Effects of ammonia nitrogen on H2 and CH4 production during anaerobic digestion of dairy cattle manure

Effects of ammonia nitrogen on H2 and CH4 production during anaerobic digestion of dairy cattle manure

Bioresource Technology 77 (2001) 9±18 E€ects of ammonia nitrogen on H2 and CH4 production during anaerobic digestion of dairy cattle manure M.C. Ster...

336KB Sizes 2 Downloads 80 Views

Bioresource Technology 77 (2001) 9±18

E€ects of ammonia nitrogen on H2 and CH4 production during anaerobic digestion of dairy cattle manure M.C. Sterling Jr., R.E. Lacey a,*, C.R. Engler a, S.C. Ricke b b

a Department of Agricultural Engineering, Texas A&M University, College Station, TX 77843, USA Department of Poultry Science, Texas A&M University, Kleberg Center, Room 101, College Station, TX 77843-2472, USA

Received 28 June 1999; received in revised form 24 August 2000; accepted 8 September 2000

Abstract A number of researchers have veri®ed the inhibitory e€ects of elevated H2 concentrations on various anaerobic fermentation processes. The objective of this work was to investigate the potential for using hydrogen gas production to predict upsets in anaerobic digesters operating on dairy cattle manure. In an ammonia nitrogen overload experiment, urea was added to the experimental digesters to obtain increased ammonia concentrations (600, 1500, or 3000 mg N/l). An increase in urea concentration resulted in an initial cessation of H2 production followed by an increase in H2 formation. Additions of 600, 1500, or 3000 mg N/l initially resulted in the reduction of biogas H2 concentrations. After 24 h, the H2 concentration increased in the 600 and 1500 mg N/l digesters, but production remained inhibited in the 3000 mg N/l digesters. Both methane and total biogas production decreased following urea addition. Volatile solids reduction also decreased during these periods. The digester e‚uent pH and alkalinity increased due to the increased NH‡ 4 formed with added urea. Based on these results, changes in H2 concentration could be a useful parameter for monitoring changes due to increased NH3 in dairy cattle manure anaerobic digesters. Ó 2001 Elsevier Science Ltd. All rights reserved. Keywords: Anaerobic digestion; Hydrogen; Nitrogen; Mesophilic; Dairy manure

1. Introduction The advantages of anaerobic digestion technology for the treatment of organic residues have been well documented. Signi®cant advances in reactor design and operation have been achieved in recent years. Nevertheless, process-control strategies currently available are those that have long been used for anaerobic digesters. Because of the complexity of microbial interactions involved, the process can be dicult to control (Boekhurst et al., 1981). Reasons for digester imbalance include excessive change in temperature, a sudden increase in organic loading, the presence of a toxic material, or a change in feed characteristics (Jeris and Kugelman, 1985). Upsets can lead to digester failure causing loss of production for extended periods of time. One means to monitor the operational health of a digester is to focus on its biochemical state.


Corresponding author. Tel.: +1-979-845-3961; fax: +1-979-8453932. E-mail address: [email protected] (R.E. Lacey).

The anaerobic digestion process is a natural biological process in which a community of bacteria cooperate to form a stable, self-regulating fermentation that converts waste organic matter into a mixture of carbon dioxide and methane gases. A manure digester community generally operates as three interdependent groups: hydrolytic bacteria, acid-forming bacteria, and methanogenic bacteria. Hydrolytic bacteria cleave polymeric carbohydrates and proteins into simple monomeric sugars and amino acids. Acid forming bacteria are composed of acetogenic bacteria, which form volatile fatty acids (VFAs) directly; homoacetogenic bacteria, which form acetate from CO2 and H2 ; and hydrogenogenic bacteria, which convert larger volatile fatty acids into acetate and H2 . Methanogenic bacteria are composed of acetoclastic methanogens, which convert acetate to methane and CO2 and hydrogen utilizing methanogens, which convert CO2 and H2 to methane. Because hydrogen is an intermediate for methane production, monitoring it should provide information on the state of a digester. In a review, Archer and Kirsop (1991) outlined the evolution of hydrogen

0960-8524/01/$ - see front matter Ó 2001 Elsevier Science Ltd. All rights reserved. PII: S 0 9 6 0 - 8 5 2 4 ( 0 0 ) 0 0 1 3 8 - 3


