L UTVD6 -TVV/^--1018 DEPARTMENT OF WATER RESOURCES ENGINEERING LUND INSTITUTE OF TECHNOLOGY, LUND UNIVERSITY CODEN: LUTVDG/(TVRL-1018)/(1997)

BIOGAS GENERATION IN LANDFILLS Equilibria, rates & yields

by

Mattias Akesson

DJSTRSUTTQN OF THK DGCUMBtfr 18 UMJ41TH) ISSN 1101-9824 I

DISCLA1liT/IW

Portions of this document may be illegible in electronic image products. Images are produced from the best available original document

t i * \

!

! BIOGAS GENERATION IN LANDFILLS

Equilibria, rates & yields

av

Manias Akesson

AKADEMISK AVHANDLING som for avlaggande av teknisk doktorsexamen vid tekniska fakulteten vid Lunds Universitet kommer an offentligen forsvaras vid Sektionen for vag- och vattenbyggnad, John Ericssons vag 1, horsal V:C, torsdagen den 15 maj 1997, kl 10.15. enl SIS 61 41 21 Recipient's Classification ISSN Distribution Supplementary Signature Keywords sources I, Manias S-221 Department cells, Title kinetics domestic on Abstract BIOGAS of distribution methanogenesis emphasis Authors) that waste LUND limitation. of coefficients, of methanogenic generation, that Organization Landfill, guidance be 13

the

P chemical undissociated waste 150-200 -19

substrate high exceeded

and and among undersigned,

00 and

UNIVERSITY permission Nm'

AKESSON

Landfilling A and

To In Comparisons Lund,

composition key subtitle test was waste . or GENERATION biogas

thermodynamic

food-rich notes

and

sum,

by pretreatment, oxygen a explain which and of of capacities. system CH,/dry

title

cell, given condition notion the inhibition

bibliographical

(name 40-70 Water VFA SWEDEN and

were

rates the bacterial acetic

refuse being to methane, the

to demand, equipped

and/or in the waste ’

results NmVdry

s publish

and supported assumed

was corresponded

Resources

of the observed tonne,yr, biocell "cells" acid,

discrepancies the

content, methane of is

^

address) mineral

characterization

investigated. fractions

encompass index

IN showed not model, acetotrophic copyright

and biogas, indicating N,

with information

is has

to tonne, ^ LANDFILLS a

a P

in

by whereas disseminate Engineering

terms equilibrate, which discrete landfill

production contents. and become

gas situ. in

that were with the thermodynamics,

respectively), which between owner

several K extraction

Six (if

the

AG^ suggests findings the were

methanogenesis designed low VFA 10 one, any)

of more

8,000

Department Lund reduction

However,

advantages

-14

at of the the - conversion

rates

VFA aspects quantified.

ranged

Equilibria, theoretical but a levels the

common abstract Nm certain oxidations the m

from systems University,

for and

5

abstract the rather

levels. test

turnover ’

ranged

of biogas of between CH/dry results

kinetics, internal possible

the

of

cells,

lower

of

the

proton

processes

of is

Biomass methane and

a rates a in Rates

Water thermodynamic

one

of of continuous reduced the

production. performance.

from between recent

with

Box bottom of

volatile limit the

conditions -11

nutritional volatile & tonne,yr above-mentioned of

gradients obtained microbial Number Security

Resources

DOCTORAL above-mentioned pools Date Document yields CODEN: and Sponsoring the commercial contents 118, to

potentials

for years. content

-15 plastic

10 most

fatty governing

S-221 of energy transition. fatty were

In and and

kj/reaction

of in observed over issue classification Different limitation biomass

this ranging

model, test acids

important

of

pages

Engineering liners,

name Lund acids, and 00 18-77 24 estimated organization

waste conservation Date

readily

cytoplasmic thesis, DISSERTATION

Lund,

g0 cells

mechanism dissertation. (VFA) quantified

and during University, limitation, were from and dissertation, 2

. g0 /l. waste

test was and with

the

biodegradable factors

SWEDEN

questions 2 from Corresponding

that /l,

(2 cells mainly monitored

investigated. decreased dependence

a mainly

respectively.

two membranes. (AG^J,

in yields methane the

showed governing

LUTVDG/

model

<7) 97-04-07 the NUTEK,

hereby

year commercial

shift the

commercial

topics are

and (in

for with Price ISBN Language explained

period, material, notion

that of yields, Pools from

this 97-05-15 grant guided

values seven

hydrogenotrophic

This biogas the

increasing of SYSAV,

the

TWR-1018 study

and bioenergetics, demonstrated to acidogenic waste of rate

growth

demonstrate years.

assimilation in to achieved

the waste all nutritional generation

English

of emissions

domestic different

found reference

relative APR to

biogas Great levels

yield were pure

by to to

Abstract

Landfilling in "cells" has become more common in recent years. Different waste streams are guided to different cells, among which the biocell is a landfill designed for biogas production. In this thesis, the dependence of biogas generation on waste composition was investigated. Six 8,000 m3 test cells, with contents ranging from mainly commercial waste to pure domestic waste and equipped with gas extraction systems and bottom plastic liners, were monitored for seven years. Great emphasis was given to the characterization of conversion processes and governing mechanism in the topics of bioenergetics, kinetics and capacities. A thermodynamic model, in which the oxidations of volatile fatty acids (VFA) (2 < C < 7) and hydrogenotrophic methanogenesis were assumed to equilibrate at a certain lower limit for energy conservation (AGmin), explained the relative distribution of VFA’s observed in situ. AGmin ranged between -11 to -15 kj/reaction and decreased with increasing levels of undissociated acetic acid, indicating the reduction of proton gradients over cytoplasmic membranes. Comparisons of methane production rates and internal conditions observed during a two year period, demonstrated that high biogas rates corresponded with low VFA levels. Rates obtained in test cells with mainly commercial waste were 13 -19 Nm3 CH4/dry tonne,yr, whereas VFA levels ranged between 10 and 24 g02/l. Corresponding values in domestic waste and food-rich waste fractions were 10 - 14 Nm3 CH4/dry tonne,yr and 18 - 77 g02/l, respectively. This demonstrate that substrate inhi bition of acetotrophic methanogenesis is one of the most important factors governing the rate of biogas generation, a notion supported by the findings from the thermodynamic model, and that the shift from acidogenic to methanogenic condition is not a discrete one, but rather a continuous transition. To explain the discrepancies between theoretical methane potentials and quantified yields (in this study found to be 150-200 and 40-70 Nm3/dry tonne, respectively), the possible nutritional limitation was investigated. Pools and emissions of chemical oxygen demand, N, P and K were quantified. Biomass pools were estimated from methane yields, growth yield coefficients, and bacterial mineral contents. However, results from commercial waste test cells showed that the assimilation of P exceeded the refusecontent, which suggests the turnover of microbial biomass and questions the notion of nutritional limitation. In sum, the results showed that the advantages of a reduced content of readily biodegradable material, achieved by guidance or pretreatment, encompass several aspects of the performance. Contents

1. Background, aim & outline 6 2. Catabolism & metamorphosis 8 Conversion paths 8 Biochemical transformations 10 3. Bioenergetics 13 Free energy changes of reactions 13 Energy conservation 14 Equilibria 15 4. Kinetics 18 Temperature 19 pH 20 Substrate 20 Inhibitors 22 Rate-limiting steps 23 5. Capacities 24 Potentials 24 Yields 25 Nutrition 26 6. Test cell methodology 28 Fixation of non-studied variables 28 Potentials 29 Biochemical conditions 30 Gas extraction 31 7. Concluding remarks 33 8. Acknowledgements 35 9. References 36 This thesis is based on the following publications, which will be referred to in the text by their Roman numerals.

I Seasonal Changes of Leachate Production and Quality from Test Cells. Mattias Akesson and Peter Nilsson, J. Envir. Eng. ASCE (in press)

II Estimation of Landfill Properties through Simulation. Mattias Akesson, (submitted)

III Thermodynamic Model of Biochemical Conditions in Landfills. Mattias Akesson, (submitted)

IV Material Dependence of Methane Production Rates in Landfills. Mattias Akesson and Peter Nilsson, Waste Manage. Res. (in press)

V Landfill Ecosystem Nutrition. Mattias Akesson, (submitted) 1. Background, aim & outline

For the last few decades the generation of biogas in landfills has been regarded as a valuable energy source receiving considerable attention. The importance of reducing the generation time period was recognized early on. Therefore, several attempts to control and enhance stabilization have been performed. These have mainly made use of physical models by monitoring the effects of different manipulations, such as leachate recirculation, pre-composting and sewage sludge addition. Biogas generating systems found in landfills are complex and involve microbial communities, which interact with their environment, and a mixture of refusematerials containing a variety of substrates. In order to determine the effects of a certain technique, it is, therefore, insufficient to simply quantify the biogas production - additional information must also be assembled to support interpretations of the microbial conversion. Such characterizations have lead to a general description of the development of biochemical conditions as a sequence of degradation states. Moreover, lab-scale experiments have improved the understanding of how different organic compounds contribute to biochemical conditions and gas yields. Still, the existing information concerning the mechanisms governing rates and yields is scarce, and the ability to predict the performance of a landfill for a given technique of operation is limited. Landfilling in separate "cells" was introduced in Sweden during the late 1980s. The most common form of implementation of this system is the use of "biocells". This technique may be characterized as a landfill designed for landfill gas production and involves such features as clay embankments shielding the cells from their surroundings; installation of horizontal gas extraction systems during the active period, and, a generally improved technique of operation. A more extensive implementation of the system is the simultaneous use of a number of different cells, although at present, this is employed in few places only. Each cell is designed to receive a certain "waste stream". This division is based on the biodegradability of the waste composition and the content of hazardous substances. Consequently, such a system includes specific cells for poorly and non-biodegradable waste material (e.g. construction and demolition waste). The implication is that the operator of the landfill is able to control the content of the biocells by guiding different truck loads to different cells. With the advent of such a system, the question will arise as to how this guidance should be conducted, since different waste compositions are likely to differ in performance regarding landfill gas production, leachate quality and biodegradability in general.

