shown that the rate of oxygen consumption in a submerged culture is independent of the actual dissolved oxygen concentration9 provided that the oxygen level exceeds a oncentration known as the critical oncenetratifn9 o Finn (32) stated that critical conoenttrations are very low for unicellular organisms. Several of these critical concentrations have been reported in the literature and are listed in Table Io

0 TABLE I CRITICAL DISSOLVED OXYGEN CONCENTRATIONS FOR MICROORGANISMS IN THE PRESENCE OF SUBSTRATE Organism Luminous bacteria Azotobacter Vinelandii Yeast Yeast Penicillium chrs ogenum Escherichia coli Saccharomyces cerevisiae Pseudomonas spo Pseudomonas ovalis Critical Dissolved Oxygen Concentrations, m./liter.. 0.31 0.56 - 1.53.O10 O011 0.25 0066 0.26 0.20 0.2 Reference (32) (32) (32) (67) (32) (32) (86) (89) 0.7 + 0.2 c;P ( 1)

9 Many efforts have been made to measure aeration efficienCyo Most often, the rate of oxygen transfer has been measured by the sulphite method (149 29 349 989 101) developed by cooper t ^ A (19) A second technique employed with larger vessels, is an unsteady state process0 The oxygen concentration in the liquid is rdue d to zero by adding sodium sulphite, by boiling, or by stripping with nitrogeno The rate of reaeration is then measured polarographically (49) or chemically (42, 102)o Recently9 instruments have become available for measuring the oxygen concentration in gaseous streams (75, 84). These instruments utilize the paramagnetic quality of oxygen. This method is fairly accurate (61)9 since the measurement of oxygen uptake is based upon the difference in oxygen coneen= trations between the inlet and outlet streamso Shu (74) used a recording manometric technique, He described an especially designed fermentor attached to a shaker in which the uptake of oxygen during a fermentation was recorded automaticallyo This was possible since a completely closed system was employedo Recently9 oonsiderable interest has been shown in measuring the dissolved oxygen contents of the fermentation mediumo Nauy instruments are now available to measure dissolved oxygen directly. these include platinum electrodes (30, 59, 829 85)., diffusion tubing (59), dropping mercury electrodes (40) 55) and polarographs (92)o The platinum electrode is small, so it can be con

10 veniently placed in a, laboratory-size fermentoro Phillips and Johnson (59) stated that it measured dissolved oxygen concentrations; other authors (319 789 85) agree that oxygen activity or the fraction of oxygen saturation of the solution is meas$ed o Progress in instrumentation has now made it possible to record both oxygen uptake rates and oxygen concentrations in actual fermentations0 Previously iit was only possible to calaulate one from the others but workers were not able to experimentally determine both values in actual fermentation syst-ems 2o 3fe.Ot of aitatio The effects of agitation and aeration have not been studied separately in fermentation systems because it is difficult to design suitable experiments to separate them (99) o Phillips and Johnson (60) reported an effe-t of agitation on mold fermentations9 but concluded that the effect was only that of mixing or:distributing oxygen to various locations in the fermentor0 In mold fermentations it was found that intraclump resistances could be reduced by adequate agitation (99)o Dion. ale. (24) studied the effect of mechanical agitation on the morphology of Pel l m hY n in stirred fermentorso When the agitation was mild, the hyphae were long with a little branohiing; shortening of the hyphae and more branching.ocurred with vigorous agitationo Vigorous agitation als increased the penicillin yieldo

11 Donovick (27) has presented data shewing increased penicillin production with increasig power input to the agitator,.Oamposano. t alo (18) found that mechanical agitation and aeration9 sufficient to produce saturated oxygen conditiions, reduced the amount of Kojic acid produced from glucose by,As eril_ s flavus in submerged fermentations. Physical damage can be done to organisms by mixing them in baffled shaker flaskso Such results have been reported by Vondraekova (96) and Smith (77)o As stated in the, discussion of controlling resistances, Tsao and Kempe (94) presented data suggesting a significant effect of agitation on mass transfer at the cell-liquid interface9 while PFin (31) disputed this viewo Phillips and Joihnson (60) werked with a variety of organismsl measuring both the oxygen transfer rates and the conecentration of dissolved oxygen in the mediao They concluded that variations in dissolved oxygen oncentrations were not important as long as these concentrations remained above the criticalo A different view has been expressed by Steel and Maxon (83) who worked with the novobiocin fermentation Here the rate of oxygen traslfer was increased when the oncentration of disselved oxygen was increased by more vigerous agitation9 even though the dissolved oxygen eoncentration was well above the liquid —ell resistance-Becomes the controlling factor for oxygen transfer in viscous9 en.oNewtonian fermentations like the novobi@ci fermentation, They did net indicate that

12 this applied to low viscosity broths, asserting that the gasliquid film still controlled in this instance. 30 Effect of Visco it In antibLitic fermentations and also in bacterial fermentations which develop products of high molecular weights such as dextran and polyglutamyl peptide, the rheological properties of the medium change as the fermentation progresseso Very little has been published ooneerning the aeration and agitation characteristics iof nanNewtonian brothso Most published data are concerned with mold mycelium and its effects on the oxygen transfer rateo For Newtonian fluids9 the shear stress is directly proportienal to the rate of shear Y7 the viscosity p is independent of the shear rateo T$; L n &(2) Any fluid not having this simple relationship is called non= Newtoniano A pseudoplastic fluid is one whose apparent viscosity decreases instantaneously with increasing shear (Figure 3) o The f llowing expression has been proposed for this case by Ostwald ( 56) and has since been fully described by Reiner (64)b r = nV (3) The apparent viscosity at any point is = r/,y ' (4)

13 |ngham co ^ / 1 oNe onian 0 co 2 Pseudopla tio / /a t atant Shear Rate Figure 3. Flow curves for various types of timeindependent, non-Newtonian fluids.

14 0~, = k y"' (5).Thus, a plot of 1gY versus log y would give n as a slope and kc as the int.erept on the j axiso Such a plot is shown in Pigure o40 Deindoerfer and Gaden (21) showed that the oxygen uptake rate was markedly dependent on broth rigidity and myoelium concentration, They used killed, reconstituted Peniaillium broth and measured oxygen conecentration polarographicallyo An 85% re duetion in the absorption rate occurred at a myeelial concentration of 13o4 go dry tissue per liter, This ooneentration of tissue produced a rigidity of 180 centipoises in the Bingham plastic brotho.Chain and Gulandi (12) showed that 1o35 peroent of dead mold mycelia reduaed the sulphite absorption rate by one-half. Solomons (78) reported that a 45% solution of sucrose, which had a viscosity of 10 centipoises, reduced the oxygen transfer rate by a fact r of 6 when measured by the sulphite method and by 12 when the polarographio method was used,.YTshida et al. (101) discussed the effect of increased viscosity due to the addition,f glycerol on the rate of oxygen uptake in stirred sulphite solutions;- at 300 RPM, an increase in viscosity from 1 to 3 6 tentipoises lowered the oxygen transfer rate by a factor of 2o4o With mold myceliium9 Solomons and Perkin (79) found significant reductiEns in the oxygen transfer rate (0TR) with increasing viscosities, Their viscosity data were based on a shear rate of 1 be og for apparent viscosities of 1, 10O

CO CO 0 u4 cd O I0. - Logarithm of Shear Rate Figure 4. Shear flow diagram for a pseudoplastic fluid, showing the flow behavior index, n, and the consistency index, kv.

16 100 and 1000 centipoisesp they found OTR's of 58, 42, 30 and 19 millimoles/(1)(hro) respectively, Solomons and Weston (80) furthered this work and found the OTR was a function of shear at the impeller according to the relationship = (IO)N (6) Bowers (7) simulated mycelia using shredded paper pulp suspended in sulphite solutions, Two percent of pulp decreased the sulphite OTR by 43 to 96%. Brierly and Steel (9) found as high as 90% reductio n in OTR when 20 g. of dry, filamentous As ergillus niger mycelia were suspended in a liter of aqueous salt solution9 and 85% reduction when shredded paper pulp was suspended at the same concentration levelO Sago pellets at 3% conoentration, had little effect on the rate of oxygen uptakeo At this concentration, the pellets had little effect on viscosity, Bartholomew et a lo (3) found a decrease in specific oxygen transfers i1,eo txygen uptake per cell, with increasing myoelial concentrations0 The rates of oxygen transfer in their studies were determined by polarographic methods. Phillips and Johnson (60)9 using Aspergillus niTe noted that oxygen deficiency in the fermentation occurred shortly after the culture began to exhibit a non-Newtonian viscosity behavi or Recently, Steel and Maxon (84) have reported on oxygen transfer studies in novabi in fermentations, They found an

17 initial decrease in the OTR as the apparent viscosity of the beer increased; then the OTR became independent of further change in viscosityo Concurrently9 with the increase in viscosity, an increase occurred in the amount of gas retained by a sample of the beero This summary of the published data shows that 'the effeet of increasing viscosity generally was examined in one of two wayso In the first method, oxygen transfer rates were measured in growing mold cultures and the change in OTR was followed with timeo In these systems, the increasing concentration of hyphae oaused the viscosity to increase as the fermentation progressed0 The second method was to measure the OTR in a system in which the viscosity was established at a constant value with suspended9 dead mycelia9 or solutions of glyeerol or sucrose. Either polarographic techniques or sulphite methods were employed in this case to measure the OTRo It can be seen thats in all of the cases reported here, the OTR fell as the viscosity increasedo However, each method experienced some side effects. In mold fermentations, there was the added resistance of diffusion into the pellets, Even though the oxygen concentration is high in the medium, it can be zero in the center of these pellets, Also, as the fermentation progressed9 both the viscosity of the medium and the characteristics. f the culture changedo The second method used a non-biological system and has been criticized as net truly representing living systems.

18 The shortcomings of the sulfite system as a method for measuring fermentation rates arise aoording to Blakebrough et al. (6)9 from two sour es0s essential differences between the processes which consume oxygen in sulphite solutions and in suspensions9 and the physical properties of solutions oThese authors also pointt out that the sulfite method measures oxygen which has reacted while the polarographic method measures oxygen in solution awaiting reactiono 4, Statement the Pro em A method for measuring the rate of oxygen uptake using Pseuco ona ovalis cells suspended in a nitrogen-free medium, was propesed by Tsao and Kempe (94) In the absence of nitrogen,.growth was arrested and the cells utilized one-half mole of txygen and one mole of glucose to form one mole of gluconic acido Hence. the rate of glueohio. acid production as measured by the amount of alkali required to control the pH at a aonstant values was proportional to the oxygen utilization rate of the. ellSo This method has many unique features when compared to other methods0 Sinae living cells are used9 conditions more closely appreximate these in commercial fermentation units than does-the sulfite systemo Alsoe this system doe*; not require sampling of the medium and measurement of oxygen transfer externally without aeration, as required by the polarographic metheod The gluconi aoid method iss howevers limited to small scale operations9 since it is impracticable to spin down

19 and resuspend large volumes of cellso Tsao and Kempe (94) found that an increase in the agitation rate increased the OTRo There was evidence to suggest that the increase in the OTR occurred even when the oxygen concentration in the liquid was high, Assuming this, they concluded that the increase in the OTR resulted from a reduction in the controlling resistance for oxygen transfer, which they postulated as being the liquid surrounding the cell, This assumption of a high concentration of dissolved oxygen in the liquid was important. If the dissolved oxygen concentration were below the critical value rather than above it, then the effect of increased agitation would more likely have been upon the gas-liquid portion of the overall oxygen transfer resistance than upon the cell-liquid portion, as they proposed, At the time this work was done, suitable instruments were not available to measure dissolved oxygen concentrations in the liquido In view of the critical aspect of this measurement for understanding more about the mechanism of oxygen transfer, it was important to further investigate the dissolved oxygen concentratonn during experiments of the type that Tsao and Kempe reported (94)o An additional mechanism for oxygen transfer, besides that occurring directly from oxygen dissolved in the liquid, has been reportedo Bartholomew et ao (3) proposed that oxygen transfer took place by two separate paths but were

20 unable to experimentally evaluate their theory at that time. With the bacterial fermentation of glucose to gluconic acid, a method for measuring oxygen transfer in living systems was availableo With thhe development of the oxygen electrode, the means of measuring dissolved oxygen concentration in the liquid was at hand0 If both of these methods were used together, it appeared that these mechanisms of oxygen transfer could be evaluated separatelyo The effect of viscosity Upon oxygen transfer has received little attentiono It is well known that long chain carboxymethyl cellulose (CM0) derivatives, glycerol and sucrose reduce OTR in non-biological systems, Some work with mold fermentations has shown that as fermentations progress, the increasing concentration of filamentous material reduces the OTRo However9 the effect of viscosity in bacterial systems has essentially been neglectedo The gluconic acid fermen= tation seemed admirably suited to this investigation, In place of growing mycelia to increase viscosity9 OMO derivatives could be used to give constant and high viscositieso

II. MATERIALS AND METHODS 1. The ermentor The equipment centered around a New Brunswick five liter fermentoro This unit consisted fundamentally of a Pyrex glass jar six inches in diameter and twelve inches high, and a head plate assembly inserted into the jar. Four baffles were attached to the head plate. The baffles were 0,625 inches wide and were positioned with their surfaces perpendicular to the walls of the jar, They extended from the top of the jar to within 0,125 inches of the bottom where they were connected by a circular band 0.625 inches high. An agitator shaft passed through a bearing on the head plate, Two impellers, 0.625 inches wide and 0,625 inches long were attached to the shaft with their flat surfaces perpendicular to the vessel's bottom. The distance from the tip of the blade to the center of the shaft was 1.50 inches, All of the components of the head plate9 baffles and agitator were constructed of stainless steel, For all runs, except the non-Newtonian viscosity series, one impeller was 2.5 inches and the other 6.5 inches above the bottom of the vessel. This resulted in the lower impeller being immersed in the liquid while the upper one contacted only the gas phase. For the series of runs with non-Newtonian broths, the upper impeller was lowered to 1, Model F-059 New Brunswick Scientific o,, New Brunswick, N, Jo 21

22 3.675 inches above the bottom of the vessel. This was done because it was necessary to have both impellers in the aqueous phase to obtain higher oxygen transfer rates and better mixing in the high viscosity broths, Compressed air was taken from the building supply. It 2 was passed through a cotton filter and a rotameter. A calibration curve for the rotameter is given in the Appendix, The air was humidified by bubbling it through water. From the bubbler, the air entered a plenum chamber, and then a sparger, which was attached to the head plate, The sparger consisted of a single orifice, centrally located and facing upwards 0.5 inches below the lower impeller. For one series of runs, the air used for sparging was diluted with nitrogeno To accomplish this, two more rotameters and control valves were required. The final flow rate of the mixed nitrogen and air was kept constant by varying the relative amounts or the other two flows, Control of the pH was effected with a pH meter3 and a 4 recorder-controller, A set of glass and reference electrodes, of the type normally used with a Beckmann model H-2 pH meter9 was installed in the fermentor. As the pH of the medium fell below a set value, the controller made contact through a micro-switcho This activated a motor which 2, Model 2F 1/4 205, Fischer and Porter Oo., Hatboro, Penn. 3, Model W industrial pH meter, Beckmann Instruments, Inc., Fullerton, Cal, 4. Wide-strip Dynamaster Millivolt iMeter, Bristol Co., Waterbury 20, Ceonno

25 powered a Sigmamotor pump that pumped alkali from the burette until the pH rose to the set pointO The alkali was introduced into the medium in the fermentation vessel through an entry port in the head plateo Sodium 'ydroxide was used as the neutralizing alkali, Depending upon the situation, its concentration was varied from Ool to lo0 N, It was standardized by titration against potassium acid phhalat s phphthaat usi hthalein as the indicator An electrical relay was installed to reduce the electrical load on the micro-switch, since the initial power surge was large when the motor started. When no relay was used, the micro-switch often stucko The use of a 50 co burette allowed accurate determinations of the amount of alkali added. This, coupled with a small diameter tube in the pump, provided pH control within Oo0 5 pH unitso The fermentor was immersed in a constant temperature bath that controlled the temperature of the fermentation broths within ~ O05o0 of the desired value, A temperature compensator was immersed in the bath to correct pH readings on the recorder-aontreller for variations in temperature, Power for stirring was supplied by a converted drill presso The transmission was effected through a V-Belt and horizontal pulleys, one being mounted on the drill press and the other on a frame above the constant temperature bath, The stirring speed was quickly and easily altered by changing

w w H c Jd E O Y M: Na o_ _be Figure 5. Equipment used in the study of the rates of oxygen transfer. R)

25 Legend for Figure X Ao Indicating pH meter B. pH recorder and controller C, Electrodes D, Temperature compensator Eo Oxygen electrode Fo Adapter for oxygen electrode Go pH meter H. Nitrogen tank I. Valve J Air supply from building K, Flow-meter Lo Cotton filter M. Bubbler containing distilled water No Plenum chamber 0, Sparger P. Baffle Q, Sample port R. Thermometer S. Pulley T, V-belt drive U. Motor V. Agitator shaft Wo Two positions of upper impeller X. Lower impeller Y, Fermentor9 glass jar Z, Head plate a. Sigma pump b. Driving motor for Sigma pump o, Burrette for sodium hydroxide d. Sodium hydroxide feed line e Relay f Constant temperature bath g. Power sources ll0V

