United States EPA-600/R-03-004
Environmental Protection • onno
Agency January 2003
v>EPA Research and
Development
A Laboratory Study to Investigate
Gaseous Emissions and Solids
Decomposition During Composting of
Municipal Solid Wastes
Prepared for
Office of Prevention, Pesticides, and Toxic Substances
and
Office of Solid Waste
Prepared by
National Risk Management
Research Laboratory
Research Triangle Park, NC 27711
-------
Foreword
The U.S. Environmental Protection Agency is charged by Congress with
protecting the Nation's land, air, and water resources. Under a mandate of national
environmental laws, the Agency strives to formulate and implement actions leading to
a compatible balance between human activities and the ability of natural systems to
support and nurture life. To meet this mandate, EPA's research program is providing
data and technical support for solving environmental problems today and building a
science knowledge base necessary to manage our ecological resources wisely,
understand how pollutants affect our health, and prevent or reduce environmental risks
in the future.
The National Risk Management Research Laboratory (NRMRL) is the Agency's
center for investigation of technological and management approaches for preventing
and reducing risks from pollution that threaten human health and the environment. The
focus of the Laboratory's research program is on methods and their cost-effectiveness
for prevention and control of pollution to air, land, water, and subsurface resources,
protection of water quality in public water systems; remediation of contaminated sites,
sediments and ground water; prevention and control of indoor air pollution; and
restoration of ecosystems. NRMRL collaborates with both public and private sector
partners to foster technologies that reduce the cost of compliance and to anticipate
emerging problems. NRMRL's research provides solutions to environmental problems
by: developing and promoting technologies that protect and improve the environment;
advancing scientific and engineering information to support regulatory and policy
decisions; and providing the technical support and information transfer to ensure
implementation of environmental regulations and strategies at the national, state, and
community levels.
This publication has been produced as part of the Laboratory's strategic
long-term research plan. It is published and made available by EPA's Office of
Research and Development to assist the user community and to link researchers with
their clients.
E. Timothy Oppelt, Director
National Risk Management Research Laboratory
EPA REVIEW NOTICE
This report has been peer and administratively reviewed by the U.S. Environmental
Protection Agency, and approved for publication. Mention of trade names or
commercial products does not constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Information
Service, Springfield, Virginia 22161.
-------
EPA-600/R-03/004
January 2003
A Laboratory Study to Investigate
Gaseous Emissions and Solids
Decomposition During Composting of
Municipal Solid Waste
by
Robert K. Ham
Dimitris Komilis
University of Wisconsin-Madison
Madison, WI 53706
Research Triangle Institute
3040 Cornwall! s Road
Research Triangle Park, NC 27709
EPA Cooperative Agreement CR 823052
EPA Project Officer: Susan A. Thorneloe
Air Pollution Prevention and Control Division (MD E305-02)
National Risk Management Research Laboratory
Research Triangle Park, NC 27711
Prepared for
U.S. Environmental Protection Agency
Office of Research and Development
Washington, DC 20460
-------
Abstract
A materials flow analysis was performed for composting municipal solid waste (MSW)
and specific biodegradable organic components of MSW. This work is part of an overall U.S.
Environmental Protection Agency (EPA) project providing cost, energy, and materials flow
information on different methods to reduce, recycle, treat, or dispose of MSW. This information
will be used by managers to optimize MSW management. Calculating energy and material
flows, emissions, and costs associated with different methods and mixes of methods for handling
MSW or for different components of MSW will provide basic information to guide
decisionmakers.
Composting is aerobic decomposition of a substrate, in this case MSW or its components.
The purpose of this work is to quantify and model energy and material flows into a typical
compost facility and material flows out of it. This work required laboratory experiments because
material flows in particular were not known for general MSW or its components.
The results indicate that MSW (at 25% inorganics) and its three largest decomposable
components (i.e., food wastes, mixed paper, and yard wastes) will lose 47, 66, 35, and 48%,
respectively, of their dry weight upon complete composting. This will produce 730, 1,340, 560,
and 800 kg of carbon dioxide (CO2) per dry U.S. ton of MSW, food wastes, mixed paper, and
yard wastes, respectively. Corresponding ammonia releases are 0.42, 49, 2.4, and 5.4 kg per dry
ton. Volatile organic compound (VOC) releases were quantified for 12 targeted VOCs, and
additional VOCs were found but not quantified. The results are modeled for facilities accepting
various combinations of MSW components (or MSW of various compositions).
11
-------
Table of Contents
Section Page
Abstract ii
List of Figures vi
List of Tables viii
Acronyms x
Acknowledgments xii
Executive Summary xiii
1.0 Introduction 1-1
2.0 Background and Literature Review 2-1
2.1 Solid Waste Composting 2-1
2.2 MSW Composting Technologies 2-3
2.3 Gaps in Knowledge and Research Objectives 2-4
3.0 Methods 3-1
3.1 Materials and Methods 3-2
3.1.1 Substrate Preparation 3-2
3.1.2 Substrate Seeding 3-4
3.1.3 Reactor Setup and Operation 3-5
3.1.4 Gas Composition Analysis 3-6
3.1.5 Carbon Dioxide Mass Loadings 3-8
3.1.6 Ammonia Mass Loadings 3-9
3.1.7 VOC Identification and Mass Loadings 3-9
3.1.8 VOC Recovery Tests 3-11
3.1.9 Control Run 3-12
3.1.10 Solids Measurements 3-13
3.1.11 Dry Weight Reduction Calculations 3-16
in
-------
Table of Contents (continued)
Section Page
3.1.12 Precision and Bias 3-16
3.2 Initial Testing of Laboratory Setup and Methods 3-16
3.3 Summary 3-23
4.0 Carbon Dioxide and Ammonia Yields During Composting 4-1
4.1 Statistical Experimental Designs 4-1
4.1.1 Full Factorial Analysis 4-1
4.1.2 Mixture Experiment 4-5
4.2 Results and Discussion 4-6
4.2.1 Seed Interaction 4-6
4.2.2 Calculating Gaseous Emissions from Seeded Runs 4-7
4.2.3 MSW Component Interactions 4-12
4.2.4 Mixture Experiment 4-18
4.3 Final Recommended Model(s) 4-20
5.0 Solids Decomposition During Composting 5-1
5.1 Results and Discussion 5-2
5.1.1 Initial Composition of Substrates 5-2
5.1.2 Substrate Degradability as a Function of Lignin and HWSM Contents 5-4
5.1.3 Reduction of Chemical Components 5-8
5.1.4 Solids Degradation Rates 5-15
5.1.5 Compost Maturity Indicators 5-20
5.1.6 Errors in Analytical Measurements 5-29
6.0 Emissions of Volatile Organic Compounds During Composting 6-1
6.1 Results and Discussion 6-2
6.1.1 Identification of VOCs from Biodegradable Fraction of MSW 6-2
6.1.2 Mass Loadings of Selected VOCs During the Composting Process . . . 6-9
6.1.3 Interaction of Seed and Component in Seeded Runs 6-11
6.1.4 Emperical Models fro Estimating VOC Yields from MSW Mixtures 6-13
6.1.5 VOC Volatilization Rates 6-16
6.2 Origin and Fate of VOCs 6-20
6.3 Spiking of VOCs to MSW Organic Substrates 6-21
iv
-------
Table of Contents (continued)
Section Page
6.3.1 Ethylbenzene Spike 6-21
6.3.2 Spike with a Mixture of Alkylated Benzenes 6-23
7.0 Conclusions 7.1
8.0 References 8-1
Appendix A - Audit Report A-l
-------
List of Figures
Figure Page
3-1 Conceptual experimental design 3-3
3-2 Digester setup 3-7
3-3 Sequential procedure for analysis of solid substrates 3-15
3-4 Solids decomposition profiles for replicate mixed paper runs MXPj and MXP0 .... 3-17
3-5 Cumulative percentage of total organic carbon emitted as C-CO2 from different
substrates during composting 3-18
3-6 Cumulative percentage of total initial organic N emitted as N-NH3 from different
substrates during composting 3-19
3-7 Cumulative production of eight VOCs during composting of seed at thermophilic
temperatures 3-24
4-1 Cumulative CO2 production (as g C) per dry kg of starting material
(including seed) 4-8
4-2 Cumulative NH3 production (as g N) per dry kg of starting material
(including seed) 4-9
4-3 CO2 and NH3 daily production rates for the FW and FWns runs 4-16
4-4 CO2 and NH3 daily production rates for the YW and YWns runs 4-17
4-5 Response surface for estimation of CO2 yields using equation 4-7 4-22
4-6 Response surface for estimation of NH3 yields using equation 4-8 4-23
5-1 Effect of initial lignin content on carbon dioxide yields 5-5
5-2 Effect of initial hot water soluble matter (HWSM) content on carbon dioxide
yields 5-9
5-3 Dry mass loss for each chemical group during composting (in % of the total initial
dry mass) 5-11
5-4 Concentrations of organic constituents and CO2 production rate during composting
of grass 5-16
5-5 Concentrations of organic constituents during composting of seeded mixed paper .. 5-17
vi
-------
List of Figures (continued)
Figure Page
5-6 Correlation plots between final average CO2 production rate and fats/lipids
reduction and C/N ratio reduction 5-23
5-7 Correlation plots between final average CO2 production rates and initial and final
solids chemical composition (% VS basis) 5-25
5-8 Correlation plot between final average CO2 production rates and initial and final
C/L and C/N ratios for various substrates 5-28
5-9 Cellulose to lignin (C/L) ratios for initial and composted materials from 15
experimental runs (plus compost from a MSW composting facility) 5-30
6-1 Cumulative production of six VOCs during composting of seeded mixed paper;
yields have not been corrected for seed 6-17
6-2 Cumulative production of nine VOCs during composting of unseeded mixed paper .6-18
6-3 Cumulative production of eight VOCs during composting of seed 6-19
6-4 Volatilization rates of ethylbenzene and CO2 cumulative production from newsprint
and yard wastes during composting 6-24
6-5 TEX cumulative volatilization profile and CO2 cumulative production during
composting of MSW 6-27
vn
-------
List of Tables
Table Page
3-1 Summary of Analytical Techniques 3-10
3-2 Experimental Results from Selected MSW Components During Composting 3-20
4-1 22 Full Factorial Experimental Runs to Investigate Interactions Between Individual
MSW Components and Added Seed 4-3
4-2 Calculated Coefficients for the Three 22 Factorial Designs 4-4
4-3 23 Full Factorial Experimental Design to Investigate Main Effects and Interactions
During Composting of the Three Major MSW Organic Components 4-4
4-4 Fractions of Components in Mixtures (dry mass basis) and Gaseous Yields for
12 Runs 4-10
4-5 Parametric Regression Analysis for Third-, Second-, and First-Order Models
Based on Equation 4-3 4-19
5-1 CO2 Yields and Initial Chemical Composition of Substrates from 17 Experimental
Runs 5-3
5-2 Correlation Matrix for Estimation of Compost Maturity Indicators Using the
Reduction of Organic Chemical Groups During Composting 5-10
5-3 Contribution of Dry Loss of Each Chemical Group to Total Dry Mass Loss of a
Substrate 5-12
5-4 Solids Decomposition Rates (As Percent of Dry Mass Decomposed/Initial Dry
Mass/Day) During Composting of Grass and Seeded Mixed Paper 5-18
5-5 Correlation Matrix for Estimation of Compost Maturity Indicators Using the
Composted Substrate Chemical Composition 5-21
5-6 Carbon, Nitrogen, and Solids Recoveries 5-31
6-1 VOCs Identified in Gaseous Emissions from Nine Runs 6-3
6-2 Yields of 12 VOCs from 12 Runs 6-10
Vlll
-------
List of Tables (continued)
Table Page
6-3 Additive Contribution of Seed to the Emissions of Each of the 12 VOCs Emitted
from a Seeded Mixture 6-13
6-4 Physicochemical Properties of VOCs Used in TEX Run 6-26
6-5 Volatilized Fraction (in %) of Initially Spiked VOC Mass During TEX Run and
Length of Time (days) for Volatilization After Each Spike 6-28
IX
-------
Acronyms
BTEX benzene, toluene, ethylbenzene, and xylenes
BVS volatile solids reduction
COV coefficient of variation
EPA U.S. Environmental Protection Agency
FA fulvic acids
FID flame ionization detector
FW food waste
GC gas chromatograph
HA humic acids
HHW household hazardous waste
HPLC high performance liquid chromatography
HS humic substances
HW hazardous waste
HWSM hot water soluble matter
ISWM integrated solid waste management
LiPs lignin peroxidases
LOTV low odor threshold value
MnPs manganese-dependent peroxidases
MS mass spectrometer
MSW municipal solid waste
MSWCF municipal solid waste composting facility
MXP mixed paper waste
NHS non-humic substances
OCC old corrugated cardboard
OFF printed office paper
ONP old newsprint
PAH polycyclic aromatic hydrocarbons
RAC ratio of area counts
SSCF sewage sludge composting facility
SWM solid waste management
-------
Acronyms (continued)
TCD thermal conductivity detector
TCFM trichlorofluoromethane
TOC total organic carbon
TLV threshold limit value
VOC volatile organic compound
YW yard waste
XI
-------
Acknowledgments
The research presented here is part of a life cycle inventory project funded by the U.S.
Environmental Protection Agency (EPA) through EPA Cooperative Agreement CR 823052 with
the Research Triangle Institute. The support of EPA and, in particular, that of the project
manager, Susan Thorneloe, is acknowledged, as is the subcontract management and help of
Keith Weitz of the Research Triangle Institute.
The carbon dioxide and ammonia capture and quantification methods were checked by an
EPA quality assurance and quality control (QA/QC) staff member, Richard Shores, and the
recovery numbers given in the paper are from the QA/QC report. The authors would like to
acknowledge the assistance of Paul Fritschel, director of the Environmental Engineering
Laboratories, and the many laboratory support personnel involved with this project. Total
organic carbon and total nitrogen measurements were performed by the Soil and Plant Analysis
Laboratory at the University of Wisconsin-Madison, and gas chromatography/mass spectroscopy
(GC/MS) analyses were performed at the State of Wisconsin Laboratory of Hygiene and at the
Department of Soil Science at the University of Wisconsin-Madison.
xn
-------
Executive Summary
Integrated Solid Waste Management (ISWM) requires making informed decisions to
optimize solid waste management (SWM) by minimizing environmental releases, energy and
resource use, and costs, while maximizing useful outputs. Judgment is required to balance these
factors for a given region. All realistic methods of SWM must be considered, including
recycling, combustion, composting, and landfilling. The U.S. Environmental Protection Agency
(EPA) initiated research with cofunding from the U.S. Department of Energy (DOE) to develop
tools and information for evaluating strategies for ISWM. The research team for this effort was
led by the Research Triangle Institute, in cooperation with North Carolina State University,
Franklin Associates, Ltd., Roy F. Weston, and the University of Wisconsin in Madison. This
document is the report of the research conducted by the University of Wisconsin to develop
degradation information for use in a process model for composting. The process model is
developed and presented in a separate report.
Composting, as commonly defined for SWM purposes and as used here, is aerobic
biological decomposition of solid waste. Factors affecting the rate and completeness of
decomposition are manipulated according to local needs and constraints to produce the desired
decomposition. These factors include waste selection or exclusion, particle size reduction,
mixing, seeding, moisture addition, and aeration. In general, more costly facilities use
mechanical methods to prepare the waste and to promote decomposition. Less costly facilities
emphasize natural processes, reducing mechanical needs.
Composting occurs in nature as organic materials degrade aerobically. It evolved as a
SWM tool when public health and sanitation became municipal functions during the 19th century
in some parts of the world. During the 20th century, composting became increasingly understood
as research, development, and experience resulted in different facility designs and operating
methods to meet local requirements. Composting has been applied most often to municipal solid
wastes (MSW), sewage sludge, and agricultural wastes, but increasingly over the last 10 years
also to yard wastes. This project focused on MSW and yard waste composting.
Although composting has a long history and has been the subject of much research and
development, little is known about the extent of decomposition. There is no information on the
xiii
-------
amount of gases produced during decomposition and only general information and theoretical
projections of gas composition. Further, there is no such information specific to MSW
components such as food waste, paper, and yard waste. Since this information tied to specific
components is required for the overall life cycle project, it was considered necessary to perform
laboratory work. The dry weight loss of waste fed to composting facilities could have been
monitored, but it would have been very difficult to use full-scale facilities to develop this
information for different waste components, and it would have been impossible to measure
gaseous emissions.
This report begins with project background and a general literature review followed by a
chapter on the laboratory methods used to simulate a compost facility, and initial testing of these
methods for reasonableness and reproducibility of the results. These methods were used to
compost the three major organic components of MSW (i.e., food waste, mixed paper, and yard
waste). Although more detailed testing was not done on specific waste sub-components (e.g.,
different types of paper), the waste components tested are deemed representative of specific
degradable sub-components as required for the overall project.
The following report presents the results for major gaseous emissions, solids reduction,
and volatile organic compound (VOC) emissions, respectively, for composting food, mixed
paper, and yard wastes. The experimental design used involved determining the need for and
impact of seed for each waste component, plus composting each component separately and in
combination with each of the other two components or with both other components. In addition,
a special experiment was performed with a mixture of all three components but at concentrations
such that these results, in combination with the results from the individual components and
mixtures, could be modeled. The result of this work includes equations relating weight loss and
quantitative carbon dioxide, ammonia, and VOC emissions to initial waste composition. The
chapter on VOCs also gives qualitative information regarding specific VOCs produced by
composting each component and mixtures of components.
The results indicate that MSW (at 25% inorganics) and the three largest decomposable
components of MSW (i.e., food wastes, mixed paper, and yard wastes) will lose 47, 66, 35, and
48%, respectively, of their dry weight upon complete composting. This will produce 730, 1,340,
560, and 800 kg of carbon dioxide per dry ton of MSW, food wastes, mixed paper, and yard
wastes, respectively. Corresponding ammonia releases are 0.42, 49, 2.4, and 5.4 kg per dry ton.
VOC releases were quantified for 12 targeted VOCs, and additional VOCs were found but not
quantified
xiv
-------
Section 1.0 Introduction
1.0 Introduction
Integrated Solid Waste Management (ISWM) requires making informed decisions to
optimize solid waste management by minimizing environmental releases, energy and resource
use, and costs while maximizing useful outputs. Judgment is required to balance these factors
for a given region. All realistic methods of managing solid waste must be considered, including
recycling, combustion, composting, and landfilling. The U.S. Environmental Protection Agency
(EPA) initiated research with cofunding from the U.S. Department of Energy (DOE) to develop
tools and information for evaluating strategies for ISWM. The research team for this effort was
led by the Research Triangle Institute, in cooperation with North Carolina State University,
Franklin Associates, Ltd., Roy F. Weston, and the University of Wisconsin in Madison. This
document is the report of the research conducted by the University of Wisconsin to develop
degradation information for use in a process model for composting. The process model is
developed and presented in a separate report.
Composting, as commonly defined for solid waste management purposes and as used
here, is aerobic biological decomposition of solid waste. Factors affecting the rate and
completeness of decomposition are manipulated according to local needs and constraints to
produce the desired decomposition. These factors include waste selection or exclusion, particle
size reduction, mixing, seeding, moisture addition, and aeration. In general, more costly
facilities use mechanical methods to prepare the waste and to promote decomposition. Less
costly facilities emphasize natural processes, reducing mechanical needs.
Composting occurs in nature as organic materials degrade aerobically. It evolved as a
solid waste management tool when public health and sanitation became municipal functions
during the 19th century in some parts of the world. During the 20th century, composting has
become increasingly understood as research, development, and experience resulted in different
facility designs and operating methods to meet local requirements. Composting has been applied
most often to municipal solid wastes (MSW), sewage sludge, and agricultural wastes, but
increasingly over the past 10 years also to yard wastes. This project focused on MSW and yard
waste composting.
The process model of composting requires information on energy and materials flowing
into a compost facility, the facility itself, and materials leaving the facility. Further, since a
major portion of the materials leaving a compost facility is compost, which is subject to further
degradation and leaching when it is applied to land or placed in a landfill, the emissions
associated with this additional degradation and leaching are included as well. The model for
composting does not include any waste processing, separation, or haul prior to the compost
plant. These issues are modeled in other parts of the overall project because they are the same as
might be used for recycling, for example. Similarly, haul and landfilling of rejects and haul of
compost are excluded from this report but are considered in the overall project.
n
-------
Section 1.0 Introduction
Although composting has a long history and has been the subject of much research and
development, little is known about the extent of decomposition. There is no information on the
amount of gases produced during decomposition and only general information and theoretical
projections of gas composition. Further, there is no such information specific to MSW
components such as food waste, paper, and yard waste. Because this information is required for
the overall life cycle project, it became necessary to conduct laboratory analyses. Dry weight
loss of waste fed to composting facilities could have been monitored, but it would have been
very difficult to use full-scale facilities to develop this information for different waste
components and it would have been impossible to measure gaseous emissions.
This report begins with project background and a general literature review followed by a
chapter on the laboratory methods used to simulate a compost facility and initial testing of these
methods for reasonableness and reproducibility of the results. These methods were used to
compost the three major organic components of MSW-food waste, mixed paper, and yard waste.
Time and budget constraints did not allow more detailed testing on specific waste
subcomponents, such as different types of paper; however, the waste components tested are
deemed representative of specific degradable subcomponents as required for the overall project.
The following three chapters present the results for major gaseous emissions, solids
reduction, and volatile organic compound (VOC) emissions, respectively, for composting food,
mixed paper, and yard wastes. The experimental design used involved determining the need for
and impact of seed for each waste component, plus composting each component separately and
in combination with each of the other two components or with both other components. A special
experiment was performed with a mixture of all three components to further investigate possible
interactions between the waste components. The result of this work includes equations relating
weight loss and quantitative CO2, NH3, and VOC emissions to initial waste composition. The
chapter on VOCs also gives qualitative information on specific VOCs produced by composting
each component and mixtures of components.
Note that care should be taken in using the laboratory data. The data represent emission
estimates from three waste components (yard waste, food waste, paper) and thus are not
necessarily representative of emissions from the entire waste stream. In particular, variability in
the composition of waste entering a compost facility can lead to significant variations in compost
products, cost, and emissions.
1-2
-------
Section 2.0 Background and Literature Review
2.0 Background and Literature Review
Composting in the United States has primarily focused on the treatment of organic
substrates such as sewage sludges and agricultural wastes. Composting is also one way of
treating and disposing of municipal solid wastes, but it has always had limited success in
achieving its ideal purpose, that is to produce a soil amendment. MSW-derived compost in the
United States is either stockpiled, landfilled, or used as a landfill cover (Kashmanian and
Spencer, 1993). At the same time, higher tipping fees in composting facilities than in landfills
(Goldstein and Steuteville, 1994) may render MSW composting a less "popular" option.
However, since landfill space is reduced and social and economic reasons focus on recycling
practices, MSW composting is gaining more attention. As of November 1995, 17 MSW
composting projects were in operation in the United States, while 27 were in various stages of
development (Steuteville, 1995). At the same time, MSW composting is studied as a
pretreatment solid waste management technique prior to landfilling. The reasoning is that the
volume of solid wastes can be reduced to up to 50% (Tchobanoglous et al., 1993) in a relatively
short period of time, which increases the active landfill life. Additionally, the concentration of
readily organic degradable material in the waste can be reduced significantly during composting,
decreasing the organic load emitted from the landfill in both liquid (leachate) and gaseous
(biogas) forms (Komilis and Ham, 2000). Thus, the capital and operating costs associated with
the design of the relevant landfill elements as well as the overall landfill postclosure monitoring
operation are expected to be reduced.
Yard waste composting is now widely practiced in the United States, with approximately
3,260 facilities in operation (U.S. EPA, 1997). The goals are to reduce landfill space utilization
and to promote productive use of yard waste as compost.
2.1 Solid Waste Composting
Composting is a method of solid waste management during which the organic portion of
the solid waste stream is biologically decomposed under controlled conditions to a state in which
it can be handled, stored, and/or applied to the land without adversely affecting the environment
(Epstein, 1997). The description of the composting process and wastes to be composted, and the
factors that affect it, are subjects covered in several reviews (Gray et al., 1971; de Bertoldi et al.,
1983; Diaz et al., 1993) and will not be discussed in detail here.
During composting, oxygen is used as the terminal electron acceptor and the organic
matter passes through a thermophilic stage. The main classes of organic compounds in MSW
belong to known biochemistry classes and have known structures (Stevenson, 1994). These
compounds—referred to herein as nonhumic substances (NHS)—are attacked by several types of
mesophilic and thermophilic microorganisms. Bacteria are prevalent in the initial stages of
decomposition, while fungi and actinomycetes are prevalent at the final stages (Epstein, 1997).
2-1
-------
Section 2.0 Background and Literature Review
Some of the organic matter is eventually mineralized to CO2 and water after passing through
several intermediate stages. Another major pathway in composting is the polymerization of a
fraction of the initial NHS and probably of the generated microbial tissues to form a group of
compounds known under the collective term of humic substances (HS) or humus, which is
refractory to biodegradation. Humified organic matter can be applied to soil to enhance plant
growth mainly because of its increased water retention potential. The physical structure of the
soil is improved and the gradual leaching of nutrients adds to the beneficial effect of humified
organic matter. Other effects of humic material on plant growth are discussed by Vaughan and
Malcolm (1987).
CO2 and ammonia (NH3) gases are the primary metabolic byproducts of the composting
process. Carbon dioxide is a well-known greenhouse gas. NH3 is an intermediate byproduct
from protein degradation. Due to thermophilic temperatures normally encountered during
composting (>50 °C), nitrification rarely occurs (de Bertoldi et al., 1983), and NH3 accumulates
and volatilizes or is bound to organic matter through biologically mediated processes
(Stevenson, 1994). NH3 is at least partially responsible for odor problems common to
composting operations (Diaz, 1987) and can cause human health impacts. Several volatile
organic compounds were identified by Eitzer (1995) in U.S. MSW composting facilities, with
alkylated benzenes, chlorofluorocarbons, and terpenes being at the highest ambient air
concentrations.
Little is known about the yields and production rates of CO2, NH3, and VOCs of different
solid waste components during composting. Information on these gaseous yields from MSW and
yard waste composting facilities is needed to evaluate the contribution of these facilities to the
global production of regulated gases. Such knowledge can be used to compare waste
management techniques and optimize waste management policies using life cycle analysis tools
(U.S. EPA, 1998).
In addition, there is limited information on weight losses and biodegradable carbon
fractions of MSW or individual MSW components during composting. Weight loss is a critical
factor for composting plant design (Haug, 1993).
Solids loss during composting of organic matter or MSW has been used to study the
effect of various parameters affecting the composting process (Regan and Jeris, 1970) and to
derive compost maturity indicators (Mathur et al., 1993). Several studies have focused on the
use of humic matter measurements to derive compost maturity indicators (de Nobili and Petrussi,
1988; Ciavatta et al., 1993; Riffaldi et al., 1986). Changes in the primary chemical constituents
of raw MSW have been suggested as indicators of compost maturity. Inoko et al. (1979)
measured the changes of such chemical parameters namely cellulose, hemicellulose, lignin and
water-soluble sugars during several stages of composting mixed MSW in actual composting
plants in Japan and suggested that, for MSW-derived finished composts, the ratio of carbon in
polysaccharides to total carbon should be less than 35%, total N should be above 2% (dry matter
basis), and C/N ratio less than 20.
Apart from the effects of various parameters (e.g., pH, C/N ratio, aeration rate) on MSW
composting, overall decomposition is expected to be affected by the heterogeneity of the MSW
stream and the percentages of each of the different MSW organic components. In one of the
-------
Section 2.0 Background and Literature Review
oldest studies, Poincelot and Day, (1960) studied the effect of combining leaves with various
organic and inorganic materials on cellulose decomposition. The authors observed that
decomposition is enhanced by materials with high contents of readily available nitrogen, such as
ammonium salts and sewage sludge. In terms of compost maturity indicators, limited
information exists on whether similar characteristics exist among composts derived from
different MSW components or from MSW of various compositions.
Compost maturity has been a subject of research for several years, yet there is no
universally accepted way of evaluating it (Mathur et al., 1993).
2.2 MSW composting technologies
Mixed MSW or the source-separated organic fraction of MSW are the substrates to
the composting facilities. Most operational projects refer to mixed MSW composting with
several degrees of pretreatment prior to the actual composting process. Pretreatment usually
comprises screening and removal of large items, shredding with hammermills or shear shredders,
and occasionally magnetic removal of some items (Gould and Meckert, 1994). There are three
basic compost process technologies:
• Windrow turned facilities. The waste is placed in windrows, which are periodically
turned using windrow turners or front end loaders until a mature material is
produced. Windrows are basically long piles of controlled cross-sectional size to
promote air contact and minimize anaerobic conditions. The turning evens
decomposition and moisture and spreads organisms. Michel et al. (1993) showed
that the turning frequency does not affect composting rates, since aeration
concentrations within the pile return to the levels prior to turning within around 3
hours.
• Aerated static pile facilities: This technology is common in sewage sludge
composting. Air is pumped from the bottom of the piles through perforated pipes.
According to Miller et al. (1982), forced aeration is better than vacuum-induced
ventilation in achieving air distribution within the pile and also in cooling a
windrow at high temperatures.
• In-vessel (digester) facilities: Several proprietary systems exist, the goal of which
is to achieve higher composting rates, since several parameters can be more
precisely controlled than with piles or windrows. In-vessel systems are associated
with higher capital costs and are almost always used in conjunction with windrow
systems to achieve maturation.
According to de Bertoldi et al. (1983), maturation times in composting processes using
closed reactors or open windrows do not differ if the processes are carried out correctly. Based
on Glaub et al. (1989), the turned windrow process has the least energy requirements (21.4
kWh/ton MSW) compared to aerated static piles (25.2 kWh/ton) and in-vessel systems (30.0
kWh/ton).
2-3
-------
Section 2.0 Background and Literature Review
2.3 Gaps in Knowledge and Research Objectives
Information is limited on the fate and source of VOCs from municipal solid wastes
during composting and the contribution of MSW composting to the production of greenhouse
gases. Furthermore, deriving a universally accepted compost maturity indicator is still being
researched. Thus, the primary objectives of this work are to
• Provide data and models to estimate CO2, NH3, and VOC yields from individual
MSW components and MSW of various compositions
• Investigate the sources and fate of VOCs in MSW composting environments
• Investigate solids decomposition and provide potential compost maturity indicators
for various organic substrates during composting.
2-4
-------
Section 3.0 Methods
3.0 Methods
To maintain control of the process and to make the necessary measurements, a laboratory
procedure was developed that simulates as closely as possible full-scale composting facilities.
Composting is an aerobic degradation process in which oxygen is used as the terminal electron
acceptor. Heterotrophic bacteria and fungi oxidize the biodegradable carbon fraction of each
substrate to obtain energy for metabolic activities and to build new biomass. Nutrient and
oxygen concentrations, substrate moisture content, the presence of seed, and temperature levels
play critical roles in determining the composting rate (Diaz et al., 1993).
The composting process was simulated based on the following findings and guidelines.
1. Moisture contents ranging from 52% to up to 60% (wet weight) do not limit
composting of most organic substrates (Schulze, 1961).
2. A minimum value of 15% (by weight) of oxygen present at the headspace of
compost vessels has been found to not limit the composting process (de Bertoldi
etal., 1988).
3. Thermophilic temperature ranges (50 to 70 °C) occur frequently in actively
composted substrates (Diaz et al., 1993). Schulze (1960), using synthetic garbage
as substrate, showed that higher oxygen uptake rates are observed with higher
compost temperatures, up to a maximum value of 70 °C. Temperatures higher
than 70 °C can significantly hinder decomposition rates (Diaz et al., 1993). In
accordance with the Arrenhius temperature dependence rule, biological reaction
rates increase with increasing temperatures up to a maximum level (Sawyer and
McCarty, 1978), and thermophilic temperatures are expected to result in higher
composting rates than mesophilic temperatures.
4. A C/N ratio of 25 to 30 has been suggested for optimum composting (Diaz et al.,
1993).
5. Seeding of substrates with microorganisms derived from active MSW composting
operations is expected to accelerate the initiation and sustainability of the
composting process for some substrates (Gray et al., 1971).
The method developed here was designed to degrade MSW or MSW components
aerobically. A closed batch system was used with defined input material to collect and measure
all gaseous emissions per unit of substrate. Water, seed, and nutrients were provided at the
beginning of the process, based on the above-mentioned guidelines, and excess air was supplied
continuously throughout the process. The system is designed to simulate well-operated (i.e.,
3-1
-------
Section 3.0 Methods
minimal anaerobic biological activity) in-vessel MSW composting plants or MSW composting
plants of the aerated static pile type with forced aeration.
Mass loadings of CO2, NH3, and VOCs were quantified continuously until "full"
decomposition was reached, as was indicated by termination of measurable carbon dioxide flow
rates. Solids analyses were performed at the beginning and end of the process. The conceptual
experimental design is shown in Figure 3-1.
The following sections describe in detail the materials and methods used to develop the
laboratory setup.
3.1 Materials and Methods
3.1.1 Substrate Preparation
The decomposable fraction of MSW was simulated by breaking it into three major
organic components: mixed paper wastes (MXP), yard wastes (YW), and food wastes (FW).
These are the three major organic components of MSW in the United States, accounting for
approximately 40%, 19%, and 9% of the overall wet MSW composition, respectively, after
recycling (Tchobanoglous et al., 1993). The use of well-defined batches of these three MSW
organic components allowed reproducibility of the experiments.
These three organic components were either composted individually or in mixtures.
Mixtures were prepared based on a typical U.S. residential MSW composition as achieved after
recycling (Tchobanoglous et al., 1993). The three major organic components were normalized to
100% by neglecting inorganic components and smaller organic component fractions.
Accordingly, when all three components were tested, the percentages were set at 59.3% mixed
paper, 27.4% yard wastes, and 13.3% food wastes, on a wet weight basis, or approximately 80%
for mixed paper, 15.4% for yard wastes, and 4.3% for food wastes, on a dry weight basis. The
relative weights put in the digesters was kept approximately the same, whether two components
were combined or a mixture of all three components was prepared.
Mixed paper wastes (MXP) were prepared by mixing old corrugated cardboard (OCC),
printed office paper (OFF), and old newsprint (ONP) at percentages of approximately 45%, 21%,
and 34%, respectively, on a wet weight basis. The percentage of each component in the mixture
was based on a typical composition for mixed paper in the United States (Tchobanoglous et al.,
1993). Other types of paper, present in smaller quantities, were not accounted for.
Cardboard was obtained from a solid waste recycling facility located in Madison,
Wisconsin. Newsprint was collected from a newsprint recycling bin, located in a Student Union
building at the University of Wisconsin-Madison. Ordinary used office paper (20.3 cm x 27.9
cm sized) was collected from graduate student offices from the Department of Civil and
Environmental Engineering at the University of Wisconsin-Madison. All types of paper used in
the experiment were shredded with an open top shear shredder (Shred Pax Corporation,
Woodale, IL) to approximately 3-cm2 to 4-cm2 pieces. Shredding was done to promote
3-2
-------
Section 3.0
Methods
C02, NH3, VOCs
Substrate
Nutrients
(C/N=25-30)
Water
(55%-60% wet weight)
Microbial seed
(10:1 dry weight substrate
to seed)
Water-soluble sugars
Fats & lipids
Cellulose
Hemicellulose
Lignin/humus
Total N
Organic C
Volatile solids
Ash
02
(02> 10%-15%v/v)
Humus,* biomass,"
undecomposed
substrate
analytes not specifically measured
Figure 3-1. Conceptual experimental design.
degradation and so paper would fit in the reactors; it should have no effect on the final extent of
degradation. All paper components were collected at the same time and stored in a 4°C room
until use.
Yard wastes were prepared by mixing grass clippings and leaves, which are common
major constituents of yard waste composting facilities but can also be found in MSW
composting as well. The ratio of grass to leaves was arbitrarily set at approximately 1.5:1 on a
dry content basis. Grass clippings were collected randomly from a yard waste drop-off area in
Madison, Wisconsin, in May 1997 and September 1997. Moisture content analysis and volatile
solid content analyses were performed for each grass batch prior to its use in a digester and
individual experimental runs were performed on the two batches of grass. There was a
difference between them, according to weight loss and CO2 evolution results. Intermediate
3-3
-------
Section 3.0 Methods
sampling and analyses of volatile solids showed that degradation during storage in the cold room
was less than 1% for grass. Leaves were collected once from a local yard. Grass, leaves, and
branches were used as is without any shredding. To aid reproducibility, all yard waste
components were stored at 4 °C until needed.
Food wastes were simulated using a typical food waste composition, as provided by the
In-Sink-Erator company (ISE, 1996). Food wastes were prepared from six common food
products: milk, cooked pasta, cooked hamburger, lettuce, raw potatoes, and carrots, mixed in
equal wet weight amounts. All food products were obtained from a local grocery store and
prepared prior to each run.