6 Background, aim & outline 7

The aim of this study was to determine the relationship between biogas generation and refuse composition. The investigation was performed with landfill test cells at the Spillepeng landfill in the City of Malmo. Six 8,000 m3 test cells, with contents ranging from mainly commercial waste to pure domestic waste and equipped with gas extraction systems and bottom plastic liners, were monitored for seven years. Great emphasis was given to the characterization of conversion processes and the governing mechanisms by substantiating observations from the test cells with conventional biochemical fundamentals. After describing the function of the leachate drainage system (I) and the waste streams directed to the test cells (II), the microbial conversion was characterized by addressing the thermodynamic condition (III) as well as the limiting factors that govern the rates (TV) and yields (V) of biogas formation. The outline of this summary was largely determined by the investigation direction. Firstly, the current understanding of landfill conversion process and degradation states is reviewed. Secondly, topics in which attempts were made to further characterize the conversion, i.e. bioenergetics, kinetics and capacities, are given a separate section each. Finally, the main framework of test cell methodology is discussed. 2. Catabolism & metamorphosis

In landfills as in nature, the presence of dead organic material enables certain microbes to grow. Both material and chemically bound energy are essential for their metabolism. The presence of water is crucial, since all conversion and transport to and from microbes take place in the water phase. A thick and compact landfill cover and the uppermost refuse layer act as an enclosure which restricts the migration of atmospheric oxygen into the landfill. The content of molecular oxygen in the landfill will, therefore, be depleted soon after the filling. If oxygen was available, as in a composting process, most of the degradable organic carbon would be oxidized to form carbon dioxide. Instead, the mean oxidation number of organic carbon is kept constant and a significant part of the carbon is released as methane. The end product is, therefore, an energy-rich gas mixture of methane and carbon dioxide- biogas. This implies that only a minor part (approximately one eighth) of the chemically bound energy in the degradable organic material is available to anaerobic organisms. Consequently, the conversion rates in landfills are much slower than in composting plants.

Conversion paths The generation of biogas in landfills is believed to proceed along pathways similar to those described in anaerobic sludge digestion (Zehnder, 1978). If organic material is to be anaerobically converted, cooperation is needed between a number of microbial groups. According to current understanding, at least four different groups partake in the conversion of the material in sequence (Figure 1): 1) primary fermenters, such as cellulolytic and hemi- cellulolytic bacteria, which use extracellular enzymes to hydrolyze polymers and ferment monomers to volatile fatty acids (VTA), alcohols, hydrogen and carbon dioxide; 2) secondary fermenters or acetogens, which convert VFA through anaerobic oxidation to acetate and hydrogen; 3) acetotrophic methanogens and 4) hydrogenotrophic methanogens. Some groups are mutually dependent in syntrophic cooperations. This is the case for acetogenic and hydrogenotrophic methanogenic bacteria. The latter group depends on the former one for its substrate, i.e. H2, but it also exerts a significant influence by the removal of products (Schink, 1991). Consequently, in an efficient biogas producing system, each microbial group must have a certain level of activity, with respect to each other. The fractions of methane formed from acetate and hydrogen were found to be approximately 70% and 30%, respectively (Jens and McCarty, 1965). The corresponding ratio between acetate and intermediates is less evident, but experiments demonstrate that fermentative bacteria tend to change their

8 Catabolism & metamorphosis 9

f BIOPOLYMERS V

Figure 1 Conversion paths of biogas generating ecosystems (adapted from Schink, 1991). fermentation pattern towards acetate, hydrogen and carbon dioxide as products if the hydrogen pressure is kept low (Schink, 1988). Acetate can either be converted directly to methane or be oxidized to form carbon dioxide and hydrogen. The latter path will only occur in an obligatory syntrophic consortium of at least two bacteria (Schniirer, 1995). This path was found to be the dominating mechanism in reactors with high levels of VFA (Schniirer, 1995). However, the proportion of acetate oxidation increased with lower acetate concentrations in a sewage sludge digestor (Petersen and Ahring, 1991). Populations of all trophic groups were enumerated in a lab-scale experiment (Barlaz et al., 1989b), as well as in samples from full-scale sites ((ones et al., 1983; Mormile et al., 1996). Little information is available regarding the forms of biomass in landfills. Syntrophic cooperations, though, are best performed over short distances. It is thus reasonable that members of such cooperations would aggregate in a mosaic arrangement (Schink, 1991). This can be achieved by the development of either surface bound films or freely suspended floes. Both may very well occur, since floes can be formed through detachment of films by hydraulic forces (Stronach et al., 1986). Films have been observed in drainage pipes (Brune et al., 1991) and cells were typically found to be associated with refuse surfaces in landfill samples (Suflita et al., 1992). However, a theoretical study of the specific surface of waste materials indicated that the methanogens are arranged as freely suspended floes (or individuals) rather than films (Ferguson, 1993). 10 Catabolism & metamorphosis Biochemical transformations During and following filling, a sequence of different biochemical conditions will evolve in the waste materials as a result of the conversion processes. The sequence has been divided into different phases or degradation states by several researchers (Farquhar and Rovers, 1973; Ehrig, 1987; Christensen and Kjeldsen, 1989; Barlaz et al., 1989b; Lagerkvist, 1995). There are some inconsistencies between the suggested divisions, mainly due to a variety of different phase definitions. The main features of the conventional view of landfill transformation are presented in Figure 2. An awareness of the variety of organic material is essential for understanding the transformation. Different compounds have been classified according to the rate by which they degrade (Hoeks, 1983). Sugars and starches are considered to be readily biodegradable materials, whereas cellulose and hemicellulose are regarded to be converted more slowly. The first phase is defined as the period when nitrate and entrained oxygen are depleted. In lab-scale experiments, this process was found to consume 2% of the sugars (Barlaz etal., 1989a). Although during filling in real operations with slow rates of filling and insufficient compaction, gas exchange will occur down to such depths that a significant amount of the organic

Figure 2 Development of gas composition and biochemical conditions in landfills (adapted from Barlaz et al., 1989b; Christensen and Kjeldsen, 1989). Catabolism & metamorphosis 11 material will be aerobically converted (Ehrig, 1991). Provided a thick top cover, this condition is diminished after closure. With the advent of anaerobic conditions, the remaining part of readily biodegradable material will be subjected to fermentations, which, in turn,lead to the rapid build-up of fermentation products (mostly VFA, carbon dioxide and hydrogen) with the effect that the pH is lowered. This condition inhibits methanogenesis and the removal of VFA and can be retained unintentionally (Farquhar and Rovers, 1973) or deliberately (Lagerkvist, 1995) for long time periods by extensive infiltration. The second phase ends with the growth of methanogens or, rather, an increased methane production (qcm)- The distinction is a small but significant one, because methanogens can grow and produce enough methane soon after filling to make the content in the voids quite high. However, a substantial methane generation does not occur until later. In lab-scale experiments, one or two months elapsed from detection to accelerated methane production (Ehrig, 1987; Barlaz et al, 1989b). In the Spillepeng test cells, methane was observed almost immediately after construction; 45 Vol% after 2-3 months (Nilsson and Edner, 1991). However, it was not until over a year after the final covering that extraction began. The decline of inhibiting conditions is not yet fully understood. The rapid shift in lab-scale experiments can be ascribed to external neutralization (Barlaz et al., 1989b). For test cells, the shift was overlooked by regarding the condition as a mosaic combination of inhibiting and non-inhibiting zones (I). However, as will be shown in Section 4, the importance of such a concept should not be overestimated, since methanogenesis can occur at high VFA levels. The third phase is characterized by high rates of biogas generation, which lead to the slow reduction of dissolved VFA, and ends with the depletion of VFA. This results in increasing pH values and alkalinities during the transformation. The process may take several years in test cells (TV) and full-scale landfills. At this stage, the gas phase is mainly a mixture of methane and carbon dioxide with proportions of 50-60 Vol% and 40-50 Vol%, respectively. The composition is determined by the mean oxidation number of the carbon in the converted material (Gujer and Zehnder, 1983). Moreover, the high solubility of carbon dioxide contributes to a shift towards a higher methane content. Methane generation will proceed during the fourth phase, but with lower rates than during the previous one, making exclusive useof more slowly degradable material. Eventually, the biogas generation will be reduced to such low levels that the intrusion of air will be possible, although no information was found in the literature describing such conditions in a "modern" sanitary landfill. 12 Catabolism &■ metamorphosis

The changes of biochemical conditions are also reflected in the leachate quality. Compiled results from a large number of leachate characterizations from drained landfills lead to the well-known classification of leachates as acetic and methanogenic (Ehrig, 1983; 1989). Some leachate parameters (pH, chemical and biological oxygen demand - COD and BOD - sulphate, calcium, magnesium, iron, manganese and zinc) exhibited different ranges in these classes, whereas others did not. One parameter in the latter case (phosphate) was, however, found to be highly correlated with parameters reflecting the content of dissolved organic material (I), indicating the phase dependence of this constituent. Acetic leachate qualities correspond to the second and third phases outlined above, whereas methanogenic leachates correspond to the fourth phase. The distinction between the second and the third phases lead to the denotation of these as the acidogenic and arid-consuming phases, respectively (I). A secondary effect of the activity of methanogens is the possible precipitation of carbonates on the bacterial surfaces. Such processes have been observed on several microbes in nature (Ehrlich, 1981). According to a model proposed by Brune etal. (1991), alkaline conditions would develop close to the active bacteria, enabling cations such as Ca2+, Mn2+ and Fe2+ to precipitate as carbonates. A similar mechanism for the formation of sulphides was also suggested. Precipitation was found to be promoted by acetic conditions (Brune et al., 1991). These processes are the main causes of incrustation in leachate drainage systems and can be a real nuisancefor the landfill operator. They also explain the phase dependency of the metals described above. 3. Bioenergetics

Anaerobic conditions in landfills restrict the prospects for microbial energy utilization of degradable material. Moreover, the cooperation of at least four different trophic groups means that the available energy has to be shared, which further confines the yields of the mediating organisms. These aspects are essential characteristics of biogas generating systems, which accordingly would be elucidated by analysis of the bioenergetical conditions of the trophic groups. In principal, such studies involve the calculationof energy yields for the mediation of certain reactions and can be performed if the substrate and product activities are known. This confines the studies to VFA oxidations and methanogenesis, since the concentrations of monomers and alcohols are usually below the detection limits. Landfills were studied in two such investigations (Mormile et al, 1996; III). In related disciplines, such as anaerobic waste water treatment, bioenergetics has previously provided a useful tool enabling a wider insight into the complex systems described in the previous section (McCarty, 1971; Smith and McCarty, 1989; Hickey and Switzenbaum, 1991; Labib et al., 1992). However, the issue of energy conservation and in particular the limits thereof, were addressed in only one of these studies (III). In the paragraphs below, the main principles of the bioenergetic conditions of biogas generating systems are presented.

