Development of low cost medium for ethanol production from syngas by Clostridium ragsdalei

Bioresource Technology 147 (2013) 508–515

Contents lists available at ScienceDirect

Bioresource Technology journal homepage: www.elsevier.com/locate/biortech

Development of low cost medium for ethanol production from syngas by Clostridium ragsdalei Jie Gao, Hasan K. Atiyeh ⇑, John R. Phillips, Mark R. Wilkins, Raymond L. Huhnke Department of Biosystems and Agricultural Engineering, Oklahoma State University, Stillwater, OK 74078, USA

h i g h l i g h t s  Design of a low cost medium for ethanol production is key for process feasibility.  We formulated and compared ten syngas fermentation media for Clostridium ragsdalei.  Nutrient effects on growth and production during syngas fermentation were examined.  Ethanol yield from CO for defined medium M9 was 36% higher than standard medium M1.  Cost of completely defined medium M9 was 5% of the cost of standard medium M1.

a r t i c l e

i n f o

Article history: Received 21 May 2013 Received in revised form 9 August 2013 Accepted 12 August 2013 Available online 22 August 2013 Keywords: Syngas fermentation Clostridium ragsdalei Ethanol Medium development Defined medium

a b s t r a c t The development of a low cost medium for ethanol production is critical for process feasibility. Ten media were formulated for Clostridium ragsdalei by reduction, elimination and replacement of expensive nutrients. Cost analysis and effects of medium components on growth and product formation were investigated. Fermentations were performed in 250 mL bottles using syngas (20% CO, 15% CO2, 5% H2 and 60% N2). The standard medium M1 cost is $9.83/L, of which 93% is attributed to morpholinoethane sulfonic acid (MES) buffer. Statistical analysis of the results showed that MES removal did not affect cell growth and ethanol production (P > 0.05). Based on cells’ elemental composition, a minimal mineral concentration medium M7 was formulated, which provided 29% higher ethanol yield from CO at 3% of the cost compared to medium M1. Ethanol yield from CO in the completely defined medium M9 was 36% higher than while at 5% the cost of medium M1. Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction Fossil fuels are important in manufacturing of transportation fuels and chemicals. The increase in energy demand worldwide has caused a rise in fossil fuels’ prices. Another concern related to the use of fossil fuels is the large amount of CO2 emissions that have negative effects on the environment (Demirbas, 2005). Therefore, new clean energy sources should be developed to supplement fossil fuels. Ethanol production from lignocellulosic biomass and other feedstocks can offer environmental and economic benefits. Hybrid gasification–syngas fermentation is a promising technology to produce ethanol. The first step in this process is gasification, which involves gasifying biomass to synthesis gas (syngas) mainly consists of CO, CO2, H2 and some contaminants (Spath and Dayton, 2003). The second step is syngas fermentation, in which anaerobic

⇑ Corresponding author. Address: Department of Biosystems and Agricultural Engineering, 214 Ag Hall, Oklahoma State University, Stillwater, OK 74078, USA. Tel.: +1 405 744 8397; fax: +1 405 744 6059. E-mail address: [email protected] (H.K. Atiyeh). 0960-8524/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.biortech.2013.08.075

bacteria consume syngas and produce ethanol and acetate. The overall reactions are shown as follows (Vega et al., 1989): 

6CO þ 3H2 O ! C2 H5 OH þ 4CO2

DG ¼ 217:4 kJ=mol

6H2 þ 2CO2 ! C2 H5 OH þ 3H2 O

DG ¼ 97:0 kJ=mol





4CO þ 2H2 O ! CH3 COOH þ 2CO2

DG ¼ 154:6 kJ=mol

2CO2 þ 4H2 ! CH3 COOH þ 2H2 O

DG ¼ 74:3 kJ=mol



ð1Þ ð2Þ ð3Þ ð4Þ

compared to other conversion technologies, syngas fermentation offers advantages such as organic biomass components including lignin are converted to syngas, mild reaction conditions, specific bacteria can be used to make desired products, and no requirement for a specific ratio of CO to H2 (Phillips et al., 1994; Wilkins and Atiyeh, 2011). Bacteria used in the hybrid conversion process are called acetogens. Examples of acetogens used are Clostridium ljungdahlii, Clostridium carboxidivorans P7, Clostridium ragsdalei P11 and Alkalibaculum bacchi strains CP11T, CP13 and CP15, which convert

J. Gao et al. / Bioresource Technology 147 (2013) 508–515

syngas to alcohols and organic acids through the acetyl-CoA pathway (Phillips et al., 1994; Liou et al., 2005; Ahmed et al., 2006; Maddipati et al., 2011; Liu et al., 2012). To improve syngas fermentation technology, researchers focused on investigating the effects of process parameters such as pH, gas flow rate and composition on bacterial metabolism and physiological response (Phillips et al., 1994; Henstra et al., 2007; Cotter et al., 2009; Liu et al., 2012; Ukpong et al., 2012). Researchers showed that the conversion of acetic acid to ethanol by C. carboxidivorans P7 occurred at a pH range from 4.5 to 4.8 (Ahmed et al., 2006). The pH of the fermentation medium could be controlled in this range using a buffer such as morpholinoethane sulfonic acid (MES). However, the presence of MES buffer increased acetic acid formation and reduced ethanol production (Kundiyana et al., 2010). In addition, the effects of trace metals on key enzymes in the metabolic pathway for ethanol production were also examined (Saxena and Tanner, 2011a). The addition of expensive reducing agents, dithiothreitol and methyl viologen, was found to enhance ethanol production with C. ragsdalei (Panneerselvam et al., 2009; Babu et al., 2010). Other studies focused on using inexpensive nutrients to replace the standard yeast extract (YE) medium for C. ragsdalei to make syngas fermentation effective and competitive on a cost basis (Maddipati et al., 2011; Saxena and Tanner, 2011b). Cotton seed extract (CSE) is one example of an alternative nutrient supplement, having a relatively low cost ($0.91/kg) compared to YE ($9.20/kg) (Kundiyana et al., 2010; Maddipati et al., 2011). A medium containing 0.5 g/L CSE produced fourfold more ethanol (2.7 g/L) with C. ragsdalei in bottle fermentations compared to a medium with 1.0 g/L YE. Corn steep liquor (CSL) is another inexpensive nutrient source ($0.18/kg) that can replace YE in fermentation processes (Maddipati et al., 2011). The use of 20 g/L CSL medium with C. ragsdalei enhanced ethanol production by 32% compared to a medium with 1 g/L YE (Maddipati et al., 2011). Based on the cost analysis of the standard YE medium M1 developed by Saxena and Tanner (2011b) for C. ragsdalei, the current cost is $9.83/L while the MES buffer accounting for 93.2% of the total cost (Table 1). Critical kinetic parameters for process development and reactor design such as specific growth rates, cell mass and product yields and syngas conversion efficiencies were not provided by Saxena and Tanner (2011b). Additionally, YE and mineral solution in medium M1 account for 2.2% and 2.8% of the total medium cost, respectively (Table 1). Thus, the concentrations of these relatively expensive nutrients must be reduced to make syngas fermentation technology economically competitive. In addition, the standard medium for C. ragsdalei is an undefined