Approximately 10% of each of the individual substrates used was sampled using a riffler
(Barlaz, 1988) and subjected to moisture content and volatile solids content analyses prior to
each experiment. Periodically samples were also analyzed for carbon and nitrogen content. All
substrates were weighed to an accuracy of ±2.5 g. If the overall initial moisture content of a
component or mixture was calculated to be less than 55% to 60% by wet weight, water was
added using a spray bottle to reach a moisture content of approximately 60%, and the material
was mixed. Food wastes and yard wastes had initial moisture contents greater than 60% wet
weight, requiring no additional moisture.
Nitrogen was added to all substrates or mixtures for which the initial average C/N ratio
was higher than 30. Nitrogen was added in the form of NH4NO3 salt that was dissolved in the
water used to raise the moisture content. No nitrogen was added to food waste and yard wastes.
Both nitrogen forms present in the NH4NO3 salt were assumed to be equally available to the
microorganisms. A distinct ammonia odor was detected during the spraying of water with the
nutrient salts on the substrates while the digesters were open. This volatilized ammonia was a
temporary phenomenon and was considered unlikely to have a measurable impact on nitrogen
mass balances. No phosphorus or other trace elements were added to the substrates to simulate
field conditions; phosphorus is not normally added to MSW or YW prior to field composting. In
contrast, nitrogen is often added to MSW during composting in the form of wastewater biosolids.
Overall dry weights of the combined materials used in all runs ranged from 130 g to
1,100 g. Overall wet weights, after addition of moisture, ranged from 1.5 kg to 2.5 kg.
3.1.2 Substrate Seeding
Seeding was considered necessary to start the decomposition process and supply an
active microbial population suitable for composting in a reproducible manner. Most of the runs,
including individual substrates and mixtures, were seeded with partially composted MSW (not
matured), hereafter referred to as seed. Seed was collected from a nearby MSW composting
facility located at Portage, WI. Approximately 15 kg of seed (wet weight) was collected
in February 1996 from the outlet of a 5-day retention time drum digester that receives raw MSW
without any preprocessing. Material leaving the composting vessel is sieved at the exit end of
the drum and so was already ground and mixed. The seed was further screened through a
12.7-mm screen to remove larger items such as glass and plastic. It was stored in a 4°C room
until the end of the experimental runs. Seeding was provided at a ratio of approximately 1:10 of
dry seed to dry substrate.
-------
Section 3.0 Methods
To evaluate the contribution of seed decomposition to total gaseous emissions of a
seeded MSW mixture, seed was composted individually. The volatile solids content of seed,
measured immediately after collection from the facility, was 94.8% ± 0.19% (dry weight basis).
A decrease of the volatile solids content of the seed was observed during storage over a year
period. The seed volatile solids content prior to use in all runs reported here had been stabilized
to 71.5% ± 0.5% (on a dry weight basis).
3.1.3 Reactor Setup and Operation
Experimental runs were done using five custom-made 25-L airtight stainless steel
digesters constructed by Hooper Corporation (Madison, WI), as shown in Figure 3-2. All
individual materials and their mixtures, after preparation, wetting, and addition of appropriate
nutrients, were placed in each digester. All MSW substrates were mixed with #1.5 HyPak
aluminum packing material (4.5 cm thickness x 4.5 cm diameter) acquired from Norton
Chemicals (Akron, OH). The packing material was uniformly mixed with the substrate to
approximately 10% to 15% of the volume occupied by the substrate. The packing material was
needed to facilitate air flow, prevent excessive channeling of air at the sides of the digester, and
ensure that aerobic conditions were maintained within the substrate. The composting process
was simulated in a batch mode.
The five digesters were operated concurrently. Each digester had a removable lid and
had one input and one output port. The digesters were kept in a controlled temperature room at a
temperature in the thermophilic (55±5 °C) or mesophilic range (32±3 °C), as desired.
Laboratory ambient air was continuously pumped into the digesters using a Barnant oil-
less air pump (Barrington, IL, Model No. 400-1901) operating at a positive pressure. The air
was first passed through a 200-g activated carbon (Orbo 32, 6-14 mesh) filter to capture ambient
air VOCs and then through a 500-mL 5 N KOH solution to capture ambient CO2. It then passed
through 10 L of distilled water, kept at incubator temperature, to humidify the substrate and
minimize excessive drying. Air was then directed through a manifold to the individual digesters.
Pumped air served as a source of oxygen for the oxidation of substrate as well as the carrier for
emitted gases.
A valve and a flowmeter (Barnant Co., Barrington, IL) prior to each digester was used to
regulate the airflow to each digester to 150 to 300 mL/min. This flow rate was used to avoid
limiting the degradation process by maintaining oxygen concentrations in the exit gas higher
than 15% by volume for all runs. A smaller additional VOC filter (Orbo 32) was placed after
each flowmeter for each digester. The smaller VOC filters were tested and replenished
periodically to check if breakthrough of VOCs from the preceding larger carbon filter occurred.
A 4.5-cm air plenum was created at the bottom of each digester using an aluminum
screen supported by aluminum packing material. The air plenum was designed to even the
distribution of air in the digester.
After exiting the digesters, the air stream first passed through an activated coconut
charcoal trap (Orbo 32, SUPELCO, Bellefonte, PA) to remove VOCs for quantification and to
partially remove organic compounds of acidic nature (e.g., acetic acid) that could interfere with
-------
Section 3.0 Methods
the CO2 quantification to follow. The coconut charcoal trap (large Orbo-32) consisted of a
primary section and a breakthrough section containing 400 mg and 200 mg of activated coconut
charcoal (20 to 40 mesh), respectively. The air stream was then bubbled through a 750-mL 5 N
KOH solution to capture carbon dioxide and then through a 500-mL 1 N H2SO4 solution to
capture ammonia. KOH solutions are more efficient than NaOH solutions for capturing CO2
(Cook et al., 1994). KOH has a water solubility of 107 g/100 mL, compared to NaOH with
a water solubility of 50 g/lOOmL, thus allowing more CO2 to be trapped as soluble carbonates.
An empty sealed jar was kept between the two traps to prevent overflow from the alkaline to the
acidic solution. The CO2 and NH3 traps were kept outside the incubator at room temperature,
while the VOC trap was placed directly after the exit port of a digester at the incubator
temperature. The laboratory setup is shown in Figure 3-2.
Concentrations of O2, CO2, CH4, and N2 were measured occasionally in the exit gas
stream, before and after the CO2 trap. These measurements were made primarily to adjust the air
flow rate according to the 15% O2 content minimum level and were especially important during
the first 10 days after initiation of a run, since this was the period of highest O2 consumption.
These measurements also aided in checking when the carbon dioxide trap became saturated if
CO2 was detected in the gas stream after the alkaline trap.
When CO2 production rates decreased and stabilized at essentially zero, the digester was
opened to check whether this rate reduction was due to a moisture limitation. If excessive drying
had occurred, moisture was uniformly added to achieve moisture levels of at least 50% wet
weight and digester operation was continued. Drying of yard wastes in particular was observed
even though moisture was added to the incoming air. This was attributed to the limited
production of water due to the low biological activity and to excessive amounts of air supplied
for this level of activity. Composting periods varied from 60 days for food wastes to 220 days
for mixed paper. The experiment was terminated when carbon dioxide production rates dropped
below approximately 0.5 g CO2 (as C)/dry kg/day and after ensuring that this was not due to a
moisture limitation.
Leachate was observed to accumulate at the bottom of the digesters during the process.
Analysis of leachate did not take place; however, all leachate was collected and dried at 75°C
along with solid matter remaining at the end of an experiment. Though volatilization of low-
volatility organic compounds might occur at this temperature, most degradable soluble
compounds were expected to have been degraded; therefore, the soluble organic matter
remaining was assumed to have been primarily of humic and fulvic acid origin.
Digesters were thoroughly cleaned with soap and hot water at the end of each run.
3.1.4 Gas Composition Analysis
The O2, CO2, CH4, H2, and N2 contents of the exit gas stream were measured by
withdrawing 0.4 mL gas from a septum connected to the tubing at the outlet port of each digester
once or twice per day during the first 10 days of each run. In addition, sampling from a septum
placed between the CO2 and NH3 traps aided in checking for breakthrough of CO2 from the
alkaline trap. Sampling was done with a 0.5-mL gas-tight Hamilton syringe. Gas samples were
injected to a Varian 3300 gas chromatograph equipped with a Molesieve 45/60 4' x 1/8" column,
-------
Section 3.0
Methods
Air-tight removable
lid with soft plastic
gasket
Teflon tube
VOC trap (Orbo™-32)
O2, CO2, CH4, H2
sampling septa
Vent
>
50 cm
t
A
45 cm
. \J wi i i
^ v
Valve
x
\
/ \ x
\
N Waste apd
aluminum packing
/ material \
\ /
\
/
\
[
/ x
/
'
T
4>
^
25
Air ni imn nn^rfltinn 3t
/AM pUlllp U[Jd CUll ly CU
positive pressure
I
Tygon tube
r
N
«
,
n
Ambient air ' /
/ /
/
- Air plenum
\
^
1 1 II
m ' /fx
/ A \ -*— / 1, \
— h^ ^ T
CO2trap Empty NH3trap
solution flask solution
(750 ml; (500 ml;
5 N KOH) 1 N H2SO4)
•* Stainless steel
digester
Air introduction port
/
/ Small VOC
/ filter
/ ^
Vxl
l/N
Valve
* »
cm Ijr1 '\
BFIowmeter
/ \
/ \ Humidifier
L
•
V- Ambient air C
*\ (500 ml; 5 r
/ \
(10 L distilled
water at
^O2 filter incubator
si KOH) temperature)
introduction Main VOC'
filter
Figure 3-2. Digester setup (not to scale).
3-7
-------
Section 3.0
Methods
a Hayesep N 80/100 6' x 1/8" column, and a thermal conductivity detector (TCD). Column
temperature programming was 30 °C for 2 min followed by an increase to 75 °C at a rate of
2 °C/min. Standard mixtures of N2, H2, CH4, and CO2, provided by Gastech (IL, USA) and
Liquid Carbonic Specialty Gases (IL, USA), were used to calibrate the gas chromatograph. In
addition, ambient air was used to calibrate nitrogen and oxygen concentrations after correcting
for local atmospheric temperature and pressure.
3.1.5 Carbon Dioxide Mass Loadings
The cumulative mass of captured carbon dioxide (expressed as total carbon) was
measured periodically by removing 3 mL of the alkaline trap solution, diluting it with 30 mL of
deionized water and performing two titrations. The total alkalinity (ALK) (pH 4.3) and
phenolphthalein alkalinity (pALK) (pH 8.3) were measured for the same sample using an Orion
pH meter. ALK and pALK are given by the formulas 2[CO3=]+[HCO3"] + [OH"] - [H+] and
[CO3=] + [OH"] - [H2CO3] - [H+], respectively (Stumm and Morgan, 1981). The carbonate
species concentration present in the solution due to the dissolution of carbon dioxide is
[CO3=]+[HCO3"]+[H2CO3] = Alk - pAlk. The mass of C-CO2 captured in the solution per unit of
dry substrate was calculated using Equation 3-1:
C-C02(t) =.
(V43.83)Nadd
V,
sample
•xl2xVtrap-C-C02control(t)
(3-1)
Dry mass
where
C-C02(t) =
' 43-83
Nacid
V,
V,
sample
trap
12
r rn
^ ^W2 control(t)
Dry mass
g of C-CO2 present in the alkaline trap per dry kg of starting material, at
timet
= titrant volume (mL) required to decrease the pH of the solution from
8.3 to 4.3
= normality of the titrant (eq/L)
= amount of alkaline solution to be titrated, mL
= recorded volume of alkaline solution at the time a measurement was
made, L
= atomic weight of C in g/mol
= amount of C-CO2 captured during operation of an empty vessel
(control) for the corresponding time t
= initial dry mass of substrate placed in the digester prior to the initiation
of a run, kg
The normality (Nacid) of the acid used for titration was standardized at 0.245 N H2SO4, based on
Sawyer and McCarty (1978).
To check on the accuracy of the CO2 measurements, 6.907 g of CO2 were passed through
the 5 N KOH trap by bubbling a 10.1% CO2 molar concentration standard gas (Scott Specialty
Gases, IL) for 202.5 minutes at 189.5 actual cnrVmin. The mass recovered by the trap was
3-8
-------
Section 3.0 Methods
6.81 g, and the bias was calculated to be 1.4% (U.S. EPA, 1997), as shown in Table 3-1. The
EPA audit report is presented in Appendix B.
Due to condensation of moisture in the exit gas stream, the CO2 trap solution volume was
observed to increase during the course of the experiment. For this reason, the exact solution
volume (Vtrap) would be measured each time a CO2 concentration measurement was performed to
accurately calculate the mass of C-CO2 present in the jar. The alkaline trap was replenished
when the C-CO2 concentration in the solution was higher than approximately 10 g C-CO/L. At
values less than 8 g C-CO2, CO2 losses were negligible, as was shown by the use of backup CO2
traps during preliminary runs. As the concentration of captured C-CO2 increased above 10 g/L,
CO2 losses would increase exponentially until the trap would become completely saturated at
concentrations of approximately 20 g C-CO2/L. An equation was developed to describe the C-
CO2 losses as a function of the C-CO2 mass contained in the trap at any time. The estimated
losses were added to the measured C-CO2 amounts during analysis when the trap was semi-
saturated. The coefficient of variation (or precision) of the CO2 measurements was 0.7% based
on triplicate samples collected at the same time.
3.1.6 Ammonia Mass Loadings
A sample of 1.5 mL of the NH3 trap solution was periodically removed and mixed with
10 mL deionized water. Dissolved ammonia was quantified using the preliminary
distillation/titration method based on Clesceri et al., (1989). Solubility of NH3 in the KOH trap
for CO2 was deemed negligible because of the low solubility and the lack of any uptake once
saturation was achieved.
Amines are expected to be captured by the acidic solution. According to Nakasaki et al.
(1998), who composted dog food under continuous thermophilic temperatures, total amine
concentrations in the ammonia trap were less than 1/1,000 of the ammonia concentration.
Therefore, the interference of amines during ammonia quantification can be considered
insignificant.
In preliminary testing during the audit by EPA (Appendix B), 3.256 mg of NH3 were
passed through the 1 N H2SO4 trap by bubbling a 26.9-ppm standard ammonia concentration gas
for 946 min at 187.1 acm3/min. The mass recovered by the trap was 3.073 mg, and the bias was
calculated to be 2.4% (U.S. EPA, 1997), as shown in Table 3-1. The coefficient of variation (or
precision) of the NH3 measurements was 4.6%, based on triplicate samples. A methyl orange
indicator was added to the acidic trap to indicate the need to replenish a trap by color change
(Clesceri et al., 1989).
3.1.7 VOC Identification and Mass Loadings
The VOC activated coconut charcoal traps were removed periodically and a new VOC
trap was installed immediately after removal of a trap. The charcoal trap was extracted with 1.5
mL high-purity CS2 (Aldrich, Milwaukee, WI) in 2-mL vials using a sample agitator
(SUPELCO), for 30 min (Eller, 1984). The liquid phase was analyzed using GC/MS, and VOCs
were quantified using a gas chromatograph (Varian 3600CX) equipped with a flame ionization
3-9
-------
Table 3-1. Summary of Analytical Techniques
Gas masses
Gas
concentrations
Solids
Analyte
VOCs
C02
NH3
O2gas
concentration
CO2 gas
concentration
Hot water soluble
matter d
Fats & lipids d
Cellulose d
Hemicellulosed
Lignin/humusd
Total organic
carbon
Total N
Volatile matter
Procedure
Sorption on activated coconut
charcoal, carbon disulfide
extraction, injection to GC/FID
KOH trap, titration using the total
and phenolphthalein alkalinities
H2SO4 trap, distillation, and HBO3
titration
Injection of 0.4 ml gaseous sample
to GC/TCD
Injection of 0.4 ml gaseous sample
to GC/TCD
Extraction with 1 50 ml hot water
(60 °C)
Extraction with 1 50 ml
toluene/ethanol solution
Acid digestion, HPLC
Acid digestion, HPLC
Acid digestion insoluble fraction &
ashing at 550 °C
Dichromate oxidation and back
titration with FeSO4 (Walkley-Black
method)
Flow injection analyzer
(colorimetric technique)
Ashing at 550 °C
Precision3
3.2%
0.7% c
4.6% c
1 .4% c
9.3% c
18.9%
32.7%
1 1 .8%
8.5%
17.0%
2.9%
12.5%
0.9%
Bias'
5.4%
1 .4% c
2.4% c
0.5% c
7.6% c
NM
NM
4.4%
NM
2.1%
NM
NM
NM
Basic reference
Partially based on Eller (1984)
Based on Zibilske (1994) and
Stumm and Morgan (1981)
Clesceri etal., (1989)
Clesceri etal., (1989)
Clesceri etal., (1989)
Partially based on Stevenson
(1965) and Nakasaki et al. (1994)
Partially based on Stevenson
(1965)
Pettersen et al. (1984); Barlaz
(1988)
Pettersen et al. (1984); Barlaz
(1988)
Based on Stevenson (1965) and
Effland (1977)
Nelson and Sommers (1996)
Soils & Plant Analysis Laboratory,
University of Wisconsin-Madison
Clesceri et al., (1989)
NM = Not measured.
a Precision is the coefficient of variation, which is the standard deviation divided by the average based on triplicate samples (solids analyses) and duplicate
samples (gas analyses).
b Bias is the deviation of measurements from accurately prepared standards (%).
0 Based on audit measurement performed by EPA (U.S. EPA, 1997).
d Precision values are based on averaging from results discussed in this paper.
-------
Section 3.0 Methods
detector (FID). The GC was equipped with a 60 m x 0.32 mm SPB-5 capillary column. The
carrier gas was helium at a flow rate of 2.0 mL/min. The temperature programming was 50 °C
for 1 min with an increase to 100 °C at 4 °C/min followed by bake out at 230 °C for 5 minutes.
The injector temperature was kept at 180 °C, while the FID temperature was kept at 300 °C. A
Varian 8200CX autosampler unit was used to inject 1.5 • L of the CS2 solution. The GC
could not detect concentrations smaller than approximately 300 ppb under the aforementioned
conditions.
A standard mixture of 25 VOCs was obtained from SUPELCO in methanol and was used
for calibration purposes. Appropriate dilutions were made in CS2 to achieve desired
concentrations, as set during calibration. A four-point calibration was done using the external
standard technique; the calibration equation was verified prior to each analysis of a batch of
samples. Calibration was repeated using newly prepared standards if, during the verification
step, the deviation exceeded 5% for at least one compound.
A solvent injection was made after every five samples to check the presence of residual
peaks. If such peaks were present, the last two runs were reinjected. The VOC traps each
contained a breakthrough section, which was analyzed to check whether breakthrough occurred.
Between 5 and 10 VOC charcoal traps were used sequentially during each experimental
run to characterize the production rate of the studied VOCs over the course of the composting
process. The total mass of captured VOCs in each trap was calculated from the volume of the
extracting agent and the measured VOC concentrations in the extraction vial.
GC/MS analyses were performed using the same temperature conditions as the GC with a
30 m x 0.25 mm HP-5MS capillary column on a Hewlett Packard (Avondale, PA) 6890 GC
equipped with a 5972 A MS (electron impact source and quadrapole analyzer). The carrier gas
was helium at a flow rate of 1 mL/min regulated by an automatic flow controller. The injection
was made in the splitless mode with a purge-on time of 1.2 min. The injection volume was
1 mL. The scan frequency was 1.5 times/s and the mass range scanned was 50 amu to 550 amu.
The identification library contained 75,000 entries and originated from the National Bureau of
Standards.
3.1.8 VOC Recovery Tests
A VOC recovery test was performed by spiking 428.5 mg of ethylbenzene (placed in a
glass vial) in an empty digester and heating it at 55 °C until no more ethylbenzene was detected
in the gaseous emissions. Of the spiked ethylbenzene, 94.2% was recovered in the gaseous
phase, while a 15.95% breakthrough of that compound was observed in one of the charcoal traps
used during the recovery run. In addition, charcoal traps were spiked directly with 1.2 mg of the
combined 25 VOCs present in the standard solution, which was a common level found during
the experimental runs involving unspiked MSW components. Spiking was done by adding a
known volume of the standard solution directly to the primary section of the Orbo-32 trap, using
a 50-» L Hamilton airtight syringe. The CS2 extraction efficiencies of these compounds were
measured approximately 1 day after extraction and ranged from 15% for naphthalene to 87.3%
for toluene. These VOC recovery values were used to adjust measured amounts for
quantification of VOCs in actual runs. The primary section of a 400-mg charcoal trap can hold
3-11
-------
Section 3.0 Methods
85 to 100 mg of combined toluene, ethylbenzene, m-xylene, and o-xylene before a breakthrough
would occur.
No chemical byproducts were observed after storage of spiked traps in a freezer for up to
2 months. It was noticed, however, that 1 day or more after extraction with CS2, toluene would
increase by approximately 80% to 100% at all concentration levels. This twofold increase of
toluene was also observed in standard solutions prepared in CS2, which contained no charcoal.
No such increase was observed in the original VOC standards that were supplied in methanol by
SUPELCO. No corresponding change of any of the other 25 VOCs was observed. This increase
of toluene is probably attributed to a breakdown of some of the other chemical compounds
present in the mixture or interactions of the compounds with the CS2. It is noted that all samples
were analyzed for VOCs within 12 hours of extraction, so this increase for toluene would be
unimportant.
3.1.9 Control Run
A control run was performed using an empty digester with Al Packing (one of the five
digesters) that was operated for a period of 48 days concurrently with digesters containing MSW
and under the same conditions as runs with substrates. During that time, a total of 0.64 g CO2 (as
C) was captured in the alkaline trap, partly because of breakthrough from the ambient air CO2
filtration trap installed prior to the digesters. From this amount, approximately 0.12 g CO2 (as C)
were a result of the preparation process during which atmospheric air CO2 was rapidly dissolved
despite efforts to minimize air contact. The CO2 amount found in the control run was subtracted
from the total carbon dioxide emitted during actual runs. It is noted that typical runs would emit
approximately 100±15 g CO2 (as C) making the CO2 mass emitted from the control run
relatively insignificant. No significant ammonia was detected during the control run.
The combined mass of 2 of 12 targeted VOCs (toluene and ethylbenzene) identified in
the control run was 23.6 mg, which was very small compared to the amounts of these VOCs
emitted from digesters with substrates. This amount was subtracted from the corresponding
masses of these two VOCs emitted during the runs. To further check cross air contamination,
two Orbo-32 traps were placed in sequence after the ambient air large carbon filter for a period
of 80 days. This was similar to the setup shown in Figure 3-1, without the digester. Negligible
VOCs were measured, indicating that the primary filtration system functioned efficiently. The
small amounts of VOCs found in the control run were therefore a result of their original presence
in the digester, rather than ambient air cross contamination, probably due to residues remaining
even after cleaning from a previous run.
The breakthrough sections of the small Orbo-32 filters placed prior to the digesters were
always analyzed during the runs. No detectable levels of VOCs were ever measured in these
sections. If there had been, the filter tube would have been replaced immediately with a new one
before continuation of the run.
3-12
-------
Section 3.0 Methods
3.1.10 Solids Measurements
Solids were sampled at the initiation and at the end of each run. Initial solid matter was
sampled using a riffler, which removed approximately 10% of the starting material that had been
collected; the sample was stored in the cold room.
At the end of each run, most of the solid material and any leachate remaining in the
digester was collected, mixed, and dried at 75 °C until constant weight. Small amounts of
residual solids remained on the sides of the digester or attached in the packing material as these
materials were not easy to remove completely. The amount of each material at the end of a run
would vary according to the mineralization extent and initial weight of material put in the
digester. Food waste runs had approximately 30 g of dry material remaining in the digester,
which was the smallest amount of finished material from all runs. All runs containing mixed
paper had the largest amounts of dry material remaining at the end of a run.
All of the dried materials were ground using a Wiley knife mill with a 2-mm screen. The
ground samples either of a starting material or of a finished material from a specific run were
each randomly divided and placed in three mason jars and stored in a freezer until analysis.
Samples from each mason jar were randomly collected after tumbling the jar. Prior to analysis,
the ground solids were dried again to constant weight at 75 °C.
A 0.1-g to 1-g sample was used for analysis of total organic carbon (TOC) using a
dichromate oxidation followed by titration (Nelson and Sommers, 1996). The detection limit
was 0.05 mg C/mg sample. Total nitrogen analysis was performed using a flow injection analyzer
(QuickChem 8000, Lachet, Milwaukee, WI) and the QuickChem method 13-107-06-2-D. The
detection limit for total nitrogen was 0.01 mg N/mg sample. Volatile matter was measured by
ignition at 550 °C (Clesceri et al., 1989) using duplicate samples, while moisture content was
measured by drying to constant weight at 75 °C.
A sequential extraction technique was used to quantify five major classes of solid organic
matter: hot water soluble matter (HWSM), fats/lipids, cellulose, hemicellulose, and
lignin/humus. A similar sequential extraction technique was suggested by Stevenson (1965) for
soil organic matter and was followed by Nakasaki et al. (1994) for MSW. A precisely weighed
ground and dried solid sample, close to 1 g, was placed in a Gooch crucible with a !-• m filter at
the bottom. This was extracted under vacuum using 150 mL of 60°C to 65°C distilled water by
adding the water slowly with occasional manual stirring. It was observed that 150 mL of water
would produce a clear filtrate for all of the substrates. Crucibles were then dried at 75°C to
constant weight and weighed. The dry weight HWSM was recorded as the weight difference
between crucible weights prior to and after extraction. According to Tenney and Walksman,
(1929), HWSM contains starches, pectins, tannins, and uric acid; according to Inoko et al.
(1979), it also contains amino acids, pigments, and some water-soluble proteins.
A second extraction of the same sample under vacuum was performed, using 150 mL 2:1
toluene:ethanol solution. The extracted substances were measured by the dry weight difference
prior to and after extraction. The fraction extracted by the toluene:ethanol solution contained
fats, lipids, waxes, oils, resins, and some tannins (Tenney and Walksman, 1929; Stevenson,
3-13
-------
Section 3.0 Methods
1965; Effland, 1977) and is referred to as fats/lipids. The compounds present in the HWSM
group and the fats/lipids group were not individually identified during this experiment.
Cellulose and hemicellulose in the residue were measured by acid hydrolysis followed by
high performance liquid chromatography (HPLC) analysis, based on Pettersen et al. (1984). A
Beckman System Gold HPLC unit was used, equipped with an AMINEX HPX-87H 300 x
7.8 mm ion exchange column kept at 30°C. The mobile phase was a 0.01 N H2SO4 at a flow of
0.60 mL/min. Detection of monosaccharides was done with a Beckman 156 refractive index
detector. Cellulose was quantified as D-glucose, while hemicellulose was quantified as
combined D-xylose, D-mannose, L-arabinose, and D-galactose. Although small amounts of
D-glucose may originate from certain types of hemicelluloses (Laver and Wilson, 1993), it is
common practice to measure hemicellulose only as the sum of the sugars mentioned above
(Laver and Wilson, 1993; Michel et al., 1993). Standards for the five monosaccharides were
obtained from Aldrich Chemical Company, Inc. (Milwaukee, WI), and injections were done
using the external standard technique. For calculations of bias, cellulose and lignin standards
were used. High purified cellulose fibers (Type 100) were obtained from Sigma Chemical Co.
(St. Louis, MO), while a standard 32.8% lignin content (dry weight) wood was obtained from the
Forest Products Laboratory (Madison, WI).
The acid-insoluble fraction contains ash and lignin/humus (Effland, 1977). Ashing of the
acid-insoluble fraction was applied to determine the lignin/humus fraction. Lignin/humus is
expected to contain lignin, humic compounds (e.g., humic acids and humin), and acid-insoluble
proteins (Tenney and Walksman, 1929; Inoko et al., 1979). Fulvic acids are expected to be
solubilized during the 72% acid extraction step, since they are soluble in all pH ranges. It is also
noted that 3% to 5% of hardwood lignin is acid-soluble, while negligible softwood lignin
dissolves (Effland, 1977). In addition, degradation of lignin may result in some additional lignin
dissolved during the acid digestion step (Effland, 1977). Therefore, some caution should be
applied in interpreting lignin/humus fractions, especially from composted substrates. The
sequential solids analytical procedure followed is shown in Figure 3-3.
The accuracy of the solids chemical composition results was checked against the volatile
solids content. The sum of the five analyzed organic classes expressed on a dry weight basis
should be approximately equal to the volatile solids content expressed on a dry weight basis.
Based on results presented in this paper, the average ratio of the sum of the dry weights of the
five chemical components divided over the volatile solids content (dry weight) was 95.0% with a
coefficient of variation of 10.9%. Ratios in the range of 70% to 75% were observed for
composted yard wastes indicating that a class of organic compounds was probably not identified
with the techniques used. This class is speculated to be the fulvic acids, since these compounds
are soluble in all pH ranges and would therefore be diluted in the acid hydrolysis step during
cellulose determination.
The HPLC method used did not measure these types of compounds. Humic acids are
soluble only under alkaline pHs and humin is insoluble in all pH ranges; therefore, they are both
expected to be part of only the acid-insoluble fraction.
3-14
-------
Section 3.0
Methods
0.1 -1 g
Dried and ground solid
sample ~~
150 ml Distilled water (at
60-65 °C)
Soluble
150 ml
Toluene:ethanol
(2:1)
Acid digestion with 72%
sulfuric acid
Insoluble
Soluble
Ashing
Total organic carbon
(oxidation, titration)
Total nitrogen
(Flow injection analyzer)
Water-soluble sugars, starch, alcohols,
amino acids, fatty acids, water-soluble
proteins, soluble ashes (gravimetric
determination)
Lipids, waxes, resins, tannins, part
of proteins and part of fulvic acids
(gravimetric determination)
Cellulose (measured as
glucose in HPLC)
Hemicellulose (measured as the sum of
xylose, mannose, arabinose and galactose
in HPLC)
Lignin, acid-insoluble proteins,
humic acids, humin
Noncombustible
Ash
Figure 3-3. Sequential procedure for analysis of solid substrates (HWSM: hot water
soluble matter; F&L: fats and lipids; SGR: polysaccharides; LGN: lignin/humus).
3-15
-------
Section 3.0 Methods
3.1.11 Dry Weight Reduction Calculations
The average dry weight of a component or mixture initially placed in a digester was
calculated by measuring the wet weights and moisture contents of the individual substances
contained in a component or mixture. At the end of a run, due to the incomplete recovery of the
solids remaining in the digester, the final dry weight was calculated based on the fact that ash
remains constant during a run. This technique requires measurement of only the volatile solids
content of sampled solids as opposed to measurement of the dry weight of all the solids in the
digester. This technique is also useful for dry weight calculations during the course of a run. By
knowing the initial dry weight and the average initial volatile solids content of materials placed
in a digester, the dry weight remaining in the digester at any time t is given by Equation 3-2:
(1 - VS0) • DW0
DWt = - - (3-2)
1 1 - VSt
where
DWt = dry weight (in g) in the digester at time t
VSt = volatile solids content (dry weight) of the solids in the digester at time t
DW0 = initial dry weight of solids placed in the digester prior to the initiation of a run
VS0 = initial average volatile solids content of the MSW components placed in the
digester before initiation of a run.
3.1.12 Precision and Bias
Table 3-1 summarizes the analytical techniques discussed in this report and includes the
average precision and average bias values recorded for all measurements. Precision is the
coefficient of variation1 from replicate samples, while bias refers to the deviation of
measurements from accurately prepared standards. Bias is interchangeably used with the term
recovery efficiency.
No bias measurements were made for the HWSM group, fats and lipids, and
hemicelluloses because no surrogate compounds were available for each of these groups.
3.2 Initial Testing of Laboratory Setup and Methods
Ten runs were done to verify the technique and to determine the reproducibility of the
measurements. Runs made at thermophilic temperatures were two yard waste runs (YW and
YWns), the first being seeded and the second unseeded; one unseeded yard waste run (YWh)
prepared from a grass clippings batch with a higher VS content than the grass used in the YW
and YWns runs; a run with seeded mixed paper (MXP); one unseeded food waste (FWns) and one
seeded food waste run (FW); and a run with seed only. In addition, two replicate runs with
seeded mixed paper (MXPj and MXP0) and a run containing yard wastes (YWmeso), prepared
'Coefficient of variation (COV) is defined as the standard deviation divided by the mean.
-------
Section 3.0
Methods
from the same grass batch as the YW and YWns runs, were performed at mesophilic
temperatures.
Figure 3-4 presents the solids composition over time for the two mixed paper replicate
runs performed at mesophilic temperatures. Fats and lipids profiles are not shown in Figure 3-4
due to overlapping with the hemicellulose profile curves. Figure 3-5 shows the cumulative
percentage of initial organic carbon of a substrate emitted as carbon dioxide (as C) for 8 runs.
Figure 3-6 shows the cumulative percentage of initial nitrogen of a substrate emitted as NH3 (as
N) from 7 runs. Runs MXPj and MXP0 are not shown in the above figures, because the CO2 and
NH3 measurement techniques were still under development when those runs were performed.
No ammonia or solids measurements were performed for the YWmeso run. Table 3-2 presents the
gaseous yields of the runs as well as solids decomposition results.
s_
cu
ts
CD
E
^
ON
— O— MXP. I
HQQO/ • IV/IVD 1
1 \J\J /O
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
i i i i i i i i i i i i i
^ — — - -•
Volatile solids
_ _
• O - -%.
- •'' X~--^ Cellulose
sv ^^^ ^^
- ^O''*.::^. \ -
":::;B- 1
"":::::::o"---
'- _ Q ^ :::i -
•*•• ** Lignin/humus -
.... ib : ::° HWSM
- 8'---— ^^^=rrrJt:^-p Hemicellulose
^^^^ m
i 1 i 1 i 1 i 1 i 1 i 1 i 1 i 1 i 1 i 1 i 1 i 1 i 1 i 1 i 1 i 1 i
o 1
-50 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80
Days
Figure 3-4. Solids decomposition profiles for replicate
mixed paper runs MXPj and MXP0.
Based on Figure 3-4, volatile solids, cellulose, hemicellulose, and lignin/ humus
measurements were reproducible. A zero fats and lipids content was recorded for all samples of
the MXP0 run, while the fats and lipids contents for the MXPj run ranged from 1.1% to 5.3% dw.
The difference is attributed to the fact that the toluene:ethanol extraction preceded the hot water
extraction for the MXPj run only. Because no fats and lipids were recorded for the MXP0
3-17
-------
Section 3.0
Methods
100%
25
50
75 100 125
Days
150
175
200
Figure 3-5. Cumulative percentage of total organic carbon emitted as C-CO2 from
different substrates during composting (temperature for the MXP run
only was gradually increased from mesophilic to thermophilic during
the first 30 days).
samples, the slightly increased HWSM content of the MXP0 run compared to the MXPj run is
attributed to the fact that substances measured as fats/lipids in the MXPj run were also removed
during the water extraction of the MXP0 run.
Assuming an additive contribution of seed to the emissions of the FW and YW runs (as
calculated based on the run with seed only shown in Table 3-2), seed contributes less than 5% to
the total CO2 and NH3 emissions for both of these runs. Runs MXPT and MXP0 are replicates
run 3 months apart. Considering the seed contribution to the final state of decomposition
negligible, the YW and FW runs with and without seed are replicates as well. (Results of
statistically designed experimental runs to determine the effects of seed are presented and
discussed in the next chapter.) The coefficients of variation for the CO2 yields from the replicate
food waste and yard waste runs are 0.8% and 1.6% respectively. The coefficients of variation for
the NH3 yields from the replicate food waste and yard waste runs are 11.9% and 3.1%
respectively.
3-18
-------
Section 3.0
Methods
0
Figure 3-6. Cumulative percentage of total initial organic N emitted
as N-NH3 from different substrates during composting.
Based on the results of Table 3-2 and Figure 3-5, the effect of temperature on
decomposition of different substrates is shown. The seeded yard waste CO2 yields at
thermophilic temperatures (YW) were approximately 67% higher than the corresponding yields
at mesophilic temperatures. This indicates that thermophilic microbial cultures present in yard
wastes achieved a higher extent of degradation of the substrate compared to the mesophilic
species. Both the YW and YWns runs had approximately 60% of their carbon converted to CO2,
while the YWmeso run had approximately 33% of its initial total carbon converted to CO2.
