United States Environmental Protection Agency National Risk Management Research Laboratory Research Triangle Park NC 27711 Research and Development EPA/6QQ/SR-97/071 September 1997 of in M. A. Barlaz, W. E. Eleazer, W. S. Odle, III, X. Qian, and Y-S. Wang The objective of this research was to characterize the anaerobic biodegrad- ability of the major biodegradable com- ponents of municipal solid waste (MSW). Tests were conducted in qua- druplicate in 2-L reactors operated to obtain maximum yields. Measured methane (CH4) yields for grass, leaves, branches, food waste, coated paper, old newsprint, old corrugated contain- ers, and office paper were 144.4, 30.6, 62.6, 300.7, 84.4, 74.3, 152.3, and 217.3 mL CH4/dry g, respectively. While there was a general trend of increasing CH4 yield with increasing cellulose plus hemicellulose (carbohydrate) content, many confounding factors precluded establishment of a quantitative relation- ship. Similarly, the degree of lignifica- tion of a particular component was not a good predictor of the extent of bio- degradation. In parallel with the decomposition ex- periments, leachate from the decom- position of each refuse constituent was analyzed for toxicity using a modified anaerobic toxicity assay. Leachate tox- icity was not found in association with the decomposition of any refuse com- ponent other than food waste. How- ever, substantial toxicity was measured in leachate from the food waste reac- tors. This toxicity was consistent with the behavior of the reactors but could not be simulated with high concentra- tions of carboxylic acids and sodium. The toxicity associated with food waste leachate is not likely to inhibit anaero- bic decomposition in U.S. landfills due to the relatively low concentration of food waste in MSW. Most probable number (MPN) tests were conducted to identify the compo- nents of refuse that carry refuse-de- composing microorganisms into landfills and to evaluate the significance of two typical cover soils as carriers of refuse-decomposing microbes. Grass, leaves, and branches were the major identifiable contributors of refuse-de- composing microbes to landfills, while the cover soils tested did not typically contain populations with the activities required for refuse methanogenesis. This Project Summary was developed by EPA's Air Pollution Prevention and Control Division of the National Risk Management Research Laboratory, Re- search Triangle Park, NC, to announce key findings of the research project that is fully documented in a separate report of the same title (see Project Report ordering information at back). Introduction Approximately 62% of the municipal solid waste (MSW) generated in the U.S. is disposed of by burial in a sanitary land- fill. The production of methane (CH4) from sanitary landfills is well documented, and there are about 119 landfill gas recovery projects currently (January 1997) in op- eration in the U.S. and Canada. While the production of CH4 from landfills is well established, there is large uncertainty in- volved in estimating the amount and rate of CH4 production. This uncertainty is in- creasing as the composition of the MSW buried changes due to increased recy- cling. ------- Development of integrated solid waste management programs, which include re- cycling and in some cases, combustion, have led to a decrease in the use of landfills. However, there is a limit to the types of waste that can be recycled, and combustion has not been the solid waste management alternative of choice for many communities. Thus, landfills will be a sig- nificant part of MSW management for the foreseeable future. Both CH4 and carbon dioxide CO2 are greenhouse gases that contribute to glo- bal climate change. CH4 traps about 20 times more infrared energy than CO2 on a volume basis. Consequently, although landfill gas contains approximately equal proportions of CH4 and CO2, CH4 is more significant with respect to atmospheric cli- mate change. Data on the amount of CH4 that can be expected from refuse already buried, as well as CH4 that will result from the decomposition of refuse buried in the future, are needed to better assess the impact of landfills on global climate change. The overall objective of this research was to develop information on the anaero- bic decomposition of refuse that will im- prove our ability to assess the impact of sanitary landfills on global CH4 accumula- tion. Three sets of experiments were con- ducted to meet this objective: (1) measurement of the CH4 potential of the major biodegradable components of MSW; (2) assessment of whether leachate toxic- ity, associated with whole refuse or some individual constituent, inhibits the onset or rate of CH4 production; and (3) identifica- tion of solid waste constituents that carry the anaerobic bacteria required for refuse methanogenesis. The results of each set of experiments are summarized separately. Experiment 1: Measurement of the CH4 