Table 1 Components and cost analysis for standard YE medium (M1). Componentsa

a

Standard YE medium (M1)

Stock solutions

mL/L

$/L

Mineral solutionb Trace metal solutionb Vitamin solutionb 0.1% Resazurin 4.0% Cysteine–sulfide 2.0 N KOH solution

25 10 10 1 2.5 10.38

0.271 0.014 0.005 0.011 0.040 0.110

% Of cost

Other nutrients

g/L

$/L

% Of cost

YE MES Total medium costc

1 10

0.216 9.160 9.827

2.20 93.21 100.00

2.76 0.14 0.05 0.11 0.41 1.12

Medium M1 recipe is based on findings by Saxena and Tanner (2011a,b). Compositions of mineral, vitamin and trace metal stock solutions are given in Section 2.2. c Overall medium cost calculated using prices of chemicals from Sigma–Aldrich and Fisher-Scientific websites in May, 2013. b

509

medium, because YE is a complex ingredient that consists of a mixture of many chemical species in unknown proportions. Thus, designing a low cost and defined medium is advantageous in syngas fermentation to ensure fermentation reproducibility and to reduce or eliminate unnecessary components that increase production cost and might interfere with product separation. No comprehensive studies were found in the literature on the development of low cost defined medium or nutrient effects on growth and product yields in syngas fermentation using C. ragsdalei. Therefore, the objective of the present study is to reduce, eliminate, or replace expensive nutrients with inexpensive nutrient supplements from the standard YE medium for C. ragsdalei, thus developing a low cost and completely defined medium for ethanol production through syngas fermentation. 2. Methods 2.1. Microorganism and syngas C. ragsdalei, also called Clostridium strain P11 provided by Dr. Ralph Tanner from the University of Oklahoma was used in the present study. The stock culture was maintained in 250 mL bottles containing 100 mL standard yeast extract (YE) medium M1 at 23 °C not shaken by replacing syngas in the headspace every two weeks to 239 kPa. The composition of syngas used was 20% CO, 15% CO2, 5% H2 and 60% N2 by volume. 2.2. Fermentation media A standard YE medium for C. ragsdalei developed by Saxena and Tanner (2011b), called in the present study medium M1 was used for preparation of the inoculum used in all experiments. The composition and cost analysis of medium M1 is shown in Table 1. The compositions of the mineral, vitamin and trace metal stock solutions used for standard medium M1 are based on findings by Saxena and Tanner (2011a,b). The mineral stock solution contained (per liter) 100 g ammonium chloride, 10 g potassium chloride, 10 g potassium phosphate monobasic, 20 g magnesium sulfate and 4 g calcium chloride. The vitamin stock solution contained (per liter) 10 mg pyridoxine, 5 mg thiamine, 5 mg riboflavin, 5 mg calcium pantothenate (B5), 5 mg thioctic acid, 5 mg p-(4)-Aminobenzoic acid, 5 mg nicotinic acid, 5 mg vitamin B12, 2 mg biotin, 2 mg folic acid and 10 mg 2-mercaptoethanesulfonic acid sodium salt (MESNA). The trace metal stock solution contained (per liter) 2 g nitrilotriacetic acid, 1 g manganese sulfate, 0.8 g ferrous ammonium sulfate, 0.2 g cobalt chloride, 1 g zinc sulfate, 0.2 g nickel chloride, 0.02 g sodium molybdate, 0.1 g sodium selenate and 0.2 g sodium tungstate. Nine other media were formulated for C. ragsdalei by reduction, elimination and replacement of expensive nutrients. A summary of all medium formulations M1 to M10 and cost is shown in Table 2. The cost of each medium was calculated based on laboratory chemical prices from Sigma–Aldrich and Fisher-Scientific in May 2013. The standard YE medium M1 with MES was compared to medium M2 without MES. Between 0.5 and 1 mL NaHCO3 solution (70 g/L) was added to each bottle without MES as needed to maintain the pH above 4.5 during fermentation. The results showed that removal of MES buffer had no negative effect on the cell mass and product concentrations. Therefore, media M2 to M10 did not contain MES buffer. In the second experiment, three media formulations M2, M3 and M4 were prepared. The composition of these media is similar to medium M1 except no MES was added and the concentration of YE was varied (Table 2). Medium M3 with 0.5 g/L YE supported C. ragsdalei growth to a similar level and

510

J. Gao et al. / Bioresource Technology 147 (2013) 508–515

Table 2 Summary of composition and cost of various media used. Mediuma

M1

M2

M3

M4

M5

M6

M7

M8

M9

M10

Yeast extract (g/L) Mineral solution (mL/L) Trace metal solutionb Vitamin solutionb YE replacementf (mL/L) Total medium costg ($/L) Cost relative to medium M1 (%)