Mixed paper composted at mesophilic temperatures had a 53% dry mass reduction
compared to a 35% dry mass reduction recorded for thermophilic temperatures. In contrast to
yard wastes, microorganisms present in the seed used in both mixed paper runs was apparently
3-19
-------
Table 3-2. Experimental Results from Selected MSW Components During Composting
(Reduction Values Calculated from the End of a Run)
£
o
*3^
o'
s
Oo
C5
Contents
Seed
MXP/
MXP0 +
MXP
YW
YWns
YWh
YW
meso
FW
FW
1 ""ns
C02
(9 C/dry
kg1)
86.0
NM
NM
153.3
217.0
221.9
265.2
121.0
368.6
364.4
NH3
(g N/dry
kg2)
1.8
NM
NM
2.0
4.4
4.6
5.2
NM
34.2
40.5
Dry mass
rdc
(%)3
19.4
53.6
53.4
35.1
46.7
47.6
53.9
NM
59.2
65.5
Crdc
(%) 4
27.4
NM
NM
37.9
63.1
62.2
66.4
NM
65.7
69.2
Fats/
lipid rdc
(%)
81.8
-85.0
ND
34.7
66.6
77.2
74.0
NM
94
82.6
HWSM
rdc (%)
-5.3
42.7
33.9
10.4
63.9
58.6
60.4
NM
53.7
45.5
Cellulose
rdc
(%)
36.2
73.2
73.7
49.6
82.1
90.7
89.8
NM
60.2
66.0
Hemicel-
lulose
rdc
(%)
23.9
66.1
76.6
71.9
96.6
99.4
100
NM
100
ND
Lignin
rdc
(%)
-7.2
-9.9
-1.2
-28.8
30.1
43.0
38.5
NM
36.9
61.6
C
closure
(%) 5
85.5
NM
NM
93.2
94.4
98.6
101.7
NM
117.9
108.5
N
closure
(%) 6
32.2
NM
NM
33.9
157.2
174.9
51.8
NM
72.7
77.5
HWSM = Hot water soluble matter.
ND = Not detected.
NM = Not measured.
ns = Nonseeded run.
rdc = Reduction mesophilic temperatures used.
CO2 yield in g C-CO2 per dry kg of mixture (seed + component).
1
2 Ammonia yield in g N-NH3 per dry kg of mixture (seed + component).
3 Dry mass reduction (%) based on combined dry masses of seed and MSW component placed initially in digester.
4 Total carbon reduction or mineralizable carbon fraction (%) based on combined dry masses of seed and MSW component placed initially in digester.
5 Carbon closure defined as: emitted C-CO2 / (total C in starting solid material including seed - total C in finished solid material).
6 N closure defined as: emitted N-NH3/(total N in starting solid material including seed + N added as nutrients-total N in finished solid material).
s-
I
-------
Section 3.0 Methods
favored by mesophilic temperatures. Mixed paper (MXP) had 35% of its initial carbon
converted to CO2. The effect of temperature on the degradability of various organic substrates
during composting has been summarized by Haug (1993).
Results show the effect of a different batch of grass in overall decomposition of yard
waste. The YWh run was prepared from grass clippings with a VS content of 89% ± 0.8% (dry
weight), while the grass used in all the other yard waste runs was 82% ± 0.5% VS (dry weight).
The CO2 yield from the YWh is approximately 20% higher than the yield from the other runs (see
Table 3-2), when expressed on an initial dry weight basis. Sixty-eight percent of the initial
organic carbon in the YWh run was mineralized to CO2. Moisture was found to temporarily limit
the CO2 production rate for the YWh run for a period of approximately 40 days, as is shown by
the flat CO2 cumulative profile for that substrate in Figure 3-5. Addition of moisture at the end
of these 40 days did result in a slight increase in CO2 yield shortly thereafter.
Food wastes had the largest carbon dioxide emissions of approximately 365 gr C-CO2 /
dry kg and the largest dry mass reduction of 65.5% among all substrates, as shown in Figure 3-5.
Approximately 75% of the initial carbon found in food waste (in both FWns and FW runs) was
converted to CO2. Although the FW and FWns runs were treated as replicates, it took
approximately 90 days for the FWns run to reach its "full" extent of decomposition compared to
56 days for the FW run. The difference in decomposition rates is attributed to a seed effect,
although seed did not appear to affect the CO2 yields.
Seed had a CO2 yield of 86 gr C/dry kg, the lowest among all components, probably
because of its partial decomposition in the MSW composting plant from which the seed was
collected. Only 23% of the initial carbon in the seed was converted to CO2.
As shown in Figure 3-6 and Table 3-2, unseeded food wastes had the largest ammonia
emissions among all substrates of approximately 40 g N-NH3/dry kg. This is an apparent result
of food waste having the largest initial nitrogen content of 6.2% (dw) among all substrates. The
emission of 19% less ammonia from seeded food wastes is attributed to the fact that the seeded
food wastes were terminated earlier than unseeded food wastes. Although carbon dioxide
production rates were relatively low for seeded food wastes at that time, ammonia production
rates were still relatively high indicating that ammonification still proceeded independent of
organic carbon mineralization. Approximately 65% of the initial N content in food wastes was
mineralized to ammonia. The rest was either not decomposed or was used for biomass
production.
Yard wastes, with an initial N content of approximately 2.0% (dw) had the next largest
ammonia emissions of approximately 4.5 gr N-NH3/dry kg. YWh had slightly larger ammonia
yields than the other runs, which can be partially explained by the fact that that run was
terminated at 110 days compared to approximately 50 days for the YW and YWns runs. For the
YW, YWns and YWh runs, 21.7%, 24.8% and 23.9% of the initial nitrogen was mineralized to
ammonia respectively.
The ammonia production from mixed paper alone (2.2 g N-NH3 / dry kg) is partially
attributed to the added nutrients. In particular, relatively high ammonia volatilization rates were
observed for mixed paper after addition of nutrient salts on day 160, indicating that added
-------
Section 3.0 Methods
nutrients were in excess and thus volatilized. It is also likely that the relatively alkaline pHs
observed at later stages of composting (de Bertoldi et al., 1983) induced volatilization of the
added ammonium ions as ammonia. Generally, the low ammonia yields from mixed
paper-regardless of nutrient addition-are expected due to the low initial N content of mixed
paper, which was 0.35% dw (including the seed). The initial N content was raised to 1.85%
(dw) after addition of the nutrient salts to mixed paper.
Seed had an initial N content similar to yard wastes (2.8% dry weight), but only 6.3% of
that nitrogen was emitted as ammonia. The corresponding ammonia yield for the seed was 1.8 g
N-NH3/dry kg. Apparently, nitrogen-containing organic compounds were less prone to
decomposition compared to the other substrates. This is attributed to the partial decomposition
of the seed prior to the laboratory composting that led to the production of metabolic byproducts
(e.g., microbial cells) that are resistant to further degradation.
Carbon mass balance closures ranged from 85.5% for the seed to 117.9% for food wastes.
The average carbon mass balance closure from all runs was 99.9% with a coefficient of variation
of 10.7%. Nitrogen mass balance closures ranged from 32.2% for the seed to 174.9% for yard
wastes. The average nitrogen mass balance closure from all runs was 85.7% with a coefficient
of variation of 67.3%. The relatively low nitrogen mass balance closure (compared to carbon)
and greater variability is attributed to: (1) possible loss of nitrogen during spiking of the nutrient
salt, (2) partial unavailability of the added nutrient to the microorganisms, (3) lack of
nitrate/nitrite measurement, (4) accumulation at the bottom of the digesters from leachate
movement, (5) volatilization of ammonia dissolved in the leachate during drying of the finished
material, and (6) relatively (compared to C/CO2) small amounts of N with the inherent resulting
variability. Saturation of the ammonia trap in the case of food wastes was also occasionally
observed.
Dry matter reductions ranged from 19.4% for the seed to 65.5% for unseeded food wastes
and were correlated to the corresponding CO2 yields. Total organic carbon reductions ranged
from 27.4% for the seed to 70.5% for the mixture of food wastes and yard wastes.
Hemicellulose was the most degradable chemical group in all substrates, ranging from
23.9% (for the seed) to 100% decomposed at the end of the runs (Table 3-2). The lower
molecular weight of hemicellulose compared to cellulose and its relative heterogeneity are the
likely reasons for its higher decomposition extent compared to cellulose (Gray et al., 1971).
Hemicellulose has also been found to be more degradable than cellulose and lignin by a variety
of fungi in soil environments (Szegi, 1988). Hemicellulose degrades rapidly at the initial stages
of mixed paper composting. The slight net increase of hemicellulose at the end of the MXPj and
MXP0 runs (as shown in Figure 3-4) is attributed to the synthesis of bacterial and fungal slimes
and gums that are known to contain hemicelluloses (Tenney and Walksman, 1929).
Cellulose was the next most degradable compound with up to 90% of it decomposed in
yard wastes. The lignin fraction was also degraded for all substrates but to a lesser extent than
cellulose and hemicellulose. This is primarily a result of the lignin resistance to microbial
attack, which has been well-documented (Crawford and Crawford, 1980; Vicuna, 1988). The
high lignin reduction for unseeded food wastes might be a result of the different structure of the
lignin fraction in that substrate, compared to the lignin fraction in mixed paper and in yard
3^22
-------
Section 3.0 Methods
wastes. Lignin in yard wastes was decomposed to the extent of 30.1% to 43%. The generation
of humic substances or microbial proteins (both accounted for in the lignin/humus fraction) to a
greater extent than lignin reduction appears to explain the net increase of the lignin/humus
fraction for mixed paper by 28.8%. A slight lignin/humus net increase was also observed for the
seed, MXPj, and MXP0 runs, probably for the same reason.
Results of the VOC identification and quantification analyses are given in Chapter 6. As
discussed there, mixed paper was the major source of various VOCs (primarily alkylated
benzenes and naphthalene) compared to yard wastes and food wastes. Seed appeared to
contribute to the emissions of seeded runs. GC/MS analysis performed for the MXPj run
identified the same types of VOCs (mostly alkanes, alkylated benzenes, terpenes) as a GC/MS
analysis performed later for the MXP run. A high volatilization rate during the first 5 days, with
a constantly decreasing trend thereafter, was observed for most VOCs during preliminary and
actual runs. VOC measurements, however, were still under development during the preliminary
runs and results must be considered preliminary.
A typical volatilization profile of eight VOCs identified and quantified in the seed is
shown in Figure 3-7. Naphthalene, toluene, and/>-isopropyltoluene were emitted in the largest
amounts from that substrate. The yield of the eight VOCs during seed composting was
8.2 mg/dry kg.
3.3 Summary
This chapter presents a laboratory method to measure CO2, NH3, and VOC emissions and
to follow solids decomposition during composting of MSW. Different runs with MSW
components were performed and the reproducibility of the materials and methods presented was
determined. Although the laboratory method was applied to the study of aerobic decomposition
of MSW and its components, it can be used for the study of aerobic decomposition of virtually
any solid organic substrate.
Results indicate that methods were reproducible for measuring CO2 and NH3 yields and
for following solids decomposition during composting. The close to 100% carbon mass balance
closures and the low biases of the CO2 and NH3 measurements verify the efficiency of the
analytical techniques for these two gases. The aggregate recovery efficiencies (or biases) during
measurement of CO2, NH3, and VOCs using the laboratory setup were 97.3%, 96.4%, and 94.6%,
respectively.
Carbon mass balance closures ranged from 85% to 117% and nitrogen mass balance
closures ranged from 32.2% to 174.9%. Precision and bias values for most of the analytical
measurements used were less than 15%.
Thermophilic temperatures resulted in 65% higher CO2 yields compared to mesophilic
temperatures using yard wastes as the substrate. When mixed paper was the substrate,
degradation at mesophilic temperatures resulted in approximately 50% more dry matter
reduction compared to thermophilic temperatures.
3-23
-------
Section 3.0
Methods
55 c
CD
CD
C/)
O)
T3
JD
E
CD
C/)
O
O
D)
3,000
2,500
2,000
1,500
1,000
500
0
-O— Tnh IPHP
-A- Ethylbenzene
-v— p/m xylene
-o— Styrene
-+-1,3,5-trimethy
-•- 1,4-dichlorobc
-•- p-isopropyltol
-T— Naphthalene
benzene
nzene
jene
20 40 60 80 100 120 140 160 180 200
Days
Figure 3-7. Cumulative production of eight VOCs during composting of seed at
thermophilic temperatures (the period between data points corresponds
to the time that one charcoal trap was continuously installed).
Food wastes produced the most CO2 and NH3 of the substrates tested, with approximately
365 g C-CO2/dry kg and 40 g N-NH3/dry kg. Yard wastes and mixed paper produced
approximately 220 and 153 g C-CO2/dry kg, respectively, and 4.5 and 2.0 g N-NH3/dry kg,
respectively. Seed emitted approximately 8.2 mg of eight VOCs per dry kg of starting material.
VOC volatilization rates followed first order kinetics and no measurable VOCs were produced
after approximately 50 days of composting at thermophilic temperatures.
3-24
-------
Section 4.0 Carbon Dioxide and Ammonia Yields
4.0 Carbon Dioxide and Ammonia Yields During
Composting
The objectives of the portion of the study presented in this chapter are to develop a tool
for predicting CO2 and NH3 yields from MSW components or mixtures of components and to
investigate whether mixing of different MSW components leads to additive gaseous yields based
on yields from the individual components.
MSW was assumed here to consist of three biologically decomposable substrates: food
wastes (FW), mixed paper (MXP), and yard wastes (YW), which are the major organic
components of MSW (Tchobanoglous et al., 1993). Although separate composting facilities
exist for residential municipal solid wastes and for yard wastes, the collective term of MSW
composting will be used to describe them both.
To investigate interactions among components during composting, a laboratory method
was developed that implemented a two-level, three-factor, full factorial experimental design.
This design is an efficient way to study effects and interactions of parameters (Berthouex and
Brown, 1994). In addition, a design based on the principles of a mixture experiment (Cornell,
1990) was used to develop a model that can estimate CO2 and NH3 yields on a per unit dry mass
basis of MSW. Twelve laboratory runs were performed using the 25-L stainless steel digesters
and methods described in the previous chapter under near optimal aerobic conditions. Carbon
dioxide and ammonia yields as well as their production rates were measured for all experimental
runs. It is noted that the same runs used in the full factorial design were also used during
analysis of the data using the principles of a mixture experiment, with one additional run.
4.1 Statistical Experimental Designs
4.1.1 Full Factorial Analysis
The full factorial design was developed at two levels and included three factors (23
factorial design). The two levels correspond to the presence or absence of a specific component
in the mixture (high and low level). The three factors correspond to the three waste components:
mixed paper waste (MXP), yard waste (YW), and food waste (FW). Eight runs were performed
for the full factorial design.
Components were mixed according to the U.S. typical MSW composition after recycling
reported by Tchobanoglous et al. (1993). Accordingly, the percentages of the components were
set at 59.3% mixed paper, 27.4% yard wastes, and 13.3% food wastes, on a wet weight basis,
which is approximately 80% mixed paper, 15.4% yard wastes, and 4.3% food wastes, on a dry
weight basis. All MSW components and their mixtures were seeded using partially composted
wastes, referred to as seed hereafter, from a MSW composting facility located near Portage, WI.
-------
Section 4.0 Carbon Dioxide and Ammonia Yields
The seed was collected from the exit end of an enclosed digester, screened, and added at a ratio
of approximately 1:10 of seed dry mass to component(s) dry mass for all runs.
The experimental responses for the full factorial design are the yields of either CO2 or
NH3 (in g C-CO2 or g N-NH3) produced from the dry mass of the corresponding component(s)
present in the digester.
To investigate the effect of seed on the overall emissions of each unseeded MSW
component, three 22 (two factors at two levels) full factorial experiments were designed. The
two factors were any of the three individual MSW components and the seed, respectively. The
high and low levels corresponded to the presence or absence either of seed or any of the MSW
components in the mixture. The three additional 22 factorial experiments are summarized in
Table 4-1 and results of the analysis are shown in Table 4-2. They required four runs in addition
to the eight runs required for the full factorial design as shown in Table 4-3. Table 4-3 also
shows two additional runs, Ywns and Fwns that are not part of the full factorial design.
The 23 factorial design model has the generic form:
Y = n + (Xp/2)xLP + (XY/2)xLY + (XF/2)xLF+ (XPY/2)xLpxLY+ (XPF/2)xLpxLF
+ (XYF/2)xLYxLF +(XPYF/2)xLpxLYxLF + e (4-1)
where
Y = experimental response (in g C-CO2 or g N-NH3)
n = experimental mean (in g C-CO2 or g N-NH3)
Xp, XY, XF, XPY,
XPF, XYP, XPYF = factor and interaction effects for mixed paper (XP), yard wastes
(XY), food wastes (XF), mixed paper/yard waste mixture (XPY),
mixed paper/food waste mixture (XPF), yard waste/food waste
mixture (X^), and mixed paper/yard waste/food waste mixture
(XPYF)
LP, LY, LF = can take values of either -1 or +1 only, representing the absence
(low level or -1) or presence (high level or +1) of each of the three
components in an MSW mixture; the three components are the
mixed paper wastes (LP), yard wastes (LY), and food wastes (LF)
e = errors (residuals) that are distributed normally with a zero mean
and a constant variance.
4-2
-------
Table 4-1. 22 Full Factorial Experimental Runs to Investigate Interactions Between
Individual MSW Components and Added Seed
Run title
CO2 response
(as g C) a
NH3 response
(as g N) a
1st 22 experiment
MXPns (0.803 dry kg) Seed (0.0856 dry kg)
Control (empty digester)
MXPns
Seed
MXP
+ ~-+
0
4.45
7.36
136.3
0
ob
0.15
1.76
2nd 22 experiment
YWns (0.1 545 dry kg) Seed (0.0240 dry kg)
Control (empty digester)
YW
1 ""ns
Seed
YW
+ ~-+
0
34.3
2.06
38.7
0
0.711
0.042
0.793
3rd 22 experiment
FWns (0.043 dry kg) Seed (0.009 dry kg)
Control (empty digester)
FW
ns
Seed
FW
+ ~-+
0
15.85
0.78
18.97
0
1.74
0.016
1.78
+ = High level, indicating presence of either an individual component or seed in the mixture.
- = Low level, indicating absence of either an individual component or seed from the mixture.
MXP, YW, FW= 3 runs with seeded mixed paper, seeded yard waste and seeded food waste, respectively.
MXPns YWns, FWns = 3 runs with unseeded mixed paper, unseeded yard wastes and unseeded food wastes, respectively.
Seed = Run containing seed only.
a As g C (or g N) produced from the corresponding amounts (in dry kg) of individual components or mixtures of component and seed.
b Nutrients were not added to the MXPns run.
Note: CO2 emissions of 0.64 g subtracted from all CO2 values (see Section 4.2). NH3 not detected.
-------
Table 4-2. Calculated Coefficients for the Three 22 Factorial Designs (based on responses shown in Table 4.1)
C02
(as g C)
Experimental mean (n)
Component (Xc/2)
Seed (Xs/2)
Component /seed
(Xcs/2)
36
33
34
31
.9
.2
.9
.2
NH3
(as
0.
0.
0.
0.
gN)
.48
.40
.48
.40
C02
(as g C)
18.8
17.7
1.63
0.60
NH3
(as
0.
0.
0.
0.
gN)
.39
.37
.03
.01
C02
(as g C)
8.90
8.51
0.98
0.59
NH3
(as g N)
0.88
0.88
0.014
0.006
Table 4-3. 23 Full Factorial Experimental Design to Investigate Main Effects and Interactions
During Composting of Three Major MSW Organic Components
Run title
Control (empty digester)
MXP
YW
YWns
MXP/YW
FW
FWns
MXP/FW
YW/FW
MXP/YW/FW
+ =
=
FW
FW/YW
MXP
MXP/FW
MXP/YW
MXP/YW/FW
YW
MXP
(at 0.803 dry YW FW
kg) (at 0.155 dry kg) (at 0.043 dry kg)
_
0.803
+
+
0 +
+
+
+ - +
+ +
+ + +
High level, indicating presence of material in the mixture.
Low level, indicating absence of material from the mixture.
CO2 response
(as g C) a
0
123.2
33.5
34.3
237.0
15.4
15.6
198.7
59.5
265.3
Seeded food waste, present at a dry weight of 0.043 dry kg, as a mixture of lettuce, carrots, cooked pasta
proportions.
A mixture of food wastes and yard wastes.
NH3 response Days to emit 50%
(as g N) a
0
1.59
0.69
0.71
0.56
1.45
1.71
0.92
2.86
0.46
, cooked meat, milk, and
CO2 yield
0
27.0
7.6
10.7
25.0
7.5
8.0
24.4
9.5
19.0
Of
raw potatoes in equal wet weight
Seeded mixed paper waste, present at a dry weight of 0.803 kg, as a mixture of office paper (OFP), old corrugated cardboard (OCC) and old newsprint (ONP),
mass ratios of 1 0'2 1'1 7 respectively
A mixture of mixed paper and food wastes.
A mixture of mixed paper and yard wastes.
A mixture of mixed paper, yard wastes and food wastes.
Seeded yard waste, present at a dry weight of 0.1545 kg, as a mixture of grass clippings and leaves set at dry mass ratios of 1.5:1
a As g C (or g N) produced from the corresponding amounts (in dry kg) of component or mixture (including seed), as
shown .
Note: CO2 emissions of 0.64 g subtracted from all CO2 values (see Section 4.2). NH3 not detected. Note also that runs YW ns and FW
text).
respectively.
at dry
ns are not part of the original full factorial design (see
-------
Section 4.0 Carbon Dioxide and Ammonia Yields
The model for the 22 factorial design used to investigate seed interaction has the form:
Y = n + (Xc/2)xLc + (Xs/2)xLs + (Xcs/2)xLcxLs + e (4-2)
where
Xc, Xs, Xcs = factor and interaction effects for each of the three MSW components
(Xc), the seed (Xs) and the interaction of component/seed (Xcs)
Lc, Ls = take values of either -1 or +1 and represent the absence (low level or
-1) or presence (high level or +1) of the component or the seed in a
mixture, where Lc corresponds to the MSW component - that is
either mixed paper, yard wastes or food wastes - and Ls corresponds
to the seed
e = errors (residuals) that are distributed normally with a zero mean and
a constant variance.
Other parameters are as defined previously.
4.1.2 Mixture Experiment
The full factorial design is based on mixtures of MSW components prepared based on a
typical U.S. MSW composition. The resulting models and the main effects and interactions may
not be applicable to MSW of a different composition. Therefore, to better describe CO2 and NH3
yields from any MSW mixture on a per dry unit mass basis, an additional design was set-up
based on the principles of a mixture experiment. The analysis used the same runs included in the
factorial analysis approach, with the addition of one run, referred to as the MSW1/3 run. In the
MSW1/3 run, all three MSW organic components were mixed in approximately equal dry weight
proportions and composted. The MSW1/3 run was considered necessary since four of the eight
runs used in the full factorial analysis contained mixed paper at percentages higher than 80% on
a dry weight basis. No seed was added to the MSW1/3 run, since yard wastes and food wastes
had been shown to contain microbial populations adequate to maintain the decomposition
process and to effectively inoculate mixed paper.
As part of the mixture experimental design analysis, the polynomial Equation (3) can be
fitted to the data:
Ykg = bpXpp + byxFY + bFxFF + bPYxFpxFY + bPFxFpxFF +
+ bYFxFYxFF + bPYFxFpxFYxFF + e (4-3)
where
Ykg = either the CO2 yield or NH3 yield of an MSW mixture expressed in g
C-CO2/dry kg or g N-NH3/dry kg of mixture, respectively
4-5
-------
Section 4.0 Carbon Dioxide and Ammonia Yields
FP, FY and FF = dry fractions of mixed paper, yard waste, and food waste,
respectively, in the mixture, with each of the FP, FY, FF values
ranging from 0 to 1 and with FP+FY+FF always equal to 1
bp, bY, bF, bPY,
bYF, bPF, bpyp = model coefficients
e = errors (residuals) that are distributed normally with a zero mean and
a constant variance.
All statistical analyses were performed with the MINITAB v!2.2 statistical package
(Minitab Inc., PA, USA).
4.2 Results and Discussion
Results of the experimental runs used in the three 22 designs are included in Table 4-1,
and results of the experimental runs used in the 23 design are included in Table 4-3. Note that
the gaseous yields shown in Tables 4-1 and 4-3 correspond to the dry masses of the materials
shown in these two tables. For example, 0.803 dry kg of seeded mixed paper (MXP) produced
136.3 g C and 1.76 g N, respectively, as shown in Table 4-1.
The CO2 emissions from the control run (empty digester) were 0.64 g (as C), measured
over a period of 48 days, and were subtracted from the CO2 yields of all experimental runs. No
significant ammonia was detected during the control run. Based on the above, the CO2 and NH3
yields for the control run were always set to 0.
4.2.1 Seed Interaction
The model shown in Equation 4-2 was fitted to the CO2 and NH3 yield data of
Table 4-1. The goal of each 22 factorial analysis was to investigate the magnitude of seed
contribution to individual component yields and the effect of the seed/component interaction.
The coefficients of Equation 4-2 are presented in Table 4-2. They are half of the values of the
estimated main and interaction effects. Since there are no degrees of freedom during model
fitting, no standard errors can be estimated for each model coefficient. The significance of each
effect should be based, therefore, on the value of the corresponding coefficient as estimated
during model fitting.
Based on Table 4-2, the seed main effects are 9.2% [(1.63/17.7) x 100] and 11.5%
[(0.98/8.51) x 100] of the yard waste and food waste main effects, respectively, when predicting
CO2 yields. The seed/component interaction effects are 3.3% [(0.60/17.7) x 100] and 6.9%
[(0.59/8.51) x 100] of the main component CO2 yield effects for yard wastes and food wastes,
respectively. This indicates that some minor additional amounts of CO2 will be produced by
seeding these two components, compared to the simple summation of the CO2 yields from the
individually composted component and individually composted seed. The fraction of the
interaction-related CO2 emissions due to additional decomposition of the component or
additional decomposition of the seed cannot be determined.
4-6
-------
Section 4.0 Carbon Dioxide and Ammonia Yields
In the case of mixed paper, seeding was necessary to initiate decomposition, since
negligible CO2 was produced from the MXPns run. This is also shown by the fact that the
interaction effects are equal to the main component effects for both the CO2 and NH3 yields (see
Table 4-2). The materials used to prepare mixed paper (office paper, cardboard, old newsprint)
do not appear to have significant amounts of microbial populations present. In the case of mixed
paper in particular, it is expected that the initial ratio of seed to substrate and the type of seed
could have affected the decomposition rates and extent. Attempts to vary seed to substrate ratios
were beyond the scope of this work. It is noted that the effect of seed to mixed paper is not a
matter of additivity of individual CO2 emissions. Seed was simply required to provide the
minimal amount of microorganisms necessary to start the decomposition process for mixed
paper.
The relatively small seed/component interaction effects for food and yard wastes (<7%
compared to the main component effect) are explained by the fact that these components are
likely to contain an indigenous microbial population capable of initiating and maintaining the
aerobic decomposition process. According to Gray et al. (1971), composting of separated waste
mixed refuse was not affected by the addition of several types of inocula due to the adequacy of
the indigenous microbial population in degrading these materials. The slightly higher CO2
seed/component interaction effect observed for food wastes than yard wastes can be explained by
the former substrate containing cooked pasta, cooked meat, and pasteurized milk. These
subcomponents account for 50% of the wet weight of food wastes and are all expected to have
negligible amounts of microbial biomass prior to composting. In addition, seed was supplied in
slightly higher ratios to food wastes (1:4.7 dry seed to dry component ratio) compared to yard
wastes (1:6.5), which can also partially explain the higher seed/component interaction in the
former substrate compared to the latter.
In the case of ammonia, the seed effect itself accounts for 8.1% and 1.6% of the yard
waste and food waste main effects, respectively. The seed/component interaction effects are
2.7% and 0.7% of the yard waste and food waste main effects, respectively. Therefore,
seed/component interaction effects for NH3 are smaller than the corresponding interaction effects
for CO2, for both yard wastes and food wastes. It is likely that a microbial population capable of
ammonification processes is originally present in the unseeded food wastes and unseeded yard
wastes. The almost negligible interaction effect of seed/food wastes (<0.7% of the main
component effect) is also explained by the relatively large ammonia emissions from food wastes
alone, compared to the ammonia emissions from other substrates.
Figure 4-1 shows the cumulative CO2 yields. Based on Figure 4-1, seeding did appear to
affect the CO2 production rates, but not yields, for food wastes. Unseeded and seeded food
wastes had similar CO2 productions rates until day 5. After day 5, unseeded food wastes had
steadily lower CO2 production rates compared to the corresponding rates of seeded food wastes.
It took 56 days for seeded food wastes to reach more or less complete degradation, as opposed to
a period of 90 days for the FWns run.
4.2.2 Calculating Gaseous Emissions from Seeded Runs
Results in the previous section indicate that the seed/component interactions for both
food wastes and yard wastes are less than 10% of the corresponding main component effects
-------
Section 4.0
Carbon Dioxide and Ammonia Yields
400
350
D)
.*:
£,300
•Q
1JJ250
'c
1T200
o
D)
150
50
0
MXP/YW/FW
N»
MXP/YW
MXP
0 20 40 60 80 100 120 140 160 180 200 220
Days
Figure 4-1. Cumulative CO2 production (as g C) per dry kg
of starting material (including seed).
when predicting CO2 and NH3 yields. The seed main effects are also relatively small (less than
10% of the main component effect). Due to the relatively small seed main effects and
seed/component interaction effects, they will both be considered insignificant. Based on that, the
gaseous yields from seeded components and seeded mixtures will be expressed as g of total C-
CO2 (or total N-NH3) emitted per dry kg of total substrate (component + seed). The same
approach is also followed for mixed paper. The cumulative CO2 (as C/dry kg) and NH3 (as
N/dry kg) emissions from all components and mixtures are given in Tables 4-3 and 4-4 and are
illustrated in Figures 4-1 and 4-2, respectively. Table 4-4 presents the fractions of each
component in the mixture and the corresponding CO2 and NH3 yields, expressed on a per dry kg
of substrate basis. This slightly differs from the expression of yields shown in Table 4- 3; the
difference simply reflects the format to which the numbers were fitted to Equation 4-1
(Table 4- 3) and Equation 4-3 (Table 4-4).
Based on the assumptions stated above, the FW and FWns runs can be treated as replicates
in terms of CO2 and NH3 yields. The average CO2 and NH3 yields from these two runs are 366.5
±2.97 g (as C)/dry kg and 37.35±4.5 g (as N)/dry kg, respectively. Similarly, the YW and the
YWns runs can be also treated as replicates with average CO2 and NH3 yields of 219.5 ± 3.46 g
(as C) / dry kg and 4.5±0.14 g (as N) / dry kg, respectively. Due to the large seed interaction
for mixed paper, the MXPns run will not be considered as replicate to the MXP run.
4-8
-------
Section 4.0
Carbon Dioxide and Ammonia Yields
D)
45
40
35
^ 30
03
*P 25
D)
CO
3?
20
15
10
5
0
FWs
20 40 60 80 100 T2040 160 180\200
MXP/YW/FW MXP/YW
Days
Figure 4-2. Cumulative NH3 production (as g N) per dry kg of
starting material (including seed).
Note: The result for MXPns was 0.0; values in parentheses indicate initial N content (in % dry
matter) for specified substrate, including nitrogen added as nutrient to the MXP, MXP/FW,
MXP/YW and MXP/YW/FW runs.
As Figure 4-1 shows, certain runs fully decomposed earlier than others, which explains
the fact that some runs were terminated at 60 days (FW) and others lasted more than 150 days
(MXP, MXP/YW). This appears to be a result of differences among hydrolysis rate constants for
the different solid fractions of each component, assuming that solids hydrolysis is the rate-
limiting step in the decomposition of all substrates, as is likely.
As illustrated in Figure 4-1, food waste is the largest producer of CO2 among all
components, producing 368 g CO2 (as g C/dry kg), when unseeded. Seed produced only 86 g
CO2 (as g C/dry kg), which was the lowest yield among all substrates. Mixtures of two
components result in CO2 yields that—on a per dry kg basis—are always between the CO2 yields
of the individual components that constitute the mixtures. The flattening of the CO2 cumulative
curve observed for the YW/FW run between days 25 and 35 is due to substrate drying. Moisture
4-9
-------
Table 4-4. Fractions of Components in Mixtures (dry mass basis) and Gaseous Yields for 12 Runs
Run
Seed
MXP
MXPns
YW
YW
1 u uns
FW
FW
1 vvns
MXP/YW
MXP/FW
FW/YW
MXP/YW/FW
MSW1/3
MXP
0
1
1
0
0
0
0
0.83
0.96
0
0.80
0.34
YW
0
0
0
1
1
0
0
0.17
0
0.78
0.155
0.37
FW
0
0
0
0
0
1
1
0
0.04
0.22
0.045
0.30
Dry seed / Dry
component
ratio
1
1:9.4
0
1:6.5
0
1:4.8
0
1:7.2
1:9.9
1:7.4
1:7.4
0
C02
(g C / dry
kg)a
86.0
153.3
5.5
217.0
221.9
364.4
368.6
246.1
236.6
301.9
265.3
266.3
NH3
(gN/dry
kg)a
1.8
2.0
0.0
4.4
4.6
34.2
40.5
0.6
1.1
14.5
0.5
6.5
Initial N content
(% dry weight) b
2.8%
1 .9% (0.35%)
0.09%
2.1%
1 .9%
5.6%
6.2%
1.5% (0.72%)
1 .9% (0.58%)
2.9%
1 .7% (0.9%)
2.6%
% of initial N
mineralized to
N-NH3C
6.3%
10.7% (56.7%)
0.0%
21 .7%
24.8%
61 .0%
65.4%
3.9% (8.4%)
5. 9% (19.1%)
50.2%
3.0% (5.6%)
25.2%
NS = No seed.
a Values represent total mass of emitted gas per dry kg of substrate (component + seed).
b Values in parentheses indicate initial total organic nitrogen content prior to addition of nutrients in the specified runs.
c Values in parentheses do not account for added nutrients in calculation of total initial N for the specified substrates.
-------
Section 4.0 Carbon Dioxide and Ammonia Yields
was added on day 35 to this run, resulting in the continuation of its decomposition. This
moisture limitation was also observed for the MXP run, between days 30 and 65, and for the
YWns run, between days 25 and 35, with moisture added in both cases thereafter. As Figure 4-1
shows, low but steady CO2 production rates were observed for the MXP, MXP/YW and
MXP/FW runs for a relatively long period of time fully decomposed, indicating that more time
might have been needed for these substrates to fully decompose. The runs were terminated,
however, because of time constraints. In all runs including mixed paper, cardboard was the only
subcomponent visually present in the digester at the termination of each run.
Figure 4-2 presents the time courses of the cumulative ammonia emissions from all runs,
except MXPns that produced zero ammonia. Table 4-4 shows the initial N contents (in % dry
matter) and the % of initial N that was mineralized to N-NH3 for all 12 runs. As shown,
ammonia yields correlate strongly with the initial N content of the substrate. Unseeded food
wastes, with an initial N content of 6.2% (dw), produced 40.5 g NH3 (as N/dry kg), which was
the highest ammonia yield among all runs. Seeded food wastes had a NH3 yield of 34.1 g N/dry
kg. The 19% lower ammonia emission from seeded food wastes is partly attributed to the fact
that the seed (which was in relatively large ratios in that run) produced minor amounts of
ammonia and was included as part of the total substrate. In addition, ammonia production rates
were still relatively high for the FW run at the time of its termination, despite the low CO2
production rates, and therefore more ammonia could have been produced if the run was stopped
at 90 days, as the FWns run, instead of 54 days.
Yard wastes were the next highest with ammonia yields of approximately 4.5 g NH3 (as
N)/dry kg. Although food wastes had an initial N content approximately three times higher than
the yard wastes, they emitted approximately 10 times more ammonia than did yard wastes. This
is likely a result of less nitrogen required for microbial biomass assimilation in the case of food
wastes compared to yard wastes, therefore resulting in a higher release of excess nitrogen as
ammonia in food wastes. Approximately 65% of the initial N of food wastes was mineralized to
ammonia, while approximately 20% to 25% of the initial N for yard wastes was converted to
ammonia.
Seed, despite its relatively high initial N content of 2.8% (dw) yielded less than half the
ammonia than did yard wastes. This is probably because nitrogen is in a less bioavailable form
in the seed compared to nitrogen forms in yard wastes and food wastes. This is likely because
the seed was already partially decomposed. Generally, the forms of nitrogen in various organic
substrates is expected to affect ammonia releases. Inoko et al. (1979) reported five forms
of nitrogen in city refuse compost, with approximately 40% being in the form of amino acids,
approximately 30% unidentifiable, and the rest in other forms.
Low ammonia emissions compared to other runs were recorded for all runs that had a
mixed paper content of higher than 80% dry matter (MXP, MXP/FW, MXP/YW,
MXP/YW/FW). It appears that N was limiting in these runs, even after the addition of nutrients
that resulted in initial N contents approximately similar to the initial N contents for yard wastes.