Potential of the Major Biodegradable Components of MSW The anaerobic biodegradability of the major biodegradable components of MSW was characterized by measurement of their CH4 yield and the biodegradation of cellu- lose and hemicellulose. The components that were tested were grass, leaves, branches, food waste, and four types of paper—newsprint (ONP), old corrugated containers (OCC), office paper (OFF), and coated paper (CP). These are the most common types of paper in MSW and also represent the range of biodegradability that could be expected from paper. At one extreme, newsprint contains all of the lig- nin of wood pulp. At the other extreme, office paper has had almost all of the lignin removed. The decomposition of mixed residential refuse was also charac- terized. Tests were conducted in 2-L laboratory reactors in quadruplicate. Each refuse component was seeded with well-decom- posed refuse to initiate refuse-decomposi- tion. CH4 yield data have been corrected for the background CH4 produced from the seed. In the case of food waste, two sets of reactors were tested. In the first (F) series, there was insufficient seed, 30% by volume, and the reactors remained in- hibited. A second set of food waste reac- tors (SF) was then initiated with 70% seed by volume, and these reactors produced measurable CH4. The test conditions were designed to measure the maximum CH4 production potential of each component. This included shredding, incubation at about 40°C, and leachate recycle and neutralization. All re- actors were monitored until they were no longer producing measurable CH4, except for the old corrugated container reactors in which the CH4 yield increased by less than 2% over the final 80 days of monitor- ing. At the termination of the monitoring period, reactors were destructively sampled for analysis of the residual sol- ids. The CH4 yield, solids composition, and extent of cellulose and hemicellulose de- composition for each MSW component and mixed MSW are presented in Table 1. As summarized in Table 1, there was sub- stantial variation in the range of CH4 yields (30.6 to 300.7 ml/dry g) and the extent of decomposition (28 to 94%) among the components tested. In previous research with mixed refuse, carbohydrates ac- counted for 91% of the stoichiometric CH4 potential of MSW. Carbohydrates were the major organic compounds analyzed in the waste components tested here, and the relationship between carbohydrate concen- tration and CH4 yield is presented in Fig- ure 1. While the data in Figure 1 show a relationship, the relatively low correlation coefficient (r2 = 0.49) and failure of the regression line to pass through zero, sug- gest that factors in addition to carbohy- drate concentration influence CH4 yield. Lignin decreases carbohydrate bioavail- ability and is expected to confound the relationship presented in Figure 1. The components with the lowest yields are the two sets of seed reactors and leaves. These are also the components with the lowest carbohydrate to lignin [(C+H)/Li)j ratio. The (C+H)/Li ratio is a measure of the degree of lignification. Values of 3 to 4 have been reported for fresh refuse, and lower values are associated with decom- posed refuse. There is a general trend of more extensive cellulose biodegradation (MC decreasing) in the less lignified sub- strates [(C+H)/Li increasing] e.g., food and office paper (r2 = 0.28). However, the quan- titative relationship is weak because the office paper (C + H)/Li is well above any of the other components tested. The trend towards increased cellulose loss with a decreasing degree of lignification is most definite among the four paper components. There is not a linear relationship be- tween (C+H)/Li and the extent of decom- position (r2 = 0.02). However, it is interesting to note that grass, which is highly lignified, underwent nearly complete decomposition as measured by either MC or the extent of decomposition (Table 1). This suggests that the lignin concentra- tion does not always reflect the degree to which lignin inhibits cellulose bioavailabil- ity. Apparently, the lignins in grass are not as restrictive to microorganisms as the lignin in other components such as branches. This result is consistent with a report that stated, ". . .the chemistry of grass lignocellulose varies considerably from that of wood." The solids decomposition (MC and MH) and CH4 yield data document the biode- gradability of even the most lignified sub- strates, leaves and branches, as well as all other components of MSW tested. The absence of good linear relationships is likely because a number of factors influ- ence CH4 production and solids decompo- sition. The biodegradation of newsprint measured here is in contrast to reports on the excavation of readable newsprint that had been buried in landfills decades ear- lier; however, these reported data did not represent average values, but rather ob- servations during an archaeological