Free energy changes of reactions For an isothermal and isobaric solution, in which some of the solutes are involved in a chemical reaction, the amount of energy that can be used to do useful work equals the change of free energy (AG) of the reaction. This property, G, is defined as H - TS, where H, T and S are the enthalpy, temperature and entropy respectively. Hence, the free energy change of the reacting system can be written as AG = AH - TAS. A reaction with a negative AG value is exergonic (as opposed to endergonic) and occurs spontaneously. On the other hand, for an endergonic reaction to occur, it has to be linked to an exergonic reaction in order to make the total reaction exergonic. A reaction has reached equilibrium when the free energy change equals zero. At this point, the entropy is maximized, which eliminates any driving forces for the reaction to proceed in any direction. The value of AG for a certain reaction can be estimated if the standard free energy change of formation (AGf°) of each reactant and product is known. Measured and estimated standard values have been tabulated for a

13 14 Bioenergetics number of reactions found in biological systems (Thauer etal., 1977). The AG value for a certain condition can be obtained if the stoichiometry of a reaction and the activity of each reactant and product are known. Such a deduction is easily performed according to the law of mass action. The activity can be estimated as equalling the concentration for dilute solutions. At higher concentrations, however, the activity can be estimated with equations in which the ionic strength is acknowledged (Stumm and Morgan, 1970).

Energy conservation All organisms using chemically bound energy are mediating reactions from which the free energy change has to be conserved for different metabolic purposes. The general currency of this conservation is the well known phosphorylation of adenosine-diphosphate (ADP) to adenosine-triphosphate (ATP). In principle, this reaction can be mediated in two ways (Thauer etal., 1977). Either through substrate level phosphorylation (SLP), where the synthesis is enzymatically coupled to a exergonic reaction, or through electron transport phosphorylation (ETP), in which reducing agents (e.g. NADH and are re-oxidized through an electron transport chain so that protons are translocated across the cytoplasmic membrane. The build-up of a proton gradient can be utilized for ATP synthesis in the membrane-bound ATP-ase complex (Figure 3a). The mechanisms of energy conservation in the reactions of interest here, i.e. anaerobic VFA oxidation and methanogenesis, are complex and not yet fully understood (Schink, 1991; Gottschalk, 1989). In general, the utilization of these substrates implies gains (through SLP or ETP) and re-investments (through ATP hydrolysis or reversed electron transport) at different intermediary steps, as in all catabolic paths. Of importance here, though, is the possibility of obtaining net energy yields as low as the energy needed for the translocation of one proton, equivalent to one-third of an ATP molecule.Such net yields have been demonstrated for butyrate and propionate oxidation (Schink and Friedrich, 1994) as well as acetate oxidation and hydrogenotrophic methanogenesis (Schnurer, 1995). The mechanism for butyrate oxidation basically consists of SLP at the terminal reaction and the re-investment of the proton motive force (A^+H) from two protons in the oxidation of one of the intermediates in the primary metabolism (Thauer and Morris, 1984). Similar pathways could tentatively apply for the oxidation of both valerate and caproate, so that each "removed" acetate would at least yield one-third of an ATP. This amount of energy equals the proton motive force and is thereby dependent on the electric potential (A4f) and the proton gradient (ApH). Any analysis of A/i+h would therefore benefit from a relationship between these cytoplasmic membrane gradients and the prevailing biochemical environment Bioenergetics 15

H+ H*" H*' H+ H* Hf

OH" OB' OH- OH- OH- OH*

Figure 3 a) Generation of A/t+H through oxidation of reducing agent SH and subsequent ATP synthesis, b) Depletion of Aji+h in the presence of undissociated weak acids (adapted from Harold, 1986). of the cell. Little information about such relationships is available. However, it is well known that undissociated weak acids equilibrate readily across the membrane, while the ionized forms are impermeable (Menzel and Gottschalk, 1985). Consequently, high concentrations of undissociated acids act as protonophores and should abolish any proton gradient (Figure 3b). It was also suggested that undissociated acids may act as uncouplers of proton gradients (Herrero et al., 1985). A number of carboxylic acid species in sour landfills could therefore be expected to reduce the proton gradients and, ultimately, the proton motive force at high concentrations and low pH values. The largest influence could be expected from acetic acid given the large proportions of this species (III).

Equilibria A later member in a syntrophic system utilizes the product from a former member to mediate a certain reaction. However, the reaction would not be mediated if the conditions (i.e. substrate-product balance) were such that the AG of the reaction was less than what could be conserved. Hence, a build-up of substrate could be expected up to a threshold after which the substrate becomes exploitable. In a similar fashion, the reaction would be mediated with a significant rate if the conditions were such that the AG exceeded the limit of conservation. A reduction of substrate could therefore be expected to reach the point where the substrate becomes non-exploitable if the total conversion rate was sufficiently small. Consequently, conditions or substrate-product 16 Bioenergetics balances will develop in such a way that the mediated reaction will yield the minimum conservable energy precisely. Hence, the reaction is said to "equilibrate at AGmin". This concept should not be mistaken for a thermodynamic equilibrium where AG equals zero. The equilibrium at AGmin or the energy minimum is rather a self regulating state or homeostasis defined by the limitations of microbial bioenergetics. Such relationships between different activities were demonstrated for syntrophic butyrate oxidation at typical sludge digester conditions (Schink and Friedrich, 1994). An indication of equilibria can be obtained by assuming that AGmin has the same valuein all members of a syntrophic cooperation (i.e. the free energy change of ATP hydrolysis is the same). With this assumption, a set of mass- action equations for each equilibrated reaction can be created. From this equation system, a number of unknown properties can be deduced - this is the number of reactions under scrutiny. Indications of equilibria were found in landfill test cells (III). The oxidation of propionate, butyrate, valerate and caproate as well as the hydrogenotrophic methanogenesis were found to equilibrate. The phenomenon gives an explanation for the relative distribution of VFA’s observed in situ. The characteristic pattern is the declining fraction of an acid with the length of its chain and the superimposed deviation with higher ratios of acids with even-numbers of carbon. Moreover, a shift towards higher proportions of propionate was observed in leachate. In a second calculation, measured values of pH, pCH4, [acetate], [propionate ] and [butyrate ] in leachate as well as solid samples were used to deduce AGmin, pm and [HC03‘] (III). In addition, redox potentials were derived from pH and pH2 values. As could be expected, AGmin was found to decrease (become less negative) with increasing concentrations of undissociated acetic acid, indicating the reduction of proton gradients (Figure 4). Values of redox potentials and [HC03‘] were found to correspond with levels that could be expected, although they illustrate the difficulties of sampling. The sudden drop of pCQ2 during sampling leads to the reduction of [HC03‘] and calculated values may therefore exceed measured values. Deduced pH2 values may, on the other hand, be substantially lower than measured values. This is characteristic of the syntrophic cooperation according to which the local partial pressure can be considerably lower than in the free pool accessible to measurements (Conrad et al., 1986). For leachate data, comparisons of AGm;n and the free energy change of acetotrophic methanogenesis indicated syntrophic acetate oxidation, since acetate yielded twice the value of AGmin. In solid samples, though, this ratio was found to range between two and three. If these results are considered, the assumption that the free energy change for the total conversion of different carboxylic acids is equally shared appears to be justified. Bioenergetics 17

Solid sample Leachate

!og[HAc] (M)

Figure 4 The minimum conservable energy vs. the concentration of undissociated acetic acid. For solid samples, calculations were performed for measured pH values (•) as well as measured values reduced with one unit (+). (From paper III) 4. Kinetics

Knowing the rates of biogas generation is essential for dimensioning gas utilization systems. Gas utilization profits and the reduction of the landfill operation time period motivate the enhancement of biogas generation. As was described in Section 2, the generation of biogas is a transient process in which the production rate exhibits a lag time, an increase, a turning point and finally a decrease (Figure 2). For a given total yield, high maximum rates imply narrow curves. Indeed, in lab experiments, over 60% of the total gas yield was generated in less than two months (Ehrig, 1987). It could therefore be incorrect to compare rates if these were taken from different points on the production curve. However, full-scale sites can be treated as conglomerates, consisting of unit batches of different ages (Augenstein and 'Pacey, 1991), so that the production curves are extended over longer time periods. Hence, measurable rates can be expected to last for years.Even in test cells, the rates remain quite constant for several years (El-Fadel et al., 1996b; IV), which enables safe comparisons. Measured rates vary to a great extent between experiments and full-scale sites. Methane production rates in full-scale landfills were found to range from 0.5 to 10 Nm3 per tonne and year with a typical value of 2.5 NmVtonne (as received),yr (Willumsen, 1996). On a dry matter basis, this would roughly correspond to 4 Nm3/t,yr. (The unit "t" is henceforth used to denote dry tonne.) Corresponding values from test cells were found to range from 10 to 40 Nm3/t,yr (Ham and Barlaz, 1989). This range covers the rates of test cells investigated here lying between 10 and 20 Nm3/t,yr (TV). In lab-scale experiments with shredded material and high moisture content, the total methane yields were generated in less than six months (Ehrig, 1987; Barlaz et al, 1989b), making the maximum rates higher than 200 Nm3/t,yr. Even if these differences reflect the gas extraction efficiency to some extent, it is obvious that any attempt to increase the formation rate should be grounded in a basic knowledge of the controlling factors. Due to the complexity of processes, any systematic description is bound to be inadequate in some respect. However, by emphasizing the conditions within the waste material and ignoring ambient factors like climate, cover and depth, the controlling factors can be grouped into macro- and micro-factors. Among the macro-factors considered to be most important are moisture, particle size, density and waste composition (additives included). In general, these factors can influence the material availability and different conductivities (gas, hydraulic and thermal) - they can also be manipulated to some extent. Micro-factors, in contrast, comprise the microbial environment, such as temperature, pH, substrate and inhi bitors.