M.C. Sterling Jr. et al. / Bioresource Technology 77 (2001) 9±18

monitoring as a potential control parameter for anaerobic digestion. Mosey (1982) built upon anaerobic microbiological studies done by Hungate (1975) and Wolin (1976, 1982) to create kinetic models for anaerobic digesters using hydrogen as a control parameter. Other authors have discussed potential advantages of monitoring hydrogen gas (Robinson and Tiedje, 1982, 1984; Bartlett et al., 1980) as well as successful trials in laboratory and industrial-scale digesters (Mosey and Fernandes, 1984; Archer et al., 1986). Although the potential for using hydrogen gas as an indicator of digester viability has been demonstrated, the feasibility of using hydrogen gas as a control parameter for digesters fed complex substrates is less clear. Hickey et al. (1987, 1989) found that monitoring hydrogen in anaerobic sludge digesters could provide rapid indication of process upsets; conversely, Kidby and Nedwell (1991) concluded that hydrogen concentration in sludge digester biogas could not be used as an indicator of incipient failure resulting from volumetric overload. A number of studies have cited the inhibitory e€ects of free ammonia (NH3 ) on the metabolism of methanogens (Angelidaki and Ahring, 1993; Braun et al., 1981; El-Hag et al., 1982; De Baere et al., 1984; Sprott et al., 1984). As ammonia is added to a digester, the pH increases until a chemical equilibrium is reached (Georgacakis et al., 1982). However, as ammonia inhibits methanogen metabolism, VFAs accumulate, resulting in a lower pH and a lower concentration of free ammonia. Assuming an adequate digester alkalinity, this mechanism tends to stabilize the digestion process at a certain VFA concentration and pH level (Georgacakis et al., 1982). With an insucient alkalinity, the digester undergoes acidosis, resulting in the cessation of methane production. The purpose of this study was to determine the feasibility of monitoring hydrogen gas as an indicator of digester upset resulting from ammonia overloading. Speci®c objectives included determining the e€ects of added ammonia nitrogen on digester H2 production, methane production, and volatile solids (VS) removal. 2. Methods 2.1. Experimental apparatus The reactors used were similar to those used by Vartak et al. (1997). The reactors were constructed from acrylic tubing, with each reactor having an internal diameter of 15.2 cm, a height of 30.5 cm, and a nominal working volume of 5 l (Fig. 1). Separate feed and ef¯uent ports made of 12.7 and 6.4 mm ID plastic tubing extended into the digester liquor. Separate ports were used for biogas sampling and collection.

Gas produced from each reactor passed through a hydrogen sensor and into a 4.5-l glass graduated gas collector (Fig. 1). The gas collector was ®lled with water and used to measure gas production by volume displacement. Displaced water from the collector was passed through an anti-siphon tubing arrangement and collected in a carboy. 2.2. Hydrogen sensing and recording system Hydrogen content of the biogas was monitored by stannic oxide sensors (TGS 821, Figaro Engineering, Wilmette, IL) housed in rectangular (3:7  7:3  10:0 cm3 ) plastic instrument boxes. Silicone sealant was used to ®ll any openings in the sensor boxes. The manufacturer warns that stannic oxide sensors may be sensitive to silicone; however, the sensors were calibrated with gas standards after construction and all 10 sensors had virtually the same calibration curve indicating that any acute sensitivity either was uniform across all sensors or was adjusted by the calibration. Additionally, the baseline resistance remained constant over the time of the experiments indicating that gradual sensitivity loss did not occur. The sensors were heated using a constant voltage (5 V DC) supplied by a power transformer (PL20-10, Microtran, Valley Stream, NY). Current from the secondary side of the transformer was distributed to the sensor heaters through parallel circuits. Sensor resistance dropped as it was exposed to hydrogen gas. Sensor resistance was measured by a data acquisition/switch unit (34970A, Hewlett-Packard, Santa Clara, CA) and downloaded to a personal computer (Packard Bell, Sacramento, CA). Because gas ¯ow rates were extremely low, convective cooling was considered negligible so that no temperature-compensation circuitry was required. 2.3. Digester operation The experiments were conducted in a walk-in environmental chamber. A set of nine digesters was used, with the temperature maintained at 35  1°C. The digesters initially were inoculated with rumen ¯uid obtained from the Dairy Cattle Nutrition Center, Texas A&M University, and had operated continuously on dairy cattle manure for over three years prior to the start of these experiments. Relatively fresh undiluted dairy cow manure was collected from the concrete surface of the holding pens at the Texas A&M Dairy Cattle Nutrition Center. Approximately 25 kg (wet basis) of manure were gathered at one time, thoroughly mixed with tap water to form a 5 g VS/l slurry, then placed in a freezer and kept frozen until immediately before use.

M.C. Sterling Jr. et al. / Bioresource Technology 77 (2001) 9±18


Fig. 1. Experimental apparatus for biogas generation, collection and monitoring. (DAQ, data acquisition unit).