The relatively high ammonia yield from mixed paper alone is attributed to the addition of
nutrients after day 160. These nutrients were probably not assimilated by the microbial biomass,
due to mixed paper being at a relatively nonactive composting stage at the time of that addition.
4-11
-------
Section 4.0 Carbon Dioxide and Ammonia Yields
Based on Figure 4-2, it appears that the FW and FWns might not have yielded all of their
releasable ammonia, since ammonia production rates were higher than zero for both runs at the
time of termination. It is likely that materials containing protein were still decomposing at the
end of these runs.
4.2.3 MSW Component Interactions
' ns
Data in Table 4-3 were fitted to Equation 4-1. The unseeded YW and FW runs (YWn
and FWns) were taken as replicates to the YW and FW runs as discussed. The MXPns run was not
used during this analysis due to the negligible degradation observed for that run because of the
lack of seed. The dry weights of materials shown in Table 4-3 are not the actual dry masses used
in the digesters, but have been adjusted to equal the dry weights of each component used in the
run MXP/YW/FW, for which the results are also presented per kg of dry mass.
4.2.3.1 CO2 Yield—Main Effects and Interactions. The best reduced CO2 yield
factorial model was fitted using Minitab v!2.2, so that errors were normally distributed with a
mean of zero. The resulting model is shown in Equation 4-4:
Yco2 = 116.2 (±4.90) + 89.8 (±4.90) LP + 33.4 (±4.97) LY
+ 17.0 (±4.97) Lp + 11.7 (±4.97) LpLY (4-4)
where
Yco2 = mass of CO2 (as g C) emitted from a mixture of MSW
LP = absence (-1) or presence (±1) of mixed paper in the mixture
LY = absence (-1) or presence (±1) of yard wastes in the mixture
LF = absence (-1) or presence (±1) of food wastes in the mixture.
Levels (L) can take values of-1 or ±1 only, defining the absence or presence of a
component in the mixture. Values in parentheses represent the standard errors for the
corresponding coefficients. The adjusted R2 for Equation 4-4 is 97.7%, while residuals have a
mean of zero and appear to follow a normal distribution. From Equation 4-4, all three main
effects are considered statistically significant at a 95% confidence level. The mixed paper and
yard wastes interaction effect appears to be statistically significant at the 90% confidence level.
According to Equation 4-4, the major contributor to the CO2 yield is actually mixed
paper, at the percentages that mixed paper is present in U.S. MSW composition. This would be
expected since mixed paper accounts for 80% of the dry mass of MSW, based on the U.S. MSW
composition used in the design of the experiment. Yard wastes and food wastes produce the next
most significant amounts of CO2, while the interaction of mixed paper and yard wastes
(MXP/YW) appeared as marginally significant. The other two two-component interactions and
the three-component interaction were insignificant.
The MXP/YW interaction significance can be explained by the fact that mixing of mixed
paper with yard wastes—at the specified ratios—results in further seeding of the paper by the
microbial population present in yard wastes. This additional seeding from yard waste provides a
population able to further degrade slowly degradable components of the mixture—especially
-------
Section 4.0 Carbon Dioxide and Ammonia Yields
mixed paper. The use of yard wastes may result also in an additional supply of nitrogen to the
total mixture. Although nitrogen was added to the MXP run in the form of salts, it might still
have been in limiting amounts. Such synergistic effects would be expected. For example, Szegi
(1988) observed that lignin by itself decomposes more slowly than lignin incorporated in cell
walls. The explanation given was that the presence of easily degradable matter in cell walls
enables quicker multiplication of microorganisms, promoting the faster decomposition of lignin.
Cumulative production profiles for the seeded mixed paper (Figure 4-1) reveal that mixed
paper decomposition could have continued, albeit at slow rates, after the experiment was
terminated. Therefore, the amount and type of seed initially added to mixed paper alone might
have simply resulted in the relatively slow decomposition rates, compared to runs with
percentages of paper and other components (MXP/YW, MXP/FW). Inadequate seeding might,
therefore, have resulted in the MXP run not reaching its "full" extent of decomposition within
the time constraints under which the experiment was performed. If more time had been used for
the MXP run—with the seed as used—the MXP yield would have been higher than that recorded
in this study. To check that, higher CO2 yields from the MXP run were input to the model. It
was observed that the higher the CO2 yields from the MXP run, the lower the interaction effect
of mixed paper and yard wastes would become, until it would finally become statistically
insignificant. It, therefore, appears that the slow decomposition rates of mixed paper that did not
allow it to fully decompose finally resulted in the interaction of MXP/YW to appear as
significant.
Exclusion of the MXP/YW interaction from Equation 4-4 results in an additive model.
The adjusted R2 for the predictive model, after excluding the interaction, becomes 96%, which is
only slightly lower than the adjusted R2 value of 97.7% of Equation 4-4. The above indicates
that an additive model can also adequately describe CO2 emissions from MSW. The additive
CO2 model is shown in Equation 4-5:
Yco2= 116.2 (±6.5)+ 89.8 (±6.5) Lp +31.5 (±6.5) Ly + 18.9 (±6.5) LF (4-5)
with parameters as defined previously. It is noted that residuals in Equation 4-5 are closer to
being normally distributed compared to the distribution of the residuals from Equation 4-4.
4.2.3.2 NH^ Yield—Main Effects and Interactions. In the case of NH3 yield, the best
reduced fit model is given as Equation 4-6:
Y^ = 1.06(±0.077) - 0.18(±0.077) LP+ 0.06(±0.079) LY + 0.37(±0.079) LF
- 0.43(±0.079) LPLY - 0.56(±0.079) LPLF (4-6)
where
YNFB = mass of NH3 (as g N) emitted from a mixture of MSW
LP = absence (-1) or presence (+1) of mixed paper in the mixture
LY = absence (-1) or presence (+1) of yard wastes in the mixture
LF = absence (-1) or presence (+1) of food wastes in the mixture.
4-13
-------
Section 4.0 Carbon Dioxide and Ammonia Yields
The adjusted R2 for Equation 4-6 is 91.6%. Values in parentheses represent the standard
errors for the corresponding coefficients, while residuals for the model are normally distributed
with a mean of zero and constant variance. As Equation 4-6 shows, the most important main
effect is due to food wastes. The interactions of mixed paper and food waste (MXP/FW) and
mixed paper and yard wastes (MXP/YW) are the only significant interactions and have negative
coefficients. This means that inclusion of mixed paper in the MSW mixture reduces overall
ammonia emissions. It is interesting to note that both the mixed paper and yard waste main
effect are statistically insignificant (at the 95% confidence level). However to omit the main
effect of either mixed paper or yard wastes from the model would be statistically wrong when
interactions that involve these components appear statistically significant. The yard waste and
food waste interaction and the three-component interaction were insignificant.
In the case of the NH3 model (Equation 4-6), more interactions were significant
compared to the CO2 model (Equation 4-4). The high ammonia yields predicted for food wastes
are a result of their high initial N content, as discussed. The addition of mixed paper in a MSW
mixture reduces ammonia yields due to the relatively low initial N content of mixed paper
despite the addition of nutrients. This addition of nutrients in mixed paper runs resulted in initial
N contents approximately similar to the initial N contents of yard wastes, for which ammonia
yields were much higher than that from the mixed paper runs. The lower ammonia yields from
mixed paper can therefore be explained by an assimilation of all the nitrogen added to or present
in mixed paper to form microbial biomass. It can also be explained by the fact that some of the
externally added nitrogen (probably nitrates) might have become only partially available to the
existing microbial population and therefore did not enter the N cycle. Leaching of the
ammonium nitrate salt to the bottom of the digester is also suspected and could contribute to a
probable low initial N content achieved for the mixed paper runs. Nitrogen losses could also
have occurred during the addition of NH4NO3 to the substrate, prior to the initiation of a run.
The nitrogen losses are also indicated by the less than 35% nitrogen mass balance closures for all
runs involving mixed paper, namely, MXP, MXP/FW, MXP/YW, MXP/FW/YW.
4.2.3.4 Carbon Dioxide Production Rate Interactions. Mixtures of components
appear to affect production rates. For example, the MXP/YW run had distinctively higher
production rates than the MXP alone, as shown by the steeper CO2 curve slope of the former
compared to the latter during the first 10 days of the experiment (Figure 4-1). Because substrate
utilization rates are usually a function of the substrate and biomass concentrations, the higher
CO2 production rates of MXP/YW compared to MXP appear to be a result of the presence of a
more readily decomposable substrate (yard wastes), compared to mixed paper, and of a higher
microbial population concentration compared to MXP alone.
Using the full factorial experimental design shown in Table 4-3, a statistical analysis was
performed to investigate the main effects of components and interactions on the rates of CO2
production. The response chosen was the number of days required for each run to emit 50% of
its corresponding CO2 yield, as shown in Table 4-3. The resulting equation is:
TC02 50% = 15.3(±0.5) + 8.6(±0.5) LP+ 0.4(±0.5) LY - 0.06(±0.5) LF - 2.3(±0.5) LPLY
-2.1(±0.5)LPLF-1.4(±0.5)LYLF (4-7)
where
-------
Section 4.0 Carbon Dioxide and Ammonia Yields
TCCG 50% = days required for the MSW mixture/component to emit 50% of its
carbon dioxide yield
LP = absence (-1) or presence (+1) of mixed paper in the mixture
LY = absence (-1) or presence (+1) of yard wastes in the mixture
LF = absence (-1) or presence (+1) of food wastes in the mixture.
Values in parentheses represent the standard errors for the corresponding coefficients,
while residuals for the model are normally distributed with a mean of zero. Equation 4-7 shows
that presence of mixed paper in an MSW mixture (Lp=+l) significantly increases overall
composting times compared to all other main and interaction effects. The main effects of both
yard wastes and food wastes appeared statistically insignificant at the 95% confidence level.
However, the two-component interactions were statistically significant. Therefore, as yard
wastes or food wastes are introduced in a mixture containing mixed paper (LY=+1 or LF=+1),
composting times are reduced. In particular, the MXP/YW interaction effect results in the
largest composting time decrease compared to the other two two-component interactions. The
seeding of mixed paper with yard wastes appears to be the reason for the faster decomposition
rates of the MXP/YW run, as discussed. In addition, mixing of yard wastes and food wastes also
reduces composting times due to "seeding" of food wastes with the microbial population present
in yard wastes as well as the additional supply of readily degradable substrate found in yard
wastes. The negative value of the food waste coefficient, as opposed to the positive coefficients
of mixed paper and yard wastes, indicates that food wastes degrade faster than either mixed
paper or yard wastes. The three-component interaction was insignificant.
Solids hydrolysis appears to be the major rate-limiting step during composting for all
substrates—except mixed paper—since nutrients, seeding, moisture, and oxygen supply were
kept at near optimum values. Therefore, the initial steep slope in all cumulative CO2 production
curves (Figure 4-1) is a result of the breakdown of the readily hydrolyzable solid fraction
contained in any of the substrates. This fraction is expected to contain starches and other water-
soluble substrates (e.g., monosaccharides and fatty acids). After readily degradable material is
decomposed, mineralization rates become a function of the concentration of the remaining
slowly hydrolyzable solid fraction and the corresponding hydrolysis rates for that fraction. The
slowly degradable fraction in all substrates is therefore responsible for the flattening of the
cumulative CO2 curves shown in Figure 4-1, as composting approaches the end. For mixed
paper alone, inadequate seeding and nutrient limitation might be additional decomposition rate-
limiting factors.
Production rate curves for CO2 and NH3 for the food waste and yard waste runs are
shown as Figures 4-3 and 4-4 respectively. The highest CO2 production rates occurred during
the first 10 days for both substrates. Figures 4-3 and 4.4 show the characteristic lag between
periods of highest CO2 and NH3 production, with NH3 production lagging CO2 production by 10
to 15 days. This NH3 lag appears to be a result of the pH increase, commonly occurring at later
stages of decomposition during MSW composting. This pH increase is a result of the
consumption of organic acids that are normally produced during the active composting stage
(Gray et al., 1971). pH values measured directly from the leachate contained at the bottom of the
digester were 8.2 at the end of the FW/YW run, 8.7 at the end of the YW run, and 8.8 at the end
of the YWns run. Ammonia is expected to be partially volatilized at such pH values and
4-15
-------
Section 4.0
Carbon Dioxide and Ammonia Yields
(C-CO2) FW
(C-CO2) FWns"
(NH3) FW -
(NH.) FW
2500
2000
CD
T3
1500^
- 1000
- 500
O)
E
25 50 75
Days from initiation of composting
Figure 4-3. CO2 and NH3 daily production rates for the FW and FWns runs.
4-16
-------
Section 4.0
Carbon Dioxide and Ammonia Yields
2500
25 50 75 100
Days from initiation of composting
Figure 4-4. CO2 and NH3 daily production rates for the YW and YWns runs.
4-17
-------
Section 4.0 Carbon Dioxide and Ammonia Yields
thermophilic temperatures. Lower production rates for CO2 and much lower rates for NH3 were
obtained for yard wastes than for food wastes.
4.2.4 Mixture Experiment
Equations 4-4, 4-5, and 4-6 can describe yields of CO2 (C) and NH3 (N) (in gr) from a
mixture of MSW depending on the relative percentages of mixed paper, yard wastes, and food
wastes on which the experiment was based. These equations are useful to investigate main
effects and interactions; however, they are based on the presence or absence of a given
component and are not necessarily applicable when estimating yields from MSW of various
compositions, which can be different from the typical U.S. MSW composition used in this
experiment.
Based on the principles of mixture experimental design (Cornell, 1990; Draper and
Smith, 1998), the results given in Table 4-3 were supplemented and then used to develop an
equation to describe gaseous mass loadings per dry kg of MSW of various compositions. For
this reason, run MSW1/3 (see Table 4-4) was also included in the data fitting. Similar to the
factorial design approach, the FWns and YWns runs were considered replicates to the FW and YW
runs, respectively. The MXPns run was not included during model fitting since negligible
decomposition was observed for that run due to lack of seed.
Equation 4-3 was fitted to the data in Table 4-4 separately for the CO2 and NH3 yields.
Table 4-5 presents the regression analysis results for the first-, second-, and third-order models.
The second- and first-order models were fitted to the data by neglecting the third-order and
seond-order terms, respectively, from equation (4-3). Note that the term "order" is used in
respect to the dependent variables (FP, FY, FF). The order of the equations with respect to the
model coefficients is 1. According to Draper and Smith (1998), the coefficient of determination
R2 is not defined for equations that have no intercept, as is the case for Equation 4-3. However,
to allow comparison with Equations 4-4 and 4-6, the adjusted R2 values were calculated based on
the definition given in most statistics texts (Draper and Smith, 1998). These R2 values are given
in Table 4-5.
Using the third-order CO2 predictive model, all model coefficients appear to be
significant. However, values higher than 650 g C-CO2 /dry kg are predicted from some mixtures
of food wastes and mixed paper. Although the 3rd order model results in a higher adjusted R2
than the other models, there are no experimental runs to support the predicted high yields. In
addition, residuals for this model were not normally distributed and the standard errors for the
second-order coefficients and the third-order coefficient were relatively large. The second-order
model is inadequate, also due to the high standard errors for the second-order coefficients. The
first-order model has a higher adjusted R2 than the second-order model, while all residuals are
normally distributed. Therefore the first-order model is considered the most adequate for
describing CO2 yields, based on the experimental runs mentioned in this paper.
In the case of NH3, the second- and third-order models have large standard errors for their
second- and third-order coefficients. In addition, the mixed paper coefficient (bp) is statistically
insignificant (at the 95% confidence level), but will be included in the final model
4-18
-------
Table 4-5. Parametric Regression Analysis for Third-, Second-, and First-Order Models Based on Equation 4-3
Model
Third-order
Second-order
First-order
Coeff/SE
CO2 coeff
SE (C02)
NH3 coeff
SE (NH3)
CO2 coeff
SE (C02)
NH3 coeff
SE (NH3)
CO2 coeff
SE (C02)
NH3 coeff
SE (NH3)
bp
154.6
15.6
2.01
2.6
202.3
26.4
2.28
1.6
217.4
19.3
-0.57
1.8
by
219.8
11.0
4.50
1.8
222.2
26.0
4.52
1.6
237.3
21.2
4.47
2.0
bF
366.3
11.0
37.3
1.8
365.8
26.0
37.3
1.6
370.5
24.1
36.8
2.2
bPY
674.0
135.3
-12.4
22.3
303.0
249.2
-14.5
15.2
bPF
1905
540.7
-57.6
89.2
-294.4
478.1
-70.1
29.1
bYF bPYF R2(adj.)
280.4 -6,858 99.7%
105.5 1563
16.2 -38.9 98.1%
17.4 257.8
202.1 98.3%
245.3
15.8 98.5%
14.9
98.5%
96.9%
SE is the standard error for the corresponding coefficient.
-------
Section 4.0 Carbon Dioxide and Ammonia Yields
since the mixed paper / food waste interaction (bPF) was found to be statistically significant in the
second-order model (at the 95% confidence level). Therefore, a reduced second-order model
was finally fit by ignoring the MXP/YW and FW/YW second-order terms and the third-order
term. Based on the above, the best reduced models expected to adequately estimate CO2 and
NH3 yields on a per dry kilogram of MSW mixture basis (in g C/dry kg and g N/dry kg,
respectively) are Equations 4-8 and 4-9:
Ykg co2 = 217.4xFP + 237.3xFY + 370.5xFF (4-8)
YkgNH3= L29 (±1.38)xFP +5.15 (±1.37)xFY +
37.6 (±1.56)xFF-68.9 (±23.4)xFpxFF (4-9)
where
Ykg C02 = mass of CO2 (as g C) emitted per dry kg of MSW mixture
Ykg NH3 = mass of NH3 (as gN) emitted per dry kg of MSW mixture
FP, FYandFF = dry fractions of mixed paper, yard waste and food waste,
respectively, in the mixture, with each of the FP, FY, FF values
ranging from 0 to 1 and with FP+FY+FF always equal to 1.
The adjusted R2 is 98.5% for both Equations 4-8 and 4-9. The negative second-order
coefficient in Equation 4-9 indicates that the mixing of paper and food wastes significantly
reduces overall NH3 emissions. This conclusion was also drawn and explained as part of the full
factorial design analysis presented earlier.
The mixture experimental approach can also estimate interactions, also known as
synergisms or antagonisms (Cornell, 1990), which are indicated by the positive or negative sign
of the coefficient of the product of any of the F; values in the equation (with I being either P, Y or
F, as noted above).
4.3 Final Recommended Model(s)
The above statistical methodologies present models that can estimate CO2 and NH3 yields
during MSW composting. The full factorial design based on Equations 4-1 and 4-2 are useful
only in understanding whether the yields are additive and what interactions are important. Their
applicability in estimating gaseous emissions for various MSW mixtures is limited, since they
are based on the specific low and high levels used during the experiment. Therefore, Equations
4-4 and 4-6 estimate the mass of CO2 or NH3 (in g) emitted from the specified dry weights of the
components (in kg) by setting the MXP/YW/FW substrate weight equal to 1 kg; Equations 4-8
and 4-9, however, are based on the grams of C-CO2 (or N-NH3) emitted per dry kg of each of the
mixtures.
The mixture-experiment-based Equations 4-8 and 4-9 are based on the same runs used in
the full factorial design, supplemented by one additional run. The additional run (MSW1/3) was
performed to investigate further the mixed paper/yard waste interaction that appeared as
4^20
-------
Section 4.0 Carbon Dioxide and Ammonia Yields
marginally significant by the full factorial analysis. The results of this additional run could not
have been included in the full factorial design.
Because errors were distributed normally for Equations 4-4, 4-6, 4-8, and 4-9, the
adjusted R2 values can be used to compare the above equations. Based on that, Equations 4-8
and 4-9 (with adjusted R2 of 98.5% each) are more adequate for predicting CO2 and NH3 yields,
respectively, than Equations 4-4 and 4-6 (with adjusted R2 of 97.7% and 91.6%, respectively).
The highest residuals of Equation 4-8 are observed for the mixed paper and the mixture
of yard wastes and food wastes (see also Figure 4-5). In the case of mixed paper, the predicted
value from Equation 4-8 is 217.4 g C-CO2 /dry kg, while the experimentally measured yield was
153.3 g C-CO2/dry kg. This deviation is probably attributable to the fact that mixed paper had
not fully decomposed, probably due to inadequate seeding; therefore, the final expected yield for
that run might have been equal to that predicted by the first order model. Based on that, the
MXP run had yielded approximately 71% of its expected ultimate CO2 yield during the
experiment. According to the gas production rates of the mixed paper run at termination of that
run, it would take approximately 200 additional days for mixed paper to reach its ultimate CO2
yield, as is predicted by Equation 4-8.
According to Equation 4-8, no component interactions are statistically significant during
degradation of MSW components to carbon dioxide, as long as these components reach their full
extent of decomposition. An additive model was also suggested during the full factorial
analysis, as shown by Equation 4-5. Figure 4-5 shows the actual and estimated CO2 yields from
the runs given in Table 4-4.
Based on the recommended Equation 4-9, yard wastes and food wastes are the only two
statistically significant components, while only one two-component interaction is significant
when predicting NH3 yields. The full factorial analysis had shown that the mixed paper and yard
wastes interaction was also significant in addition to the mixed paper and food waste interaction.
The full factorial derived equation had shown that the yard waste main effect is insignificant, as
opposed to the mixture equation predictions. These partially different conclusions are attributed
to the different formats of Equations 4-6 and 4-9. Equation 4-6 estimates the yields of gases
produced from the specific amounts of each component present in the mixture (e.g., 0.69 g N
produced from 0.155 dry kg), while equation 4-9 estimates the yields per dry kg from each
component or mixture. Equation 4-9 estimates NH3 yields relatively well for all runs except for
the runs that contain relatively large percentages of mixed paper (MXP, MXP/FW, MXP/YW,
MXP/YW/FW). This lack of fit is probably due to the relatively high ammonia yield from the
MXP run resulting from the late addition of nutrients. Figure 4-6 shows the actual and estimated
NH3 yields from the runs mentioned in Table 4-4.
Figures 4-5 and 4-6 are the response surfaces for the prediction of CO2 and NH3 yields
according to Equations (4-8) and (4-9), respectively. Black dots indicate the "location" of
experimental runs, showing the percentages at which components were combined in the mixture
on a dry weight basis. The "flat" response surface of Figure 4-5 indicates that the predictive
model is linear. The highest yield is observed when food waste is the sole component of MSW
with a maximum yield of approximately 366 g C-CO2 /dry kg of food wastes, which is the
average value of the FW and FWns run CO2 yields. The CO2 yield predicted for mixed paper
-------
Section 4.0
Carbon Dioxide and Ammonia Yields
(MXP/YW)
246.1 [219.8]
(MXP)
153.3 [217.4]
(MXP/FW)
236.6 [224.0]
(YW*) (YW/FW)
219.5 [237.3] 301.9 [266.6]
(FW*)
366.6 [370.5]
Figure 4-5. Response surface for estimation of CO2 yields using Equation 4-8
(contour lines represent points of equal carbon dioxide yield expressed
in g C-CO2 per dry kg of substrate).
* Average experimental value from seeded and unseeded runs considered as replicates for
corresponding component.
Note: Experimental data values shown left of the brackets; Equation 4-8 predictions shown in
brackets.
4-22
-------
Section 4.0
Carbon Dioxide and Ammonia Yields
(MXP)
2.0 [1.3]
(MXP/YW)
0.6 [1.9]
(MXP/FW)
1.1 [0.0]
(jyrxp/FW/YW)
0.5 [1.1]
(YW*)
4.5 [5.2]
(YW/FW)
14.5 [12.3]
(FW*)
37.4 [37.6]
Figure 4-6. Response surface for estimation of NH3 yields using Equation 4-9
(contour lines represent points of equal ammonia yield expressed
in g-N-NH3 per dry kg of substrate).
Note: Experimental data values shown left of the brackets; Equation 4-9 predictions shown in
brackets. *: indicates average experimental value from seeded and unseeded runs
considered as replicates for corresponding component.
4-23
-------
Section 4.0 Carbon Dioxide and Ammonia Yields
(shown at the top of the triangle in Figure 4-5) is the expected ultimate CO2 yield from that
component, which would be observed had longer composting times or a better seed been used.
As Figure 4-6 shows, a plateau exists over approximately the top one third of the triangle
representing negligible NH3 yields from MSW mixtures with a dry weight mixed paper content
higher than approximately 80%. A "slope" in the middle of the triangle toward food waste
exists, reflecting the increase in NH3 yields as the food waste percentage in MSW increases. The
yield of 2.0 g N-NH3/dry kg of mixed paper represents a separate peak at the top of the triangle.
Figures 4-5 and 4-6 can be used for rapid estimation of CO2 and NH3 yields, respectively,
from MSW of various compositions.
It is noted that all of the above analyses and conclusions are based on the use of mixed
paper, yard wastes, and food wastes as the only organic components of MSW. Therefore, it is
assumed that MSW has negligible amounts of other decomposable wastes, such as agricultural
wastes, sewage sludge, etc. To express gaseous yields per unit mass of actual MSW that
contains inorganic components, the yields predicted here—which are based only on the organic
fraction—have to be reduced according to the percentage of inorganics in MSW.
In addition, Equations 4-8 and 4-9 were derived from fitting the data resulting from the
experimental runs given in this paper. In this sense, Equations 4-8 and 4-9 are empirical models
that describe (or estimate) these data but cannot be considered proven predictive tools.
4-24
-------
Section 5.0 Solids Decomposition During Composting
5.0 Solids Decomposition During Composting
The objective of this chapter is to describe the decomposition of solids during MSW
composting using the laboratory digesters. Seventeen experimental runs were performed using
food wastes, yard wastes, and mixed paper, which are the primary degradable components in
MSW. These components were composted individually and in mixtures, and selected
subcomponent runs of grass, leaves, branches, and office paper were also performed. The loss of
solid matter was followed by measurement of total carbon content, total nitrogen content,
volatile solids content, and five organic chemical classes: hot water soluble matter (mostly
sugars and starches), fats and lipids, cellulose, hemicellulose, and lignin/humus. This chapter
also considers the concentrations of some of the solid parameters found in the composted
substrates from the various runs in an attempt to derive compost maturity indicators.
The experimental runs performed were: seeded food wastes (FW), unseeded food wastes
(FWns), seeded mixed paper (MXP), unseeded mixed paper (MXPns), seeded yard waste (YW),
unseeded yard waste (YWns), mixed paper + yard wastes (MXP/YW), mixed paper + food wastes
(MXP/FW), yard wastes + food wastes (YW/FW), mixed paper + yard waste + food waste
(MXP/YW/FW), mixed paper + yard waste + food waste combined at approximately equal dry
weight percentages (MSW1/3), and the seed. The gas yields for these runs are reported and
discussed in the previous chapter.
Additional individual subcomponents composted and considered in this chapter were
grass, leaves, branches, and office paper (OFF); grass and leaves were used to prepare the yard
waste mixtures (YW and YWns). Grass and leaves were composted without any prior shredding
and without addition of seed. Because composting facilities for disposal of wood materials do
exist in the United States, branches were composted individually without any shredding.
Branches consisted of small twigs, 5 to 10 cm long, and were seeded with leaves at a ratio of
5.4:1, dry branches to dry leaves. Office paper, a subcomponent of mixed paper, was also
composted individually. In contrast to cardboard and newsprint, which are mechanically treated
wood pulps, a chemical process is associated with office paper manufacture that could affect its
decomposition. Office paper was seeded with both seed and leaves in one run at a dry mass ratio
of 10:1:1. Leaves were added to office paper to investigate the effect of (probably) different
microbial populations on office paper decomposition.
In addition to the four individual subcomponents, a yard waste mixture (YWh) was
prepared from a batch of grass with a higher volatile solids content than the grass used in the
YW and YWns runs. The high volatile solids grass had a volatile solids content of 89.8% dry
weight (dw) as opposed to the latter batch with a volatile solids content of 81.6% (dw).
Moisture and nutrients were added to all the runs as needed. The materials and methods
associated with setup of the laboratory experiment, substrate selection and preparation, analytical
5-1
-------
Section 5.0 Solids Decomposition During Composting
techniques, and quality assurance and quality control results performed during operation of the
laboratory runs are discussed in Chapter 3.
5.1 Results and Discussion
5.1.1 Initial Composition of Substrates
Table 5-1 provides initial chemical composition data for all runs. The five chemical
classes presented are hot water soluble matter, fats and lipids, cellulose, hemicellulose, and
lignin. Table 5-1 also includes initial volatile solids contents and initial organic carbon contents
of the substrates used in the runs. The CO2 yields are shown in Table 5-1.
Table 5-1 presents initial composition including the dry mass of seed present in the
seeded runs. As Table 5-1 shows, cellulose is the dominant building block for almost all
substrates. In particular, all runs that included paper had the highest average cellulose contents
(>60% dry mass). Food waste runs were next highest with more than 42% cellulose content
(dw). Mixed yard waste runs had cellulose contents of approximately 27% (dw), since grass,
which had a 39.7% cellulose content, was the primary constituent. Leaves had the lowest
cellulose content of 9.5% (dw) among all substrates. Leaves appear to have been partially
decomposed naturally, as was indicated by the relatively low (compared to the other substrates)
initial volatile solids content of 63.3% (dw). The highest cellulose content of 69.7% was
measured for mixed paper (unseeded), which had a volatile solids content of 94.8% (dw).
The next most dominant component for almost all substrates was the lignin/humus
fraction, which will be referred to as "lignin" hereafter. The measured lignin fraction includes
lignin, humic acids, and acid-insoluble proteins (Tenney and Walksman, 1929; Inoko et al.,
1979). Branches and leaves were the most lignified of all substrates with lignin contents of
42.9% and 33.9% on a dry weight basis, respectively. The three yard waste mixtures had lignin
contents of 24.3% to 26.0% (dw). Food wastes had a lignin content of 12.03% and 14.3% (dw),
with office paper (including the added leaves and the seed) having the lowest value of 6.5%
(dw).
Hemicellulose, HWSM, and fats/lipids, in that order, were the next most dominant
groups of compounds for almost all substrates. Hemicellulose was assumed to be a complex of
two hexoses (D-galactose and D-mannose) and two pentoses (L-arabinose, D-xylose) only.
Glucose was assumed to be derived only from cellulose. The highest hemicellulose content at
16.9% dry weight was found in grass and the yard waste mixtures. A 12.9% (dw) hemicellulose
content was found in branches, while food wastes appeared to have negligible amounts of
hemicellulose. All mixed paper mixtures had hemicellulose contents from 6.9% to 8.5% (dw).
The HWSM fraction is expected to be mostly sugars and starches but its composition
might differ from component to component. Food wastes had the highest amount of HWSM, at
13.6% (dw), with grass and yard waste mixtures following at 10.6% and approximately 7.0%
(dw), respectively. Mixed paper mixtures, leaves, and branches had HWSM contents of less
5-2
-------
Table 5-1. CO2 Yields and Initial Chemical Composition of Substrates
from 17 Experimental Runs
CO2 Fats/ Hemicel-
(in g C/ lipids HWSM Cellulose lulose
Contents
FW
1 ""ns
FW
YWns
YW
MXPns
MXP
MXP/YW
MXP/FW
YW/FW
MXP/YW/FW
MSW1/3
Seed
YWh
Grass
Leaves
Branches'
Office paper'
FW
FW =
rvvns
MXP
MXPns
MXP/YW
MXP/FW
MXP/YW/FW =
MSW1/3
YWns
YW
YWh
YWns
YW/FW
dry kg)a (% dw)b (%
368.6 12.89
364.4 11.57
221.9 2.49
217.0 2.87
5.5 .018
153.3 0.68
246.1 1.15
236.6 1.15
301.9 4.80
265.3 1.58
266.3 4.79
86.0 5.31
265.2 2.48
202.0 2.41
81.1 2.59
82.7 1.04
112.5 0.74
Seeded food wastes
Unseeded food wastes
Seeded mixed paper waste
Unseeded mixed paper
Mixed paper + yard wastes
Mixed paper+food wastes
dw) (% dw) (% dw)
13.61
12.26
6.93
6.74
3.54
3.76
4.32
4.14
8.39
4.68
7.85
5.86
7.23
10.62
1.68
3.20
3.72
46.09
42.51
27.20
26.82
69.66
65.41
57.97
64.62
31.65
57.84
47.39
25.51
27.73
39.67
9.48
14.71
68.13
0.00
0.73
11.25
10.23
7.79
7.45
8.49
7.16
8.75
7.56
6.89
4.20
16.05
16.89
3.24
12.87
6.71
a
b
c
d
Lignin/
humus
vs c
C/N Cellulose/
(%dw) (%dw)c (%dw)d
12.03
14.33
24.34
24.54
15.90
16.80
18.57
16.71
21.53
18.02
17.66
25.21
26.00
17.63
33.88
42.89
6.50
CO2 (as g C)
95.9
91.7
73.8
73.9
94.8
92.5
89.7
92.7
78.8
89.3
87.7
71.8
79.8
81.7
63.3
81.7
86.9
49.2
47.1
36.2
36.5
44.1
43.4
42.5
43.7
39.3
42.4
42.8
37.0
39.2
40.6
30.1
40.4
39.2
ratio6 lignin ratio
7.9
8.4
18.7
17.8
505.6
23.5h
27.5h
23.2h
13.6
25.5h
16.6
13.1
18.0
17.5
21.6
13.1h
27.6h
3.83
2.97
1.12
1.09
4.38
3.89
3.12
3.87
1.47
3.21
2.68
1.01
1.07
2.25
0.28
0.34
10.48
FHCHU
VSf Duration9
88.2
88.7
97.8
96.3
102.4
101.7
100.9
101.1
95.3
100.4
96.5
92.1
99.6
106.8
80.4
91.5
98.8
91
57
69
77
88
198
179
129
114
170
47
62
110
27
71
54
61
per dry kg of total substrate.
% on a dry weight (dw) basis.
Volatile solids.
Total organic
carbon.
e Includes N added as N salt.
f
Mixed paper+yard wastes+food wastes
Mixed paper+yard wastes+food wastes
equal dry weight proportions
Unseeded yard wastes
Seeded yard wastes
Yard wastes with high volatile
Unseeded yard wastes
Yard wastes+food wastes
mixed each at
g
h
'
The ratio of the sum of fats,
lignin/humus
HWSM, cellulose,
concentrations to the
hemicellulose, and
volatile solids concentration.
Experiment length (days).
Run to which
Seeded with
NH4NO3 salt was added.
leaves and seed.
solids content
-------
Section 5.0 Solids Decomposition During Composting
than 5% (dw). The fats and lipids fraction is expected to also contain waxes, resins (Tenney and
Walksman, 1929), and probably alcohol-soluble proteins. Fats and lipids were at the smallest
percentages of all five chemical groups for all substrates, except food wastes, which contained
approximately 12% (dw) of fats.
Table 5-1 shows that the volatile solids contents of all substrates used ranged from 63.3%
(dw) for leaves to 95.9% for food wastes. The volatile solids content was approximately two
times higher than the corresponding total organic carbon content for all substrates. The factor of
2 differs slightly from the factor of 1.8 suggested by New Zealand researchers (after Diaz et al.,
1993) and used frequently since then. Differences are probably attributed to the use of different
analytical techniques for the estimation of organic carbon or to the types of substrates used. The
C/N ratio (also shown in Table 5-1) includes the weight of nitrogen added as NH4NO3 salt to
some of the runs in order to achieve the suggested C/N optimum range of 25 to 30. The lowest
C/N ratio of 7.9 was measured for food wastes, which had an initial N content of 6.1% (dw).
The seed and the grass were the next most nitrogenous subcomponents, with N contents of 2.8%
and 2.1% (dw), respectively.
It is noted that, because of time constraints, office paper had not fully decomposed, as
was shown by the relatively low but steady CO2 production rates prior to terminating that run.
Mixed paper is suspected not to have reached its fully decomposed because of inadequate
seeding, as discussed in Chapter 4.
5.1.2 Substrate Degradability as a Function of Lignin and HWSM Contents
Linear regressions (including an intercept) were made between the initial concentrations
of the five chemical groups measured in the solids and the total carbon dioxide yields (Table 5-1)
from 15 runs. Office paper was not included in these regressions because it had not reached its
full extent of decomposition. The MXPns results were also not included since no decomposition
occurred for this run due to lack of seed. Based on the regressions (multiple regressions were
not tried), the correlation coefficients (adjusted R2) for fats/lipids, HWSM, cellulose,
hemicellulose, and lignin/humus were 0.33, 0.53, 0.14, -0.01 and 0.50, respectively. Therefore,
the initial lignin content and initial HWSM content had the strongest correlation to
mineralization extent. Components with high HWSM contents had high CO2 yields, while
components with high lignin content had low CO2 yields. Figure 5-1 presents the carbon dioxide
yields (in g C/dry kg) versus initial lignin contents.
