'0 .\ .hn 0‘ I‘H' ‘1'er 1. “(mggmp 1...‘ . A . .r ,_ 'hHu‘ mwn‘u ~r-r‘n n. 112-. v 3...... '4 :- .4- "3'1!'- . .w , .21. <'.‘..’.:~- :J- ”H“. ”1..”- (”F-- ... "...""."""* . ,.- n."0‘"‘h’avt , V.,,...' . , rad ....n........- nu... .W,....,.,A..m-v; . .0. ....... ‘ ”- w "I :l‘nvn-r-rd: Date ‘Ja’v‘ ¥ 0-7639 IBRARIES lll'cllllj Illill'lm WIN/ll r 31 La till/lime l This is to certify that the thesis entitled The Production of Calcium Magnesium Acetate from Renewable Biomass via Microbial Fermentations presented by Kurt P. Rindfusz has been accepted towards fulfillment of the requirements for M.S. degree in Chemical Engineering 3% a % . Major professor 13 /??£ MS U is an Affirmative Action/Equal Opportunity Institution F MIRA!" Michigan “ate 1 University Li J __~ PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. DATE DUE DATE DUE DATE DUE @9101 125107301 ll 44;] 'TT-TW t 5 ill—T fl ,4 —J—TI IL fil l MSU Is An Affirmative Action/Equal Opportunity Institution cmmut PRODUCTION OF CALCIUM MAGNESIUM ACETATE FROM RENEWABLE BIOMASS VIA MICROBIAL EERMENTATIONS BY Kurt Patrick Rindfusz A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Chemical Engineering 1991 ABSTRACT TEE PRODUCTION OF CALCIUM MAGNESIUM ACETATE FROM RENEWABLE BIOMASS VIA MICROBIAL PERMENTATIONS by Kurt Patrick Rindfusz Sodium chloride has been used for years as an agent to remove snow and ice from public highways. Unfortunately, studies have shown that NaCl contributes greatly to corrosion of vehicles, bridges, concrete, and underground utilities. Its use as a deicer also has adverse effects on the environment and human health. These side effects cost the American public an estimated 5.72 billion dollars each year. This prompted a search to find a suitable deicing substitute. Calcium magnesium acetate (CMA) has been determined to be the substitute deicer of choice. Unfortunately, current CMA prices are 29 times that of NaCl. This paper investigates a novel microbial fermentation which utilizes renewable biomass to produce CMA. An economic analysis revealed that the proposed fermentation can produce CMA for approximately 25.76 cents per pound. A similar hypothetical fermentation, using technology from the vinegar industry, can produce CMA for 16.34 cents per pound. PAGE LIST OF TABLES .......................................... iv LIST OF FIGURES ......................................... v LIST OF ABBREVIATIONS USED ............................ viii CHAPTER I: INTRODUCTION ................................ 1 CHAPTER II: BACKGROUND ................................. 3 l. Deicers Today ...................................... 2. Disadvantages and hidden costs of sodium chloride.. 3. Alternatives to sodium chloride .................... l4 4. Calcium magnesium acetate .......................... l8 5. Advantages and disadvantages of CMA ................ l9 6. Summary ............................................ 25 CHAPTER III: LITERATURE REVIEW ......................... 27 1. Current use of CMA ................................. 27 2. Production techniques .............................. 3O 1. Acetic acid production via acetic acid bacteria ....................................... 32 a. History of vinegar/acetic acid production .. 32 b. History of the acetic acid bacteria ........ 37 c. Advantages and disadvantages of acetic acid production via acetic acid bacteria (motives behind the research) .............. 44 2. CMA production ................................. 46 CHAPTER IV: THE RESEARCH PLAN AND PROJECT GOALS ........ 51 CHAPTER V: MATERIALS AND METHODS ....................... 55 1. Bacterial acquisition and storage .................. 55 2. Inoculation procedure .............................. 56 ii iii 3. Sample preparation and HPLC analysis ............... 57 4. Optical density measurements ....................... 6O 5. Dry weight analysis ................................ 6O 6. Growth rate determination .......................... 61 7. Fermentor operation ................................ 64 CHAPTER VI: EXPERIMENTAL PROCEDURES AND RESULTS ........ 66 1. Shake flask experiments ............................ 66 1. Screening and determination of the two most promising bacterial strains .................... 67 2. Determination of optimum conditions for growth and production ................................. 70 a. Substrate inhibition ....................... 70 b. Addition of corn steep liquor .............. 71 c. Addition of lactose and glucose ............ 76 d. Effects of temperature ..................... 80 e. Effects of pH .............................. 86 f. Product inhibition ......................... 89 2. Experimentation in bench scale fermentors .......... 94 1. Batch experiments .............................. 94 a. Growth rate studies ........................ 94 b. Comparison of sparging with oxygen vs. air ........................................ 99 c. Comparison between Ca/Mg OH and NaOH for pH control ................................. 101 2. Semi-batch experiments ......................... 103 3. Continuous flow experiments .................... 107 CHAPTER VII: DISCUSSION ................................ 112 1. Summary of Results ................................. 112 2. Economic evaluation ................................ 119 3. Recommendations for further study .................. 132 CHAPTER VIII: SUMMARY .................................. 137 BIBLIOGRAPHY ............................................ 143 TABLES Ba-Bf' Reported use of sodium chloride, calcium chloride, and abrasives by states in the winter of 1966-1967 (tons). National annual costs of NaCl related damage (Millions of 1991 dollars). Potential alternative methods for reducing NaCl use as a road deicer. Performance data for 28 promising deicing chemicals. Comparative substrate inhibition studies for A. oxydans and G. oxydans subsp. suboxydans. Summary of shake flask results. Summary of assumptions for economic analysis of fermentation processes. Capital cost and operating cost estimates for proposed CMA production plants. iv 13 14 16 70 115 124 125 FIGURES 53-5h. FIGURE 7i'2d' The Fring's Acetator. Detailed metabolic pathways of various acetic acid bacteria. Absorbance vs. time. A view of the phases of bacterial growth. ln(absorbance) vs. time. Another view of bacterial growth. Diagram of the bench scale fermentor used in experimentation. Lactic and acetic acid profiles of A. oxydans, G. oxydans subsp. suboxydans and A. cerinus. Acetic acid production by A. oxydans and G. oxydans subsp. suboxydans upon addition of low levels of corn steep liquor in various lactic acid media. EIGHEE_§; Acetic acid production by A. oxydans upon FIGURES 9a-9d' addition of high levels of corn steep liquor. Acetic acid production by A. oxydans upon addition of 10% lactose or 10% glucose to various lactic acid media. 36 40 63 63 65 68' 73 75 78 EIGUEE 11b; FIGURES JZE‘JZQ' Specific growth rates of A. oxydans at 26 and 33 degrees Celsius. Acetic acid production by A. oxydans at 26 and 33 degrees celcius (on a 50 mM lactic acid medium). Growth of A. oxydans at 26 and 33 degrees Celsius (on a 50 mM lactic acid medium). Comparison of growth and acetic acid production by A. oxydans at 33 and 37 degrees Celsius. Growth profiles of A. oxydans for times greater than 4 hours at varying pH levels. Growth profiles of A. oxydans for times less than 4.5 hours at varying pH levels. Effects of product inhibition on growth rate of A. oxydans. Effects of product inhibition on acetic acid production by A. oxydans. Total carbon profile for A. oxydans growth Growth of A. oxydans on varying levels of peptone and yeast extract. vi 82 83 83 84 87 88 9O 92 92 97 FIGURES ZZQ'ZZD' FIGURES 23a’23h‘ Specific growth rate of A. oxydans in varying levels of lactic acid medium. Comparison of acetic acid profiles of A. oxydans in a batch fermentation grown on air vs. pure oxygen. Comparison of acetic acid profiles of A. oxydans in a batch fermentation grown with NaOH vs. Ca/Mg OH. Lactic and acetic acid profiles of A. oxydans in a semi-batch fermentation grown on 495 mM lactic acid. Lactic and acetic acid profiles of A. oxydans in a semi-batch fermentation grown on 577 mM lactic acid. Effects of altering dilution rate on lactic and acetic acid profiles in a continuous flow system. Effects of altering medium lactic acid concentration in a continuous flow system. Calcium magnesium acetate production from the proposed "lactate" process. CMA production with fermentation technology from the vinegar industry. vii 98 100 102 104 106 109 111 121 122 American Tissue and Culture Collection Clarified corn steep liquor Calcium magnesium acetate Calcium magnesium lactate Corn steep liquor Continuous stirred tank reactor Federal Highway Administration High performance liquid chromatography Revolutions per minute Tri-carboxylic acid United States Environmental Protection Agency Volume/volume/minute viii CHAPTER I: INTRODUCTION The size of the nation's highway system and the number of automobiles traversing it has increased dramatically in the latter half of the twentieth century. Along with this increase comes the growing necessity to provide safe traveling conditions for the millions of Americans using these roads every day. To furnish such conditions in the northern and mountainous regions, state governments must combat the seasonal forces of mother nature - snow and ice. Rock salt (sodium chloride) has long been used as a roadway deicing agent due to its availability, low initial cost, and good performance as a deicer. Recently, however, it has been shown that sodium chloride has numerous adverse side effects which limit its utility and indirectly increase its cost to society by millions, perhaps billions, of dollars each year. Among these side effects are: health hazards where highway runoff from.melting snow finds its way into water supplies, damage to roadside vegetation and aquatic ecosystems, corrosion of highway cement, corrosion of the steel found in bridges and reenforcing rods in 2 concrete intrastructures, and probably most important to the general public - corrosion of vehicles. Due to these drawbacks, the Federal Highway Administration (FHWA) along with the United States Environmental Protection Agency (USEPA) and many state highway commissions have searched for suitable replacements which would reduce the extreme corrosion problems and health risks of NaCl while retaining its deicing capability and economic feasibility. While many substances have been reviewed and used in isolated instances (including sand, calcium chloride, and urea) the substitute deicer of choice is calcium magnesium acetate (CMA). It has been shown that CMA can provide adequate deicing while eliminating the health and environmental risks invoked by NaCl. Further, corrosion tests involving CMA prove that it is non-corrosive toward steel and concrete. Unfortunately, the current market price of CMA is nearly thirty times that of rock salt. This high market price, whichmakes complete replacement of NaCl with CMA unrealistic, is a direct result of the cost associated with current production techniques. The research conducted here investigates a method of producing CMA via a microbial fermentation of renewable biomass which has the potential of reducing its production cost so that CMA may become an economically feasible substitute for NaCl. CHAPTER II : BACKGROUND Since 1960, sodium chloride has been the principal agent used by state highway agencies for deicing the thousands of miles of public roadways. The principal reason for its extensive use is that NaCl serves as an efficient deicer by lowering the freezing point of water 21 degrees Celsius. Further, it accomplishes this task quickly enough to melt snow and ice within minutes (1). Also, NaCl is present in the form of rock salt throughout most of the Snow Belt in enormous deposits near the Earth's surface. This makes it extremely accessible and inexpensive, as large quantities are available at a moments notice without any production costs other_than those incurred by mining, shipping, and handling. When considering the miles of public roadways that must be kept clear of ice (Michigan alone has over 117,000 miles of interstate, county, and local roads (2)) these factors become extremely important. Other deicing substances have been used in large quantities in past years. Calcium chloride (CaClz) is one such substance. Over 150,000 tons of CaClz are used on an 4 annual basis for deicing purposes. This amounts to approximately one third of the calcium chloride produced in the United States. The advantage of using CaC12 is that it lowers the freezing point of water one and one half times more than NaCl. This allows for a more complete and faster clearing of snow and ice in extremely cold weather. Unfortunately, the cost of CaC12 is over $100/ton delivered as opposed to $25/ton delivered for NaCl (3). Thus, although it is frequently used in extremely low temperature situations, NaCl has remained the primary deicing chemical. Abrasives, such as sand, have also been used to provide traction during inclement weather. Often these abrasives are applied in conjunction with chemical deicers allowing for both the melting of ice and the added bonus of increased traction on the resulting wet roadway. These abrasives, however, cause problems in populated areas, especially large cities, because they do not dissolve in the runoff water. They simply pile up on road sides or find their way into water drainage systems and sewers, clogging and damaging many of these important city waterways. Also, the effectiveness of abrasives is generally considered to be less than that of using NaCl. Table 1 gives a state-by-state breakdown of the usage of each of these deicers in the winter of 1966—1967. It is 5 easily seen that the use of NaCl far outweighed that of either CaC12 or abrasives (4). As high as these figures are, the use of NaCl has significantly increased each year so that today, a staggering nine million tons of rock salt is spread on the nations highways each year - one million tons in New York State alone (5). This accounts for over 10% of the world wide NaCl consumption and over 20% of the consumption in the United States (6). 6 TABLE 1 Reported use of sodium chloride, calcium chloride, and abrasives by states in the winter of 1966-1967 (tons). EASTERN STATES STATE Neg}: C8912 nus IVES Maine 99,000 1,000 324,000 New Hampshire 118,000 - 26,000 Vermont 89,000 1,000 89,000 Massachusetts 190,000 6,000 423,000 Connecticut 101,000 3,000 335,000 Rhode Island 47,000 1,000 86,000 New York 472,000 5,000 1,694,000 Pennsylvania 592,000 45,000 1,162,000 New Jersey 51,000 6,000 70,000 Delaware 7,000 1,000 2,000 Maryland 132,000 1,000 40,000 ,ZEEQinia 77,000 22,000 204,000 TOTAL .__l.975.000 3.000 4,455,000 NORTH-CENTRAL STATES Ohio 511,000 12,000 43,000 West Virginia 55,000 9,000 230,000 Kentucky 60,000 1,000 - Indiana 237,000 6,000 77,000 Illinois 249,000 10,000 60,000 7 TABLE 1 (continued) NORTH -CENTRAL STATES sm'rr. Nag, oac12 nmsms Michigan 409,000 7,000 6,000 Wisconsin 225,000 3,000 102,000 Minnesota 398,000 14,000 84,000 North Dakota 2,000 1,000 13,000 TOTAL I__3,;46,000 63,000 615,000 I_ L' :s 1 :un ||,L e. 0 Ol_lu .,O: I The amount of rock salt needed to supply adequate highway deicing is enormous. For instance, 600 pounds of salt is required to remove 0.2 inches of ice from a one mile section of roadway which is 20 feet wide. This yields runoff water with a salt concentration between 69 and 200 grams per liter (4). This high concentration of NaCl has caused many unforeseen problems ranging from health risks to vehicular corrosion. One such problem occurs when street runoff water finds its way to local water supplies and lakes. Over time, sodium and chloride ions from the runoff can accumulate to 8 levels which are high enough in many communities to be considered health risks. Salt intake has been cited as a causative factor in hypertension, heart disease and other circulatory problems, liver and kidney disorders, and metabolic disorders. Due to these facts, it has been estimated that 29% of Americans should reduce their salt intake. Also, the United States Public Health Service has set the limit for recommended chloride concentrations in groundwater supplies at 0.25 g/l and a recommended limit of 20 mg/l of sodium for those on a sodium restricted diet (7). There are many examples of salt-laden runoff causing local water supplies to exceed these limits. Studies of municipal sewers in Milwaukee, for instance, indicated that daily chloride loads were 40 - 50% higher in the winter months than in summer. Other studies showed that street runoff changes the yearly salt concentration in farm ponds and lakes. A survey of 27 farm ponds in Maine showed salt concentrations increasing on a yearly basis where road salting occurred. Similar studies in Wisconsin, Michigan, and New York attributed density stratification of chlorides to salt runoff (4). Perhaps the most severe case of salt contamination has (occurred in Massachusetts. Over a seven year period, from 1983 through 1990, 100 of the 351 Massachusetts 9 municipalities had complained of salt contamination (21). Further, in 1978, it was estimated that 25% of the peOple in the Snow Belt were drinking water contaminated with road salt, and the cost of supplying fresh water to these people would be approximately 150 million dollars per year (6). Environmental issues must also be considered when using large quantities of rock salt. Sodium chloride can have toxic effects on roadside vegetation and has significantly altered the aquatic ecosystems of streams and lakes in areas where heavy salting is practiced. Most studies dealing with plant injury and death due to deicing have focussed on the sugar maple decline in a 16 state region in and around New England. In this area, it was observed that the number and health of sugar maple trees has significantly declined since the onset of heavy salt usage in the early 1960's. Several studies have linked this demise directly to the use of NaCl as a road deicer. Among the symptoms studied were: leaf scorch, early coloration and defoliation, reduced growth, browning of twigs, branching, and death of trees. These studies observed that defoliation and serious damage were more prominent on the side of the tree fronting streets, and on trees located on the side of the street receiving the salt-laden drainage. Also, they noted that trees greater than thirty feet from salted roads were found to be "almost always healthy". The 10 conclusion derived from these results was that wind blown salt residue was accumulating on and killing the trees (4). Not only is the airborne residue harmful, but it has also been determined that NaCl in runoff water can cause damage to roadside foliage. Sodium ions from the salt can cause the water absorbing capability of soils to be significantly decreased. Also, the soil often becomes compacted when exposed to high concentrations of sodium. This prevents rain water and air from penetrating to the plant roots. Ultimately, these problems result in death of roadside vegetation (8). The cost to society of this roadside damage is surprisingly high. It has been determined that a 15 foot tree is valued at $1,767. Using this as a basis, one study estimated the annual economic loss from tree death alone to be in excess of 50 million dollars (6). High concentrations of salt in runoff waters have also been responsible for altering the aquatic ecosystems of streams, ponds, and even large lakes. Lake wide dissolved chloride concentrations in Lake Michigan, for example, have increased from 4 mg/l to 9 mg/l in the latter half of this century, with recent rates being 0.1 mg/l/yr. In fact, parts of Lake Michigan have been so contaminated, that they are now considered to have a saltwater ecology (9). 