To maintain a 10-day hydraulic retention time (HRT), 500 ml/day of e‚uent was removed from each digester and replaced with the same volume of feed slurry. This gave an organic loading rate of 0.5 kg VS/ m3 d. Before sampling, the digester was isolated from the gas collection system and the contents were mixed by repeated inversion of the reactor. Each digester was inverted 10 times over a period of about 15 s to obtain complete mixing. After opening both the gas outlet line and liquid ef¯uent port, 500 ml of digester e‚uent was collected in a 600-ml beaker. The liquid e‚uent port was then closed, the feed inlet port was opened and feed slurry was added to the digester. After the addition of feed, the feed inlet and gas outlet ports were closed and the reactor was mixed as before sampling. After mixing, the digester was reconnected to the gas collection system. Stable operation was achieved before these experiments were begun. A 10-day period just prior to collecting data was considered sucient for de®ning stable

operation. The system was considered to be operating stably when the coecient of variation for daily gas production was less than 10% (Vartak et al., 1997). 2.4. Nitrogen addition After the digesters reached stable operation, the digesters were switched to batch mode for the ammonia overloading experiment. For the experimental digesters, urea was added as a solid to the manure slurry to provide di€erent ammonia nitrogen concentrations as shown in Table 1. Urea was selected as the nitrogen source due to its ease of degradation to ammonia by a variety of microorganisms. After an initial (t ˆ 0 h) removal of e‚uent and addition of feed slurry (500 ml each), e‚uent samples of 125 ml were taken from each digester at 3, 6, 9, 12, 24, 48, 72, and 96 h. To maintain constant volume, 125 ml of tap water was added to each digester after each sampling period and concentrations were adjusted for dilution e€ects. As during continuous


M.C. Sterling Jr. et al. / Bioresource Technology 77 (2001) 9±18

Table 1 Experimental design for urea addition experiment Reactor group

Digesters in test group

Urea concentration in digester (test variable)

Control Group A Group B Group C

Reactors Reactors Reactors Reactors

Manure Manure Manure Manure

1, 2, 5, 3,

4, 9 6 7 8

operation, digester contents were mixed before sampling and after additions. 2.5. Chemical analysis Gas volumes were recorded daily. All biogas and methane production volumes were converted to STP (0°C, 1 atm) conditions. Gas samples were collected in gas-tight syringes and analyzed by gas chromatograph for methane, carbon dioxide and nitrogen composition. Gases were separated using a 2.4 m by 3 mm stainless steel column packed with 100/120 mesh HayeSepâ D (Hayes Separations, Bandera, TX). The column was operated at 35°C with helium as the carrier gas (30 cc/ min). Hydrogen concentrations were measured on-line with stannic oxide sensors. Sensor voltage was converted to hydrogen concentration using the following calibration equation:     1 R0 b ‡ log ; log…H2 concentration† ˆ ÿ Rexp a …1† where a and b are the correlation coecients, R0 the sensor resistance for 100 ppm H2 at 35°C, and Rexp is the sensor resistance measured at experimental conditions. For the sensors used, a, b, and R0 values were ÿ0:9, 1.484 and 100 X, respectively. Gas standards were obtained from Scott Specialty Gases (Plumstead, PA). H2 standards of 1% in N2 and 100 ppm in N2 were used in calibrating selected hydrogen sensors. Pure (99+%) samples of CH4 , CO2 , and N2 were used in calibrating the gas chromatograph. Feed and e‚uent samples were analyzed for total solids (TS), VS and alkalinity using standard procedures (APHA, 1991). In addition, biosolids samples were assayed for hemicellulose, cellulose, lignin, and silica ash using Neutral Detergent Fiber (NDF) (Van Soest and Wine, 1967), Acid Detergent Fiber (ADF) (Van Soest, 1963), and permanganate lignin (Van Soest and Wine, 1968) procedures. Colorimetric assays (Hach Company, Loveland, CO) were used to measure chemical oxygen demand (COD), total nitrogen (an analog of Total Kdeldhal Nitrogen), and total ammonia nitrogen. For each colorimetric assay, a 5 ml aliquot of e‚uent sample was diluted to 200 ml with deionized water. Free or unionized ammonia concentrations were calculated using the following equation (McCarty and McKinney, 1961):

‰NH3 Š ˆ 1:13  10ÿ9

slurry only slurry ‡ 600 mg N/l slurry ‡ 1500 mg N/l slurry ‡ 3000 mg N/l

‰T-NH3 Š ; ‰H‡ Š


where [NH3 ] is the concentration of un-ionized ammonia (mg/l), [T-NH3 ] the concentration of total ammonia (mg/l), and [H‡ ] is the concentration of hydrogen ions (mole/l). 2.6. Statistical analysis Data reported are averages of the reactors for each nominal urea concentration. Analysis of variance on changes in biogas production, methane production, pH, alkalinity, VS, COD, and ®ber content data were conducted. Duncans Multiple Range test (Milton and Arnold, 1990) was used to do overall comparisons for the experiments. Duncans multiple comparisons were conducted only on analysis of variance main e€ects that were signi®cant. For comparisons at speci®c times during an experiment, t-testing was used to determine signi®cance. Signi®cance was reported at an a of 0.05. The SAS (Statistical Analysis System, Cary, NC) system for Windows (release 6.11) was used for statistical analyses.