The 15 data points were fitted by least squares to the linear equations shown as
Equations 5-1, 5-2, and 5-3. Equation 5-1 relates the carbon dioxide yield to initial lignin
content (dry weight basis), Equation 5-2 relates volatile solids reduction (expressed as a fraction)
to initial lignin content (volatile solids basis), and Equation 5-3 relates carbon dioxide yield to
initial HWSM content (dry weight basis). The equations, with standard errors and coefficients of
correlation (adj. R2), are:
5-4
-------
Section 5.0
Solids Decomposition During Composting
400
350
300
250
Q 200
^ 150
T3
o 100
CO
-i— <
c/)
.0
"ro
"CD
(/> 50
o
FW
Data points
Equation 1
Office paper *
Seed Leaves \Branches
MXP
_i_
_L
0% 10% 20% 30% 40% 50% 60%
% initial lignin content (dry weight basis)
* Not included in fitting to Equation 5-1.
Figure 5-1. Effect of initial lignin content on carbon dioxide yields.
Ligninfit:
CO2yield = 409.9 (±50.3) - 841.8 (±215.6)x LGNdiy (adj. R2=0.50)
BVS = 0.80 (±0.09) - 0.91 (±0.30)x LGNVS
HWSMfit:
C02yield = 87.6 (±37.2)+ 2,019 (±496)xHWSMdiy
(adj.R2=0.37)
(adj.R2=0.53)
(5-1)
(5-2)
(5-3)
where
BVS
LGNdiy
LGN,,
HWSMdiy
carbon dioxide yield (in g C / dry kg of component excluding seed)
volatile solids reduction (as fraction)
the substrate's initial lignin content (dry weight expressed as a fraction
of the total dry weight)
the substrate's initial lignin content (dry weight expressed as a fraction
of the dry volatile solids)
initial hot water soluble matter content (dry weight basis expressed as
fraction ).
5-5
-------
Section 5.0 Solids Decomposition During Composting
Equations 5-1 and 5-3 can be used for estimating CO2 yields from different
lignocellulosic substrates by knowledge of their initial lignin and HWSM contents, respectively.
As suggested by Figure 5-1 and Equations 5-1 and 5-2, lignin retards overall substrate
decomposition. This retardation is thought to be due primarily to physical inhibition. Lignin is
present between cellulose fibrils, decreasing the available surface area and preventing ready
access to cellulose by invading microbes and enzymes (Szegi, 1988). Most of the information on
the inhibition of lignin to decomposition in aerobic environments is found in studies of soil
environments. Szegi (1988), in particular, showed that degradation of cellulose by certain fungi
is actually retarded by lignin in soil environments as a result of physical inhibition rather than
the simple presence of lignin. Lignin, however, is degradable aerobically, especially by white-
rot fungi and brown-rot fungi (Kirk, 1984) and by bacteria to a much smaller extent (Vicuna,
1988).
The retardation of organic matter degradation due to lignin has been studied more
extensively for anaerobic environments. Unaltered lignin is known to be nondegradable in
anaerobic environments because its initial fragmentation requires molecular oxygen (Kirk,
1984). The retardation by lignin on the anaerobic degradability of several types of organic
matter has been shown by Chandler et al. (1980). Retardation by lignin to the anaerobic
degradability of paper wastes and various solid waste components has been shown by Stinson
and Ham (1995) and Eleazer et al. (1997), respectively. Stinson and Ham (1995), in particular,
showed that the retardation of cellulose decomposition in paper wastes was due to physical
inhibition related to the sheathing of cellulose by lignin rather than to a chemical inhibition (e.g.,
adsorption of cellulolytic enzymes onto lignin).
Equation 5-2 allows comparison to a similar equation developed by Chandler et al.
(1980) that applies to anaerobic environments. Chandler's predictive model is:
BVS = 0.830-2.8 LGNVS, (R2 = 0.94), with terms defined as in Equation 5-2. The intercepts in
Equation 5-2 and in Chandler's predictive model are very similar, suggesting that an organic
substrate with no lignin content would achieve a maximum volatile solids reduction of 83% in
anaerobic environments and 70% to 90% (0.80±0.09) in aerobic environments. Based on the
slope of the linear Equation 5-2—which is 0.91—and the corresponding slope of Chandler's
equation—which is 2.8, it appears that the degree of retardation of substrate degradability due to
lignin is approximately three times higher in anaerobic environments than in aerobic
environments. This is probably a result of the negligible anaerobic decomposition of lignin
compared to aerobic decomposition.
Care should be taken in comparing these two equations because different organic
substrates were used as a basis to fit each equation. In addition, the Chandler equation was
based on a range of lignin contents from 2.0% (VS basis) for corn meal up to 20.9% (VS basis)
for newsprint, while the corresponding range for Equation 5-2 was 12.5% (VS basis) for food
wastes to 53.5% (VS basis) for leaves. The lignin fraction, however, was determined using the
72% sulfuric acid extraction technique in both studies.
The variability observed in Equations 5-1 and 5-2 can be partly explained by the fact that
the analytically measured lignin/humus fraction includes, apart from lignin, other chemical
components that can have different degrees of availability or biodegradability and therefore have
different degrees of inhibition to overall degradability. Lignin/humus in food wastes, for
-------
Section 5.0 Solids Decomposition During Composting
example, was most degradable with an overall reduction of 74.1% during composting, while
lignin/humus in branches was the least degradable with a reduction of 9.9%. This might
partially explain the fact that food wastes had CO2 yields higher than predicted by the model.
In addition, different types of lignin have different degrees of biodegradability.
According to Shevchenko and Bailey (1996), grass lignin is different from softwood and
hardwood lignins because of differences in the ratios of guaiacyl, syringyl, and/?-hydroxy-
phenylpropane units present in each type. Kirk (1984) mentions that a broader range of
microbial populations is involved in the degradation of lignin in herbaceous plants—such as
grass and leaves—compared to lignin in wood, making the former class of lignin more
degradable than the latter. Allison (1973), by incubating woods and barks in soil aerobic
environments, showed that there is a significant difference in degradability between hardwoods
and softwoods and even between species within the softwood and hardwood families by factors
up to 5.
In contrast to Equation 5-2, the Chandler formula does not have such a large variability,
despite the range of organic substrate types used. This is probably because all types of lignin are
similarly refractory to degradation in anaerobic environments but vary in degradability in aerobic
environments.
The CO2 yield from office paper is the major outlier to the lignin predictive models and
was not used in curve fitting. Office paper's low lignin content should, according to the model,
result in the highest CO2 yield. Although office paper was provided additional seed at day 21 and
phosphorus (P) and nitrogen (N) nutrients were added on day 54, the small increase in CO2
production rate observed was typical after opening all digesters for inspection or moisture
addition. It is likely seed, N, and P were not limiting decomposition. Explanations for the low
decomposition rate of office paper are:
• Fillers and sizing agents of organic nature are added to office paper to prevent the
feathering (spreading) of inks into the sheet. These agents are usually resin acids
(e.g., abietic acid) that create a more hydrophobic surface, less prone to attack by
microorganisms (Young, 1998).
• The moisture content ranged between 60% to 65%, with the office paper forming
wet clods. This probably reduced oxygen access and therefore decomposition
rates. A longer period of time would have resulted in more CO2 generation, but
the CO2 rate at the conclusion of the run was very low, suggesting much
additional time would be required to reach "full" decomposition.
The seed had also lower CO2 yields compared to the model prediction from its lignin
content. Since seed was partially composted MSW, it could contain materials such as plastics,
textiles, leather or other organics, that can have a different degradation behavior than the other
components (food waste, yard waste, paper) used in this experiment. If plastics were present, for
example, they would be quantified as lignin/humus and could explain the relatively low
CO2 yield from seed, since plastics are generally considered nondegradable.
5-7
-------
Section 5.0 Solids Decomposition During Composting
The correlation of CO2 yield and initial HWSM content (adj.R2=0.53) is illustrated in
Figure 5-2. This correlation is partially explained by the fact that the HWSM group contains
relatively easily decomposable organic compounds, such as starches, monosaccharides, and
amino acids (Inoko et al., 1979). Office paper was not included in the fitting since it may not
have reached its full extent of decomposition, as discussed earlier. The seed appears to behave as
an outlier in Equation 5-3. The presence of components in the seed other than food wastes,
mixed paper, and yard wastes might partially explain the relatively large deviation of its value
from Equation 5-3 estimations.
The correlation of total CO2 yields to initial lignin content and initial HWSM content
indicates that these values can be used as general predictors of ultimate CO2 yields from various
substrates when composted. The use of lignin and HWSM as CO2 yield predictors does not
mean, however, that these groups are the dominant biodegradable groups of the substrates used.
This is because lignin has a low degradation extent, while HWSM has a relatively small initial
concentration that ranges from 1.7% to 13.6% (dry weight) for all substrates. As will be shown
in the next section, it is the degradability of other fractions, such as cellulose and hemicellulose,
that controls the extent of substrate degradation. Based on the above, the variability of
Equations 5-1 to 5-3 can be explained by the fact that substrates with similar initial lignin
contents, or HWSM contents, can have different degradabilities not only because of
degradability differences of the lignin, or HWSM, but also because of differences in the contents
of the cellulose that the lignin sheaths.
5.1.3 Reduction of Chemical Components
The objective of this section is to discuss the extent of reduction of each of the five
chemical groups during composting and the contribution of each component to total substrate
degradability. Table 5-2 presents the extent of reduction for each of the five chemical groups for
all runs. Table 5-2 also includes the dry matter reduction and the total organic carbon reduction.
Figure 5-3 illustrates the total dry matter reduction for each run as shown by the length of each
bar (positive values). Total dry matter reduction is further broken down to the dry loss of each
chemical group. Table 5-3 shows the contribution of dry mass loss of each of the five chemical
groups to total dry mass loss recorded for each run. The sum of the dry mass losses of the five
chemical groups for each substrate is expected to be close to the total dry mass loss of that
substrate. Deviations are primarily because certain chemical groups were not accounted
for during solids measurements. Table 5-3 includes the dry mass reduction closure (in %). A
positive closure value indicates an underestimation of the total dry mass loss, while a negative
value indicates an overestimation of the total dry mass loss. Figure 5-3 also illustrates the dry
mass reduction closure.
High solids loss is particularly apparent for food wastes; further, the unmeasured solids
class was relatively large compared to other runs. The unmeasured dry fraction for food wastes
is probably a water-insoluble chemical group (e.g., water-insoluble proteins) that is solubilized
during acid hydrolysis but is not measured during HPLC analysis as any of the monosaccharides
used for cellulose and hemicellulose determination. Negative values in Figure 5-3 indicate that
an increase of the corresponding chemical group was observed during the process. For example,
a lignin increase was observed for mixed paper, probably due to generation of humus and
5-8
-------
Section 5.0
Solids Decomposition During Composting
o
o
400
350
300
250
200
150
100
50
Data points
Equation 5-3
FW
FVV
FW/YW
MXP/YW/FW
MXP/YW^
MXP/FW
MXP,
YWh
"1/3
YWYWns
Grass
_ Leaves Branches
Office paper
Seed
0% 1% 2% 3% 4% 5% 6% 7% 8% 9% 10%11%12%13%14%15%
% initial HWSM content (dry weight basis)
*Not included in fitting to Equation 5-3.
Figure 5-2. Effect of initial hot water soluble matter (HWSM)
content on carbon dioxide yields.
biomass (measured in the lignin/humus fraction) that was greater than overall lignin loss. An
overestimation of the dry mass reduction was observed, especially for grass, MXP/FW, and
YWns, probably due to overestimation of the cellulose and hemicellulose contents during HPLC
analysis.
Figure 5-3 illustrates that cellulose was the major contributor to the total dry matter loss
for most substrates. Cellulose dry loss contributed between 66.5% and 92.4% of the total dry
losses recorded for the MSW1/3, MXP/YW, grass, office paper, MXP/YW/FW, MXP/FW, and
MXP runs, while it contributed approximately 50% of the dry loss of all yard waste runs (see
Table 5-3). Based on Table 5-2, cellulose was degradable for all substrates at levels that ranged
from 25% loss for office paper to 91% for yard wastes. The low cellulose decomposition of
5-9
-------
Table 5-2. Correlation Matrix for Estimation of Compost Maturity Indicators
Using the Reduction of Organic Chemical Groups During Composting
o'
s
Substrate3
MXPnsb
MXP
Leaves
MXP/YW
Seed
MXP/YW/FW
FW/YW
YW
YWh
FW/MXP
FW
Branches
Ywns
FWns
MSW1/3
Office paper
Grass
Correlation
coefficient (R2)
Average
final CO2
production
rate
(g C/dry
kg/day)
0.02
0.25
0.26
0.26
0.26
0.28
0.31
0.45
0.50
0.64
0.75
0.93
1.02
1.03
2.47
2.88
3.70
Fats/lipids
reduction
(%)
0.0
34.7
69.6
67.1
81.8
77.1
88.7
66.6
74.0
61.1
94.0
42.0
77.2
82.6
93.9
100.0
80.4
0.18
HWSM
reduction
(%)
0.0
10.4
-489.4
58.9
-5.3
56.6
47.0
63.9
60.4
99.5
53.7
-80.6
58.6
45.5
62.5
39.2
-27.6
0.01
Lignin
reduction
(%)
0.0
-28.8
18.7
26.7
-7.2
24.2
33.4
30.1
38.5
0.1
36.9
9.9
43.0
61.6
12.8
28.5
12.1
0.002
Cellulose
reduction
(%)
0.0
49.6
53.9
82.1
36.2
87.8
89.5
82.1
89.8
60.4
60.2
-54.8
90.7
66.0
82.0
25.0
88.0
0.00
Hemicel-
lulose
reduction
(%)
0.0
71.9
73.5
91.5
23.9
87.4
95.0
96.6
100.0
100.0
100.0
13.7
99.4
0.0
79.4
100.0
93.8
0.00
Dry mass
reduction
(%)
0.0
35.1
19.0
62.4
19.4
62.5
55.1
46.7
53.9
45.3
59.2
15.4
47.6
65.5
58.4
24.0
44.7
0.02
C reduction
(%)
0.0
37.9
26.9
66.3
27.4
68.1
70.5
63.1
66.4
46.8
65.7
18.0
62.2
69.2
65.5
28.8
54.7
0.006
Cellulose/
lignin
reduction
(%)
0.0
60.9
43.3
75.6
40.5
83.9
84.3
74.4
83.3
60.4
36.9
-71.8
83.7
11.6
79.4
-5.0
86.3
0.005
C/N ratio
reduction
(%)
0.0
9.2
64.1
31.3
9.8
38.1
29.4
57.2
38.0
11.2
-112.8
-31.2
56.3
-95.8
NM
NM
NM
0.29
NM = Not measured.
a Substrates are placed in ranking order from most mature (top of table) to least mature (bottom of table).
b Negligible mineralization occurred for the MXPns run and therefore no solids measurements were performed for this run; all solids reduction
values were set to 0 for this run.
o
s
o
'
-------
Section 5.0
Solids Decomposition During Composting
Office paper
Branches
Leaves
Grass
YWh
Seed
MSW1/3
MXP/YW/FW
FW/YW
MXP/FW
MXP/YW
MXP
MXPns
YW
YWns
FW
FWns
I 1 Dry mass reduction closure
R<\<3 Hemicellulose
BSS1 Cellulose
V//A I innin/hiimng
HWSM
Fats/lipids
solids analysis was not performed
-10%
10% 20% 30% 40% 50% 60% 70% 80%
Dry mass loss (%)
Figure 5-3. Dry mass loss for each chemical group during
composting (in % of the total initial dry mass).
office paper is attributed to reasons discussed earlier for this substrate. Apart from the physical
inhibition of lignin to cellulose decomposition, the binding of cellulose to humic matter during
humification can also result in retardation of cellulose decomposition. Cellulose reduction in
seed and leaves was 36.2% and 53.9%, respectively. The relatively low cellulose
decomposition for these substrates, both of which were partially decomposed prior to
composting in the laboratory, might be explained by the binding of cellulose to the generated
humic matter. This is also indicated by the fact that seed and leaves have the highest initial
lignin/humus contents of 25.2% (dw) and 33.9% (dw), respectively. Hanninen et al. (1995)
suggested that carbohydrates in compost are covalently bound to the structures of humic acids
and fulvic acids during composting. According to Hanninen et al. (1995), the binding of
polysaccharides to humic matter during composting can be due to ester or ether bonds formed
between the hydroxyl group of polysaccharides and phenolic acid groups of the humic matter.
The polysaccharides to humic or fulvic acids are still subject to slow degradation, with
polysaccharides associated with the humic acid fraction being more biologically stabilized
compared to those associated with the fulvic acid fraction.
5-11
-------
Section 5.0
Solids Decomposition During Composting
Table 5-3. Contribution of Dry Loss of Each Chemical Group to Total Dry
Mass Loss of a Substrate (in %)
FW
ns
FW
YWns
YW
MXP
MXP/YW
FW/MXP
FW/YW
MXP/YW/FW
MSW1/3
Seed
YWh
Grass
Leaves
Branches
Office paper
Fats/
lipidsa
16.3C
18.4
4.0
4.1
0.7
1.2
1.6
7.7
1.9
7.7
22.4
3.4
4.3
9.5
2.8
3.1
HWSMa
9.5
11.1
8.5
9.2
1.1
4.1
9.1
7.2
4.2
8.4
-1.6
8.1
-6.6
-43.2
-16.8
6.1
Lignin/
humus3
11.3
8.9
22.0
15.8
-13.8
7.9
0.1
13.1
7.0
3.9
-9.3
18.6
4.8
33.3
27.6
7.7
Cellulose3
46.5
43.2
51.8
47.1
92.4
76.2
86.2
51.5
81.2
66.5
47.7
46.2
78.1
26.9
-52.3
70.7
Hemicellulose3
0.0
1.2
23.5
21.1
15.2
12.4
15.8
15.1
10.6
9.4
5.2
29.8
35.4
12.6
11.4
27.9
Dry mass
closure13
16.5
17.2
-9.9
2.6
4.3
-1.9
-12.7
5.5
-4.9
4.1
35.7
-6.1
-16.1
61.0
127.2
-15.5
a A positive value indicates reduction and a negative value increase.
b A positive value indicates a dry mass loss underestimation and a negative value indicates a dry mass loss
overestimation. For example, the sum of dry mass losses of the five chemical groups in food wastes was 16.5%
less than the measured total dry mass loss for that substrate, while the sum of dry mass losses of the five chemical
groups in grass was 16.1% higher than the measured total dry mass loss for that substrate.
c This means that 16.3% of the total dry loss recorded for food wastes was due to the reduction of fats/lipids.
Cellulose reduction in the MXP and MXP/FW runs was 49.6% and 60.4%, respectively.
These values are lower than the corresponding cellulose reductions in MXP/YW (82.1%),
MXP/YW/FW (87.8%), and MSW1/3 (82.0%). Cellulolytic microorganisms appear to be present
in yard wastes, resulting in more vigorous degradation of cellulose present in mixed paper in
the MXP/YW, MXP/YW/FW, and MSW1/3 runs. A lower amount of cellulolytic microorganism
activity appears to be present in the seed as indicated by the low cellulose reduction values for
the seed itself as well as for the seeded mixed paper (MXP). The relatively low (60.4%)
cellulose reduction in the MXP/FW mixture is probably a result of the low concentrations of
food in the MXP/FW mixture, which, therefore, did not adequately function as a cellulolytic
seed. The presence of cellulolytic microorganisms in yard wastes is also indicated by the 91%
cellulose reduction measured for unseeded yard wastes and the 88% cellulose reduction for the
grass alone. The negative value of cellulose reduction for branches (as shown in Figure 5-3 and
Table 5-2) is attributed to the low initial cellulose content and the resulting multiplier effect of
analytical variability.
Based on Figure 5-3 and Table 5-3, hemicellulose appears to be the second largest
contributor to total dry matter reduction for the yard waste runs, grass, leaves, and all mixtures
with mixed paper. Hemicellulose contribution ranges from 12.6% to 35.4% of the dry losses for
the above runs. Food wastes contained no hemicellulose while only 5% of the seed total dry
loss was due to hemicellulose degradation. Based on Table 5-2, hemicellulose was more
-------
Section 5.0 Solids Decomposition During Composting
degradable than cellulose in all components. Hemicellulose generally aids in the physical
bonding of lignin to cellulose fibrils (Szegi, 1988) and has a lower molecular weight than
cellulose (Gray et al., 1971). Szegi (1988), using several strains of soil fungi, showed that
hemicellulose in soil was more degradable than cellulose and much more degradable than lignin.
Hemicellulose was approximately fully degraded in the yard waste mixtures and the grass.
Hemicellulose, in the MXP/YW, MXP/FW, and MXP/YW/FW runs, was decomposed at levels
higher than 85% and was always more degradable than cellulose. The same microorganisms
that degrade cellulose have been found to also degrade hemicellulose (Szegi, 1988). The higher
degradation extent of hemicellulose compared to cellulose is explained by its lower molecular
weight and its greater heterogeneity compared to cellulose. Branches had one of the lowest
hemicellulose reduction extents, probably due to lignin inhibition.
Slightly lower cellulose reductions were observed by Michel et al. (1993) during
composting of mixtures of grasses and leaves. Cellulose reduction values ranged from 61% to
78% and hemicellulose reduction ranged from 57% to 82%. The lower values could be
attributed to the fact that all experiments were stopped after 43 days and, based on the CO2
cumulative profiles, substrates had not reached their full extent of decomposition.
As shown in Figure 5-3 and Table 5-3, lignin/humus was degradable in almost all
substrates. Lignin/humus loss contributed between 15.8% and 33.3% of the dry weight loss for
yard waste runs, branches, and leaves. It contributed less to food wastes and the grass. Lignin
is degradable under aerobic environments primarily by fungi and certain actinomycetes
(Crawford and Crawford, 1980; Kirk, 1984). Based on Table 5-2, lignin reductions varied from
9.9% for branches to 61.6% for unseeded food wastes. Lignin reduction in seeded food wastes
was 36.9%, probably because the lignin in the seed combined with the food wastes was less
degradable than the lignin in food wastes. The variations in lignin degradability might be
attributed to the different structures of lignin and different chemical groups included in the
"lignin/humus" fraction, as discussed. Generally, lignin reduction values should be interpreted
with caution, since the lignin/humus fraction that was quantified will contain humic matter and
portions of microbial biomass, both of which are generated during composting (Tenney and
Walksman, 1929). According to both the lignin theory and polyphenol theories of humus
formation, lignin is the primary source of humus in soil (Stevenson, 1994). The contribution of
polysaccharides to humus formation is still a subject for research (Shevchenko and Bailey,
1996).
As mentioned in the review by Szegi (1988), cellulolytic bacteria may be responsible for
humus formation. This pathway might partially explain the negative value for lignin reduction
observed for the MXP run and the seed if more humic matter and biomass were formed than
lignin was decomposed. The same reason might explain the zero reduction of lignin/humus in
the MXP/FW run. Since both the MXP and MXP/FW runs were seeded, it appears that no
ligninolytic fungi were present in the seed used. This is further supported by the fact that a
negative lignin reduction was recorded for the seed run alone. Seed was collected at the end of
a 5-day retention time from a MSW composting facility. According to the succession and
concentration of microorganisms during MSW composting (de Bertoldi et al., 1983),
ligninolytic fungi and actinomycetes that are responsible for lignin degradation are present in
high concentrations in the latter stages of composting. Though direct microbial population
5-13
-------
Section 5.0 Solids Decomposition During Composting
measurements did not take place here, the microbial population in the seed used is expected to
be dominated by heterotrophic bacteria rather than ligninolytic microorganisms.
According to Effland (1977), the degradation of lignin by fungi will result in some lignin
being dissolved in the 72% sulfuric acid. Soluble lignin in hardwoods also can be up to 3% to
5%. Partial acid dissolution of the lignin fraction in food wastes might explain the high lignin
reduction values for that substrate.
The addition of yard wastes to mixed paper in the MXP/YW or MXP/YW/FW runs
appears to further "seed" mixed paper with a microbial population able to degrade lignin. This
is shown by the fact that lignin reductions for the MXP/YW and MXP/YW/FW runs are on the
order of 25%. The presence of ligninolytic fungi in yard wastes is further supported by the 43%
lignin reduction measured for unseeded yard wastes, which was one of the highest lignin
reduction values. Ligninolytic fungi are likely to be present in two subcomponents of yard
wastes—grass and leaves—as was indicated by the 12% and 18.7% lignin reduction values for
these two unseeded components, respectively. Michel et al. (1993) also recorded lignin/humus
reduction values of up to 40% for mixtures of grass and leaves during composting.
Fats and lipids were in small amounts in almost all runs except food wastes.
Decomposition of fats and lipids contributed approximately 16% of the total dry weight loss in
food wastes (see Table 5-3). Fats and lipids in the seed were the greatest contributors to the
degradation of that substrate (22.4%) because of the advanced decomposition stage of that
substrate and the presence of relatively large concentrations of microbial biomass that is further
depleted during endogenous decay. Fats/lipids contributed 9.5% of the dry loss of leaves, which
was also at an advanced composting stage.
FEWSM loss was responsible for the smallest contribution of the total dry loss of nearly
all substrates, since it was in small concentrations initially. An exception was food wastes,
since FEWSM loss contributed approximately 15% to the total loss of that substrate. The
increase of HWSM in leaves and grass is attributed to the generation of water-soluble humic
compounds during composting (e.g., fulvic acids). The increase of FEWSM in branches is
partially attributable to the presence of leaves that were mixed with branches as seed.
Dry matter reduction values ranged from 15.4% for branches to 65.5% for unseeded food
wastes. Volatile solids (VS) and total carbon (C) reduction values were similar. Carbon
reduction ranged from 18.0% for branches to 70.5% for the mixture of yard wastes and food
wastes. The relatively low reduction for branches undoubtedly contributed to the poor dry mass
closure for branches (see Table 5-3) because of the multiplier effect of analytical variability.
The correlation of CO2 yield (as g C/dry kg) with dry mass reduction is as follows:
YCO2 = 489.3 (±17.9) x DryRed (adjusted R2=0.79) (5-4)
where
YCO2 = CO2 yield (as g C / dry kg)
DryRed = dry weight reduction (expressed as a fraction).
5-14
-------
Section 5.0 Solids Decomposition During Composting
Equation 5-4 was developed with a zero intercept so no carbon dioxide will be produced
at zero dry matter reduction. Equation 5-4 indicates that the maximum amount of carbon
dioxide theoretically produced from a substrate, if all volatile solids were completely
decomposed (100% reduction), would be approximately 490 g C/dry kg. However, the
maximum dry matter reduction is less than 100% for all substrates tested in this report,
indicating the formation of refractory organic residuals, such as humic matter and microbial
biomass, that cannot be further decomposed and accumulated (Haug, 1993). This was also
indicated by Equation 5-2, which predicts that the total decomposition of volatile solids during
composting cannot exceed 80%. Equation 5-4 already accounts for the accumulation of biomass
and humic matter and was developed based on a dry mass reduction range of 15.4% to 65.5%
(see Table 5-2).
Dry mass reduction of a substrate during composting can be calculated on a constant ash
basis by knowledge of the volatile solids contents of the starting and composted materials, as
discussed in Komilis and Ham (1999). Using Equation 5-3, cumulative CO2 production can be
estimated this way at different locations within a composting facility.
5.1.4 Solids Degradation Rates
The objective of this section is to discuss solids decomposition rates during composting
of grass and seeded mixed paper. Intermediate solid samples were collected and analyzed for
two special runs performed under conditions similar to those used for results presented
previously. The organic component concentrations, expressed as a fraction of the initial dry
mass, are shown in Figures 5-4 and 5-5 for grass and mixed paper, respectively.
As Figure 5-4 illustrates, approximately 80% of the overall VS reduction of grass
observed throughout the process occurred during the first 5 days, which coincides with the peak
CO2 production rates observed during that time. The initial high VS reduction is primarily due
to the decomposition of cellulose. Overall cellulose reduction during the process was 88%, with
71% reduction occurring during the first 5 days. Glucose—the primary hydrolysis product from
cellulose degradation—appears to be rapidly consumed, since no glucose was detected in HPLC
analyses of the hot water extracts during any of the five decomposition sampling stages for
grass.
The decomposition rate of each chemical group was calculated as the percentage of the
loss of each component over time (i.e., mass of component decomposed over a time period /
initial mass of component), as shown in Table 5-4. Table 5-4 indicates that initial cellulose
degradation rates were 12.6% per day during the first 5.6 days and were the highest among all
solid components. The cellulose decomposition rate between day 5.6 and day 14.7 was 0.91%
per day, after which it stabilized at 0.7% per day until the end of the run. The relatively
constant cellulose degradation rates after day 5 indicate that cellulose enzymatic hydrolysis was
not a function of the cellulose concentration any more, but rather a function of other factors such
as the crystalline structure of residual cellulose or availability of cellulose to hydrolysis if
limited by lignin or other sheathing. Cellulose appears to have reached a constant final value of
8.6% (dw) at the end of the run. This could be due to a suggested binding of polysaccharides to
humic or fulvic acids (Hanninen et al., 1995).
5-15
-------
Section 5.0
Solids Decomposition During Composting
c 1-0
o
'•6 0.9
CD
*= 0.8
CD CO
CO
£ CD 0.7
CD p
o ^ 0.6
1 i °-5
8 ;1 0.4
» 0.3
0.2
0.1
0.0
c
g
—•— Dry matter
-•- HWSM
-A- Fats/Lipids
—V~ Cellulose
—O— Hemicellulose
—I— Lignin/humus
-X-Volatile solids
—*—CO,
30
28
26
24
22
20
18
16
14
12
10
8
6
4
2
0
10
15 20 25 30 35 40 45
Days
50
CD
D)
^^
^^^
5s
O
o
O)
Figure 5-4. Concentrations of organic constituents and CO2
production rate during composting of grass.
The easily degradable cellulose has apparently been degraded by day 5; hemicellulose
degradation rates were 6.7% per day and 5.3% per day during the first 15 days and decreased to
0.74% per day and 0.37 per% per day until the end of the run, as shown in Table 5-4.
Hemicellulose had therefore relatively constant decomposition rates during the first 15 days,
after which it continued to degrade more slowly as it approached complete loss. Hemicellulose
reduction was 93.8% by the end of the run.
HWSM, fats/lipids, and dry matter degraded at similar rates as hemicellulose during the
first 5 days. Dry matter rates stabilized to relatively constant values after day 5.6, as was also
observed for cellulose. An increase in the decomposition rate of the fats and lipids was
observed after day 15, but the concentration of that chemical group was very low. The increase
of HWSM at the end of the process is attributed to the generation of water-soluble humic matter,
probably of fulvic acid origin. This is likely since fulvic acids are generated during the
composting process (Riffaldi et al., 1986) and are soluble in all pH ranges (Stevenson, 1994).
5-16
-------
Section 5.0
Solids Decomposition During Composting
0.0
0 5 10 15 20 25 X 35 40 45 50
Days
Figure 5-5. Concentrations of organic constituents during
composting of seeded mixed paper.
5-17
-------
oo
Table 5-4. Solids Decomposition Rates (As Percent of Dry Mass Decomposed/Initial Dry Mass/Day)
During Composting of Grass and Seeded Mixed Paper
Grass
Period (days)
0-5.6
5.6-14.7
14.7-19.8
19.8-27.0
Dry matter
6.14
0.43
0.62
0.40
Fats/lipids
7.17
-3.60
4.19
5.33
HWSM
6.96
0.51
1.13
-3.98
Cellulose
12.58
0.91
0.71
0.70
Hemicellulose
6.66
5.32
0.74
0.37
Lignin/humus
2.25
-1.20
1.68
0.22
Mixed paper
Period (days)
0-10
10-18
18-45
Dry matter
0.39
3.37
0.84
Fats/lipids
-22.74
-0.81
5.51
HWSM
1.58
1.99
0.40
Cellulose
0.61
4.87
1.05
Hemicellulose
10.00
-4.20
-0.01
Lignin/humus
-5.29
3.33
0.61
-------
Section 5.0 Solids Decomposition During Composting
Lignin/humus decomposition rates were highest during the first 5 days. A slight net increase of
the lignin/humus fraction was observed from day 5 to 15, which can be explained by a net
production of humus and biomass, accounted for in the lignin/humus fraction. Lignin,
however, did not change much during the process, having an overall net reduction of 12.1%
(see Table 5-2).
Decomposition rates for mixed paper are included in Table 5-4 and shown in
Figure 5-5. For mixed paper, the initial cellulose decomposition rate was 0.61% per day,
approximately 1 order of magnitude less than the initial decomposition rate for cellulose in
grass. Cellulose decomposition rates increased between days 10 and 18 to 4.87% per day and
decreased to 1.05% per day after day 18, which is a rate similar to the rate of cellulose
decomposition in grass during the latter stages of composting. The initial cellulose persistence
is probably because of the lag time required for the acclimation of cellulolytic microorganisms.
Although mixed paper was seeded, the type and concentration of seed is expected to affect
decomposition rates, as discussed earlier. A 30% residual cellulose content appears to remain
at the end of the experiment, indicating potential residual cellulose binding to humus or lignin.
The structure of cardboard might also retard mixed paper decomposition. Cardboard consists
of sheets of paper attached together, which could make the invasion of the microorganisms
between the sheets difficult.
Hemicellulose was highly degradable in the first 10 days with a decomposition rate of
10% per day. A slight increase in hemicellulose was detected after day 18 and is attributed to
the synthesis of bacterial and fungal slimes and gums that are known to contain hemicelluloses
(Tenney and Walksman, 1929).
HWSM decomposed slowly and at relatively constant rates during the first 18 days.
Fats and lipids increased during the first 18 days, probably due to the generation of cells and
cell lysis byproducts, which have a lipophilic nature. A slight decrease of fats and lipids was
observed after day 18, indicating decomposition. An increase of lignin/humus was observed
initially and reduction prevailed thereafter.
Dry mass reduction rates in mixed paper were similar to the rates of cellulose reduction
since that is the dominant organic component. It is noted that this mixed paper run was
performed at a mesophilic temperature, while an ascent from mesophilic to thermophilic
temperature was used for the MXP run shown in Table 5-2 for the first 20 days after which
thermophilic temperatures were maintained. The net overall reduction extents for the mixed
paper run shown in Figure 5-5 were 53% for dry mass, 73.2% for cellulose, and 66.1% for
hemicellulose; increases were 85% for fats/lipids and 9.9% for lignin/humus. Therefore, mixed
paper was decomposed at higher extents in mesophilic compared to thermophilic temperatures
and had a larger dry mass reduction extent compared to the grass.
This degradation profile of mixed paper is similar to observations made by Inoko et al.
(1979), who followed solids decomposition in city refuse during composting in actual
composting plants. Inoko et al. (1979) observed that hemicellulose was the component that
decreased at the highest rates among all other solid constituents during the first 10 days of
composting. Hemicellulose concentrations (in % dry mass basis) remained constant thereafter.
They observed that cellulose percentages (expressed on a dry matter basis) increased during the
-------
Section 5.0 Solids Decomposition During Composting
first 10 days, indicating that negligible cellulose decomposition actually occurred during that
time. Cellulose reduction started to occur after approximately day 10, similar to that observed
for the mixed paper run illustrated in Figure 5-5. The similarities are probably because city
refuse is expected to contain high percentages of paper.
5.1.5 Compost Maturity Indicators
Whether finished composts derived from different initial substrates can have certain
similar properties is a question of interest and a topic of research for decades (Mathur et al.,
1993). Such similarities would aid in deriving commonly accepted compost maturity
indicators. This would particularly apply to MSW for which differences in composition exist
among countries or even municipalities.
The objective of this section is to discuss certain properties of the composted substrates
used in these runs as to whether they could be used as potential compost maturity indicators.
The average carbon dioxide production rate during the last 10 days of each of the runs, shown
in Table 5-2, was selected as the basis for comparing the decomposition status of each
substrate. It was therefore assumed that runs with low final average CO2 production rates are
closer to maturity than runs with higher final CO2 production rates. This assumes that CO2
production rates are not limited by factors such as nutrients, moisture, or seeding but only by
the solids composition, that is the presence of refractory organics.
Based on the average values of CO2 production rates, all runs were ranked from the
most "mature" (top of Tables 5-2 and 5-5) to less "mature" (bottom of Tables 5-2 and 5-5). As
Table 5-2 shows, mixed paper (MXP) had one of the lowest extents of VS reduction, but is
defined as mature due to the minor CO2 production rates recorded at the end of the run. As
discussed earlier, inadequate seeding may be the reason for the low decomposition rate for that
substrate. It was estimated in Chapter 4 that this component had reached 71% of its ultimate
CO2 yield at the time of termination of that run. The ultimate yield for that run would have
been reached sooner had mixed paper been combined with other more degradable substrates.