18 Kinetics 19

The actual rate is determined by several factors. Extensive measurements are necessary in order to obtain thorough descriptions, but evaluations are generally confined to physical models in which all factors but one are fixed (see Section 6). Resultsfrom such measurements have been interpreted as if the rate exhibits a general dependency, which can be estimated with certain functions in the form q CH4 = % f;(X). The conventional approach for recognizing several dependencies is to use the product of each one; q CH4 = q m II f;(Xj). Such models have been used for landfills (Young, 1990), as well as for wood chip storages (Ernstson and Rasmuson, 1988). Problematically, the effect of different factors may interact (Haas, 1994; Clymo, 1995), making the use of products in many cases inadequate. The task of finding a "unified" model, according to which the rate is described by a specific function of all factors, i.e. q CH4 = % f(Xb X2,...), should be regarded as one of the goals of this and analogous research fields. Such a mechanistic approach should, first and foremost, be applicable to factors describing the microbial environment. This far, only one case of unification relevant to landfills has been reported which will be described later. The importance of macro-factors is not entirely self evident. For example, an increased moisture content were reported to stimulate biogas generation (Rees, 1980; Ham etal., 1993; Bogner and Spokas, 1993). Yet, high initial moisture contents (Barlaz et al., 1987) and high hydraulic loads (Farquhar and Rovers, 1973) can lead to the retainment of acidogenic conditions with minor gas production rates. The highest methane production rates obtained in the studies reported in this thesis occurred in the test cell with the lowest moisture content (TV). The effect of macro-factors, such as moisture and waste composition, probably acts indirectly to a large extent. When they are altered the microbial environment will be changed and, consequently, so will the conversion rate. Therefore, the review below emphasizes the micro-factors relevant for landfills and their effects on microbial growth. Nutrients, such as N and P, could also be included, but since the effect of these were expected to be more associated to the methane yields, this issue is instead reviewed in the next section. Moreover, the cell concentration is neglected due to the scarcity of information regarding this issue. The main focus is on methanogenesis. Rate-limiting steps and the role of hydrolysis are discussed in the last paragraph.

Temperature The rates of most reactions increase with temperature. Empirically, many reactions have rate constants that follow the Arrhenius equation. This expression can be justified by collision theory of chemical reactions (Atkins, 1986). The growth rate of cells with enzymatically mediated reactions also increases with raising temperatures, but over certain temperatures this will lead 20 Kinetics to the denaturation of enzymes and reduced cell viability. Procaryotes are specialized to have growth optima within certain temperature ranges and are generally classified as psychrophilic (obligate 15-18°C; facultative 25-30°C), mesophilic (30-45°C) and thermophilic (55-75°C). Methanogens in each of these classes have been identified, although most species are mesophilic (DoE, 1988). No psychrophiles have been found among acetate-utilizing methanogens. Therefore, the growth optima of 30-45°C could be considered best for biogas formation in landfills. Even higher growth rates could be expected in the thermophilic range, but such temperatures are unlikely to occur in temperate climates. The enhancing effect of a temperature increase has been observed in lab-scale experiments. Ehrig (1987) found a three-fold decrease in methane production half-life when he compared results from two containers kept at 20 and 30°C, respectively. In the test cells of this study, with temperatures well within the psychrophilic range, no correlation was found between the temperature and the methane production rate (IV).

P* The effect of pH on microbial activity is primarily related to the state of ionization of particular groups in the enzymes or the substrate or both, furthermore extreme pH values will lead to protein denaturation (Barker, 1992). As with temperature, different methanogenic species exhibit (Efferent optimal pH ranges. Whereas the pH optima of individual methanogenic species is quite narrow, approximately half a pH unit (Zehnder, 1978; DoE, 1988), it ranges between pH values of 6 and 8 for mixed cultures (Zehnder et al, 1982; Clark and Speece, 1971). This could be the main cause of the development of acidogenic conditions, such as those in the leach cell in S. Sunderby, Sweden. In this cell, the pH value was kept below 6. At the same time, no methane could be detected (Lagerkvist, 1995). Such conditions are likely to occur locally in most landfills, which is supported by the positive effect on methane production of buffer addition (Kinman et al., 1987; Barlaz etal, 1987). An observation that to some extent contradicts these findings is the growth, albeit minor, of methanogens under acidogenic conditions (Lagerkvist and Chen, 1993; Barlaz etal., 1989b). Moreover, any significant pH difference within the optimal range of mixed cultures should not in itself be expected to result in any significant differences in methane production rates, since the range is quite "top-hat" shaped. In the next paragraph, though, it will be shown that the pH value has another indirect effect on the methane production rate.

Substrate A large number of enzymatically mediated reactions exhibit a relationship Kinetics 21

between substrate concentration and reaction rate that follows the theoretically derived Michaelis-Menten equation. Under the assumptions that the rate controlling step in a cell is mediated by an enzyme displaying Michaelis-Menten kinetics and that the concentration of the enzyme is proportional to the concentration of the bacteria, the growth of the cell exhibits a similar substrate dependence as the enzyme. This relationship is generally termed Monod kinetics. The equation for this dependency has been applied to the growth of methanogenic bacteria (Gujer and Zehnder, 1983). Furthermore, several reported numerical models of microbial conversion in landfills have applied Monod kinetics (Straub and Lynch, 1982; El-Fadel et al., 1989; Young, 1990; El-Fadel et al., 1996a). However, the use of Monod kinetics is questionable at high acetate concentration levels, which can act in an inhibitory manner. A probable I mechanism for inhibition is that undissociated organic acids can equilibrate ! across the membrane or act as uncouplers of proton gradients (Andrews, 1969; Fukuzaki et al., 1990a). This idea is supported by the bioenergetical studies described in the previous section. Fukuzaki et al. (1990a) measured growth rates of Methanosarcia barken for acetate concentrations of up to 250 mM

qcH4 (NrrWt.yr) 20

Commercial waste 15 i Food-rich wastefractions F=i- 10 Domestic waste

5

0 ' , , , 1 0 50 100 ThOD sc//c/ sa/rp'cs (9/1)

Figure 5 Methane production rates vs. solid sample aqueous THODvfa concentrations, measured during a two yearperiod. (From paper IV; Volumes are adjusted to STP conditions.)

! 22 Kinetics

(equivalent to 16 g COD/1) for pH values between 6 and 7.5. They found a relationship between the growth rate and the concentration of undissociated acetic acid that followed a second order substrate inhibition model (Haldane type). According to measurements at pH 6.6, the optimal growth rate would occur at 80 mM acetate. This is equivalent to a total theoretical oxygen demand concentration of 17 g/1 according to the acetate/VFA ratios presented in paper III, which is consistent with the relationship between methane production rate and organic strength presented in paper IV (Figure 5). Domestic waste test cells with high content of readily biodegradable material were found to exhibit high VFA levels and low methane production rates and conversely, commercial waste (low content) resulted in low VFA levels and high production rates (IV). Hence, the content of readily biodegradable material is a crucial macro-factor which promotes the development of acetic conditions. Similar resultswere obtained from lab-scale experiments in which different waste materials were added in different batches (Wolffson, 1985). The same mechanisms are probably responsible for the enhanced reduction of leachate organic strength by aeration (Stegmann, 1995) and landfilling in thin layers (Ehrig, 1983).

Inhibitors Many substances, nutrients included, act as inhibitors when at sufficiently high concentrations. Different mechanisms exist, of which one is the inhi bitory effect of undissociated acetic acid on methanogenesis described above. Similar characteristics were demonstrated for undissociated propionic acid (Fukuzaki etal. 1990b). The inhibitory effects of salts are attributed to the cations (like alkali & alkaline-earth cations) rather than the anions (McCarty, 1964). Ammonium nitrogen levels above 1.7 g/1 may inhibit the biogas production (see review by Schnurer, 1995), probably due to the depletion of intracellular potassium content (Sprott et al., 1984). Ammonium levels in leachate from test cells suggest that this factor could have some influence on differences in methane production rates (TV). Finally, acetotrophic methanogens were found to be inhi bited at partial pressures of carbon dioxide above 0.2 bar (Hansson and Molin, 1981). This can be explained by the thermodynamic conditions. Increased activitiy of the product would make the reaction less beneficial, unless the acetate activity increased accordingly. However, only a minor effect was observed for methane pressures up to 1 bar (Hansson and Molin, 1981). Even so, results from a computer controlled landfill gas extraction system show that the methane production rate can be increased by intensive extraction (Birkeland, 1996). Kinetics 23 Rate-limiting steps It is difficult to define the real rate-limiting reaction of the overall process of biopolymer conversion (Kaspar and Wuhrmann, 1978). The accumulation of dissolved organic material during the initial stage indicates that the methanogenesis is less extensive than the primary fermentation and hydrolysis. After reaching maximum levels of dissolved organics, the decreasing concentrations indicate that methanogenesis surpasses hydrolysis. At this stage, hydrolysis would appear to be rate-limiting. Indeed, this view is widely accepted (Eastman and Ferguson, 1981; Pavlostathis et al., 1988; Tong et al., 1990). In mathematical models, the rate of hydrolysis has conventionally been regarded as a first order reaction: dC/dt = -kC, where C is the remaining amount of the substrate and k is a time constant (El-Fadel etal., 1989; Young, 1990). Measured methane generation rates in full-scale landfills indicate rather a zero order kinetic model: dC/dt = -k (Barlaz et al, 1990). However, the rate differences between hydrolysis and methanogenesis appears to be minor, because the depletion of VFA extends over severalyears (TV). Thus, observed differences in methane production rates at different VFA levels (TV) are also valid for the rate of hydrolysis, which appears to be influenced by the concentration of VFA and, ultimately, by the rate of methanogenesis. This is consistent with results from lab-experiments (Barlaz et al., 1989b) suggesting feedback inhibition of hydrolysis owing to acid accumulation. 5. Capacities