26 the diameter of the pulleys, The stirring speed was measured with a revolution counter and a stop watchO 2o The Ox en Eleetrode Since the time when Olark (15) first introduced his oxygen electrode, it has been used in a number of biological systems (579 82, 97). The oxygen measuring device used in this study5 is shown in Figure 6o It consisted of a silver anodead nd a platinum athode covered with a polyethylene membrane two to four mils in thickness. The cathode and the silver anode were immersed in an electrolyte of 3% KOLo The electrodes were encased in a dip-cell made of glass; the membrane was held securely against the cathode by a plastic sleeveo This sleeve was altered so that it slipped compietely up, onto the body of the cell; no over-hanging edges were present to trap air bubbleso The electrical circuit was such that a potential of o06 to 0,7 volts was supplied to the electrode by a size D dry cell9 housed in an adapter unite The adapter was connected 6 -- to a pH meter, At this operating voltage, dissolved oxygen was the chief substance reduced at the cathode0 The reactions taking place were, 2 01 + 2 Ag - 2 AgOl + 2e (anode) I02 + H20 + 2e - 20H (cathode) 5, NEo 11098 Oxygen Electrode and No 18902 Adapter unit, Beckmann Instruments9 Inco, Fullerton, Oal, 6. Model H-29 Beckmann Instrunients Inc,, Fullerton, Cal,

27 lectrode Leads to the Adapter Electrode Cap — Glass Dip Cell -— Potassium Chloride Solution ir Silver Anode -Membrane Sleeve I / A, L —Membrane Platinum Cathode Figure 6. Schematic drawing of the oxygen electrode,

28 Thus, the current flowing in the system was a function of the number of oxygen molecules being reduced at the surface of the cathodeo The potential drop of this current flowing through a known resistance was indicated on the pH meterO This potential drop9 in millivolts. was therefore a measure of the concentration of dissolved oxygen in the solution, It is generally agreed (31, 789 85) that the oxygen electrode measures oxygen activity in solution rather than concentration of dissolved oxygeno The electrode cannot be calibrated by observing the response when placed in various air-saturated salt solutions, as is done with dropping mercury electrodeso To test this statement by Finn (31), the electrode was placed in distilled water whioh had been saturated with oxygen by bubbling air through it for one hour, Then the electrode was placed in a solution of sodium chloride or sucrose which had similarly been saturated with oxygen. All of these solutions were at the same temperature, The electrode reading was independent of the concentration of the additive and did not change from its initial reading in the pure water0 To determine the dissolved oxygen concentration of a solution, it was first necessary to standardize the electrode, With the elebtrode disconnected9 the pH meter was set at zero on the 0-800 millivolt scale, Then the cell was placed in water at 250 00 which had been saturated with oxygen by bubbling air through it0 The reading on the 6001400 milli

29 volt scale was adjusted to 1400. After the meter reading remained steady for 5 to 10 minutes, the electrode was disconnected and fastened to one of the baffles in the fermentors so that the membrane was approximately one inch below the liquid surface during sparging and stirring of the liquid. The percentage of oxygen saturation was calculated by dividing the pH meter output, in millivolts, by the fullscale value for saturation, Occasionally, the set point for a saturated solution was less than 1400 mvy In these cases, the reading at saturation, rather than 1400, was used to calculate the percentage of saturation. The manufacturer reported a linear relation between the instrument reading and the dissolved oxygen concentration. This claim was checked by placing the probe in solutions containing various concentrations of dissolved oxygeno By mixing saturated and boiled water, the amount of dissolved oxygen in solution was varied, The oxygen electrode reading, in millivolts, was recorded and plotted against the dissolved oxygen concentration determined by the Winkler method (81), The data showing linearity, as reported, are recorded in Appendix B. To measure the rate of oxygen uptake with the electrode, the air supply to the fermentor was stoppedo Then the rate of change of the dissolved oxygen concentration in the liquid was recorded, The slope of the plot of dissolved oxygen concentration against time then became the apparent rate of

350 oxygen uptake by the cellse Before calculation of the oxygen utilization rates however9 it was necessary to account for surface reaerationr Reaeration occurred when oxygen diffused through the surface of the liquid into the medium at the same time as the cells were removing oxygen from the solution, Blanketing the surface of the liquid with nitrogen did not step surface reaeration, To completely purge oxygen from the space above the liquid took about 15 minutes9 during which time the- solution was exposed to a gas of varying oxygen conoentrationo When pure nitrogen finally overlaid the surface9 there was counter-diffusion of oxygen from the liquid surface into the gas., Hence, there was a two-fold loss of oxygen to the cells and to the atmosphere. It seemed best to allow air to cover the liquid surface and to correct for surface reaeration, To determine the amount of oxygen supplied through the surface9 the fermentor was charged with a glucuo$ephesphate solution at 2500o The liquid was then stripped of oxygen by sparging nitrogen through it, When the solution was free of oxygen the nitrogen flow was stopped, the nitrogen above the surface of the liquid was replaced by air and exygen was allowed to diffuse into the solution through the surface, The short time taken to replace nitrogen with air was not critical in this case9 since reaeration continued for at least an hour before the liquid became saturated with oxygen0 The oxygen ooncentration in the liquid was recorded as a function of time,

31 From these data, it was possible to calculate a mass transfer coefficient for reaeration, The method of calculation of the coefficient is given in a later section. Using an electrode of their own design, Phillips and Johnson (59) found a significant lag in meter response to changes in dissolved oxygen concentration. To check this, the electrode previously described was placed in a saturated solution of oxygen in water which was being stirred. A mixture of sodium ulfitte and copper sulfate was added in order to quickly reduce the dissolved oxygen concentration to zero0 The meter reading fell from 1400 to 100 in 18 seconds and to 0 in 30 secondso This test was repeated with a sodium sulfite solution containing cobalt ions with similar results, 35 The RoLational Vicoete Since sugar and glycerol form Newtonian solutions, their viscosities were measured with a capillary tube viscometer at 25,000o Garboxzm thyl celluloses (OMO) however, form nonNewtonian solutionso Because their flow curves are not linear, another instrument was necessary to measure the viscosity of the OMO solutions, For this purpose a rotational viscometer (70) was used, The material was confined between long vertical coaxial cylinders, The outer one could be rotated at various speeds, while the inner one was attached to a resistance wire strain gauge that. converted the shear force caused by contact of the fluid anthe fe ixed internal cylinder into a direct current electrical signala A tachometer generator geared to

32 the drive of the outer cylinder provided direct current signals proportional to the rate of shear These two currents were fed to an X=Y recorder, which plotted the shear stress from the strain gauge9 as a function of the shear rate from the tachometer generator Thus, a continuous flow curve was plotted for the sample as the speed of the rotating cup was progressively increasedo 4. Bacterial Secie A strain of Pseudomonas ovalis NRRL B-8S was obtained from the Northern Utilization Researoh and Development Branch of the United States Department of Agriculture located at Peoria, Illinois, Lockwood et ale (47) had previously reported that this strain oxidized glucose to glueonic acid without the development of side products This organism was also used by Tsac and Kempe (44) to measure oxygen uptake by Pseudomonas ovalis in resting cell suspensions0 5o Media A medium of the following composition was used for cultivating the microorganism Glucose 50 grams Yeast extract 5o0 grams Urea 2o0 grams KH2PO4 0o6 grams MgSO4 7H20 Oo25 grams Distilled water lo0 liter When a solid medium was required9 2% agar was added. About

33 ten mlo of the agar medium were slanted in each test tube. Twelve hundred mlo of the liquid culture were placed in a two-liter Erlenmeyer flask which was stoppered with a cotton plugo Both the slants and the liquid culture were autoclaved at 15 pounds gauge pressure for 25 minutes, To resuspend the cells9 a nitrogen-free medium of the following composition was employed: Glucose 50 grams K2HPO4 09 grams KH P04 0 6 grams Polyglycol P-20007 0025 cc Distilled water 2o0 liters pH 7~0 Since there was no need to maintain aseptic conditions, this medium was not autoclavedo 6o Stock Cultures In order to preserve the original culture9 the organism was grown in a liquid medium in test tubes for 24 hours at 3C~0, after which the cultures were frozen in a dry icealcohol bath and stored at -1000o Every three months, one of these cultures was quickly melted and used to inoculate fresh tubes of sterile media0 After incubation, these cultures were added to the collection, which was stored at a low temperature, 7, QDow Chemical Oompany Midland Micho

34 7o General Experimental Conditions The organism was carried on agar slants which were incubated at 3000 ~ 2 Co Daily transfers were made from a 24-hour slant to a fresh slantIn preparation for a run, a slant culture was streaked and incubated for 24 hourso About ten mli of a sterile liquid medium were aseptically poured into the slant tubeO The cells on the surface of the agar were mixed into this medium with a sterile transfer needle, and the suspension was then poured into a two-liter Erlenmeyer flask that contained 1200 mlo of fresh medium, The inoculated medium was incubated for 24 hours at 30~C on a rotary shakero8 A setting of 3,0 on the shaker dial resulted in a shaking action of 180 RPM with a circular motion lo5 inches in diametero 8,o hisal and Chemical Measurements All measurements were made on cells grown in liquid medium at 3000 for 24 hourso The dimensions of 10 cells were measured with an American Optical A0 Spencer microscope using a lOX eyepiece and a 97X9 oil immersion, obJectiveo The ocular micrometer in the instrument was calibrated against a stage micrometer. One slide was made from a lO'2 dilution of the cells in physiological salineo It was fixed with heat9 stained with crystal violet and examined under oil immersion0 Another smear of cells in their growth medium was not heat 8 Model VS, New Brunswick Scientific Co, New BrunswickO NL J.

35 fixed, but was also stained with crystal violet and examined on a wet mountO A third measurement of cell size was made using unstained cellso In thI s instanie, the growth medium was diluted with physiological saline and the cells were examined on a wet mount. The moisture content of cells harvested in a Sharples continuous centrifuge9 was determined by placing about one gram of packed cells in tared, dry, aluminum-foil disheso The cells were dried at 1100C until a constant weight was obtained To measure the ash content, about one gram of packed cells was placed in a tared, dry crucibleo After combustion for one hour at 18000F9 the crucible was cooled in a desiccator and reweighedo In the following four chemical analyses, a stock solution of 10% by weight of cells in distilled water was usedo The stock solution was further diluted to llO)09 l lO0 and lslOJ000 to give readings within limits of the tests o To analyze for protein9 the Folin phenol method (48) was employed, The method was a celorimetric one, in which the sample was treated with an alkaline copper solution and then with the Folin reagento The intensity of celer of the blue solution which developed was determined in a colorimeter using a red, 660 mR filtero The actual amount of protein was then obtained when the reading was compared to a standard curve prepared using crystalline bovine albumin, Distilled

36 water was used to set the zero reading on the colorimeter, The diphenylalanine method (72) was employed to analyze for deoxyribonucleic acido Here the sample was extracted with cold 5% trichloroacetic acid, centrifuged and the supernatant discardedO The precipitate was treated with diphenylalanine reagent and the solution boiledo After the solution was cooled, the intensity of the color was determined with the colorimeter, using a green, 540 mp filter. Distilled water was used to set the zero reading on the colorimeter, The amount of deoxyribonucleic acid was obtained when the reading was compared to a standard curve prepared using Herring sperm DNA which was run concurrently with the Pseudomonas ovalis sampleo The orcinol method (16) was used for the determination of the ribonucleic acid content of the cells, This technique involved an initial extraction with cold, 50% acetic acid followed by centrifugation and treatment of the precipitate with an HOl FeOl3 reagent and an alcoholic orcinol solutiono The solution was boiled9 cooled and the intensity of the color evaluated in a colorimeter using a red 660 mu filtero Distilled water was used to set the zero reading on the oolorimetero The readings were compared with a series of solutions of known concentrations that were run concurrentlys and the amount of ribonucleic acid determinedo The amount of carbohydrate was determined by the an te-sulfuric acid method (71)o The cells were boiled with -s-te csulfuric acid reagent, cooled:, and the color

37 intensity was determined with a colorimetero A red, 660 mu filter was employedo The blank used to set the zero reading on the colorimeter was a sample of distilled water carried through the test concurrentlyo Results were obtained when these readings were compared to standard glucose solutions tested at the same time, The growth characteristics for Pseudomonas ovalis in a shaken culture were determined at 30~0o The test was performed with 100 ml. of medium in a 250 ml. Erlenmeyer flask, to which a side-arm colorimeter tube had been attached, The flask was inoculated with 0,01 ml. of a 24 hour liquid cultureo This inoculum was obtained by aseptically diluting one ml, of a 24 hour broth in 100 mlo of fresh, sterile mediao As growth occurred9 turbidity and protein content were frequently determined. Viable and total cell counts were made at only a few point o To analyze for protein9 one ml, of the culture was aseptically removed and centrifuged9 for 15 minutes at about 1800 RPMo The supernatant was discardedg the cells were resuspaedeandrecentrifuged under the same conditions o One ml, of water was added to the drained cell pack and the cells were resuspended and stored at 00C for a period not exceeding 24 hourso Storage at such low temperatures does not affect the protein ontent (28)o The Folin reagent method described 90 Model 0CL International Clinical Centrifuge, International Equipment 0o,, Boston, Masos

38 previously, was used to determine the protein content of the cells o To measure the rate of thermal killing of the microorganism9 one mlo of a 24 hour culture ef Pseudonomas ovalis was placed in 99 mlo of a gelatin-saline solutieno The saline in each dilution bottle was equilibrated with the water in a constant temperature batho Temperatures employed were 25,09 48o0, 53.5 and 60o5Q0 (+ OoS~0)o The decrease in the number of viable cells was determined at each temperature by making pour plates with trypticase soy agaro These plates were incubated in an inverted position at 3500 for 24 hour and countedo 9 0 Procedure for a Run The culture was removed from the shaker9 and about one mlo of loO N sodium hydroxide was added per 100 mlo of broth to raise the pH from 403 to 7o00 Then the cells were vigorously aerated for 30 minutes by means of air blown through a sintered glass spargero After this neutralization and aeration process9 the cells were separated from the broth by centrifuginglO at 1300 G for 20 minuteso The supernatant was discarded and the cells resuspended in two liters of the nitrogen-free glucose medium described previouslyo This- cell suspension was poured into the fermentor: the agitator and air flow were started and a sample of the medium takeno.The -pH of thisn initial sample was determined with a 10, Model V. Size 29 International Centrifuge, International Equipment 0o00 Boston9 Mass0

39 Beckmann model H-2 pH meter and used to standardize the recorder-controller. This sample was also used for measuring the initial turbidity. The rate of gluconic acid production was obtained from the rate of sodium hydroxide addition. The level of sodium hydroxide in the burette was read every 5 to 10 minutes; readings were continued until a steady rate of acid production was established which usually required about one hour. When the concentration of dissolved oxygen in the liquid medium was desired, the agitator was stopped and the head plate removed. The oxygen electrode was then mounted on one of the baffles and the run started again. A steady reading for the oxygen concentration in the solution usually established itself within 5 minutes. To measure the rate of oxygen utilization with the electrode, the air was shut off, but the agitator was not. As the dissolved oxygen was used by the cells, the level in the fermentor fell; then the data for the concentration of dissolved oxygen in the solution was plotted against time to determine the rate of removal. 10. Bacterial Densities The concentration of bacteria was measured with a KlettSummerson photoelectric colorimeterll using a blue, 420 ml filter, against a distilled water blank. In order to compare data from several runs, it was desirable to find a correlation 11. Klett-SummerSon Photoelectric Colorimeter, Klett Manufacturing 0o., New York.