11 Lake Erie and Lake Ontario have seen chloride concentrations increase from less than 8 mg/l to nearly 30 mg/l in this same time span. These increases were also on a lake wide basis, suggesting that the chloride is coming from a large, widespread source (such as road deicing) and not from an isolated contamination (9). Rock salt contamination of fresh water streams and lakes is particularly damaging to invertebrate communities. To illustrate this, a study was conducted on several streams in New York's Adirondack Mountains. Samples of invertebrate populations were taken at points both upstream and downstream from salted roadways. The results of this work showed that invertebrate populations (particularly insect species) were significantly lower downstream from the salt contamination in every case (10). While no one will deny that health concerns and environmental hazards stemming from NaCl usage are important issues, many would argue that they are necessary evils since NaCl usage potentially saves countless lives each winter by preventing accidents. One side effect of NaCl remains however, which pits every American against it. It reaches deep into the pocket book by causing tremendous amounts of rust and corrosion. This corrosion nearly doubles the depreciation rate of vehicles where it is used. In fact, the estimated annual economic loss from this depreciation is greater than any other cost of capital equipment in the 12 United States today (11). In 1978, the state of Michigan alone lost nearly 200 million dollars from depreciation of vehicles caused by rock salt. This was nearly equal to the total national cost of NaCl purchase and application (6). Rock salt corrosion creates even more expenses than the 200 dollars per year the average car owner pays for salt related damage to their vehicles. NaCl increases the damage of repeated freezings and thawings of cement causing "scaling"; corrodes underground utilities such as cables, transformers, and water mains; and contributes to the deterioration of bridges by rusting their steel members and corroding reenforcement bars in the bridge decks. Table 2 gives the results of a 1976 study which summarized the annual economic loss due to usage of NaCl (updated to 1991 dollars). This table shows that the total cost of salt-related damage is in excess of 5.72 billion dollars per year. Also, the indirect costs account for over 93% of the total, outweighing that of salt purchase and application more than 14 times (4). 13 TABLE 2 National annual costs of NaCl related damage (millions of 1991 dollars). sooner. §_x_1,05 Utilities 20 Vegetation 100 Water Supplies and Health 200 Salt Purchase and Application 400 Highway structures 1,000 Vehicles 4,000 T_0'I'_A_L 5 . 72 BILLION 14 21li__ALIEBNAIIMES.IQ_SQDIHM;CHLQEIDE Clearly, the hidden costs to society due to the use of NaCl are so great that alternative deicing methods and alternative deicers need to be considered. Table 3 gives several alternate deicing techniques currently under consideration by the United States Environmental Protection Agency (USEPA) (4). TABLE 3 Potential alternative methods for reducing NaCl as a road deicer. I ME_THOD IExternal and/or in slab melting systems. obile thermal "snow melters". Compressed air snow plows and sweepers. Inclusion of snow and ice adhesion-reducing substances in the cement or application of such a substance on the road surface as needed. Pavement that will store and release solar energy. Electromagnetic energy to shatter ice. Road and drainage modification to enhance runoff. Salt retrieval and treatment possibilities (for reuse) enhanced by addition of chelating agents. Improved tire and vehicle designs to reduce deicing requirements. 15 Although each of these methods can theoretically reduce the need for chemical deicers, it is clear that none can eliminate or even substantially reduce their use. Thus, the USEPA and the FHWA have also sponsored a great deal of research to find and test chemicals which could be used as alternate deicers. In 1980, the FHWA sponsored a study which would systematically evaluate all known chemicals for this purpose. This research eliminated chemicals that were costly, hazardous, and rare. The remaining chemicals were then tested for their deicing characteristics and compared to NaCl for their economical feasibility. Table 4 gives a summary of the results from tests run on the 28 most promising deicing candidates (12). 16 TABLE 4 Performance data for 28 promising deicing chemicals. EUTECTIC WITH WATER AMT. TO LIQUIFY ICE AT -10 C CANDIDATE m. 9. TEMP cos'r Lia/100 case To pH DEICER mucus (C) ($/Lba) Lb ICE mm- or N22; INORGANIC CHEMICALS NaCl 23.3 -21.1 .014 16.0 1.0 7 NaHC03 >6 -2 .08 b 8 Nazco3 7 -3 .024 b 12 NaH2P04 35.6 -9.7 .107c 55 26.4 3-4 NazHPO4 1.56 -.5 .101c b 9 KHCO3 19 -8.8 .14 b 8 K2C03 41 -36 .1 19 8.5 12 KH2P04 11.6 -2.7 .134c b 4—5 K2HPO4 36.8 -13.7 .145c 47 30.4 9 K4P207 60 -39 .35 37.4 58.4 10-11 NH4H2PO4 18.5 -6 .08 b 4—5 (NH4)2HPO4 35 -14 .067 32.5 9.7 8-9 NH4HC03 10.6 -9.5 .03d 11.9 1.6 8 (NH4>2C03 30 —14.6 .04d 31.6 5.6 9 3 <20 <-77 .09 8.1 3.3 12 17 TABLE 4 (continued) EUTECTIC WITH WATER AMT. TO LIQUIFY ICE AT -10 C CANDIDATE WT. 96 TEMP COST Lb/ 100 COST TO pH DEICER DEICER (C) ($/Lb~‘-|) Lb ICE THAT or N391 onenggp CHEMICALS ethanol 83 -125 .065 19 5.5 7 Ethanol 93 -131 .17 23.5 17.8 7 Iso- >80 <—42 .127 29.9 16.9 7 Propanol Acetone >60 <<-27 .15 50 33.5 7 Urea 33 -13 .08 34 12.5 7 Formamide 65 -45 .35 34 53.5 7 imethyl >50 <-50 .47 35.1 73.7 7 Sulfoxide Ethyl 6O -6 .32 b 7 Carbanate a All prices are bulk and derived from the 7. 5.76 issue of CbemicaLMarketinsLRenerter except that for C02 which was obtained 10/6/76 from an Airco representative in Chicago. No liquid phase at —10 degrees Celsius. C From H3PO4 and the carbonate. d Mixed gases C02 and NH3. 18 The authors of this study concluded that the most logical deicing candidate was a mixture of calcium acetate and magnesium acetate. Further, they noted that: "On two counts, corrosion inhibition and soil building, far from being detrimental they (calcium and magnesium acetate) are beneficial. For this reason, it is hardly accurate to compare these salts with sodium chloride on the basis of materials cost alone for the equivalent deicing performance. A proper comparison must include as well the attendant cost (or benefit) to the taxpayer's person and property (including that which he owns through the state) resulting from the application of the respective deicers. On this basis, the calcium and magnesium salts of the lower organic acids, instead of five times as costly, may prove to be half as costly as sodium chloride on an equivalent deicing bases.“ Calcium Magnesium Acetate (CMA) is a mixture of calcium acetate and magnesium acetate. It is a noncorrosive, nonpolluting, roadway deicer, which can be made by combining acetic acid with dolomitic lime (CaO-MgO) and magnesia (MgO). Tests have shown that mixing magnesium acetate and calcium acetate in a molar ratio of 2.3:1 yields the lowest eutectic temperature (approximately -35 degrees Celsius), thus creating the optimum.mixture for use as a deicer. 19 CMA has shown such promise in repeated testing that it has been selected in FHWA-sponsored research as the most viable substitute for NaCl. In 1985, 24 states became involved in the "CMA Research and Development Effort". This program consists of a number of parallel research projects which address three main concerns. The first concern is determination of the environmental acceptability of CMA. Next, the FHWA desired to study the development of technology to manufacture CMA from non—oil and non-gas feedstocks. Finally, the evaluation of technical benefits and problems with CMA use along with its economic feasibility were studied. The results of the CMA Research and Development Effort (along with other active studies) have given insight to its potential use. Data has been gathered in the areas of health, environmental concern, economic feasibility, and other assorted aspects of its use. This data has shown that while drawbacks to CMA exist (predominantly in economics) the advantages are widespread. First, while health considerations are a growing concern with the use of NaCl (as detailed above), CMA is 20 effectively harmless to human life. Neither calcium nor magnesium have been identified as being harmful upon ingestion. Acetate also has no deleterious effects, and in fact is frequently consumed in the form of vinegar. CMA also has advantages over NaCl in environmental issues. Theoretically, the acetate anion degrades to form carbon dioxide and water, which do not cause adverse effects to plant or invertebrate ecology, as the chloride anion from NaCl does. Further, it has been predicted that since CMA use as a deicer occurs at low temperature, the evolution of C02 and H20 will not effect the biological oxygen demand in lakes or streams. Also, the cations of CMA cause no damage because calcium and magnesium form carbonates which precipitate and therefore should have no effect on water density or lake turnover rates (13). Field tests have been run to corroborate these claims. For example, bio—assays of ten northern California lakes revealed that addition of varying levels of CMA showed no response in algal biomass in eight of the ten lakes. Only in one instance did the addition of high amounts of CMA result in any notable change (an increase in phosphorous uptake was observed) (14). Thus, it seems that CMA is in no way detrimental and in fact, it exhibits certain traits which benefit the environment. One such characteristic is that its widespread 21 use may alleviate the effects of acid rain by forming hydroxides (of Ca and Mg) which act as buffers to neutralize sulfuric and nitric acids in the environment adjacent to roadways. Also, CMA has the reverse effect on roadside vegetation as that of NaCl in at least one area. As discussed earlier, sodium causes the breakdown of soils, resulting in compaction. This compaction lowers the permeability of the soil to air and water making it difficult for nutrients to reach plant roots. In contrast, the divalent calcium and magnesium ions tend to reverse this effect (12). Perhaps the most direct advantage of replacing NaCl with CMA is its ability to act as a deicer without instilling the negative efiects of corrosion. Corrosion occurs in an electrolyte solution via two reactions. The first is the oxidation of iron at the anode and the second being the reduction of hydrogen at the cathode. 1) Anode: Fe ==========> Fe+2 + 2e“ 2) Cathode: H+ + e‘ ==========> 1/2H or 2H+ + 1/202 + 2e' =======> H20 These reactions result in the rusting of steel alloys as the iron component in them deteriorates. If, however, the pH of the electrolyte solution is high, a relative lack of 22 hydronium ions must exist. Thus, the cathodic reaction cannot proceed and corrosion will be decreased. Solutions of sodium chloride exist at a neutral pH, while both calcium acetate and magnesium acetate are basic salts. Thus, solutions of CMA contain fewer H+ ions than comparative solutions of NaCl. Theoretically, this gives CMA a corrosion inhibition character (8). Experimental data has been acquired to support this theory. One study, for example, immersed steel strips into solutions containing various concentrations of CMA or NaCl for two weeks. The results revealed that solutions with concentrations of CMA ranging from 1.4 to 13% gave increasingly less corrosion (with a minimum of 30% less) than a 7% solution of NaCl. (measured by the per cent weight loss of steel). Further, pure solutions of both calcium acetate and magnesium acetate showed zero and “trace” amounts of corrosion respectively (4). A second set of experiments conducted on most of Michigan's currently used highway metals revealed that, "a qualitative observation of the CMA and NaCl specimens from only several hours of exposure onward have consistently indicated considerably worse corrosion occurring in the NaCl environment... and at no time during the total 12 months of exposure was there any indication that the CMA environment was as corrosive as the NaCl environment" (15). 23 The conclusion is that CMA will not corrode exposed steel in vehicles or bridges. Another significant study compared the corrosive characteristics of CMA and NaCl on the reenforcing "rebars" of steel which are imbedded in most concretes. These tests showed that rebars imbedded 2.5 inches from the surface of concrete blocks were uneffected when submerged in solutions of CMA for 90 days. Comparative tests with NaCl showed substantial corrosion (measured by electrical 1/2 cell potentials of the steel) (12). The conclusion of these tests was that CMA is far less corrosive to exposed steel and concrete than NaCl. In fact, some tests have shown CMA to be a corrosion inhibitor. One final advantage CMA offers is that it appears to retain its deicing capabilities long after the effects of NaCl have worn off. It was observed that when snow was compacted atop limestone-based cement, CMA or NaCl was applied and then washed off, and compacted snow was reapplied, the NaCl treated specimen required over twice the shear stress to remove the reapplied snow. This was due to the fact that CMA was detected in the concrete at "noticeable depths", creating a hypothesized reservoir of CMA which aided the deicing in subsequent snow applications. Conversely, NaCl was only found in trace amounts below the limestone surface (16). 24 Unfortunately, CMA has one major drawback, its cost. The 1991 cost of CMA exceeded $640/ton which is approximately 29 times that of NaCl. Also, due to its higher molecular weight, 1.8 pounds of CMA are required to produce the equivalent deicing effect as one pound of NaCl. This raises the cost of CMA to nearly 50 times that of NaCl for an equivalent deicing capacity. This cost is increased further because CMA is less dense than NaCl. Thus, a larger volume must be stored, transported, and applied to produce the desired effect. In an attempt to reduce these costs, NaCl/CMA mixtures have been studied. Results of such studies show that a 0.46:1 CMAzNaCl mixture results in a product that has "almost as good“ corrosion performance as pure CMA. A mixture of this composition would lower the cost of the deicer to approximately $240/ton. Also, further testing may reveal that a mixture containing even less CMA is adequate (15). 25 In the mid 70's, it was estimated that the damage done by sodium chloride usage as a road deicer cost between six and twenty times that of the salt itself. The lower estimate considered only salt accelerated corrosion damage to vehicles, bridges, guardrails, sign structures, and reenforcing steel in concrete; while the higher estimate took environmental issues into consideration as well. Since this time, the cost of NaCl has remained relatively constant while other costs have risen. Thus, salt related damage today is estimated to be between twelve and forty times its COSt . This enormous indirect cost of NaCl prompted the Federal Highway Administration to sponsor research to find a suitable deicing substitute. This substitute had to fulfil two criteria. First, it needed to alleviate many (preferably all) of the environmental dangers and concerns brought on by NaCl. Second, it had to be producible for less than the monetary amount of damage done by NaCl. As a result of this research, calcium magnesium acetate (CMA) was chosen as the most suitable, substitute deicer. Past and continuing studies on CMA show that its cabability to prevent corrosion, reduce health risks, and aid many environmental concerns, warrants its use as a road 26 deicer. The only drawback to 100% replacement of NaCl with CMA is its economic feasibility. Current methods of production result in a market price for CMA which is approximately 29 times that of NaCl. Thus, if CMA is to be used on a large scale, current production methods must be improved upon, or new methods must be developed. Current production techniques for CMA (from combining acetic acid with dolime) warrant prices many times that of NaCl. This high initial cost must be weighed against the long term benefits of lowered corrosion and health and environmental safety when determining when and where to use CMA. Clearly, unless health and environmental concerns become more serious, or the production cost of CMA is drastically reduced, it is unrealistic to predict that CMA will ever totally replace NaCl. However, it may be economically wise to use CMA in isolated instances where economic or health concerns dictate. One such instance is the case of new and expensive highway structures where use of CMA could add years to the life of a structure by decreasing corrosion. The state of Michigan has already employed this strategy with respect to its new zilwaukee Bridge. The state determined that the life span of the 135 million 27 28 dollar structure could be dramatically increased by using CMA on the bridge and on one mile stretches of the adjoining highway (this would prevent NaCl from being carried onto the bridge from vehicles). After two years of use, highway officials have reported that CMA has been at least as effective as NaCl in preventing hazardous driving conditions, and has in fact shown a residual effect which helps prevent icing in subsequent storms (17). Due in part to this success, the "Van Regenmorter Bill” has been submitted to the Michigan Legislature. The provisions of this bill state that an advisory committee shall be formed to coordinate a series of studies by Michigan's four state departments that share responsibility for the different aspects of highway deicing (the departments of: Transportation, Commerce, Natural Resources, and Management and Budget). This committee will then site areas of the state which are particularly "salt sensitive" (whether it be due to environmental concerns such as high sodium concentrations in the water supply, or economic reasons such as seen with the Zilwaukee Bridge). Finally, the committee will submit a plan providing for an alternative deicing phaseein period for these highly sensitive areas (18). The state of Massachusetts has also experimented with CMA on a large scale due to its drinking water contamination problem. The state began using CMA in designated areas as 29 early as 1983 and has continued to record data on its use throughout chosen regions in the state. Although this data is yet inconclusive, many localities show promising early results. Exclusive use of CMA form 1987 through 1990 in the cities of Lakeville and Freetown, for instance, resulted in a reduction of sodium levels in local wells from 75 mg/l to 34 mg/l. Similar results were observed in Goshen from 1983 through 1988 (19). Field testing of CMA has also been conducted in Iowa and Washington. Further, in 1987, New York State began studies to determine the environmental impact of large scale CMA use (20). This ever increasing use of CMA, along with pending legislation such as the Van Regenmorter Bill has prompted Chevron to begin production of CMA for commercial use 0 Currently, Chevron produces a CMA product called "Ice— B—Gon" which is made from glacial acetic acid, dolomitic lime (CaO-MgO), and magnesia (MgO). The product's 1991 market price was 30 cents per pound. Reportedly, most of the manufacturing costs came from production of glacial acetic acid (which lists for 29 cents per pound) (21). 