As with rates, knowing the amount of biogas that will evolve after landfilling is essential to efficientlydimension treatment systems (Augenstein and Pacey, 1991). Even if the site is too small to sustain profitable yields, the gas has to be vented since the generation always entails the risk of lateral migrations and possible explosions. A third aspect of gas potentials is the global impact. Not only are landfills regarded as a source of greenhouse gases, but they were recently treated as carbon sinks (Bogner and Spokas, 1995). The methane yield is always smaller than the corresponding potential (Figure 6), if the potential is the amount that can be expected (using more or less sophisticated means) from the waste composition and the yield is the amount that is actually emitted. Some parts of the organic compounds, such as cellulose, that usually are subjected to microbial attacks, appear to be preserved (Barlaz et al., 1989a). This can possibly be due to the enclosure of recalcitrant material, such as lignin, or through nutritional limitation. Whereas the former explanation is difficult to investigate, but is supported by long-term batch fermentations of different lignocellulosic materials (Tong et al, 1990), the latter have been addressed several times. In the paragraphs below, estimated potentials and measured yields are presented and the issue of possible nutritional limitation is reviewed.

Potentials In the simplest forms of theoretical estimates, all carbon is converted to methane or carbon dioxide with the proportions determined by the mean oxidation number of the degraded carbon. For instance, the total mineralization of pure cellulose (C6 H10O5)n will correspond to a methane potential of 415 NmVt (Cossu etal., 1996). Calculations based on elementary data from "typical" US municipal refuse resulted in a potential of 350 NmVt (Ham and Barlaz, 1989). More sophisticated estimates can be obtained by acknowledging the biodegradabilities of different organic compounds. A direct method is to quantify the content of cellulose, hemicellulose, proteins and sugars. By calculating the potential of each of these compounds, with the Buswell equation or any of its derivatives, the total value of the waste mixture can be readily obtained. Barlaz et al. (1989a) used potentials for carbohydrates (CnH2nOn) and proteins (CpH5ON0.86) °f 373 and 517 Nm3/1, respectively.The reason for using a stoichiometry for hydrolyzed polymers, instead of the cellulose potential above, is not known. Certainly, the content of lignin and synthetic organics is significant, but these do not contribute to the biogas generation thanks to their recalcitrant properties.

24 Capacities 25

CH4 (NrriVdiy t) Biogas (Nnf/wet 11 400 ■q 500

- 400 300

MouriianView toti - 300 Spillepeng 200 Wsccnsin

mass losses - 200

* Uhlv.of .... Test cells • Wisconsin Mountain 100 View TediUnrv.of t 1 Spillepeng 100 Braunschweig j Brogborough Lab experiments]

Figure 6 Potentials and yields of methane (dry mass) and biogas (wet mass, assumed 55% CH, content and 30% moisture content). Data from Wisconsin by Barlaz et al. (1989a); Braunschweig by Ehrig (1987), Mountain View by El-Fadel et al. (1996b), Brogborough by Campbell et al. (1995) and Spillepeng from paper V.

A second method is to use the carbon content of each waste fraction and to introduce a biodegradability factor (mass degradable C/mass total C). Values of this factor were reported to be 0.8 for food waste, 0.7 for yard waste, 0.5 for mixed paper and wood and 0.2 for textiles (Cossu et al., 1996). Biodegradability factors were used to estimate the potential of the test cells in this study, which were found to range between 150 and 200 NmVt (V). Calculated potentials of the test cells at Mountain View, California, came to 230 NmVt (El-Fadel et al., 1996b). Similar values can be obtained from the content of cellulose and hemicellulose in the lab-scale experiment by Barlaz et al. (1989a). However, a significant portion of these compounds remained after that the methane production had ceased: 29% and 23% for cellulose and hemicellulose, respectively. By assuming a loss of 10% for biomass production, this would correspond to a potential of 153 NmVt (Barlaz et al., 1989a).

Yields The total actual methane yields after approximately 300 days in lab-scale 26 Capacities experiments with municipal solid waste were found to be around 80-130 NmVt (Ehrig, 1987; Barlaz et al. 1989a). After little more than four years of extraction, the test cells at Mountain View had yielded approximately 40 to 90 NmVt (El-Fadel et al., 1996b). For a period of the same duration, the test cells at Brogborough, UK, had yielded 20 to 40 NmVt (Campbell etal., 1995), assuming a moisture content of 30%. Corresponding values after five years in Spillepeng were found to be 40-70 NmVt (V), although these values can be expected to increase over the years to come. As a result from the rate differences presented in Figure 5, the methane yields from the commercial waste test cells exceeded those containing domestic waste. The yields from full- scale sites were found to range from 0.3 to 60 NmVt (Barlaz et al., 1990). Possible explanations for the apparent decline in yield with increasing scale include: aerobic degradation during the filling; a low content and heterogenous distribution of moisture; and diffuse emissions. Naturally, the poor kinetics of conditions resulting from conventional filling techniques implies that the generation is extended over long time periods, making quantified yields inferior to lab-scale results.

Nutrition The way in which nutrients can possibly limit landfill conversion has been addressed in a number of studies (Rees, 1980; Kinman etal., 1987; Christensen and Kjeldsen, 1989; Barlaz et al., 1989c; V). It was observed that ammonium and phosphate were present at concentrations sufficient for methanogenesis (Rees, 1980; Barlaz et al. 1989c). However, these levels, which are typical for the initial phases of landfill development, may reflect the distribution of nutrients to materials that are converted rapidly. "When pools of slow cellulosic materials, such as paper, are to be subjected to decay, the inherent content of nutrients may be too low to enable growth. Among all growth factors, the one most likely to restrict the conversion processes is phosphorous (Christensen and Kjeldsen, 1989). Nutritional requirements were estimated from relationships between amounts of converted organics and nutrient contents in the waste material and biomass pools (Rees, 1980; V). Essential for such estimates is the information on the nutrient fraction in the waste material, the corresponding value in the biomass, the growth yield coefficient (Y) of the mediating organisms, and the turnover of dead biomass. Whereas the nutrient content of biomass is fairly constant (Haug, 1993), the concentration in the waste material has to be quantified. The most uncertain factors are the growth yield and the turnover. Rees (1980) used a Y value of 0.05 gVSS/gCOD. Compiled results from in vitro studies ranged between 0.05-0.07 for methanogens and 0.1-0.5 for acidogens (El-Fadel et al., 1996a). The upper limit of the ranges can be attributed to conditions with high growth rates when only a minor part of Capacities 27 conserved energy is used for maintenance (Thauer et al., 1977). Values in the lower part of the ranges should therefore be used for conditions in landfills. In addition, growth yields of the different groups in the food-chain should be added (McCarty, 1971). Rees (1980) concluded that neither N nor P were limiting, even without considering the turnover of biomass. The Spillepeng results strongly indicated the turnover in the test cells with commercial waste (V), since the assimilation of P clearly exceeded the amount in the waste. The P content in domestic waste corresponded fairly well with the requirements for obtaining gas yields with the same magnitude as those in lab-scale experiments. Some empirical evidence of nutrient limitation can be found in the literature, although more information is needed to finally resolve the issue. Experiments with nutrient addition in landfill simulators were reported to result in increased rates of methane formation (Kinman etal., 1987). Nutrient media added to landfill samples contained in serum bottles and then incubated for 700 days showed increased means of methane yields (27 Nm3/t) in comparison to samples with only moisture addition (18 NmVt) (Bogner and Spokas, 1993). However, the ranges showed only a limited difference. 6. Test cell methodology

Attempts to improve the descriptions of conversion processes in landfills and to develop the technology have mainly made use of physical models. This is natural, considering the practicality of repeating experiments and examining a whole sequence of setups in parallel. The objectives of full-scale sites may also be in direct conflict with the goals of an investigation in which filling takes place over an extended time period. Similary, gas collection systems are designed and operated with incentives other than those required for quantification of the gas generation. The test cell is thus a model by which the designer tries to mimic the full-scale case, but with an almost total control of content and emissions. The main benefit over smaller experiments is, of course, the similarity with full- scale sites, which makes the results more applicable. The disadvantage is the difficulty in maintaining full control over the system and the long time periods required for relevant evaluation, factors which lead to high costs. Aerobic conversion can be minimized by increasing the rate of filling. This improves the prospects of keeping track of the carbon fluxes, but will also promote the development of unusually high VFA levels. Moreover, the test cell tends to transform as a complete unit, rather than as an intermix of biochemical conditions, as is the case in a full-scale landfill (Ham, 1996). Thus, results should be applied with caution, whether the test cell designer and operator succeed in keeping full control or not. The main objective of test cell experiments is to investigate the effects of landfilling techniques on biogas production. The first requirements are to enable the documentation of the construction and to ensure that the generated gas can be collected and quantified. Apart from this, measurements should be emphasized to provide a foundation that can support interpretations and theories regarding the most relevant processes, i.e. microbial conversion, diffuse emissions and interfering processes. These fundamental goals of information retrieval are listed below together with the properties for quantification in test cell experiments (Table 1). It is beyond the scope of the thesis to try to thoroughly review all the ideal methodologies. Descriptions of equipment and methods of analysis are presented in papers (I-V). Instead, the discussion below emphasizes the main framework of test cell investigation by addressing issues such as construction characteristics as well as types, purposes and extents of sampling.