1.0 25b 10 10 0 9.83 100

1.0 25b 10 10 0 0.73 7.4

0.5 25b 10 10 0 0.58 5.9

2.0 25b 10 10 0 0.98 10.0

0.5 25c 10 10 0 0.38 3.9

0.5 25d 10 10 0 0.33 3.4

0.5 25e 10 10 0 0.31 3.2

0.5 25e 10 10 29.2 0.70 7.1

0.0 25e 10 10 29.2 0.53 5.4

0.0 25e 10 10 0 0.18 1.8

a M1 contained MES, M2–M10 contained no MES while pH was maintained above 4.5 by using NaHCO3; all media contained 1 mL/L of 0.1% resazurin and 2.5 mL/L of 4.0% cysteine–sulfide solutions. b Compositions of stock solutions are shown in Section 2.2. c Revised mineral solution I. d Revised mineral solution II. e Revised mineral solution III as shown in Table 3. f Composition of nutrients to replace YE is shown in Section 2.2. g Overall medium cost was calculated using prices of chemicals from Sigma–Aldrich and Fisher-Scientific websites in May 2013.

Table 3 Composition of minerals in the standard and three revised solutions I, II and III.

a

Mineral solution components

Standard solution for medium M3 (g/L)

Revised solution I for medium M5 (g/L)

Revised solution II for medium M6 (g/L)

Revised solution III for medium M7 (g/L)

Ammonium chloride Potassium chloride Potassium phosphate monobasic Magnesium sulfate Calcium chloride

100a 10a 10a

33.3 0.0 10.0

16.7 0.0 5.0

1.8 0.0 1.2

20a 4a

2.0 1.3

1.0 0.7

1.0 0.5

Composition is based on findings by Saxena and Tanner (2011a,b).

Table 4 Elemental analysis and predicted maximum OD for media M3, M5, M6 and M7 based on elements added from various mineral solutions and other media components. Element

N P S K Na Ca Mg Cl Fe a

% Dry wta

Predicted maximum OD M3 with standard mineral solution

M5 with revised mineral solution I

M6 with revised mineral solution II

M7 with revised mineral solution III

14.0 3.0 1.0 1.0 1.0 0.5 0.5 0.5 0.2

12.91 6.07 28.10 55.56 198.35 13.67 24.85 928.13 1.43

5.12 6.07 13.47 22.78 198.35 4.58 2.65 300.35 1.43

3.17 3.70 12.65 13.80 198.35 2.32 1.42 158.31 1.43

1.43 1.90 12.65 6.98 198.35 1.74 1.42 34.11 1.43

Data for E. coli as a model microorganism (Bailey and Ollis, 1986).

produced more ethanol compared to media M2 and M4. Thus, 0.5 g/L YE was used in further media formulations. In the third experiment, three media M5, M6 and M7 were formulated to contain similar components of medium M3 except that the content of the mineral solution was varied as shown in Table 3. The design of the revised mineral solutions I, II and III for media M5, M6 and M7, respectively, (Table 3) was based on an elemental composition of Escherichia coli as a model microorganism, which is conveniently available (Bailey and Ollis, 1986; Phillips et al., 1993). The new medium recipe was used to predict potential OD from elemental nutrients in the medium (Table 4). According to Table 4, the predicted maximum OD from elemental nutrients in medium M3 with standard mineral solution revealed that most of the elements were in excess and iron seemed to be the limiting element. Thus, more balanced media were designed by altering the concentrations of

nutrients contained in the mineral stock solution. The concentrations of the elements in the media with revised mineral solutions were selected to support cell growth to an OD of at least 1.4. The control medium used in this experiment was medium M3 that contained 0.5 g/L YE, which had similar concentrations of minerals as in standard YE medium M1. The medium M7 with revised mineral solution III had similar cell growth and product profiles as medium M3 with standard mineral solution. Therefore, revised mineral solution III was used in a subsequent experiment for the development of completely defined medium to replace YE. In the fourth experiment, several nutrients were used to replace the function of YE. Based on the analysis of nutrients contained in yeast extract (Difco 212750, Detroit, MI, USA), 18 amino acids, choline chloride, inositol and thymidine are identified as the nutrients provided by YE for C. ragsdalei in the standard medium M1. These nutrients are expected to completely replace YE in fermentation with C. ragsdalei. Based on the analysis of YE, medium M7 with 0.5 g/L YE contains 329.25 mg/L of the 18 amino acids, 0.15 mg/L choline chloride, 0.7 mg/L inositol and 8.75 lg/L thymidine. A two by two-factorial statistical design with two-levels and twofactors was used to examine the effect of replacing YE with defined nutrients. Media M7 to M10 contained revised mineral solution III, trace metal and vitamin solutions (Table 2). Medium M7 also contained 0.5 g/L YE. Medium M8 contained both 0.5 g/L YE and YE nutrient replacement. Medium M9 contained YE nutrient replacement. Medium M10 had neither YE nor YE nutrient replacement. The YE nutrient replacements added to each bottle containing 100 mL of media M8 and M9 were 2.5 mL amino acids, 0.17 mL choline chloride solution (90 mg/L), 0.17 mL inositol solution (420 mg/L) and 0.08 mL thymidine solution (10.5 mg/L). The amino acid stock solution was added to media M8 and M9 after sterilization. However, choline chloride, inositol and thymidine stock solutions were added before sterilization. The amino acid stock solution contained (per liter) 1072 mg alanine, 604 mg arginine,