30 Current supplies of dolomitic lime and limestone across much of the United States (including the Snow Belt region, where most of the CMA manufacturing would take place) are high enough that their availability for CMA production is not viewed as a problem. Acetate sources, however, could be potentially lacking. If, for instance, CMA were to replace a mere 10% of the NaCl used nationwide (at a substitution ratio of 1.5:1.0 weight percent CMA to NaCl), the demand for acetic acid would be approximately one million metric tons per year. This is over 75% of the 1980 national consumption. To accommodate such a large scale use of CMA, the FHWA specified that the size range for CMA production facilities would have to be between 100 and 1000 tons per day. This would require acetic acid feedstock supplies of 240,000 tons per year. Thus, each CMA plant would require its own acetic acid production plant (3). Current production methods of glacial acetic acid require large quantities of petroleum. The FHWA however, has specified a desire to produce CMA from feedstocks other than petroleum or natural gas. This requirement stems from the national concern to conserve these fuels, along with the possibility of highly fluctuating oil and gas prices in 31 times of crisis. One potential substitute feedstock is coal. Coal can be used to produce methanol and carbon monoxide, which can then be converted to acetic acid using the Monsanto methanol carbonylation scheme. Although large scale plants using coal to produce methanol are rare, they do exist. Tennessee Eastman for example, used Texaco coal gasification technology to produce over 200,000 tons of methanol per day from 1,600 tons of coal. Although this plant is not used for acetic acid production, similar facilities of its size could support the FHWA's suggested CMA production (3). Although production of acetic acid via chemical routes such as the Monsanto process is feasible, a second production route which is potentially more cost effective is being investigated in great detail. This method produces acetic acid by using microorganisms in conjunction with renewable biomass (wood or field crop residues). Acetic acid production by such fermentation processes has been used in the vinegar industry for quite some time, and the possibility of using similar processes for CMA production is very promising. Corn, which is in great excess in the United States, can be used as the feedstock for these acetic acid fermentations. Technology for this method of acetic acid production has been available for years. In the first step of the process, glucose (which is acquired by hydrolyzing 32 starch in the corn) is used as a substrate for S. cerevisae, a strain of yeast, and is converted to ethanol and carbon dioxide. The resulting ethanol is then fermented by the bacterium Acetobacter aceti which produces acetic acid and water. S. cerivisae l) C6H1206 :::::::::=:::> 2C2HSOH + 2C02 A. aceti 2) 2C2H5OH + 202 =============> 2CH3COOH + 2H20 E 2 1 . . E . : . .3 i . Production of acetic acid from fermentation is far more involved than described above. It has a long and storied history dating back at least seven thousand years to the time of ancient Babylonia when vinegar (of which acetic acid " is the main constituent) was first fermented from wine (22). Almost certainly, vinegar was initially considered an annoyance whose taste signalled the spoilage of wines and 33 beer. Over the centuries however, it was discovered that vinegar had many advantageous properties. It was, for instance, one of the first known antibiotics, with its medicinal use dating back to 400 B.C.. Vinegar has also been used for thousands of years as a food preservative and a condiment. Also, its value in the sauces of fine French cooking is immeasurable. The first known commercial production of vinegar originated in Orleans, France in 1670 (22). This production method, which is still used in some parts of the world today, was termed the "Orleans" or "Slow“ process. The technique used for production was quite simplistic, whereby a 50 pound wooden barrel was filled with low grade wine and fresh vinegar. This fresh vinegar, which accounted for approximately 20% of the volume, served as an inoculum for the batch. Holes were then drilled in the barrel for oxygen transfer and product removal. After approximately two months, a semi-batch exchange was made with fresh wine and a portion of the newly formed vinegar. This basic process has remained unchanged over the past three centuries. The length of time required for the Orleans process inspired development of the "Quick Process”. The initial ideas of this improvement were conceived in 1823 by a German chemist named Schutzenbach (23). He hypothesized that since vinegar production was highly dependent on exposure to air, 34 the production rate could be increased if a method to raise the contact surface of the vinegar stock with air could be found. The method he proposed was to trickle the vinegar stock over twigs or wood shavings. It was found that this technique drastically reduced the time required for a batch fermentation. The ensuing 120 years saw a great number of improvements to this quick or "trickle" process. Among these were refinements in the method of recirculation, temperature control, continuous and semi-continuous methods, and changes in the packing material used. These modifications have yielded an improved quality of vinegar with a marked increase in both the speed of conversion and the final acetic acid concentration of the product. Trickle method vinegar plants today generally use about 2,000 cubic feet of beechwood shavings as a packing material and produce 500 to 1,000 gallons of 100 grain vinegar a day (with batch sizes varying between 2,000 and 4,000 gallons) (23). A third process for the making of vinegar has been developed in the twentieth century. This process, termed "Submerged Fermentation", again shows an improvement to the Orleans method by increasing the contact area of the fermentation broth with oxygen. The basis behind this method of vinegar production, which originated in 1923, was to force air through a tube into the bottom of the 35 fermentation vessel. Air would then be bubbled through the fermentation broth, allowing for oxygen exchange throughout the batch. Studies and improvements to early submerged fermentations were made for several years before the process became economical. These studies led to the development of the Frings Acetator in the early 1960's. The design of this acetator allowed for control of such paramaters as temperature, pH, and oxygen transfer rate. This transfer rate is maximized via the inclusion of a device called a cavitation pump. The pump was designed to, "bring air effectively in solution in the vinegar mash and to supply the oxygen needed for the biological conversion of alcohol to acetic acid.“ It was an important improvement because of observations that merely bubbling the air through the mash was not effective. Figure 1 shows a schematic drawing of the Frings Acetator used in submerged culture vinegar production today (23). Continuous development of the submerged fermentation process has made it the most efficient method of vinegar production. Currently, such fermentations are approximately ten times more productive than comparable trickle bed processes. 36 FIGURE 1: The Fringe Acetator. 37 Vinegar production has grown over the years into a multi million dollar, world—wide industry. This has prompted detailed investigation of both the processes used and the biological mechanisms which convert ethanol to acetic acid. Improvements to these processes are constantly being made so that both the production rate and final concentration of acetic acid are increasing. Hopefully, these improvements, along with studies of novel production methods (such as that presented later in this paper) will provide a more economical method for producing acetic acid and ultimately CMA. 3 2 1 1. . E 1 . .1 l . Acetic acid bacteria are polymorphous with cells that are ellipsoidal to rod—shaped, and straight or slightly curved. These cells are 0.5 - 0.8 by 0.9 - 4.2 um and can occur singly, in pairs, or in chains. There are both non— motile forms and motile forms with polar or peritrichous flagella. They are obligatorily aerobic and some produce pigments while others produce cellulose (24). It was observed as early as 1822 that when vinegar was produced via the Orleans method, a film seemed to form on the surface of liquids undergoing acetic fermentation. This film, called "mother of vinegar", was not proven to be living matter until 1867 when Pasteur showed that it was 38 responsible for absorbing oxygen from the air and producing acetic acid during growth on an alcoholic medium. He termed this organism Mycoderma aceti. Over the next three decades, various researchers showed that acetic fermentation could in fact be brought about by several species of bacteria, and in 1901 Beijerinck first suggested the name Acetobacter (25). Today, well over 60 Acetobacter species have been identified. Each species has its own favored growth conditions, yet the basic mechanism for converting substrate to acetic acid is the same for each. It consists of an oxidative fermentation in which diluted solutions of ethanol are oxidized with oxygen from air to produce acetic acid and water. The mechanism for this oxidation proceeds through a two-step process, with the major intermediate being acetaldehyde. 1) CH3CH20H + 1/2 02 ==========> CH3CHO + H20 2) CH3CHO + 1/2 02 ==========> CH3COOH A simultaneous pathway occurs in many of the acetobacter species where two moles of acetaldehyde react to form one mole of ethanol and one mole of acetic acid. 2 CH3CHO + H20 ==========> CH3CH20H + CH3COOH 39 The mole of ethanol is then again oxidized via the first pathway (26). Most of the acetic acid bacteria also have the ability to metabolize several carbohydrates and sugars. Included here is the ability to convert lactate to acetate via a three step mechanism with pyruvate and acetaldehyde being intermediates (it should be noted here that most strains utilize D-lactate at least four times faster than L- lactate). Figure 2 shows a detailed pathway of the major groups of acetic acid bacteria (27). For a more comprehensive description of acetic acid bacteria catabolism see reference 28. The genus acetobacter has grown so large that taxonomical classification of the subspecies has been desired. Unfortunately, the similarities of these' subspecies overlap with one another so frequently that this classification has been a difficult, if not impossible task. The first attempt at classification was by Asai in 1934. He divided the acetic acid bacteria into two genera. The first retained the name Acetobatcer, while the second was termed Gluconobacter. The distinguishing characteristic of this new genus was that it produced acetic acid from glucose only, not from ethanol. E. w DRAW M 1.1.... 141/ .. I 11...} .1. . mum mubelhuelthemm m W pad-Ilene!“ m~m~w11.humu&£) m ' 3 'mflum‘ E , /\. . .1... Hate Me \ m— w m ‘ 2.5-daze M” /. least; flung-to. J, M—OW—flm Wu tact-W of ARM (M “It MJWMHMMJzfirme-JW’“” 1W FIGURE 2: Detailed metabollc pathways of verIoue acetlc add bacterla. 41 Asai's division drew little argument until 1950 when Frateur realized that further classification was needed. He used five criteria to divide the original genus Acetobacter into four subgroups. These criteria were: presence of catalase, ”overoxidation" of ethanol through acetic acid to carbon dioxide and water, oxidation of lactate to carbonate, oxidation of glycerol to dihydroxyacetone, and production of gluconic acid from glucose. Using these criteria, Frateur placed every known species into one of the following groups: ,peroxydans, oxydans, mesoxydans, and suboxydans. Four years later, Leifson proposed to split Frateur's four groups into two separate genera. The distinction would be a combination of the ability to oxidize acetate and the type of flagellation. Leifson's first genus, Acetobacter, was peritrichously flagellated and could overoxidize acetate. It consisted of three of Frateur's subgroups: peroxydans, oxydans, and mesoxydans. Acetomonas were to comprise the second genus (consisting of Frateur's suboxydans) which were polarly flagellated and could not overoxidize acetate. Unfortunately, a 1959 study by Shimwell showed that nearly every species of Leifson's Acetobacter bacteria, produced progeny which could acquire different characteristics than the parent culture. Occasionally, these progeny even lost the ability to oxidize ethanol to 42 acetic acid. This prompted much debate in the microbiology world, so that some retained the Acetobacter/Acetomonas distinction while others reverted back to the Acetobacter/Gluconobacter classification. The latter opinion was adapted by the 8Lh edition of Bergey's Manual which defines Gluconobacter as a genus that has polar flagellation or none, does not oxidize acetate and lactate, and oxidizes glucose to gluconate and 1— and 5- ketogluconate. Acetobacter was defined to have peritrichous flagellation or none, oxidize acetate and lactate, and have the enzymes of the tri-carboxylic acid (TCA) cycle. Work by Gillis and De Ley in 1980 showed that these two genera were more closely related to each other than to any other genus. Thus, they proposed the formation of the family Acetobacteriacae which would be comprised of the two genera Acetobacter and Gluconobacter. This work, which identified each species by C-l4 labelled RNA denaturation studies, along with related studies seems to be producing a classification system which will finally identify all species and eliminate the many misunderstandings found in the literature both past and present. After reviewing the above history of the acetic acid bacteria, it should be easy to see why there is still great confusion over the past and present naming of individual species. Some strains have acquired the new, proper names 43 while others have retained names from older classification systems. An unscientific (and strictly speaking, incorrect) distinction between the species of these bacteria is as follows: Those species retaining the name "Acetobacter" can oxidize ethanol to acetate but do not oxidize glucose. They also have enzymes of the TCA cycle and therefore can overoxidize the acetate to carbon dioxide and water. "Gluconobacter" species do not oxidize ethanol, but they can oxidize glucose to gluconate. Further, they lack the TCA cycle enzymes and cannot overoxidize acetate, lactate, or gluconate. Species termed "Acetomonas" (a classification which may soon be eliminated) have characteristics of both previous classifications. Depending on the particular species, they can oxidze ethanol, glucose, or lactate. Acetomonas, however, have either no ability to overoxidize acetate, or a reduced ability to do so. This informal classification may aid the reader when researching various acetic acid bacteria (it certainly helped me!). It should be stressed however that this classification is very general and by no means should it be considered complete or correct in a strict sense. Modern day vinegar production is conducted on a large scale by companies throughout the world. These vinegar producing facilities, whether they use modern submerged fermentations or the ancient Orleans method, employ various Acetobacter strains to oxidize alcoholic substrates and create acetate. Unfortunately, several drawbacks are incurred by the use of these bacteria. One disadvantage with the use of Acetobacters is that they posses the TCA cycle enzymes. Thus, they are able to use acetic acid as an intermediate in metabolism. This of course results in loss of a significant amount of the product as the organisms consume acetic acid for energy. Another drawback to the current production method stems from the fact that all acetic acid bacteria are dependent on large supplies of oxygen. In fact, as stated earlier, this was the impetous behind the development of the "quick process" and "submerged fermentation" as methods of vinegar production. These processes expose the fermentation broth to large amounts of air, via trickling over wood chips or bubbling. Unfortunately, the substrate used for vinegar production today is ethanol, which is highly volatile. This 45 high volatility causes a significant loss of potential acetic acid production as substantial amounts of ethanol are lost in the contact with air. Various methods, such as employment of condensation columns and gas recycle systems, have been used to decrease this loss, but they are expensive and are not 100% effective. A goal of the research presented in this paper was to eliminate these two drawbacks from the current process of acetic acid production. Attainment of this goal would make the proposal of producing CMA via a microbial fermentaion more realistic. To eliminate loss of substrate due to the high volatility of ethanol, it was proposed to change the organism used from an Acetobacter to a Gluconobacter (or Acetomonas) species. This would allow the use of non- volatile substrates such as glucose, lactose, or lactate in place of ethanol. The use of Gluconobacter species would simultaneously solve the overoxidaton problem since the bacterium.would lack the enzymes required to degrade the acetate product. This proposed change in the production of acetic acid would have an added bonus in that it would provide a use for the huge surplus of corn that currently exists in the United States. 