Applied Microbiology and Biotechnology

, Volume 71, Issue 5, pp 587–597

Bacterial acetone and butanol production by industrial fermentation in the Soviet Union: use of hydrolyzed agricultural waste for biorefinery

Authors

  • V. V. Zverlov
    • Institute for MicrobiologyTechnische Universität München
    • Institute of Molecular GeneticsRussian Academy of Science
  • O. Berezina
    • Institute of Molecular GeneticsRussian Academy of Science
  • G. A. Velikodvorskaya
    • Institute of Molecular GeneticsRussian Academy of Science
    • Institute for MicrobiologyTechnische Universität München
Mini-Review

DOI: 10.1007/s00253-006-0445-z

Cite this article as:
Zverlov, V.V., Berezina, O., Velikodvorskaya, G.A. et al. Appl Microbiol Biotechnol (2006) 71: 587. doi:10.1007/s00253-006-0445-z

Abstract

Clostridial acetone–butanol fermentation from renewable carbohydrates used to be the largest biotechnological process second only to yeast ethanol fermentation and the largest process ever run under sterile conditions. With the rising prices for mineral oil, it has now the economical and technological potential to replace petrochemistry for the production of fuels from renewable resources. Various methods for using non-food biomass such as cellulose and hemicellulose in agricultural products and wastes have been developed at laboratory scale. To our knowledge, the AB plants in Russia were the only full-scale industrial plants which used hydrolyzates of lignocellosic waste for butanol fermentation. These plants were further developed into the 1980s, and the process was finally run in a continual mode different from plants in Western countries. A biorefinery concept for the use of all by-products has been elaborated and was partially put into practice. The experience gained in the Soviet Union forms a promising basis for the development of modern large-scale processes to replace a considerable fraction of the current chemical production of fuel for our future needs on a sustainable basis.

Introduction

It is generally accepted that the fossil sources for our energy supply have to be gradually replaced by renewable substrates, and scientific scenarios for sustainable alternatives have been developed (Stout 1982; Dale 1987; Lynd et al. 1991; Claassen et al. 1999; Wyman 2001; Thomas et al. 2002; Greene et al. 2004; Lynd et al. 2005). For the future production of commodity products such as organic fuels, chemicals, and materials as well as food and animal feed within the frame of a biorefinery concept, lignocellulosic biomass from plants provides the only readily available and quantitatively sufficient renewable substrate (Lynd et al. 1999). In this context, the most valuable and versatile commodities in our mobility and chemical technology are liquid energy carriers such as ethanol followed by biodiesel, which together already greatly contribute to reduce the release of the greenhouse gas CO2 as demanded by the so-called Kyoto protocol (Galbe and Zacchi 2002; Greene et al. 2004; Demain et al. 2005; United Nations Framework Convention on Climate Change).

However, other solvents including butanol are equally useful and, for special purposes, superior to ethanol. These solvents have very little public recognition. Butanol could, for example, gradually replace gasoline as well as diesel due to its high energy content, miscibility, octane-improving power, low volatility, and other positive qualities (Schwarz and Gapes 2006). The economics of butanol fermentation are favorable even with present day technology (Gapes 2000a), and considerable improvements in every single step of production are foreseeable with stringent and consistent research and development efforts (Ezeji et al. 2003; Greene et al. 2004).

In addition to its usefulness for the biofuel sector, butanol is a valuable C4 compound for chemical synthesis for which it is presently chemically synthesized from fossil-oil-derived ethylene, propylene, and triethyl-aluminum or carbon monoxide and hydrogen. Despite steadily growing production, the market for butanol remains tight and the price high because of low investment and hefty demand. An alternative source for n-butanol is anaerobic bacterial (Clostridium) fermentation in conjunction with acetone and a small amount of ethanol, the so-called AB or ABE fermentation. However, for establishing the production of the quantities needed for our fuel market, a radical expansion of the available substrates has to be realized, e.g. by including cellulosic plant biomass.

A number of reviews have been published on AB fermentation (Jones and Woods 1986; Awang et al. 1988; Dürre 1998; Nimcevic and Gapes 2000) and attempts have been made to revive the AB fermentation with modernized technology for fermentation and downstream processing as well as with alternative renewable substrates like lignocellulosic biomass hydrolyzates (Nativel et al. 1992; Gapes 2000b; Ezeji et al. 2005).

The development of AB fermentation

The conversion of plant biomass into solvents for fuel and chemical industry is in principle an old technology. In fact, the fermentation of sugar and starch to ethanol can be regarded as the oldest biotechnology (for beer and wine production) and probably always has been the biotechnology of the largest scale. After earlier attempts by L. Pasteur and others, the fermentation of starch to the solvents acetone, butanol, and ethanol (AB fermentation) was developed by C. Weizman in 1912 at Manchester University into an industrial process (Dürre 1998). He isolated the strain which was later called Clostridium acetobutylicum and ran the first production plant for acetone production from starch. Because of the strategic need for large amounts of acetone for cordite production and constant problems with substrate delivery during World War I, the first large-scale industrial plants were erected in Canada and USA. Butanol was an unnecessary by-product during the war. But with the increasing needs for butanol after the war, the fermentation industry changed the substrate to molasses and screened for new clostridial strains. After the Weizman patent died in the 1930s, large production facilities were erected in USA, Japan (Kyowa Hakko), and South Africa (National Chemical Products in Germiston), among others. One of the best production strains isolated was NCP262 which is now the type strain of Clostridium saccharobutylicum (Keis et al. 2001).