J. Gao et al. / Bioresource Technology 147 (2013) 508–515

1338 mg aspartic acid, 148 mg cysteine, 2840 mg glutamic acid, 650 mg glycine, 240 mg histidine, 646 mg isoleucine, 938 mg leucine, 1030 mg lysine, 210 mg methionine, 506 mg phenylalanine, 520 mg proline, 568 mg serine, 590 mg threonine, 272 mg tryptophan, 240 mg tyrosine and 758 mg valine. All medium components were mixed with deionized water in a round bottomed flask. The initial pH of the medium was adjusted to 6.1 using 2 N KOH. Medium was heated to boiling and degassed by bubbling with N2 in the liquid for 20 min to maintain anoxic condition. After the medium was cooled to 20 °C, 100 mL of medium was dispensed into 250 mL serum bottles that were purged with N2. The bottles were sealed with No. 1 butyl rubber stoppers (VWR Scientific, Radnor, PA, USA) and capped with aluminum caps (Wheaton, Millville, NJ, USA). The bottles containing the medium were then autoclaved for 20 min at 121 °C. After the medium in the bottles was cooled to 20 °C, 0.25 mL of 4% cysteine–sulfide solution was added to each bottle. The medium was then inoculated with 10% (v/v) of C. ragsdalei stock culture. The inoculum culture was passaged twice (i.e., transferred to fresh medium when optical density, OD, was above 0.4) to reduce lag phase and ensure cells viability. After the medium was inoculated, the bottles were maintained at 37 °C and shaken upright at 150 rpm using orbital shakers (Innova 2100, New Brunswick Scientific, NJ, USA). Syngas was replaced in the headspace every 24 h to 239 kPa. All bottle fermentations were performed in triplicate for 15 days. Liquid and gas samples were collected every 24 h to measure pH, cell mass and product concentrations and gas utilization.

511

3. Results and discussion 3.1. Effect of MES buffer Similar growth profiles of C. ragsdalei were observed in standard YE medium M1 with MES and medium M2 without MES during 360 h of fermentation. Cell mass in both media increased from 0.02 g/L to about 0.17 g/L in the first 96 h. Cells entered the stationary phase after 96 h. No significant differences were observed in cell mass concentrations in media M1 and M2 after 48 h to 336 h (P > 0.05). The differences between the maximum cell mass concentrations and the specific growth rates of C. ragsdalei in media M1 and M2 were not significant (P > 0.05) as shown in Fig. 1. In medium M1 with MES buffer, the pH decreased from 5.4 to 4.6 in the first 144 h due to production of acetic acid. After which, there was minimal change in pH due to the decrease in acetic acid production. The pH decreased from 5.5 to 4.5 in 48 h in medium M2 due to acetic acid production and lack of the MES buffer. To avoid a pH decrease below 4.5 in medium M2, which can inhibit C. ragsdalei, 1 mL of the 7% NaHCO3 solution was added after 48 h. The pH in media M1 and M2 remained about 4.5 after 168 h of fermentation. Acetic acid production in media M1 and M2 was mainly growth related. More than half of the total acetic acid production occurred during cell growth in both media. In medium M1, there was an increasing trend in acetic acid production until 168 h after which acetic acid was converted to ethanol. The conversion of acetic acid to ethanol was also reported in previous studies for A. bacchi, C.

2.3. Analytical methods 2.3.1. Fermentation products and gas utilization Cell mass concentration was determined by measuring optical density (OD) at 660 nm using a UV–visible spectrophotometer (Cole-Parmer Company, Vernon Hills, IL, USA). The cell mass concentration equals to 0.34  OD (Panneerselvam et al., 2009). Before solvent analysis, liquid samples were centrifuged at 16,000g for 10 min using Accuspin Micro centrifuge (Fisher Scientific, Pittsburgh, PA, USA). The solid-free supernatant was used for GC analysis. Ethanol and acetic acid concentrations were measured using a flame ionization detector (FID) in an Agilent 6890N Gas Chromatography (GC) with a DB–FFAP capillary column (Agilent Technologies, Wilmington, DE, USA) as described previously (Liu et al., 2012). Gas samples were withdrawn from the headspace of each serum bottle using a gas tight syringe (Hamilton Company, Reno, NV, USA). The injection volume of gas sample was 100 lL. Agilent 6890N GC (Agilent Technologies, Wilmington, DE, USA) was used for CO, CO2, H2 and N2 analysis with a thermal conductivity detector (TCD) and Supelco PLOT 1010 column (Supelco, Bellefonte, PA, USA) as described previously (Liu et al., 2012). The compositions of gases in the sample were recorded in mole percentages. 2.3.2. Statistical analysis and calculations Analysis of variance (ANOVA) was determined using the GLM procedure of SAS Release 9.2 (Cary, NC). A Duncan’s multiple range test was used to determine whether there were statistically significant differences in pH, cell mass, ethanol and acetic acid concentrations, CO utilization, H2 utilization and CO2 production between the treatments with the different media. The significance level tested was at a = 0.05. The cell mass and ethanol yields from CO and gas utilization were calculated as described previously (Liu et al., 2012). To show the potential of media M1 to M10 to support production of ethanol with C. ragsdalei on a cell mass basis, ethanol yield was also calculated as follows:

Ethanol yield ðg=gÞ ¼

Maximum EtOH produced : Cell mass

ð5Þ

a

b

Fig. 1. (a) Maximum cell mass, ethanol and acetic acid concentrations and (b) specific growth rate and maximum cell mass yield obtained during syngas fermentation by C. ragsdalei in ten various media with composition in Table 2. Error bars (n = 3 for M1, M4–M6 and M8–M10; n = 6 for M2, M3 and M7) represented ±1 standard deviation.