46 MW Once acetic acid is produced, it must somehow be combined with calcium and magnesium in the proper concentrations to make CMA. Three major processes for this reaction exist. First, dolomitic lime (CaO-MgO) can be combined with acetic acid to form calcium acetate, magnesium acetate, and water. Second, hydrated lime (Ca(OH)2°Mg(OH)2) can be combined with acetic acid to form calcium acetate, magnesium acetate, and water. Finally, dolomite (CaCO3-MgCO3) can be combined with acetic acid to form the acetates of calcium and magnesium, water and carbon dioxide. 1) CaO'MgO + 4CH3COOH =======> (CH3COO)2Ca + (CH3COO)2Mg + 2H20 2) C3(OH)2‘MQ(OH)2 + 4CH3COOH =====> (CH3COO)2Ca + (CH3COO)2Mg + 4H20 3) CaCO3°MgCO3 + 4CH3COOH ==> (CH3COO)2C& + (CH3COO)2Mg + 2H20 + 2C02 As previously stated, an ideal mixture of CMA would show a magnesium acetatezcalcium acetate molar ratio of 2.3:1.0. Mixtures containing lower magnesium acetate concentrations would probably still be acceptable for highway deicing, but will exhibit higher eutectic points, and therefore be less effective at extremely cold temperatures. 47 It has been shown that both dolomite and dolomitic lime readily dissolve in concentrated acetic acid. Unfortunately, acquiring solid CMA pellets from the resulting solution requires the removal of excess water via drying. This evaporation process yields a CMA containing traces of free acetic acid which in turn, causes the effectiveness of CMA on concrete to become virtually non- existent. Raising the pH of the CMA solution to approximately 9.0 before drying can eliminate this problem because nearly all of the free acetic acid (which has a pKa of 4.8) will be converted to acetate. This alkaline pH would also help to keep acetic acid losses due to evaporation at a minimum (12). To help eliminate contamination of the CMA with other acetate salts, it would be advantageous to use dolomite as a pH control during the actual fermentation (as opposed to bases such as NaOH that could dissociate and form these other salts). Unfortunately a problem exists with this method. Dolomite is virtually insoluble in alkaline solutions, and its rate of dissolution is a negative exponential function of pH. That is, for every decrease of one pH unit, there is a tenfold increase in the dissolution rate. It has been shown that even finely pulverized dolomite (100 mesh) requires a pH less than 6.0 before any substantial carbon dioxide is evolved. Thus, the bacterium used for fermentation must be viable and able to produce 48 acetic acid at'a pH of 6.0 or below if dolomite is used as a pH control (3). Light-burned dolomitic limestone gives more flexibility as a pH control, as it is more reactive than dolomite. As the pH increases above six, however, the reactivities of CaO and MgO begin to diverge. CaO continues to dissolve rapidly until a pH of 12 is reached, while MgO becomes inert at pH values above 6.0. Thus, if dolomitic limestone is used in systems with pH control above 6.0, the resulting solution would be that of calcium acetate, not CMA. It has been seen that the acetates of calcium and magnesium crystallize quite differently upon drying. Pure calcium acetate forms tangled, dendritic clusters of microscopic needles which cling tightly to each other. It also exhibits a negative temperature coefficient of solubility. Conversely, magnesium acetate has a positive temperature coefficient of solubility and crystallizes as a tetraydrate. It is, however, difficult to form these crystals by simply drying a pure magnesium acetate solution, as they do not self nucleate. Instead, an amorphous glass forms which becomes sticky, apparently from absorbing moisture out of the air. Equimolar solutions of calcium acetate and magnesium acetate can result in either a uniform or nonuniform crystal 49 structure depending on the methodology used to dry the solution. Tests have shown that equimolar solutions can crystallize and form platelets, under the proper conditions, instead of either the dendritic or amorphous shapes of pure calcium or magnesium acetate. These platelets (which have a 3:1 mole ratio of magnesiumzcalcium) only form.when the drying process is very rapid. If the drying process is slow (for example, if a large pool of the solution is set out to dry for several months), the acetates tend to crystallize separately. This causes nonuniformity of the resulting solid, with some regions containing the dendritic calcium salt and other regions existing solely of the amorphous magnesium acetate crystals (3). The potential drawback of a slow drying process is that the resulting solid may need to be crushed and subsequently mixed to assure that a uniform blend of the acetates is acquired. Large scale fast drying methods, such as drying in a rotary steam heated drum, are initially more expensive, but avoid these added steps and are probably more economical in the long run. Another advantage of a fast drying method is that the resulting flakes can be designed to have any degree of coarseness by altering the drum speed. Also, energy (and therefore money) can be conserved if the drum temperature is lowered, allowing some of the CMA to be left in a hydrated form. This strategy, however, must be carefully scrutinized as there would be an economic trade- 50 off between the cost of removing the water of hydration and the cost of shipping and handling excess water in the hydrated CMA. Thus, the formation of CMA from acetic acid is extremely sensitive to, and dependent upon, pH and the methods used for crystallization and drying. Light-burned dolomitic limestone should be used as a pH control for fermentations which contain bacteria that can survive and produce acetic acid in an acidic environment of pH not exceeding 6.0. Also, the crystallization technique should be chosen after an economic analysis of the shipping/handling procedures and the necessity of acquiring a uniformly mixed CMA is conducted. A fast drying method such as a rotary steam-heated drum system is the most flexible and probably the most economical method for large scale production. The production of acetic acid via bacterial fermentation is by no means a new concept. However, using bacteria which are not true Acetobacters, and using lactate (or lactic acid) as the substrate is a novel route for such a fenmentation process. Thus, any information acquired regarding the bacteria's growth patterns and production rates would be used to build a data base. The goal of this project was to create a strong data base which could then be used to determine if an industrial scale fermentation would be economically feasible for the production of acetic acid and ultimately CMA. Ideally, a positive conclusion would be reached, and future research would optimize the process developed here and employ it at the pilot plant level. To make a competent judgement on the validity of using this fermentation route, a systematic research plan had to be developed. The following is a synopsis of that plan. The first step was to screen several strains of the most promising bacteria (chosen from a literature review of the acetic acid bacteria) and choose the two best for 51 52 further study. Determination of these species would depend upon their ability to completely convert lactic acid to acetic acid at high substrate concentrations. The rate of this conversion would also be taken into consideration. After the pool of potential acetic acid producers was narrowed by the above screening, it would be necessary to optimize the conditions for growth and production. To accomplish this task, an attempt to optimize the growth medium would have to be made. First, tests using lactate as the sole substrate (in conjunction with constant levels of peptone and yeast extract as nitrogen sources) would be performed to determine substrate inhibition levels. These initial tests would then be followed by similar studies using glucose, lactose, corn steep liquor and combinations of the three as supplements and partial supplements for lactate. The goal of these studies would be to learn the approximate levels of lactate (with or without amounts of an added secondary metabolite) which would yield the optimum growth and production rates. As previously stated, an industrial process for the production of CMA using dolime as a pH control would require operation at a pH below 6.0. Thus, the next step would be to gather information regarding the effects of pH on the fermentation. Further optimization would require studies to 53 determine the range of acceptable operating temperatures, as bacterial growth is also highly dependent on this parameter. Tolerance of the bacteria to high concentrations of calcium and magnesium would then have to be determined. These experiments would show if dolomitic limestone could be used as a pH control during the fermentation. If the bacteria failed to remain viable, or if it showed substantially reduced acetic acid production in the presence of either ion, the final process would require a post- fermentation addition of dolomitic limestone to create CMA. Finally, product inhibition studies would be conducted. The results provided by this set of experiments would give insight to the possibility of using cell recycle and/or continuous fermentation systems. Results of the aforementioned experiments, conducted in 250 ml shake flasks, would be used to determine which bacteria would be studied at the next level. The same results would also determine many of the operating conditions needed to optimize the bench scale process. Selection of the best-choice bacteria, and fermentor operating conditions would be followed by bench scale studies in a five liter fermentor. These experiments would add a new dimension to the data base by studying the effects of such variables as feed substrate concentrations, power 54 input to the fermentor's impeller, oxygen input via a sparger system, and feed dilution rate. These variables would be studied in batch, semi-batch, and continuous mode fermentations to determine which system would yield the best acetic acid production rate. Finally, an economic review of the optimum process would be made and compared to that of currently available production techniques. A recommendation will then be made as to whether this fermentation route warrants consideration at the pilot plant level. A literature review of the bacterial strains Acetobacter, Acetomonas, and Gluconobacter produced four bacteria which appeared to be potential candidates for CMA production. A freeze-dried culture of each of these was acquired from.the American Tissue and Culture Collection (ATCC). These strains and their accompanying ATCC numbers are as follows: Acetobacter pasteurianus, subspecies orleanensis (9432); Acetobacter cerinus (12303); Acetomonas oxydans (19357); and Gluconobacter oxydans, subspecies suboxydans (33448). The freeze-dried cultures were asceptically revived, near a flame, and grown for 36 hours in 10 ml culture tubes containing the ATCC recommended growth medium. This sterilized growth medium contained: 5 grams yeast extract (Difco), 3 grams bacto-peptone (Difco), and 25 grams D- mannitol (Sigma) per liter distilled water. Continuous 55 56 agitation was applied for the entire period by placing the culture tubes in a Lab-Line "Orbit Shaker Bath" at 220 rpm and 26 degrees Celsius. The bacteria were then asceptically transferred to slants (consisting of the same growth medium) via a loop and refrigerated for storage. To prevent bacterial death from "drying out" of the slants, each culture was transferred to fresh slants every three months. Also, as a precaution, a sample of each bacteria was freeze— dried and stored by The Michigan Biotechnology Institute. Prior to every experiment, the desired culture had to be revived from its refrigerated storage. First, the appropriate slant was removed from the refrigerator and allowed to reach room temperature. Next, a loop of the bacterium was transferred to a culture tube containing 10 ml of the ATCC growth medium. This culture was then grown overnight (approximately 12 hours) in the shaker bath at aforementioned conditions. At this point, the culture (which has taken on a cloudy appearance due to the increased bacterial population) could be used as an inoculum for the 57 desired experiment. Each experiment used a 10% inoculum (by volume) of a culture prepared in this manner. This and all subsequent procedures was performed under asceptic conditions, with media which was autoclaved for 15 minutes at 121 degrees Celsius. Also, unless otherwise stated, all shake flask experiments were conducted at 26 degrees Celsius, with constant revolutions per minute (rpm) of 220, and were titrated with H2804 to a final pH of approximately 5.5. These experiments were conducted in 50 ml of medium placed in 250 ml shake flasks in the orbital shaker bath noted above. Determination of substrate, intermediate, and product concentrations was accomplished by high performance liquid chromatography (HPLC analysis). The system used consisted of an ISCO model 2350 pump with an ISCO V4 Absorbance Detector. The chosen column was a Bio-Rad Aminex Ion Exclusion HPX—87H column with a Bio-Rad 125-0129 guard column. This system was connected to an Anspec AN-728 58 autosampler and a Waters 745 Data Module/Integrator for interpretation. Sample preparation consisted of the following steps. First, the sample was diluted with deionized water to a point such that the chromatogram results would be "on- scale". Sulfuric acid was then added to reduce the pH to levels between one and three (this step was necessary to prevent damaging the column). Samples were then centrifuged at 10,000 rpm for ten minutes, and one ml of the supernatant was passed through a 0.45 um filter to remove any particulate matter. The samples, along with several standards, were then placed in the auto sampler and run through the column at a flow rate of 0.4 ml/minute. The mobile solvent consisted of 0.005 M H2SO4 and the column was heated to 60 degrees Celsius to aid in the separation. Chromatographs were then analyzed by the integrator and concentrations of lactic, pyruvic, and acetic acids were calculated. On occasion it was necessary to store samples for several days before analysis. This was accomplished by first adding 3 N H2304 to the sample in a 1:4 ratio (this lowered the pH for HPLC analysis and prevented further metabolism by the bacteria). The samples were then frozen until ready for analysis. To determine if this process 59 would alter the results, test samples were frozen, thawed, run on the column, and then compared to unfrozen duplicate samples. At no time did this procedure appear to yield obscure results. 60 EM T Optical density readings were determined by measuring the absorbance of visible light at 640 nm. These measurements were conducted on a Varion 634 Spectrophotometer using the double beam setting with the ATCC growth medium as an internal standard for each measurement. Whenever necessary, samples were diluted with a stock of this medium (cell free) to keep the reading on scale. Dry weight measurements were only conducted for the bench scale experiments, where it was possible to acquire large sample volumes without altering the experiment. Fifty ml samples were taken and placed in centrifuge vials. The samples were then spun at 7,000 rpm for one hour and removed from the centrifuge. The cell free supernatant was 61 carefully poured off, and the remaining wet pellet was removed and placed in a drying oven over night (approximately 12 hours). Finally, the dried pellet was weighed on an analytical balance for dry weight determination. The specific growth rate can be determined from the slope of a plot of ln(cell concentration) vs. time during the exponential growth phase. Cell concentration was found to be proportional to absorbance at 640 nm. Thus, a plot of ln(absorbance) vs. time will also yield the specific growth rate (29). Samples of the growing culture were taken at 30 minute intervals. Optical density was then measured as explained above, and absorbance was plotted against time. This plot showed the various phases of bacterial growth (lag, exponential, stationary, and death). A plot of ln(absorbance) vs. time was then constructed. The specific growth rate was determined to be the slope of the linear portion of this second curve. Slopes were determined by 62 applying a linear regression program to the appropriate points on the plot. Figures 3 and 4 show examples of these curves for growth of A. oxydans on 50 mM lactic acid, 0.5% peptone, and 0.5% yeast extract. ABSORBANCE AT 640 nm LN (ABSORBANCE) AT 540 nm 63 +05% PEPTOhE/YEAST EXTRACT TIME (hoLirs) ' FIGURE 3: Absorbance vs. time. A VICW of the phases of bacterial growth H—osx PEPTmE/YEASI' EXTRACT |§ 4 . TWIE (hours) FIGURE 4: h (absorbance) vs. time. Another View of bacterial growth. 64 Bench scale experiments were conducted with two New Brunswick Scientific "Bio-Flo IIc Batch/Continuous Fermentors". The capacity of these fermentors was 2.5 and 5.0 liters respectively. Each had the capability of using the following features: pH control via measurement by an Ingold 465-90 Steam Sterilizable pH meter, internal acid and base pumps, and input ports; dissolved oxygen control via measurement by a Phoenix galvanic oxygen probe and subsequent internal control of the impeller rate; temperature control using a shell and tube heat exchanger with continuous water flow; antifoam control via a level sensor accompanied by an internal pump and input port (the antifoam used was "Antifoam A" acquired from Sigma); variable speed motor for agitation via an impeller; sparger for the input of any desired gas along with a flowmeter to control the input rate; variable rate nutrient pump for optional use as a CSTR, and an asceptic sampling port (Figure 5). 65 nowuuen ", «3 run —(} : ’eunee ’ unmet (a) ‘9? Hum / "U" acre 3; use E r / “I moeuunou / new" ,) oueeueen SAN'LS II““ c (.’ Iw J I L‘ uunteut i i J II union . one“ 1'," "d 2;} I m I I I neurone —E:3 > , : . I . _ nine/contact. CU ' ‘ mu "we 2 3' T T— ' vetve at «seven . tutu -—-—— ' I A I cum ~ FIGURE A HARVEST! fl) HARVEST VESSEL VSSSIL FIGURE 5: Dlegrem at the bench scale fermentor used In eicperlmentetlon. CHAPTER VI: EXPERIMENTAL PROCEDURES AND RESULTS Most preliminary experimentation was conducted in small, 250 ml shake flasks in a batch culture system. The purpose of these experiments was to study the growth and production rates of the bacterial strains in various mediums and growth conditions. Specifically, the goal of these studies was to gather preliminary information pertaining to substrate inhibition, effects of secondary metabolites, effects of temperature, effects of pH, and product inhibition. The information acquired was used to choose the most promising strain, and its optimal growth conditions for work in bench scale fermentors. 