40 between cell concentrations and turbidity. A concentrated suspension of Pseudomonas o liin the nitrogen-free medium described previously, was diluted with an equal volume of the same solution, The turbidity of the suspension was read with the colorimeter before and after dilutiono The diluted sample, in turn, was mixed with an equal volume of solution, and the turbidity was determined again. This process was repeated until the suspension was almost clear. To relate turbidity and viable cell numbers, counts were made at two different cell concentrations of Pseudomonas ovalis in the buffered mediumo The cells in the glucosephosphate medium were spread on the surface of trypticase soy agar, inverted and incubated for 24 hours at 3500, Dry weights were also determined for several cell concentrationso Twenty-five mli of the cell suspension in phosphate solution at pH 7.0 were pipetteted into tared aluminum disheso The phosphate solution contained 0,9 go K2HP04 and 0,6 KH2PO4 per liter of distilled water, The dishes were placed in an oven at 10500 and reweighed after 24 hourso There were three samples for each concentration including the phosphate solution used as the buffering agent, The weight of the blank was subtracted from the dry. ieghto 11. Alteration of Visosity For these studies it was desirable to use a nonNewtonian fluid having an easily varied viscosity. It was also desirable to add as little solid material as possible

41 to the water so as not to alter the oxygen solubility or the osmotic pressure0 ~~13~~~ ~12 The materials chosen were Natrosel 250 H12 and Methocel 65 HG9 4000 OP13 Both of these compounds are high molecularweight cellulose etherso They are readily soluble in watero Small amounts of them increase the viscosity markedlys and neither is affected by moderate concentrations of acids, bases or saltso They are fairly resistant to microbiological degradation- and their visoesities show no time dependency up to 2% by weight of the dry material in waterO The density of these solutions is approximately o10 g/mlo Natrosol is cellulose etherified with hydroxyethyl groupso In a 2% solution, the viscosity measured by a Brookfield visoometer may be as high as 259000 centipoises at 250 (39)o Methoeel is a propylene glycol ether of celluloseo A 2% solution of 65 HG results in a viscosity of about 4000 centipoises when measured at 25000 To initiate a run9 the cells were grown and harvested in a manner described previouslyo Since turbidimetric readings were affected by the viscosity additives9 a suitable way to control cell concentration was to spin down the same volume of a 24 hour broth culture each time, About lo:2 liters of a 24 hour cell suspension were used for each run? the Klett reading was about 100 if the cells were resuspended in 2 liters of distilled watero 12o Hercules Potder qompany Wilmington9 Delaware 13o Dow Chemioal Company9 Midland9 Michigano

42 For a run, the cell pack was resuspended in a small volume of distilled water, This concentrated cell suspension and a polution of 50 grams of glucose in water were added to a solution of the viscosity agent which had been prepared the night before o The total volume was made up to 2o00 literso This suspension was poured into the fermentation vessel, and the ae ate of glueonic acid production determined using a stirring speed of 570 RPM and an air rate of lo77 VVM in order to maintain a high dissolved oxygen level in the liquido By placing the oxygen probe into the medium at the end of the run, it was possible to determine the dissolved oxygen content of the brotho A saple of the vistous broth of each run was saved and stored in a refrigerator in order to determine its viscosity at a later time0 The viscosity test required about 300 mlo which were placed in the central compartment of the rotational viscometer allowed to equilibrate at 30~0 + 0 olO0 During the viscosity test9 the speed of the outer cylinder was increased automatically in steps from 32 to 438 RPMo After raching maximum velocitya the process was reversedo A curve f shear rate against shear sr tress was obtained, 12~ Sul fite Oxidation Many oxygen transfer studies have been conducted using oxidation of sodium sulfite to sulfate in the presence of a suitable Oatalyst9 such as cuprio or cobalt ionSo The method

43 was intro4uced by Miyamoto et a. (53, 54), and adopted for the evaluation of agitated gas-liquid contactors by Cooper et ale (19)o In this series9 no living cells were employedo Three liters of an aqueous solution containing 0.25 mole of sodium sulfite per liter were placed in the fermenter, Air was bubbled through the liquid at a constant rate of 1.0 VVM* The agitator was set at 300 RPM, and a solution of OuS04 was added to provide a concentration of cupric ions of 103 moles/lo Ten mlo samples were withdrawn at appropriate intervals and pipetted into 25 ml, of an iodine solution of known concentrationo The excess iodine was back-titrated with a sodium thiosulfate solutiona In this way, the concentration of the sulfite was determined; the rate at which sodium sulfite disappeared was a measure of the rate of oxygen transfer0 Details of this method, including equations for the chemical reactions inTvlved, have been described more fully by Tsao (93)o The visaosity was increased by incorporating sucrose or glycerol into the original three literso The viscosity of the medium was measured at 25o000 with an Ostwald visoometer calibrated against liquids of known viscosity. 139 Measurement of Oxgen Transfer Rates Us the Oxten The rate of oxygen transfer to cell-free solutions was determined by measuring the increase in the dissolved oxygen

44 concentration in the liquid with the oxygen electrode, Before each aeration test, the water was deoxygenated; in the series of runs in which the agitation rate was studied, nitrogen was bubbled through the solution; in the series to determine the effect of viscosity and air flow rate, a small amount of sodium sulfite was added, All tests were conducted at 25,0o, No ohanges were made in stirring speed, air rate or medium composition during each test. In the investigation of the effect of agitation, no air was supplied to the sparger, The only supply of oxygen to the liquid was from air overlaying the surface. The liquid medium was the same glucose-phosphate solution used to resuspended the Pseudomeas valis cultures In the other two tests also conducted at 25 ' 00, the agitator was consta tly turning at 300 RPM. Air was sparged into the liquid at the rate of 0,2 VVM in the series in which the effect of viscosity was examinedo In these tests, Methoeel was disselved in distilled water, In the last series the air rate ewas altered The medium contained 2,5 g/1 of Methoeel dissolved in distilled water, The operation of the electrode has been described previously in detailo In this series? it was set at a zero reading in the deoxygenated liquid and at a full scale deflection of 1400 millivolts in water saturated with oxygen when in contact with air Both of these settings were accomplished at 25,0 0,

II1, EXPERIMENTAL RESULTS A0o Prelma Ex erim 1o Ph sLcal and Chemical Measurements Microscopic observation of the cells showed them to be rods with rounded endSo The dimensions9 which varied somewhat according to the methods of staining and fixing, were: l16 x 005 microns when stained with methylene blue and the cells were dry mounted 2 5 x 0 5 microns when not stained and the cells were wet-mounted 2o5 x 0o7 microns when crystal violet was added and the cells were wet-mountedo These data agree with those of Aiba et al. (1)9 who reported values ranging from (lo4 to 2,5) x 0.6 microns for his 8-hour cultures of Pseudomonas ovalis. An electron photomicrograph of these bacteria magnified 10000 X is shown in Figure 7, Chemical analyses of the bacteria are reported as percentages of the dry weight in Table II; moisture content of packed cells was found to be 71o4% of the wet weight, 2, Growth Ourve Readings of turbidity and protein content defined the growth curve, Viable cell counts9 at a limited number of points9 provided date for demarcation of the stationary and declining pha-seso The generation time was found to be 42 minutes during the exponential growth periodo This value for the generation 45

46 TABLE II CHEMICAL ANALYSIS OF Pseudo onas ovali Determination Percentage by Weight....................(dry basse) Protein 34 Carbohydrate 20 Ribonucleic acid 901 Deoxyribonueleic acid 7 8 Ash 6o7

47 Figure 7. Electron micrograph of Pseudomonas ovalis NRRL B-8S, grown for 2i hours on an agar surface (X 10,000).

48 time is in general agreement with reported values of 30-40 minutes for most microorganisms, The initial lag phase lasted 7,5 hours. A maximum viable population of 1.6 x 109 oellsal,. occurred at about 12 hours, following inoculation, From these data, it can be seen in Figure 8 that the cells entered the stationary phase after approximately twelve hours, There was no significant decline in the number of viable cells at 24 hours; however, at this point the culture was in the stationary phase of growth, In the stationary period, there should be very little change in the characteristics of the cells with time this should result in good reproducibility between runs in which such cells are used for the conversion of glucose to glueonic acid as described previously~ 3. Thermal Death Rate The thermal destruction of bacteria is generally considered to be caused by coagulation of cellular protein (43). As such, the death rate can be represented by a first order reaction, where the rate of death is directly proportional to the number of living cells present at the beginning of any interval of times d1 3 -knt (7) where No is the number of bacteria in the volume under consideration, t is the time of exposure to the killing temperature and k is the rate constant for thermal deo structiono

400 - 2.5 C- ro;o 3 /0 4 0 Qr-..I // i7 <Da 9-4i ~~~~~~~~~d0 0 0 H I \D ro X I \ -21.5 04.,4 0 -P JjZZ mn_ r^ I ~~~o Viable Cells 1200 - $ Klett Reading -. 1.0 bo OH 5 0 0 5 10 15 20 25 30 Time (hours) Figure 8. Growth of Pseudomonas ovalis, NRRL B-8S in a 5% glucose medium at 30"C.

50 TABLE III GROWTH OP Pseudomonas ovalis NRRL B-8S A 5% GLUCOSE MEDIUM AT 300C IN Time Hours 4.5 8.5 9,2 9.5 9.9 10. 5 1100 12.3 12.5 22 8 23.3 24.3 30.5 47.0 52.0 Klett Reading 28 49 67 80 105 140 167 212 218 350 355 360 385 420 425 Protein Contents ____9/ml. 22 33 44 60 84 118 166 233 233 259 290 300 Viable Cells _ number/l 1.1 x 105 1.0 x 106 6.8 x 106 7.0 x 108 1.6 x 109 1.3 x 109 1.3 x 109 1,9 x 108

51 At room temperature, there was no measurable decrease in the concentration of viable cells. At the other four temperatures, the number of cells decreased quickly. The data for cell numbers were plotted in Figure 99 and a slope, k, was determined for each temperature. Q10 values, which are defined as the ratio of the velocity constant at one particali lar temperature to the velocity constant at a temperature ten degrees lower, were calculated using the formula: k10 log Qlo = T log ( (8) These Q10 values are found in Table VI. Lamanna and Mallette (46) reported that the Q10 values of spores were in the range 3,8 c 10.70 with higher values occurring as the temperature of killing increasedo Comparable values for vegetative cells may reach 70. Both Deindoerfer (20) and Rahn (63) reported that the Q1 values decrease as the temperature decreases, as was found in this worko Rahn (63) gave several values for temperature coefficients in the range 55 6000, These rose as high as 42 for Salmonella s:cotmuelleri, but usually were around 20 for the Salmonella species, Sa nella sottmuelleri, like Pseudomonas oyalis, is a grah negative rod, In any given environmen t, values for a species generally vary with temperature in a fashion which can be represented by the Arrhenius equation. This equation is usually used to represent the large changes in velocity of chemical reactions with temperature (4):

rH H H 0 r-1 a) 0 6 -0 0 Rl (D Q4-I CO I4 10 Time (minutes) Figure 9. Effect of temperature on survival of Pseudomonas ovalis in a gelatin-saline medium.

53 TABLE IV EFFEOT OP TIME AND TEMPERATURE ON THE NUMBER OF VIABLE CELLS OF A 24 HOUR CULTURE OF Pseudomonas ovalis CELLS IN GELATIN-SALINE BROTH Time of Exposure.minutes __ Number of Viable Oells at the Fallowing.Tem erature., of -Exposr e., __0. -5,j6 48io _ - 5 - 0 1/2 1 1 1/2 2 3 -5 6 8 10 7;8xlO6 2. 5x05 9 5xl06 4.3x106 6.0x106 4.5xlo 0 OxO2 1 o OxlO 9 6xlO6 8. 0x106 AVo 8.5x106 9.x oz10 3.6x106 7. 0x05 1.6xlO5 7.6x10 1.5x106 1.0xl15 7.2x103 7,8x108 lo1xl06 9,0xl02 7.5x106 5.Oxl05 60Qx0l 6 3.o0x0o5 6.0xl~6 3,Ox105 0

54 TABLE V EFFECT OF TEMPERATURE ON THE THERMAL DEATH RATE CONSTANT FOR A 24 HOUR CULTURE OF Pseudomonas ovalis IN GELATIN-SALINE BROTH Temperature 48 53.5 55,5 60 5 0.022 0.18 0.66 3.9 k is the thermal death rate constant for Equation 7. which relates number of surviving cells to time,

55 TABLE VI TEMPERATURE COEFFICIENTS FOR THERMAL DEATH RATES OF A 24 HOUR CULTURE OF Pseudomonas ovalis IN GELATIN-SALINE BROTH Temperature Range Q, 48 ~ 60.5 96 5355 - 60.5 82 55.5 60.5 35

56 k = AA e /RT (9) where A is the frequency factor, R the universal gas constant, ~A T the absolute temperature and EA is an apparent activation energy for heat of destruction of the microorganism, Taking logarithms: E log k = log AA - A (10) 2.3 RT For all practical purposes, log AA may be treated as a constant (20)o Then the equation is of the form y = mx + b and a plot of log k versus 1/T for any species should yield a straight line with a slope of -E /2 3 RT. The data were plotted in Figure 10. Using the slope of the resulting straight line, the activation energy was found to be 90000 calories when the data for the thermal death rate of Pseudomnas.alis were examined by the method of least squareso The intercept, log i, was found to be 62.2. Chick and Martin (13) reported values for protein denaturation and the lethal action of heat on bacterial spores to be 60s000 and 1359000 respectivelyo Pfeiffer and Vojonovich (58) reported the range of Arrhenius activation energies to be 50,000 to 100 000 cal/moleo 4o Products of the Fermentation ef Glucose b Pseudomonas ovalis Pseudomonas ovalis has been reported (47) to yield only gluconic acid when supplied with glucose and oxygen as nutrientso The equation for the reaction has been given by

57 10. r-l H I r-1 H (D 0 0 C) qr4 a) 0) 4-P 0 a) CD 0 0) 1.0 0Q1 0.011 3.00 3.02 3.04 306 3.08 310 312 Figure 10. l/Temperature (O~K x 103 Arrhenius plot for the effect of temperature on the survival of Pseudomonas ovalis in a gelatinsaline medium.

58 Tsao (93): 06H1206 + 02 6-' %12 7 glucose gluconic acid Tsao (93) reported conversion of 99% of the glucose to acid at pH 5.5. Aiba et al. (1) found 95% conversion at pH 6.0 in a resting cell suspension. A resting cell suspension of Psead o valis was produced by separating cells from their growth medium by centrifugation. The liquid was discarded and the cell pack was resuspended in a glucose solution containing no nitrogen. The absence of nitrogen prevented the cells from reproducing, Therefore, the only process taking place was respiration, and all oxygen utilized was used for this purpose, A suspension of this type will be referred to as a "resting" cell suspension. The rates of glucose utilization and of acid production were measured in resting cell fermentations at pH 5.5 and 7.0. The concentration of glucose was measured by the anthrone method; the acid formation rate was determined by measuring the rate of sodium hydroxide addition. Both values were plotted against time with the degree of conversion being taken as the slope of the glucose utilization rate divided by the slope of the acid production rate, A conversion of 96% was found for the single run made at pH 5o5. At pH 7, three runs were made with 93, 98, and 100% conversions being found. To determine the pKa of the acid produced in the fermena I

59 tation of glucose by Pseudomonas ovalis, cells were resuspended in a solution of glucose and water. The phosphate buffer was used, The cells were allowed to ferment the sugar until approximately 0O08 moles acid had been producedo This suspension was centrifuged to remove the cells; the supernatant liquid was decanted and acidified. It was then titrated with sodium hydroxide from its initially low pH of 3 to a pH of 11. The buffer index, 9, was determined for each addition of base by dividing the amount of base added by the change in pH resulting from that addition. Van Slyke (95) has shown that the pK is the value of the pH at the maximum value of the buffer inde~x ioeos the point at which the addition of either acid or base results in a minimum pH changeo The supernatant liquid from the experimental run yielded a value of 3.56o The Merck Index (50) lists the pKa value for gluconi ai aid as 5360 The pK values for reagent grade sodium and potassium gluconate were also testedO They were found to be 3,58 and 3,70 respectively. 5. T rbidit n the Estimation of. Bacterial Populations The most widely used procedure for the measurement of bacterial density is the determination of transmitted or scattered light (46)o Data for a series of two-fold dilutions of a concentrated suspension of Pseudomonas ovalis with a glucosephspphate medium are recorded in Table VIIo The turbidity9 measured in Klett units9 has been plotted against cell concentration in Figure 12, Arbitrarily, the

24 20 16 CO_ x12 PKa - O 4 3 4 5 6 7 8 9 10 11 pH Figure 11. Graphical determination of the pKa of the acid formed during the fermentation of glucose by Pseudomonas ovalis, NRRL B-8S in a resting cell Suspensi n.

TABLE VII RELATIONSHIP BETWEEN TURBIDIf AND THE LOGARITHM OF THE MULTIPLE OF DILUTION OF A SUSPENSION OF Paed^monlas i WITH A GLUCOSE-PHOSPHATE MEDIUM 9it-hm of the tTt td Iy. o Ce 11I Fra ti onl 1 Met ameters of D tip of e oef te.... Concentration Trans mitt Ianre (Optical Density) Logar Mul DiuTton t-to the Base Two S u e.; JLial%:, of Light Klett Re ading Ar bit:ay Uni D log D or 0.6+log (1-~D)_ 0 1 2 3 4 5 6 7 8 9 10 11 975 810 630 465 333 206 85 46 24 10 7.5 5.0 512 256 128 64 32 16 8 4 2 1.0 0.5 0.25 0.0167 O,024 5 0,0549 0.112 0.216 0.387 o.667 0.810 0.896 0*955 0.966 0,9775 1.777 1.611 1.260 0.951 0.666 0,412 0.109 -0.121 -0,383 -0.747 -0.869 -1.048 H-C

62 400 300 0 -< 4 -43 "' 200 4Pa -H oH "7:-,Q E 100 0 1 2 3 Cell Concentration jcells/ml) x 101o 4 Figure 12. Relationship between turbidity and cell concentration for a suspension of Pseudomonas ovalis in glucose-phosphate medium.