512

J. Gao et al. / Bioresource Technology 147 (2013) 508–515

carboxidivorans and C. ragsdalei (Hurst and Lewis, 2010; Maddipati et al., 2011; Liu et al., 2012). Large variability in measured acetic acid concentrations was observed after 240 h in medium M1. In medium M2, acetic acid production was observed until 264 h, after which no more acetic acid was formed. Similar concentrations of acetic acid (3.8 g/L) were produced in both media at 360 h. In medium M1, no ethanol production was observed during cell growth. After 168 h when cells entered the stationary phase and pH dropped to about 4.5, ethanol production started. However, in medium M2, ethanol production started after 48 h because the pH quickly decreased to 4.5. Ethanol production by C. ragsdalei started in both media when the pH was around 4.5, which is similar to previous studies (Maddipati et al., 2011). The statistical analysis showed that the effect of MES was not significant on the maximum ethanol and acetic acid concentrations (P > 0.05) as shown in Fig. 1a. No significant difference was observed in ethanol production after 288 h (P > 0.05). In addition, the differences in acetic acid concentrations in these media after 216 h were not significant (P > 0.05). In both media M1 and M2, the consumption of CO and H2 by C. ragsdalei was observed from the beginning of the fermentation. The difference in cumulative CO and H2 utilized in both media was not significant (P > 0.05). The total percentages of CO utilized during 360 h in media M1 and M2 were 58%, and 54%, respectively (Fig. 2a). However, the total H2 utilized in media M1 and M2 were 54% and 63%, respectively. Based on these results, it can be seen that removing the MES buffer from the standard YE medium M1 and maintaining the pH above 4.5 with NaHCO3 had no negative effects on C. ragsdalei growth or product formation. The differences in ethanol yield from CO or ethanol yield (g ethanol/g cells) in media M1 and M2 were

a

not significant (P > 0.05) as shown in Fig. 2. Thus, MES buffer could be eliminated from the standard medium. The cost of medium M2 was 93% lower than medium M1 (Table 2). No MES was used and the pH was maintained above 4.5 with NaHCO3 in further experiments. 3.2. Effect of YE Three media formulations M3, M2 and M4 that contained 0.5, 1.0 and 2.0 g/L YE, respectively, and no MES buffer were compared. The composition of each medium is provided in Table 2. The pH profiles during fermentation in the three media were similar. A fast decrease in pH from 5.5 to 4.6 was measured during the first 48 h because of absence of MES buffer. After 48 h, up to 1.0 mL of NaHCO3 solution was added to each medium to maintain the pH above 4.5. No lag phase was observed during growth in all three media. The specific growth rates and cell mass concentrations increased as the YE concentration in the medium increased. Medium M4 with 2.0 g/L YE provided the highest growth rate (0.042 h1), maximum cell mass concentration (0.28 g/L), maximum acetic acid concentration (4.9 g/L) and maximum cell mass yield (1.24 g/mol CO), which were significantly higher (P < 0.05) than in media M2 and M3 as shown in Fig. 1. However, medium M3 with 0.5 g/L YE produced over twofold more ethanol than in medium M4 with 2.0 g/L YE. Only 16% more ethanol was produced by C. ragsdalei in medium M3 with 0.5 g/L YE compared to medium M2 with 1.0 g/L YE. Although slightly higher CO and H2 conversion efficiencies were obtained in media M2 (1 g/L YE) and M4 (2.0 g/L YE), ethanol yield from CO in medium M3 with 0.5 g/L YE was over 20% and twofold higher than in media M2 and M4, respectively (Fig. 2a). In addition, ethanol yield (g ethanol/g cells) was the lowest for medium M4 with 2.0 g/L YE (Fig. 2b) due to production of more cells and less ethanol than media M2 and M3 (Fig. 1a). The difference in the cumulative amounts of CO2 produced in the three media was not significant (P > 0.05). Medium M3 with 0.5 g/L YE sufficiently supported C. ragsdalei growth and produced more ethanol than media M2 and M4. Thus, medium M3 with 0.5 g/L YE was used in further experiments as the control medium providing 94% and 21% lower cost compared to standard medium M1 and medium M2, respectively (Table 2). 3.3. Effect of minerals

b

Fig. 2. (a) Ethanol yield from CO and conversion efficiencies of CO and H2, and (b) ethanol yield (g ethanol/g cell) after 360 h of syngas fermentation by C. ragsdalei in ten various media with composition in Table 2. Error bars (n = 3 for M1, M4–M6 and M8–M10; n = 6 for M2, M3 and M7) represented ±1 standard deviation.

Three revised mineral solutions I, II and III (Table 3) were used in the preparation of media M5, M6 and M7, respectively, that contained 0.5 g/L YE and other components (Table 2). Medium M3 with 0.5 g/L YE and standard mineral solution was the control medium. The medium recipe was formulated based on elemental composition of E. coli (Bailey and Ollis, 1986; Phillips et al., 1993) that allowed the prediction of potential cell growth from elemental nutrients in the formulated media (Table 4). Similar pH and growth profiles were observed in all media. The pH decreased from 5.6 to 4.6 in 72 h due to production of acetic acid, then the pH changes were very small until the end of fermentation. During the first 72 h, the cells were in the growth phase. Then, cell growth rate decreased and cells entered the stationary phase. Cell mass concentration decreased in all media after 240 h. Although medium M7 contained the least concentrations of minerals, a maximum cell mass concentration of 0.22 g/L was obtained in medium M7 (Fig. 1a), which was not significantly different from media M3, M5 and M6 (P > 0.05). Additionally, the biomass yield was the highest in medium M7 at 1.14 g/mol CO (Fig. 1b). Acetic acid was produced in media M3, M5, M6 and M7 until about 216 h after which time the acetic acid concentration decreased due to its conversion to ethanol. The maximum acetic acid production was obtained in medium M7 with revised mineral