66 As stated above, a literature review resulted in the acquisition of four bacterial strains which showed promise for the formation of acetic acid. One of these strains however, A. pasteurianus subsp. orleanensis (ATCC # 9432) is a true acetobacter, meaning that it will convert ethanol to acetic acid, but will not convert lactate. A. pasteurianus was therefore eliminated from the screening process. Cultures of the three remaining strains (Acetobacter cerinus (12303); Acetomonas oxydans (19357); and Gluconobacter oxydans, subspecies suboxydans (33448)) were prepared, and used to inoculate 100 ml of medium which contained approximately 200 mM lactic acid, 1% bacto-peptone, and 1% yeast extract. Duplicates of these cultures were grown for 96 hours, in the orbital shaker with samples taken daily for HPLC analysis. Results showed that both A. oxydans and G. oxydans, subsp. suboxydans converted nearly 100% of the lactic acid to acetic acid within a few days while A. cerinus showed only a 50% conversion (Figure 6). Thus, the latter strain was dropped from further study. CONCENTRATION (mNI) CONCENTRATION (mm) 68 -D- II. urine 4— a oxydcre sbsp. aboxydms so ' ‘2 TIME (hours)- FIGURE 60: Lactic acid profile of A. oxydans. G. oxydans subsp. suboxydons, and A. cerlnus. 0 TM: abours) . FIGURE 66: Acetic acid profile of A. oxydcns. G. suboxydans subsp. suboxydans, and A. cerlnus. This series of experiments had two objectives. The first was to determine which of the two remaining strains converted lactic acid to acetic acid more quickly and completely. This bacterium would then be studied in subsequent bench scale fermentation. The second objective was to approximate the growth conditions and medium content that yielded optimal conversion. 5 1 2 . E 1 1.1. . Several sets of experiments were run to determine the conditions needed for subsequent bench scale studies. The first set consisted of testing the conversion of lactic acid to acetic acid at varying initial substrate levels. These substrate inhibition studies were conducted in 50 ml of medium which contained varying levels of lactic acid (400 mM, 600 mM, 800 mM, and 1000 mM), 1% bacto-peptone, and 1% yeast extract. Duplicate flasks of each initial lactic acid concentration were prepared for each bacteria and sampled daily for five days. After five days, both strains showed 90% conversion of the 400 mM flasks, and 85% conversion of the 600 mM flasks. The acetic acid production began to decline significantly 70 for each species when initial lactic acid concentrations rose above 600 mM, with A. oxydans being the superior strain. Table 5 summarizes these results. Data revealed that as the initial concentration of lactic acid was increased, the lag time for acetic acid build-up was increased. Also, the final percent conversion of lactic acid to acetic acid was substantially lowered. This shows the inhibitory effect of lactic acid. Also, due to the observation that acetic acid levels reached near complete conversion at the lower substrate levels and remained constant it appears that overoxidation of acetic acid is not a concern with either of these bacteria. TABLE 5 Comparative substrate inhibition studies for A. oxydans and G. oxydans subspecies suboxydans. FINAL ACETIC ACID CONCENTRATION (IN) Initial lactic acid A. oxydans 447 a. oxydane subsp. concentration (mu) euboxydane 400 I 379 " 362 600 531 529 800 625 529 1000 747 ' 208 71 6.1.2.b; Additign Qfi gan steep liggg; The next several experiments were designed to determine if the addition of a secondary metabolite would increase the growth rate and/or production rate of acetic acid. The first metabolite tested was corn steep liquor (CSL). CSL is a residue obtained from the milling of corn which contains nutrients, sugars, starch and a host of polypeptides. To remove many of these polypeptides, which are potentially damaging to the organic acid column used in HPLC analysis, solutions of "clarified" CSL were made. This was accomplished by diluting 375 ml CSL in 1,125 ml distilled water. The solution was then titrated to a pH of 8.0 with 10 M KOH and autoclaved for 20 minutes. Finally, the resulting solution was filtered with 5.0 Whatneys filter paper. This clarified CSL was then used for experimentation. Preliminary experiments were run to determine if small amounts of CSL would inhibit the final conversion in initial lactic acid concentrations of 200 mM, 400 mM, 600 mM, and 800 mM. A secondary objective of these experiments was to determine which species was the superior strain as far as complete conversion to acetic acid was concerned. Flasks were prepared which contained 100 ml of medium and a 10% inoculum of either A. oxydans or G. oxydans, 72 subsp. suboxydans. The flasks contained varying amounts of lactic acid (200 mM, 400 mM, 600 mM, or 800 mM). For each lactic acid medium, four concentrations of CSL were prepared (either 10, 20, 40, or 60 ml of clarified CSL were added per 200 ml medium). This resulted in 16 distinct media to be tested for each bacteria. Replicates of each media were sampled daily. Results showed that these low concentrations of CSL were not inhibitory, and in fact slightly increased the final concentration of acetic acid. This may be misleading however, since lactate and other sugar substrates are present in CSL. Because of this, more carbon is available to the bacteria. Thus, although slightly more acetic acid was produced, the percent conversion of substrate to product was probably uneffected. These experiments also showed that A. oxydans was the superior strain, as it produced significantly more acetic acid after two days in every case (Figures 7a through 7d). WRATUNI ACE'IIC ACD AFTER 45.5 HEIRS WRATW ACE'IIC Aw AFTER 46.5 HELRS 8 EE . ,L........ Fro-A. I E rat-G. OXYDAIB SUBSP SUBOXYDANS 5 e .8. ml mCLARFIED CSL ADDED FIGURE 70: Acetic acid production by A. oxydans and G. oxydans subsp. subaxydans an addition of low levels of corn steep Ilquor. -u-A. OXYDANS ate-G. OXYDMB SUBSP.SUBOX‘I1:)ANS an 40 an rTI CLARFIED CSL ADDE FIGURE 7b: Acetic acid production by A. oxydans and G. oxydans subsp subaxydan's an addition of low levels of corn steep llcpor. WRATDN ACETIC ADD IFTER 45.5 HQJRS WRATDN ACE'I'IC ADD AFTER 45.5 HCLRS 74 -@ W .44. OXYDANS ..-x-G. OXYDANS suasp. suaoxvaNs an an an rri CLARFIED CSL ADDED FIGURE 7c: Acetic acid production by A. oxydans and G. oxy dons subsp. subaxydans on the addition of low levels of corn steep liquor. -o-A. OXYDANS I s -x-Gt mums gasp. sueoxvoms 5 i .' i L i 1 i 1 i 1 i ml allCLARiL—IED CSL ADDED FIGURE 7d: Acetic acid production by A. oxydans and G. oxyda'Is SLbap. subaxydcns on the addition of low levels of corn steep ,llqupr. 75 Next, experiments were conducted to determine the maximum amount of CSL that can be added before inhibition is seen (these experiments were conducted on A. oxydans only as it was shown to be the superior strain). Preliminary studies showed no inhibition with clarified CSL up to 120 ml/200 ml medium. This amounted to 6% CSL, thus, these tests included CSL concentrations of 12%, 14%, 16%, 18%, 20%, and 22% (by volume). Again, 100 ml of each media was prepared and sampled daily. The media contained 600 mM lactic acid, 1% bacto-peptone, 1% yeast extract, and the appropriate concentration of clarified CSL. A plot of acetic acid production vs. initial concentration CSL (by volume) showed that CSL seemed to become inhibitory at approximately 18-20% (Figure 8). I'll Irti rrNAC£TC ACD‘PRooImDNSDAYS § IITT' '3 3 3 3 3 3 3 mIn;n-l1.1.1111:lgnnnlnnnnlnnnllnnnnlnm a II I: III 20 22 24 'CORN STEEP LIQUIOR (VOLUME 76) FIGLRE 8: Acetic acid production by A. oxydcro on addtlon of hid) levels of corn steep HQJOl'. 6 76 2 . ... E J i 1 These experiments were designed to establish whether the addition of 10% glucose or 10% lactose to the medium would result in higher acetic acid production rates. It was hypothesized that addition of glucose and/or lactose may stabilize or enhance growth of the organism, especially in the early stages of growth. These tests were again conducted only on A. oxydans. Medium was prepared that consisted of 1% bacto-peptone, 1% yeast extract, and lactic acid concentrations of either 200 mM, 400 mM, 600 mM, or 800 mM. A second batch of medium was then made which consisted of the above ingredients with the addition of 10% glucose (weight/volume). Finally, a third batch was made substituting lactose for glucose. All media were titrated to a pH of approximately 5.5 with Ca/Mg OH. Fifty ml flasks of each medium were inoculated with A. oxydans and samples were taken for 10 days. Plots of acetic acid concentration vs. time showed that addition of 10% lactose to the medium was slightly inhibitory to product formation in every case. Also, addition of 10% glucose was slightly inhibitory for initial concentration of lactic acid at or below 400 mM. Higher initial lactic acid concentrations showed slight inhibition 77 from glucose for the first three days, and increased acetic acid formation afterward (Figures 9a through 9d). In no case was there a substantial increase or decrease in production rate. The conclusion, therefore, is that none of the secondary metabolites tested (corn steep liquor, glucose, or lactose) have a significant effect on the production of acetic acid from lactic acid. 78 25° -a-FLRE um: ACO 4.10;; GLUCOSE ADDED ; 440% LANE A0950 5 CCNCENTRATDN ACETIC ACID {HIM} ' .1501 zoo .250 sec .350 TIME (hours) FIGURE 9a: Acetic acid production by A. oxydans on oddltlon of 107. lactose or 10% glucose. '0'FLRE LAC": Aw """"""" .r vvvvvvvvvvvv -. , *‘DX (IUCOSE ADDED : ; . : ‘v'm LANE ADDED ...; ............ ° ......:. .......... 3 ............ é CCNCENTRATION ACETIC ACID {mM) FIGURE 98: Acetic acid production by A. oxydans on oddltlon of 10% lactose or 10% glucose. CONCENTRATION ACETIC ACID (mid) CONCENTRATION ACETIC ACID (mM) 500 79 ~a—PURI: mart ACO """" '" """""" :«mx euros: ADDED 5 . 5 : : 3.137;, “07085 ADDED 5 5 5 ........... 5 ............ Z. ........................ 5. ............ , ................................. : «’2 ._ .................................. .................................................... : ................................ g ............ .. ....................................... 1 1 1 1 I 1 1 1 1 I 1 I L 1 I I 1 1 1 1 I 1 1 1 o 50 no 150 200 250 500 $50 TME (hows) FIGURE 9c: Acetic acid production by A. oxydans on odditlon of 10% lactose or 10% glucose. -a-FU?E MC": ACO _+on GLUCOSE ADDED +139: allows: ADDED 5 _ ........................ °, ............ ; ......... . .................... , ............ O0.00.0003...OOOCOI'OIOIOOIOOOOI.OOOUOOOOOOCO~OOOOO00.00.00.0000000000IO. OOOOOOOOOOOOO 0 ~50 ‘00 45104 2:)0 Ll2;0-..3;0 550 TIME (hours) FIGURE 9d: Acetic acid production by A. oxydans on oddltlon of 10% lactose or 10% glucose. 80 WW Bacterial metabolism is also effected by temperature. Thus, these experiments investigated the effect of temperature on growth and production rates of A. oxydans for cultures containing several different concentrations of lactic acid as the substrate. Ideally, the growth and production rates of identical cultures could be tested at several temperatures run in simultaneous, parallel experiments. Unfortunately, only two orbital shaker baths were available. Thus, experiments could be conducted at only two temperatures at a time. The first set of experiments compared results from cultures grown at 26 and 33 degrees Celsius. Flasks were prepared containing 100 ml of medium consisting of 1% bacto-peptone, 1% yeast extract, and varying concentrations of lactic acid (5, 25, 50, 100, 400, 600, and 800 mM respectively). These flasks were inoculated and placed in the shaker at 26 degrees Celsius. An identical set of flasks was inoculated and placed in a 33 degree Celsius bath. Samples were then taken at 30 minute intervals and analyzed to determine growth rates, production rates of acetic and pyruvic acids, and consumption rates of lactic acid. After the results of this experiment were scrutinized, a similar experiment was run comparing the effects on 81 cultures grown at 33 and 37 degrees Celsius. This experiment followed the same procedure except that the number of flasks at each temperature was reduced to three, with the concentrations of lactic acid being 0, 57, and 110 mM. All flasks were titrated with Ca/Mg OH and H2804 to a pH of 5.5. Comparison of the 26 and 33 degree Celsius cultures showed that in every case (with the exception of the 600 and 800 mM cultures where substrate inhibition severely limited growth), the specific growth rate was higher at the increased temperature (Figure 10). HPLC analysis revealed that the production of acetic acid and the corresponding consumption of lactic acid were also greater in the 33 degree cultures. As an example, figures 11a and 11b show comparisons of growth and acetic acid production for the 50 mM lactic acid medium. SPECIFIC GROWTH RATE (1/hr) 82 - . . was; egg Ioo zoo zoo LACTIC ACD CONCENTRATION (mi/I) FIGURE 10: Specific growth rates of A. oxydans at 26 and 33 degrees celslus. WIWIN Into maroon. 83 INITIHLUICTICFBIO-Sflnfl 6° ozsmmazscaCm /' +33mmm 48 /’ 35 ' ./ 1/ 24 ////‘// 12 - =,/ 0 0 5 I) 15 ' 3 new.) 4 mm: uuwwwimouu udfldcmonlduluafl-Hbduddum INITIHL LflCTIC PCID - 50 III" _+$ mm *Bnm I- Whore) FIGURE 11.: Mdkw‘mdu and” «mu casino. 84 These results prompted the study comparing cultures grown at 33 and 37 degrees Celsius. Figures 12a through 12c Show that there was little or no difference in growth of the organism at either temperature. Thus, it appears a ”window" of temperatures exists between 33 and 37 degrees Celsius where maximum growth and production occur. This window may extend to higher temperatures as no higher values were studied. ABS URBANCE AT 640 nm 5 [+35 DEGREES cows +37 DEGREES (nous TIME (hows) FIGURE 120: Growth of A. oxydans at .33 cnd 37 degrees celcius on peptone and yeast extract. 20 ABSGQBANCE AT 64-0 nm ABSW AT 640 nm 85 I '0‘53 [IGREES CELCLB +37 DEGREES CELCLS ‘D TIIVE (hours) FIGURE 12b: Growth of A. oxydans at 33 and 37 degrees celcius on 57 mM lactic acid. 5 [ +53 DEGREES coats -ae-37 DEGREES cacus TIle (hours) FIGLRE 12c: Growth of A. oxydans at .33 and 37 degrees celcius on 110 Inlvl lactic acid 86 Elites—Emma Another factor which significantly effects bacterial growth is pH. Studies were conducted to determine the pH values that resulted in the optimum growth. These experiments were run at 33 degrees Celsius due to the results of the previous experiments. Also, it was previously determined that growth was not effected by substrate inhibition for lactic acid concentrations of 100 mM. Thus, these experiments were conducted with 100 ml of medium consisting of 100 mM lactic acid, 1% bacto-peptone, and 1% yeast extract. Each culture was asceptically titrated with NaOH so that the pH before inoculation was 4.5, 5.0, 5.5, 6.0, 6.5, 7.0, or 7.5. Hourly samples were then taken for growth rate study. Absorbance values seen in Figure 13 revealed that growth was uneffected when the pH was 6.0 through 7.5. When the medium's pH was lowered to 5.5 however, a significant decrease in the growth rate was seen. Also, no growth was observed at pH values of 5.0 or 4.5. Unfortunately, data was not recorded for times less than 3.25 hours. Thus, a second set of experiments following the identical procedure was run with samples taken over the first 4.5 hours of growth (the only exception being that no cultures were grown with pH values below 5.5). The 87 data acquired in these experiments confirmed the optimal pH to be 6.0, and that pH is not inhibitory to the growth of A. oxydans at values between 6.0 and 7.5 (Figure 14). 9: : pill-4.5 u:+pH-5° 00000000 § 0000000000000000000000000000000 I ooooooooooooooooooooooooooooo fw-pH-ub 7*Dn-w oooooooo 1 ooooooooooooooooooooooooooooooo ' oooooooooooooooo 1 oooooooo ABSORBANCE AT 64-0 rm 13 9 13 'ITlI/E (hours) . FIGURE 13: Growth profiles of A. oxydans for times greater than 4 hours at varying pH levels. 88 LN (AesoeANCE) AT 640 nm TIME (hours) 3 - FIGURE 14: Growth proflcs of A. oxydans for times less than 4.5'hours at varying pH levels 89 i 1 2 E' I . 1.]. . 3' It is a well established fact that the growth of many bacteria is inhibited by large concentrations of their metabolic products. Thus, it was necessary to investigate the effects of acetic acid on the specific growth rate and production rate of A. oxydans. One hundred ml of medium were prepared in seven different shake flasks. Each flask contained 0.5% bacto— peptone, and 0.5% yeast extract. Six of these flasks contained 100 mM lactic acid as the substrate. These six flasks also contained concentrations of acetic acid ranging from zero to 150 mM. The seventh flask was used as a control and contained neither lactic acid nor acetic acid. After sterilization and titration, the flasks were placed in the orbital shaker bath at 33 degrees Celsius. They were then sampled for growth rate studies and HPLC analysis every 30 minutes for 17 hours. As expected, the addition of acetic acid resulted in a marked decrease in growth rate, ranging from 0.322/hour at 10 mM to zero growth at 150 mM (Figure 15). Curiously, the addition of 10 mM acetic acid seemed to improve the growth rate when compared to the internal standard which contained no acetic acid (only lactic acid as a substrate). It seems that acetic acid is not significantly inhibitory until WHITE (1hr) mmmm 150 FIGURE 15: Eflects of product lnhlbltlon on growth'rate' at A. oxydans. 91 concentrations exceed 50 mM, and that it is not critical until concentrations exceed 100 mM. These concentrations are below the desired final concentration for an industrial process. Thus, if growth and production of acetic acid follow this same pattern in larger bench scale fermentations, this could be a limiting factor in an industrial situation which may make the process unfeasible, particularly for continuous fermentations. The production of acetic acid, consumption of lactic acid, and buildup of pyruvic acid, was also monitored. Results showed the rate of acetic acid production to be greatest when little or no acetic acid was added to the medium (Figure 16). It appeared to progressively drop off as more acid was added for initial concentrations ranging from zero to 100 mM. At an initial concentration of 150 mM, virtually no acetic acid was produced. Figure 17 shows a plot of total carbon vs. time. The term "total carbon" refers to the added concentrations of lactic, acetic, and pyruvic acids (as stated earlier, pyruvic acid is the expected primary intermediate in the metabolism of this bacterium). These results show that the total carbon concentration remained virtually constant for each flask. The significance of this result is that it supports the claim that acetic acid is not lost to "overoxidation" via a TCA Cycle with this bacterium. 92 mowers... : times *Wmm y l l :j><7 I.._../ / ”Ti/Er I I I I I j I O S 15 I) mm FIGURE 16: Eticctc oi product inhibition on acetic acid . production byA. oxydans. momentum, :Rfifittfin I&%%% 3m M ' . M /'\._—-—o——._—. E I. .