After a peak in the 1950s, the productivity of the AB fermentation plants in the Western industrialized countries declined constantly due to persistent problems with fermentation reliability because of frequent bacteriophage infections and decreasing quality of molasses by improved sugar processing technology. Moreover, the price of molasses was increasing as it was used as a feed additive for pig breeding. Basic chemicals and fuels were also produced cheaper from petroleum oil by the upcoming petrochemical industry. As a result, AB fermentation was no longer profitable and was discontinued in the industrialized Western countries during the 1960s. The production facilities, the research laboratories, and most of the strain collections and production notes were destroyed. Only a few plants survived into the 1980s, including the plant in Germiston, South Africa. Biological production of AB was probably pursued even later in the Republic of China and in Egypt. Not much was known about the large production facilities in the Soviet Union except a short report (Nakhmanovich and Shcheblykina 1959).

In spite of the unfavorable economic situation, interest in the AB fermentation has not ceased and research has continued in the assumption that alternative substrates, especially lignocellulosic biomass, have a bright future. This was based on consideration of forthcoming oil crisis and increase of greenhouse carbon dioxide caused by fossil oil consumption (Kyoto protocol, Amendment to the United Nations Framework Convention on Climate Change). Hydrolyzates containing the pentoses and hexoses of a wide range of agricultural waste materials such as wood chips or corncobs would alleviate these problems (Yu et al. 1984; summarized in Jones and Woods 1986). Research has progressed in the microbiology, biochemistry, and genetics of AB-producing bacteria as well as in substrate preparation, sterilization, fermentation and downstream processing technology. In addition, improved hydrolysis technology, now under development for bioethanol production, would probably make the AB production cost effective again.

Butanol-producing strains of Clostridium

Over the years, a large number of solventogenic clostridia have been described. They were given a variety of names without regard to phylogenetic characterization. In spite of the high number of isolates, only about 40 solventogenic Clostridium strains have survived in public strain collections. However, for the screening of new traits like use of other substrates, phage resistance, other product ratios or products, etc., a huge pool of strains with a wide genetic variety would be a prerequisite for a revival of a future AB fermentation from renewable plant biomass (other and cheaper than starch).

Several solventogenic clostridia have been investigated on the molecular level. They belong to cluster I of the clostridia and were classified into four species by similarity in their 16S rDNA sequences and DNA–DNA homology (Collins et al. 1994; Wilkinson and Young 1993; Jones and Keis 1995; Keis et al. 1995, 2001; Dürre 1998). Shaheen et al. (2000) have found that members of the four species differed significantly in solvent-producing ability (between 10 and 24 g l−1) and solvent yield (between 6.8 and 33.2 %). Strains also differed considerably in the utilization of starch and sucrose (molasses). The clostridia isolated in the USSR are not yet characterized phylogenetically but would add favorably to widen the genetic basis of solvent-producing industrial bacteria.

C. acetobutylicum harbors a large plasmid which carries the genes for solventogenesis. Loss of the plasmid causes the instability leading to degeneration of the bacteria during long fermentations which is characterized by acid accumulation without a switch to solventogenesis (Kashket and Cao 1995; Cornillot et al. 1997). As sporulation genes are also located on the plasmid, repeatedly growing freshly germinated cultures from spores has been a means for maintaining AB production in the cultures. In Clostridium beijerinckii (and probably also other butanologenic strains), the solventogenic genes are located on the chromosome (Wilkinson et al. 1995). The degeneration observed with these strains may have another cause and/or be less dominant.

Aside from the acetobutylicum group of solventogenic clostridia, other species have been shown to produce acetone and butanol. Clostridium ljungdahlii is able to utilize carbon monoxide and hydrogen from synthesis gas (reformer gas) as energy and carbon source. More recently, solvent-producing strains of Clostridium butyricum have been described, which show enhanced hemicellulolytic activity (Montoya et al. 2001).

The chromosome and the megaplasmid of C. acetobutylicum have been sequenced (Nölling et al. 2001) and the genes involved in acid and solvent production have been identified (Dürre 1998). The genomes of C. beijerinckii and C. ljungdahlii are presently being annotated. An intriguing finding in the C. acetobutylicum sequence is the presence of a seemingly complete operon with a set of genes coding for a cellulosome (Nölling et al. 2001; Schwarz 2001). This should enable the organism to degrade crystalline cellulose and, thus, permit the production of solvents directly from cellulosic biomass in a consolidated bioprocess (Lynd et al. 2005). However, only fairly active mini-cellulosomes were found in C. acetobutylicum ATCC 824 cultures (Sabathe et al. 2002; Lopez-Contreras et al. 2004) and attempts to correct reading frame errors and improvement of promoters did not lead to substantial cellulase production (Sabathe and Soucaille 2003).