Table 5-2 shows that the most mature component is MXPns since it had negligible
degradation due to lack of seed. The MXPns run was not included in the analysis to follow
since decomposition was not induced for that run. Table 5-2 shows that office paper, grass, and
the MSW1/3 had still relatively high CO2 production rates at the time of the termination of the
runs, indicating that they had not reached complete maturity by that time. The grass and
MSW1/3 runs were terminated due to time constraints; it was estimated that they had yielded
more than 90% of their CO2 ultimate yield based on first order modeling in Chapter 4.
To derive indicators that could be useful for determining compost maturity, linear
regressions were made between the average final CO2 production rates for each run (Table 5-2)
and various parameters for each substrate. The parameters were total reduction extents (in %)
for the five chemical groups, the total organic carbon reduction, the C/N ratio (%), and the
cellulose-to-lignin ratio (%) during composting. Linear regressions were also made with the
concentrations of the five chemical groups in the composted (finished) substrates (expressed as
5-20
-------
Table 5-5. Correlation Matrix for Estimation of Compost Maturity Indicators Using the
Composted Substrate Chemical Composition
o'
s
Substrate3
MXP'ns
MXP
Leaves
MXP/YW
Seed
MXP/YW/FW
FW/YW
YW
YWh
FW/MXP
FW
Branches
Ywns
FWns
MSW1/3
Office paper
Grass
Correlation
coefficient (R2)
Cured MSW
compost8
Fats/
Lipids
(% VS)
0.8
1.8
1.4
1.8
1.4
2.3
3.5
2.5
0.9
2.1
0.9
2.2
7.4
1.0
0.0
1.3
0.05
0.9%
HWSM
(% VS)
5.9
22.3
6.5
11.8
7.6
18.7
8.9
11.0
0.0
17.5
8.7
10.9
24.4
10.1
3.6
36.6
0.13
7.2%
Lignin
content
(% VS)
37.7
62.2
50.0
51.6
51.0
60.3
63.2
61.7
35.2
27.8
58.3
52.9
15.2
52.7
7.4
41.9
0.17
46.4%
Cellulose
(% VS)
57.4
9.9
38.1
31.1
26.4
13.9
17.7
11.0
53.9
52.1
34.3
9.7
51.5
29.1
81.4
12.9
0.024
24.1%
Hemicel-
lulose
(% VS)
3.6
1.9
2.7
6.1
3.6
1.9
1.3
0.0
0.0
0.0
16.8
0.3
0.0
4.9
0.0
2.8
0.0
4.2%
VS
(% dw)
88.5
54.7
72.5
65.0
71.5
52.8
51.0
56.2
86.7
79.7
78.3
50.0
88.1
70.4
82.7
66.9
0.026
63.1%
C
(% dw)
41.6
20.2
38.1
33.3
36.1
25.8
25.2
28.5
42.5
39.7
39.2
26.1
43.8
35.5
36.8
33.3
0.017
31 .4%
C/N ratio"
21.31
7.77
18.88
11.82
15.78
9.62
7.60
11.15
20.63
17.87
17.26
8.17
15.55
N/m
N/m
N/m
0.002
Cellulose /
lignin ratio
1.52
0.16
0.76
0.60
0.52
0.23
0.28
0.18
1.53
1.87
0.59
0.18
3.39
0.55
11.01
0.31
0.004d
0.52
Substrates are placed in ranking order from most mature (top of table) to least mature (bottom of table) as judged from the average
CO2 production rates during the last 10 days of each run.
No N measurements were performed for the grass, MSW1/3, office paper, and the cured MSW compost.
Cured MSW compost chemical composition is given for comparison and was not included in the statistical analysis.
Office paper excluded from analysis (see text).
No solids measurements were performed for the MXPns run.
o
s
o
'
-------
Section 5.0 Solids Decomposition During Composting
% of the volatile solids content), shown in Table 5-5. Regressions with final C/N ratio and
final cellulose-to-lignin ratios were also made (Table 5-5). The linear regressions provided
coefficients of determination (R2) that were used to compare parameters in their ability to
predict compost maturity. These correlation coefficients are included in Tables 5-2 and 5-5;
they are useful only for comparison and are not meant to indicate the accuracy of predictive
models
In addition to the above and as part of comparing the laboratory results with values
observed in the field, cured compost was collected from an MSW composting facility near
Portage, WI, and its solids composition measured. The chemical composition of the cured
MSW compost is included in Table 5-5.
The correlation coefficients of Table 5-2 indicate that the C/N ratio reduction and the
fats/lipids reduction correlate more strongly with the final CO2 production rates than other
reduction parameters, as illustrated in Figure 5-6.
A reduction in C/N ratio is generally observed as the substrate matures. This is evident
from Figure 5-6 as well as from the initial and final C/N values for the various substrates
shown in Tables 5-2 and 5-5. Generally, the C/N ratio decreases during composting due to
greater reduction of carbon compared to the reduction of nitrogen (Epstein, 1997). Increases of
C/N ratio during composting were observed only for food wastes (an increase from
approximately 8.0 to approximately 16.0 at the end of the process). This is due to the high
initial N content of food wastes. The reduction of C/N ratios during composting is generally
well-documented (Mathur et al., 1993). Consideration of both the initial and final C/N ratios
are required to calculate this reduction.
A C/N ratio of 10 has been suggested as a compost maturity indicator (Mathur et al.,
1993). However, this was not shown in the above analysis. There is a generally wide range of
C/N ratios for composted substrates, such as 7.8 for composted leaves and 21.3 for composted
mixed paper (MXP) and no particularly good correlation was observed between the CO2
production rates and the C/N ratio (as will be shown in Figure 5-8). Food-waste-derived
compost C/N ratios are higher than 10 due to the relatively high residual carbon concentration
present. Most of the mixed paper mixtures, however, had C/N ratios between 15 and 20, also
due to a high residual carbon, mostly in the form of residual cellulose. Yard wastes and leaves
had values lower than 10.
The reduction of fats and lipids during composting also had a stronger correlation than
other parameters. However, as Figure 5-6 shows, this higher correlation is governed by the
three data points with the highest residual CO2 production rates, while the other data points are
clustered together. The indicated increase of fats and lipids during composting might be due to
the generation of biomass and cell lysis byproducts that are lipophilic. An increase of fats and
lipids during composting was also shown earlier for mixed paper composting. Fats/lipids
reduction extents should be used with caution as compost maturity indicators because of the
relatively small concentrations of that organic group in both initial and final substrates.
5-22
-------
Section 5.0
Solids Decomposition During Composting
110
100
90
80
c
o
•| 7°
T3
2 60
50
40
30
Fats/lipids reduction
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
C-CO / dry kg /day
4.0
120
100
80
60
40
20
0
-20
-40
-60
-80
-100
-120
**
C/N ratio reduction
0.0 0.5 1.0 1.5 2.0 2.5 3.0
C-CO / dry kg /day
3.5
4.0
Figure 5-6. Correlation plots between final average CO2 production
rate and fats/lipids reduction and C/N ratio reduction.
5-23
-------
Section 5.0 Solids Decomposition During Composting
Based on Table 5-5, final lignin/humus (as % VS) and final HWSM (as % VS) correlate
more strongly with final CO2 production rates than with other parameters describing the final
waste composition. The correlation plots for lignin/humus and HWSM are included in
Figure 5-7. Also shown are correlation plots for the other three chemical constituents
(cellulose, fats/lipids, and hemicellulose). The final lignin/humus content appears to be a
reasonable indicator of the concentration of the stabilized organic matter in compost, since
it includes the slowly degradable lignin as well as the generated humic matter. Expressing
lignin/humus content on a VS basis rather than a dry matter basis gives a better quantification
of this organic group because the composted dry matter may have a high inorganic content,
which is not degradable. The average lignin content for all composted materials shown in
Table 5-5 is 48% (±3.4%) on a volatile solids basis. Most of the final lignin/humus
concentrations, however, are above a threshold value of 50% (VS basis). Cured MSW, in
particular, had a lignin/humus content of 46.4% (VS basis).
Figure 5-7 indicates that the lignin/humus concentration increases with composting. As
materials mature, final lignin/humus concentrations appear to become greater than 50% (% VS)
and approach a value of between 50% and 60% (VS). A threshold value of at least 50% (VS
basis) for lignin/humus is suggested as a compost maturity indicator based on this study.
The FW and FWns runs had the lowest lignin contents in their composted substrates at
27.8% and 15.2% (VS basis), respectively. The low lignin values for food wastes are probably
because lignin in food wastes is more degradable than lignin in the other MSW substrates or
because the lignin/humus in composted food wastes was more soluble in the 72% sulfuric acid.
In addition, humification might not have proceeded to a great extent in food wastes, making
mineralization the primary pathway for conversion of the organic carbon found in the primary
solid constituents. The composted substrates from the FW/MXP and MXP runs had lignin
contents less than 40% because of the high residual cellulose content found in both composted
substrates. Office paper had also not reached its full extent of degradation, as shown in
Table 5-2. Therefore, the low decomposition rates probably resulted in limited formation of
humic matter and biomass, so the residual lignin/humus content was only 7.4% (VS basis).
As shown by the correlation analysis, the extent of lignin/humus reduction did not
provide a reasonable correlation and cannot be used as a compost maturity indicator.
The relatively high final HWSM content correlation with CO2 production rates is
governed primarily by the three least mature substrates (office paper, MSW1/3, and grass). This
is attributed to the formation of water-soluble humic matter, since these substrates still had not
decomposed completely at the end of the runs. As Figure 5-7 shows, HWSM final contents
range between 5% and 25% for the more mature substrates, while generally an increase of the
HWSM content is observed during composting. Despite the correlation, HWSM final contents
may not be suitable to indicate compost maturity due to their wide range in mature substrates.
Figure 5-7 also shows that a reduction in fats/lipids, cellulose, and hemicellulose occurs
during composting. More apparent is the reduction of hemicellulose during composting, with
final hemicellulose contents close to 0 for most of the substrates. Food wastes (unseeded)
contained no hemicellulose in both the initial and final substrate.
5-24
-------
Section 5.0
Solids Decomposition During Composting
0.5
0.4
0.3
"o
w o.2
> 0.1
0.0
Lignin/humus
'(/>
.2
p
"o
JD
"5
^
0.9
n Q
U.o
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
'_ D Initial |
• Final [
%V
y u •
.
• .
" ID" n
n •
Ecu. D D -
' n "
_
i . i . i . i . i . i . i . i .
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
g C-CO2 / dry kg / day
HWSM
D Initial
• Final
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
g C-CO2 / dry kg / day
Figure 5-7. Correlation plots between final average
production rates and initial and final solids
chemical composition (% VS basis).
(continued)
5-25
-------
Section 5.0
Solids Decomposition During Composting
Fats
u.o
0.4
ro 03
"5
w o.2
JD
iS
0
> 0.1
s£
0.0
D Initial | •
• Final |
_
•
-
n D
-
n '-'
. & tf# . D i-
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
w~wv^/ / ui i\u / udV
gC-CO2/dry kg 7 day
Cellulose
.52
co
r2
"o
.CD
"o
^
I .U
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
I
n n
D
H
% • D
n " D "
gl n
~ • " "
• n
- ?• • • -
- :
0.0
0.5
1.0
1.5 2.0
2.5
3.0
3.5
4.0
g C-CO2 / dry kg / day
Figure 5-7. (continued)
5-26
-------
Section 5.0
Solids Decomposition During Composting
Hemicellulose
1.0
0.9
0.8
.w 0.7
co
.0 0.6
~ 0.5
O
CD 0.4
0.2
0.1
0.0
D Initial
• Final
0.0 0.5 1.0
1.5
2.0
2.5
3.0
3.5
4.0
Figure 5-7. (concluded)
Although cellulose-to-lignin ratio (C/L ratio) in finished composts did not show any
strong correlation, a decrease of this parameter was generally observed during composting (see
Figures 5-8 and 5-9). Because this ratio has been used as an indicator for characterizing fresh
and matured solid wastes in anaerobic environments (Bookter and Ham, 1982), it will be
discussed here. Figure 5-8 shows the correlation plot between residual CO2 production rates
and initial and final C/L and C/N ratios. Figure 5-9 shows the initial and final C/L ratios for all
substrates in a bar graph format. Based on Figure 5-8, it appears that most substrates approach
a constant C/L ratio less than 0.5 during composting. The reduction of this ratio is explained
by the fact that, although lignin does partially decompose during the process, the cellulose
reduction is greater than that of lignin, resulting in a reduction of the C/L ratio from an average
initial value of 2.3 (±0.4) (Table 5-1). Higher than average initial C/L ratios were recorded for
food waste and wastes containing mixed paper. Office paper is not shown in Figure 5-9
because of its significantly high C/L ratios compared to the other substrates. Yard wastes and
its subcomponents had relatively low initial C/L ratios.
The composted food wastes had a relatively high C/L ratio compared to other
composted materials. This is a result of the low lignin content of composted food wastes and
the relatively high residual cellulose.
The increase of C/L ratio for branches might be explained by the uniquely low cellulose
and high lignin contents of this material. In addition, branches had an increase in lignin during
composting, probably due to a net increase in lignin/humus as a result of minimal degradation
of lignin in branches and the generation of humus from both lignin and cellulose. This was
also true for the mixed paper run.
5-27
-------
Section 5.0
Solids Decomposition During Composting
O
30
25
20
15
10
C/N ratio
D
. D
D Initial
• Final
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
g C-CO2 / dry kg / day
Cellulose / lignin ratio
12
10
c
O) g
-------
Section 5.0 Solids Decomposition During Composting
MSW-derived cured compost had a C/L ratio of 0.52, which is close to the value that
most of the substrates appear to converge to during composting, as shown in Figure 5-8.
Based on the above it appears that a C/L ratio less than 0.5 could be used as a maturity
indicator for various substrates. Ratios of approximately 0.8 have been recorded for 8-year-old
refuse in landfills, and ratios of 0.16 to 0.24 have been recorded in older landfills (Bookter and
Ham, 1982). In addition, Bookter and Ham reported fresh refuse with C/L ratios up to 4.04, a
value close to the initial C/L ratio of several of the substrates used in this experiment. It should
be noted that Bookter's analytical methods were different than those used in this paper,
probably overestimating the amount of cellulose measured in his substrates.
The above analyses were based on regressions between the final average CO2
production rates and final chemical composition of all substrates. The goal of the regressions
was to compare potential compost maturity indicators and not to provide adequate predictive
models. It is recommended that the above analysis be performed based on regressions of
chemical composition and CO2 production rates of materials not only at the end of a run—as
was done here—but by using values throughout a run, based on intermediate sampling.
5.1.6 Errors in Analytical Measurements
The objective of this section is to evaluate sources of error in analytical measurements
and to consider mass balances. Eleazer et al. (1997) have suggested that, in the case of
anaerobic solid waste treatability studies, the ratio of the sum of cellulose, hemicellulose, and
lignin to volatile solids concentration (all expressed on a dry weight basis) should equal 1.0.
This was suggested by assuming that only cellulose, hemicellulose, and lignin are the dominant
organic materials in the substrate. However, based on the methods used here, the dry mass of
solid waste components was further fractionated to two additional chemical groups: fats and
lipids and the hot water soluble matter. If only these five groups of organics are present in the
substrate, the ratio of their sum to volatile matter concentration (all expressed on a dry mass
basis) should equal 1.0. Table 5-1 shows that this ratio (defined as FHCHL/VS) ranges from
0.88 for food wastes to 1.07, with an average of 0.964 ± 6.7%, for all starting materials. The
relatively low value for food wastes (0.88) is attributed to dissolution of organic groups other
than polysaccharides during acid hydrolysis as discussed previously. The low FHCHL/VS
ratio for leaves (0.80), might be explained by partial dissolution of lignin, since leaves were
already partially decomposed prior to initial analyses, making the lignin fraction more
susceptible to acid solubilization (Effland, 1977). Table 5-6 includes the FHCHL/VS ratio for
the composted materials. Most of the values are close to 1.0, indicating good accuracy in the
determination of the five chemical groups. The low value for YWns (0.76) and the high value
for branches (1.19) are attributed to analytical errors. The fact that the FHCHL/VS ratio for
leaves increased from 0.80 in the starting material to 0.98 in the composted material might
indicate the polymerization of compounds to humic compounds that are acid insoluble and
quantified in the lignin/humus fraction. The average FHCHL/VS value from the composted
substrates is 0.955 ± 9.9%, a value very similar to the FHCHL/VS ratio reported for the starting
materials.
5-29
-------
Section 5.0
Solids Decomposition During Composting
03
5.0
4.5
4.0
3.5
.g 3.0
C
j? 2.5
(U 2.0
I 15
o 10
0.5
0.0
Initial
Composted
LI
1
1 UT>^
* Office paper am not included due to high C/L values; no
measurements were performed for the MXR^ run
Figure 5-9. Cellulose to lignin (C/L) ratios for initial and composted materials from
15 experimental runs (plus compost from a MSW composting facility).
Table 5-6 also presents the carbon and nitrogen mass balance closures performed for all
runs. The yields of carbon and nitrogen emitted during the process were quantified by CO2 and
NH3. Carbon recovery values ranged from 85.5% to 117.9%, where recovery is defined as the
amount of C captured as CO2 divided by the carbon reduction in the solid phase. The average
carbon recovery value was 100.5% ± 9.91% from all runs. Nitrogen recoveries had large
variabilities ranging from 5.7% for MXP/YW/FW to 173.7% for yard wastes, where recovery
is defined as the amount of N emitted as NH3 divided by the nitrogen (TKN) reduction in the
solid phase. Nitrogen added as nutrients was accounted for in the calculations. If nitrates were
unavailable to the microbial population, and since they would not be measured during the final
TKN analysis, nitrogen recoveries less than 100% were recorded, which may partially explain
the low N recoveries for the mixed paper runs.
5-30
-------
Section 5.0 Solids Decomposition During Composting
Table 5-6. Carbon, Nitrogen, and Solids Recoveries
Contents C recovery (%)a N recovery (%)b FHCHL/VSc
FW
r vvns
FW
YW
1 vvns
YW
MXPns
MXP
MXP/YW
FW/MXP
FW/YW
MXP/YW/FW
MSW1/3
Seed
YWh
Grass
Leaves
Branches
Office paper
Cured MSW compost
108.5
117.9
98.6
94.4
NM
93.2
90.2
115.9
109.0
91.8
95.0
85.5
101.7
91.0
100.7
114.3
100.0
77.5
72.7
173.6
157.2
NM
33.9
7.4
14.5
86.4
5.7
NM
32.2
51.8
NM
-23.3
37.9
NM
0.98
0.99
0.76
0.95
NM
1.05
0.99
0.90
0.97
0.90
0.98
1.02
0.86
0.95
0.98
1.19
0.92
0.83
3 CO2 yield (as C) / (total C in starting solid material - total C in composted solid
material).
b NH3 yield (as N) / (total N in starting solid material + N added as nutrients - total N in
composted solid material).
c The ratio of the sum of fats, HWSM, cellulose, hemicellulose, and lignin/humus
concentrations to the volatile solids concentration.
NM = Not measured.
5-31
-------
Section 6.0 Emissions of Volatile Organic Compounds During Composting
6.0 Emissions of Volatile Organic Compounds
During Composting
Certain runs presented in Chapters 4 and 5 were selected and used to provide VOC
emissions data. VOC emissions were quantified for 11 runs. These runs were: seeded mixed
paper (MXP), unseeded mixed paper (MXPns), seeded yard wastes (YW), unseeded yard wastes
(YWns), seeded food wastes (FW), unseeded food wastes (FWns), a seeded mixture of mixed
paper and yard wastes (MXP/YW), a seeded mixture of mixed paper and food wastes
(MXP/FW), a seeded mixture of food wastes and yard wastes (FW/YW), and an unseeded
mixture of food waste, mixed paper, and yard wastes (MSW1/3). A run with only seed (obtained
from an MSW composting facility) was also performed. Seed was added at approximately 10%
of the component's weight (on a dry weight basis), and the relative amounts of each component
in the mixture runs were set according to their relative amounts in the MSW.
VOCs emitted were captured on activated coconut charcoal traps and extracted with
carbon disulfide (CS2). Compounds were identified using GC/MS and quantified using a GC
equipped with a flame ionization detector. Detailed methods are presented in Chapter 3.
The compounds identified may vary depending on the type of sorbent used during
capture. The activated coconut charcoal used can effectively capture a wide range of organic
molecules, but especially VOCs with two to five carbon atoms (Supelco, 1997). The porous
surface of activated coconut charcoal makes it a more efficient sorbing material compared to
graphitized carbon black. Wilkins (1994) and Tolvanen et al. (1998) used Tenax tubes, Eitzer
(1995) used Tenax and graphitized carbon black traps, and Kim et al. (1995) and Brown et al.
(1997) used charcoal traps.
The same experimental setup was also used to quantify emissions after spiking known
amounts of VOCs. VOCs were spiked onto different MSW substrates in both liquid and vapor
phases and under mesophilic and thermophilic temperatures. Two runs were performed by
spiking ethylbenzene in liquid form to seeded newsprint and then to a mixture of yard wastes.
Spikes were introduced through the exit port at the top of each digester. This was designed to
simulate the introduction of VOCs originally introduced in MSWs through ruptured containers
or broken bottles.
A mixture of toluene, ethylbenzene, w-xylene, and o-xylene (TEX) was spiked to
simulate an MSW (mixture of food wastes, yard wastes, and mixed paper) in the vapor phase. A
125-mL glass bulb was connected to the tubing that directed air into the digester. Through a
rubber septum located on the glass bulb, a known volume of a TEX liquid standard solution,
prepared in methanol, was injected into the bulb via an airtight syringe. Both the digester and
the glass bulb were kept in the thermophilic temperature range from 50 to 55 °C. The VOC
mixture was rapidly vaporized at such temperatures and was carried by the air stream into the
-------
Section 6.0 Emissions of Volatile Organic Compounds During Composting
MSW mixture. This setup might simulate the case where a VOC present in the center of a
continuously aerated windrow will volatilize due to the development of thermophilic conditions
and move through the compost mass by advection and/or diffusion.
The air flow rate was kept at approximately the same levels during the three spiked runs.
During the spiked runs, VOCs were analyzed with GC/FID, using the internal standard
technique.
6.1 Results and Discussion
6.1.1 Identification of VOCs from Biodegradable Fraction of MSW
Results of the identification of VOCs for 9 of the 12 runs are presented in Table 6-1. One
Orbo charcoal trap from each run was used for VOC identification by GC/MS. Most charcoal
tubes used for GC/MS analysis were collected during the first 5 days of the composting of a
substrate, since most VOC volatilization was observed to occur during that time, as was shown
by preliminary runs not presented here. In the case of the MXP run, the sample used for GC/MS
was placed on the digester between day 29 and day 48 from the initiation of composting. This
was done to collect VOCs for identification over the thermophilic temperature range used after
day 29. Prior to day 29, a gradual ascent from mesophilic to thermophilic temperatures was used
for that run only. For the MXP/YW run, the sample used for GC/MS VOC identification was
collected between days 10 and 28 because of partial volatilization of the samples until day 10 for
that run. The time period for VOC sampling by the charcoal tube used for VOC identification
might be important because it reflects the decomposition stage for that run during that time
period.
No GC/MS measurements are given in Table 6-1 for the seeded yard wastes (YW), the
food waste and yard wastes (FW/YW), or the (MSW1/3) runs. VOCs were not identified for these
runs because the compounds emitted were not expected to differ from those emitted from the
individual components that made up the mixtures or from the unseeded yard wastes.
Table 6-1 presents only the VOCs with a purity index (PI) higher than 70%, which is
close to the common threshold value of 80% used by the State Laboratory of Hygiene (Madison,
WI) for considering a compound as positively identified by MS only (personal communication.
John Mathew, Director of GC/MS laboratory). Note that a purity index of 100% indicates the
mass spectrum of a known standard organic compound (acquired from pertinent mass spectra
databases) precisely matches the mass spectrum of an organic compound found in the mixture.
VOCs with a purity index less than approximately 70%, although reported by the GC/MS, were
not included in Table 6-1. Coelution of compounds usually confounds proper identification of
compounds. Twenty-five VOCs were further targeted and tentatively identified using both
GC/MS and retention times. Discussion of the selection of the 25 VOCs follows in a later
section.
Table 6-1 presents the fraction (RAC) of the area counts of that compound, as measured
during the GC/MS analysis, to the total area counts of all compounds identified in the charcoal
6-2
-------
Section 6.0
Emissions of Volatile Organic Compounds During Composting
Table 6-1. VOCs Identified in Gaseous Emissions from Nine Runs"
Unseeded mixed paper (MXPns) RACb
Cyclopentasiloxane, decamethyl 5.0
Dodecane 4.9
2-ethyl-1 hexanol 4.1
1,2,4-trimethylbenzene 4.0
Cyclotetrasiloxane, octamethyl 3.7
Undecane 3.3
Toluene 2.9
Naphthalene 2.6
Acetic acid, trifluoro-, octyl ester (or) 1 octanol 2.5
1 ethyl-4-methyl benzene 2.4
3-cyclohexene-1-methanol, alpha 2.0
2,6 dimethyldecane 1.8
Ethylbenzene 1.6
Benzene, 1-methyl-2-(1-methylethyl) 1.6
3,8 dimethyl decane 1.6
1-ethyl-2,3-dimethyl benzene 1.5
2 methyl decane 1.5
Camphor 1.4
2,6 dimethyl undecane 1.4
1,2-dimethoxy-benzene 1.3
4-methyl undecane 1.3
Cyclotrisiloxane, hexamethyl 1.2
Docosane 1.2
1 dodecene 1.1
1,4-dichlorobenzene 1.1
Benzene, 1-methyl-2/4-(1-methylethyl) 1.1
Cyclopentane, butyl 1.0
3,4 dimethyl undecane 1.0
Cyclohexene, 1-methyl-4-(1-methylethyl) (or)
Limonene 0.9
Cyclohexanone 5-methyl-2-(1-methyl-) 0.9
Cyclohexanol, 5-methyl-2-(1-methyl) 0.9
Decane 0.8
2-butoxy-ethanol 0.7
• -pinene 0.7
1,3,5-trimethyl benzene 0.7
p-xylene 0.7
1-ethyl-2-methyl-benzene 0.6
1,2,3-trimethylbenzene 0.6
Camphene 0.5
Propyl benzene 0.5
Beta ninene 0.4
Seeded mixed paper (MXP) RAC
2-ethyl-1-hexanol 60.0
2 cyclohexen-1-one, 3,4,4-trimethyl 3.7
Benzole acid, methyl ester 2.5
1,2-dimethoxy-benzene 2.2
Cyclotrisiloxane, hexamethyl 1.9
Toluene 1.5
Seeded mixed paper (MXP) (cont.)
Undecane 1.1
Limonene 1.0
2-pentyl furan 0.9
Cyclopentasiloxane, decamethyl 0.9
Bicyclo[2.2.1] heptan-2-ol, 1,7,7-trimethyl
ester) 0.8
Dodecane 0.8
Camphor 0.8
Naphthalene 0.8
1-octanol 0.7
1,4-dichlorobenzene 0.7
Heptanoic acid, 3,5-dimethyl-methyl (ester) 0.5
3-furancarboxylic acid, 2,5-dimethyl (ester) 0.4
2-methyl 3 hexanone 0.3
2,3-dimethyl butanoic acid, methyl ester 0.3
3-heptanone 0.3
Benzene, 1-methoxy-4-methyl- 0.3
Benzaldehyde 0.3
Benzene, 1-methyl-3-(1-methylethyl)- 0.2
3-penten-2-one, 3-methyl 0.2
Ethylbenzene 0.2
p-xylene 0.1
2-4 dimethyl pentanone 0.07
Pyridine 0.03
Unseeded yard wastes (YWns) RAC
A-pinene 13.5
Toluene 12.6
3-methyl-2-pentanone 5.5
Camphene 5.3
Bicyclo [3.1.0]hexane, 4 methylene 4.8
Beta-pinene 4.7
Limonene 4.2
2 cyclohexen-1-one, 3,5,5 trimethyl- 3.5
2-cyclohexen-1-one 3.5
Camphor 2.5
Dimethyl-disulfide 2.4
Cyclohexanone, 2-2-6 trimethyl 2.1
2-Pentyl furan 2.1
Tricyclo[2.2.1.02,6] heptane,
1,7,7 trimethyl 2.0
Tricyclo-heptane 2.0
1,7-Nonadiene, 4,8-dimethyl 1.7
4-methyl-3 hexanone 1.7
Benzene, 1-methyl-3/4-(1-methylethyl) 1.7
1.7-Nonadiene 1.7
Octane 1.2
Cyclohexanone, 3,3,5-trimethyl 0.9
Bicyclo [3.1.0] hex-2-ene, 2 methyl 0.9
Cyclopentane, 2-ethylidene-1,-1 dimethyl 0.9
(continued)
6-3
-------
Section 6.0
Emissions of Volatile Organic Compounds During Composting
Table 6-1. (continued)
Unseeded yard wastes (YWns) (cont.)
Styrene 0.8
Pyrazine 2,5-dimethyl 0.7
2,5-Dimethyl pyrazine 0.7
1,4-Dichlorobenzene 0.6
1,3-Cyclohexadiene, 1-methyl-4-(1-?) 0.6
(+)-2-carene 0.6
Cyclopentasiloxane, decamethyl 0.5
Bicycle [3.1.1 ]heptan-3-one, 2,6,6-trimethyl 0.5
Methyl isobutyl ketone 0.3
Nonane 0.3
3-cyclohexen-1-ol, 4-methyl- 1-(1-methyl?) 0.3
Dimethyl trisulfide 0.2
p/m xylene 0.2
Pyridine, 2-6 dimethyl 0.1
3-cyclohexen-1-methanol, • -, 0.1
Unseeded food wastes (FWns)c RAC
Dimethyl disulfide 10.8
Cyclopentasiloxane, decamethyl 3.5
Dimethyl trisulfide 3.2
Phenylethyl alcohol 3.1
Butanoic acid, 3-methyl, 3-methyl 2.8
Camphor 2.2
2-nonanone 1.7
Benzene, 1-methyl-3-(1-methylethyl) 1.5
Acetic acid, butyl ester 1.4
1-hexanol, 2-ethyl 1.3
Butanoic acid, butyl ester 1.2
Butyl 2-methylbutanoate 1.1
Limonene 1.1
Pyrazine, trimethyl 1.0
1-undecene 0.9
Phenol 0.9
A-pinene 0.4
Butanoic acid, 2-methyl 0.3
Pyrazine, 2-5 dimethyl 0.3
4-heptanone 3 methyl 0.2
Ethylbenzene 0.2
Butanoic acid, 3-methyl 0.2
Pyrazine, 2-6 dimethyl 0.2
Toluene 0.2
2-hexanone, 5-methyl 0.14
Octane 0.09
Butanoic acid ethyl ester 0.08
Butanoic acid, 3-methyl, ethyl ester 0.05
1 butanol 0.01
Toluene 37.6
Dimethyl disulfide 29.2
Cyclotetrasiloxane, octamethyl 3.0
Benzene, 1-methyl-4-(1-methylethyl)b 2.0
Seeded food wastes (FW) RAC
Dimethyl trisulfide 1.5
Cyclopentasiloxane, decamethyl 1.4
A-pinene 1.2
Limonene 1.1
Benzene, 1-methyl-4-(1-methylethenyl) 0.9
1 hexanol, 2 ethyl 0.7
Ethylbenzene 0.4
p-xylene 0.4
Beta-pinene 0.4
Furan, 2-pentyl 0.4
1-butanol, 3 methyl acetate 0.3
Bicyclo[3.10]hexane, 4-methylene 0.3
Methoxy-benzene 0.2
1,4-dichlorobenzene 0.2
1-undecene 0.2
Nonane 0.1
Bicyclo[3.1.0] hex-2-ene, 2-methyl 0.1
Camphene 0.1
Undecane 0.1
Butanoic acid, 3-methyl ester 0.1
2-nonanone 0.09
3-octanone 0.08
Methyl ethyl disulfide 0.06
Mixed paper and food wastes (MXP/FW) RAC
Toluene 9.6
2-pentyl furan (or) 1,2,4-trimethylbenzene 3.7
Cyclopentasiloxane, decamethyl 3.4
Dodecane 3.0
2-heptanone 2.5
p-xylene 2.5
Naphthalene 2.5
1 hexanol, 2-ethyl 2.3
Undecane 2.1
Limonene 2.0
Benzene, 4-ethyl-1, 2-dimethyl 1.9
A-pinene 1.8
2-nonanone 1.8
3-octanone 1.7
1,4-dichlorobenzene 1.7
Benzene, 1-ethyl-2-methyl 1.6
3-cyclohexene-1-methanol, alpha 1.6
Benzene, 2 ethenyl-1,4 dimethyl 1.3
Benzene, 1,2,4,5-tetramethyl 1.3
Benzene 1,2,3,4-tetramethyl 1.3
Undecane, 2-6 dimethyl 1.3
Tetratriacontane 1.1
Benzene, 1-methyl-4-1(1-methylethyl)2 1.1
Camphor 1.1
Ethylbenzene 1.0
(continued)
6-4
-------
Section 6.0
Emissions of Volatile Organic Compounds During Composting
Table 6-1. (concluded)
Mixed paper and food wastes (MXP/FW) (cont.)
Benzene, 2-ethyl-1,4-dimethyl 1.0
Decane 0.9
Decane, 3 methyl 0.9
Octane 0.8
Naphthalene, decahydro-2-methyl- 0.8
Benzene, 4-ethyl-1, 2-dimethyl 0.8
Butyl benzene 0.8
Beta-pinene 0.7
1,2,3-trimethyl benzene 0.7
Benzene, 1-methyl-2-(1-methylethyl) 0.7
Camphene 0.6
Propyl-benzene 0.5
2-heptanone, 6-methyl 0.5
1,3,5-trimethylbenzene 0.5
Benzene, 1-ethyl-3-methyl 0.5
Cyclotrisiloxane, hexamethyl- 0.4
Benzene, 1 methylethyld 0.4
Methane, isothiocyanato 0.1
Disulfide, dimethyl 0.1
Mixed paper and yard wastes (MXP/YW) RAC
Hexanoic acid, 2-ethyl-methyl ester 20.1
Camphor 10.3
2-cyclohexen-1-one, 3,4,4 trimethyl 7.1
2-cyclohexen-1-1one, 3,4,4-trimethyl 7.1
Limonene 5.2
Bicyclo[2.2.1]heptan-2-one, 1,3,3-methyl 5.0
Nonanoic acid, methyl ester 3.3
Cyclopentasiloxane decamethyl 2.9
Cyclopentasiloxane, decamethyl 2.9
1,2-dimethoxy- benzene 2.7
2-pentyl furan 2.1
1-octanol, 2-butyl 1.7
Benzene, 1-methyl-4-(1-methylethyl) 1.6
Toluene 1.6
2-ethyl-1-hexanol(PI53%) 1.4
1,4-dichlorobenzene 1.4
Cyclotetrasiloxane, octamethyl 1.3
Styrene 1.2
A-pinene 0.7
Ethylbenzene 0.7
p-xylene 0.6
Seed RAC
Cyclopentasiloxane, decamethyl 5.1
2-cyclohexene-1-one, 3,4,4 trimethyl 4.6
Camphor 3.7
Limonene 3.5
Toluene 3.1
Naphthalene 2.8
1 hexanol, 2-ethyl 2.5
Cyclotetrasiloxane octamethyl 2.3
1,1-Tricosene 2.0
Dodecane 1.9
Undecane 1.5
Heptanoic acid, dimethyl, methyl ester 1.4
Hexanoic acid, methyl ester 1.3
Bicyclo[2.2.1]hepten-2-one, 1,3,3 trimethyl 1.3
Dimethyl disulfide 1.3
Cyclohexanone, 5-methyl-2-(1 -methyl) 1.3
1,4-dichlorobenzene 1.3
Benzene, 1,2,4,5 tetramethyl 1.3
2 pentanone 1.0
Benzene, 2-ethenyl-1,4-dimethyl 1.0
2,4-dimethyl 3 pentanone 0.8
4-heptanone 0.8
Benzene, 1-methyl-2-(1-methylethyl) 0.8
Styrene 0.8
Methyl isobutyl ketone 0.7
A-pinene 0.7
Decane 0.7
1-ethyl-2-methyl benzene 0.6
p-xylene 0.5
Beta-pinene 0.5
1-ethyl-4-methyl benzene 0.3
1,3,5-trimethylbenzene 0.3
Cyclotrisiloxane, hexamethyl- 0.2
Ethylbenzene 0.2
2 hexanone 0.1
Control run RAC
2-ethyl 1 hexanol 38.3
1-octanol 17.4
Toluene 10.9
Cyclotrisiloxane, hexamethyl 6.2
Cyclotetrasiloxane, octamethyl 5.3
1-hexanol 3.7
Benzene, isothiocyanato 2.7
Cyclotetrasiloxane, decamethyl 1.6
3-dodecene 0.8
Ethylbenzene 0.3
RAC is the ratio of area counts of corresponding
compound to sum of area counts of all compounds
measured in the specific run.
p-lsopropyltoluene.