J. Gao et al. / Bioresource Technology 147 (2013) 508–515

solution III, which was not significantly different than in media M3, M5 and M6 (P > 0.05) (Fig. 1a). Ethanol formation was observed after 192 h until the end of the fermentation. The maximum ethanol concentration was measured in medium M5 (1.36 g/L), which was not significantly different from values observed in media M3, M6 and M7 (P > 0.05). The reduction in the concentrations of minerals in the various media did not affect the final concentrations of acetic acid or ethanol after 360 h (P > 0.05). The reduction 3 2+ of the concentrations of NHþ by 95% and 4 by 98%, PO4 by 88%, Mg Ca2+ by 87.5% in medium M7 from concentrations suggested by (Saxena and Tanner, 2011b) in medium M3 had no negative effects on C. ragsdalei growth, product formation and yields (Figs. 1 and 2). This suggests that the concentrations of these components in medium M7 were adequate to sustain C. ragsdalei growth and keep the enzymatic pathway of ethanol synthesis active. For example, a complete removal of PO3 4 from C. ragsdalei was reported to reduce growth and ethanol production substantially (Saxena and Tanner, 2011b) because phosphate is a major source of phosphorus required to build nucleic acids, phospholipids and nucleotides for cells (Gottschalk, 1986). However, Saxena and Tanner (2011b) found that reduction of NHþ 4 concentration by 50% from the standard medium did not affect growth or ethanol production by C. ragsdalei; however, complete removal of NHþ 4 reduced C. ragsdalei growth by 33% and ethanol production by 41%. The profiles of cumulative CO and H2 utilized and CO2 produced were similar in all media. Over 85% of the utilized CO and H2 by C. ragsdalei occurred during the first 240 h of fermentation. Although the CO and H2 conversion efficiencies in medium M3 were slightly higher than in media M5, M6 and M7 (Fig. 2a), the differences in the CO and H2 conversion efficiencies between these media were not significant (P > 0.05). The highest ethanol yield from CO was 65% in medium M5, which was not significantly different from media M3, M6 and M7 (P > 0.05). The results showed that any of the developed media M3, M5, M6 or M7 with any of the revised mineral solutions can support C. ragsdalei growth and ethanol production. The statistical analysis showed that various concentrations of minerals in the media used did not have a significant effect on the maximum cell mass and ethanol concentrations, ethanol yield or gas conversion efficiencies (P > 0.05). Medium M7 should be selected because it is the least expensive compared to media M3, M5 and M6 (Table 2). Medium M7 with revised mineral solution III (Table 3), in which potassium chloride was completely eliminated and concentrations of other minerals were reduced by over 87%, provided C. ragsdalei with enough nutrients that resulted in comparable cell growth and ethanol production as in the standard medium M1. Medium M7 was also 97% less expensive and provided 29% higher ethanol yield from CO compared to standard medium M1. However, medium M7 is still undefined medium due to presence of YE that consists of a mixture of many chemical species in unknown proportions. Low cost and completely defined media are investigated in the next section with medium M7 as the control medium.

513

at the beginning of the fermentation. This resulted in an increase in the pH between 5.8 and 6.3 in the first 24 h. Then, the pH decreased after 24 h in media M7, M8 and M9. However, the pH decreased after 48 h in medium M10 with neither YE nor YE replacement due to a 48 h lag phase (Fig. 3a). After 96 h, the pH profiles were similar in all media. C. ragsdalei grew in media M7, M8 and M9 immediately after inoculation (Fig. 3a). However, there was a 48 h lag phase in medium M10, after which cells grew and entered the stationary phase after 144 h. Medium M10 resulted in significantly lower (P < 0.05) cell mass concentrations compared to other media. C. ragsdalei grew on syngas in YE free medium M10 without YE replacement, which is unlike previously reported that C. ragsdalei could not grow in YE free medium (Saxena and Tanner, 2011b). Medium M9 that contained YE replacement supported C. ragsdalei growth to a maximum of 0.15 g/L. The statistical analysis showed that cell mass produced in medium M9 was significantly higher (P < 0.05) than that in medium M10. Thus, the YE replacement added to M9 supported cell growth. Cell growth profiles were similar in media M7 and M8 until 168 h (Fig. 3a). Then, cells kept growing in medium M8 until 240 h. However, cells concentration decreased in medium M7 after 168 h. There were significant differences (P < 0.05) in cell mass concentrations between media M7 and M8 after 168 h. The addition of both YE and YE replacement into medium M8 had a ‘‘synergistic’’ effect on cell growth due to presence of more nutrients than in medium M7. In addition, cell growth in medium M8 was longer (0–240 h) compared to medium M7 (0–168 h), due to the additional nutrients in medium M8. The addition of YE replacement had a significant effect (P < 0.05) on the maximum cell mass

a

b

3.4. Effect of YE replacement Four media formulations M7, M8, M9 and M10 were used. The composition of each medium is given in Table 2. Media M7 and M8 contained 0.5 g/L YE. However, medium M8 also contained other nutrients as YE replacement. Media M9 and M10 did not contain YE. M9 contained YE replacement. The composition of YE replacement is given in Section 2.2. All media contained revised mineral solution III (Table 3). The pH and cell growth profiles in media M7, M8, M9 and M10 are shown in Fig. 3a. The initial pH in all media after inoculation was about 5.3 (Fig. 3a). One milliliter of 7% NaHCO3 solution was added to each bottle in all media to avoid a rapid decrease in pH

Fig. 3. (a) pH (open symbol and dash line) and cell mass (solid symbol and solid line), and (b) acetic acid (open symbol and dash line) and ethanol (solid symbol and solid line) profiles for C. ragsdalei during fermentation in media (s) M7 with 0.5 g/L YE, (h) M8 with 0.5 g/L YE and YE replacement, (D) M9 with only YE replacement and, (}) M10 without YE or YE replacement. Detailed media composition is in Table 2. Error bars (n = 3) represented ±1 standard deviation.