Direct fermentation of cellulose-containing biomass to butanol in a consolidated bioprocess with a cellulolytic, solventogenic bacterium remains a future task. Repeated screenings in a number of laboratories have shown that none of the available strains was able to hydrolyze native cellulose (P. Soucaille, personal communication; V. Zverlov, unpublished results). On the other hand, a process using extracellular hydrolysis by enzymes or by chemical or physical pre-treatment (as is presently under development for production of bioethanol) could be applied to furnish lignocellulosic substrates for microbial conversion to solvents.

Fermentation of pentoses

In contrast to most ethanol-producing microorganisms, AB-producing bacteria are able to produce solvents from hemicellulose-derived pentoses. The solventogenic clostridia described thus far ferment glucose, sucrose (in molasses), and starch via the Embden–Meyerhof pathway (summarized in Andreesen and Gottschalk 1969). In addition, they also utilize glycerol and a number of other hexoses, pentoses, and oligosaccharides like cellobiose, lactose, raffinose, mannose, xylose, and arabinose (Hazlewood and Gilbert 1993; Keis et al. 2001; Andrade and Vasconcelos 2003). The pentoses are metabolized by the pentose phosphate pathway via pentose-5-phosphate, resulting in the production of fructose 6-phosphate and glyceraldehyde-3 phosphate which enter the glycolytic pathway.

A number of strains have been shown to produce AB readily from xylose (Langlykke et al. 1948; Nakhmanovich and Shcheblykina 1959; R. Gapes, unpublished data). It is, thus, not surprising that solventogenic clostridia can be used to ferment hydrolyzates of hemicellulose, including pentosans, to butanol as was suggested by Waksman and Kirsh (1933). Solventogenic clostridia could thus be promising for the conversion of hemicellulose and pectin, as well as starch or sucrose to solvents at an industrial scale.

Continuous fermentation

A truly continuous process is of particular interest for the industrial-scale fermentation because substrate preparation and sterilization can be run continuously as well as downstream processing of the fermentation products. Truly continuous fermentation processes are usually more energy efficient. However, the continuous AB process is difficult and was not achieved in industrial plants. In textbooks, the AB process is generally regarded as a batch process with an acidogenic phase during which short-chain fatty acids, mainly acetate and butyrate, accumulate, followed by the solventogenic phase where the acids are converted to the reduced end products butanol and acetone. The solventogenic phase is initiated by the accumulation of undissociated fatty acids, and the bacterial culture stops growing and enters the sporulation phase (Terracciano and Kashket 1986; Maddox et al. 2000; Dürre et al. 2002). The molecular signal for the switch has not been identified. Overproduction of acids is frequently observed and results in acid crash, i.e., inactivation of the culture and absence of solvent formation. The entire batch is lost (Maddox et al. 2000). These phenomena were observed by C. Weizmann with his early production efforts.

Attempts to develop a continuous fermentation have been made in a number of ways: pH control, feeding of butyrate or acetate, phosphate or nitrogen limitation, or cell immobilization have been used to control culture stability (Iarovenko et al. 1960; Bahl et al. 1982; Gottschal and Morris 1982; Haggstrom 1985; Groot et al. 1992; Girbal and Soucaille 1994; Mutschlechner et al. 2000; Huang et al. 2004). Chemostat cultures have been kept stable for up to 70 days with growth on a glucose/glycerol mixture (Andrade and Vasconcelos 2003). The degeneration grade of a culture can be monitored online (Schuster et al. 2001). Although these methods would improve an industrial process considerably, none of them has yet been used in a production plant. However, parts of these advancements in research have been realized in the Russian Dokshukino process which we are reporting below.

AB fermentation in the former USSR

Very little information regarding the industrial AB fermentation process which operated in the former USSR has been published in international English language literature (some are summarized in Jones and Woods 1986; Nimcevic and Gapes 2000). After Glasnost, the international exchange of scientific information was intensified and a considerable amount of publications, patents, and internal reports for both production details and research carried out on the AB fermentation process is translated and discussed here (e.g., Yarovenko et al. 1963).

The era before and after World War II in the Soviet Union was characterized by a central planned economy. The state-run agriculture produced large amounts of waste from barley, rye, wheat, and potato processing which was exploited as a biomass resource for the production of basic materials for the strategically and politically important chemical industry.

An AB industry was first established in the USSR during the period 1929–1935. The initial process was utilizing C. acetobutylicum in a starch-based fermentation that was essentially similar to the Weizmann process developed in England, Canada, and the USA during the First World War: butanol and acetone from wheat and rye starch. A number of plants were erected throughout the country. They initially operated in batch mode. However, this fermentation method was inefficient and the large amounts of starch necessary for an industrial butanol production resulted in shortages of substrate supply. Part of the starch was consequently later replaced with molasses from sugar beet refineries (Table 1).
Table 1

Yield of solvents in three production plants (1961–1962)

 

Average value 1960

Sept. 1961

Oct. 1961

Nov. 1961

Dec. 1961

Jan. 1962

Feb. 1962

Mar. 1962

Apr. 1962

May. 1962

Jun. 1962

Average value for 10 months

Dokshukino

 Molasses (%)

67.2

69.9

69.1

70.0

69.2

68.4

66.4

60.7

52.4

43.5

63.7

 Hydrolysate (%)