exca- vation. The presence of readable newsprint that had not undergone biodegradation may be due to its isolation from other factors required for biodegradation such as bacteria, moisture, and nutrients. The biodegradability of a newspaper buried in a bag that did not break during waste compaction would differ from the biode- gradability of newsprint exposed to other refuse components. Based on the CH4 yields presented in Table 1, a model was constructed to esti- mate CH4 yields based on assumed com- positions of buried refuse. These results are summarized in Table 2. The actual methane yield per wet kg of refuse buried decreases by only 10% between the base case (64.9 L CH4/wet kg) and the case with the most recycling (58.6 L CH4/wet kg). However, the appropriate way to evaluate changes in methane yield is to calculate the change in methane potential ------- Table 1. CH4 Yield and Initial and Final Solids Composition Data Summary3 Reactor Series Seed sd Seed-2d sd Grass sd Raleigh grass sd Leaves sd Branch sd Yield mL CH4/ dryg 25.5 5.7 5.8 0.6 144.4 15.5 127.6 21.8 30.6e 8.6 62.6e 13.3 Cellulose 23.4 18.3 26.5 25.6 15.3 35.4 Hemi- cellulose 4.7 3.7 10.2 14.8 10.5 18.4 Lignin 22.5 22.1 28.4 21.6 43.8 32.6 MCb 0.18 0.02 0.34 0.01 0.19 0.01 0.43 0.05 0.52 0.05 MHb 0.36 0.03 0.69 0.11 0.42 0.06 0.68 0.10 0.59 0.02 Extent of Decomposition0 21.8 6.3 94.3 75.5 28.3 27.8 Food 46.1 6.2 8Q .6 Second Food sd ONP sd OCC sd OFF sd CP sd MSW sd 300.7e 10.6 74.33 6.802 152.3 6.7 217.3 14.96 84.4 8.1 92.0e 4.1 55.4 48.5 57.3 87.4 42.3 28.8 7.2 9 9.9 8.4 9.4 9 11.4 23.9 20.8 2.3 15 23.1 0.24 0.02 0.73 0.05 0.36 0.01 0.02 0 0.54 0.01 0.25 0.03 0.58 0.04 0.46 0.06 0.38 0.01 0.09 0.01 0.58 0.06 0.22 0.05 84.1 31.1 54.4 54.6 39.2 58.4 " Data represent the average for each reactor set. Standard deviations (sd) are presented below the average where data are the average of all reactors in a set. b The ratio of the cellulose (MC) or hemicellulose (MH) recovered from a reactor divided by the mass added initially. c The measured CH4 yield divided by the yield calculated by assuming conversion of 100% of the cellulose and hemicellulose (and protein in the case of food waste) to CH4 and CO2. d Seed used for second set of food waste reactors. 8 Yield data for the leaf reactors exclude L2, data for the branch reactors exclude B4,and data for the second food and MSW reactors were corrected for leakage. based on the yield multiplied by the mass landfilled. Using this calculation, the po- tential reduction in methane production is 25.5% and 38% for the recycling cases based on national averages and local re- cycling rates, respectively. Thus, these data suggest that recycling can have a substantial impact on the volume of meth- ane available for recovery over the de- composition period. Where CH4 is released to the atmo- sphere, recycling clearly reduces the amount of CH4 released from landfills. However, at landfills where there is an active program to compare the relative benefits of recycling and energy recovery. Given the CH4 potential data for individual constituents measured here, this analysis could be done on a component-specific basis because the results may be differ- ent for two different types of paper or between yard waste and paper. The calculated composite CH4 yields in Table 2 range from 58.6 - 64.9 L CH4/wet kg of MSW. These values are low relative to landfill gas models that generally as- sume a yield of 62.3 to 112.2 L/kg. This is surprising in that the CH4 yields measured here were measured under optimal condi- tions and should be considerably higher than values assumed for field conditions. There are two potential explanations for this discrepancy. The first explanation is that the assumed waste composition is in error. The data presented in an EPA re- port represent an estimate of MSW gen- eration and exclude a number of wastes that are buried in landfills. Some of these other wastes have high CH4 yields (waste- water treatment plant sludge and agricul- tural and food preparation wastes), while others have little or no CH4 potential (wa- ter treatment plant sludge and construc- tion and demolition debris). A second explanation for the discrepancy in yield calculations pertains to the assumptions used by the landfill gas models. The range of values used, 62.3 to 112.2 L/kg, is based on field measurements and an esti- mate of the mass of waste buried in a landfill. While this mass is accurately known in newer landfills where all waste received is weighed, this mass represents only an estimate at older facilities and errors of 20 to 30% would not appear to be unreasonable. Thus, the values as- sumed in practise may be inaccurate. ------- 350 -i Second Jk food ^ r2 = 0.493 100 Cellulose + Hemicellulose (%) Figure 1. CH4 yield vs. carbohydrate concentration. Table 2. Calculated CH4 Yield Based on Measured Yields and Assumed MSW Composition Case Base Case-No Recycling Recycling at National Average