3. Results and discussion 3.1. Total nitrogen and ammonia concentrations Addition of urea to the experimental digesters caused nitrogen and ammonia concentrations to increase (Fig. 2). Total nitrogen content increased after the addition of urea and remained virtually constant throughout the remainder of the experiment. Unanticipated low nitrogen concentration values at 3 and 24 h (Fig. 2) likely resulted from insucient heating or reaction time during those analyses. This conclusion is consistent with material balance calculations on digester volatile solids described later. Variations in the nitrogen content values in control digesters and in the experimental digesters after 24 h were attributed to experimental error. Increases in total ammonia concentration were slower than for total nitrogen (Fig. 2). Lagging ammonia production was expected due to the time required for the digester micro¯ora to convert urea to ammonia. The ammonia concentration approached the total nitrogen concentration within 24 h in the 600 mg N/l digesters and within 48 h in the 1500 and 3000 mg N/l digesters.

M.C. Sterling Jr. et al. / Bioresource Technology 77 (2001) 9±18


iment. Conversely, values for the 1500 and 3000 mg N/l groups increased to over 200 mg/l by 12 h and the ®nal values of 630 and 3143 mg/l, respectively, were above inhibitory values reported in the literature. 3.2. Biogas composition To determine biogas production, gas volumes (reported at standard conditions of 1 atm and 273.15 K) were corrected for CO2 produced by urea degradation. Since urea degradation produces 2 mole of ammonia and 1 mole of carbon dioxide (Eq. (3)), increases in ammonia concentration were used to calculate the resulting increases in carbon dioxide. Urea ‡ H2 O ! CO2 ‡ 2NH3


For a given time interval, volumetric CO2 production (1 atm, 273.15 K) due to urea degradation was calculated using Eq. (4):   D‰NH3 Š ; …4† DCO2 ˆ 15:821 Dt

Fig. 2. Total nitrogen (closed symbols) and ammonia nitrogen (open symbols) concentrations in digesters after urea addition ((a) control; (b) 600 mg N/l; (c) 1500 mg N/l; (d) 3000 mg N/l added).

Subsequently, no signi®cant changes in either total nitrogen or ammonia concentrations in any of the digesters occurred, suggesting that equilibrium had been reached. The concentration of un-ionized ammonia was related to digester pH and total ammonia concentration using Eq. (2). Calculated values of free ammonia are given in Table 2. Values for the control and 600 mg N/l groups remained below 200 mg/l throughout the exper-

where DCO2 (ml/l d) is the average daily production rate of CO2 during a sampling interval, D‰NH3 Š (mg/l) the increase in digester ammonia concentration during a sampling interval, and (Dt) is the length of the sampling interval (h). For the control and 600 mg N/l digesters, biogas production followed the expected trend for an anaerobic batch process (Fig. 3). After addition of feed, the biogas production rate increased. Following a peak in the production rate between 9 and 12 h, the rate of biogas production decreased through the rest of the experiment. Throughout the experiment, the ratio of gas produced per amount of VS degraded remained within typical values (2.5±4.0 ml/mg) for dairy cattle manure digesters (Loehr, 1984). Biogas production from the digesters with higher amounts of urea added followed a signi®cantly di€erent trend. In the 1500 and 3000 mg N/l digesters, biogas production rates decreased by approximately 30% and 50%, respectively, during the ®rst 3 h after urea

Table 2 Free ammonia concentrations (mg/l) in digesters after urea addition Time (h) 0 3 6 9 12 24 48 72 96 a

Free NH3 (mg N/l)

Control a

1:8  0:15 2:24  0:71 2:66  0:36 2:53  0:30 4:60  1:28 3:99  0:58 3:33  0:18 3:94  0:49 4:10  0:43

Standard deviation of the mean.




4.61  0.69 9.74  0.89 23.47  1.93 28.75  1.03 126.60  21.08 142.58  11.11 163.02  1.98 189.98  4.06 208.96  4.14

9.19  0.74 31.14  2.71 51.37  1.78 80.55  18.08 248.77  21.35 390.10  7.49 568.4  10.58 584.74  10.60 629.88  11.41

8.49  1.96 26.65  4.04 72.39  6.07 99.51  23.54 573.44  3.20 1534.75  11.45 2724.68  21.88 3032.03  12.64 3143.30  5.87


M.C. Sterling Jr. et al. / Bioresource Technology 77 (2001) 9±18

(t ˆ 0 h) values. During this period, the decrease in VS reduction rates in these digesters contributed to the reduction in methane production. However, since the VS reduction was only about 10% lower than for the control digesters, methanogen inhibition likely was the dominant contributor to reduced methane production. Koster and Lettinga (1984) provided data showing that acetogenic methanogens, which produce 70% of the total methane, were inhibited more than the CO2 -utilizing methanogens. 3.4. Hydrogen production Fig. 3. Biogas production in digesters after urea addition (j, control; d, 600 mg N/l; r, 1500 mg N/l; N, 3000 mg N/l).

addition. There was a brief increase in gas production between 12 and 24 h, then the biogas production rates for both groups decreased through the end of the experiment. The ratio of gas produced per quantity of VS degraded (ml/mg) was well below typical values for dairy cattle manure throughout the experiment. These di€erences were due to the combined e€ects of increased solubility of CO2 in the high ammonia digesters due to increased pH and inhibited biogas production as highlighted by decreased methane production.