3.5

3.6

3.6

3.6

4.0

4.2

5.0

4.9

6.3

5.5

4.4

 In all (%)

 

70.7

73.5

72.7

73.6

73

72.6

71.4

65.6

58.7

49.0

68.1

 Acetone

124

103

108

108

105

93

96

99

94

102

96

100

 Butanol

215

218

214

209

203

207

206

203

198

213

205

208

 Ethanol

40

38

36

42

46

58

60

56

65

75

86

56

 Total (kg/t starch)

379

359

358

359

354

358

362

358

357

390

387

364

Grosnyi

 Molasses (%)

61.2

44.0

40.9

28.8

60.0

56.9

34.2

51.1

62.4

57.6

49.7

 Acetone

121

105

108

104

110

99

91

110

98

101

86

101

 Butanol

218

208

205

196

205

204

201

208

208

223

199

206

 Ethanol

43

35

44

37

41

46

49

40

45

53

48

44

 Total (kg/t starch)

382

348

358

337

356

349

341

358

351

377

333

351

Talitsk

 Molasses (%)

61.6

63.0

67.0

63.1

50.6

56.5

52.8

54.0

55.4

50.8

57.5

 Acetone

120

97

105

102

92

99

101

96

100

101

100

99

 Butanol

215

204

213

215

191

198

204

198

208

209

207

205

 Ethanol

45

44

45

46

44

46

43

45

42

48

48

45

 Total (kg/t starch)

380

345

363

363

327

343

348

339

350

358

355

349

For comparison, the average yield from 100 % flour for the year 1960 is included. The values show kilograms of solvents produced per ton of sugar added (in starch equivalents). In addition to flour, the plants in Grosnyi and Talitsk used molasses, the plant in Dokshukino was run with wheat flour and addition of molasses and pentosane hydrolysate. Percentages indicate the amount of flour replaced by the other sugar sources. Molasses and hydrolysate are given in percentage of sugar content per starch equivalent in flour

The subsequent development of the AB fermentation technology in the USSR occurred largely independent of that in other countries. After the Second World War, the AB fermentation process in the Western countries was not developed much further and, by the 1960s, most of the industrial plants had been shut down. The Russian AB industry accumulated considerable experience with the handling of bacterial strains and with the fermentation technology under the guidance of a central research institute run by the Dokshukino plant. New technology and bacterial isolates were developed and tested in the full-scale production plant in Dokshukino. Until recently, the process technology was further improved based on scientific investigations. This experience may help to re-establish AB fermentation in the near future.

At least eight industrial AB fermentation plants were working in the USSR and some of them were used up to the late 1980s. The storage facility and substrate preparation of the Dokshukino plant is shown in Fig. 1 and part of the fermentation vessels in Fig. 2. Because the USSR lost economical power and because cheaper alternative petrochemical processes became available, all the acetone–butanol plants were shut down at the end of the 1980s. The fermentation facilities were demolished and the sites converted into other production.
https://static-content.springer.com/image/art%3A10.1007%2Fs00253-006-0445-z/MediaObjects/253_2006_445_Fig1_HTML.jpg
Fig. 1

Storage tanks for sulphuric acid and mounds of plant waste (background) at the Dokushukino AB production plant (from Yarovenko et al. 1963)

https://static-content.springer.com/image/art%3A10.1007%2Fs00253-006-0445-z/MediaObjects/253_2006_445_Fig2_HTML.jpg
Fig. 2

Two of the 225-m2 fermenters at Dokshukino plant. The water jackets for heating the fermenters are visible (from Yarovenko et al. 1963)

The basic Russian AB process

The basic Russian AB process was initially carried out as follows: A fermenter was filled with sterilized substrate (Fig. 2), a bacterial culture was added, the fermentation proceeded from the acidogenic phase, through a transition to the solventogenic phase, and to distillation of the fermentation broth. The fermentation vessels were kept at 37 °C (not 32 or 34 °C). The pH and the content of volatile fatty acids and solvents were assayed at intervals. When the maximum solvent concentration had been reached, the cultures were pumped into a separation line to separate the liquid and insolubles fractions. The liquid was processed by continuous distillation into separate solvent fractions, and the insolubles sludge was transferred to a fermenter for thermophilic methanogenic fermentation or into a yeast–protein plant for fodder production (Fig. 3).
https://static-content.springer.com/image/art%3A10.1007%2Fs00253-006-0445-z/MediaObjects/253_2006_445_Fig3_HTML.gif
Fig. 3

The biorefinery concept used for AB fermentation plants in Russia. The major substrates and products are shown in bold boxes. Only the processes of the biorefinery concept marked in bold print were realized (modified after Yarovenko et al. 1963)

This basic outline of the process was gradually modified to optimize the production and to make the overall process more economically viable (Fig. 3). Especially with the establishment of the yeast production, the overall process reached profitability and at the same time reduced the amount of organic sludge to be disposed (Yarovenko et al. 1963).