Several acids and esters of acids with purity index
less than 60%.
Isopropyl benzene.
6-5
-------
Section 6.0 Emissions of Volatile Organic Compounds During Composting
extract. For example, a ratio value of 1.5 indicates that the area counts for that compound
account for 1.5% of the sum of the area counts of all compounds (excluding the solvent)
measured during the specific run. This ratio value is useful only for relative comparisons of the
concentrations of the compounds found in the specific sample. Compounds shown in Table 6-1
are ranked from highest to lowest relative concentrations in each sample, on an area count basis.
Since only compounds at a purity index of 70% or higher are listed, the sums of the RACs for
each substrate are less than 100.
Based on the results shown in Table 6-1, alkanes were frequently identified compounds,
especially in three runs containing mixed paper (MXP, MXPns, MXP/FW). Undecane and
dodecane, in particular, were found at high relative concentrations. Decane and methylated
undecanes were found in unseeded mixed paper only. Decomposition or volatilization of these
compounds might explain their absence from the MXP/YW run, since the charcoal trap used for
VOC identification for that run was placed at an advanced composting stage (after day 9).
Octane was detected in unseeded food wastes and the mixture of mixed paper and food wastes,
indicating that food wastes might be a potential source of this hydrocarbon. All of the
aforementioned alkanes were also detected in mixed waste headspaces but not in biological
waste headspaces by Wilkins (1994).
Several alkylated benzenes were also detected, primarily from the same three mixed
paper runs (MXP, MXPns, MXP/FW). A higher number of alkylated benzenes were present in
the MXPns run compared to the MXP run. This might be attributed to partial decomposition of
such compounds in the MXP run compared to MXPns run. This relates to the respective amounts
of decomposition of the basic substrates in these two runs, since 153.3 g CO2 (as C/dry kg) were
produced during the MXP run compared to only 5.5 g CO2 (as C/dry kg) from MXPns.
Of the alkylated benzenes, toluene and/?-isopropyl toluene (cumene or l-methyl-4-1-
methylethyl benzene) were found at some of the highest relative concentrations for almost all
runs shown in Table 6-1, while ethylbenzene was present in relatively high concentrations from
the MXPns and MXP runs only. Toluene and ethylbenzene, in particular, have also been found in
the highest ambient air concentrations at MSW composting facilities (Eitzer, 1995) compared to
other alkylated benzenes. Styrene was found to be present in both yard waste runs and the
MXP/YW run and the seed, but not in other runs. Styrene might be associated therefore with
yard waste components. 1,4-Dichlorobenzene was the only chlorinated compound detected in
the emissions of all runs, except unseeded food wastes. Naphthalene was detected at some of the
highest relative concentrations in MXP, MXPns, and MXP/FW. The origin of naphthalene is
suspected to be the carbon black contained in newsprint (Reinhart, 1993). Benzene could not be
identified with the GC/MS techniques used, due to coelution with the extraction solvent.
All of the alkylated and chlorinated benzenes shown in Table 6-1 were also identified by
Eitzer (1995) and by Wilkins (1994), primarily in the headspace of mixed wastes, as well as by
Tolvanen et al. (1998). It is interesting to note that several compounds that Eitzer (1995) did not
identify, or identified in negligible amounts, were also not identified in this work. These
compounds were several chlorinated aliphatic and chlorinated aromatics, such as methylene
chloride, 1,1-dichloroethane, chloroform, carbon tetrachloride, trichloroethene, chlorobenzene,
chlorotoluene isomers, the 1,2- and 1,3-dichlorobenzene isomers, and all trichlorobenzene
isomers. However, trichlorofluoromethane, perchloroethane, and 1,1,1-trichloroethane were
-------
Section 6.0 Emissions of Volatile Organic Compounds During Composting
quantified in relatively high concentrations by Eitzer (1995), but not in this study. This indicates
that the source of these chlorinated compounds may be hazardous and industrial wastes and not
the MSW components used here.
Methoxy-benzenes were detected from the MXP/YW, MXP, and MXPns runs. These
compounds might be associated with the removal of lignin during paper manufacturing, since the
structure of these VOCs is similar to moieties found in the lignin polymer.
Based on Table 6-1, the dominant VOCs from the yard waste and food waste runs were
terpenes, alcohols, acids, esters of acids, and ketones. Most of these compounds are generally
considered of biogenic origin (Wilkins, 1994). From the terpenes particularly, • »pinene and
limonene were identified in all runs. These compounds have been found widely in several
degradation environments, such as MSW composting facilities (Eitzer, 1995), sludge composting
facilities (Van Durme et al., 1992), and landfills (Young and Parker, 1983). Wilkins (1994) also
identified several terpenes, including • «pinene, • »pinene, and limonene. Terpenes have usually
been associated with decomposition of wood chips found in composting (Wilber and Murray,
1990; Miller, 1993), but as results show here, woody material is apparently not the only source of
such compounds.
Dimethyl disulfides and dimethyl trisulfides were found primarily in food waste
emissions but also in yard wastes. According to Miller (1993), dimethyl disulfide can be a result
of both biological (decomposition of proteins) and non-biological reactions. Butanoic acid and
esters of butanoic acid were detected in food wastes and are probably partially responsible for
the "strong" odor of decomposed food wastes (Miller, 1993). 2-Ethyl-hexanol was one of the
compounds found in the largest amounts from the MXPns, MXP, and MXP/FW runs.
Interestingly, the dominant compound in the MXP/YW run (as Table 6-1 shows) was the 2-ethyl
hexanoic acid, indicating that partial oxidation of the 2-ethyl hexanol might have occurred. This
could be a result of the more active decomposition in the MXP/YW run, compared to the other
three runs. This is further suggested by the fact that the MXP/YW charcoal tube used during
GC/MS was collected during an advanced composting stage compared to the other runs. The
oxidation of alcohols to their corresponding carboxylic acids is a probable aerobic pathway.
Ketones were also dominant in almost all runs, with camphor found at some of the largest
concentrations. Ketones have been associated with the decomposition of wood chips in
composting facilities (Miller, 1993). Ketones are generally intermediate oxidation products of
alcohols to acids and would therefore be expected to be present in aerobic environments. The
presence of pyrazine in unseeded food wastes and yard wastes can be expected due to the
abundance of nitrogen in both substrates, at least compared to the smaller nitrogen content of
mixed paper, from which this compound was not detected. The lack of volatile acids in the
MXPns run can be explained by the lack of significant decomposition for that run. The heptanoic
acid found in the seeded mixed paper can be, for example, an oxidation product of the heptanal
found in the MXPns run, since much more decomposition occurred in the first run compared to
the latter.
Siloxane compounds (cyclopentasiloxane, cyclotetrasiloxane, and cyclotrisiloxane) were
found at large relative concentrations in all runs except unseeded yard wastes and unseeded food
wastes. The origin of these compounds is not known. Several of the alcohols, ketones, alkanes,
-------
Section 6.0 Emissions of Volatile Organic Compounds During Composting
alkylated benzenes, and sulfuric compounds found in this study were also detected by Wilkins
(1994) and Tolvanen et al. (1998).
Seed might be partially responsible for some of the emissions of seeded components.
The seed would be expected to emit various VOCs because it was collected from an MSW
composting facility and would therefore be expected to contain some hazardous and industrial
wastes. Several xenobiotic compounds were identified in the seed, with toluene, limonene, and
naphthalene being at the highest levels. The number of compounds identified in the seed were
fewer than the number of compounds identified in the MXPns run. As will be discussed later,
seed is not totally responsible for the amounts of selected VOCs produced from seeded MSW
mixtures.
The emission of various VOCs from unseeded mixed paper (MXPns), for which
negligible CO2 production was recorded, indicates that these compounds were somehow
embedded on the individual mixed paper components (cardboard, office paper, newsprint) and
released upon simple wetting, heating, and probably air stripping. The ink in office paper and
newsprint could be partially responsible for these emissions. Most of the compounds found in
the MXPns run emissions appear to be in a less oxidized stage (e.g., several hydrocarbons and
alkylated benzenes) compared to compounds found in mixed paper mixtures in which
decomposition occurred (e.g., alcohols, ketones, acids). This indicates that VOCs were partially
biologically oxidized in the latter runs as a result of decomposition of the basic substrate.
To further investigate the above hypotheses, leaching tests were performed separately for
newsprint, office paper, and cardboard using the same dry mass of each component, with 60 °C
water being the extracting agent. Water, however, may not completely dissolve all compounds,
especially the nonpolar compounds such as alkanes and alkylated benzenes. Each component
was placed in a jar filled with 250 mL of water to minimize headspace and let stand for
approximately 1 hour at 60 °C. Using a sample size of 20 mL, a purge and trap instrument
(Tracer LSC-2, sample concentrator), and a GC/FID, several chromatographic peaks were
observed in water extracts from newsprint and no peaks were recorded from extracts of
cardboard and office paper. Although no further identification of these compounds took place,
it is suggested that newsprint was at least partially responsible for the production of VOCs from
mixed paper. A similar extraction using carbon disulfide as a non-polar solvent on newsprint
revealed no chromatographic peaks.
The presence of toluene and/>/m-isopropyltoluene from unseeded food wastes is
interesting. In distillation experiments designed to identify compounds responsible for the flavor
of certain food products, some xenobiotic compounds, such as aromatic hydrocarbons and
halogenated aromatics, were identified (Coleman et al., 1981; Heydanek and McGorrin, 1981).
Some of the aforementioned VOCs were identified in a control run (Table 6-1).
However, the total number of VOCs emitted from the control run, as well as the amounts, was
smaller than the number of VOCs identified in all other runs. The VOCs in the control run
appear to be a result of imperfect cleaning of the digester and the tubing prior to reuse.
6-8
-------
Section 6.0 Emissions of Volatile Organic Compounds During Composting
6.1.2 Mass Loadings of Selected VOCs During the Composting Process
Twenty-five VOCs were selected for quantification, including 13 VOCs identified in
Table 6-1 that are commonly found at the largest ambient air concentrations at MSW composting
facilities (Eitzer, 1995). Because/?- and w-xylene coeluted in the GC, they will be treated
together; so, in effect, 12 VOCs were found at quantifiable concentrations. Some of these
compounds are categorized as priority pollutants under the Clean Water Act, while others are
hazardous air compounds under the 1991 Clean Air Amendments (see Table 6-2). The goal of
the quantification was to provide mass loading information for these VOCs for each substrate
separately (expressed on a per dry kg basis of substrate), and to study the rate of production of
these VOCs during the composting process. The production rate was followed by using three or
more charcoal traps in sequence for each run and by measuring the mass of each VOC in each
trap. It is noted that except for the targeted VOCs, no quantification of the other VOCs listed in
Table 6-1 was done, primarily due to time and budget limitations and the lack of readily
available standards. Quantification was done using a standard mixture of 25 VOCs, supplied in
methanol by Supelco. Appropriate dilutions in CS2 were made to achieve the concentrations
used during calibration. Multiple injections of the samples gave a coefficient of variation of
between 1% and 3% for all VOCs.
Table 6-2 presents the VOC quantification results. Twelve of the 25 VOCs present in the
standard were not identified by GC/MS (Table 6-1) and were not detected at reportable
concentrations by GC/FID. They are not shown in Table 6-2. The 12 nonreported VOCs were:
benzene, chlorobenzene, o-xylene, bromobenzene, 2-chlorotoluene, 4-chlorotoluene, 1,3-
dichlorobenzene, 1,2-dichlorobenzene, tert-butylbenzene, sec-butylbenzene, 1,2,4-
trichlorobenzene, and 1,2,3-trichlorobenzene. The following results and discussion are based
only on the 12 identified VOCs.
Table 6-2 presents the yields of the 12 targeted VOCs emitted from 12 runs (including
the seed and the control run) over the course of the composting process expressed in • g per dry
kg of component or mixture initially placed in the digester. The units for the control run are in
micrograms. Yields of the 12 VOCs reported in Table 6-2 have been reduced by the amounts of
the corresponding VOC produced from the control run and from the seed present in each run by
simple subtraction.
Table 6-2 indicates that seed produced the highest total yield of the 12 VOCs on a per dry
kg basis. Toluene, />-xylene, styrene, 1,4-dichlorobenzene, 1,3,5-trimethylbenzene, and
naphthalene were the major compounds produced from the seed. As noted earlier, seed would be
expected to emit VOCs because it is partially composted MSW. It is worth noting that Kim et al.
(1995), as part of another study, performed leaching tests with CS2 directly on screened compost
that had been collected from the same location as the seed used here, and they did detect both
toluene and ethylbenzene in the extract.
6-9
-------
Table 6-2. Yields of 12 VOCs from 12 Runs (in • g/dry kg
Compounds FWns FWc'd MXPns MXPC YWns
Toluene"9 (-)h 1,164 151.3 1,150.2 8
Ethylbenzene''9 (-)h 254.3 236.8
p/m-xylene9 23 56.6 282.3 16
Styrene9 62
Isopropylbenzene
Propylbenzene 23.0
1,3,5-Trimethylbenzene 31.4
1,2,4-Trimethylbenzene' 608.0
1,4-Dichlorobenzenef 42 79.1 175.2 7
p-lsopropyltoluene 327.8 1549 4,179.7 3,272.3 736
n-Butylbenzene
Naphthalene' 1,127.8 943.5
Sum of 12 VOCs 327.8 2,778 6,511.2 6,060.3 831
a VOC yields expressed per dry kg of component (only) after excluding the amount
correcting for the two VOCs emitted from the control run.
b No entry indicates that this VOC was not identified by GC/MS.
c Seeded runs.
YWC MXP/YWC
.5 903.0 955.1
(-)h 219.8
.6 326.1 218.9
.7 517.8 195.9
(-)h 29.6
(-)h
.7 315.0 191.6
.4 59.9
.9 2,121.8 1,810.9
of VOCs emitted from the
Substrate)3 b
MXP/FWC FW/YWC MSW1/3 Seed
221.0 (-)h 75.5 1,906.4
72.4 34.0 70.5
162.8 137.9 55.0 504.4
196.6 140.2 563.2
73.4
50.9
55.5 70.8 151.3
149.6
116.3 45.5 35.0 519.8
169.6 155.9 401.6 1,803.9
90.1
492.8 321.6 2,680.5
1,654.4 569.9 1,099.7 8,199.9
Control6
7.3
16.3
23.6
seed present in seeded mixtures and after
d Seed contribution was not subtracted for this run due to the relatively high component/interaction effect shown in Equation (6-3).
e Values for the control run are expressed in • g and not in • g/kg.
f Priority pollutant (designated under the Clean Water Act).
9 Hazardous air pollutant (designated under the 1 991 Clean Air Act Amendments).
h (-) indicates that value became negative after subtracting the corresponding VOC
seed.
' 2-Pentyl furan was identified at the same retention time of 1 ,2,4-trimethylbenzene
yield from the control run
except for the MXPns and
and correcting for the additive contribution from the
MXP/FW runs.
S'
S
fei
1
S'
5,
g
a.
TO"
0
"I
g
!?
c?
1
0
S
S
b
2
S'
n
0
0
S'
-------
Section 6.0 Emissions of Volatile Organic Compounds During Composting
Relatively high emissions of the 12 VOCs were produced from all seeded runs (except
the MXPns), indicating that the seed is partially responsible for these emissions. To better
investigate the potential seed and component interactions, a statistical analysis was performed as
discussed in the next section.
6.1.3 Interaction of Seed and Component in Seeded Runs
A 22 factorial design approach, as described in Chapter 4, was implemented to investigate
the interactions between seed and each MSW component using the VOC yield as a response.
This was done because seed was indicated to be partially responsible for the VOC emissions of
seeded individual components, as discussed earlier. On the basis of 1 dry kg of unseeded MSW
component (mixed paper or yard waste or food waste), Equations 6-1, 6-2, and 6-3 present the
results of the 22 factorial analyses.
YVOCP = 3,568 + 3127LP + 313Ls-128LpLs (6-1)
YVOCY = 1,168 + 581 LY + 778 Ls + 191 LYLS (6-2)
YVOCF = 1,353 + 492 LF+1,189 Ls +328 LFLS (6-3)
where
Y Y
ivocp, iyocY,
YVOCF = sum of the yield of 12 VOCs (in |j,g) emitted from a seeded mixture
of mixed paper, yard waste, and food wastes, respectively
LP = absence (-1) or presence (+1) of mixed paper in a mixture
LY = absence (-1) or presence (+1) of yard wastes in the mixture
LF = absence (-1) or presence (+1) of food wastes in the mixture
Ls = absence (-1) or presence (+1) of seed in the mixture.
Because no replication was used, no standard errors can be provided for the model
estimates; whether an estimate is statistically significant or not is based only on the absolute
value of that coefficient in the corresponding equation. Equation 6-1 indicates that LP is more
important than Ls, and the interaction of LP and Ls is the smaller. Therefore seed contributes less
than 10% of the sum of the yields of the 12 VOCs emitted from mixed paper alone in the seeded
mixed paper run. The high VOC yield of unseeded mixed paper resulted in the interaction of
mixed paper with the seed to appear to be relatively small (and negative), as shown from the LpLs
coefficient in Equation 6-1. Based on Equation 6-2, yard waste alone and seed contribute
approximately similar amounts of VOCs in the seeded yard waste run. Based on Equation 6-3,
Ls is more important than LF indicating that the presence of VOCs in the seeded food waste runs
is primarily due to seed. An exception is/?-isopropyltoluene, which was identified in unseeded
food waste emissions. The interaction factor in Equation 6-3 indicates that additional VOCs are
emitted in seeded food wastes compared to the sum of the amounts emitted by the seed or food
wastes individually. This is likely because the high extent of degradation of food wastes
compared to the other components probably induces the degradation of the seed itself, resulting
in additional release of VOCs attached to the seed. It is also likely that the high moisture content
of food wastes aids in the solubilization and easy release of sorbed VOCs from the seed. The
same explanation can be used for the yard waste and seed interaction, though this interaction is
smaller compared to food wastes.
-------
Section 6.0 Emissions of Volatile Organic Compounds During Composting
Equations 6-1 and 6-2 indicate that, because of the relatively small seed and component
interactions for mixed paper and yard wastes, respectively, an additive contribution of seed to
each of these components in the corresponding seeded run can be assumed. This cannot be done
for seeded food wastes, however, because it might overestimate the amount of VOCs produced
from that component alone. This is because it is not clear if the additional amount produced
from the interaction is due to the seed or to the food waste alone, although the former is likely.
Table 6-2 includes the yields of the 12 VOCs from all runs excluding seed; that is, the
amount of seed present in a seeded run and the VOC yield per dry kg of seed are calculated from
the run with seed only. This was not done for food waste, as discussed. In addition, Table 6-3
presents the additive contribution of seed to total emissions. Table 6-3 shows that seed can
contribute VOCs over a range of 10% to 20% for most of the compounds. Values higher than
100% indicate that seed might be solely responsible for the emissions of that VOC from the
component/mixture. The fact that some compounds identified in the seed were not identified in
some seeded mixtures (e.g., styrene in seeded food wastes) indicates that the compound might
have been decomposed.
According to Table 6-2, the MXPns run had the highest amounts of the 12 VOCs among
all other substrates except the seed, with/>-isopropyl toluene and ethylbenzene accounting for
64% and 3.9% of the sum of the 12 VOCs, respectively. MXPns produced 1,2,4-trimethyl
benzene, which was not emitted from other mixed paper mixtures, indicating that it might have
been decomposed in the other mixed paper runs. It is noted that 2-pentyl furan would be
reported at the same retention time as 1,2,4-trimethyl benzene so coelution is also a possible
explanation. MXPns had the highest ethylbenzene and naphthalene yields followed by MXP,
MXP/YW, MXP/FW, and MSW1/3 in that order. Because the amount of decomposition (loss of
dry weight or CO2 produced) followed the same ranking order for these four runs, it is likely that
these two compounds are decomposed as a result of basic substrate decomposition. The
probable presence of ligninolytic fungi in yard wastes, as discussed previously, could have
induced the "complete" degradation of naphthalene in the MXP/YW run. BTEX compounds and
polycyclic aromatic hydrocarbons (PAHs) are known to be efficiently degraded by a common
ligninolytic fungus (Paszczynski and Crawford, 1995).
Although seed contribution was subtracted assuming an additive contribution, different
seed/component interactions might exist separately for each of the 12 VOCs than the interactions
predicted by Equations 6-1, 6-2, and 6-3 based on the sum of the 12 VOCs. As Table 6-2 shows,
toluene was emitted at some of the largest amounts from runs that were seeded (footnoted in
Table 6-2). Because toluene was at high amounts from the seed alone and in relatively low
amounts from the unseeded runs (e.g. FWns, YWns, MXPns), it is expected that seed was partially
responsible for these toluene emissions.
Butylbenzene and isopropylbenzene were found in the MXP/FW run, but not in the
MXP/YW run. Assuming that these compounds are due to mixed paper only (as shown by the
MXPns run), it is likely that the microbial culture present in the MXP/FW mixture did not
6-12
-------
Section 6.0 Emissions of Volatile Organic Compounds During Composting
Table 6-3. Additive Contribution of Seed to the Emissions of Each of the
12 VOCs Emitted from a Seeded Mixture3 b
MXP YW MXP/YW MXP/FW FW/YW
Toluene
Ethylbenzene
p-Xylene
Styrene
Isopropylbenzene
Propylbenzene
1 ,3,5-Trimethylbenzene
1 ,2,4-Trimethylbenzene
1 ,4-Dichlorobenzene
p-lsopropyltoluene
n-Butylbenzene
Naphthalene
11.3
2.3
12.1
18.1
4.2
17.5
11.6
9.1
6.8
9.6
38.8
19.4
3.8
21.6
25.4
37.0
24.4
466.1
43.8
8.4
22.4
0.0
0.0
20.3
29.2
48.7
0.0
33.3
43.6
8.3
12.6
10.6
23.2
23.3
a FWwas excluded from this analysis due to the high seed/food waste interaction effect, as
shown in Equation 6-3.
b Assuming that X is the mass of a VOC emitted from a seeded mixture and Y is the amount of
VOC expected to have been produced from the specific mass of the seed used during
seeding (as can be calculated from the "seed" run), contribution is defined as 100 x (X-Y)/X.
degrade these compounds, while decomposition occurred in the MXP/YW run. Another
potential explanation is that the presence of isopropylbenzene, found only in the MXP/FW run,
was a result of decomposition of other alkylated benzenes.
/>-Isopropyltoluene was present in all runs and at some of the highest amounts (as shown
in Table 6-2). There was difficulty in the separation of that compound for the MXP and MXPns
runs, due to the presence of 2-ethyl-l-hexanol, which was in very high concentrations and eluted
close to/>-isopropyltoluene during chromatographic analysis.
Styrene was found only in runs that included yard wastes, as discussed earlier, but was
also detected in large amounts from the seed.
Isopropylbenzene, w-propylbenzene, and butylbenzene were generally present in
relatively small amounts in the runs shown in Table 6-2, which coincides with the reported low
concentrations for these VOCs by Eitzer (1995) in MSW composting facilities.
6.1.4 Empirical Models for Estimating VOC Yields from MSW Mixtures
To better investigate the production of these 12 quantified VOCs from an MSW mixture
comprising three organic components only (food wastes, mixed paper wastes, and yard wastes),
the results in Table 6-3 were modeled. The model is empirical, based on the principles of
mixture experimental design, as discussed in Chapter 4. The resulting equation expresses the
-------
Section 6.0 Emissions of Volatile Organic Compounds During Composting
sum of the yields of the 12 VOCs per dry kg of MSW, as a function of the fractions of each of
the three organic components in any MSW mixture. The VOC yields from the unseeded
individual components and for the seeded mixtures, as shown in Table 6-2, were used. The
equation for estimating the production of the 12 VOCs from MSW is Equation 6-4 and was fitted
using the MINITAB v!2.2 statistical package (Minitab Inc., PA, USA).
Yvoc = 4,162 (±1,701) FP + 831 (±1,890) FY + 458 (2,340) FF
- 7,558 (±17,662) FPFY - 6,006 (±28,770) FPFF c (6-4)
where
Yvoc = sum of 12 VOCs volatilized from an MSW mixture, expressed in »g
VOCs/drykgofMSW
FP, FY and FF = dry fractions of mixed paper, yard waste, and food waste,
respectively, in the mixture, with each of the FP, FY, and FF values
ranging from 0 to 1 and with FP±FY±FF always equal to 1 .
The adjusted R2 for Equation 6-4 is 0.33. The relatively low R2 and the high standard
errors of the coefficient indicate that Equation 6-4 may not be a good predictor for VOC
emissions from MSW mixtures because various VOCs that may each behave differently were
summed together. The seed correction may also contribute to the low coefficient of
determination (R2) of Equation 6-4. Equation 6-4 can still illustrate, however, that mixed paper
is the greatest source of these 12 VOCs in an MSW mixture, as compared to the other two
organic components. Combinations of yard wastes or food wastes with mixed paper result in a
significant decrease of the sum of these 12 VOCs. As discussed earlier, this is likely a result of
the decomposition of these VOCs due to degradation of the basic substrate. Mixed paper
reached its full extent of degradation faster when combined with either food wastes or yard
wastes, compared to when composted individually, as discussed previously. This probably
affects the decomposition of VOCs as well.
Separate equations were also individually fitted by least squares to the results for each
VOC. Parameters for which their confidence intervals contained zero were not included in the
models; therefore, the following equations are the best reduced models. Errors were distributed
normally and had a zero mean. Values in parentheses are the standard errors for the
corresponding model coefficients.
Toluene fit (adj. R^O.
YTOLUENE = 306.2 (±79.2) FP + 4,663.5 (±812.5) FpFY - 5,322 (±1,305) FpFF
Ethvlbenzene fit (adj. R^
YETHYLBENZENE = 214.7 (±34.9) FP - 1,090 (±535.4) FPFF
p/m xylene fit (adj. R2=0.564)
YXYLENE= 151.6 (±47.8) FP
-------
Section 6.0 Emissions of Volatile Organic Compounds During Composting
Stvrene fit (adj. R^O.948)
YSTYRENE = 67.2 (±27.1) FY+ 1,218.5 (±187.8) FPFY - 1,012 (±349.2) FPFF± 786.3 (±197.5) FYFF
n-propylbenzene fit (adj. R2=0.758)
YNPROPYLBENZENE = 35.9 (±7.3) FP- 165.8 (±65.7) FPFY
1.3.5-trimethvlbenzene fit (adj. R2=0.97n
Y135 = 14.8 (±8.5) FP± 160.2 (±86.7) FPFY + 1,405.6 (±139.3) FpFF
1.2.4-trimethylbenzene fit (adj. R^O.
Y124 = 375.6 (±104.7) FP - 1,734 (±937.8) FpFY
1.4-dichlorobenzene fit (adj. R^
Y14 = 109.2 (±27.7) FP + 408.2 (±248.4) FPFY
p-isopropyltoluene fit (adj. R2=0.302)
YP-ISOPROPYLTOLUENE ~~ 1,643 (±819.2) Fp
n-butvlbenzene fit (adj. R2=0.227^
* NBUTYLBENZENE ~ 3 1 . / (± 1 0. 1) rp
Napthalene fit (adj. R^O.
= 765.7 (±200.4) FP - 2,401 (±1,794) FPFY
where
Yvoc = mass of VOC volatilized per unit dry mass of substrate (in mg/dry
kg)
FP, FY and FF = dry fractions of mixed paper, yard waste, and food waste,
respectively, in the mixture, with each of the FP, FY, and FF values
ranging from 0 to 1 and with FP±FY±FF always equal to 1 .
The above models show that mixed paper is the major VOC source except for styrene, for
which yard wastes appears to be its primary source. Interactions between mixed paper and either
yard wastes or food wastes are important and, in most cases, reduce VOC emissions due to
decomposition of the basic substrate. This indicates that degradation of VOCs takes place after
combining mixed paper with a more degradable component.
In the case of toluene, the interaction of mixed paper and yard wastes increases toluene
yields. This could be due to additional toluene produced by the seed present in the MXP/YW
run. A decrease of toluene is, however, observed when mixed paper is combined with food
wastes. Degradation of ethylbenzene, w-propylbenzene, and 1,2,4-trimethylbenzene is observed
when mixed paper is combined with either food wastes or yard wastes. In the case of styrene,
only the interaction of mixed paper and food wastes results in a decrease of styrene emissions;
increased emissions are observed when yard wastes are present in the mixture.
6-15
-------
Section 6.0 Emissions of Volatile Organic Compounds During Composting
Additional release of 1,3,5-trimethylbenzene and 1,4-dichlorobenzene is observed upon
combining mixed paper with food wastes or yard wastes, indicating that decomposition might be
the reason for the production of that compound.
Reduction of naphthalene is observed upon combining mixed paper with yard wastes,
probably because of ligninolytic microorganisms present in yard wastes.
6.1.5 VOC Volatilization Rates
The volatilization rates for the 12 VOCs were developed for some of the runs shown in
Table 6-2. These profiles are helpful in further investigating the origin and behavior of the 12
targeted VOCs during composting. Volatilization rates for the MXP, MXPns, and the seed runs
are shown in Figures 6-1, 6-2, and 6-3. The rates for these runs are more or less typical of all
runs.
For the MXP run (Figure 6-1), the incubator temperature was gradually increased from
30 to 55 °C during the first 25 days of the run, after which thermophilic temperatures were
maintained. This was done to simulate the typical temperature increase observed in the center of
MSW compost windrows during active composting (Diaz et al., 1993). This gradual temperature
increase was used for the MXP run only. Thermophilic temperatures were kept continuously for
all other runs to accelerate decomposition. As seen in Figure 6-1, a relatively low volatilization
rate was observed during the increasing temperature period for most VOCs. VOC emission rates
started to rapidly increase at temperatures higher than approximately 45°C and stabilized to a
maximum value. This profile indicates that these compounds are somehow attached to the solid
matrix and released upon heating. The VOC yield might be affected by temperature; that is,
higher VOC yields might have been measured if temperatures higher than 55°C had been used.
In any case, this profile indicates that a fixed amount of VOCs is originally present in the solid
matrix and released.
As shown in Figure 6-1, toluene had a relatively high volatilization rate from the
beginning of the composting process due to its being the most volatile of the compounds with a
boiling point (b.p.) of 110.6 °C. Ethylbenzene, the second most volatile after toluene, had the
second highest volatilization rates during the initial 25-day period. Naphthalene, being less
volatile than both compounds with a b.p. of 217.9 °C, had close to zero volatilization rates during
that initial period. Between days 30 and 40, naphthalene rapidly increased to rates similar to
those of toluene.
Most VOCs rapidly volatilized during the first 20 days for unseeded mixed paper
(Figure 6-2) because this run was performed directly at thermophilic temperatures throughout.
VOC emissions reached a constant value after approximately 20 days, except forp-
isopropyltoluene, naphthalene, and 1,2,4-trimethylbenzene. These three compounds stabilized
after approximately 60 days.
6-16
-------
Section 6.0
Emissions of Volatile Organic Compounds During Composting
30Cto55C
55 C
3,000
T
2,500 -
D)
.*
2,000 -
-O 1,500
o—o
_L
-O— Toluene
-A- Ethylbenzene
-V~ p-xylene
-•- 1,4-dichlorobenzene
-•- p-isopropyltoluene
-T- naphthalene
-O
40 60 80 100 120 140 160 180 200
Days
Figure 6-1. Cumulative production of six VOCs during composting of seeded mixed
paper; yields have not been corrected for seed (the period between two data
points corresponds to the time that one charcoal trap was in use
continuously).
6-17
-------
Section 6.0
Emissions of Volatile Organic Compounds During Composting
55 c
3,000
2,500
D)
2,000
^ 1,000
E
0
(/)
O
D)
500
XXX
1
—O-Toluene
—A- Ethyl benzene
—V— p-xylene
— I - propylbenzene
—I— 1,3,5-trimethylbenzene
—X- 1,2,4-trimethylbenzene
—•— 1,4-dichlorobenzene
—•- p-isopropyltoluene
—T~ naphthalene
i . i . i . i
0 20 40 60 80 100 120 140 160 180 200
Days
Figure 6-2. Cumulative production of nine VOCs during composting of unseeded mixed
paper (the period between two data points corresponds to the time that one
charcoal trap was in use continuously).
6-18
-------
Section 6.0
Emissions of Volatile Organic Compounds During Composting
55 c
0)
0)
o
D)
3,000
2,500
2,000
1,500
0 1,000
O
O
500
0
—O-Toluene
—A- Ethyl benzene
—v— p/m xylene
—o— styrene
—\— 1,3,5-trimethylbenzene
—•— 1,4-dichlorobenzene
—•- p-isopropyltoluene
—Y- naphthalene
20 40 60 80 100 120 140 160 180 200
Days
Figure 6-3. Cumulative production of eight VOCs during composting of seed (the period
between two data points corresponds to the time that one charcoal trap was
in use continuously).
In the case of seed (Figure 6-3), only/?-isopropyltoluene appears to have a different
volatilization profile compared to the other compounds. />-Isopropyltoluene appears to have a
lower initial volatilization rate compared to the rate between days 15 and 40. Decomposition of
this compound is rather unlikely during the initial time because it is water insoluble. It is likely
that/Msopropyltoluene is more strongly attached to the solid phase than the other compounds,
which explains its initial volatilization lag. The octanol-water partition coefficient log Kow for/>-
isopropyltoluene is 3.66, while the log Kow for naphthalene is 3.29, which indicates that/?-
isopropyltoluene will tend to sorb more onto the solid phase compared to naphthalene, despite
the higher boiling point of the latter. As decomposition of the seed proceeds, however, the
bound/Msopropyltoluene is released and volatilized.
6-19
-------
Section 6.0 Emissions of Volatile Organic Compounds During Composting
6.2 Origin and Fate of VOCs
The results presented above do indicate that the 12 targeted VOCs are somehow bound to
or embedded in the solids of all MSW components, but primarily mixed paper, and released
upon heating and probably other mechanisms that affect their fate (i.e., sorption, water
solubility). Decomposition of the basic substrate can reduce the yields of VOCs as indirectly
indicated in Equation 6-4. However, definite conclusions cannot be made since the initial
amount of VOCs present in each substrate is not known.
Some of the VOCs emitted from mixed paper—especially those frequently referred to as
xenobiotic—might be produced due to the ink present in office paper and newsprint, as
discussed earlier, or due to paper processing (e.g., wood bleaching) during which dissolution and
degradation of lignin occurs that could give rise to some aromatic chemicals. However, this
hypothesis needs more careful investigation, since lignin consists of oxidized aromatic rings,
such as phenolic and methoxy groups (Kirk, 1984). The production of alkanes and aromatics,
which are VOCs in a reduced oxidation stage, from lignin moieties would require anaerobic
environments and high pressures that are not expected to occur during composting. Only the
presence of methoxy benzene can be explained by the previous mechanism. Another potential
explanation of the presence of VOCs in mixed paper could be atmospheric deposition.
The apparent decomposition of xenobiotic VOCs within the MSW environment is most
likely a result of cometabolic processes; that is, carbon and energy sources are abundant for the
composting cultures owing to the presence of the basic substrate. These cometabolic processes
occur due to the generation of the required catabolic enzymes, which are used during basic
substrate degradation as well as VOC decomposition. For example, oxygenases, which are
commonly found in lignin degradation environments (Kirk, 1984), have also been shown to
catalyze the biodegradation of several alkylated benzenes and PAHs (Gibson and Subramanian,
1984).
Decomposition of the basic substrate is expected to produce various VOCs, usually
referred to as biogenic, such as terpenes, acids, acid esters, alcohols, and ketones. However,
based on the results of this study, the identified biogenic VOCs also appear to be produced
during the initial stages of composting as well as from substrates with negligible degradation
(MXPns). Therefore, it is likely that production of some biogenic VOCs (e.g., limonene, which
was found in some of the largest amounts in unseeded mixed paper) is also a result of simple
volatilization. Biogenic VOC—though byproducts—could be also subject to further degradation
depending on contact with the compost matrix.
Generally all types of VOCs measured in this study decreased as composting progressed
and no VOCs appeared to be produced after a certain period, following the trend shown in
Figures 6.1, 6.2, and 6.3. Volatilization rates appeared to follow first-order kinetics based on the
12 quantified VOCs. This decreasing trend in VOC emissions during composting was also
observed by Eitzer (1995) for several VOCs, but not ketones, for which an opposite trend
was observed. Eitzer (1995) did observe that all types of VOCs measured were emitted during
shredding, where no extensive biodegradation usually takes place.