- .. aim ..... . _T WT-‘g WW”; _ :—_ 6°: m: I: m. o. I I I I I I I I I I I I I I 0 5 D 15 mm FIGURE 17: Total carbon proliic tor A. oxydans growth. 93 It should be noted that the flask containing only peptone and yeast extract (no added lactic acid or other substrates) did produce some acetic acid. This is due to the fact that some lactic acid and residual sugars are present in the peptone and yeast extract which could be used for substrates until they are depleted. This also explains why growth was repeatedly observed in experiments where the medium consisted solely of these ingredients. Also note that the 10 mM acetic acid flask has been omitted from these results. This was due to an unfortunate lab accident causing the loss of this data. At this point in the research, it had become apparent that use of A. Oxydans gave the most optimistic results for an industrial scale fermentation. As stated in the research plan, the next level of experimentation was to study its growth and production rates in bench scale fermentations. These experiments were conducted in 2.5 liter or 5.0 liter fermentors described earlier. Three main series of experiments were conducted. These consisted of studies in batch, semi-batch, and continuous flow systems. LLLW Several batch experiments were run in the two fermentors. These consisted of tests designed to detenmine growth rates, comparison between use of sparged oxygen and air, and comparison between use of Ca/Mg OH and NaOH for pH control. WW Growth and production rates were previously observed for A. oxydans in 250 ml shake flasks. These studies showed very little growth when the substrate concentration was 95 raised above 400 mM. Preliminary studies in bench scale fermentors however, showed considerable growth up to 600 mM lactic acid (seen by visual inspection of the culture). The hypothesis was that oxygen transfer somewhat limits growth in the shake flasks, as the liquid-gas interface is the only possible point for oxygen exchange. Conversely, oxygen transfer in the fermentors is not a limiting factor due to the presence of a sparger system and an internal impeller which greatly increases the liquid-gas surface area. Thus, the growth rates were seen to vary significantly between small shake flasks and the larger fermentors - much the same way vinegar production is increased in submerged fermentation vs. the Orleans process. This set of experiments determined the growth rate of A. oxydans for various substrate concentrations in bench scale fermentations. In each experiment the predetermined optimum growth conditions were maintained - pH was controlled with Ca/Mg OH and H2504 at a constant level of 6.0 and temperature was maintained at 33 degrees Celsius. Also, oxygen was supplied by sparging air (which was first sterilized through a .45 um filter) through a humidifier and into the culture at a constant rate of one volume per volume per minute (v/v/m). The impeller speed was controlled at 500 rpm. A ten per cent inoculum (250 ml grown overnight) was introduced to 2250 ml medium and samples were taken for optical density analysis every thirty minutes for several 96 hours. The fermentor was then sterilized and the procedure was repeated several times for media of varying substrate concentrations. Two main sets of experiments were run. First, growth rates were studied for media containing only peptone and yeast extract in varying proportions (0.1%, 0.25%, and 0.5% on a weight to volume basis). A second set of studies were then conducted on medium containing equivalent amounts of peptone and yeast extract (0.5% each) and varying amounts of lactic acid. Studies of varying peptone and yeast extract concentrations showed that both the growth rate and the total amount of growth were lowered when less peptone and yeast extract was added (Figure 18). The 0.5% fermentation showed little or no lag phase and gave a growth rate of 0.706/hr over the first 6 hours. When the content was lowered to 0.25%, a lag phase of approximately 1 hour was observed and the growth rate was reduced to 0.407/hr. Finally, when the concentration was reduced to 0.1%, a lag phase of nearly 3 hours was seen. The growth rate, however, did not significantly change, as it was determined to be 0.419/hr. Also, the final optical density was reduced from approximately 12.0 to 6.0 and 3.0 when the concentrations were lowered from 0.5% to 0.25% and 0.1% respectively. 97 r—o— 0.1% PEPTGE /Y EAST EXTRACT _—l— 0.25% PEPTONE/YEAST EXTRACT ABSORBANCE AT 640 nm -"- (LOX PEPI'ONE/YEAST EXTRACT TME (hod's) FIGURE 18: Growth of A. oxydcns on varying levels of peptone and yeast extract. 98 Figure 19 shows the effect of altering the medium's lactic acid concentration. The initial growth rate appeared to increase almost linearly with substrate concentration for the 0.0, 25, 50, and 100 mM lactic acid fermentations. Higher concentrations of lactic acid, however showed significant substrate inhibition. The growth rate was lowered from 1.1/hr at 100 mM lactic acid to 0.87/hr at 200 mM and continued to fall until no growth was observed at 800 mM. It should be noted that although optimal growth (and presumably production rate) was observed at 100 mM lactic acid, substantial growth was still seen at concentrations up to 600 mM. Thus, it should be possible to run subsequent fermentations at the higher substrate levels. 15 - - - - - - I I O O I C I C I C I I I C I O O C SPECIFIC GROWTH RATE (1 /HR) ”0“”..5.””2.5;“..35".I.””sio”‘3§.””“"2... CONCENTRATION LACTIc ACID (AI) FIGURE 19: Specific growth rate of A. oxydans in varying levels of lactic acid medILm . l 9' 019;,i_0 9‘ .‘-n -pcno:1o .: 1 0 0+1 .' Acetobacter species are highly aerobic. Consequently, it was hypothesized that acetic acid production by these bacteria may be more efficient when grown on pure oxygen rather than air. Preliminary experiments appeared to show however, that the growth of A. oxydans in air fermentations equaled or surpassed that in similar oxygen fermentations (by visual inspection of turbidity in the fermentation broth). To confirm this result, growth studies were conducted in two side-by-side fermentations. Each fermentation broth contained an initial lactic acid concentration of 400 mM, 0.5% peptone, and 0.5% yeast extract. The pH was titrated and held constant at 6.0 with Ca/Mg OH and H2804. Temperature of each fermentor was held constant at 33 degrees Celsius, gas flow rates through each were 1.0 v/v/m, and the impeller rotation was 500 rpm. The only difference between the two fermentors being that the first was sparged with oxygen while the second used air. A 10% inoculum was introduced to each fermentor, and daily samples were acquired for HPLC analysis. Surprisingly, the results showed that the air fermentation produced acetic acid slightly faster, and at a higher final conversion than the oxygen fermentor 100 (Figure 20). Production in the air system was approximately 8.5 mM/hr for samples taken between 13.5 and 47 hours. The corresponding oxygen system produced acetic at the slower rate of 6.9 mM/hr. Each system appeared to reach a ”near final” conversion after only 2 days (89% for the air and 76% for the oxygen culture). The conclusion of these results is that little difference is observed in acetic acid production whether A. oxydans is grown on air or pure oxygen, and that growth on air may in fact be advantageous. 5w 1 4L- CRUNHICWLNR ‘flF£RUWTH(}IDXYGBI ”£400 ................... ‘ ....... ; ......................... . .......................... . ............ 5 2 93m .................................................... ., ................. 5 520° ........................................................................................ gm ...................... :. ......................... ._ .......................... , ............ 01 l I l I l I O 20 FIGURE 20: Comparison of acetic acid profiles of A. oxydans in a batch fermentation grown on air vs. pure oxygen 0 ° 0119'. i .7 Q‘ “‘ a U. .' 0". \g.’ 0 Q A problem had arisen because Ca/Mg OH repeatedly clogged the base input port of the pH controller. To eliminate this problem, it was proposed to substitute NaOH as the base control since previous tests in shake flasks seemed to show that the bacterium grew well on NaOH. Thus, it was necessary to compare the relative production rates from use of NaOH vs. Ca/Mg OH. Further, it was necessary to determine if Ca/Mg OH inhibited production, as one hope of this project was to use it as a pH control at the pilot plant and industrial levels. Once again, 2.5 liters of medium were prepared. This medium contained approximately 500 mM lactic acid, 0.5% peptone, and 0.5% yeast extract. Pure oxygen was sparged through the broth at a rate of 1.0 v/v/m and the pH control was set at 6.0. All other conditions were identical to previous fermentations. A second fermentation was prepared using the identical procedure with the exception that NaOH was used in place of Ca/Mg OH as a pH control. Samples were again taken at daily intervals for 4 days. Figure 21 shows the resulting acetic acid production profiles. The NaOH production rate over the first two days was 7.1 mM/hr compared to 6.8 mM/hr when Ca/Mg OH was used. 102 The conclusion was that NaOH and Ca/Mg OH could be used interchangeably in subsequent fermentations without significantly altering the results. :w—GROIVI'HCNCA/NCO-I - ; 4—o ._.._ GRWlT-IONNm-l ' = = CONCENTRATDN (mu) to 8 'TTTIIIIIIIIII|IIII[ FIGURE 21: Acetic acid profiles of A. oxydans in a batch fermentation gown with NaOH vs. (30/ Mg OH. 103 5 2 2° ._1 1 . The goal of these experiments was to find the acetic acid production rate of A. oxydans in a semi-batch fermentation. A medium of approximately 500 mM lactic acid, 0.5% peptone, and 0.5% yeast extract was prepared. Inoculation procedures were identical to previous experiments with the following fermentor conditions: temperature = 33 degrees Celsius, pH control (with NaOH and H2804) set at 5.75, oxygen sparged at 1.0 v/v/m, and impeller rate = 500 rpm. Daily samples were taken for HPLC analysis. After four days, approximately 50% of the fermentation broth was removed and replaced with fresh, cell free medium. Several cycles of this type were repeated. A plot of lactic acid vs. time showed an initial concentration of 495 mM. The substrate was rapidly converted to acetic acid so that after only 2 days, 65 mM lactic acid remained and 414 mM acetic acid were formed (Figures 22a and 22b). This corresponded to a production rate of approximately 7.5 mM/hr. After 3 days, the acetic acid concentration appeared to peak at 479 mM (97% conversion). After day four, the first medium exchange was performed. Unfortunately, although lactic acid was continually degraded, acetic acid was not produced. This indicated a possible contamination. Within two cycles, a 104 visible contaminant was observed, and the experiment was terminated. someIIRAToN (mu) CONDENI'RATDN (mM) I4......0.00.0.0...O‘COOOOOOOOOOOOOOOO 000000000000000000 Irlrrirlrllllrrrrlrr 111111111 [1111l111li1111 100 ‘50 TME (hairs) 250 FIGURE 220: Acetic acid profle of A. oxydans in a sent—batch fermentation grown on 495 mM lactic acid. WilllIIIIWIITIIIIIIIT 11111l1111l1111 11111111 0 The (new?) 51) 250 FIGLRE 22b: Lactic acid profile of A. oxydans in a serri—batch fermentation grown on 495 NM lactIc acid. ’ 105 A second semi-batch fermentation was attempted following the same procedure, with the exception that the medium contained 577 mM lactic acid. Results of this experiment are seen in Figures 23a and 23b. Initial acetic acid buildup was again fast, reaching 404 mM (70 % conversion) after 45 hours and peaking at 518 mM (90%) after 96 hours. This buildup seemed to lag slightly compared to the previous experiment when the initial concentration of lactic acid was only 495 mM. This may be due to the inhibitory effect of the increased lactic acid concentration. Acetic acid buildup after subsequent medium exchanges, showed a slower production rate and lower final acetic acid concentration. This phenominum.makes the use of a semi- batch fermentation undesirable. The production rate, however, seemed to increase with each cycle as did the final concentration. This was expected, as the bacterium acclimated itself to the fermentation conditions. Each cycle seemed to show a one day lag followed by a day of rapid production, a day of slow production, and a day of little or no production. Lactic acid consumption profiles showed similar results. The first cycle was very fast and complete with lactic acid concentrations falling from 577 mM to 100 mM 106 after three days (17%). Each cycle showed a lag period followed by a rapid degradation period, and then a slower degradation period. comaITRATm (my) I TivE MS) FIGURE 230; Acetic acid profle of A. oxydcns in a serri-batch fermentation grown on 577 mM lactic acid. TME 2(liars) FIGLRE 23b: Lactic acid profile of A. oxydcns in a semi-batch fermentation grown on 577 mM lactic acid ‘ 107 5 2 3_ : l' E] E . Careful scrutinization of previous growth and production rate studies indicated that the production of acetic acid was growth associated. That is, most of the production occurred when the culture was in its exponential growth phase. This suggested that fermentation in a continuous flow system may yield higher production, as the culture would constantly remain in the exponential phase. Two side—by—side experiments were prepared in the Bio— Flo IIc fermentors. Initially, one fermentor was filled with medium consisting of approximately 400 mM lactic acid and the second with 200 mM. Each medium also contained, 0.5% peptone, and 0.5% yeast extract. Fermentor conditions were as follows: pH control set at 6.0 (using HZSO4 and NaOH), sparged air for oxygen supply at 1.0 v/v/m, impeller rate set at 250 rpm (this lowered rate was incorporated to help reduce a foaming problem previously observed), and temperature set at 33 degrees Celsius. A ten percent inoculum was introduced to each system and grown in batch fermentation until significant cell concentration buildup was observed. Daily samples were again taken for HPLC analysis, dry weight analysis, and optical density analysis. After substantial cell growth was observed, seen by monitoring optical density of the broth, a 108 continuous fermentation was started with the substrate concentration being 300 mM lactic acid in each fermentor. The initial dilution rate in each culture was set at 0.05/hr, giving a 20 hour residence time. The cultures were allowed to grow in this environment until steady-state conditions were observed. At this point, one fermentor was used to study the effects of altering the dilution rate (while keeping the substrate concentration constant), and the second fermentor was used to determine effects of altering the substrate concentration (while keeping a constant dilution rate). Figure 24 shows the lactic acid and acetic acid profiles when the dilution rate was varied between 0.05/hr and 0.02/hr (corresponding to residence times ranging between 20 and 50 hours). Analysis of this plot shows a relatively high rate of conversion during the first two days when the system was in batch growth. The fermentation was then changed to a continuous flow system.with a dilution rate of 0.05/hr for the next six days. This resulted in steady state acetic and lactic acid concentrations of approximately 44 mM and 200 mM respectively. The subsequent five days saw mechanical problems with the fermentors feed pump. Thus, concentration profiles showed erratic results over this time period. After the 109 problem was solved on day 13, the dilution rate was reduced to 0.03/hr. A steady state was eventually obtained so that after 560 hours acetic and lactic acid concentrations of approximately 80 mM and 135 mM were observed. Finally, the dilution rate was reduced to 0.02/hr for the last five days of the fermentation. Steady state values of approximately 165 mM acetic acid and 105 mM lactic acid were achieved for this condition. 4w . . , , W 5 I-oflxnmcaad jo-Aafibfnm -f-%I&mmaan1 r? w .......... ........... ........... i ........... .... ......... .. m ... mum... moves-m “mom" i zoo rnM lactic acid ° 1 5 3 ° 5 " .8 E 2” l ..... ......... . ................... 3 ........... .... :' ......... .. m g g ll 0 . g E . g E m l o 1 1 3. :33... ......uézooo w £2, g 2 . : 4“ N . : : 5 3 V7 : ' 8 w I) in”; ........... : ........... : .......... «Fr-f." ...... ......... .. 20 I W: V“: s s ‘ U : °Q¥Qe ° : : : ' o IJJJ i I I I I i I I I I LI I I I i I I I I i I I I II I I I I - o O 100 200 500 $0 7m :00 400 TIME (hours) FIGURE 24: ‘ Effects of altering diLItion rate on lactic a'Id acetic acid profiles In a conthuous flow system 110 A plot of acetic and lactic acid profiles with varying concentrations of lactic acid in the feed is shown in figure 25. Batch growth was ended after three days and a continuous flow system containing 300 mM lactic acid medium and a dilution rate of 0.05/hr was initiated. A foaming problem in the fermentation broth caused lactic acid concentrations to be erratic for the next five days. Eventually, this problem was corrected with the incorporation of an internal level sensor and an antifoaming agent. The resulting steady state conditions of 44 mM acetic acid and approximately 180 mM lactic acid were observed after 330 hours. The level of lactic acid in the feed was then decreased to approximately 200 mM for the next four days. This resulted in a decreased steady state value for lactic acid (approximately 100 mM) and a slightly increased acetic acid concentration (49 mM). Finally, the feed content of lactic acid was lowered to 100 mM. The corresponding steady state values of lactic and acetic acids were approximately 30 mM and 50 mM respectively. CONCENTRATION (mM) 111 4m 0 O i C O m [jo-Lactlc acIcl *Afzetlc add . +2001me . I 2 ' 1 2 v - m I... .......... o ooooooooooooo o ooooooooooo dance I8.“ {.6 oooooooooooooooooooo I. 2” .... ........... .'... . ...... i ........ ...i. g l- b ICSTR: m _ dilutieo'nnrntc-OJSIM. 8 _ 300 MM lactic odd & m u....... ............. ....... ,.:.. ' U M 1 11 1 I 1 1 1 1 I 1 1 1 1 I 1 1 1 1 I 1 1 1 1 d 0 1m - soo KI) 200 300 400 TME (hows) FIGURE 25: Effects of altering medtm lactic acid concentration in a continuous flow system. CHAPTER VII: DISCUSSION A literature survey of acetic acid bacteria and a preliminary screening of the four strains judged most promising, narrowed the pool of species to be studied in detail to two. Both of these bacteria; Gluconobacter oxydans, subspecies suboxydans (ATCC number 33448) and Acetomonas oxydans (ATCC number 19357); was found to convert nearly all of a 200 mM lactic acid medium.to acetic acid in three days and showed no loss of prdduct due to overoxidation of acetic acid. A series of tests was conducted in 250 ml shake flasks to determine levels of substrate inhibition, effects of adding secondary metabolites to the medium.