Russian research approach

The Dokshukino plant (Fig. 2) was a full-scale production plant with initially moderate production yields. It was used for the development and testing of new technologies which were introduced step by step into the industrial process and, once successful, integrated in the other industrial plants. The central research laboratory in Dokshukino was fully engaged in all aspects of the process, including the isolation of new strains, the development of more effective substrate preparation, and the introduction of new substrates. Fermentation technology improvements included equipment design, downstream processing of the solvents, and by-product utilization (see below).

At the same time, accompanying research was focused on problems that arose during industrial-scale fermentation, such as infections by bacteriophages and acid-tolerant bacteria, toxic by-products generated during substrate sterilization and hydrolysis, and foam development during fermentation. Among other improvements, a separation line to extract vitamin B12 from the archaebacteria grown up in the thermophilic methanogenic fermentation was established in 1967 (Evremovo) and in 1969 (Grosnyi). A total of 400–600 μg of cobalamin (excluding cobalt ion) was obtained per liter of fermentation broth (Fig. 3). This process was economically viable in itself and contributed to the profitability of the AB plants.

Two major improvements of the AB fermentation from biomass were obtained: (1) a continual flow process which had great advantages over the batch mode, and (2) use of agricultural waste material by hydrolyzing the hemicelluloses; this extended the amount of raw material for production.

Russian continual fermentation

During the heyday of the acetone–butanol fermentation in the Western world and also in the Soviet Union, the preparation and sterilization of the fermentation broth was done in continuous mode with heat recovery to reduce energy requirements to a minimum. This technology was adapted in the USSR from 1958 to 1960. Continuous distillation was similarly used, although batch facilities were also in operation. However, the traditional batch mode fermentation over 35 to 40 h in ca. 100 m3 fermenters used in the Western world was developed further in the Soviet Union where fermenters as large as 275 m3 working volume were used. Still larger working volumes of more than 2,000 m3 were achieved since 1939 by linking batch fermenters in series whereby only the first fermenter was inoculated with a fresh culture (Fig. 4). The flow through the battery was halted after about 50 to 60 h before the onset of fermenter acidification by culture degeneration or breakdown due to infections (Fig. 5).
https://static-content.springer.com/image/art%3A10.1007%2Fs00253-006-0445-z/MediaObjects/253_2006_445_Fig4_HTML.gif
Fig. 4

Schematic drawing of a single battery at the Dokshukino plant. 1, flour medium preparation, mixing vessel; 2, preheating; 3, solid separation; 4, delivery pump; 5, column heater; 6, pressured incubation vessel for sterilization; 7, steam separator; 8, cooler; 9, seed fermenter; 10, fermenter; 11, molasses and hydrolyzate preparation, mixing vessel (from Yarovenko et al. 1962)

https://static-content.springer.com/image/art%3A10.1007%2Fs00253-006-0445-z/MediaObjects/253_2006_445_Fig5_HTML.gif
Fig. 5

Flow diagram of the continuous flow fermentation indicating the sterilization, loading, fermentation and unloading of each fermenter in the four batteries at the Dokshukino plant. The direction of fermenter loading is reversed at the ends of the batteries. sf, seed fermenter (from Yarovenko et al. 1963)

Although a length of 50 to 60 h is too short and “steady-state” conditions may not have been present in all fermenters, this was perhaps a first step towards a truly “continuous” acetone–butanol fermentation. To increase overall site production, parallel batteries of reactors connected in series were used. As an example, the Dokshukino plant consisted in year 1962 of four parallel batteries—one of four, two of seven and one of eight fermenter vessels each with 225–275 m3 working volumes (Figs. 2, 4, and 5). In comparison, the ca. 100 m3 batch fermenters used in the West were run in batteries of 20 or more and were run in parallel but staggered with respect to time. This enabled truly continuous substrate preparation and truly continuous distillation and this batch process can be termed “continual” fermentation (as opposed to “continuous” fermentation). The final design of the process was tested at the Dokshukino plant in year 1962 and transferred to other plants.

A flow scheme indicating the cleaning, filling, and fermentation times of the four fermenter batteries run in parallel at the Dokshukino plant is shown in Fig. 5. In comparison to the batch mode, the productivity of the continual fermentation was increased by 31 %. The total volume processed during one cycle of continual fermentation in one battery was approximately 2,000 m3. Table 1 shows the average fermentation results of the first 10 months when the continual mode of fermentation was introduced.

Problems with the fermentation

A commonly reported problem with a C. acetobutylicum culture is strain degeneration (see above). The fermentation scheme of the continual fermentation as performed in Dokshukino was adapted to the number of generations possible in one cycle so that degeneration was not a problem in the Russian AB fermentation.

Another problem reported frequently in the Western fermentation plants were bacteriophage infections. An advantage of Russian AB industry was the use of independently isolated C. acetobutylicum strains which seem to be rather stable to bacteriophage infections. They also produced solvents at higher temperatures (37 °C) than most other industrial strains. The strains were isolated from grain and improved during a multi-cycle selection. Problems with bacteriophages were reported infrequently in the Russian files: this is probably also a result of a rigorous sterilization scheme and the hot steam sterilization of tubings and fittings (150 °C). However, problems with infection by lactic acid bacteria have been reported repeatedly. This was the most frequent cause of low yield and limited the time a cycle could be run reliably to a maximum of 60 h.