H2 normally is present in digester biogas, with researchers reporting normal concentrations ranging from 60 to 200 ppm (Harper and Pohland, 1986). Hydrogen data for the control digesters (Fig. 5) shows that there was little variation in the biogas H2 concentration. The addition of urea resulted in a series of changes in biogas H2 concentration for all experimental groups (Fig. 5). For all the experimental groups, the H2 concentration initially decreased after the addition of urea. These reductions partially resulted from the increased pH, which shifted the equilibrium of the digester system. These reductions also resulted from the inhibition of

3.3. Methane production Methane production trends followed the biogas production trends for all digesters (Fig. 4). At 24 h, methane production rates had increased over initial production rates by 175% and 300% in the control and 600 mg N/l digesters, respectively, as a result of increased VS reduction rates in these digesters. After 24 h, methane production rates for these digesters decreased to the initial (t ˆ 0 h) rates as the VS reduction rates decreased to initial values. At 24 h, methane production rates in the 1500 and 3000 mg N/l digesters decreased to 60% of their original

Fig. 4. Methane production in digesters after urea addition (j, control; d, 600 mg N/l; r, 1500 mg N/l; N, 3000 mg N/l).

Fig. 5. Hydrogen concentrations in biogas after urea addition ((a) control; (b) 600 mg N/l; (c) 1500 mg N/l; (d) 3000 mg N/l added).

M.C. Sterling Jr. et al. / Bioresource Technology 77 (2001) 9±18

hydrolytic and acetogenic bacteria due to increased ammonia concentrations in the digesters. For the 3000 mg N/l digester group, the excess ammonia resulted in limiting the H2 concentrations to below 20% of the control values for the entire experiment. For the 600 and 1500 mg N/l digesters, the biogas H2 concentrations increased after approximately 24 h. At 36 h, these increases resulted in peak concentrations of 230 and 150 ppm for the 600 and 1500 mg N/l digester groups, respectively. This indicated that the methanogens were inhibited more at this time than the hydrolytic and acid-forming bacteria. After these peaks, the H2 concentrations for these digester groups decreased over the course of the experiment, consistent with the batch operating mode. 3.5. E€ect on pH The pattern of pH changes exhibited by the control group was typical of a digester undergoing stable operation (Fig. 6). After feeding, digester pH decreased due to increased VFA production. However, as VFAs were metabolized, the pH increased to its normal operating value. The addition of urea caused an increase in pH in all experimental groups (Fig. 6). During the ®rst 12 h, digester pH values increased proportionally with the amount of urea added. After this period, the rise in pH values was signi®cant though small in the 600, 1500 and 3000 mg N/l digesters while no signi®cant increase was observed in the controls. As expected, the ®nal pH values for each of the digesters increased in proportion to the amount of urea added. The ®nal pH values were 8.2, 8.5, and 9.0 for the 600, 1500, and 3000 mg N/l digesters, respectively (Fig. 6). These pH values were below the pH limit of the ammonia bu€er (9.4) (Georgacakis et al., 1982).

Fig. 6. pH in digesters after urea addition (j, control; d, 600 mg N/l; r, 1500 mg N/l; N, 3000 mg N/l).


3.6. Alkalinity The control digesters exhibited an expected alkalinity cycle, with the addition of feed causing alkalinity to decrease immediately as VFAs were formed from soluble substrates in the manure slurry (Fig. 7). As the newly formed VFAs were slowly metabolized and removed from the digester liquor, the digester alkalinity began to return to its stable operating value. The addition of urea resulted in increased digester alkalinity due to the increased ammonium ion concentration (Kroeker et al., 1979; Georgacakis et al., 1982). While all groups had similar initial alkalinity values (2500±3000 mg CaCO3 /l), the initial sample after urea addition (3 h) indicated that the alkalinity values for the experimental digesters remained signi®cantly higher than for the controls. Between 6 and 24 h, the alkalinity of the controls increased to its original value. However, during that same period, the increase in alkalinity for the experimental digesters was 3±4 times that of the controls. In addition, there were no signi®cant di€erences in alkalinity increase among the experimental digesters during this period. After 24 h, the alkalinity increases for the 600 and 1500 mg N/l experimental digesters were not signi®cantly di€erent. However, the increase for each of these was signi®cantly higher than for the controls and signi®cantly lower than for the 3000 mg N/l digesters. Thus, at the end of the experiment, the overall increase in alkalinity was not signi®cantly di€erent between the 600 and 1500 mg N/l digesters. Providing excess ammonia contributed to the increased alkalinity of the experimental digesters in two ways. As shown in Eqs. (5)±(7), one e€ect of adding ammonia was increased bicarbonate concentration in the digesters through the formation of an ammonium salt with bicarbonate taken from dissolved CO2 (Georgacakis et al., 1982). ÿ NH‡ 4 ‡ OH NH3 ‡ H2 O