514

J. Gao et al. / Bioresource Technology 147 (2013) 508–515

concentration in medium M8 compared to medium M7 as seen in Fig. 1a. On the other hand, the addition of YE replacement in medium M9 was not sufficient to provide similar growth potential as in medium M7 with 0.5 g/L YE (Fig. 3a). The highest cell mass concentration measured in medium M9 was 32% lower than that in medium M7. This indicates that other components in YE, such as carbohydrates and vitamins, not added in the YE replacement, could have contributed to better growth of C. ragsdalei. Although medium M10 did not contain YE or YE replacement, it supported cell growth to 0.1 g/L, which indicates that C. ragsdalei can construct components needed for growth directly from syngas, but at a slower rate than if provided from external sources such as YE. Previous studies reported that corn steep liquor (CSL) and cotton seed extract (CSE) can be used as nutrient replacement for YE to support cell growth and ethanol production (Kundiyana et al., 2010; Maddipati et al., 2011). Therefore, a comparison of the compositions of the nutrients in YE, CSL and CSE could help in designing a fully defined medium with nutrient replacement that can support higher cell mass concentration than the YE replacement modeled based on YE. Acetic acid and ethanol profiles in media M7, M8, M9 and M10 are shown in Fig. 3b. Acetic acid was produced during the growth and early stationary phase. The maximum acetic acid concentration (4.68 g/L) was obtained in medium M8 that contains both YE and YE supplement (Fig. 1a). The maximum acetic acid concentrations measured in the fully defined media M9 and M10 were significantly lower than in medium M7 with 0.5 g/L YE (P < 0.05). Ethanol production in all media occurred when cells were in stationary phase (Fig. 3b). In medium M7, ethanol was produced after 168 h, followed by an increase in ethanol production rate after 264 h due to acetic acid conversion to ethanol. The maximum ethanol concentration measured in medium M7 was 28% and 53% higher (P > 0.05) than in the defined media M9 and M10, respectively, as seen in Fig. 1a. Ethanol production in the fully defined medium M10 with the least nutrients started earlier than in other media due to conversion of syngas to ethanol rather than to support growth and acetic acid formation (Fig. 3b). Ethanol yield from CO in medium M7 with 0.5 g/L YE was about 60%, which was similar to the defined media M9 and M10 and double the yield measured in medium M8 (Fig. 2a). Ethanol yield (g ethanol/g cell) in medium M10 was 45%, 195% and 20% higher than in media M7, M8 and M9, respectively (Fig. 2b). This shows that C. ragsdalei produces more ethanol per gram cell mass in media with less nutrients, which could indicate that limitation of nutrients enhances ethanol formation. This is similar to a study reported by Phillips et al. (1993), in which the removal of YE increased ethanol production by C. ljungdahlii. In addition, limitations of phosphates and sulfate in acetone–butanol–ethanol (ABE) fermentation resulted in production of more butanol than the other products (Bahl et al., 1986; Jones and Woods, 1986). The cumulative CO and H2 utilized and CO2 produced by C. ragsdalei are shown in Fig. 4. CO and H2 were consumed during growth and early stationary phases. The CO consumption profiles in media M7, M8 and M9 were similar during the first 192 h. The maximum CO and H2 conversion efficiencies of C. ragsdalei were 58% and 68%, respectively, measured in medium M8 and were significantly different (P < 0.05) than in media M7, M9 and M10 as seen in Fig. 2a. This was primarily due to the additional nutrients in medium M8. Although, C. ragsdalei utilized the largest amount of CO in medium M8, ethanol yield from CO (30%) was the lowest, as most of CO was used for acetic acid production. The cumulative CO and H2 utilized after 360 h were significantly different (P < 0.05) in the four media. The lowest percentages of CO and H2 utilization by C. ragsdalei were measured in medium M10. In addition, C. ragsdalei produced the most amount of CO2 in medium M8 (Fig. 4b) during growth and production of acetic acid and ethanol (Fig. 3). CO2 pro-

a

b

Fig. 4. Cumulative (a) CO (open symbol and dash line) and H2 (solid symbol and solid line) utilized, and (b) CO2 produced during fermentation by C. ragsdalei in media (s) M7 with 0.5 g/L YE, (h) M8 with 0.5 g/L YE and YE replacement, (D) M9 with only YE replacement and, (}) M10 without YE or YE replacement. Detailed media composition is in Table 2. Error bars (n = 3) represented ±1 standard deviation.

duction by C. ragsdalei in media M7 and M10 were similar, which were higher than in medium M9 after 216 h. Based on the cost analysis, the completely defined M9 with only YE replacement costs 71% more than medium M7 with 0.5 g/L YE due to the high cost of amino acids contained in the YE replacement. The completely defined medium M10 with neither YE nor YE replacement performed similarly to medium M9 and was 42% lower in cost than medium M7 with 0.5 g/L YE. When cell growth, cell mass, and product concentrations and yields by C. ragsdalei were compared in all examined media, the best performing medium was M7 with 0.5 g/L YE and revised mineral solution III (Figs. 1 and 2). Although a fully defined medium M9 can be developed for C. ragsdalei, the cost of that medium should be much lower than medium M7 to justify its use. Medium M10 was the least expensive medium ($0.18/L, Table 2) and provided the highest ethanol yield (11.3 g ethanol/g cell) among all media investigated (M1–M10) as seen in Fig. 2b. However, C. ragsdalei growth was slow and 35% lower ethanol production was observed in medium M10 compared to medium M7. Thus, the recipes for media M9 and M10 should be further modified to reduce their costs and improve their performance for syngas fermentation by C. ragsdalei. Elimination and/or reduction in the concentration of amino acid and cysteine–sulfide below levels used in the completely defined media M9 and M10 warrant further investigation. The costs of all media were based on laboratory chemical prices from Sigma–Aldrich and Fisher-Scientific in May 2013 (Table 2). Media costs are expected to be much lower when chemicals are purchased in bulk for industrial operation. Bulk chemical prices depend on grade and quantity of the chemical delivered, supply

J. Gao et al. / Bioresource Technology 147 (2013) 508–515

and demand, contracts and cost of raw material and energy (Hodges, 2009; Braun and Moxon, 2010). Medium cost estimation based on bulk chemical prices was beyond the scope of this study. However, it is important to note the relative costs among the media for comparison purposes. As such, the cost of the M7, M9 and M10 media were 3%, 5% and 2%, respectively, of the M1 medium’s cost. The use of commercial components at bulk commodity prices is expected to further reduce the costs of the developed media M7, M9 and M10 for potential use in large-scale ethanol production from syngas.