(corn steep liquor, lactose and glucose), effects of temperature, effects of pH, and levels of product inhibition. Table 6 summarizes the results of these experiments. 112 113 TREK“! 6 Summary of shake flask experimental results. 37 degrees Celsius where optimum growth and acetic acid production occurs. The window may extend to even higher temperatures. TEST ACETOMONIS OXYDANS GEUCONOBACTER OXTTHUWS, EflHflSRECflHHS SNHNIXYIUUWS SUBSTRATE Began to show inhibition above 600 Similar to A. oxydans, INHIBITION mM lactic acid. but inhibition was more pronounced. Showed (substantial inhibition above 600 mM lactic acid. CORN STEEP Low concentrations did not effect ow concentrations did LIQUOR production. Produced significantly not effect production. ADDITION more acetic acid in two days than Produced less acetic G. oxydans, subsp. suboxydans. CSL acid in two days than became slightly inhibitory when 18% A. oxydans at these by volume was added. concentrations. LACTOSE 10% addition was slightly Not tested ADDITION inhibitory to product formation at Isubstrate levels ranging from 200 RM through 800 mM. GLUCOSE 10% addition was slightly Not tested. ADDITION inhibitory to acetic acid formation below 400 mM lactic acid. Cultures containing 600 mM and 800 mM lactic acid showed an initial inhibition followed by slightly increased production. TEMPERATURE An apparent I'window" of Not tested. ‘EFFECTS temperatures exists between 33 and 114 TABLE 6 (continued) TEST ACETOMONBS OXIDANS GEUCQNOBACTER CDCZDMUWS, EflIBSflUICUJMS ASUEKHYYIUUMS Growth rates were uneffected when Not tested PH EFFECTS PH was 6.0 through 7.5. A pH of 5.5 showed decreased growth rate and no growth was observed at pH values of 5.0 and 4.5. PRODUCT 10 mM acetic acid yielded an Not tested INHIBITION increased growth and production rate. Further addition of acetic acid, however, showed inhibition to growth and production. This inhibition appeared to become significant at 50 mM and critical at 150 mM where no growth or production was observed. As seen in this table (and detailed in the previous chapter), studies of substrate inhibition and secondary metabolite addition show that A. oxydans produces acetic acid at a substantially faster rate and converts a higher percentage of substrate to product than G. oxydans, subspecies suboxydans. For this reason, all subsequent experimentation focussed on this bacterium. Bench scale fermentors were used to study the growth and production characteristics of A. oxydans in batch, semi— batch, and continuous systems. Batch experiments consisted 115 of studies designed to examine the effects of substrate inhibition, sparging with pure oxygen as a replacement for air, and controlling pH with NaOH or Ca/Mg OH. Varying the level of substrate concentration seemed to have a significant effect on growth of the bacteria. As the substrate level was increased, a corresponding, proportional increase in growth rate was observed until a maximum rate of 1.1/ hr was reached at 100 mM lactic acid. Above this concentration, lactic acid became inhibitory, and the growth rate dropped to 0.87/hr at 200 mM. The growth rate continued to fall with increasing substrate concentration until no growth was observed at 800 mM. Although optimal growth is observed at 100 mM lactic acid, batch fermentations should be feasible up to 600 mM, where a substantial growth rate was still seen. Subsequent batch fermentations compared the production rates of air vs. pure oxygen fermentations, and fermentations with pH control using NaOH vs. Ca/Mg OH. These studies revealed a production rate of acetic acid with the use of air to be approximately 8.5 mM/hr compared to 6.9 mM/hr with oxygen (when the medium contained between 400 and 500 mM lactic acid). After two days the conversion to acetic acid for the air system was 89% and that of the oxygen system was 76%. Apparently, pure oxygen slightly inhibits the growth of this bacterium. 116 Results of the comparison in bases used for pH control showed little or no difference when either NaOH or Ca/Mg OH was used. The production rates over 48 hours were 7.1 mM/hr and 6.8 mM/hr respectively. This result prompted the assumption that NaOH and Ca/Mg OH could be used interchangeably as a pH control without significantly effecting the results. This assumption is significant in that all future experiments used NaOH, due to experimental difficulties with Ca/Mg OH. It was stated previously, that pH control with Ca/Mg OH would eliminate the costly steps of adding dolomitic limestone to the completed fermentation broth in the industrial scale production of CMA. Thus, characteristics of A. oxydans grown on Ca/Mg OH or a substitute base that yields similar results were desired. After these batch experiments were completed, semi- batch and continuous fermentations were tested. Results of semi-batch experiments grown on either 500 or 600 mM lactic acid showed a maximum production rate of approximately 7.5 mM/hr (this occurred over the first 48 hours). Also, a conversion ranging from 70 to 85 per cent after two days was observed in the initial batch growth. However, neither this production rate nor the final acetic acid concentration was duplicated after even one medium exchange (approximately 60% exchanges by volume were made). 117 A possible explanation for this behavior is that the bacterial metabolism is severely limited.due to product inhibition. If this were the case, a high initial production rate would be expected, due to a lack of acetic acid in the broth. However, after every semi—batch exchange, a relatively high acetic acid concentration would exist. Thus, subsequent growth and production would be limited. This was in fact the case observed in each semi- batch experiment. This hypothesis is also supported by the fact that shake flask experiments showed severe product inhibition when the acetic acid concentration exceeded 100 mM. One could theoretically alleviate this problem by raising the fermentation's pH to a value as high as 7.5, since no inhibition was seen up to this level in earlier experiments. Such action would convert most of the free acetic acid, which has a pKa of 4.8, to acetate. Unfortunately, as previously stated, magnesium hydroxide is insoluble at pH values above 6.0. Thus, raising the pH to 7.5 may help to alleviate the acetic acid inhibition (assuming there is not a corresponding acetate inhibition), but the product would be calcium acetate, not the desired mixture of calcium and magnesium acetate. Results of experimentation using a continuous phase fermentation were disappointing. These experiments altered 118 either the concentration of lactic acid in the medium (from 100 mM to 400 mM), or the dilution rate of the fermentation broth (from 0.02/hr to 0.05/hr). The highest concentration of acetic acid (163 mM) was observed with a 300 mM lactic acid medium and a dilution rate of 0.02/hr. This corresponds to a meager 3.3 mM/hr acetic acid production rate with a residence time of 50 hours. This rate is less than half that observed in batch fermentation. Further, as the dilution rate increased to 0.05/hr (a 20 hour residence time), the corresponding product concentration was decreased, to approximately 43 mM at steady state. This behavior was expected since lowering the residence time would decrease the available time to convert lactic acid to product. Initially, it was believed that the low conversion was due, at least in part, to substrate inhibition. Tests that lowered the medium's lactic acid concentration, however, failed to significantly change the production rate. Thus, the most probable explanation for the low conversion in semi-batch and continuous systems is the aforementioned product inhibition. These experiments indicate that the most efficient means of production involve using A. oxydans in a batch fermentation grown on air with a medium of approximately 500 mM lactic acid, a pH of 6.0, and a temperature between 119 33 and 37 degrees Celsius. This system exhibits acetic acid production rates between 7.5 mM/hr and 8.5 mM/hr over the first two days of operation. The substrate concentrations yielding these production rates were 495 mM and 577 mM lactic acid, and the final conversion to acetic acid reached 97% and 90% respectively, after four days. The current market price for CMA is thirty cents per pound, with over 95% of this cost being directly attributed to the production of petroleum based glacial acetic acid. These figures must be reduced if CMA is to be used on a large scale basis for roadway deicing. In an attempt to achieve the desired reduction, it was proposed here to incorporate the use of renewable biomass, in the form of surplus corn, in a two stage fermentation to form CMA. The process, shown in figure 26, would be initiated by corn milling and hydrolysis of the created starch to form glucose. This glucose would then be utilized as a substrate for a lactate fermentation. Proprietary research reports that this fermentation, which replaces NaOH with industrial 120 dolime for pH control, produces 82 g/l lactate from 90 g/l glucose. This lactate exists in a 1:1 molar ratio of the calcium and magnesium salts with 95% of the calcium magnesium lactate (CML) in solution. It should be noted that a potentially limiting factor in this process is that CML is only soluble up to 900 mM. This may not be a problem, however, since production of a 1.5 M CML slurry has been reported. This slurry could be used as a substrate for subsequent steps with the particulate CML coming out of solution as the dissolved CML is utilized by the bacteria. CML would then be fed to a submerged culture, acetate fermentor. As discussed above, research proved the optimal scheme for this step to consist of a batch fermentation using Acetomonas oxydans. This system converted 97% of a 495 mM lactic acid medium to acetic acid (on a molar basis) in 91 hours and 90% of a 577 mM medium in 96 hours. At this point the slurry would enter a centrifugation stage for the removal of protein fiber, germ oil, and cells. Ideally, these components could be recovered for byproduct credit. This stage would be followed by the addition of bentonite and a second centrifugation to remove the remaining particulate matter. The dilute, solid-free CMA solution would then undergo evaporation to a near saturation 121 condition of approximately 30 weight per cent by a six effect evaporator. In the last step of the process the CMA would be dried in a steam heated rotary drum, and pelletized to form the final product. “flflER ACEUUE FERMENTKHON I A CENTRIFUGING DRYING EVAPORATION + I WASHING 1 WATER = m WATER ’ ' ANIMAL FEED “MWER noun 2‘: an production 1m tho peopoood -ucuu- prooou. 122 An alternate process, not tested here, could use conventional acetic fermentation technology from the vinegar industry. This process, seen in the schematic drawing of figure 27, would differ from the above proposal only in the fermentation steps. It would utilize well known technology to ferment ethanol from glucose in the first stage, and acetic acid from ethanol in the second stage. This fermentation process is documented in the vinegar industry. Thus, an economic analysis of the proposed "lactate process” must be compared to this "vinegar process" to determine which proposal deserves consideration at the pilot plant level. WATER AcID ETHANOL , ACETATE * FERMENTATTDN FERMENTATTDN DOUME cENTRIFUGE DRYING EVAPORATION UTRALIzATIDN H - l I WAsFIING "E » WATER ANIMAL FEED “MWER “gun 21: on production with mm (to- tho m: industry. 123 Typically, acetic acid is produced in the Vinegar industry by various acetic acid bacteria at a rate of 1.7 g/l/hr (28.3 mM/hr) and a final acetic acid concentration of 100 g/l (approximately 2.2 M). The process involves a batch or semi-batch fermentation requiring approximately 58.8 hours, and it is possible to increase the yield to nearly 200 g/l (4.4 M) by lowering the process temperature after an acetic acid concentration of 100 g/l is achieved, and increasing the cycle time. Analysis of experimental results shows that the lactate process gives a 97% molar yield of acetic acid from a 495 mM lactic acid medium and a 90% yield from a 577 mM medium. Coupled with the aforementioned proprietary lactate fermentation, and an assumed 100% recovery of CMA from the acetic acid broth, this results in 0.75 pounds CMA formed per pound glucose and 0.70 pounds CMA formed per pound glucose, respectively. The corresponding vinegar process yields 0.554 pounds CMA per pound glucose. Unfortunately, the concentration of acetic acid via the vinegar process after fermentation is over three times that in the two lactate processes. This, and the realization that the fermentation of the lactate process requires nearly twice the time, drastically decreases its economic feasibility. Table 7 gives a summary of the assumptions made in the analysis of each process. Further, tables 8a through 8f 124 show the estimated capital cost and operating costs for CMA plants incorporating the lactate process with either 577 mM or 495 mM lactic acid medium, and vinegar process. This data was based on 1988 price estimates for a plant utilizing the ethanol process with a 200 ton per day CMA capacity (30). IHUELB 7 Process assumptions for economic analysis of fermentation processes. W W Glucose to lactate: .911 lb. Glucose to ethanol: 95% of theory, lactate per pound glucose. or .485 lb ethanol per lb. glucose. Lactate to acetic acid: determined Ethanol to acetic acid: 98% of experimentally. theory or 1.278 lb acetic acid per lb. ethanol. Acetic acid to CMA: 100} of theory cetic acid to CMA: 100‘ of theory Lzr 1.25 lb. CMA per pound acetic r 1.25 lb. CMA per pound acetic cid. cid. 100% starch to glucose. 100‘ starch to glucose. ,Dilute CMA free of suspended solids Dilute CMA tree of suspended solids evaporated to 30 wt% by a six Pvaporated to 30 wt% by a six effect evaporator using 1 lb. steam effect evaporator using 1 lb. steam per 5 lb. vapor evaporated per 5 lb. vapor evaporated Heavy solids dried by same drying eavy solids dried by same drying technique used for Corn gluten. technique used for corn gluten. Non-sterile operation . ran-sterile operation . Ethanol oxidation identical to that sed for'vinegar production: 1.7 ll/hr,100 g/l in 58.8 hours. 125 TABLE 8a CAPITAL COST FOR PROPOSED "LACTATE" PROCESS (1991 dollars with 495 mM lactic acid medium) COMPONENT NUMBER SIZE 3 (THOUSANDS) Corn dry mill 1 10,645 bu/day 3,680 Acid hydrolyzer 1 324L235 lb/day 360 [Lactate fermentors with heat exchangers + 3 225,000 gal 1,035 circulation pumps 1,000 ft2 Acetic acid fermentors with heat 7 500,000 gal 38,687 exchangers. power agitators, air filters and 6,000 ft2 compressors 1700 HP 8,900 scfm JNeutralizers with agitators 0 0 0 entrifuges 3 875 g/min 4,902 350 HP [Protein feed dryer + loader l 25 .200 lb 4,434 HZO/hr 6 effect evaporator 1 61,250 lb 5,369 H20/hr ICMA dryer 4» agglomerator 1 140,000 lb 6,462 H20/11r SUB TOTAL 64,930 ICapital for off sites + start up 64,930 TOTAL CAPITAL INVESTMENTS 129,858 CAPITAL COST FOR PROPOSED 126 TABLE 813 " LAC TATE " PROCESS (1991 dollars with 577 mM lactic acid medium) COMPONENT NUMBER SIZE S (THOUSANDS) Corn dry mill 1 10,053 bu/day 3,299 Acid hydrolyzer 1 306,198 lb/day 323 Mme fermentors with heat exchangers 3 200,000 gal 965 changers + circulation pumps 880 ft2 Acetic acid fermentors with heat 6 500,000 gal 33,161 exchangers, power agitators, air filters and 6,000 ft2 compressors 1700 HP 8%”sdm [Neutralizers with agitators 0 0 0 entrifuges 3 750 g/min 4,496 300 HP [Protein feed dryer + loader 1 21,600 lb 4,042 HZO/hr 6 effect evaporator 1 52,500 lb 4,894 HzO/hr ICMA dryer 4» agglomerator 1 120,000 lb 5,890 HZO/hr SUB TOTAL 57,043 [Capital for off sites + start In) 57,043 TOTAL CAPITAL INVESTMENTS 114,085 127 TABLE 8c CAPITAL COST FOR PROPOSED " VINEGAR " PROCESS I COMPONENT NUMBER SIZE S (THOUSANDS) lCorn dry mill 1 15 ,255 bu/day 4,728 lAcid hydrolyzer 1 464,640 lb/day 462 actate fermentors with heat exchangers + 4 250,000 gal 1,471 circulation pumps 1,100 ft2 Acetic acid fermentors with heat 2 500,000 gal 11,054 exchangers, power agitators, air filters and 6,000 ft2 compressors 1,7000HP 8,900 scfm INeutralizers with agitators 3 8,000 gal 210 entn'fugg 3 250 gpm 2,332 IPrDtein feed dryer + loader 1 7200 lb H204" 2,091 effect evaporator 1 17,500 lb 2,532 HZO/hr ICMA dryer + agglomerator 1 40,000 lb 3,047 HZO/hr SUB TOTAL 27,928 [Capital for off sites + start up 27,928 TOTAL CAPITAL INVESTMENTS 55,928 OPERATING COSTS FOR PROPOSED (with 495 mM lactic acid medium) 128 TABLE 8d " LAC TATE " PROCESS I ITEM QUANTITY UNIT COST $lHR CENTS/LB ICDrn 837.7 $2.00/LB 1,675.4 10.05 IDolime 7,333 lbLhr 0.04 $/1b 293 8/hr 1.1752 lAcid 62.87 lb/hr 0.09 Mb 5.66 0.034 ICaustic 68.54 lbjhr 0.14 $/Ib 9.6 0.058 ICom steep liquior 0 0 0 0 IBentonite 726 0.0218 Mb 16 0.096 IKWH 1,832 0.05 KW/hr 92 .552 lCoaI 119.4 x 1063mm: 231106 BTU 238 1.428 ital recovery .119 crf 1.356 19.066 actor (10113., 15%) r 20 $50,000 128 .768 ITOTAL COSTS 3,418 22.88 IFIber credit 3.874.9lb/hr 0.0226 87.6 0.526 [Germ credit 5,192.9 lb/hr 0.0527 273.7 1.64 Animal feed credit 7,515.2 lb/hr 0.0733 550.9 3.306 NET PRICE 28.16 OPERATING COSTS FOR PROPOSED (with 577 mM lactic acid medium) 129 TABLE 86 " LACTATE " PROCESS l ITEM QUANTITY UNIT COST SIT-IR CENTS/LB lCorn 697.9 lb/hr $2.00jbu 1,395.8 8.38 lDolime 7,333 lb/hr 0.04 S/lb 293 Slhr 1.752 cid 52.37 lb/hr 0.09 $/lb 4.71 0.0283 lCaustic 57.1 lb/hr 0.14 $/lb 7.99 0.0479 ICDrn steep liquior 0 0 0 0 entonite 726 0.0218 $llb 16 0.096 IKWH 1,832 0.05 KW 92 .552 [Coal 119.4 x 1063mm: 25/106 BTU 238 1.428 pital recovery .119 crf 1,356 16.60 actor (10 yrs., 15%) or 20 $50,000 128 .768 [TOTAL COSTS 21.86 lFIber credit 3,228 lb/hr 0.0226 73 0.438 Germ credit 4,3261b/hr 0.0527 228 1.368 Animal feed credit 6260.8 0.0733 458.9 2.754 NET PRICE 25.76 130 TABLE 8f OPERATING COSTS FOR PROPOSED "VINEGAR" PROCESS I ITEM AUANTITY UNIT COST SIHR CENTS/LB ICDrn 635.5 bu/hr $1.00/bu 1,271 7.626 IDolime 7,333 lb/hr 0.04 $/Ib 293 1.752 IAcid 47.71b[hr 0.09 $/lb 4 0.024 austic 52.0 lb/hr 0.14 $/lb '7 0.042 om steep liquior 260 lb/hr 0.05 $/lb 13 0.078 entonite 726 Win 0.0218 S/lb 16 0.096 IICW'H 1,832 0.05/KWI-1 92 0.552 ICoal 119.4:106 BTU/hr 2 8/106 BTU 238 1.428 apital recovery 0.119 crf 1,356 8.136 actor (10 yrs., 15%) abor 20 $50,000/year 128 0.768 ITOTAL COSTS 3,418 20.50 IFIber credit 2,940 lb/hr 0.026 66 0.396 [Jenn credit 3,9401b/hr 0.0527 207 0.01242 Animal feed credit 5,702 lb/hr 0.0733 418 2.51 NET PRICE 16.36 131 The first step for estimating capital equipment costs was to determine the required equipment size for the current plant. If the desired equipment size varied from the 1988 study, a determination was made to either increase the number of equipment units, or to increase the size of the units. If the number of units was increased, the cost estimate was increased linearly from the 1988 study. Whenever the equipment size was changed, the ratio of the current size to 1988 equipment size was raised to the power of 0.6 and multiplied by the 1988 cost estimate: Current estimate = (1988 estimate) * (current size/1988 size)°6 This resulted in the values seen in Table 8a. All operating cost estimates were based on a plant capacity of 200 tons CMA produced per day, annual operation of 7,800 hours, and 20 full time operators. Further, all estimates were updated to 1991 dollars using Marshall and Swift price indices. These estimations predict the total capital investments of the vinegar process to be $55,928,000. The capital investments of comparative lactate process plants were found to be approximately $129,858,000 and $114,085,000 for facilities utilizing 495 mM and 577 mM lactic acid feeds. The corresponding operating costs result in a predicted price for finished CMA of 16.36 cents/lb, 28.16 cents/lb and, 25.76 cents/lb. 132 It is evident from these conclusions that CMA production using corn as a feedstock can be less expensive than production with currently used petroleum based techniques which result in CMA prices of 30 cents/lb. It is also apparent that without improvements to increase the final acetic acid concentration, the proposed two—step route utilizing a lactate fermentation and an acetic fermentation studied here does not result in the most economical method for CMA production. That distinction seems to be reserved for fermentation based on current vinegar technology. Research and data presented here indicates that the use of A. oxydans in the acetic fermentation of lactic acid does not result in the most cost effective method of CMA production at the present time. Clearly, for this process to become economically competitive, improvements must be made which would increase the production rate and the final concentration of acetic acid. I feel that future research has a good potential of leading to such improvements. 