Fixation of non-studied variables If the effect of one variable, such as material or density, is to be investigated in a test cell experiment so that the value of the variable is gradually changed

28 Test cell methodology 29

Table 1 Construction and process characteristics needed for evaluation of test cells.

Technique of Construction (time, volume, surfaces, angles) landfilling Bottom construction Material (amount, moisture, composition) Manipulations and additives Effects of filling (composting, wetting)

Gas extraction Extraction system construction Optimal operation Fluxes and methane content

Degradation Biochemical conditions states Remaining potentials and moisture content

Emissions Diffuse gas emissions and methane oxidation Leachate production and quality

Interferences Precipitation Water locks Settlements in the array of objects, it is of great importance that other variables are kept constant. It was discussed whether the observed differences in performance of the Spillepeng test cells were a result of factors other than those emphasized - among which the most important were the rate of filling and density (TV). Both could possibly influence the extent of the composting. In addition, large density differences between the cells could create discrepancies in gas extraction conditions. Still, the results strongly indicated that the gas generation was primarily a function of the VFA levels and, ultimately, the refuse content. Influences of other factors were thus subordinate. It is of great importance to have a clear idea of the conditions at the site where the experiments are to be performed. Information of access roads, available machinery and frequency of entering waste material must all be acknowledged during the planning of construction and time schedules.

Potentials Ideally, used waste materials should be characterized concurrently with filling through sampling, sorting and physico-chemical analyses. Such a procedure requires that a surface, preferably roofed, and labor can be provided. The objective of characterization is to quantify the moisture content and the potentials of methane generation and rapid VFA formation (i.e. cellulose, hemicellulose etc.). No material analysis was performed at the onset of the studies presented here. Instead, moisture contents and potentials were estimated on the basis of entrance registrations, visual inspection and sortings performed on similar waste streams in the region (II, V). Balance calculations are strengthened if moisture content changes and 30 Test cell methodology potentials can be monitored. The most dramatic change occurs during the initial phase when oxygen is depleted with the secondary effect that the content of readily biodegradable material and water is reduced. The moisture content can also be affected by precipitation and evaporation during construction, but the effects are probably inferior at high build-up rates. To quantify these effects, thorough sampling programs are required after the oxygen depletion, i.e. almost directly after final covering. Moisture contents from measurements made at this time were compared with moisture contents in waste materials entering the test cells and were estimated to correspond to a mass reduction of 3-10% (II). Monitoring potential changes under anaerobic conditions can be performed with long time intervening periods. Here, samples were collected annually with an auger. This frequency was deemed unnecessarily high considering the conversion rates observed, and could safely have been reduced to once every two years. On the other hand, the sampling technique probably suffers from an over representation of fines (V). Future studies would therefore benefit by increasing the precision by taking larger samples, for examples through use of a bucket auger.

Biochemical conditions The growth of microbes is generally highly related to their environment. It is, therefore, advantageous if interpretations of biogas generation rates in test cells can be related to the biochemical condition. However, there are no obvious guidelines as to the characterizations which should be included, but, as was shown in Section 3 and 4, the concentration of undissociated carboxylic acid and temperature appear to be two of the key factors. Therefore, VFA and pH should be monitored. In order to bring down the costs, VFA could, perhaps, be estimated from measured COD concentrations. Two different sampling techniques have been used at Spillepeng: solid sample extracts and leachate samples. A third technique was employed in the test cells at the Filborna landfills outside of Helsingborg, Sweden. These were equipped with lysimeters (Meijer, 1992), a technique that, in contrast to the traditional landfill literature, is defined as a well drained container filled with refuse and placed inside the test cell. The disadvantage of solid samples and lysimeters is that these are localized and, therefore, a whole suite of sampling points are needed. Leachate can, on the other hand, be regarded as a unified sample which represents all the internal conditions. The problem with leachate is that it will be affected on its way out of the test cell (I), unless a coarse drainage material is employed. With high organic strength leachate, this will inevitably lead to incrustation and possible obstructions downstream of the test cells. These facts give preference to the local methods. In order to minimize labor (cleaning and changing tubes and pipes), the drainage layer Test cell methodology 31 should be protected with a compost layer. In choosing between solid samples and lysimeters, the lysimeter is preferred due to the high labor intensity and costs associated with solid sampling. It is recommended that lysimeters are employed for monitoring biochemical conditions. The use of solid samples should be confined to measurements of moisture contents and remaining potentials. Finally, leachates should be used exclusively for mass balance calculations and treatment issues. The optimal number of lysimeters in each test cell is dependent on the expected heterogeneity and the needed resolution. Simple hypothesis testing with a Student t-test can indicate whether the means of two test cells differs significantly. The coefficient of variation (c.v.)of internal COD concentrations in solid samples was found to be approximately 40% (TV). A reasonable resolution to show differences between two test cells, could be one having a COD concentration of less than 70% of the value in the other. This would imply that a 20% and a 10% level of significance could be obtained with 8 and 12 lysimeters, respectively.This estimate can be regarded as conservative, since the c.v. values from lysimeters are expected to fall below the value for solid samples, due to the larger surface area of lysimeters. The sampling frequency should, in a similar way, be based on the expected temporal variation of the conditions. The dominant variation in temperate climates is the one that arises from the winter infiltration and, therefore, has a periodicity of one year. Here, the infiltration was overshadowed by the trends originating from the conversion in the drainage layer (I), but the biogas ratio (methane/carbon dioxide) did exhibit pronounced seasonal changes (Akesson and Nilsson, 1993). In order to cover a periodicity of one year, six samples a year were considered appropriate.

Gas extraction The extraction system used in this experiment was characterized by high extraction rates, noticeable air intrusion predominating at the upper extraction levels and minor diffuse emissions. This implies that the production rates deviatefrom the quantified extraction rates, dueto the maintenance of minor aerobic zones and the loss of some carbon fluxes. Nevertheless, the use of identical systems in the array of objects would still produce results which could be used for comparisons. The need of identical systems strengthens the importance of fixating non-studied variables as discussed above. Since the experience of construction and operation of extraction systems reported has been confined to one type of system, the advice given below is generalized. A more extensive investigation would be needed in order to give more precise information about the extraction conditions. Extraction systems should be designed so that the surfaces of the systems are isolated from the open air. The central gas wells in Spillepeng’s test cells 32 Test cell methodology intersected the cover material. This in itself would perhaps be of minor concern if the wells had not been constructed during filling, which, of course, is impossible with the horizontal systems. The difficulties in obt aining even and thorough compaction around the wells probably result in the noticeable air intrusion. However, these obstacles could be reduced with vertical wells inserted after filling. Secondly, a large and well distributed extraction surface should promote the efficiency of the gas collection. Indeed, the ability of collecting those production rates observed in this study can be attributed to the large surfaces of the systems. However, with the star-shaped layout of the gas drains, the surfaces are condensed in the vicinity of the central gas well, increasing the load in the most exposed part - possibly contributing to the air intrusion. Finally, systems should be designed so that condensate gatherings can be avoided. This could be achieved by inclining the gas tubes and installing dewatering equipment. Otherwise, much effort may be put into extracting condensate by extensive pumping. The main rule for operating the system should be to maximize the extraction whilst maintaining an acceptable gas quality. Naturally, the acceptable limit can be set higher with higher qualities of the constructions. Attention should also be given to the risk of artifacts caused by differing extraction conditions between two cells, since the gas generation was found to increase by the intensity of the load (Birkeland, 1996). In this case, it was decided to set the limit of acceptable methane contents to 50% and 30% for the lower and upper levels, respectively. Thus, a ratio between the gas fluxes from the lower and upper levels of 60/40 gives a maximum air content in extracted gas of approximately 20%. Still, as long as leaks are restricted to minor zones, the effect on the overall anaerobic conversion is negligible. Therefore, the main disadvantage of air intrusion is the need to reduce the load and, consequently, the increased diffuse emissions. Sporadic measurements of diffuse emissions have shown that less than 10% of the carbon fluxes are lost to the atmosphere (mostly in the form of COj, although higher ratios have been observed (Lagerkvist etal, 1996). Of course,the risk of fires should also be taken into account. The sampling frequency should be selected according to requirements. Measurements which aim to quantify the methane flux should be performed weekly so that unforeseen events, such as water locks and changing downstream conditions, can be taken care of. If variations stemming from changing air pressures is to be covered, a periodicity of a few hours has to be applied (Young, 1992). Biogas ratios are more related to the biochemical conditions and should therefore be covered six times a year. 7. Concluding remarks

The results presented above show that the advantages of a reduced content of readily biodegradable material, achieved by guidance or pretreatment, encompass several aspects of the performance: high methane production rates; leachates with low organic strength; reduced extent of incrustation and; after five years of gas extraction, increased methane yields. Optimizing methane production can thus be considered equivalent to optimizing the reduction of readily biodegradable material. Even if the domestic waste test cells eventually attain the same gas yields as those containing commercial waste, the results indicate that the methane production half-life of the latter test cells would still be lower. If this is not the case, these findings are inconsistent with results from lab-scale experiments, which show that initially inhibited setups actually yielded more than the zero test (Wolffson, 1985). The discrepancy could perhaps be explained by the thermal differences between these two experiments. While the lab-experiment was performed at a constant temperature of 30°C, the Spillepeng test cells were left to acclimatize to ambient lower temperature conditions. Moreover, the performance of a test cell with several thousands tonnes of material was shown to have a theoretical foundation. Equilibria, stemming from the limit for energy conservation, were demonstrated, enabling the establishment of relationships between key properties. In addition, substrate inhibition of acetotrophic methanogenesis was found to be one of the most important factor governing biogas generation. The depletion of proton motive forces with high levels of undissociated acetic acid, demonstrated with the thermodynamic model, strengthens the notion of substrate inhibition. These findings indicate that the shift from acidogenic to methanogenic condition is not a discrete one, but rather a continuous transition. Finally, the results indicate that methane yields are dependent on the turnover of biomass, if not the phosphorous content of the refuse. As is often the case, the findings pose new questions. In order to address the reduction of the content of readily biodegradable material, greater emphasis needs to be placed on imitating the filling procedures of those landfills the models are intended to simulate. However, perhaps the time is ripe for a general mathematical model of these processes to be created. New directives have recently been suggested. Some of these directives are intended to promote and steer the waste management towards pretreatment (biological or thermal) of organic refuse before final storage. In such a situation, the technique of biocells would be permitted for a limited duration only, after which the cells would be excavated. Securing a market of materials from such ventures necessitate a high quality of the incoming refuse

33 34 Concluding remarks fractions. However, this confines the application to pure source separated fractions with high contents of readily biodegradable materials, which were found to reduce the performance of biocells. The technique would therefore have to be extended to include new elements, such as aeration or neutralization, leading to the reduction of methane yields and increasing emissions of the used additives, respectively. Moreover, such a system are bound to require extensive monitoring and operation. Finally, discrepancies between methane yields and theoretical potentials remain to be investigated. Explanatory models such as lignin enclosure, nutrient limitation, heterogeneities and a combination of these, have to be further evaluated. Even if nutrients are limiting, it is uncertain whether measures should be taken to improve conditions, not only due to economic concerns, but also as landfills constitute one of the few resistances against the conventional carbon fluxes of our culture. 8. Acknowledgements

I would like to express my gratitude to:

Lars Bengtsson, my supervisor, for his support and patience.