Acid hydrolyzates of alternative substrates

In an attempt to reduce the utilization of food-grade material such as rye flour and molasses, an effort was made to use pentose hydrolyzates of agricultural waste material like hemp waste, corncobs, and sunflower shells. These posed a waste problem in some agricultural areas of the Soviet Union and were abundantly available.

The complete hydrolysis of lignocellulosic biomass (e.g., wood) to sugars by treatment with high sulfuric acid concentrations at high temperature according to Scholler (Riehm 1962) has a long tradition in Russia and is still a process performed at large scale for ethanol production from wood chips. However, it yields mainly glucose from cellulose for fermentation with yeast to ethanol (Table 2, “hexose hydrolyzate”), whereas pentoses are largely degraded at the high temperature used (160–180 °C) and with the concentrated sulfuric acid treatment. Pentoses are useless for traditional yeast fermentation to ethanol in any case.
Table 2

Pentose and hexose content of a sample hydrolysate

Content of sugars by chromatography analysisa

Corncobs

Sunflower shells

Hemp waste

Pentose hydrolysateb

Hexose hydrolysatec

Pentose hydrolysate

Hexose hydrolysate

Pentose hydrolysate

Hexose hydrolysate

Glucose

5

90

2–3

91

7–10

87

Xylose

75

8

70

6

78–85

10

Arabinose

15–17

0

22–25

0

5–7

0

Galactose

2–3

0

5

0

2–4

0

Mannose

0

2–3

0

3

0

3

Sugars from dry mass, % (w/w)

39

nd

21

nd

22

nd

For conditions of hydrolysis, see text. Adapted from Logotkin (1958)

nd Not determined

aIn percentage (w/w) of total sugar

bLow acid and temperature (hemicellulose hydrolysis)

cHigh-temperature hydrolysis (cellulose hydrolysis)

For AB fermentation, the technology of pentose hydrolyzates from hemicellulose with dilute sulfuric acid at moderately high temperature was developed from April 1959 to January 1961 at the Doshukino plant research laboratory for the production of pentose syrup containing mainly xylose and arabinose with minor amounts of glucose and galactose (Table 2, “pentose hydrolyzate”). The partial hydrolyzate containing the pentoses gave better fermentation results than the complete hydrolyzate using the harsher conditions which contained mostly glucose, but obviously also more toxic by-products.

The plant biomass was ground to powder, diluted 1:10 (g/ml) with 1 % (v/v) sulfuric acid and heated to 115–125 °C. The time of hydrolysis (1.5 to 3 h) depended on the substrate and the process temperature, with the shorter times for the corncobs which are the easiest hydrolyzed and the longer for sunflower shells. Hydrolyzates were neutralized by addition of lime (CaOH2). In addition to the kind of substrate, the mode of hydrolysis had an influence on the fermentation quality. The sugar yield in “pentose hydrolyzates” obtained with this method is indicated in Table 2. In laboratory trials with a high concentration of hydrolyzate (up to 50 % w/w of the sugar content in the medium), the maximal yield of solvents reached 36 % (36 g of solvents from 100 g of initial sugars), with an average yield of 32 %. In comparison, the yield of solvents from flour starch was 35–36 % (Yarovenko et al. 1963). However, fermentation times were increased by addition of hydrolyzates. But replacement of less than 20 % of flour starch by pentosan hydrolyzates did not decrease the speed of fermentation and the yield of solvents. Compared with the other hydrolyzates, C. acetobutylicum fermented most readily the corncob hydrolyzates.

The composition of solvents from the fermentation of pentose hydrolyzates shifted slightly towards ethanol production with an increase of 15–30 % compared to the fermentation of pure starch (Table 1). This is probably due to the fermentation of calcium acetate present in hydrolyzates (about 3 %).

Selected production data of the years 1960 to 1962 are shown in Table 1. For comparison, the fermentation yields from pure starch are also shown (cumulated data of 1960). The slight decrease in production yield by addition of molasses and molasses + hydrolyzate after September 1961 is more than compensated for by the lower cost and broader substrate basis. These data show that over 70 % of flour starch could be substituted with molasses and pentosane hydrolyzate with consistent and reliable results in solvent production. From May 1962, the fermentation yields improved considerably with growing experience in handling the fermentation schemes and with reduction of the added molasses, whereas hydrolyzate additions increased (Table 1). It should be mentioned that hydrolyzate-containing substrate mixtures were fermented more easily in continual flow mode than in batch fermentations.

The cellulose hydrolyzates from a second hydrolysis step with harsher hydrolysis conditions (Table 2 “hexose hydrolyzate”) were planned for use at a later stage of the development but, to our knowledge, this change was not implemented in an industrial scale up.