4. Conclusions Elimination of morpholinoethane sulfonic acid (MES) buffer and KCl, and reduction of yeast extract (YE) by 50%, NH4Cl by 98%, KH2PO4 by 88%, MgSO4 by 95% and CaCl2 by 87.5% in medium M7 for C. ragsdalei resulted in similar growth and cell mass yield as in standard YE medium M1. Ethanol yield from CO was 29% higher at 3% of the cost of medium M1. The completely defined medium M9 provided 36% higher ethanol yield from CO at 5% of the cost of medium M1, which shows its potential for use in large-scale syngas fermentation. Acknowledgements This research was supported by USDA-NIFA 2010-34447-20772, the Oklahoma Bioenergy Center and the Oklahoma Agricultural Experiment Station. References Ahmed, A., Cateni, B.G., Huhnke, R.L., Lewis, R.S., 2006. Effects of biomass-generated producer gas constituents on cell growth, product distribution and hydrogenase activity of Clostridium carboxidivorans P7T. Biomass Bioenergy 30, 665–672. Babu, B.K., Atiyeh, H., Wilkins, M., Huhnke, R., 2010. Effect of the reducing agent dithiothreitol on ethanol and acetic acid production by Clostridium strain P11 using simulated biomass-based syngas. Biol. Eng. 3, 19–35. Bahl, H., Gottwald, M., Kuhn, A., Rale, V., Andersch, W., Gottschalk, G., 1986. Nutritional factors affecting the ratio of solvents produced by Clostridium acetobutylicum. Appl. Environ. Microbiol. 52, 169–172. Bailey, J., Ollis, D., 1986. Biochemical Engineering Fundamentals. McGraw-Hill Book Co., New York, pp. 984. Braun, F., Moxon, C., 2010. Dealing with price volatility. Chem. Ind., 22–24.

515

Cotter, J.L., Chinn, M.S., Grunden, A.M., 2009. Influence of process parameters on growth of Clostridium ljungdahlii and Clostridium autoethanogenum on synthesis gas. Enzyme Microbial. Technol. 44, 281–288. Demirbas, A., 2005. Potential applications of renewable energy sources, biomass combustion problems in boiler power systems and combustion related environmental issues. Prog. Energy Combust. Sci. 31, 171–192. Gottschalk, G., 1986. Bact. Metab., second ed. Springer Verlag, New York, pp. 359. Henstra, A.M., Sipma, J., Rinzema, A., Stams, A.J.M., 2007. Microbiology of synthesis gas fermentation for biofuel production. Curr. Opin. Biotechnol. 18, 200–206. Hodges, P., 2009. Coping with oil price volatility. Chem. Ind., 17–19. Hurst, K.M., Lewis, R.S., 2010. Carbon monoxide partial pressure effects on the metabolic process of syngas fermentation. Biochem. Eng. J. 48, 159–165. Jones, D.T., Woods, D.R., 1986. Acetone–butanol fermentation revisited. Microbiol. Rev. 50, 484–524. Kundiyana, D.K., Huhnke, R.L., Maddipati, P., Atiyeh, H.K., Wilkins, M.R., 2010. Feasibility of incorporating cotton seed extract in Clostridium strain P11 fermentation medium during synthesis gas fermentation. Bioresour. Technol. 101, 9673–9680. Liou, J.S.C., Balkwill, D.L., Drake, G.R., Tanner, R.S., 2005. Clostridium carboxidivorans sp. nov., a solvent-producing clostridium isolated from an agricultural settling lagoon, and reclassification of the acetogen Clostridium scatologenes strain SL1 as Clostridium drakei sp. nov. Int. J. Syst. Evol. Microbiol. 55, 2085–2091. Liu, K., Atiyeh, H.K., Tanner, R.S., Wilkins, M.R., Huhnke, R.L., 2012. Fermentative production of ethanol from syngas using novel moderately alkaliphilic strains of Alkalibaculum bacchi. Bioresour. Technol. 104, 336–341. Maddipati, P., Atiyeh, H.K., Bellmer, D.D., Huhnke, R.L., 2011. Ethanol production from syngas by Clostridium strain P11 using corn steep liquor as a nutrient replacement to yeast extract. Bioresour. Technol. 102, 6494–6501. Panneerselvam, A., Wilkins, M., DeLorme, M., Atiyeh, H., Huhnke, R., 2009. Effects of various reducing agents on syngas fermentation by ‘‘Clostridium ragsdalei’’. Biol. Eng. 2, 135–144. Phillips, J., Klasson, K., Clausen, E., Gaddy, J., 1993. Biological production of ethanol from coal synthesis gas. Appl. Biochem. Biotechnol. 39, 559–571. Phillips, J.R., Clausen, E.C., Gaddy, J.L., 1994. Synthesis gas as substrate for the biological production of fuels and chemicals. Appl. Biochem. Biotechnol. 45, 145–157. Saxena, J., Tanner, R.S., 2011a. Effect of trace metals on ethanol production from synthesis gas by the ethanologenic acetogen, Clostridium ragsdalei. J. Ind. Microbiol. Biotechnol. 38, 513–521. Saxena, J., Tanner, R.S., 2011b. Optimization of a corn steep medium for production of ethanol from synthesis gas fermentation by Clostridium ragsdalei. World J. Microbiol. Biotechnol., 1–9. Spath, P., Dayton, D., 2003. Preliminary Screening – Technical and Economic Assessment of Synthesis Gas to Fuels and Chemicals with Emphasis on the Potential for Biomass-Derived Syngas. National Renewable Energy Lab, Golden, CO. (US), NREL/TP-510-34929. Ukpong, M.N., Atiyeh, H.K., De Lorme, M.J.M., Liu, K., Zhu, X.G., Tanner, R.S., Wilkins, M.R., Stevenson, B.S., 2012. Physiological response of Clostridium carboxidivorans during conversion of synthesis gas to solvents in a gas-fed bioreactor. Biotechnol. Bioeng. 109, 2720–2728. Vega, J., Prieto, S., Elmore, B., Clausen, E., Gaddy, J., 1989. The biological production of ethanol from synthesis gas. Appl. Biochem. Biotechnol. 20, 781–797. Wilkins, M.R., Atiyeh, H.K., 2011. Microbial production of ethanol from carbon monoxide. Curr. Opin. Biotechnol. 22, 326–330.