Peter Nilsson, for inviting me into the test cell project and for encouragement and a fruitful cooperation.

All my colleagues at the department, for creating a stimulating academic environment. The contributions from those who have been willing to discuss, criticize, encourage and help me in different ways are also deeply appreciated and not forgotten.

My friends and family, especially Anna, for their support in this work.

This study was mainly financed through a doctoral grant from Lund University and through funding from the Swedish National Board for Industrial and Technical Development (NUTEK), Southwestern Scania Solid Waste Company (SYSAV Co). The support of the Swedish Waste Research Council (APR) is also appreciated.

The construction of the test cells was mainly financed by SYSAV Co. Solid samples were obtained by GeoSyd. Analyses were performed by Anox AB/ Cenox AB (solid sample: VFA, pH and DS), AgroLab AB (solid sample: N, P and K), Department of Microbiology, Swedish University of Agricultural Sciences/Ultuna (leachate: VFA), Department of Plant Ecology, Lund University (leachate: IC), and Water and Environmental Engineering, Lund University (all other leachate analyses).

35 9. References

Akesson, M. and Nilsson, P. (1993) Emissioner fran testceller med olika avfalls- sammansattning. (Emissions from test cells with different refuse compositions) In Symposium om svensk upplagsforskning1993 (Edited by Lagerkvist, A.), RVF rapport 1993:8, Malmo, Sweden, 93-106 (in Swedish) Andrews, J.F. (1969) Dynamic model of the anaerobic digestion process. J. Sanit. Eng. Div. ASCE 95, 95-116. Atkins, P.W. (1986) Physical chemistry. Oxford University Press, Oxford, UK. Augenstein, D. and Pacey, J. (1991) Modeling landfill methane generation. In Proceedings SARDINIA 91, Third International Landfill Symposium, Vol. I, CISA, Cagliari, Italy, 115-148. Barker, R.D.J. (1992) Enzymes. In Principles of Cell Energetics (Edited by Weston, G.D.), Butterworth-Heinemann Ltd., Oxford, UK, 90-129. Barlaz, M.A., Ham, R.K. and Schaefer, D.M. (1989a) Mass-balance analysis of anaerobically decomposed refuse. J. Environ. Eng. ASCE 115, 1088-1102. Barlaz, M.A., Ham, R.K. and Schaefer, D.M. (1990) Methane production from municipal refuse: A review of enhancement techniques and microbial dynamics. CRC Crit. Rev. Environ. Contr. 19, 557-584. Barlaz, M.A., Milke, M.W. and Ham, R.K. (1987) Gas production parameters in sanitary landfill simulators. Waste Manage. Res. 5, 27-39. Barlaz, M.A., Schaefer, D.M. and Ham, R.K. (1989b) Bacterial population development and chemical characteristics of refuse decomposition in a simulated sanitary landfill. Appl. Environ. Microbiol. 55, 55-65. Barlaz, M.A., Schaefer, D.M. and Ham, R.K. (1989c) Inhibition of methane formation from municipal refuse in laboratory scale lysimeters. Appl. Biochem. Biotechnol. 20, 181-205. Birkeland, K.H. (1996) Automated control of LEG abstraction. In Nordic Landfill Energy Technology Network Seminar, Pite Havsbad, 17/1 1996, Landfill gas in the Nordic countries, Report 1996:01, The Landfill Group, Lulea University of Technology, Lulea, Sweden, 51-66. Bogner, J. and Spokas, K. (1993) Landfill CH4: Rates, fates and role in global carbon cycle. Chemosphere 26, 369-386. Bogner, J. and Spokas, K. (1995) Carbon storage in landfills. In Soils and Global Change (Edited by Lai, R. et all), Advances in Soil Science Series, CRC Lewis Publishers, Boca Raton, FI, US, 67-80. Brune, M., Ramke, H.G., Collins, H.J. and Hanert, H.H. (1991) Incrustation processes in drainage systems of sanitary landfills. In Proceedings SARDINIA 91, Third International LandfillSymposium, Vol. I, CISA, Cagliari, Italy, 999-1035. Campbell, D., Caine, M., Meadows, M. and Knox, K. (1995) Enhanced landfill gas production at large-scale test cells. In Proceedings SARDINIA 95, Fifth International Landfill Symposium, Vol. I, CISA, Cagliari, Italy, 593-601. (errata distributed at the presentation) Christensen, T.H. and Kjeldsen, P. (1989) Basic biochemical processes in landfills. In Sanitary Landfilling: Process, Technology and Environmental Impact (Edited by Christensen, T.H., Cossu, R. and Stegmann, R.), Academic Press, London, UK, 29- 49.

36 References 37

Clark, R.H. and Speece, R.E. (1971) The pH tolerance of anaerobic digestion. In Proceedings, 5th International Conference on Water Pollution Research 11-27,1-4. (cited in El-Fadel et al., 1996a) Clymo, R.S. (1995) Nutrients and limiting factors. Hydrobiologia 315, 15-24. Conrad, R., Schink, B. and Phelps, T.J. (1986) Thermodynamics of ^-consuming and H2-producing metabolic reactions in diverse methanogenic environments under in situ conditions. FEMS Microbiol. Ecol. 38, 353-360. Cossu, R., Andreottola, G. and Muntoni, A. (1996) Modelling landfill gas production. In Landfilling of waste; Biogas (Edited by Christensen, T.H., Cossu. R, and Stegmann, R.), E &FN SPON, London, UK, 237-268. DoE (Department of Energy) (1988) A basic study of landfill microbiology and biochemistry. Contractor Report, ETSU B 1159, UK. Eastman, J.A. and Ferguson, J.F. (1981) Solubilization of particulate organic carbon during the acid phase of anaerobic digestion. /. Water Pollut. Control Fed. 53, 352- 366. Ehrig, H.-J. (1983) Quality and quantity of sanitary landfill leachate. Waste Manage. Res. 1, 53-68. Ehrig, H.-J. (1987) Anaerobic degradation of municipal solid waste - Laboratory-scale tests. In Global Bioconversion, Vol II (Edited by Wise D.L.), CRC Press, Boca Raton, Fl, US, 121-152. Ehrig, H.-J. (1989) Water and element balances of landfills. In The Landfill: Reactor and Final Storage, Lecture Notes in Earth Sciences (Edited by Baccini, P.), Springer-Verlag, Berlin, Germany, 83-115. Ehrig, H.-J. (1991) Prediction of gas production from laboratory scale tests. In Proceedings SARDINIA 91, Third International Landfill Symposium, Vol. I, CISA, Cagliari, Italy, 87-114. Ehrlich, H.L. (1981) Geomicrobiology. Marcel Dekker Inc., New York, US. El-Fadel, M., Findikakis, A.N. and Leckie, J.O. (1989) A numerical model for methane production in managed sanitary landfills. Waste Manage. Res. 7, 31-42. El-Fadel, M., Findikakis, A.N. and Leckie, J.O. (1996a) Numerical modelling of generation and transport of gas and heat in landfills: I. Model formulation. Waste Manage Res. 14, 483-504. El-Fadel, M., Findikakis, A.N. and Leckie, J.O. (1996b) Numerical modelling of generation and transport of gas and heat in sanitary landfills: II. Model application. Waste Manage Res. 14, 537-551. Ernstson, M.-L. and Rasmuson, A. (1988) Mathematical modelling of transport processes and degradation reactions in piles of forest fuel material. In Production, storage and utilization of wood fuels, Proceedings of IEA/BE Conference Task III/Activity 6 and 7, Vol II, Uppsala, Sweden, ISSN 0282-2377, 1-17. Farquhar, G.J. and Rovers, F.A. (1973) Gas production during refuse decomposition. Water Air Soil Pollut. 2, 483-495. Ferguson, C.C. (1993) A hydraulic model for estimating specific surface area in landfill. WasteManage Res. 11, 227-248. Fukuzaki, S., Nishio, N. and Nagai, S. (1990a) Kinetics of the methanogenic fermentation of acetate. Appl. Environ. Microbiol. 56, 3158-3163. Fukuzaki, S., Nishio, N., Shobayashi, M. and Nagai, S. (1990b) Inhibition of the fermentation of propionate to methane by hydrogen, acetate, and propionate. Appl. Environ. Microbiol. 56, 719-723. 38 References