By-products of hydrolysis

The major by-products of acid hydrolysis were acetic and formic acids and furfurol. These substances can be toxic to C. acetobutylicum cultures, but their concentration depends on the hydrolysis method and the substrate used. Wood hydrolyzates obtained with concentrated sulphuric acid were found to be toxic for Clostridium fermentations but could be used for Saccharomyces fermentations (“wood-ethanol”). However, with mild hydrolysis of corncobs as described above, the concentration of toxic substances is low: 0.27 % formiate and 0.1 % furfurol (v/v). The maximum level for formiate tolerated was 0.02 %, and 0.25 % for furfurol as shown experimentally by gas formation (Table 3). Acetic acid was not toxic and was obviously metabolized to ethanol.
Table 3

Effect of furfurol on gas production during AB fermentation

Furfurol concentration (% w/v)

0

0.05

0.10

0.20

0.30

0.30

Time  (h)

Development of gas (g/l)

11

4.60

3.65

1.50

1.65

0.50

0.20

19

12.10

12.85

6.75

7.15

1.10

0.25

23

18.6

21.85

13.25

12.50

3.35

35

28.5

28.35

25.00

23.25

17.35

2.60

40

29.0

28.75

26.75

25.00

21.90

63

29.5

28.85

28.50

27.15

26.10

8.70

83

29.9

29.60

28.75

28.25

27.50

9.70

Aside from formic acid and furfurol, other inhibitory substances after acid hydrolysis were melanoids which were formed by Maillard reactions of amino acids with sugars. They affected the fermentation negatively by resulting in over-acidification of the culture. Hydrolysis and heat-sterilization at lower temperature decreased the formation of melanoids in the substrate mixture. Arsenate above 0.001 % was highly inhibitory; it was probably introduced as a contaminant of the technical grade sulphuric acid. No problems of culture inhibition by the acrylamide formed by boiling or autoclaving the starch were seen, in contrast to other reports (Ezeji et al. 2003).

The amino acids arginine, aspartic acid, histidine, glutamine, glutamic acid, α- and β-alanine, lysine, proline, serine, tyrosine, and cysteine were by-products of hydrolysis. They were beneficial for the growth of the bacteria.

In spite of the mild conditions used and the attempts of detoxification, the concentration of toxic by-products of acid hydrolysis was too high for usage of undiluted sugar syrups in C. acetobutylicum fermentations. Detoxification by overliming had an optimal effect, but only up to 7.5 % of the sugars in the molasses/flour starch mixture could routinely be replaced in fermentations without causing problems or diminished yields. But efficient laboratory-scale fermentations with up to 75 % (v/v) pentose hydrolysate have been reported in the Dokshukino files.

Fermentation efficiency

The sugar content of feeding solutions was determined as sugar equivalents, i.e., by determining the reducing sugars after complete (analytical) hydrolysis of all oligo- and polysaccharides. During AB production, about 50 % (w/w) of the fermented sugars were converted to the gases CO2 and H2, and 33–39 % to solvents. At the end of an average fermentation, the broth contained 6.4 g acetone, 10.0 g butanol, and 1.5 g ethanol per liter. The 4.7 % (w/v) of sugar equivalents in the incoming fermentation substrate (a mixture of starch, maltodextrins, sucrose, and pentoses) were reduced to about 0.5 %. At the Dokshukino plant, about 925 t of corncobs (dry weight) were hydrolyzed per month in 1962 from which syrup containing 312 t of pentosan sugar was produced.

The use of 100 % hydrolyzates would theoretically reduce substrate costs for an identical solvent yield by about 45 % compared to the cost of wheat flour by Soviet market prices of 1960 for raw material, sulphuric acid, lime, and energy. In laboratory experiments, the clostridia fermented pentose hydrolyzates containing the relative amount of sugars in 2.7 t starch to 1 t of solvents. The use of the cellulose-derived glucose (Table 2) would further increase the output of solvents from agricultural waste, but the toxic products derived from the high-temperature acid hydrolysis prevented the AB fermentation at the time. A sample calculation using an average yield from the data in Table 1 is shown in Table 4, although the sugar content (in “starch equivalents”) for a mixture of hexoses and pentoses is difficult to quantify. The usable combustion energy from the solvents and the hydrogen actually produced in a Russian AB plant was comparable to the average combustion energy produced in a traditional yeast-based ethanol fermentation plant (assuming 460 kg ethanol per ton of starch) if the possible by-product biogas is not taken into account.
Table 4

End products of the AB process and the combustion energy of the products

Product

Weight (kg)a

Combustion energy (kJ/g)

Net calorific value per ton of substrate (MJ/t)b

Butanol

207±10

33.1

6,852

Acetone

100±8

28.5

2,850

Ethanol

56±15

26.7

1,495

CO2

≈600

H2

≈19

119.8

2,276

Total AB

  

13,273

aKilogram of product per ton of starch equivalent

bKilojoules of combustion energy (net calorific value) in end product per ton of starch equivalent

For 5 t of dry flour necessary for the production of about 1 t of solvents, 58.8 t of liquid substrate had to be sterilized. The Evremovo plant produced about 15,000 t of solvents per year. The Grosnyi plant was about 1.5 times as big. No production numbers of the other plants were available. It is of special interest that the production yield per ton of substrate (starch equivalents) in Table 1 was very consistent in three different plants and no breakdown of fermentation was documented from 1960 to mid 1962. This confirms the reliability of the Dokshukino process which is crucial to operate the AB process economically.