63 cell suspension having a Klett reading of 10 was taken as a concentration of cells equal to 1. Each succeeding measurement had a cell concentration twice that of the previous one. In the lower range of cell concentrations (with the Klett reading less than 100), the graph was linear, indicating that Beer's Law was obeyed, Beer's Law is an expression which relates the absorbance of light to the density of a given bacterial suspension. It can be expressed as: logl =s 0 (11) where I is the intensity of an incident beam of light striking a solution, Io the intensity after passing through the solution, r the turbidity coefficient and O is the number of microorganisms in cells/ml. When Beer's Law is followed, as is the case for lower concentrations of Pseudemos ovalis a plot of log (Ia/I), or Klett reading which is directly related to it, is linear, In this range, all corrections for deviations in cell concentration were made using Klett readings directly. For example, in the study of the effect of pH on the rate of production of gluconic acid, all rate data were corrected to a cell concentration corresponding to a Klett reading of 100, If in one run the Klett reading was 110 and acid was produced at the rate of X meq/(l)(hr.), then the corrected acid rate for a run having a cell concentration corresponding to a Klett reading of 100 would be (100/110) X me/(l) (hr.).

64 At higher cell concentrations (greater than a Klett reading of 200) a considerable departure from a straight line occurred Hence, the usual method to determine the number of cells in a medium containing a high cell population is to dilute the sample until a more suitable range is reached, Kurokawa ei aIo, (45) however, showed that the data for turbidity could be represented by a linear plot over the whole range of measurable values, if (-)log D for (0 < D < 0.5) and 0,6 + log (1-D ) for (0.5 < D ~ 1,0) were plotted against the logarithm to the base two of the multiple of dilution of the suspension, The term (-)log D has been defined as optical density, and D is the fractional transmittance of light. These data are plotted in Figure 13, Oorrections for deviations in cell concentrations at higher levels can be made using this plot, For example, it can be seen that a medium which has a turbidity corresponding to a Klett reading of 333 has twice as many cells/ml. as one having a Klett reading of 206. For intermediate values, corrections were made by changing the Klett reading to transmittance, calculating (-)log D or 0.6 + log (1*D) and getting the arbitrary concentration value which could be related to any other Klett reading, In order to relate these arbitrary cell concentrations to some definite value, several cell suspensions were evaporated to dryness. The data relating dry weight and Klett reading are found in Table VIII,

65 1.8 ii 1.4 *-1 \ 1.0 0` *o 0.6 0 \0 0 0.2 - 0.2-6 0 1.0. -02 0 \0 -1.4 0 2 4 6 8 10 12 Logarithm of the Multiple of Dilution to Base 2 Figure 13. Relationship between transmittance (D) or its metameters and the logarithm of the multiple of dilution of a suspension of Pseudomonas ovalis with glucose-phosphate medium.

66 TABLE VIII RELATIONSHIP BETWEEN TURBIDITY AND DRY CELL WEIGHTS FOR Pseudoaona. s ovalif Klett Reading 97 100 206 206 286 317 Dry Weight g/liter 0,22 0.23 0.53 0.57 0.78 0.83 406 1.54

67 Numbers of viable cells were also measured as a function of turbidity. At Klett readings of 159 and 233, the number of viable cells/ml. was 1.6 x 107 and 3,5 x 107 respectively. B. Effect of pH, Tempratre and 0oncentration of Cells on the _Rte of Production of Gluconic Acid b Pseudomonas ovalis 1, Effect of H Two series of runs were made. The first series used cells still in their growth medium while, in the second, the cells were centrifuged from the growth medium and then resuspended in the nitrogen-free glucose-phosphate medium previously described. Two liters of media inoculated 24 hours previously and incubated on the shaker at 30~0 were placed in the fermenter with 50 grams glucose and 0.025 cc of the antifoam agent. The suspension was agitated at 300 RPM and sparged with air at a rate of 1.16 VVMo To determine the cell concentration, 10,0 ml, of the suspension were centrifuged in an International Clinical Centrifuge, model CL, at approximately 4000 RPM for 15 minutes, The clear supernatant liquid was discarded, and 100O ml, of distilled water were used to resuspend the cells, The turbidity of the suspension was determined with a Klett colorimetero Several runs were made with the pH being held at a constant value in each run. Initially, the pH of the medium

68 was about 4,3 due to acid production during growth of the cellso Sodium hydroxide, 1.0 N, was used to raise the pH to the desired value for the run, During each run, the pH was kept constant, Determinations of the rate of acid production were made at pH values ranging from 5.5 to 100, Concentration of the base used for control of the pH was varied from Ool to loO normalo The choice of concentration was governed by the fact that dilute alkali increased the precision of the determination of the rate of acid productions, However too dilute a solution altered the volume of the liquid in the fermentor beyond an acceptable valueo It was necessary to apply two corrections to the data; one accounted for variations in the cell concentration, while the other corrected for the effect of undissociated acid, The Kurokawa plot (Figure 13) relating turbidity and cell concentration was used or the f th irst correction, The second, and minor correction was developed for weak acids in solution as follows: in solutions weak acids ionize to give: HA 9 H + A (12) for which an ionization equilibrium constant can be written 1H+] [A-] a [HA [ (13) Taking derivatives at a constant pH9 d [[H d[Ai] 4) t p ~K ~d ~~t ].(14) dt K dt a

69 From equation 14, one can see that the rate of production of dissociated acid is proportional to the total acid production. The rate of production of gluconic acid, both dissociated and undissociated, can be calculated when both the pH and ionization constants are known. For gluconic acid, the correction becomes only important below a pH of 6. For a sample calculation of these corrections, see Appendix D. The data for variation in acid rate with pH are plotted in Figure 14. The maximum rate of acid production was 14.5 meq/(l)(hro) at a pH of 7.15. The Michaelis-Menten equation can be used to examine the data (10): Am - 1H+J ([] l+ 2 (15) RA K [Ho [H+ I Here, RA is the velocity of acid production at a hydrogen ion concentration of [H+] p and R. is the maximum acid production rate at a hydrogen ion concentration of [Ho]. Using the optimum values given above for v and [Ho, K was found to be 15. The resulting equation represents the data with an average arithmetical deviation of 3.1% over the pH range from 5.75 to 8.0, Suspended cells were used in a similar series. In this case, the initial pH was established by adding either 1.0 N hydrochloric acid or 1,0 N sodium hydroxide. The air rate, agitation speed and temperature were the same as those used for the cells when they were prepared in their growth medium. Turbidity of the suspended cells, which was usually within 10 Klett units of 100, was measured before the pH was adjusted.

70 14 - O 0 o 0 P ro 0 0 0 OH a o <n 0 PC9 12 10 8 6 4 2 0 5 6 7 8 9 10 pH Effect of pH on the rate of production of gluconic acid by Pseudomonas ovalis in a growth medium at 25~C an KIett readig 300. Figure 14.

71 These data, shown graphically in Figure 15, may also be represented by the Michaelis-Menten equation (15). For a maximum rate of acid production of 2,85 meq/(l)(hr.) at a pH of 7,35, K is 18o4. The Michaelis-Menten equation represents the data with an average arithmetical deviation of 8.5% over the range of pH values from 5.85 to 8.5. When the two plots for the rate of acid production are examined9 it is seen that cells resuspended in a nitrogenfree medium become inactive at pH 5 and 9.7, while cells in their original growth medium will still produce acid. For suspended cells, the K value in Equation 11 is larger. The larger value of K1 indicates a greater effect of pH on the reaction. This is reasonable, since the cells in their growth medium would be better protected by components of the medium such as yeast extract which acts as a buffer. 2, Effect of TemPerature To study the effect of temperature on the rate of acid production, Pseudomonas ovalis cells were suspended in the nitrogen-free medium described previously. The rate of agitation was 300 RPM and the liquid was sparged with air at the rate of 1,16 VVMo The cell concentration was measured with the Klett photo electric colorimeter, Usually, the reading on this instrument for a sample of the liquid from the ferment or was close to 100o To account for any deviation in concentration from this figure, a correction was applied to the rate of gluconic acid production of the cells by multiplying

' 001 BUTpeea qa.TI pUe8 On,,g a mnTlpeu ee iJ-ueSo4.:u e UT stTBAO ssuomopnesJ &q PTo oTTuoonTS jo uotqonpoJd jo eqeaJ eq4 uo Hd jo qoejjz- *0'5 inST Hd 01 6 9 L 9 g 0 ct 0 17 Q 0\~ ~ ~ ~ ~ F H'-d \ / <: \ / p<~~~~~~ \ / ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~d \ s/ s~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~-, \. ^ <p ^~~~ ~

73 it by 100/Klett reading. Runs were made at temperatures between 90C and 400C, with the rate of acid production being measured (Figure 16). The data can be correlated using the Arrhenius equation, In Figure 17, the activation energy was found to be 9,600 cal/msle while the Ql values for these runs are reported in Table X and range between 1,7 and 1,8, These data agree fairly well with those reported in the literature~ Sizer (76) stated that Ql values for enzymes vary from 1,3 to 3o5 while Sumner and Somers (88) reported that the Qlo values vary between 1,1 and 2.0, Sultzer (87) reported activation energy values for the oxidation of saturated fatty acids by Pseudomonas species and Pseudomonas aeniclata ranging from 6,870 to 14,666 cal/mole, Ingraham and Bailey (41) studied the effect of temperature on the rate of oxidation of glucose by Pseudomonas species and Pseudomonas perolens, They found values of 119600 and 9,040 cal/mole for the respective activation energies, Kempe et al (44) also used a controlled pH to study the rate of acid production by Lactobacillus delbrueckii, From steady-state operation at various temperatures, they calculated values of 17,000 cal/mole for the activation energy in the conversion of glucose to lactic acid 3. Effect of Cell oncentration The effect of cell concentration on the rate of production of gluconic acid was investigated over a 30-fold

74 TABLE IX EFFECT OF TEMPERATURE ON THE RATE OF GLUOONIC ACID PRODUCTION BY Pseedomnas valis IN A NITROGEN-FREE MEDIUM AT pH 7.09 WITH AN AGITATOR SPEED OF 300 RPM9 AN AIR SPARGING RATE OF l1 6 VVM AND A KLETT READING OF 100 Run Number -T 10 T 7 T 8 T 6 T 5 T 1 T 6 T 3 T 4 Temp Erature 9o0 14.0 19o5 24,9 300 3000 34 8 34 3 40 o Acid Production Rate mea/ (1) hr) o1.09 1067 1,92 2 51 3069 4.20 5.20 4.00 2.82

75 TABLE X Qlo VALUES FOR THE RATE OF PRODUCTION OF GLUCONIC ACID BY RESTING CELLS OF Pseudomonas ovalis AT pH 7.0 WITH AN AGITATOR SPEED OF 300 RPM AND AN AIR SPARGING RATE OF 1.16 VVM Temperature Range Qlo _____OC _____ __ 10 - 20 1.80 15 - 25 176 20 - 30 1.73 25 a 35 1.70

76 6 5 Or-I H 3 0 0 1 ro 2 01 H III I 0 10 20 30 40 50 Temperature (OC) Figure 16. Effect of temperature on the rate of gluconic acid production by Pseudomonas ovalis in a nitrogen-free medium.

77 0.5 0.4 r-4) 0 4-, 0 0 o o <: o r, 0 0 H 0 0 0.3 0.2 0.0 Figure 17. 3.20 3.30 3.40 3.50 360 jl/Temperature (OKj x 10 3 Arrhenius plot for the effect of temperature on the rate of gluconic acid production by Pseudomonas ovalis in a nitrogen-free medium.

78 range of concentrations. At a constant stirring speed and rate of air flow, acid production of Pseudomonas ovalis cells was proportional to cell concentration over a very wide portion of this range, The scale on the abscissa on Figure 18 is the concentration of cells in number of viable cells per ml,, obtained from turbidity values using a Klett-Summerson coLorimeter and plate counts PFrom this plot, it can be seen that cell concentration increased as the oxygen concentration decreased, or vice versa depending upon which was designated as the independent variable. The values of the rate of gluconic acid production began to deviate from the straight line when the oxygen concentration in the medium fell to 40% saturation.. Direct Use of the Oxylen Electrode 1. Measurement of the Critical Oxygen oncentration All respiration, according to Finn (32), proceeds at a rate which is independent of the dissolved oxygen concentration so long as the latter remains above a critical value, Above this critical oxygen concentration (Cc), the rate of oxygen uptake by cells is unaffected by an increase or a decrease in dissolved oxygen concentration, Most measurements of the critical oxygen concentration are made by withdrawing samples from the fermenting medium and placing them in a polarographic cell. As the cells respire, the oxygen content of the liquid falls, The rate of oxygen utilization is constant until the oxygen concen

H 1100 ''0 -P )14 A 80 I ~,12 - \ 0g - 60. - lO - ~'10 s' ro -p 0 - 40;48 A ~ o Q) OH 0 O 6 A 20 04 Ir - ^ Io I $o 2 -Oa. ' ): r'2 0n "'0 2 4 6 8 10 Cell Concentration (cells/ml) x 10i7 Figure 18. Effect of cell concentration on the rate of gluconic acid production and on the equilibrium oxygen concentration of the liquid during the fermentation of glucose to gluconic acid by resting cells of Pseudomonas ovalis.

80 tration reaches a very low value (3, 32). In this work, a direct measurement was made without sampling, since the electrode was immersed in the medium which was being aerated and agitated, The value of 0% was obtained by stopping the air flow to the fermentor while still continuing to stir, The value of dissolved oxygen concentration9 measured by the electrode9 was recorded as a function of time, In a system in which microorganisms are respiring, the level of the dissolved oxygen concentration is a balance between the rate of removal f o xygen by the cells and the rate of supply. The supply comes from two sources: (1) the gas sparged below the surface, (2) oxygen dissolved through the surfaceo When the air supply to the sparger was turned off, the major supply of oxygen stopped, As a result, the dissolved oxygen concentration fello Unless the agitation rate was in excess of 400 RPM9 there was insufficient oxygen diffusing through the surface to supply the needs of the cells, and the dissolved oxygen concentration fell to zero, Figure 19 shows the change in dissolved oxygen concentration with timeo The straight line indicates a constant rate of cell respiration, The slope of this line, the rate of change of oxygen concentration with time (dO/dt)9 is the apparent rate of oxygen uptake by the cells, and it is constant, The oxygen cone@ntration at which this plot deviates from a straight line is the critical oxygen concen

81 1400 1200 4-4 01000 - ' H rEl l 800 ro 0 600 r-. 400 200 0 4 8 12 16 Time (minutes) Figure 19. Measurement of the rate of oxygen utilization by resting cells of Pseudomonas ovalis using the oxygen probe with agitation at 114 RPM and no aeration.

82 trationo Most previous workers used polarographic cells to measure the rate of change of oxygen concentrationo In these cells, the surface of the liquid was covered or the surface area was small when compared to the volume, In both cases, the amount of oxygen dissolving into the liquid sample was probably smallo With the method of measurement of oxygen uptake described here, the amount of oxygen entering the liquid through the surface was significanto Moreover, this source of oxygen became more important as the turbulence of the surface increased due to greater agitationo In the present study, an attempt was made to blanket the surface with nitrogen to eliminate this surface supplyo It was impossible to quickly eliminate all the oxygen from the head space of the fermentor by flushing the space over the liquid with nitrogeno Also, when pure nitrogen was involved9 oxygen passed from the liquid into the gas by counter-diffusiono Thus, a correetion factor was applied to obtain the true oxygen uptake by the bacteriao This factor became more important as the stirring speed was increased and as theconcentration of oxygen in the liquid decreasedi at 114 RPM9 the correction factor was very smallo The slope (dE/dt) of the curve in Figure 19 was plotted against oxygen concentration in Figure 20o The rate of oxygen uptake (dE/dt) was initially represented by a horizontal line which fell off rapidly to zero as the

83.r 120.Il I 4 -H 0 H 100 X 0600 d 0/0 0 0 0 0 C) 4o 0 CD O 20 r) 0 25 50 75 Dissolved Oxygen Concentration (% saturation) Figure 20. Effect of dissolved oxygen concentration on the rate of oxygen utilization by resting cells of Pseudomonas ovalis with no aeration at pH 7.0 and at pH 7.0 and ^CI. --

84 dissolved oxygen concentration approached zero, The point at which the oxygen utilization rate became dependent on oxygen concentration was the critical value9 OCo 0 was determined in 15 separate runs; values were measured a total of 20 timeso Stirring speed was kept constant for each determination; speeds employed were 114, 217 and 300 RPMo These data were analyzed statistically, and the critical oxygen concentration was found to be 11o 0,2 mg/l It was not dependent on stirring speedo For comparison with these data9 several values of Oc are listed in the introduction, They are lower than the value of 1,1 mg/lo reported here, 0c, however, can vary widely for the same microorganismo Finn (14) has suggested that discrepancies may arise from differences in methods of measurement, Winzler (100) reported that the presence of a substrate such as glucose increases the critical level of dissolved oxygeno His data were restricted to one organismo 2, Measurement the Rate of Qxf en Utilization 1 Pseudomonas ovalis The slope of the u9ve dE/dt, with appropriate corrections for transfer of oxygen through the surface, is a measure of the rate at which the microorganisms consume oxygeno Figure 19 represents oxygen uptake by Pseudomonas ovalis in an agitated but unsparged9 systemo Measurements of the two rates of oxygen utilization may be compared, The first rate was obtained by the continuous titration of the gluconic acid9 while the second was calculated