Fig. 7. Total alkalinity in digesters after urea addition (j, control; d, 600 mg N/l; r, 1500 mg N/l; N, 3000 mg N/l).


M.C. Sterling Jr. et al. / Bioresource Technology 77 (2001) 9±18

CO2 ‡ H2 O H‡ ‡ HCOÿ 3


ÿ ÿ ‡ NH‡ 4 ‡ OH ‡ H ‡ HCO3 ÿ

…NH‡ 4 ‡ HCO3 †salt ‡ H2 O


Another e€ect of high ammonia concentrations is inhibition of both the hydrolytic and acetogenic groups of bacteria (Hill and Bolte, 1987; Angelidaki et al., 1993), thereby reducing VFA concentrations in the digesters. Lower VFA concentrations would decrease the amount of alkalinity used for acid nutralization in the digesters. 3.7. VS reduction

Fig. 8. Manure VS concentrations in digesters after urea addition (j, control; d, 600 mg N/l; r, 1500 mg N/l; N, 3000 mg N/l).

To analyze changes in manure VS reduction after the addition of urea, the following assumptions were used in determining material balances: only urea and ammonia were measured by the colorimetric total nitrogen analysis, ammonia present in the samples volatilized during sample drying, and urea present in the samples remained during drying and volatilized during ashing. Based on these assumptions, the following material balance equations were obtained: ‰Urea±NŠ ˆ ‰Tot-NŠ ÿ ‰NH3 ±NŠ


TS…manure† ˆ TS…measured† ÿ 2:15  ‰Urea±NŠ


VS…manure† ˆ VS…measured† ÿ 2:15  ‰Urea±NŠ


where TS…measured† is the experimentally determined concentration of total solids (mg/l), TS…manure† is the concentration of total solids due to manure (mg/l), VS…measured† the experimentally determined concentration of volatile solids (mg/l), VS…manure† the concentration of volatile solids due to manure (mg/l), [Tot-N] the total nitrogen concentration determined by colorimetry (mg/l), and [NH3 ±N] was the total ammonia concentration determined by colorimetry (mg/l). A conversion factor (2.15) was included to convert urea nitrogen concentrations to urea concentrations. Trends in manure VS reduction in the experimental digesters were similar to that in the control digesters (Fig. 8). VS concentration initially increased in each set of digesters by approximately 5000 mg/l, the amount of feed added at the beginning of the experiment. During the experimental period, VS decreased rapidly during the ®rst 24 h and then more slowly for each set of digesters. For the control digesters, the VS reduction rate was 90:5  9:5 mg VS/l h. This led to a 29:6  3:4% VS reduction during the experiment. For the 600 mg N/l digesters, the VS reduction rate (88:3  9:7 mg VS/L h) was not signi®cantly di€erent than for the control digesters. This led to a 29:4  2:5% VS reduction. However, for the 1500 and 3000 mg N/l digesters, the rate of decrease was lower than for the control digesters, 79:4  6:9 and 74:6  5:1 mg VS/l h, respectively. The

overall VS reductions for these digesters were 25:5  2:5% and 23:7  2:9%. Based on these results, ammonia seemed to have di€erent impacts on VS reduction by anaerobic digestion. At a relatively low concentration (600 mg N/l), ammonia did not a€ect VS reduction and may have served as an easily accessible source of nitrogen. However, higher VS concentrations in the digesters with more urea added suggest that the digester hydrolytic bacteria were inhibited by higher ammonia concentrations. In a similar experiment, Angelidaki et al. (1993) proposed that one-half of the loss in methane yield was attributable to inhibition of the hydrolytic bacteria. However, no mechanism for this inhibition was presented. In a study of anaerobic ruminant bacteria, Ricke and Schaefer (1996) noted that non-limiting growth concentrations of NH3 ±N resulted in lower fermentation product formation, increased lactate formation, and less acetate and propionate reduction. 3.8. Fiber composition An analysis of ®ber composition in the biosolids present in the digester e‚uent was performed to examine whether any particular group of hydrolytic bacteria was more impacted by increased ammonia concentrations. Biosolid samples taken at 3, 6, 9, 12, and 24 h were combined and analyzed using ®ber reduction analysis (Van Soest and Wine, 1967, 1968; Van Soest, 1963). These samples were compared to samples taken during the stable operating period, one week before experimentation. Changes in ®ber composition for the various digester groups are shown in Table 3. Weight percentages for cellulose and hemicellulose were determined analytically, with the remainder of solids categorized as lignin and ash. Pairwise t-tests showed that there were signi®cant decreases in the cellulose concentrations of all digesters after urea addition, while there were no signi®cant decreases in hemicellulose and lignin concentrations. However, because the control digesters