Recycling-Typical Local Program Yield (L CH4/ wet kg) 64.9 59.9 58.6 Methane Reduction3 (%) na 25.5 38.0 Recycle Rate (%) na 19.4 30.9 1 Calculated from the CH4 yield multiplied by the mass buried after recycling relative to the CH4 yield and mass buried in the base case. Experiment 2: Measurement of the Anaerobic Toxicity of Leachate Associated with the Decomposition of Individual Refuse Components The anaerobic toxicity of leachate asso- ciated with the decomposition of each refuse component tested above was mea- sured in parallel with the decomposition experiments. Leachate was sampled three times from each reactor. For food waste, four samples were collected from the F reactors, but no samples were collected from the SF reactors. Six samples were collected from the MSW reactors. The ini- tial strategy was to sample each reactor twice during the acid phase and twice during the decelerated CH4 production phase. However, except for the first set of food reactors (F), the acid phase was very brief. As a result, only one sample was collected from most reactor sets dur- ing the acid phase. Leachate toxicity was evaluated using a modified anaerobic toxicity assay (ATA). The ATA included anaerobic medium, ground refuse as a carbon source, and an inoculum. The inoculum was a methano- genic consortium enriched from decom- posed refuse with ground refuse as a carbon source. CH4 production from the ground refuse was measured in triplicate in the presence and absence of leachate. Leachate was tested at final concentra- tions in the ATA of 25 and 80% of its original strength. Two sets of controls were also inoculated. Controls containing inocu- lum and medium but no refuse were used to measure background CH4 production from the inoculum. Controls containing in- oculum, medium, and ground refuse were used to compare CH4 production in the presence and absence of leachate. Leachate toxicity was not measured in association with the decomposition of any refuse component other than food waste. However, leachate associated with the food waste reactors containing 30% seed and 70% food waste (F) exhibited sub- stantial toxicity, and this toxicity was gen- erally consistent with the behavior of the reactors. The toxicity of the food waste leachate could not be simulated with synthetic leachate containing high concentrations of carboxylic acids and sodium. ATAs with 20, 5, 15, and 12 g/L of acetate, propi- onate, butyrate, and sodium, respectively, suggested that high concentrations of bu- tyric acid and sodium inhibited the onset of CH4 production, but that refuse micro- organisms could acclimate to these con- centrations within 5 to 10 days under the conditions of the ATA. The corresponding concentrations of undissociated acetic, pro- pionic, and butyric acids were 113, 27, and 96.8 mg/L, respectively. Comparison of carboxylic acid concentration data from the S and SF reactors series indi- cated that the refuse ecosystem was able to tolerate and recover from 142, 35, 24, and 305 mg/L of undissociated acetic, propionic, i-butyric, and butyric acids, respectively. These concentra- tions of undissociated, carboxylic acids are higher than concentrations reported to be inhibitory in previous research with anaerobic digesters. Experiment 3: Identification of Solid Waste Constituents that Carry the Anaerobic Bacteria Required for Refuse Methanogenesis The objective of part of this study was to identify the components of refuse that carry refuse-decomposing microorganisms into landfills. A second objective was to evaluate the significance of two typical cover soils as carriers of refuse-decom- posing microbes. Refuse buried in a sani- tary landfill is typically covered with 15 cm of soil daily. Recently, geotextile sheets and foams have been proposed as alter- natives to soil to minimize the volume of soil in a landfill. While soil may contribute refuse-decomposing microorganisms to landfills, the proposed alternatives almost certainly do not. The total anaerobic population and the subpopulations of cellulolytic, hemicellu- lolytic, hydrogen-producing acetogenic (based on butyrate catabolism) bacteria ------- and acetate- and hydrogen (H2)/CO2-uti- lizing methanogenic bacteria were enu- merated by the most probable number (MPN) technique on several waste com- ponents: grass, branches, leaves, food waste, whole refuse, and landfill cover soil. For each component, the objective was to enumerate microbial populations on a representative sample in the form in which it would typically enter a landfill. Although paper represents 37.6% of refuse, it was not tested because it is likely populated with bacteria originating in wet components of refuse. Microbial enumerations were performed by MPN tests. Thus, it was necessary to form a liquid inoculum from solid samples. The technique used here was similar to a technique developed previously to process smaller samples. In the laboratory, refuse samples were