Gottschalk, G. (1989) Bioenergetics of methanogenic and acetogenic bacteria. In Autotrophic Bacteria (Edited by Schlegel, H.G. and Bowien, B.), Science Tech Publishers, Madison, WI, US, 383-396. Gujer, W. and Zehnder, A.J.B. (1983) Conversion processes in anaerobic digestion. Water Sci. Technol 15, 127-167. Haas, C.N. (1994) Unified kinetic treatment for growth on dual nutrients. Biotechnol. Bioeng. 44, 154-164. Ham, R.K. (1996) Field testing for evaluation of landfill gas yields. In Landfilling of waste: Biogas (Edited by Christensen, T.H., Cossu, R. and Stegmann, R.), E & FN SPON, London, UK, 297-315. Ham, R.K. and Barlaz, M.A. (1989) Measurement and prediction of landfill gas quality and quantity. In Sanitary Landfilling: Process, Technology and Environmental Impact (Edited by Christensen, T.H., Cossu, R. and Stegmann, R.), Academic Press, London, 155-166. Ham, R.K., Fritschel, P.R. and Robbins Norman, M. (1993) Refuse decomposition at a large landfill. In Proceedings SARDINIA 93, Fourth International Landfill Symposium, Vol. II, CISA, Cagliari, Italy, 1045-1054. Hansson, G. and Molin, N. (1981) End product inhibition in methane fermentations: Effects of carbon dioxide and methane on methanogenic bacteria utilizing acetate. Eur. J. Appl. Microbiol. Biotechnol. 13, 236-241. Harold, F.M. (1986) The vital force: A study of bioenergetics. W.H. Freeman Sc Co, New York, US. Haug, R.T. (1993) The practical handbook of compost engineering. Lewis Publ., Boca Raton, FI, US. Herrero, A.A., Gomez, R.F., Snedecor, B., Tolman, C.J. and Roberts, M.F. (1985) Growth inhibition of Clostridium thermocellum by carboxylic acids: A mechanism based on uncoupling by weak acids. Appl. Microbiol. Biotechnol. 22, 53-62. Hickey, R.F. and Switzenbaum, M.S. (1991) Thermodynamics of volatile fatty acid accumulation in anaerobic digesters subject to increases in hydraulic and organic loading. Res. J. Water Pollut. Control Fed. 63, 141-144. Hoeks, J. (1983) Significance of biogas production in waste tips. Waste Manage. Res. 1, 323-335. (cited in Cossu et al., 1996) Jeris, J.S. and McCarty, P.L. (1965) The biochemistry of methane fermentation using CM tracers. J. Water Pollut. Control Fed. 37, 178-192. Jones, K.L., Rees, J.F. and Grainger, J.M. (1983) Methane generation and microbial activity in a domestic refuse landfill site. Eur. J. Appl. Microbiol. Biotechnol. 18, 242- 245. Kaspar, H.F. and Wuhrmann, K. (1978) Kinetic parameters and relative turnovers of some important catabolic reactions in digesting sludge. Appl. Environ. Microbiol. 36, 1-7. Kinman, R.N., Nutini, D.L., Walsh, J.J., Vogt, W.G., Stamm, J. and Rickabaugh, J. (1987) Gas enhancement techniques in landfill simulators. Waste Manage. Res. 5,13- 25. Labib, F., Ferguson, J.F., Benjamin, M.M., Merigh, M. and Ricker, N.L. (1992) Anaerobic butyrate degradation in a fluidized-bed reactor: Effects of increased concentrations of H2 and acetate. Environ. Sci. Technol. 26, 369-376. References 39

Lagerkvist, A. (1995) Tvastegs anaerob nedbrytning - en alternativ behandlingsteknik for upplagt hushallsavfall (Two-step anaerobic degradation - an alternative treatment technique for landfilled MSW). PhD thesis, Inst, for Samhallsbyggnadsteknik, 1995:177D, Tekniska Hogskolan i Lulea, Lulea, Sweden, ISSN 0348-8373. (in Swedish) Lagerkvist, A. and Chen, H. (1993) Control of two step anaerobic degradation of municipal solid waste (MSW) by enzyme addition. Water Sci. Technol. 27, 47-56. Lagerkvist, A., Forsberg, T. and Maurice, C. (1996) Kolbalanser vid testceller (Carbon balances of landfill test cells). In Symposium om svensk upplagsforskning 1996 (Edited by Lagerkvist, A.), RVF rapport 1996:01, Malmo, Sweden, 119-145. (in Swedish) McCarty, P.L. (1964) Anaerobic waste treatment fundamentals. Part three. Public Works 11, 91-94. McCarty, P.L. (1971) Energetics and kinetics of anaerobic treatment. Adv. Chem. Ser. ACS 105, 91-107. Meijer, J.-E. (1992) Design of cells for landfill gas generation. In ISWA 6th International Solid Waste Congress and Exhibition, June 14-19, Madrid, Spain. Menzel, U. and Gottschalk, G. (1985) The internal pH of Acetobacterium wieringae and Acetobacter aceti during growth and production of acetic acid. Arch. Microbiol. 143, 47-51. Mormile, M.R., Gurijala, K.R., Robinson, J.A., Mclnerney, M.J. and Suflita, J.M. (1996) The importance of hydrogen in landfill fermentations. Appl. Environ. Microbiol. 62, 1583-1588. Nilsson, P. and Edner, S. (1991) Test cell study of methane production by controlled waste deposition. In Proceedings SARDINIA 91, Third International Landfill Symposium, Vol. I, CISA, Cagliari, Italy, 149-159. Pavlostathis, S.G., Miller, T.L. and Wolin, M.J. (1988) Kinetics of insoluble cellulose fermentation by continuous cultures of Ruminococcus albus. Appl. Environ. Microbiol. 54, 2660-2663. Petersen, S.P. and Ahring, B.K. (1991) Acetate oxidation in a thermophilic anaerobic sewage-sludge digester: the importance of non-aceticlastic methanogenesis from acetate. FEMS Microbiol. Ecol. 86, 149-158. Rees, J.F. (1980) The fate of carbon compounds in the landfill disposal of organic matter. J. Chem. Technol. Biotechnol. 30, 161-175. Schink, B. (1988) Principles and limits of anaerobic degradation: Environmental and technological aspects. In Biology of Anaerobic Microorganisms (Edited by Zehnder, A.J.B.), John Wiley and Sons, New York, US, 771-846. Schink, B. (1991) Syntrophism among prokaryotes. In The Prokaryotes (Edited by Balows, A., Triiper, H.G., Dworkin, M., Harder, W. and Schleifer, K.H.), Springer- Verlag, New York, US, 276-299. Schink, B. and Friedrich, M. (1994) Energetics of syntrophic fatty acid oxidation. FEMS Microbiol. Rev. 15, 85-94. Schniirer, A. (1995) Syntrophic acetate oxidation in ammonium-enriched methanogenic systems - organisms and pathways. PhD thesis, Department of Microbiology, Rapport 62, Swedish University of Agricultural Sciences, Uppsala, Sweden, ISSN 0348-4041. Smith, D.P. and McCarty, P.L. (1989) Energetic and rate effects on methanogenesis of ethanol and propionate in perturbed CSTRs. Biotechnol. Bioeng. 34, 39-54. 40 References

Sprott, G.D., Shaw, K.M. and Jarrell, K.F. (1984) Ammonia/potassium exchange in methanogenic bacteria. J. Biol. Chem. 259, 12602-12608. Stegmann, R. (1995) Concepts of waste landfilling. In Proceedings SARDINIA 95, Fifth International LandfillSymposium, Vol. /, CISA, Cagliari, Italy, 3-12. Straub, W.A. and Lynch, D.R. (1982) Models of landfill leaching: Organic strength. J. Environ. Eng. Div. ASCE 108, 251-268. Stronach, S.M., Rudd, T. and Lester, J.N. (1986) Anaerobic digestion processes in industrial wastewater treatment. Springer-Verlag, Berlin, Germany. Stumm, W. and Morgan, J.J. (1970) Aquatic Chemistry. Wiley-Interscience, New York, US. Suflita, J.M., Gerba, C.P., Ham, R.K., Palmisano, A.C., Rathje, W.L. and Robinson, J.A. (1992) The world’s largest landfill: A multidisciplinary investigation. Environ. Sci. Technol. 26, 1486-1495. Thauer, R.K., Jungermann, K. and Decker, K. (1977) Energy conservation in chemotrophic anaerobic bacteria. Bacteriol. Rev. 41, 100-180. Thauer, R.K. and Morris, J.G. (1984) Metabolism of chemotrophic anaerobes: Old view and new aspects. In The Microbe (Edited by Kelly, D.P. and Carr, N.G.) Part II, Procaryotes and eucaryotes, Cambridge University Press, Cambridge, UK, 123-168. Tong, X., Smith, L.H. and McCarty, P.L. (1990) Methane fermentation of selected lignocellulosic materials. Biomass 21, 239-255. Willumsen, H.C. (1996) Actual landfill gas yields. In Landfilling of waste: Biogas (Edited by Christensen, T.H., Cossu, R. and Stegmann, R.), E 8t FN SPON, London, UK, 293-295. Wolffson, C. (1985) Untersuchungen iiber den Einfluss der Hausmiillzusammensetzung auf die Sickerwasser- und Gasemissionen - Untersuchungen im Labormasstah (Investigation of the influence of household waste composition on leachate and gas emissions - lab-scale investigations). Sickerwasser aus Mulldeponien - Einflusse und Behandlung. Fachtagung Martz 1985, Braunschweig, Veroffentlichungen des Institute fur Stadtbauwesen, Heft 39, Technische Universitat Braunschweig, Germany. 119-46. (in German) Young, A. (1990) The landfill ecosystem. IMA J. Math. Med. Biol. 7, 199-217. Young, A. (1992) The effects of fluctuations in atmospheric pressure on landfill gas migration and composition. Water Air Soil Pollut. 64, 601-616. Zehnder, A.J.B. (1978) Ecology of methane formation. In Water Pollution Microbiology, Volume 2 (Edited by Mitchell, R.), John Wiley & Sons, New York, US, 349-376. Zehnder, A.J.B., Ingvorsen, K. and Marti, T. (1982) Microbiology of methane bacteria. In Anaerobic Digestion, 1981 (Edited by Hughes, D.E. et al), Proceedings of the Second International Symposium of Anaerobic Digestion, Elsevier Biomedical Press, B.V., Amsterdam, The Netherlands, 45-68. (cited in Christensen and Kjeldsen, 1989)