M.C. Sterling Jr. et al. / Bioresource Technology 77 (2001) 9±18


Table 3 Fiber content of digester e‚uent before and after urea additiona Control Cellulose (%) Hemicellulose (%) Lignin ‡ Ash (%) a b

Before After Before After Before After

12.11  2.36 5.96  0.66 12.06  1.89 13.16  2.13 75.84  4.25 80.88  2.79


600 mg N/l

1500 mg N/l

3000 mg N/l

11.27  2.46 10.28  1.97 11.55  1.21 12.24  1.45 77.19  3.67 77.49  3.42

13.12  1.51 9.37  2.29 8.38  1.23 8.39  1.63 78.5  2.74 82.25  3.92

11.64  1.94 7.96  2.20 11.61  2.62 11.89  1.65 76.76  4.56 80.16  3.85

n ˆ 6 for control, n ˆ 4 for other experimental groups. Standard deviation of the mean.

behaved the same as the experimental digesters, it appears that ammonia inhibition was not the cause of the decrease in cellulose concentrations. 4. Conclusions The amount of ammonia nitrogen in the digester feed impacted digester hydrogen production, methane production, and volatile solids removal. Small increases in ammonia nitrogen resulted in increased H2 and methane production. Larger increases caused the inhibition of H2 and methane production. Total biogas production was una€ected by small increases in ammonia nitrogen while higher increases reduced biogas production to 50% of the original rate. Addition of urea led to nearly immediate increases in pH and alkalinity and a somewhat slower increase in free ammonia concentration. Throughout the experiment, the free ammonia concentration remained below inhibitory levels in the 600 mg N/l digesters but reached inhibitory levels within 12 h in the 1500 and 3000 mg N/l digesters. E‚uent VS concentrations were increased by the two largest increases in ammonia nitrogen concentration while the smallest increase resulted in no signi®cant increase in e‚uent VS concentration relative to the control digesters. References Angelidaki, I., Ahring, B.K., 1993. Thermophilic anaerobic digestion of livestock waste: the e€ect of ammonia. Appl. Micro. Biotech. 38, 560±564. Angelidaki, I., Ellegaard, L., Ahring, B.K., 1993. A mathematical model for dynamic simulation of anaerobic digestion of complex substrates: focusing on ammonia inhibition. Biotech. Bioeng. 42, 159±166. APHA, 1991. Standard Method for the Examinations of Water and Wastewater. American Public Health Association, Washington, DC. Archer, D.B., Hilton, M.G., Adams, P., Wiecko, H., 1986. Hydrogen as a process control index in a pilot scale anaerobic digester. Biotech. Lett. 8, 197±202. Archer, D.B., Kirsop, B.H., 1991. The microbiology and control of anaerobic digestion. In: Wheatley, A. (Ed.), Anaerobic Digestion:

A Waste Treatment Technology, vol. 31. Elsevier, New York, 43± 91. Bartlett, K., Dobson, J.V., Eastham, E., 1980. A new method for the detection of hydrogen in breath and its application to acquired and inborn sugar malabsorption. Clin. Chim. Acta 108, 189±194. Boekhurst, R.H., Ogilvie, J.R., Pos, J., 1981. An overview of current simulation models for an anaerobic digester. In: Livestock Waste: A Renewable Resource. American Society of Agricultural Engineers, St. Joseph, MI, pp. 105±108. Braun, R., Huber, P., Meyrath, J., 1981. Ammonia toxicity in liquid piggery manure digestion. Biotech. Lett. 3, 159±164. De Baere, L.A., Devocht, M., Van Assche, P., Verstraete, W., 1984. In¯uence of high NaCl and NH4 Cl salt levels on methanogenic associations. Wat. Res. 18, 543±548. El-Hag, M.G., Vetter, R.L., Kenealy, M.D., 1982. E€ects of silage additives on fermentation characteristics of corn silage and performance of feedlot heifers. J. Dairy Sci. 65 (2), 259±266. Georgacakis, D., Sievers, D.M., Iannotti, E.L., 1982. Bu€er stability in manure digesters. Agri. Wastes 4, 427±441. Harper, S.R., Pohland, F.G., 1986. Recent developments in hydrogen management during anaerobic biological wastewater treatment. Biotech. Bioeng. 28, 585±602. Hickey, R.F., Vanderwielen, J., Switzenbaum, M.S., 1987. The e€ects of organic toxicants on methane production and hydrogen gas levels during the anaerobic digestion of waste activated sludge. Wat. Res. 21 (11), 1417±1427. Hickey, R.F., Vanderwielen, J., Switzenbaum, M.S., 1989. The e€ect of heavy metals on methane production and hydrogen and carbon monoxide levels during batch anaerobic sludge digestion. Wat. Res. 23 (2), 207±218. Hill, D.T., Bolte, J.P., 1987. Modeling fatty acid relationships in animal waste anaerobic digesters. Trans. ASAE 30 (2), 502±508. Hungate, R.E., 1975. The rumen microbial ecosystem. Ann. Review Ecol. Syst. 6, 39±66. Jeris, J.S., Kugelman, I.J., 1985. Secrets to the success of anaerobic digestion. Wat./Eng. Man. 132, 32±35. Kidby, D.W., Nedwell, D.B., 1991. An investigation into the suitability of biogas hydrogen concentration as a performance monitor for anaerobic sewage sludge digestion. Wat. Res. 25 (8), 1007±1012. Koster, I.W., Lettinga, G., 1984. The in¯uence of ammonium±nitrogen on the speci®c activity of pelletized methanogenic sludge. Agri. Wastes 9, 205±216. Kroeker, E.J., Schulte, D.D., Sparling, A.B., Lapp, H.M., 1979. Anaerobic treatment process stability. J. WPCF 51 (4), 718±727. Loehr, R.C., 1984. Anaerobic treatment. Pollution Control for Agriculture. Academic Press, Orlando, FL. McCarty, P.L., McKinney, R.E., 1961. Volatile acid toxicity anaerobic digestion. J. WPCF 33 (3), 223±232. Milton, J.S., Arnold, J.C., 1990. Introduction to Probability and Statistics: Principles and Applications for Engineering and the Computing Sciences, second ed. McGraw-Hill series in probability and statistics, New York, pp. 480±484.


M.C. Sterling Jr. et al. / Bioresource Technology 77 (2001) 9±18

Mosey, F.E., 1982. New developments in the anaerobic treatment of industrial wastes. Water Poll. Cont. 81, 540±550. Mosey, F.E., Fernandes, X.A., 1984. Mathematical modelling of methanogenesis in sewage sludge digestion. In: Grainger, J.M., Lynch, J.M. (Eds.), Microbiological Methods for Environmental Biotechnology. Academic Press, London, pp. 159±168. Ricke, S.C., Schaefer, D.M., 1996. Nitrogen-limited growth response of Selenomonas ruminantium to limiting and non-limiting concentrations of ammonium chloride. Appl. Microbiol. Biotechnol. 46, 169±175. Robinson, J.A., Tiedje, J.M., 1982. Kinetics of hydrogen consumption by rumen ¯uid, anaerobic digester sludge and sediment. Appl. Environ. Microbiol. 44, 1374±1384. Robinson, J.A., Tiedje, J.M., 1984. Competition between sulphatereducing bacteria and methanogenic bacteria for H2 under resting and growing conditions. Arch. Microbiol. 137, 26±32. Sprott, G.D., Shaw, K.M., Jarrell, K.F., 1984. Ammonia/potassium exchange in methanogenic bacteria. J. Biol. Chem 259, 12602± 12608.

Van Soest, P.J., 1963. Use of detergents in the analysis of ®brous feeds. II. A rapid method for the determination of ®ber and lignin. J. Assoc. Ocial Anal. Chem. 46 (5), 829±846. Van Soest, P.J., Wine, R.H., 1967. Use of detergents in the analysis of ®brous feeds. IV. The determination of plant cell wall constituents. J. Assoc. Ocial Agr. Chem. 50, 50±62. Van Soest, P.J., Wine, R.H., 1968. Determination of lignin and cellulose in acid-detergent ®ber with permanganate. J. Assoc. Ocial Agr. Chem. 51, 780±792. Vartak, D.R., Engler, C.R., McFarland, M.J., Ricke, S.C., 1997. Attached-®lm media performance in psychrophilic anaerobic treatment of dairy cattle wastewater. Bioresource Technol. 62, 79±84. Wolin, M.J., 1976. Interactions between H2-producing and methaneproducing species. In: Schlegel, G., Gottschalk, G., Pfenning, A.N. (Eds.), Microbial Production and Utilization of Gases (H2, CH4, CO). Goltze KG, Gottingen, pp. 141±150. Wolin, M.J., 1982. Hydrogen transfer in microbial communities. In: Bull, A.T., Slater, J.H. (Eds.), Microbial Interactions and Communities. Academic Press, London, pp. 323±356.