133 One area of the process which deserves further study involves biological manipulation of the bacteria itself. It is well known that bacteria are resourceful organisms which will quickly adapt to changing environments. Thus, by exposing cultures to selective pressure, one can produce a bacterial cell line that may exhibit desired traits. In this case, for example, it appears that bacterial growth is highly limited by product inhibition. Use of selective pressure may be able to decrease this inhibition. This would be accomplished by first establishing cultures grown in medium containing relatively high levels of acetic acid. These would then be used as inoculums for second generation cultures grown in medium containing an even higher acetic acid concentration. Presumably, this new generation would be better adapted for survival in the high acid environment. It would of course be used to inoculate a third generation in even stronger acetic acid medium. Eventually, it is feasible that an acetic acid tolerant culture may be developed which could be used in the fermentation. Mutation is another potential for microbial strain improvement. This method of manipulation is different than that previously mentioned in that instead of focussing on species adaptation, it actually alters the bacterial DNA. A common method of mutation is to expose the culture, 134 typically grown on solid agar medium, to ultra violet light. This produces the effect of changing the bacterium's DNA which may result in a mutant capable of living in environments that are toxic to the wild type strain. It is probable that adaptation, mutation, or a combination of both will lead to more resilient bacteria which exhibit increased production rates and higher final acetic acid concentrations. Only exhaustive research in this area, however, will show the extent of improvement. Further study of the process itself could also lead to higher conversion rates. Research in batch systems seemed to show that acetic acid production was growth associated. This finding suggested that maximum production rates should be seen in continuous fermentations where the culture would continually exist in an exponential growth phase. Unfortunately, CSTR operation resulted in lower than expected cell concentration and a correspondingly low conversion, possibly due to product inhibition. One potential method for increasing the cellular biomass of the fermentation broth during CSTR operation would be to install a cell recycle system. Such a system would use a series of tangential flow filtrations which would allow the product to be removed in the effluent stream, while recycling active cells back to the fermentor 135 for further production. To prevent the culture from increasing its cell mass to a point where it no longer exhibits exponential growth, it may be necessary to use only a partial rather than a complete recycle system. It would also be helpful to determine whether the main stumbling block for metabolism in continuous fermentation is due to the presence of free acetic acid. If this were found to be the case, increasing the fermentation's pH may be advantageous. As discussed earlier, raising the pH to levels of 7.5 appeared to have little or no effect on metabolism in batch systems. However, it may lead to increased growth and production in a CSTR by decreasing the concentration of free acid. A potential drawback of this action is the insolubility of MgO at pH values above 6.0. This would lead to production of a calcium acetate solution only, not one consisting of calcium and magnesium acetate. However, if this action were found to substantially increase production, it may be economically feasible to use it to create calcium acetate. A subsequent lowering of pH would then cause the dissolution of MgO and the formation of a CMA mixture. I feel that although chances of success with this concept are low, it deserves consideration. 136 A final concept that has the potential of increasing productivity involves splitting the acetic fermentation into a two—stage process. Research showed that overoxidation was not observed by A. aceti when the combined concentration of ethanol and acetic acid exceeded 5% (by volume) (31). Thus, a two-stage fermentation could be used with A. oxydans converting lactic acid to acetic acid in the first fermentation, and A. aceti (possibly using added ethanol as a second substrate) in the second phase. This scheme would utilize A. oxydans in the first fermentation to raise the acetic acid concentration to levels which would inhibit overoxidation in the second fermentation. This process would provide a use for surplus corn while simultaneously reducing losses from overoxidation. Hopefully, extensive research of these ideas will lead to improved acetic acid production. These improvements would ideally increase the capability of A. oxydans to produce acetic acid in a final concentration comparable to or exceeding that observed in the vinegar industry. If such process improvements can be found, the ”lactate process" described above will become more economical for CMA production and will deserve further analysis at the pilot plant level. Sodium chloride has been used for over thirty years as the primary highway deicer in northern and mountainous regions of the United States. It has the ability of lowering the freezing point of water 21 degrees Celsius, and acts quickly in removing snow and ice from roadways. Further, sodium chloride is found in plentiful supplies in the form of rock salt, at or near the earth's surface in much of the Snow Belt. This results in an easily obtained and inexpensive chemical for large scale use as a deicer. Unfortunately sodium chloride produces several adverse side effects which have invoked concern and apprehension about its use. A primary fault of NaCl is that it causes massive amounts of corrosion damage to vehicles, highway cement, bridges, and underground utilities. The monetary loss from this salt induced corrosion has been estimated to exceed five billion dollars each year, well over ten times its purchase and application costs. Another chief concern is that the extensive use of NaCl can result in health risks when salt laden run-off invades city water supplies. These health risks are particularly high in densely populated 137 138 areas such as Massachusetts, where nearly one third of the state's municipalities have complained of salt contamination in drinking water. Finally, environmentalists have lobbied against the over use of NaCl as a deicer for years due to its deleterious effects on roadside vegetation, and aquatic ecosystems. Combined, these side effects have cost the public an estimated 5.7 billion dollars annually. This astronomical figure prompted the Federal Highway Administration to sponsor research to find a substitute deicer. The result of this exhaustive search was that calcium magnesium acetate (CMA) was the most suitable candidate for large scale replacement of sodium chloride. CMA was found to attain the deicing capabilities of NaCl, while eliminating all of its undesirable traits. Numerous studies have proven that CMA is noncorrosive to concrete and steel, and in fact may be a corrosion inhibitor. Further, the results from experimental programs and theoretical literature reviews indicate that CMA has toxicological effects which are less severe than NaCl, has little or no effect on aquatic species, has little or no phytotoxic effect on roadside vegetation, and may provide some benefits to soil structure. Hypothetically, the wide scale use of CMA may even reduce the effects of acid rain by neutralizing sulfuric and nitric acids. 139 The only drawback to the use of CMA is found in the economics involved in its production. Current production techniques combine dolomitic lime and magnesia with glacial acetic acid to form CMA. This results in a product which sells for $640/ton delivered. The corresponding market price of rock salt is approximately $25/ton delivered. Thus, for CMA to be used as a replacement for NaCl on a national scale, methods must be developed to reduce its production costs. The goal of the research conducted here was to determine if these production costs could be lowered using a novel microbial fermentation in place of current production techniques. It was proposed to use corn as the feedstock for production of lactate. This lactate would then be used as the substrate for an acetic fermentation, similar to that currently used in the industrial production of vinegar. The proposed acetic fermentation would eliminate the problems of substrate loss due to high volatility, and overoxidation of product, two major concerns in vinegar production. This process would also provide a use for large amounts of excess corn seen in the United States today. To make an accurate economic analysis, it was necessary to find a suitable bacterial species for fermentation and determine its growth and corresponding acetic acid production characteristics. After a determination of the 140 conditions needed to optimize this production, it was necessary to analyze the various systems available (batch, semi—batch, and continuous flow) and determine which system was most productive at the optimal conditions. Once the most efficient process was determined, a first order economic study was conducted for comparison with current production techniques. Numerous tests have shown that of the organisms investigated, Acetomonas oxydans was the most promising species for the microbial production of CMA. Its growth on a 495 mM (29.7 g/l) lactic acid medium at 33 degrees Celsius, a pH of 6.0 (titrated with Ca/Mg OH and H2504), in a batch system sparged with air resulted in a final acetic acid concentration of 479 mM (28.7 g/l) in 73 hours (97% conversion of lactic acid to acetic acid). An identical fermentation grown with a 577 mM lactic acid medium for 96 hours yielded the highest final concentration of acetic acid (518 mM or 31 g/l) but showed a reduced production rate of 0.32 g/l/hr (5.33 mM/hr). Economic analyses of these two fermentations in conjunction with a proprietary lactate fermentation, compared with a hypothetical fermentation involving documented acetic acid fermentation technology from the vinegar industry, revealed that the microbial fermentation 141 of corn using each of these three processes has the potential of lowering the production cost of CMA. The most cost effective method is that of the hypothetical vinegar fermentation route where it is estimated that CMA can be produced for 16.34 cents per pound. Data and economic estimates showed that this price was increased to 25.76 cents per pound using the lactate fermentation in conjunction with the best-case acetic fermentation. Further research involving genetic engineering, bacterial acclimation, cell recycle systems, and two-stage fermentations, may reduce this estimate until it becomes competitive with the vinegar process. The conclusion of this study is that while calcium magnesium acetate cannot be supplied at a price comparable to NaCl, its production costs can be significantly lowered by replacing current production techniques with the microbial fermentation of renewable biomass. The most cost effective fermentation process would instil technology used in the vinegar industry. However the proposed fermentation route using Acetomonas oxydans to convert lactic acid to acetic acid is feasible and may become more economical if future work can increase the final concentration of acetic acid. 142 The ultimate decision of when and where to use CMA as a replacement for NaCl must take into consideration not only the market prices of the two deicers, but also the enormous indirect costs incurred by NaCl use, health concerns, and environmental considerations. When the initial cost of CMA is weighed against these issues, it is clear that replacement of NaCl by calcium magnesium acetate is the logical solution, especially if the results presented here can be duplicated and/or improved upon at the pilot plant level. BIBLIOGRAPHY / M3 BIBLIOGRAPHY Bryan, W. L.: Van Cauwenberge J. E.; Bathast R. J.: "Preliminary Evaluation of C. thermoaceticum tolerance to calcium magnesium acetate (CMA).', USDA, Agricultural Research Service 1815 N. University Street, Peoria, IL 61604. Gales, J. E.; VanderMeulen, J.: “Deicing Michigan's Highways: The Policy Dilemma“, Conference- Alternative Deicing Technologies and the Environment, Michigan State University, March 25- 26, 1991. Marynowski, Chester W.; Jones, Jerry L.: Tuse, D.; Boughton, R. L.: “Fermentation as an advantageous route for the production of an acetate salt for Roadway Deicing.'. WW. 1985, 24, pg. 457-465. Field, R.; O'Shea, M. L.: “The USEPA Research Program on the Environmental Impacts and Control of Highway Deicing salt pollution.', Conference-Alternative Deicing Technologies and the Environment, Michigan State University, March 25-26, 1991. Bertram, B.M.: Salt Institute Statistical Report Analysis, 1977-1987. Salt Institute, Alexandria, VI. ’ Wurster, C. F.; “Of Salt...', Neu_XQrk_Iimes, March 4, 1978, p. 21. . 10. 11. 12. 13. 14. M4 Murray, D. M.; Ernst, U. F. W. : “An Economic Analysis of the environmental impact of highway deicing.', EPA series EPA 600/2-76-105, EPA, Municipal environmental Research Lab., Cinc. OH, 45268, 1976. Hsu, M. T.; "Production and Testing of CMA in Maine", Transportation Research Record 962 pg. 77-82. Moll, R. A.; Rossmann, R.; Barres, J.; Horvath, F.: "The Historical Trends of Chlorides in the Great Lakes", Conference-Alternative Deicing Technologies and the Environment, Michigan State University, March 25-26, 1991. .Demers, C. L.: "Effects of road deicing salt on aquatic invertebrates in four adirondack streams", Conference-Alternative Deicing Technologies and the Environment, Michigan State University, March 25- 26, 1991. Hudson, L. R.: ”CMA from low—grade biomass.', Proceedings from IGT Conference XI on Energy from biomass, Orlando FLA., March 18, 1987. Dunn, S. A. ; Schenk, R. U.: "Alternative Highway Deicing Chemicals“, Report # FHWA-RD-78-108, Dept. of Transportation, March 1980. McFarland, B. L.: O'Reilly, K. T.: "Environmental Impact and toxicological characteristics of calcium magnesium acetate", Conference-Alternative Deicing Technologies and the Environment, Michigan State University, March 25—26, 1991. Mulligan, T. J.: 'Calcium.magnesium acetate biodegradation and its impact on surface waters“, Conference-Alternative Deicing Technologies and the Environment, Michigan State University, March 25-2 6, 1991. 15. 16. 17. 18. 19. 20. 21. 22. M5 McCrum, R. L.: "CMA and NaCl as deicing salts - a comparative study”. Enxironmental_Ireatment_and_ ggntrgl, Dec. '89. pg. 24-28. ' Adams, E. E.; Alger, R. G. Chekan, J. P. Williams, F. D. Valverde; R.: ”Persistence of reduced snow to pavement shear strength for two aggregate materials treated with calcium magnesium acetate (CMA) and sodium chloride(NaCl)”, Conference-Alternative Deicing Technologies and the Environment, Michigan State University, March 25-26, 1991. Miller, W. L.: "Calcium magnesium acetate use on I-75 and the Zilwaukee bridge in Michigan.', Conference- Alternative Deicing Technologies and the Environment, Michigan State University, March 25-2 6, 1991. Van Regenmorter, W.: "Winter highway maintenance: A legislative perspective“, Conference-Alternative Deicing Technologies and the Environment, Michigan State University, March 25-26, 1991. Pollock, S. J.: 'Remediating highway deicing salt contamination of public and private water supplies in Massachusetts“, Conference-Alternative Deicing Technologies and the Environment, Michigan State University, March 25-26, 1991. Hudson, L. W.: New York State Energy Research and Development Authority, Albany N. Y. personal communication (1987). "Chevron deicing chemicals. Ice-B-Gon questions and answers.', Chevron Chemical Co., San Ramon, Cal. 1988. Allgeier, R. J.; Nichol, G. B. ;Conner, H. A.: “Vinegar: History and Development”, Eggd Eroduction_and_neyelcnment. 8(5), 1974. pg. 69-71. 23. 24. 25. 26. 27. 28. 29. 30. 31. M6 Mayer, E.: "Historic and modern aspects of vinegar making (acetic fermentation)", EQQd_Iecthngy, 532, May 1963 vol. 17. ' ' ' , Williams & Wilkins, Baltimore/London 1984. Rainbow, C.: “The biochemistry of the Acetobacters', WWW Vol III London Heywood Company, 1961. Greenshields, R. N. “Primary products of metabolism: Acetic acid: Vinegar" Wham Vol 2, 1978. Dale, H. W.: Bacterial_Metabglism, Academic Press, New York, 1969. pg. 307-327. De Ley, J. ; Schell, J. ”Lactate and Pyruvate Catabolism in Acetic Acid Bacteria. ' J9nrna1_gf_ W 1962 29 pg 589- 601 Bailey, J. E. - Ollis, D. F. W McGraw- Hill Book Company, 1986. Grethlein, H.; Jain, M.: presented at the “Second National Corn Utilization Conference of the National Corn Growers Association", Columbus, Ohio, Nov. 17-18, 1988. Asai. T.: Acetic_Acid_Bacteria1__Classification_and_ .Biochemical_Actiyities University Park Press, Baltimore, 1968.