85 from the oxygen electrode measurement. The rate of oxygen utilization was first calculated from the data on the rate of production of gluconic acid: NB (dVB/dt) 1 millimoles oxygen HRG A x 2 liter-hour (16) The factor of one-half is necessary because the conversion of glucose to gluconic acid requires one-half mole of oxygen per mole of glucose, Oxygen consumption for the cells without aeration of the liquid medium can be calculated from the electrode measurements: (dE/dt) C0 millimoles oxygen R H (1/60)(32)E B) liter-hoUr (17) In the above expression, E is the measure of oxygen concentration expressed in millivolts. It is the direct reading from the pH meter, The difference in the reading on the millivolt scale of the pH meter between two oxygen electrodes placed in separate glucose-phosphate solutions of the same composition and temperature, one solution being saturated with oxygen from the air and the other being devoid of oxygen, was (E E) The saturation concentration of dissolved oxygen in distilled water at 2500 has been reported to be 8018 mg/l. (17)o The medium used in this work contained glucose, sodium mono- and di- hydrogen phosphate, sodium gluconate,

86 antifoam and bacteriaO The amount of oxygen needed to saturate the liquid was reduced by the presence of these soluteSo Solomons (78) has reported data for variations in dissolved oxygen coneentration. in water with various concentrations of glueoseo These data have been used to prediot the 6o8 mg/lo used as the saturation concentration for the medium in the worko 30 3Efect of 0xjgn 0encentration on the Rate of Gluconic o d Production Preliminary experiments indicated that the rate of acid production might be affected by the concentration of dissolved oxyge n in the medium as well as by the rate of air flow and stirring speed,, even at values above 0O This phenomenon was investigated in a separate series of runso It was shown that the acid production rate could be raised if agitation speed were increased from 114 to 400 RPM; changes in the air supply rate were unimportant0 However, the reverse of this change could not be accomplished9 i oeo lowering the stirring speed did not immediately reduce the rate of acid productiono Whether the rate would eventually have dropped was net establishedo The data recorded in runs D-3 to D-7 (Table XI) illustrate this effecto As the oxygen concentration was raised by changing the stirring speed to 400 from 114 RPM, the rate of and production increased, However, lowering the speed did not decrease the gluatoni acid production rate within a reasoneble length of timeo

87 TABLE XI EPFEOT OF VARIATIONS IN THE AGITATION i SPXED ON THE RATE OP GLUC0NI0 ACID PRODUCTION BY RESTING CELLS OF Pse udona: ovalis AT pH 7,0 A 250Q udov Run Turbidity of.ell Suspensions Klett Air.Flow Rate VVWM Produetion Rate of Glueonic Acid at Low and High Agitation Speeds meq/ (1)(hro) 114RPM 400RPM 114RPM 400RPM D=3 D-4 D-5 D-6 DB7 90 66 78 140 100 - 075 1,77 3,19 2~35 o177 2.35 1.27 0,88,083 3044 1,89 1,07 o1,04 4077 2,94 1o65 1o05 1,27 4,36 3 o 01 4,36 3,17 2,59 2064 D-8 D=9 99 98 1,89 2,89 2,39 2.14 2001

88 In an attempt to obtain this decrease in rate of acid production between the high and the low speeds, the following experiment was attemptedg the fermenting medium was aerated and agitated at 400 RPM? the air was shut off for 30 minutes the agitator was set at 114 RPMo As a result, the oxygen concentration fell to zero During this period, the only oxygen supply to the liquid was by diffusion from air contacting the upper surface, Shutting off the air to the sparger, in this experiment9 allowed the initial rate of gluoenio acid production to be re-established (runs DQ8 and D-9)o In another run, D:109 several levels of dissolved oxygen were attained by altering both the stirring speed and the gas rateo A plot of rate of glueonic acid formation against the dissolved oxygen concentration is ihown for this experiment in Figure 21o At the end of this- run the "reducing technique" just described was tried and a point was inserted on the plot at an acid -preduction rate of lo92 meq/(1)(hro) o It coincided with the curve drawn through the other data taken as the dissolved oxygen concentration was progressively increased o This last run showed a direct relati on between aoid production-and the concentration of adissolvd oxygeno The alteration in oxygen concentration in the run was effected by changing the stirring speed and the rate of air supply to the cell suspensiono To remove these two variables, bthre r uns, D12 1 Dm13 and D 14, were conductedo In the first

89 3 I I I H 4r 0 -0 -> C) o 0 oH 0 C) 0 0 P4 0 QH <D 2 H 1 I I I 0 20 40 60 80 Dissolved Oxygen Concentration (% saturation) Figure 21. Effect of oxygen concentfration on the rate of gluconic acid production (speed and air flow rate varied) by resting cells of Pseudomonas ovalis at pH 7.0 and 25~C.

90 two, the stirring speed was maintained at 300 RPM and the air was supplied to the fermentor at a rate of 2o0 liters per minuteo However, the oxygen concentration in the gas stream was altered by mixing various ratios of air and nitrogeno The last run was similar to thefirst two, but the speed was increased to 400 RPMo The rate of gluconic acid production at both 300 and 400 RPM increased with increasing oxygen concentration (Figure 22) until the concentration reached approximately 40% of saturation~o The rate of glueonic acid production did not increase further with increases in dissolved oxygen concentrationo It should be noted that the rate of acid production at 400 RPM exceeded the rate at 300 RPMo Purthermore. the rate of acid production continued to rise with an increase in dissolved oxygen concentration, although the dissolved oxygen concentration was well above the criticalo This last phenomenon has also been reported by Steel and Maxon (83)o: The slope of the curve showing the rate of gluconic acid formation at 400 RPM is steeper than the curve for 300. RPMo Another run was made to determine the independent effect of stirring on the rate of gluconic acid productiono In this case- the air sparging rate was held constant at 0046 VVM, while the stirring speed was increased, in steps9 from 114 to 570 RPMo Both the acid production rate and the oxygen concentration in the medium were measured during the fermenn tation; at the end of the run the rate of oxygen utilization

TABLE XII EFFECT F D OF DISSOLVED OXYOEN CONCENTRATION ON THE RATE OF PRODUCTION OF GLUCONIC ACID BY RESTING CELLS OF Pseudomonas ovalis AT pH 7.0 AND 25~C Run Turbidity" Of Cell Suspension Klett Reading D-10 104 Agitation Speed RPM, I1.4 157 217 300 114 300 D-12 Rate' or Air Flow 1.00 1.77 2,35 i 77 0.23 2.00 1.00 1.00 1 00 Rate of Gluconic Acid Production aeq/(l)(hr,__ 1'14 1053 2.16 2.37 2.49 1o92 1.48 1.87 2.17 2.22 2.26 1.00 1.87 2.34 2.61 2.52 2.52 2.52 0.11 0.88 1.57 1.90 2 13 2.24 Concentration of Di s olve d. Ox-.en, % Saturat ion 7 19 45 57 73 34 13 29 45 60 81 4 14 31 43 62 73 93 6 15 29 50 63 82 H D-14 114 400 300 D-13 107

92 3 0 FO 0 0 0 o 0 4) cz c) 0 p-4 4, m~ 2 1 0 20 40 60 80 100 Dissolved Oxygen Concentration (% saturation) Figure 22. Effect of oxygen concentration on the rate of gluconic acid production by resting cells of Pseudomonas ovalis at 25.C, a constant air rate of 1.o00o-PV nd a pH of 7.0.

93 was determined using the oxygen probeo It can be seen in Figure 23 that as the stirring speed was increased9 both the concentration of oxygen and the rate of glueonic acid production increasedo The lower, horizontal line represents the rate of oxygen uptake as measured by the oxygen probeo It is about 55% of the maximum oxygen transfer rate determined by measurement of the rate of acid production at 570 RPMo Finally9 the air flow rate to the medium containing suspended cells was varied to observe the effect on the dissolved oxygen concentration in a fermenting medium, Dissolved oxygen concentrations were measured in the medium at two agitation speeds, These data are recorded in. Table XIT and are plotted in Figure 24, The results indicate that below 217 RPMs the air rate must be very high to maintain reasonable levels of dissolved oxygen in the liquid~ Oonm versely9 above 217 RPM only very low air flow rates are needed to supply the needs of resting cells of this concentrationo D, Effect of Viscosity 1o Sulfite System Sucrose and glycerol, which develop Newtonian solutions in water (78)9 were used to conduct the initial experiments with increased viscosityo However, in order to attain significant increases in viscosity, large concentrations of these chemicals were neededo

94 TABLE XIII EFFECT OF AGITATION SPEED ON THE RATE OF OXYGEN UTILIZATION AND ON THE DISSOLVED OXYGEN CONCENTRATION: RESTING CELLS OF Pseudomonas ovalis,, IN A GLUCOSE MEDIUM AT 25,09 pH 7o0 AND AIR FLOW RATE 0,46 VWM IN RUN D-15 Agitation Speed 114 204 300 570 Rate of Oxygen Utilization (mmoles/(1) (hr.) 0.o56 (1) 1o39 (1) o163 (1) 1,82 (1) Concentration of Dissolved Oxygen ~ Saturation 6 79 95 100 114 0.99 (2) deorepsing (1) Calculated from the rate of gluconic acid production (2) Calculated from the oxygen electrode measurements

95 2.0 Ft 0.r4 O Hr Co1 0 0 0 r-I " S El 4 ~r4 o a 0 1.5 1.0 rate measured / by gluconic aci / method / rate measured with electrode ~ / at 114 RPM Data from run D-15 I I 0.5 0 200 400 600 Stirring Speed (RPM) 0,-I 0.r-I O (d k N, r-t — CO 0 -W0 0 4r O 0 0 H 0 1.5 1.0 I I rate measured with electrode I I 0.5 0 100 200 300 Stirring Speed (RPM) Figure 23. Effect of agitation speed on the rate of oxygen utilization by resting cells of Pseudomonas ovalis in a glucose medium at 25~0 and pH 7.0. Rates determined with aeration (gluconic acid) and without aeration (electrode).

96 TABLE XIV EFFECT OF AIR FLOW RATE AND AGITATOR SPEED ON THE DISSOLVED OXYGEN CONCENTRATION IN A GLUCOSE MEDIUM CONTAINING RESTING CELLS OF Pseudomonas ovalis AT 2500, pH 7.0 AND KLETT READING 100 Rate of Air Flow Dissolved Oxygen Ooncentration, % Saturation 4.6 4,3 4.1 3.5 3,0 2.4 1 8 1o3 0.8 0,5 0.25 217 RPM 75 73 73 71 68 63 58 50 114 RPM 70 56 48 36 25 22 8 41 30 6

97 75 O 4 -50 - 0 4-, 40 0 I3 /< / ~~A 114 RPM IQ (I / 0 217 RPM 0 1 2 3 4 5 Air Flow Rate (VVM) Figure 24. Effect of stirring speed and air flow rate on the dissolved oxygen concentration in a suspension of resting cells of Pseudomonas ovalis at 250C, pH 7.0 and Klett reading-i0.-

98 Concurrent with the increase in viscosity was a large increase in osmotic pressure. If the osmotic pressure in the viscous medium was higher than in the cells, tbee microorganisms were dehydratedo It was, therefore, not possible to use Pseudomonas ovalis with high concentrations of glycerol and sucrose; thus, the chemical system in which sulfite.is oxidized to sulfate-was chosen, The data for the effect of viscosity on the rate of oxygen transfer are recorded in Table XV and are plotted in Figure 25~. Both glycerol and sucrose significantly reduced the rate of oxygen transfer, With sucrose solutions, reductions of 93% in the rate of oxygen transfer were ac-, complished by increasing the viscosity of the solution from 1 to 12 Cp, Using glycerol, the OTR was decreased by 72% when the viscosity was raised from 1 to 4,75 cpO These data are in agreement with published results6 Solomons (78) found a 45% solution of sucrose, equivalent to a viscosity of 10 cent.-poises, reduced the OTR by a factor of 6 (84%) as measured by the sulfite method and 12 (92%) by the polarographic method, Yoshida et alo (101) found a 60% reduction in the OTR when the viscosity was increased to 306 centipoises with glycerolo Plotted on logarithmic scales (Figure 26), the data for OTR against viscosity yielded straight lines for both substanceso Slopes of these lines are (-) loll for sugar and (-) 0097 for glycerol, which means that the OTR varied inversely as the viscosity raiaed to the power 1.11 or 0:,97 dependiMg t. he chemical added,

99 TABLE XV EPFEOT OF VISOOSITY ON THE RATE OF OXYGEN TRATS PER TO SPARGED, AGITATED SOLUTIONS OF SODIUM SULfI AIR PLOW RATE 1.0 WM, AGITATION SPEED 300 RPM, TEMPERATURE 25~C0 Run Number V13 vi. V14 V17 V16 V18 V15 V19 V9 Viseosity 4,78 3,07 2034 lo77 10o36 llo75 3036... 2057 1.061 1o15 0o89 Amount of Vis8osity Additive in 3 Liters of Solution glyoerlol155. mlo glyoerol-1200 alo glycero-1000 rol glyeeroel750 mlo glyeerol 500 mio suerose-1990 go buorose=1285 go buorose-1075 go sucrose-1075 go Sucrose=380 g.,none= Oxidation Rate eg/)(hro)f ) 27.5 42.0 55.8 50.8 88.9 151

100 160 0 Glycerin 120 Sucrose 4-\ 4 80,d 0 0 ed Os 4-. 05 T11 4012 0 2 4 6 8 10 12 ViscositY (centipoise) Figure 25. Effect of increasing viscosity, due to the addition of glycerin or sucrose on the rate of odim sulfte oxidtion. The liquid was sparged with air at a rate of 1.0 VVM and agitated at 300 RPM at 250C.

101 %rl w-4 0 H Cl) 0 0 0 -, oH 0 a 0 *H 0 43 OH Pf 200 I I I I I I 100 50 40 30 W 20 - ON 0 Glycerin A Sucrose 10 - 51 07 I I I I I I 1 2 345 10 20 Viscosity (centipoise) Figure 26. Logarithmic plot showing the effect of increasing viscosity, due to the addition of glycerin or sucrose, on the rate of sodium sulfite oxidation. The liquid was sparged with air at a rate of 1.0 VVM and agitated at 300 RPM at 25~C.

102 It is possible to estimate the effect of decreased oxygen solubility due to increasing.kgar concentrations on the rate of mass trkasfero Solomons (78) has published data for the equilibrium solubility of the oxygen from the air in glucose at 37a0o If it is assumed that oxygen concentrations would be equ@ialent in solutions of glucose and sucrose of the same molarity9 then a correction can be made for oxygen saturation reductionso Although these data are for 3700, the percent reduction in solubility is the same for all oxygen and sugar concentrations over a given range (81), Thuss these figures indicate that some mechanism other than solu= bility is importanto 2. Production Mf Glugoni Aci in noniewtonian Brronths Small concentrations of ether celluloses impart non= Newtonian characteristics to aqueous solutionso Two su@h commercial preparations were used in this studyo (1) NatrBsol Ho a non-ionics hydroxyethyl ether of cellulose (2) Methotel 65 HG9 4000 OP; a propylene glycol ether of cellulose Initial experiments showed the necessity for increasing beth the stirring speed to 570 RPM and air flow rate to 1,85 M in order to maintain the dissolved oxygen (DOo.) concentration at a level above the critical so that it did not influence the rate of respiration. In addition to increasing the air flow and stirring speed, the second impeller was lowered so that it, toos was immersed in the

103 TABLE XVI SULFITE OXIDATION RATE IN SUOROSE SOLUTIONS CORRECTED FOR REDUCED SOLUBILITY OF OXYGEN IN EQUILIBRIUM WITH AIR Run Number V17 V16 V18 V15 V19 V9 Suorose Concentration 1,o 94 1,25 1005 0.~73 0037 0.00 Equilibrium Oxygen c one entrati on pB. 3l7o 285 3,8 4.2 4~7 5.2 60.9 Corrected Oxidation Rate meg/(l) (h.) (p) 2,8 6,9 8.7 11 17 22

104 liquid to improve gas dispersion and increase turbulence. Several runs were made in which different concentrations of chemicals were used to alter viscosityo During each run, the rate of production of glueonic aoid was measured, and the level of dissolved oxygen was determinedo When each run was terminated, a sample was taken for viscosity measurement, Thee flow curve for the medium used in the run was determined with a rotational viscemetero The recorder plotted shear rate against shear stresso An example curve as plotted by the machine is shown in Figure 270 Logarithmic plots of shear stress against shear rate were linear for all concentrations of Natrosol and Methocel, The alope of the straight line wai eo A curve is shown for each compound in Figure 28, These data do not agree with the published works of Godleski and Smith (39)0 who reported that Natrosol was a pseudoplastic fluid that did not follow the simple power law equation Their flew curves were established with a Brookfield, synehoelectric viscometer, Metzner and Otto (52) employed Natrosol in concentrations of 1,0% and 2o0%o The 2o0% solution gave a straight line when the logarithm of shear stress was plotted against shear rateo Although there was some deviation for the 1% solution from the straight line9 that deviation9 according to the authors was less t:h the scattering of the data pointso Acoordingly values of 0,32 and 0 40 were found for the exponent no

20 18 16 02 P4 03 -p 0 aC C) 14 12 10 8 6 4 2 0 2 4 6 8 10 12 Shear Stress (arbitrary units) Figure 27. Flow curve for a medium containing 15 grams of Natrosol per liter. The plot of shear rate versus shear stress was determined with a rotational viscometer.