Gas production

Major side products of the acid and solvent formation are large amounts of the gases CO2 and H2, obtained from the fermenters in molar (and volume) ratio of approximately 1.5:1. Per mole of hexose (glucose), about three moles of CO2 and H2 are formed; during the fermentation, 1.7 t of gases are formed per 1 t of solvents (1,649 kg CO2, 97 % w/w, and 51 kg H2, 3 % w/w). Thus, an average Russian AB plant produced per day approximately 85 t of gases per 50 t of solvents, i.e., 3.43 t/h CO2 and 0.106 t/h H2 (Logotkin 1958). This summed up to 8.7 million m3 of H2 and 13.1 million m3 of CO2 per year for the Evremovo plant (about 23,000 t CO2). Gases were separated and CO2 was sold as dry ice and liquid CO2. A 1 m3 of gas contained 0.01 kg of solvents (55 % of acetone, 40 % of butanol, and 5 % of ethanol) which were extracted by condensation. It was planned to convert the gas mixture into methanol and formaldehyde, but it is not clear to date if this part of the biorefinery concept was realized (Fig. 3).

Despite the vigorous gas production, foam was unexpectedly not a problem with the type of AB process developed in Dokshukino. This may be due to the relatively flat fermenter geometry or the steady flow through the vessels resulting in vigorous mixing by the incoming broth. Four fermenter-sized tanks for foam collection turned out to be unnecessary and were later converted to a fourth fermenter battery in Dokshukino (fermenters 23–26 in Fig. 5).

Besides the fermentation gases, 30,000 m3 of biogas were produced per day from the fermentation sludge in a thermophilic methanogenic fermentation. It was used to provide process heat in sterilization and distillation.

Conclusions and perspectives for a modern AB process

Because of the longer time for research and scientifically guided process modification up to the late 1980s, the Russian industrial plants were further developed than Western plants. It was shown in the Russian AB plants that the conversion of agricultural biomass into liquid fuels such as butanol is a feasible technology, which could contribute considerably to make a national economy less dependent on imported fossil oil. The technology of fermenting biomass hydrolyzates to solvents is a promising link in the chain of advances toward the use of lignocellulosic biomass for fuel production. Necessary improvements for a modern process would include breeding suitable plant varieties, growing the energy plants in large scale (with attention to the socioeconomic and environmental effects), pretreatment and hydrolysis, fermentation, downstream processing, and marketing. Although a lot of further research and development, especially in substrate pretreatment, strain selection and genetic modification, as well as in the downstream processing of the solvents have to be made to reach process maturity, the biorefinery concept underlying the Russian process points to the future of a modern solvent production based on renewable lignocellulosic biomass.

The process presented here as the “Dokshukino process” is to our knowledge the first using agricultural waste materials routinely in a full-scale industrial process, with the exception of the ethanol fermentation from woody materials after acid hydrolysis, which is still carried out in Russia. Attempts by US and Canadian companies (BC International, Arcenol, Iogen, and others) to use both, pentoses from hemicelluloses and glucose from cellulose, for ethanol production are not yet mature industrial processes. The production of butanol and acetone is technologically more challenging than that of ethanol but also more advantageous if the substitution of gasoline as well as diesel is the target (Schwarz and Gapes 2006).

The mass flow of the Evremovo plant can be calculated approximately from the numbers partially specified in the above text: about 40.500 tons of starch equivalents (about 90.000 t of flour dry weight, partially substituted by molasses and hydrolyzate) per year were sterilized in 880.000 t of liquid substrate and fermented to 15.000 t of solvents containing 4.140 t acetone, 8.550 t butanol, 2.310 t ethanol, and to 8.7 million m3 hydrogen gas, and 13.1 million m3 carbon dioxide. The fermentation broth was used for the production of 11 million m3 biogas (containing methane gas) and an unknown amount of vitamin B12. In addition, fodder yeast was produced.

The experience gained with the industrial AB fermentation in Russia shows a lot about the perspectives which lie in the fermentation of lignocellulosic biomass. The transfer of research ideas into an industrial-scale solvent production paves the way into a modern praxis of sustainable fuel production from renewable material. Because an advanced process (continual fermentation) for solvent production has been adapted to the usage of alternative, abundant, and renewable substrates, the Russian AB process as developed at the Dokshukino plant offers a number of advantages over the AB fermentation scheme known from the former Western plants:
  • A continuous feeding to a battery of fermentation vessels (continual fermentation)

  • Decreased incidence of bacteriophage infections by strain selection and stringent sterilization of substrate and appliances

  • Replacement of starch and/or molasses by hydrolyzates of agricultural waste material

  • Use of pentose hydrolyzates in addition to hexoses

  • Full integration in a biorefinery concept making as much use as possible of by-products (see Fig. 3).

It should be possible to use this experience and to build up a modern mature biofuel production which is economically viable and at the same time obeys ecological necessities of an after-petrol-oil era.

Acknowledgements

The authors thank D. Antoni, H. Bahl, R. Gapes, R. Igelspacher, D. Jones, E. Kashket, J. Puls, D. Schieder, W. L. Staudenbauer, and S. Yarotsky for numerous comments, corrections, suggestions, and critically reading the manuscript. The support of an INTAS-YSF fellowship to OB is gratefully acknowledged.

Copyright information

© Springer-Verlag 2006