placed in a 113-L plastic garbage can which had been wiped with ethanol and purged with sterile argon. A measured volume of filter sterilized anaero- bic phosphate buffer (23.7 mM, pH 7.2) was then added to a sample to form a slurry. The sample was then stirred by hand (covered with arm length gloves). Next, four samples were removed using a 1-L beaker, and the liquid was poured into a sterile, 4-L, argon-purged, Erlenmeyer flask. The liquid in this flask served as the inoculum for MPN enumerations. Inocula were serially diluted in phosphate buffer (23.7 mM, pH 7.2). For soil, 250 to 300 g of each sample was added directly to a nitrogen-purged flask, 2.5 L of sodium py- rophosphate (0.1%, pH 7) was added, and the slurry was shaken for 2 minutes. The slurry was then allowed to settle for 1 minute after which a liquid sample was removed for use as an inoculum. Microbial populations on each waste component and whole refuse are reported in Table 3. Total anaerobic and hemicellulolytic populations were present on all components tested, while the pres- ence of cellulolytic, acetogenic, and methanogenic bacteria was more limited. Thus, identification of the waste compo- nents that are the major contributors of cellulolytic, acetogenic, and methanogenic bacteria is evaluated here. Yard waste (grass, leaves, and branches) most con- sistently carried the microorganisms re- quired for refuse methanogenesis. Surprisingly, food waste did not carry ei- ther cellulolytic or methanogenic bacteria, and one of two food waste samples con- tained only one acetogen per gram. Popu- lations of cellulolytic, acetogenic, and methanogenic bacteria were generally lower in the mixed refuse samples com- pared to the grass, leaves, and branch samples. Table 3. Anaerobic Microbial Populations on Refuse Components (Most Probable Number—Iog10 cells/dry g)a Trophic Total Hemicellu- Methanogen Methanogen Group Anaerobes lolytic Cellulolytic Acetogen Acetate H2/CO2 Grass (April 92) Grass (July 92) Branches 9.8 9.8 6.5 7.9 9.5 4.2 1.4 t> 2.5 0.7 1.8 1.3 1.6 t> 1.1 1.8 1.3 0.8 Leaves (Nov. 91) Leaves (Nov. 92) Food (Mar. 92) Food (Aug. 92) Refuse (July 92) Refuse (Sept. 92) 5.8 6.9 >8.0C 9.4 9.3 8.4 4.1 4.4 5.3 6.2 6.6 6.3 1.0 <0.4 1.0 1.8 4.4 3.0 <-0.4 <-0.1 <-0.1 <-0.4 0 <0 0.4 0.2 3.6 <-0.2 <-0.1 <-0.1 0.7 3.8 <-0.1 <0 5.0 0.8 Grass, leaves, and branches were the major identifiable contributors of refuse- decomposing microbes to landfills. About 9% of the refuse stream is characterized as "miscellaneous" and contains many dif- ferent items. In addition to diapers and pet wastes, there may be other compo- nents in the miscellaneous fraction that carry refuse-decomposing microbes. How- ever, their presence is small, and they are likely to be poorly distributed. The impor- tance of yard waste should be considered as solid waste management programs are implemented. Where there is interest in CH4 recovery from landfills, banning yard waste from landfills may be self-defeating. Unless, of course, the landfill is receiving substantial volumes of other wastes known to carry refuse-decomposing microbes. The cover soils tested did not typically contain populations with the activities re- quired for refuse methanogenesis. Thus, efforts to develop lower volume alterna- tives to cover soil will not adversely im- pact the input of refuse-decomposing microbes to landfills. Financial support for this research was provided by the Climate Change Research Program of the U.S. Environmental Pro- tection Agency, Waste Management In- corporated, the National Science Foundation, and S. C. Johnson Wax & Son. This support is gratefully acknowl- edged. We are also grateful for the assis- tance of Kathi McBlief in editing, typing, and figure preparation. " Data reported as less than a number indicate that no positive tubes were detected. The number reported assumes one positive tube in the first dilution. b MPN results code was anomalous and not reported. 0 All tubes were positive at the highest dilution tested. ------- M. A. Barlaz, W. E. Eleazer, W. S. Odle, III, X. Qian, and Y-S. Wang are with North Carolina State University, Raleigh, NC 27695-7908. Susan A. Thorneloe is the EPA Project Officer (see below). The complete report, entitled "Biodegradative Analysis of Municipal Solid Waste in Laboratory-Scale Landfills," (Order No. PB97-189674; Cost: $35.00, subject to change) will be available only from National Technical Information Service 5285 Port Royal Road Springfield, VA 22161 Telephone: 703-487-4650 The EPA Project Officer can be contacted at Air Pollution Prevention and Control Division National Risk Management Research Laboratory U.S. Environmental Protection Agency Research Triangle Park, NC 27711 United States Environmental Protection Agency Center for Environmental Research Information Cincinnati, OH 45268 Official Business Penalty for Private Use $300 BULK RATE POSTAGE & FEES PAID EPA PERMIT No. G-35 EPA/6QO/SR-97/071 ------- |