106 20 10 4C - 4, 0 0 0 43 CO 0 CO) aO CO 5 1.0 0.5 0.2 1 Shear Rate (arbitrary units) Figure 28. Relationship between shear stress and shear rate for Natrosol and Methocel.

107 TABLE XVII VALUES OF THE EXPONENT n IN THE POWER LAW EQUATION FOR SEVEAL JOOEN0TRATIONS OF NATRODSOL AND METHOOEL Run Number Ell E12a E12b E12c Oencentration of Viseosity Agent-,gr-ams/lite 5 (Natrosol) 15 11 " Power Law.Exponent 0.57 0~286 0.392 0,466 8 5 ii'f E14 E13 Ml M2 M3 M4 M5 10 (Methocel) 20 w 5 o.637 0.426 0.968 0.885 0.721 0.616 0.458 7.5 10 15 Q19

108 Oalderbank and Moo-Young (11) have also reported on work with earboxymethyl cellulose For several. different concentrations of that substance, they reported values of I: ranging from 0o44 to,0649 indicating a linear plot of the logarithm of shear stress versus the logarithm of shear rate. The power-law exponent, A^ which is the slope of the logarithmic curve of shear stress versus slhear rate, is a measure of the magnitude deviation from Newtonian solutions, As the solution' became mere concentrated, the viscosity rose and a decreasedo A plot' of versus concentration of the vk:iT.: 'ty agent (Figure 27), yielded straight-lines for both Natrosol and Methooelo Since the apparent viscosity of non-Newtonian fluids depends on the shear rate, the data were extrapolated to a standard of one seo p1 as suggested by Solomons and Perkin Y a.. (79 ) The data are reported in. Table XVIIIo These viscosity agents were then studied as components of the medium in the fermentero As the viscosity increased, the turbulence of the liquid decreased, and the efficiency of gas distribution fell off markedlyo Large bubbles formed and rose to' the surface around the impellero There was very little movement of the liquid near the baffleso The oxygen electrode was installed behind one of the baffles; the values of the dissolved oxygen ooneentratieo indicated that the suspension, even at that point, was supplied with sufficient oxygen o It was noted that as the broth became more viscous, the bubbles rose much more slowlyo

109 01 )-I Cl 0 H 0 OH Ca r4 1.0 0.8 0.6 0.4 0.2 0 5 10 15 20 Concentration (g/l) Figure 29. Relationship between the concentration of Natrosol and Methocel added to water to increase the viscosity and n, the slope of the logarithmic plot of the flow curve for that medium.

Run Number E10 E4 E8 Ell E7 E9 E12 E13 E14 TABLE XVIII PRODUCTION OF GLUCONIC ACID BY Pseudomonas ovalis IN SOLUTIONS HAVING PSEUDOPLASTIC VISCOSITYO AGITATION RATE 570 RPM, AIR RATE 1,85 VVM, TEMPERATURE 250C Viscosity Apparent Viscosity Oxygen Rate of GI Agent at Shear Stress Concentration Acid Prodi of 1 sec,1 Measured at 30~C c @entipolse8 % Saturation meq/(l)s O 1 2.27 O 1 2o13 luconic action (hr __ 5 (Natrosol) 5(Natrosol) 15 (Natrosol) 15 (Natrosol) 15(Natrosol) 20 (Methocel) 10 (Methocel) 88 460 2.45 2,12 2,82 3,14 2.43 2,51 2,07 H 0 17 000 19, 000 738 39 73 94 2o44 Average Value

Ill This agrees with reasoning based on Stokes Law, which indicates that as the viscosity increases, bubble velocity decreases, and the contact time increaseso The rate of production of glueonie acid was not affected by viscosity The data are reported in Table XVIII for 9 separate runs; concentrations of Methocel as high as 2% produced an apparent viscosity of 19,000 op measured at a shear rate of one sece 1o The average rate of acid production for these runs was 2044 meq/(l)(hro) and all runs (except one) were within 20% of that figure 0 30 Effect of Viscosityn the Rate of Oxygen Transfer in a 0ell-free S stem The rate of oxygen transfer to a solution from the atmosphere has been expressed as: H = KL A (g 0- ) (18) If the oxygen concentrations 0% is measured with the, oxygen electrode9 and if it is assumed that ~ is proportional to the meter reading then 0 = m(EE BO ) (19) when m is a constant9 E3, is the zero reading on the meter when no oxygen is present in the liquid and E is the meter reading at concentration 0o.The following expression is obtained if Equation 19 is differentiated: dt dt(2)

112 Substituting the above expression in Equation 18, it is found that H 3 KLA (KES E) (21) The reaeration method for measurement of OTR first involved stripping all of the dissolved oxygen from the solutiono Then, as the oxygen was replenished by bubbling air through the medium or by diffusion of air through the surface of the liquid, the meter reading was recorded as a function of time, Figure 30 gives the results of such an experiment. Initially, the meter reading, which was proportional to the dissolved oxygen concentration, was zero. As absorption proceeded, the oxygen concentration increased and asymptoticalt ly approached the saturation value, ES. The rate of oxygen transfer, at any time, is the slope of the curve in Figure 30. The rate decreased as oxygen uptake progressedo Equation 21 can be integrated (42) using E = 0 at t = 0 and E E at t = t as limits, giving: KLA lO1 iE ( 22) L 1t lO $ ~ allowing the mass transfer coefficient KLA to be evaluated. Since the unit of KLA is the reciprocal time, it applies without alteration to both Equations 18 and 21, The values for the mass transfer coefficient have been measured for the transfer of oxygen to the glucose-phosphate medium used to resuspend the cells of Pseudomonas ovalis. The data

113 1400 -- 1200,P-. 0, 1000 800 0 600 0 ID 400 0 200 0 10 20 30 40 50 60 Time (minutes) Figure 30. Oxygen transfer by diffusion through the surface of a glucose-phosphate medium stirred at 300 RPM at 250C.

114 showing the effect of stirring speed on KL A are found in Table XIXo These values have been used to calculate the rate at which oxygen diffused through the surface from the atmosphere when the electrode was being used to measure oxygen transfer rateso The correction for diffusion may be calculated for any concentration of dissolved oxygen according to the equations,(\s' = KLA (2 3 E) (23) The net rate at which the cells consumed oxygen was the sum of the measured rate of removal plus the rate of oxygen 'diffusing through the surfaceg ~(a^OTAL (Hxs (24) An example calculation is shown in Appendix Eo To determine the effect of viscosity on the rate of oxygen transfers the oxygen electrode was employed to measure transfer rates in solutions containing different concen= trations of a high mPlOular weight viscosity additiveo As the concentration of Methcel was increased9 a dere ase in OTR occurred conuerrently with the increase in viscosity, From the data reto.rded in Table XX, it can be seen that the major reduction in OTR occurred when the Methocel concenr tration was increased from 5 to 10 grams per litero This oorresponds to the point at which the viscosity shows a significant.ha nge o

115 TABLE XIX EFFEQT OF AGITATION SPEED ON THE SS OEPFIOIENT POR OXYGEN.DIPFUSING THRROUGH THE SURPACE OP A GLUOOSE' SO3LUTIONo MEASUREMENTS MASE WITH THE OXYGEN ELECTRODE Agitation Speed Mass Transfer Coeffioient RPM (K A) x 100 min"1 114 1.07 217 1.88 ~30~0 5.03,78 412 16,6

116 TABLE XX EFFECT OF INCREASING CONCENTRATIONS OF METHOOEL ON THE OXYGEN TRANSFER RATE MEASURED BY THE OXYGEN PROBES AGITATION RATE, 300 RPM; AIR FLOW RATE TO SPARGER, 0,2 VVM, TEMPERATURE, 2500 Run Number 2-0,,2 2-D-2 >zCCC1 120B1~ 2wC2D2 2aD 2 2QD2B Concentration of Methocel 0 2,5 5,0 10 Mass Transfer Coeffioient mint0.55 0048 0.40 0092 Power Law Exponent n 1oO 0,968 0,885 0.616

117 The effect of the rate of air flow was determined by sparging air into distilled waters The data are recorded in Table XXI and the logarithmic plot, Figure 31, shows that for agitatio aat 300 RPM, the OTR varies at the 2~3 power of the air rateo

118 TIBLE XXI EPPEOT OF THE AIR FLOW R3TZ ON THE OXYGEN TRANSFER RATE MEASURED BY THE OXYGEN ELEOTRODE IN A FERMENTOR AGITATED AT 300 RPMo THE METHOOEL CONCENTRATION IN THE LIQUID WAS 2o5 g/10 Run Number 2-B-1 2,uB-aK4 2 B3L 8 ab3S 2B4~~ Air Rate 0005 0O2 0O75 lo25 Mass Transfer.oeffic ent 0031 o,51 0059 0.73

119 2.0 1.0 0.5 0.3 I I I I I 0 O @1= 4al 0 H frx *r-., 1 0 0.4 0.05h 0.03 - I 1 2 Mass Transfer I I I I 3 4 5 7 Coefficient (min'1) 10 x 10 Figure 31. Effect of the air flow rate on the oxygen transfer rate measured by the oxygen electrode in a fermentor agitated at 300 RPM. The Methocel concentration in the liquid was 2.5 g/1.

Ivo~, DISCU~-SSIONE lo Mechanism of Cell Adr Ron on Bubbles The supply of oxygen to microorganisms, its subsequent reaction within the cells and the later removal of the metabolic products is carried out by an integrated series of physical9 chemical and biochemical reactions, The resistances to oxygen transfer from sparged air to the respiring cells in the liquid are many9 and the mechanism of transfer is still not completely understoodo As an example of the controversy surrounding this subject, the multitude of articles about the seemingly simple oxidation of sodium sulfite to sodium sulfate can be cited; even in this cell-free system, the mechanism has not been clearly detailedo Aeration and agitation are coupledo Indeed, Solomons (78) has stated thato "The effects of agitation independent of aeration have not been studied a great deal9 since it is often difficult to devise experiments where it is possible to differentiate between. agitation and aerationo" Theprocess of agitation promotes the transfer-of oxygen through the surface and9 in fact, vortex aeration (6) has been used as a means of supplying oxygen to fermentations, Vortex aeration occurs if a turbine impeller is employed in a unbaffled system, The liquid moves in wide, cylindrical paths, but when the vortex touches the impeller, air is distributed into the liquid as small bubbles: the system becomes an efficient gas-liquid contactor at high agitation speeds~ 120

121 Similarly, aeration by sparged air provides agitation, since the passage of swarms of bubbles through the liquid aids in mixing besides replenishing the oxygen supplyo Many studies have been made to determine oxygen transfer rates in nonbiological systems9 both with and without the use of sparged air~ With respiring organisms, however, it has been very difficult to separate these two effectso Most studies of this type have relied upon removing a sample of the broth from the medium and measuring the rate of oxygen utilization in,a polarographic cello In this instance9 the microorganisms are in a quiescent liquid that is being neither aerated nor agitatedo Tsao (93) developed a method of measuring the oxygen transfer rates in an actual fermentation by using the Pseudomona oalis conversion of glucose t gluconic acido Since he knew that each mole of acid produced required onehalf mole- of oxygeng he was able to calculate the OTR, provided he measured the rate at which sodium hydroxide was supplied in order to keep the pH constanto He concluded that the OTR was affected by the stirring speed and that the major resistance to oxygen transfer lay at the cell-liquid interfaceo His conclusions were based upon the fact that the,OTR was proportional to oell.oncentration and, hence9 to the area of the. cell-liquid interfaceo He also noted that the OTR measured by the sulfite method was very much larger than that found by the Pseudomonas ovalis fermentation of glucoseo Thus he concluded that the

122 dissolved oxygen concentration in the liquid was high during his fermentations and that the cell-liquid resistance must control the OTRo The present research supported his data and provided further proof of his conclusions9 but the new data also suggested that an additional mechanism of oxygen transfer was operative in the overall transfer scheme, The oxygen electrode was used to measure dissolved oxygen (D, 0,) concentrations in the medium as the cells respiredo Indeed, as Tsao predicted, the Do 0o levels were high for most cell concentrations and stirring speeds, Only at agitation speeds below 200 RPM did the Do 0O concentration fall below the critical value0 The present research also confirmed Tsao s finding that the rate of acid production was proportional to the cell concentrationo Figure 18 illustrates this fact, but the data presented in this thesis cover a thirty-fold range of cell concentrations, compared to the seven-fold range that Tsao investigated. It will be noted, however, that the concentration of Do 0o fell below the critical value at rates of gluconio acid production well above the levels attained by.Tsaoo Both of these curves9 one for the rate of gluconic acid production and the other for Do 0o concentrationS leveled out at about the same cell concentration0 By shutting off the air supply9 the OTR was studied at various agitation rates without aerationo The oxygen probe was used to follow the fall of the D, 0O concentration

123 in the medium as the cells removed itt The slope of this curve is a measure of the rate of oxygen consumption by the ells o Since oxygen also entered the liquid by diffusion through the surface of the liquidp it was necessary to apply a correction faotor0 This method of measuring the OTR was limited to stirring speeds equal to or less than 300 RPM since9 above this speeds the reaeration rate by diffusion of oxygen through thhe surface of the liquid became excessively higho Above 300 RPM9 the dissolved oxygen concentration value fell very slowly, and it was necessary to oaloulate the utilizati n rate mainly from the correction factor for surface reaeratien- rathae than from the slope of the ourve of concentration vre'.s -1me. It was found that the-dissolved oxygen content of the broth fell at a rate whi h was independent of stirring speed for 1149 217 and 300 RPMo This is evidence for the view that the controlling mechanism is not restricted to the liquid film around the cello This conclusion is supported by FiP (31) and Aiba et a1o (1)9 who have independently calculated the resistance of the film to be insignificanto However9 it is important to recognize that the rate of producti n of glu oni acid has been shown, in the present study to rise witho (1) an increase in air rates (2) an increase in dissolved oxygen concentration in the liquid9 bPve the critical value, (3) an increase in agitation speedo

124 Also, it is important to observe that in run D-10, a single batch of cells was employed to investigate the combined effect of stirring speed and air flow rate on the rate of production of gluconic acido In the same manner, a single batch of cells was used in run D-15 to evaluate the effect of agitation rate alone on the rate of oxygen transfer, In this manner, differences in physiological characteristics between different batches of cells were eliminatedo It is to be noted in Figure 23, which contains the results of Run D5, that the OTR increased with each increase in speed, even at Do 0' concentrations well above the critical value. Steel and Maxon (83) have reported the same phenomenon in novobiocin fermentations0 Moreover, in the study reported here, the rate of oxygen uptake was less when measured with the probe than when calculated from the rate of acid productiono The transfer of oxygen from air bubbles to a cell wall encounters a number of resistances which have been described in detail previously. These resistances are found at the gas-liquid and the cell-liquid interfaces, as well as in the bulk or The liquido All of these restrict the availability of oxygen to the cellso Another concurrent path was suggested by Bartholemew et alo (3), out direct evidence to support their model could not be obtained due to limitations of the system and equipment available at that timeo They proposed that a shorter but parallel path for oxygen transfer existed when cells were

125 adsorbed a bblese, Direct contact between the two would combine the two liquid films into oneo This would eliminate the path through the bulk of the liquid as well as reduce the length of the diffusion path in the liquid films surrounding the bubbles and the ce lls The combination of films would result in a cell being exposed to high concentrations of oxygen across a shortened patho Such a system would be favored, according to the originators of the theory (3),9 'by a large population of small bubbles having a large interfacial area coupled with thorough mixing between bubbles and cellso.t It was noted in Tsaoes work (93) that increasing the stirring speed increased the gas holdupo In this research, it was found that increasing the air rate, or the stirring speed9 increased the OTRo Oxygen concentration was another variable whose effect on the rate of production of gluconic acid was measuredo High oxygen concentrations in the gas phase led to higher rates of oxygen transfer as shown in Figure 229 even though the concentration of dissolved oxygen in the liquid was well above e the riticlal vaue In this work, the OTR was measured by two methodso during aeration and agitation9 it was calculated from the rate that sodium hydroxide was added to maintain the pH at a constant values when the air flow was stopped, the rate of decrease of oxygen concentration was the measurement used to calculate the OTRo When the two rates were compared9 it was found that

126 in Run D-l59 the OTRS measured under conditions of no aeration, was 53% of the rate when the cells were being both aerated and agitated t ioeo the OTR was less when there was no direct bubble-oell contacto Aibas et alo (1)9 using the oxidation of glucose to glueonti acid as one method to measure OTR and a polarograph for the second measurement9 agree with this findingo They found the 0TR without aeration to be about 66% of the rate with aeration0 Siegell and Gaden (75) also observed a similar phenomenon in yeast Ifrmentations and stated that it was likely due to a change in the environment from active aeration and agitation in the fermentor to static conditions in the sample chamber of the polarographo Phillipl et al (59) have expresse dissatisfaction with the accepted theories and mechanisms that describe the effeocts of liquid behavitr on the rate of absorption of a gas by a liquido They decided that no single mechanism would explain their datao They proposed that oxygen transfer would result from no mal -iffusional processes together with the inctrpr ation Lt tho medium of an adsorbed layer of molecules o The theorits presented il this thesis also propose that no single mechanism is able to aooount for oxygen transfero The view of Bartholomew et alo (3) that oxygen transfer occurs across the ommin film around a 'cell and a bubble, can be used to represent the data in b th this work and that of the previous authorso Moreover, the "film" mechanism does

127 not rely upon getting more oxygen into solution in the bulk of the liquid9 but assumes direct transfer from the air bubble to,the solid cell adsorbed on the air bubble The film Joining the two act as a path for direct transfero Strohm and Dale (85) compared oxygen transfer in the oxygen electrode and in living cells, They believed that a high resistance to diffusion existed at the polyethylene surface covering the oxygen electrode and that a similar resistance was provided by the common water film between an air bubble and the surface of the cello Thus, the significant variable would be partial pressure of oxygen in the gas phase, rather than its concentration in the solutiono This presentation by Strohm and Dale (85) is consistent with the data reported hereO Partial pressure of oxygen in the gaas wa important to the rate ef gluconic acid production (Figure 22) when the cells were being aerated because there was a direct transfer of oxygen across the filmo When this shortened path was removed by ceasing aeration9 the OTR dropped since oxygen was drawn only from that dissolved in the liquid. In this oase9 the rate of oxygen removal from the water was independent of the agitation rates thus indicating that the major resistance lay at the cell membraneo It is well known that solids tend to concentrate at interfacial boundarieso For example, froth flotation has long been-used to oonoentrate metallic ores and to separate useful minerals from wrthless rook (35), Air bubbles adhere to mineral particles in water and bring them to the surface

128 as a froth. It is also well known that microorganisms adhere to bubbleso Gaudin et a10 (36) have presented data for froth flotation of Echerih i o Many ether organisms, such as Sta hlo 1.ous albus S hie s-c.haroye g (25 26)9,.ia marceseens (8) Bacillus subtilis varo nier spores (37 38) and Baclu cerceus var, t.erminais spores (5) also have been concentrated by froth flotationo An experiment was also performed in the present study with Ps$eudmonas ovalis suspended in a nt.trogen-free glueose phosphate medium9 which showed that they9 to9, eould be concentrated at the gasliquid interface by this meanso 2,. Effect of Visloos n th e rOxen Utake Rate Rheological properties of mold and actinomycete fermentatlon broths change -as the fermentation progresseso During growth9 yield stress ionreases with the aging of the culture9 and the meda t md end toward plastic behavior due to the nee inrease in the concentration of the organism (21)o Steel and Maxon (84) also illustrated the change in rheo logical properties of a fermentation medium as the mold grew; they reported9 hewever9 that the apparent viscosity reached a maximm' and then declinedo In studies of power requirements for gas-liquid eontaot rs9 the data have been correlated by the following equation Pg0 (i 2-o) (25),pXJ D

129 This shows that the power number is a function of the Reynolds numbero It is the same basic equation that is applied to Newtonian fluids but the value of the visaosity used to calculate the Reynolds number in non-Newtonian systems is that value measured at the appropriate shear rate for the impellero Oalderbank and Moo-Young (11) give the following equation for the oaloulation of the rate of the amount of shear produced by an impelier ina stirred vessel containing either a Bingham or a pseudoplastie fluid 10 (26) Richards (65) found 1000 to be the lower limit for turbulent flow regions in which the power required is directly proportional to the fluid density and is independent of viscosityo He also found 10 to be the upper limit of the Reynolds number for viscous flow9 although Godleski and Smith (39) used 40. for pseudoplastic fluidso Below this figure9 the pow.er input is directly proportional t the fluid density and independent of viscosityo In the studies with non-Newtonian broths., which have been rep rted here9 the Reynolds number varied from a high value of 33,000 for water9 to a low of 51 with the maximum concentration of Natresolo Detailed cal clations for these figures are found in Appendix G, In the latter case9 the Reynolds number was close to the boundary value for the viscous regions this was also indicated by visual observationo The air sparger emitted large bubbles which rose to the

130 surface and burst. There were, however, a large number of small bubbles immersed in the liquid; these rose very slowly. Samples of the viscous medium contained many small, finely dispersed bubbles, and these often persisted in the liquid for a period as long as onechalf houro This prolonged residence of small bubbles increased the volume of air held in the mediumo Similar results were reported by Steel and Maxon (84), who found an approximate correlation between apparent viscosity and gas retention etraentage. The shear rate in the fluid falls rapidly with distance from the tip of the impellero For this reason Solomons and Weston (80) believed that, in viscous mold broths, virtually all of the oxygen transfer took place within the impeller envelope and that little resulted from the slow-moving bubbles outside of this envelopeo It would be expected that the large increases in viscosity produced by carboxymethyl cellulose compounds would result in a decreased OTRo However, the data presented here show that there is little if any, effect upon the OTR measured by the gluconic acid fermentation when the viscosity is raised by adding carboxymethyl cellulose0o The data for the effect of viscosity on the OTR, measured in this work by the sulfite method or by use of the electrode in a cell-free system9 agree with published works When the OTR was measured by the oxidation of sedium sulfite in the presence of glycerol and sucrose9 the data for the decrease in the OTR duplicated that of Yoshida et alo (101)o.OO.

131 When carboxymethyl cellulose was added to distilled water, the OTR decreased if measured with the oxygen electrode. Analogous results9 indicating reductions in OTR with an increase in viscosity, were reported by a number of investigators, These include Ohain and Gulandi (12), who used killed Penicillium ehrysogenum and measured the OTR with a rotating platinum electrode; Solomons and Weston (80), who used Aspeillu myelium inhibited by sodium azide to increase the viscosity and the platinum electrode to measure OTR; Deindoerfer and Gaden (21), who used the non-steady state polarographic technique to study OTR in a suspension of azideinhibited Penicillium 0hrysogenum. Tie effect of viscosity on the OTR has been measured in the noavbiocin fermentation by Steel and Maxon (84), They reported an initial deerease in the OTR as the apparent viscosity rose. However, as such fermentations progressed, the cell weight and rigidity increased concurrently, The resultant decrease of the 0TR in media in which the viscosity increased, was due not only to the increased mycelial concentration, but also to the structure produced by intertwining of the hyphae, Much of the resistance to oxygen transfer in such a system is due to the slow diffusion of oxygen within the mycelial clumpso Bartholomew et al. (3) noted that this resistance could be a major factor in affecting diffusion into the center of such structures. However9 when cells are adsorbed on the surface of a bubble, the length of the path for oxygen transfer and total

132 resistance to such transfer is less than that for cells freely suspended in the liquid; the driving force which is the difference in dissolved oxygen concentration between the almost saturated film surrounding the bubble and the very low concentration at t urfa ofhe surf f e cell, is at a maximum, Agitation of the fluid will not affect the transfer of oxygen across the film between adsorbed cells and air in a bubble, except as agitation may increase the gas-liquid interfacial area upon which adsorption can occuro Thus, both sets of data, one showing reduction of the OTR with increasing iscosity, and hthe other showing no effect, are consistent;with the concept of liquiddfilm coontrolled- mass-transfer operations in which liquid turbulence or film thickness depends upon viscosity characteristicso

V SMSMARY The study of oxygen transfer rates in the Pseudomonas ovals. fermentation of gluconio acid has suggested that two independent paths of oxygen transfer are operative in aerobic fermentations. Conventionally, it is thought that oxygen dissolves in the liquid and then is transported through the bulk of the liquid to the cells, where it reacts with the enzymeso In this process, the oxygen encounters a number of resistances to its movemento these are the film on the inside of the bubble9 the gas-liquid interface, the film on the outside of the bubble, the bulk of the liquid and the liquid film surrounding the cello Another path appears to exist and operate concurrently as a result of the adsorption of microorganisms on air bubbleso Due to this adsorption, the liquid films surrounding the cell and the bubble are united and oxygen transfer takes place directly across this common film from the air to the cello The common film can be regarded as a membrane and the mechanism designated -as direct oxygen transferO Recognition of the existence of the direct path in the gluconic acid fermentation was based upon a number of experimental observations, three of which are now noted: (1) It was found that increasing the area of gas available for transfer either by sparging more air or by turning the agitator at a higher speed, increased the rate of oxygen transfer. (2) Increasing the oxygen content of the gas used for 133

134 aeration increased the rate of oxygen transfer, even though the dissolved oxygen concentration in the liquid was well above the critical value, (3) The rate of transfer of oxygen dissolved in the liquid to suspended cells, was unaffected by the agitator speed if the liquid was not being aerated, The rate of oxygen transfer has been reported to decrease as the viscosity of a fermentation medium increased. Sucrose, glycerol and carboxymethyl cellulose were added to distilled water and the rate of oxygen transfer determined in a nonliving system by the sulfite method or by means of the oxygen electrode, The Pseudomonas ovalis fermentation was used to study the effect of carboxymethyl cellulose on the oxygen transfer rate in a resting cell preparationo In this living system, the viscosity had no effect on the rate of oxygen transfero However9 in the homogeneous chemical systems9 the rate of oxygen transfer decreased as the viscosity of the medium increasedo The effect of environmental conditions on the rate of gluconic acid production by resting cells of Pseudomonas ovalis was investigatedo The maximum rate was found to occur at pH 7~35 and 37~0o An activation energy of 99600 cal/mole was calculated for this fermentation, The critical dissolved oxygen concentration for resting cells of Pseudomonas ovalis was ol mg/lo

135 APPENDIX A Cali ration of Rotameter The rotameter was calibrated as follows: The rate of air flowing through the rotameter at each reading was determined by passing the gas through a Precision Wet Test Meter, This meter recorded the total volume of air saturated with water at atmospheric pressure, The time taken for a given volume of air to flow through the meter was measured with a stop watch. The wet test meter was calibrated by the volumetric displacement of water and found to be accurate to within + o,5%. The data are found in Table XXII, The flow rate was read by noting the position of the top of the sapphire ball in the rotameter tube, The air flow rate was measured at 730,3 mm atmospheric pressure and 78F~ while saturated with water,

136 TABLE XXII CALIBRATION DATA FOR FISCHER AND PORTER FLOWMBTER No~ 2 F 1/4-20-5, SAPPHIRE FLOAT; CALIBRATED USING A WET TEST METER AT 78~F AND 730.3 mm PRESSURE Rotameter Scale Reading 19,9 19,1 18,0 17.0 16.0 15.0 14.0 13.0 12.0 11,0 10.0 9.0 8,0 7,0 6.0 4,9 4.0 3,0 Rate of Air Flow 1/min. 9.25 8.85 8.12 7,52 6.89 6.38 5.86 5.28 4.69 4.12 3.54 3,.08 2.53 1.99 1,51 0.92 0.46 0.16

137 APPENDIX B Calbration of the Oxyen Electrode The following calibration was run to determine the relationship between the magnitude of oxygen activity in solution, which is read on the millivolt scale of the pH meter, and the chemically measured dissolved oxygen concentration in the liquid: several samples, each containing different concentrations of dissolved oxygen, were made by mixing boiled, distilled water which contained no oxygen and water saturated with oxygen from the air. The meter was first set at zero by placing the electrode in the solution devoid of oxygen; the maximum reading of 1400 millivolts was set by placing the meter in the solution saturated with oxygen from the air, Then the electrode was placed in each of the samples and the reading was recordedo The samples were stirred during the five minute period required for the meter reading to become steady, and the temperature was maintained at 25.0 + 0.2~0, After the reading on the millivolt scale of the pH reter was determined, the dissolved oxygen concentration was determined chemically by the Winkler Test, as outlined in Standard Methods (81), The data are given in Figure 32, A straight line has been drawn through the points for the plot of millivolt reading against dissolved oxygen concentration,

138 1400 0 1200 4-,H / 800 o 600 400 0 200 - 5Z by th Winkl60 method. C3 1 Oxygen Concentration (mg/l) Figure 32. Relationship between the response of the oxygen electrode and the dissolved oxygen concentration in distilled water measured at 25 ~C by the Winkler method.

139 APPENDIX 0 Effect of Ts~rature on the 0O eln Electrode To study the effect of temperature on the measurement of dissolved oxygen concentration with the oxygen electrode, the following test was run: the electrode was placed in samples of distilled water saturated with oxygen from the air. The temperatures of these samples ranged from 1300 to 2400 and were controlled at the set value during the five minute period allowed for the electrode reading to become steadyo At the end of this period during which the samples were stirred, the reading on the millivolt scale of the pH meter was recordedo The data are plotted in Figure 33, This curve shows that a linear relationship exists between saturated dissolved oxygen concentration and temperature. There is approximately a 7% increase in reading per degree Centigrade rise in temperature, This figure agrees very closely with the results of Eye et ale (30), who measured the effect of temperature on the reading over a range from 1300 to 2800,

43 H H 1200 0 1 1000 H go I / I 0 0 0 800 4 -0 600 600 0 15 20 25 Temperature ( C) Figure 33. Effect of temperature on the oxygen electrode response. The electrode was placed in distilled water saturated with oxygen from the air.

141 APPENDIX D Calculation of the Corection for the Rate of Production Undissooiated Gluconic Acid During the presentation of the results showing the effect of pH on the rate of production gluconic acid, it was noted that some of the acid did not ionize-as it was produced, As a result, no sodium hydroxide was needed to neutralize it, Equation 14 was developed to relate the rate of production of undissociated acid to the rate of actual addition of base needed to keep the pH constant: d(HA) = (H+) d(A-) = (H+) N dV 2 fjsj NB dV (27) dt Ka dt d dt In order to give an example calculation, the following data has been used: Run 28A pH 5.25 Rate of addition base (dV /dt) 14,2 ml/hr. B Strength of base (NB) 0.500 N B Ionization constant of gluconic acid 2,4 x 10 Thus the rate of production of undissociated acid was: d(HA) (5.63 x o106) dt 4(2.4 x 104) (14.2) (0o500) = 0.16 meq/(l)(hr.) Then the corrected rate of production of gluconic acid was: (14.2) (00500) + 0.16 = 7.26 meq/(1)(hr.)

142 APPENDIX E Com arison of Oxeen Transfer Rate Measured b the Rate of Production of Gluconoic Acid -to t. h Rate Measured with the Oxy2en Electrode The rate of oxygen transfer, as measured by the rate of gluconic acid production, can be calculated by using Equation 17: NBdvB/dt RG G V F x - mmoles/(l)(hro) 2 (17) In run D-15, the rate of addition of sodium hydroxide 0,2560N was 29ol ml/hr. while the medium was agitated at 570 RPM and air flow rate was maintained at 1,86 VVMo The volume of medium in the fermentor was 2050 mlso o R (0256).(29.1) 182 meq/(l)(hro) G (2,05)(2) After the-rate of addition of sodium hydroxide had attained a steady flow, the stirring speed was quickly changed to 114 RPM, and the air flow to the sparger ceased0 The effect on the oxygen concentration has been shown in Figure 199 and data are given in Table XXIII The rate of decrease of oxygen concentration, dE/dt, was plotted in Figure 209 after a correction was added to account for leakage of oxygen through the surface during removal of dissolved oxygen from solution by the cellso This correction was calculated in the following manner:

143 From t=5 to t=6 min,, the reading changed from 810 to 715 mvy or a decrease of 95 mv/mino In this run, the reading for saturation was 1345 myV and 0 at zero oxygen concentration. The leakage rate waso (d) sKLA (Es E) (28) 1.07 x 10 (1345 810 715) =5 mv/mino ~o ()az = 95 + 5 = 100 mv/mino otal Then Equation 17 was used to calculate the rate of oxygen utilization by the resting cells; (dE/dt) C (60) RE ( ) (s - )mmoles/(l)(hr.) (17) (60) (32) (1345) 0.96 moles/(l)(hr) The ratio of the two rates was 0,96/1,82 = Oo53 showing that more oxygen was transferred in the sparged system, 1. The value of KLA used in Equation 28 was obtained from Table XIX. 2o Rather than integrating, it was deemed sufficiently accurate to use an average value of E for this small interval in Equation 28, 3, The value of dE/dt substituted in Equation 24, was the slope of the curve plotted in Figure 19,

144 TABLE XXIII CHANGE OF DISSOLVED OXYGEN CONCENTRATION IN A SUSPENSION OF RESTING CELLS OF Pseudomonas ovalis AGITATED AT 114 RPM AFTER THE AIR FLOW WAS STOPPED Times Oxygen Concentration mimntes 0 1 2 3 5 6 7 8 9 10 11 12 13 13.5 14 15 millivolts 1290 1185 1100 1010 910 810 715 620 525 430 345 258 180 103 67 38 16 Corrected Rate of Change of Oxygen Concentration mv/min. 106 87 93 104 105 101 102 103 104 95 98 90 90 43 36 41

145 APPENDIX F Calculation of the Parameters in the Power-Law Ecuation for Non-Newtonian Viscosity The rotational viscometer plotted shear rate against shearo An example curve is shown in Figure 27. To interpret the data in terms of standard units, the following machine calibration constants were used: 33,3 grams force/unit for shear stress which was plotted on the y axis, 1028 sec0 /unit for shear rate9 which was plotted on the x axis. To obtain shear stress or shear rate at any point, the scale reading was multiplied by these quantities. The apparent viscosity at any point was obtained by dividing the shear stress by the shear rateo An example of this calculation has been made at the point represented by shear stress 903- and shear rate 10s ta I =T 93 x33.3 x 4.379 dynes/cm2 133 cp a 13 1 TY 1028 x 10 zseo1 The data from Figure 27 were plotted logarithmically in Figure 28~ The slope of the line in the figure was n, the exponent in the power-law equation0

146 APPENDIX G Calculation of Reno1ds Number:r Ai tation The Reynolds number has been used in the -o~etJowi-ng system to correlate power measurements (51), mixing efficiency (39), and scale-up calculations (69), If the Reynolds number is above l000, turbulent flow is presents if below 40, viscous flow is present. In view of the importance of this quantity, calculations of the Reynolds number are shown below for the following systems: (1) no carboxymethyl cellulose added; Newtonian solution in which the viscosity was 1 centipoise (2) the most viscous solution employed; 15 g. Natrosol per literT non-Newtonian In both cases the solutions were stirred at 570 RPMo (1) Por Cells in Gluce-Phosphate Solution BN - (29) NRe 2 3 (3/12) (570/60) (62.4) x 1488 33,000 NRe (1.0) (2) l Cells in WUgsePhos. te Rlusi g rams Natrosol In this- case the apparent viscosity must be the value measured at the appropriate shear rate for the impeller. Metzner and Otto (52) proposed that the average rate of shear in an agitated vessel varies linearly with the rotational speed of the impeller according to the equation:

147 Y a ksN (30) They originally suggested that ks = 13 but recently altered this to 11. Oalderbank and Moo-Young (11) found 10 to be a suitable value. A value of ks = 10 was used in this calculation to compute the apparent viscosity. Using the flow curve with TY 1O = 10(570/60) = 95 sec., the apparent viscosity was found to be 722 centipoises. Calderbank and Moo-Young (11) have shown that the density of carboxymethyl cellulose compounds is close to 1,0.. =. (3/12)2(570/60)(62.4) x 1488 51 Re 722

148 APPENDIX H Nomenclature AA Frequency factor in the Arrhenius equation - 'AB Equation 9 IA COoncentration of anion in solution C Concentration of dissolved oxygen in a liquid solution C0 Critical concentration of dissolved oxygen Concentration of oxygen in the liquid phase in equilibrium with air D Fractional transmittance of light Di Diameter of impeller log D Optical density E Dissolved oxygen activity in solution as read on the millivolt scale of the pH meter E Reading of millivolt meter when electrode was placed in a solution devoid of oxygen E Apparent activation energy for heat deerjuction of the microorganisms E Reading a millivolt meter when electrode is placed in a solution saturated with oxygen from the air g9 Gravitational conversion factor [HA] Concentration of undissociated acid in solution [Ha ] Hydrogen ion concentration at which the rate of glueonic acid production is R A [Q]- Hydrogen ion concentration giving the maximum rate of glueonic acid production, RM II Intensity of a beam of light striking a solution I Intensity of a beam of light after passing through a solution K Constant in Equation 15 Ka Ionization constant for a salt in an aqueous solution

149 K A Mass transfer coefficient for oxygen absorption, L -1 sec. k Rate constant for the thermal destruction of bacteria in Equation 7 k Constant in the shear rate equation, Equation 30 s k Constant in the power-law equation, Equation 3 m Constant of proportionality in Equation 19 N Agitator speed, RPM N Normality of sodium hydroxide used for maintaining B constant pH in the fermentor No Number of viable cells per unit volume of medium n Constant in power-law equation for pseudoplastic materials, Equation 3 0 Total number of microorganisms per unit volume of medium OTR Oxygen transfer rate P Power input to an agitated fermentor Q10 Ratio of the velocity constant at one particular temperature to the velocity constant at another temperature ten Centigrade degrees lower R Universal gas constant RA Rate of production of gluconic acid at any pH R Rate of oxygen utilization in meq/(l)(hr.) E measured by the oxygen electrode R Rate of oxygen utilization, meq/(l)(hr.) G measured by the rate of gluconic acid production RM Maximum rate of production of gluconic acid at RM the optimum hydrogen ion concentration [Ho] t Time T Absolute temperature, OK VB Volume of sodium hydroxide added to medium to maintain constant pH during the fermentation of glucose to gluconic acid

150 VF VVM dO/dt dE/dt (dE/d t) (dE/dt)T 7) '0 Ia "Ir NRe Volume of nitrogen-free glucose-phosphate medium in the fermentor Volumes of air per volume of medium per minute Rate of change of oxygen concentration with time Rat e of change of oxygen activity with time, millivolts/min, Correction factor for oxygen leaking through the surface of the liquid during removal of dissolved oxygen by cells Rate of fall of millivolt reading due to removal of dissolved oxygen by cells minus the supply through the surface of the liquid, as noted above Shear rate Turbidity coefficient in Beer's Law, Equation 11 Newtonian viscosity Apparent viscosity, ratio of shear stress to shear rate Density of a solution Shear stress 2 D% Np Reynolds number for agitation equal to -- Pa Pg0 Power number equal to c D N DiN

BIBLiO GRAPHY 1, Aiba, S,, Mo Hara, and J. Someya, Progress Report No. 21, Biochem. Eng. Lab,, Inst. Applo Microbiol.9 Univ. of Tokyo 9 Japan, 1962. 2, Barker, J. J,, and R. E. Treybal, A, I, Oh. h E J, 6, 289 ( 1960). 3. Bartholomew, W. H., E. 0. Karow, Mo R. Sfat, and R, H. Wilherlm, Ind E. COhem,,, 4 1801 (1953). 4, Benson, S., W he Poundationas of Chemical Kinetics, MeGraw-Hill Book Co., Ino., New York, N. Ye, 1960Q. 5, Blaok, S. H. Ro E. MaoDonald, T, Hashimoto, and P. Gerhardt9 Nature, 9, 782 (1960). 6. Blakebrough, N,, G. Hamer, and M, W. Walker, Trans, Inst Che~m EoD,9 A71 (1961). 7. Bowers, R, H,, J. Ai. Chem., f, 542 (1955), 8. Boyles, W. A,, and R. E. Lincoln, Apl, Miorobiolo. 327 (1958)o 90 Brierley, M. R., and Ro Steel, Ap1. Microbiol. 1, 57 (1959), 10. Oalderbank, P. H., Trans, Inst. h. Eg. 37, 173 (1959) 11.o alderbank, P. H., and Mo M. B. Moo-Young, Transo Inst. Ohm. En. o9, _ 26 (1959). 120 Chain, EB B., and G. Gulandi, Rendicti Institut Superiore Di Sanita, l, 5 (1954. 13. Chick, H., and 0, J. Martin, J. Ph siol, (London), 40, 405 (1910), 14 1Clark, D. S,, Bioteoh. and Bioen,, 4, 241 (1962). 15. Clark, L, C., Jr., Trans. Am. Soc. Artificial Inst, Organs, 2 '41 (1956). 16. Oolwick, S., and N. Kaplan, Methods in Enz ymolo Vol. III, Academic Press, New York, N. Y,, 1957o 17, Committee on Sanitary Eng. Rsch, of the Sanitary Eng. Divo, J E 86, 2556 SA4 (1960). 151

152 18..amposano, A,, E, B. Ohain, and Go Gulandi, Proo. VIIth Int. 0ong. Miorobiol. Sweden, 1958, 19. Oooper, 0. M., G. A. Fernstrom, and S. A. Miller, Ind Ego Ohem.,, 504 (1944), 20. Deindoerfer, PF H., Apl,. Miro biol. 5, 221 (1957). 21, Deindoerfer, P. H,, and E, L. Gaden, Jr., Ar. Microbiol., 1, 253 (1961). 22, Deindoerfer, F. H., and A. E. Humphrey, Ind. Eng. Ohem., 5, 755 (1961). 23, Deindoerfer, F. H,, and J, M. West, J, Biochem, Microb. th. and gT.,, 165 (1960)o 24, Dion, W. M., A. Carilli, G. Sermonti, and E. B. Chain, E ndionti Snettuto..Sp..eri.e Di Saa, della Roma, Engiz sh ed i 17. 187 (195 4) e 25. Dognon, A., Bu. s a. chim. biol., 2, 249 (1941). 26. Dognon, A., ti.j Sientifiaue, 12, 613 (1941). 27, Donovick, R., hpI. Micro 1., 8, 117 (1960). 28, Eoker R. E., and W. R. Lookart, i. Baoteriol. 82, 511 (19615 29 Elsworth, R., V. Williams, and R. Harris-Smith, J. Ap. cem. I, 261 (1957). 30, Eye, J. D,, L, H. Reuter, and K. Keshavan, Water ew. 0o12 1o 8P 231 (1961), 31, Pinn, R. K. 136th Meeting, AOS, Atlantio Oity, N. J., Sept. 1959* 32. Pinn, R. K., aete o. toa,.,;, 254 (1954), 33* Priedenwald, J. S., and G. D. Maengwyn-Davis, jA.mposium t he Mhanisam j Bsnzymie~ Aciln, W. X D. MoElroy, and B. Glass, ds, Th Johns HpnsPress, Baltimore, Md., 1954. 34. Priedman, A. M, and S. N. Lightfoot, Jr.,. E. Qem., i, 1227 (1957), 355 Gaudin, A. M., lt to1, 2nd ed., MoGrawi-Hill Book 0., New York, N, Y., 1957, 36. Gaudin, A. M,, AN. S. Davis, and S, E, Bangs, oBilath. a 0i3^a&~, 4, 211 (1962),

153 37. Gaudin, A. M., A, L. Mular, and R. O'Cnnor, Appl. Microbiol., 8, 84 (1960). 38. Gaudin, A. M., A. L. Mular, and R. F. O'Connor, AjpP. Mierobio1., 8, 91 (1960). 39, Godleski, E. S., and J. 0. Smith, A. I. Ch. E., 8, 617 (1962). 40. Hixson, A. W., and E. L. Gaden, Jr., Ino En. Ohem., 4, 1792 (1950). 41. Ingraham, J. L., and G. F. Bailey, j. BDactriol., 77, 609 (1959). 42. Ippen, A. T., and 0. E. Oarver, Sew. and Ind Wastes, 28, 813 (1954). 43. Kempe, L. L., Adv. A 'Miorobiol., 2, 313 (1960). 44. Kempe, L. L., R. A. Gillies, and R. B. West, Ap1. Micrebiol., 4, 175 (1956). 45. Kurokawa, M., M. Hatano, N, Kashiwagi, T. Saito, S. IShida, and R. Homma, j. Bacteriol., 8, 14 (1962). 46. Lamanna, 0., and M. F. Mallette, Basic Baoteriology, The Williams and Wilkins 0o., Baltimore, Md., 1959. 47. Lockwood, L. B., B. Tabenkin, and G. E. Ward, J. Bacteriol., 42, 51 (1941). 48. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J. Randall, i. Biol. Ohe. l 195 265 (1951). 49. Mancy, K. H., and D. A. Okun, I. Watr Pollution Oontrol PFderation, 2, 351 (1960). 50. Merck & Co., Inc,, The Merc' k Index, 7th ed., Rahway, N. J., 1960. 51. Metzner, A. B., R. H. Feehs, H. L. Ramos, R. E. Otto, and J. D. Tuthill, A. I. Oh. E. J.,, 3 (1961). 52. Metzner, A. B., and R. E. Otto, A. I. h. E. J., ~, 3 (1957). 53, Miyamoto, S. and T. Kaya, Bull. hem. Soc. Japan, 5, 123 (1930). 54, Miyamoto, S., T. Kaya, and A. Nakata, Bu ll. hem. So. Japan, 5, 229 (1930).

154 55. Morgan, P. F., and J. K. Bewtra, J. Water Pollution Control Federation, 4, 363 (1962T. 56. Ostwald, W., Kollaldzschrd, 8, 261 (1926). 57. Ough,. S., and M. A. Amerine, Food Rsch, 24, 744 (1959). 58. Pfeifer, V. F., and 0. Vojonovich, Ind. En. Ohem 1940 (1952). 59. Phillips, D. H., and M. J. Johnson, Biohe Miorobiol. Teh. and.a, 2, 261 (196l)... 60. Phillips, D. H., and M. J. Johnson,,. Biochem. Microbiol. Tech. and E., 3, 277 (1961). 61. Phillips, K. L., H. R. Sallans and J. F. T. Spencer, Ind. Eng. Ohem., _Z, 749 (1961). 62, Porter, J. R., Bacterial Ohemistry and PhySiology, John Wiley and Sons, Inc., New York, N. Y., 1946 63. Rahn, 0., Bacteriol. evs., ~, 1 (1945). 64. Reiner, M., Deformation and Flow, Lewis, London, 1949. 65. Richards, J. W., Br. Ohem. En.,, 454 (1961). 66. Richards, J. W., Progress in Indstrial l robiolo Vol. III, D. J. DP. Hoekenhull, Ed., Interscience Publishers, New York, N. Y., 1961, p. 140. 67. Rolinson, G. N., J. Gen. Microbiol., 6, 336 (1952). 68. Rose, J. B., and G. F. Somers, Ohemistry and Methods of Enzymes, Academic Press, New York, N. Y,, 19553 69. Rushton, J. H., E. W. 0ostioh, and H. J. Everett, Chem. na.. Prog., 4, 467 (1950). 70, Saymn, J. 0. and A. M. Mattocks, J. AmPham. p Assoc. ~, 310 (1957). 71. Seifter, S., S. Dayton, B. Novic, and.. Muntwyler, Arch. Biochem., j, 191 (1950). 72. Sevag, M. G., J. Smolens, and D. B. Lackman, J. Bio1. Ohem.1, 523 (1940). 73. Shedlovsky, L., Ann.. Y. Aad. Sci., 4_, 279 (1948). 74. Shu, P., AE. and Food Chem., 1119 (1953). i. _~~~~~~~~~~~~~o

155 75. Siegell- S. D,, and E. L. Gaden, Jr,, Biotech. and Bigeng, 4 345 (1962). 76. Sizer, I. W., Advances in Enzymoloy, Vol. II, Interscienoe Publishers, Inc., New York, N. Y,, 1943, p. 38. 77. Smith, 0. G., Appl. Bacteriol.,,8 42 (1960). 78. Solomons, G. L., Th.e S of hem. d., Monograph No. 12k 233. 79, Solomons, G. L., and M. P P.erkin,. Appl. Chem, (London), 8, 251 (1958). 80, Solomons, G. L., and G. 0. Weston,. Bioohem. Micobiol. Teh. and En.,, 1 (1961). 81. Standard Methods for the Examination o Water and Waste WAter, Ilth ed., American Public Health Assoc., Inc., New York, N. Y., 1960. 82. Staub, IN C.,,,. A_1. Physiol., 16, 192 (1961). 83, Steel, R., and W. D. Maxon, 139th Meeting, ACS Convention, Atlantic Oity, N. J., Sept. 1962. 84, Steel, R. and W. D. Maxon, Bioteh. and Bieng., 4, 231 (1962. 85. Strohm, J. A., and R. F. Dale, Iln. g_. Ohem., 5, 760 (196 L). 86. Strohm, J. A., R. P. Dale, and Ho J. Peppler, Apl. Morobiol., Z, 235 (1959). 87, Sultzer, B. M., tJ Bacteriol,, 492 (1961). 88. Sumner, J. B., and G. F. Somers, ohemistry and Methods o ERz.vmes, Academic Press, Inc., New York, N. Y., 19553. 89. Tengerdy, R. P., J. Biochem Microbiol. Teh. and En., 2, 241 (1961). 90. Terui, G,, and N. Konno, Tech. Reports of the Osaka Universitv9, 0, 889 (196o0)T" 91. Terui, G., and N. Konno, Tech. Reports f the 0saa University, 1, 447 (1961). 92. Terui, G., N. Konno, and M. Sase, Teh,. Reports of the Osaka University, 10, 527 (1960). 93. Tsao, G. T. N., Ph.D. Thesis, University of Michigan, Ann Arbor, 1959.

156 94. Tsao, G. T. N., and L. L. Kempe, J Biochem. and Miorobiol. Teoh. and _., 2, 129 (1960). 95. Van Slyke,, D. D.,. Biol. Ohem. 2, 525 (1922). 96. Vondrackova, J,, Oeskoslov, Microbiol., _, 37 (1957). 97. Warshowskly, B., and B. J. Sehantz, Anal. hem. 26 11 (1954). 98, Wasserman, A. E., and J. W. Hampson, APp. MiErobiol Te. and., i, 163 (1959). 99. West,, J M, and E. L. Gaden, Jr., J. Biochem. Merobiol. Teh. an Eng., i, 163 (1959)T 100. Winzler, R. J, J. Oell. oom. Physiol., 1_, 263 (1941). 101, Yoshida, F., A. Ikeda, S. Imakawa, and Y. Miura, mInd, Eng, Ohem,, ^g, 435 (1960). 102. Ziemls-ki, S. A., and R. L. Hill, J. Ohem. En. Data, 51 (1962).

UNIVERSITY OF MICHIGAN 1 1I11 III 11 11 1111111111111 1111111111111 3 9015 02223 2220 THE UNIVERSITY OF MICHIGAN DATE DUE itl/ 5;oZf