6-20
-------
Section 6.0 Emissions of Volatile Organic Compounds During Composting
Equation 6-4 was developed based solely on the three organic MSW components (and the
seed) used in this study. Other MSW components, such as inorganics or plastics, were not used
and therefore their contribution to VOC emissions is not known. In addition, no hazardous or
industrial wastes were used in the study, which would be expected to contribute to VOC
emissions. Equation 6-4 predictions (given on a per unit dry kg basis of the sum of food wastes,
yard wastes, and mixed paper) have to be multiplied by the percentage of the combined three
components in typical MSW in order to express results on a per unit dry mass of "real" MSW.
The results given here are important because they indicate that hazardous and industrial
wastes might not be the sole source of xenobiotic VOCs in MSW. Comparisons between studies
of VOCs from solid wastes is important; however, attention should be given to the conditions
under which VOCs are measured, the types of VOCs quantified, and the analytical techniques
(such as VOC traps) used during measurement.
6.3 Spiking of VOCs to MSW Organic Substrates
To further investigate the effect of basic substrate decomposition on the fate of VOCs in
MSW composting, separate runs were performed based on spiking known amounts of specific
VOCs onto MSW components of different degradability.
6.3.1 Ethylbenzene Spike
The objective of this work was to investigate whether the decomposition of the basic
substrate influences the decomposition and the fate of VOCs in a composting environment. This
information would be important for predicting the fates of VOCs in MSW of various
compositions, thus aiding in the development of standards concerning the disposal of household
hazardous wastes in the MSW stream from different localities or countries.
Ethylbenzene was spiked separately to seeded newsprint and seeded yard waste at the
same approximate initial concentration levels. Ethylbenzene was selected because it is the
alkylated benzene found at the highest ambient air concentrations at MSW composting facilities
(Eitzer, 1995) and can be used as a model for alkylated benzenes. These two runs were
performed independent of the 12 runs described earlier in this chapter. Newsprint and yard
wastes were selected due to their different degradation behaviors during composting, which
might affect the fate and decomposition of the spiked VOC. Newsprint was chosen as one of the
substrates instead of mixed paper to provide a relatively uniform substrate, allowing easier
sampling of solids compared to mixed paper. Newsprint was seeded at the same proportions as
the runs presented previously. Ethylbenzene was not spiked to food wastes because of solids
sampling difficulties. Both runs were performed at mesophilic temperatures ranging from 30
to 35 °C. Mesophilic temperatures were selected to reduce extensive direct volatilization of the
compound and, therefore, to allow better investigation of the potential degradation and
metabolite production from this compound.
A total dry mass of 0.47 kg newsprint and 0.51 kg of yard wastes were used in the
experiments. Ethylbenzene was spiked at concentrations of 9,100 ppm and 10,000 ppm (mg
VOC/dry kg) onto the newsprint and yard wastes, respectively, by adding the VOC directly in
liquid form through the exit port at the top of each digester. These values correspond to
6^21
-------
Section 6.0 Emissions of Volatile Organic Compounds During Composting
approximately 3,600 ppm and 3,500 ppm (mg VOC/ wet kg), which are approximately 11 times
the amounts of hazardous chemicals present in MSW, based on EPA estimations (Brown et al.,
1997). The relatively high concentrations used in this run were selected to allow better
identification of potential metabolites produced during the process.
Air was supplied intermittently during the first 5 days of the experiment to minimize
volatilization of ethylbenzene and to induce acclimation of the biomass present in the solid
wastes to the added VOC. Because of this inadequate aeration, however, relatively low CO2
production was observed during the first 5 days from both substrates. Continuous air flow was
applied beyond day 5 and a rapid increase in CO2 production rates was observed as a result of
these increased oxygen levels in the digesters. Air flow rates ranged from 200 to 400 mL/min,
corresponding to approximately 400 to 800 mL/dry kg/min.
As shown in Figure 6-4, high ethylbenzene volatilization rates were observed
immediately after spiking the VOC onto newsprint during the intermittent aeration period.
Volatilization rates from yard wastes were lower than from newsprint, indicating that sorption or
degradation of the ethylbenzene occurred rapidly after spiking. The potential rapid partitioning
of ethylbenzene onto the solid matrix could be a result of the hydrophobic nature of
ethylbenzene as well as the presence of hydrophobic groups (e.g., fats/lipids) in yard wastes.
The fats and lipids in yard wastes, at a content of 2.5% dw, could therefore retain the
hydrophobic VOC to a greater extent than newsprint, which has a fat/lipid content of 0.5% dw.
Such hydrophobic binding has been suggested for humic substances, the hydrophobic surfaces of
which can sorb nonpolar organic compounds (Sparks, 1995). It is also likely that ethylbenzene
was more easily solubilized in the moisture present in yard wastes compared to newsprint if the
initial moisture adjustment for newsprint provides less moisture compared to yard wastes. Note,
however, that initial moisture contents were similar. Such initial dissolution would aid
subsequent degradation of the compound.
The initiation of continuous aeration after the first 5 days of the experiment increased the
mineralization rates for both substrates, as indicated by the rapid increase of the CO2 production
rate after 5 days (Figure 6-4). However, continuous aeration rates also induced temporary
stripping of ethylbenzene from both substrates, as shown by the rapid increase of the
volatilization rate of ethylbenzene after the fifth day. Volatilization rates did decrease thereafter,
indicating that combined degradation and sorption mechanisms determined the fate of the
remaining VOCs.
When ethylbenzene volatilization rates became approximately zero, VOC emissions
sampling was stopped. Approximately 10 g of the wet mass of the solid material from each run
was randomly collected and placed in 125 mL jars. One hundred milliliters of carbon disulfide
was added to each jar and, after letting stand for 30 minutes, the extract was analyzed using
direct injection to GC/FID. No ethylbenzene was measured during these tests. Such leaching
tests, however, are not capable of measuring VOCs that may have been irreversibly sorbed onto
the solid wastes as a result of potential biological humification processes or irreversible abiotic
binding. Binding of chlorinated compounds onto humic acids during dehalogenation processes,
based on biologically mediated radical formation, has been reported by Dec and Bollag (1994).
Radical formation during degradation of alkylated benzenes by ligninolytic fungi (Barr and Aust,
1994) is possible since BTEX compounds have been shown to be degraded by a ligninolytic
6^22
-------
Section 6.0 Emissions of Volatile Organic Compounds During Composting
fungus (Yadav and Reddy, 1993). Ligninolytic fungi are present in MSW composting (de
Bertoldi et al., 1983), although primarily at later stages of composting. Biologically mediated
radical formation of ethylbenzene could therefore result in covalent bonding with humic
material. Further research is needed in this area as it applies to MSW composting environments.
From the initially spiked mass of ethylbenzene, 30.5% and 12% was volatilized from the
newsprint and yard waste runs, respectively. Based on the MXP run discussed earlier, the
ethylbenzene yield expected from newsprint without any spiking (baseline) should be
approximately 200 to 300 • g/dry kg (Table 6-2), which is several orders of magnitude less than
the approximately 3,000 mg/dry kg volatilized from the newsprint run after the spike. Since no
ethylbenzene was detected in the final product, the nonvolatilized ethylbenzene must have been
degraded to either intermediate byproducts or CO2. Irreversible adsorption is also likely.
Therefore, 88% and 69.5% of the initially spiked ethylbenzene was apparently decomposed in
yard wastes and newsprint, respectively. Because the two substrates produced 121 g CO2 (as
C)/dry kg and 93 g CO2 (as C)/dry kg, respectively, it appears that the degradation of
ethylbenzene was influenced by decomposition of the basic substrate. However, a better
comparison can be made by accounting for the yield of CO2 exerted until the time the VOCs
were completely volatilized, which was about 15 days from the initiation of the composting
process. During these initial 15 days, 45.4 g CO2 (as C)/dry kg and 35.8 g CO2 (as C)/dry kg
were produced from yard wastes and newsprint, respectively. This also suggests that more
decomposition of the basic substrate (yard waste) is partially responsible for the corresponding
increase in decomposition of ethylbenzene.
Therefore, yard wastes have either a higher concentration of microbial population,
compared to newsprint, capable of degrading ethylbenzene or the microbial population in yard
wastes is capable of degrading ethylbenzene to a greater extent than the population in newsprint.
A cometabolic process is likely for such decomposition, as discussed earlier.
According to Gibson (1996), ethylbenzene can be oxidized to styrene by the presence of
naphthalene dioxygenase; however, no styrene was detected in this study. This pathway may not
be dominant during aerobic decomposition of ethylbenzene. Less volatile metabolites, such as
2,3-dihydroxyethylbenzene, are usually the preferred decomposition metabolites (Gibson, 1996).
The exact metabolic pathway cannot be concluded from this study. The use of
radiolabeled compounds could have aided in the determination of the fraction mineralized and of
the fraction attached to the solids; however, a radiolabeled compound could not be used,
primarily because of concerns for safety and potential laboratory contamination from the highly
volatile compounds studied.
6.3.2 Spike with a Mixture of Alkylated Benzenes
A separate run (TEX run) was performed to further assess the behavior of VOCs during
MSW composting. The goal of this run was to assess the behavior of different alkylated
benzenes when each is spiked at the same initial concentration levels during different stages of
MSW composting. The selected VOCs were toluene, ethylbenzene, m-xylene, and o-xylene
(referred to as TEX), which are commonly found in MSW composting facilities (Eitzer, 1995).
6-23
-------
Section 6.0
Emissions of Volatile Organic Compounds During Composting
continuous air flow
, maintained after day 5
600
—O— VOC from newsprint
—O— VOC from yard wastes
—•— CO2 from newsprint
—^— CO2 from yard wastes
150
125
100
75
30
Days
40
50
60
T3
er
O
50 O
25
Figure 6-4. Volatilization rates of ethylbenzene and CO2 cumulative production from
newsprint and yard wastes during composting. (Ethylbenzene was
spiked at initial concentrations of 9,100 and 10,000 mg per dry kg of
newsprint and yard wastes, respectively.)
6-24
-------
Section 6.0 Emissions of Volatile Organic Compounds During Composting
Certain key properties of these VOCs that are expected to affect their fate during composting are
given in Table 6-4. These properties are boiling point, vapor pressure, solubility, octanol-water
partition coefficient, and biodegradation half-lives based on aqueous (and soil) aerobic
environments. VOCs were spiked to the mixture in the vapor phase, as discussed earlier.
As shown in Figure 6-4, decomposition of ethylbenzene in both substrates might have
been limited during the first 5 days due to delayed aeration. The rapid increase in ethylbenzene
volatilization after 5 days indicates that there is still a residual amount of VOC attached to the
solid matrix of both substrates. Because this initial release was higher for newsprint than yard
wastes, it is likely that the decomposition rate was higher in yard wastes compared to newsprint
during the first 5 days.
Generally, decomposition of ethylbenzene in both substrates took place within the first
15 days after initiation of the experiments. No acclimation of the microbial population present in
either substrate had taken place, which would probably have reduced decomposition times.
The temperature during the TEX run was kept at thermophilic levels to better simulate
actual conditions, since such temperatures commonly occur during composting (Diaz et al.,
1993). Since the goal was not to compare the effect of different substrates on the fate of VOCs,
the basic substrate was set to be a mixture of mixed paper, yard wastes, and food wastes
in proportions based on a typical U.S. MSW composition. Therefore, the organic fraction of
MSW was simulated using 80% mixed paper, 15.4% yard waste, and 4.2% food wastes (all on a
dry content basis), as used previously.
Since both microbial populations and organic matter change during composting (de
Bertoldi et al., 1983), several spikes were performed to investigate potential effects of the
different composting phases on VOC fate.
6-25
-------
Section 6.0
Emissions of Volatile Organic Compounds During Composting
Table 6-4. Physicochemical Properties of VOCs Used in TEX Run
(La Crega et al., 1994; Howard et al., 1991)
Compound
Toluene
Ethylbenzene
m-xylene
o-xylene
Boiling point
(°C)
110.8
136.2
139.0
144.4
Solubility
(mg/L)a
515
152
175
130
Vapor pressure
(mm Hg)a
22
7
10
10
Log Kow
(dim/less)
2.73
3.15
2.95
3.26
Biodegra-
dation half-
lives b
96 / 528
72 / 240
168/672
168/672
3 At20°C.
b Low and high biodegradation half-lives (hours), respectively, in aerobic aqueous (and soil)
environments (unacclimated), from Howard et al. (1991).
Five spikes of the mixture of these four VOCs were performed on days 0, 7, 23, 28, and
35 from the initiation of the run. Each VOC was spiked at a concentration of approximately
65 mg/dry kg of substrate, so that a total VOC concentration of approximately 265 mg/dry kg or
approximately 150 mg/wet kg was achieved. This value is near the lower level of the VOC
concentration range found in MSW (200 mg/kg to 1,500 mg/kg), as found in a study in King
County, WA (Kissel et al., 1992). Because dry matter reduction occurred during composting, the
masses of VOCs added during the fourth and fifth spike were approximately half of the masses
used during the first three spikes. This was done in an attempt to maintain the spiking
concentration at approximately 265 mg/dry kg of substrate at the time the spike was performed.
The dry matter reduction during the process was calculated indirectly by measurement of the
carbon dioxide yields. A continuous aeration regime was used during the first four spikes, and
intermittent aeration was used during the fifth spike to minimize volatilization and to study the
effect of such a change.
The cumulative volatilization profiles of all four VOCs during all four spikes are shown
in Figure 6-5, and Table 6-5 summarizes the fractions (in %) of each of the initial masses of
spiked VOCs volatilized. Table 6-5 also shows the time required for volatilization of 99% of the
totally volatilizable mass of each VOC during each spike.
Toluene was detected in the breakthrough section of some of the traps during the second
and third spike, indicating that some was lost. Therefore, the actual toluene volatilization
fractions would have been higher than the values shown in Table 6-5. The significant
breakthrough for toluene is probably attributable to the fact that it has the highest vapor pressure
and lowest boiling point among the compounds tested.
6-26
-------
Section 6.0
Emissions of Volatile Organic Compounds During Composting
•-
"o
^ O
-s
O
110%
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
—O~ Toluene
—A— Ethylbenzene
—A— m-xylene
—D— o-xylene
Breakthrough
occured
for toluene .+'
u-+~
-+-"1
Cumulative CO2
production
-5 0
^H-6-1
250
200
150
D)
o
100 cfl
50
10 15 20 25 30 35 40 45 50
Continuous aeration
Days
Delayed aeration
Figure 6-5. TEX cumulative volatilization profile and CO2
cumulative production during composting of MSW.
O
o
s_
D)
6-27
-------
Section 6.0 Emissions of Volatile Organic Compounds During Composting
Table 6-5. Volatilized Fraction (in %) of Initially Spiked VOC Mass During TEX
Run and Length of Time (days) for Volatilization After Each Spike
Spike 1
Spike 2
Spike 3
Spike 4
Toluene
87%
(3.5) a
50% b
26% b
104%
(1.8)
Ethylbenzene
87%
(4.5)
90%
(2.6)
98%
(0.95)
100%
(1.9)
m-Xylene
89%
(5.5)
93%
(2.7)
103%
(1.0)
105%
(2.2)
o-Xylene
91%
(5.9)
100%
(2.75)
105%
(1.2)
93%
(2.4)
Spike 5c 29% 25% 29% 25%
a Values in parentheses indicate days after each spike, that 99% of the VOC mass produced after that
spike was volatilized.
b Toluene breakthrough (loss) occurred and therefore a higher value would be expected.
c Delayed aeration regime was used during this spike.
During the initial spike, lower volatilization percentages were observed for all VOCs compared
to spikes 2, 3, and 4. The difference is attributed to degradation, because, during the first 5 days,
the basic substrate was at its most active degradation stage (as shown by the relatively steep CO2
cumulative curve during that time), which is primarily a result of decomposition of the readily
degradable components (e.g., food wastes). Rapid sorption of the VOCs onto the substrate
during the first spike, primarily onto food wastes and yard wastes that have a higher content of
fats/lipids compared to paper, must have facilitated the retention of VOCs and their subsequent
decomposition. All VOCs appear to be volatilized at similar levels of between 87% and 91%
during the first spike. As Table 6-5 shows, the days required for volatilization of each VOC
follow the same ranking order with the corresponding boiling points of these four VOCs, which
is probably expected. It took from 3.5 to 5.9 days for these VOCs to volatilize during the first
spike. Sorption might be partly responsible for the initial non-volatilized fraction, however,
since all VOCs used are considered relatively decomposable, and decomposition is the likely
dominant mechanism for their removal.
More than 90% of the VOCs, excluding toluene, volatilized during the second spike, and
it took less time for the full volatilization of the four VOCs compared to the first spike. This is
probably a result of dry matter reduction, which resulted in increased void space and a potential
for increased channeling of the VOCs from the side of the digester, therefore reducing their
contact with the solid matrix. It took approximately 2 to 3 days for all VOCs to volatilize during
this spike and the same order of volatilization rates was observed as with spike 1. Assuming that
the difference from 100% is decomposed VOCs, ethylbenzene appears to be the most
degradable. Generally, ethylbenzene is one of the most degradable alkylated benzenes in various
environments, as shown in Table 6-4 and as was reported by Yadav and Reddy (1993) for
ligninolytic environments. w-Xylene and o-xylene had relatively similar apparent
6^28
-------
Section 6.0 Emissions of Volatile Organic Compounds During Composting
decomposition extents during the second spike, which is in accordance with their similar
degradation half-lives. It is noted, however, that o-xylene is generally considered the least
degradable of BTEX compounds in several environments (Yadav and Reddy, 1993), which may
explain its full volatilization during the second spike, despite it being the one with the highest
boiling point and the highest log Kow.
Higher volatilization percentages (approximately 100%) and even smaller volatilization
times were observed during the third spike compared to the second and first spikes for all VOCs.
The additional dry matter reduction was probably responsible for this, as discussed. Because the
same mass of VOCs was spiked during the second and third spike compared to the first and
additional dry matter reduction occurred, an increasingly higher mass of VOCs was spiked per
unit dry mass of substrate in the digester as composting progressed.
In the fourth spike, the mass of VOCs added was reduced and corresponded to a total
concentration of 149 mg (sum of VOCs) per dry kg of initial substrate. It is worth noting that the
three most volatile VOCs (toluene, ethylbenzene, and w-xylene) totally volatilized, with only
o-xylene having a lower volatilization fraction of 93%. Decomposition and sorption are both
likely operable mechanisms reducing the release of this compound. o-Xylene is the least volatile
and has the highest log kow among all VOCs, indicating a stronger retention in the solid phase.
During the fourth spike, composting is already advanced and the corresponding generation of
humic materials might be responsible for sorption of o-xylene. Decomposition is also likely,
however, since acclimation of the biomass must have taken place as a result of the previous three
spikes. The high volatility of the other three VOCs resulted in their total release.
During the fifth spike, intermittent aeration (5 minutes every 1 hour) was maintained to
minimize the excessive volatilization that occurred during the previous spikes. This resulted in a
decrease of more than 60% in the volatilization extents of all VOCs compared to the previous
spikes. All VOCs volatilized within 2.1 days during the fifth spike. The potential for
decomposition and sorption of the VOCs was increased during that spike due to delayed
aeration. Ethylbenzene and o-xylene had the lowest volatilizable fractions at 25% among the
four VOCs. A higher sorption is likely for both, compared to the other two compounds, due to
their having relatively high log Kow partition coefficients. However, a high sorption coefficient
would retard degradation because the compound is probably retained in hydrophobic organic
surfaces of the substrate. If partitioning between the solid and water phases takes place during
composting, the initial sorption of compounds may result in relatively more subsequent
degradation after all. Decomposition is more likely for ethylbenzene because it is generally the
most degradable of these VOCs, and the previous experiment showed that no residual
ethylbenzene remains on the solid. Generally the delayed aeration and the potential acclimation
of the biomass by the fifth spike made decomposition the likely mechanism for the lower
amounts volatilized. Ethylbenzene and o-xylene appear to be the most degradable of the four
VOCs.
No absolute conclusions can be made regarding the decomposition of the four VOCs
during this run since no solids analyses were performed. Heterogeneity of the material and the
multiple spikes might have confounded such analyses. Based on the fifth spike, and assuming no
VOCs were finally retained on the solids, it appears that 71% of toluene and w-xylene and 75%
of ethylbenzene and o-xylene were decomposed when aeration was kept low so that no excessive
6^29
-------
Section 6.0 Emissions of Volatile Organic Compounds During Composting
release of the compounds in the volatile phase occurred. Kim et al. (1995) reported that 6.2%
and 0.2% of ethylbenzene and toluene, respectively, remain on solids during composting of
MSW.
The close to 100% volatilization percentages recorded during the first four spikes of the
TEX experiment (continuous aeration) are comparable to the findings by Brown et al. (1997).
They determined that 100% of selected VOCs were volatilized in times similar to those reported
here. Differences in the spiking techniques should be accounted for between the studies, since a
vapor phase addition of the VOCs was used in the TEX run. Kim et al. (1995) concluded that
0% and 12.4% of ethylbenzene and toluene volatilizes during composting of MSW in a
continuous aeration in-vessel plant, which partially contradicts the findings of this study. The
different conclusions of Kim et al. (1995) might be attributed to the higher retention time of
VOCs in the actual composting vessel compared to smaller retention times used in this study and
by Brown et al. (1997).
6-30
-------
Section 7.0 Conclusions
7.0 Conclusions
This report presents a laboratory method to measure CO2, NH3, and VOC emissions and
to characterize solids decomposition during composting of MSW. Different runs with different
MSW components and mixtures of components were performed, and the reproducibility of the
materials and methods was validated. The MSW components used (mixed paper, yard waste,
and food waste) normally are the largest decomposable fractions of MSW. The close to 100%
carbon mass balance closures and the low biases of the CO2 and NH3 measurements verify the
usefulness of the methods and analytical techniques for these two gases.
The results of this study suggest the following conclusions.
Chapter 3
1. The laboratory method is reproducible in measuring yields of CO2 and NH3 and solids
decomposition during composting.
Chapter 4
2. Seeding of yard wastes or food wastes did not significantly affect their decomposition.
Seed is necessary for the decomposition of mixed paper.
3. The results using the mixture experimental design showed that CO2 yields (in g C/dry kg)
can be estimated by the additive model: Ykg_c02 = 217.4xFP + 237.3xFY + 370.5xFF,
where FP, FY and FF are the dry fractions of mixed paper, yard waste, and food waste,
respectively. The model indicates that all interactions were insignificant.
4. NH3 yields (in g N/dry kg) from mixtures of MSW components can be estimated by the
nonadditive model: Ykg ^ = 1.29xFP +5.15xFY + 37.6xFF-68.9xFpxFF, with
parameters as defined in number 3, above. Inclusion of mixed paper in MSW mixtures
resulted in a decrease of ammonia emissions because nitrogen is limited in mixed paper.
5. The interactions of mixed paper with either yard wastes or food wastes were significant
in terms of rates of CO2 production. Mixing mixed paper with either yard wastes or food
wastes significantly reduced the composting time of the mixture compared to composting
mixed paper alone.
6. The response surfaces shown in Figures 4-5 and 4-6 are based on the equations in
conclusions 3 and 4 above, and can be used for rapid estimation of CO2 and NH3 yields
from MSW of various compositions.
7-1
-------
Section 7.0 Conclusions
Chapter 5
7. Initial lignin and HWSM contents (dry weight basis) were important in determining CO2
yields from different MSW substrates.
8. Lignin was less inhibitory to substrate decomposition in an aerobic environment than in
an anaerobic environment. This is due to the physical association of lignin with cellulose
(sheathing) and because lignin is relatively degradable in aerobic environments but
refractory in anaerobic environments.
9. Cellulose and hemicellulose were responsible for more than 50% of the total dry mass
loss for most of the substrates tested. Lignin/humus losses were responsible for up to
22% of the total dry mass loss of MSW substrates.
10. Hemicellulose was almost totally degraded in all substrates. Cellulose reduction ranged
from 53.9% to 91.1% for all substrates that had approached "complete" degradation.
11. Lignin/humus reductions were 61.6% in food wastes, 43% in yard wastes, and
approximately 26.7% in mixtures of yard and mixed paper wastes. A net increase of
lignin/humus dry mass was observed in the mixed paper and the seed at the end of
composting. A lignin/humus content of 50% or more of the volatile solids content was
an indicator of maturity for most substrates.
12. Cellulose-to-lignin and C/N ratios decreased during composting. Cellulose-to-lignin
ratios less than 0.50 indicated maturity for most of the MSW substrates.
Chapter 6
13. Various VOCs were identified in the gaseous emissions of mixed paper, yard waste, and
food waste during composting. These waste components specifically excluded external
sources of hazardous or industrial wastes.
14. Undecomposed mixed paper was a source of various VOCs, primarily alkanes and
alkylated benzenes. From the latter group, />-isopropyltoluene, naphthalene, toluene,
ethylbenzene, and 1,3,5-trimethyl benzene were found at the highest concentrations
among the 12 quantified VOCs. Limonene and 2-ethyl-l hexanol were also found at
some of the highest relative concentrations in mixed paper. These VOCs are apparently
in the solid matrix and are released upon wetting and heating.
15. Except for 1,4-dichlorobenzene, no chlorinated VOCs were found in the gaseous
emissions from mixed paper, yard waste, or food waste.
16. The composting of seeded mixed paper and unseeded mixed paper produced
approximately 6.1 and 6.5 mg of the targeted 12 VOCs, respectively, per dry kg of
material.
7-2
-------
Section 7.0 Conclusions
17. Most of the identified VOCs in the emissions of food wastes and yard wastes during
composting are biogenic, including organic sulfur compounds, terpenes, fatty acids,
alkanes, ketones, and alcohols. Dimethyl disulfide and dimethyl trisulfide were dominant
in food wastes emissions, while limonene was present in relatively large amounts from
both substrates. Styrene was produced from unseeded yard wastes and the seed only.
18. The seed, as collected from an actual MSW composting facility, emitted 8.2 mg/dry kg of
combined toluene, ethylbenzene, p/m-xy\ene, styrene, 1,3,5-trimethylbenzene, 1,4-
dichlorobenzene,/?-isopropyltoluene, and naphthalene.
19. All measured VOCs were emitted early in the composting process and followed a first-
order-like decreasing trend as composting progressed.
20. Based on Equation 6-4, mixing of mixed paper with either yard wastes or food wastes
resulted in reductions of VOC yields due to decomposition of the VOCs with the basic
substrate. VOCs with more oxidized functional groups were found in the emissions of
mixtures of mixed paper and other substrates, compared to the emissions of
undecomposed mixed paper, indicating that VOCs were oxidized during the process.
21. Of the ethylbenzene, 88% and 69.5% was decomposed during composting when spiked
to yard wastes and newsprint, respectively, at concentrations of approximately 10,000
mg/dry kg and at mesophilic temperatures. Ethylbenzene decomposition was apparently
influenced by decomposition of the basic substrate, since yard wastes produced more
CO2 than mixed paper.
22. Approximately 90% of four alkylated benzenes spiked to MSW at total levels of 265
mg/dry kg were volatilized under continuous aeration regimes and at thermophilic
temperatures at the beginning of the composting process, with the rest being decomposed
or sorbed. The use of delayed aeration resulted in apparent decomposition of
approximately 75% of the four VOCs, with ethylbenzene and o-xylene being the most
degradable.
23. Significant decomposition and sorption of VOCs occurs during composting; most VOCs
to be emitted are volatilized and stripped early in the composting process as a function of
temperature and aeration rate.
7-3
-------
Section 8.0 References
8.0 References
Allison, F.E. 1973. Soil Organic Matter and Its Role in Crop Production. Elsevier Scientific,
Amsterdam (The Netherlands), New York, NY.
Barlaz, M. 1988. Microbiological and Chemical Dynamics During Refuse Decomposition in a
Simulated Sanitary Landfill. Ph.D. thesis, Dept. of Civil and Environmental Engineering,
University of Wisconsin-Madison, Madison, WI.
Barr, D., and S. Aust. 1994. Mechanisms white rot fungi use to degrade pollutants.
Environmental Science and Technology, 28(2):79-87.
Berthouex, P.M., andL.C. Brown. 1994. Statistics for Environmental Engineers. Lewis
Publishers, Boca Raton, FL.
Bookter, T.J., and Ham, R.K. 1982. Stabilization of solid waste in landfills. Journal of
Environmental Engineering., 108(6): 1089-1100.
Brown, K.W., J.C. Thomas, and F. Whitney. 1997. Fate of volatile organic compounds and
pesticides in composted municipal solid waste. Compost Science and Utilization,
5(4):6-14.
Chandler, J.A., WJ. Jewell, J.M. Gossett, PJ. Van Soest, and J.B. Robertson. 1980. Predicting
methane fermentation biodegradability. Biotechnology andBioengineering Symposium
No. 10. John Wiley and Sons, Inc., New York, NY.
Ciavatta, C., M. Govi, and P. Sequi. 1993. Characterization of organic matter in compost
produced with municipal solid wastes: An Italian approach. Compost Science and
Utilization, 1(1):75-81.
Clesceri, L.S., A.E. Greenberg, and R.R. Trussel (Eds.). 1989. Standard Methods for the
Examination of Water and Wastewater (17th Ed.). American Public Health Association,
Washington, DC.
Coleman, E.G., Chi-Tang Ho, and Stephen S. Chang. 1981. Isolation and identification of
volatile compounds from baked potatoes. J. Agric. FoodChem., 29:42-48.
Cornell, J.A.I 990. Experiments with mixtures: Designs, models, and the analysis of mixture
data (2nd ed.). John Wiley and Sons, New York, NY.
Crawford, D.L., and R.L. Crawford. 1980. Microbial degradation of lignin. Enzyme
Microb. Technol. ,2:11-22.
-------
Section 8.0 References
de Bertoldi, M., A. Rutili, B. Citterio, and M. Civilini. 1988. Composting management: A new
process control through oxygen feedback. Waste Management and Research, 6(5):239-
259.
de Bertoldi, M., G. Vallini, and A. Pera. 1983. The biology of composting: A review. Waste
Management and Research, 1(3):157-176.
Dec, J., and J.M. Bollag. 1994. Dehalogenation of chlorinated phenols during oxidative
coupling. Environmental Science and Technology, 28:484-490.
de Nobili, M., and F. Petrussi. 1988. Humification index (HI) as evaluation of the stabilization
degree during composting. J. Ferment. Technol, 66(5):577-583.
Diaz, L. 1987. Air emissions from compost. Biocycle, 28(3):52-53.
Diaz, L., G. Savage, L. Eggerth, and G. Golueke. 1993. Composting and Recycling Municipal
Solid Waste. Lewis Publishers, Boca Raton, FL.
Draper, N.R., and H. Smith.1998. Applied Regression Analysis (3rd Ed.). John Wiley and Sons,
Inc., New York, NY.
Effland, MJ. 1977. Modified procedure to determine acid soluble lignin in wood and pulp.
TAPPI, 60:143-144.
Eitzer, B.D. 1995. Emissions of volatile organic chemicals from municipal solid waste
composting facilities. Environmental Science and Technology, 29(4):896-902.
Eleazer, W.E., W.S. Odle, Y.S. Wang, andM.A. Barlaz. 1997. Biodegradability of municipal
solid waste components in laboratory scale landfills. Environmental Science and
Technology, 31:911-917.
Eller, P.M. 1984. NIOSH Manual of Analytical Methods, 2. U.S. Department of Health and
Human Services, Washington, DC, February.
Epstein, E. 1997. The science of composting. Technomic Publishing Company, Inc., Lancaster,
PA.
Gibson, D.T., and V. Subramanian. 1984. Microbial degradation of aromatic hydrocarbons. In
Microbial Degradation of Organic Compounds. Marcel Dekker, New York, NY.
Gibson, L.K. 1996. Toluene and ethylbenzene oxidation by purified naphthalene dioxygenase
from Pseudomonas sp. Strain NCIB 9816-4. Appl. Environ. Microbiology, 62(9):3101-
3106.
Glaub, J.C., L.F. Diaz, and G.M. Savage. 1989. Preparing MSW for composting. The Biocycle
guide to compostingMSW. JG Press, Inc. Emmaus, PA. 1989.
8-2
-------
Section 8.0 References
Goldstein, N., and R. Steuteville. 1994. Solid waste composting seeks its niche: Parti,
Biocycle, 77: 30-35.
Gould, M., and W. Meckert. 1994. Materials separation systems for solid waste composting.
Biocycle, 35(9):69-74.
Grant, W.D., and P.E. Long. 1981. Environmental Microbiology. John Wiley and Sons, New
York, NY.
Gray, K.R., K. Sherman, and G. Biddlestone. 1971. A Review of Composting - Part 1. Process
Biochemistry., 6(10):32-36. June.
Hanninen, K.I., J.T. Kovalainen, and J. Korvola. 1995. Carbohydrates as chemical constituents
of biowaste composts and their humic and fulvic acids. Compost Science and Utilization,
3(4):51-68.
Haug, R. 1993. The Practical Handbook of Compost Engineering. Lewis Publishers, Boca
Raton, FL.
Heydanek, M.G., and R.J. McGorrin. 1981. Gas chromatography - Mass spectroscopy
investigations of the flavor chemistry of oat groats. J. Agric. Food Chem., 29:950-954.
Howard, P.H., R.S. Boethling, W.F. Jarvis, W.M. Meylan, and E.M. Michalenko. 1991.
Handbook of Environmental Degradation Rates. Lewis Publishers, Boca Raton, FL.
Inoko, A., K. Miyamatsu, K. Sugahara, and Y. Harada. 1979. On some organic constituents of
city refuse composts produced in Japan. Soil Sci. Plant Nutr., 25(2):225-234.
In-Sink-Erator Company (ISE). Personal communication with Wayne Riley, Racine, WI, May
1996.
Kashmanian, R., and R.L. Spencer. 1993. Cost considerations of municipal solid waste compost
production versus market price. Science and Engineering of Composting. Proceeding of
the First International Composting Research Symposium, Columbus, OH, May 27-29,
1992. H.A.J. Hoitink and H.M. Keener, eds. Renaissance Publishers, Worthington, OH.
695-719.
Kim, J.Y., J.K. Park, B. Emmons, and D.E. Armstrong. 1995. Survey of volatile organic
compounds at a municipal solid waste co-composting facility. Water Environment
Research, 67(7): 1044-1051.
Kirk, K.T. 1984. Degradation of lignin. In Microbial Degradation of Organic Compounds, pp.
399-437. Marcel Dekker, New York, NY.
Kissel, J.C., C.L. Henry, and R.B. Harrison. 1992. Potential emissions of volatile and odorous
organic compounds from municipal solid waste composting facilities. Biomass and
Bioenergy, 3, 3-4:181-194.
-------
Section 8.0 References
Komilis, D.P., and R.K. Ham. 2000. A laboratory method to investigate gaseous emissions and
solids decomposition during composting of municipal solid wastes. Compost Science
and Utilization, 8(3):254-265. June.
La Crega, M.D., P.L. Buckingham, and J.C. Evans. 1994. Hazardous Waste Management.
McGraw-Hill, New York, NY.
Laver, M.L., and K.P. Wilson. 1993. Determination of carbohydrates in wood pulp products.
TAPPI, 76(6): 155-159.
Mathur, S.P., G. Owen, H. Dinel, and M. Schnitzer. 1993. Determination of compost
biomaturity, I. Literature review. Biological Agriculture & Horticulture, 10:65-85.
Michel, F.C., C.A. Reddy, and LJ. Forney. 1993. Yard waste composting: Studies using
different mixes of leaves and grass in a laboratory scale system. Compost Science and
Utilization, 1(3): 85-96.
Miller, F.C., S.T. MacGregor, K.M. Psarianos, J. Cirello, and M.S. Finstein. 1982. Direction of
ventilation in composting wastewater sludge. Water Pollution Control Federation,
54(1): 111-113.
Miller, F.C. 1993. Minimizing odor generation. In Harry AJ. Hoitink & Harold M. Keener
(eds.), Science and Engineering of Composting: Design, Environmental, Microbiological
and Utilization Aspects (pp. 219-241). Ohio State University, Renaissance Publications,
Worthington, OH.
Nakasaki, K., A. Nobuto, and K. Hiroshi. 1994. Accelerated composting of grass clippings by
controlling moisture level. Waste Management and Research, 12(1): 13-20.
Nakasaki, K., H. Kuratomi, H. Wakizaka, R. Hiyama, andN. Akakura. 1998. Quantitative
analysis of ammonia and odorous sulfur compounds evolved during thermophilic
composting. Waste Management and Research, 16(4):514-524.
Nelson, D.W., and L.E. Sommers. 1996. Total carbon, organic carbon and organic matter. In D.
L. Sparks et al. (eds.), Methods of Soil Analysis, Part 3. Chemical Methods (pp. 995-
996). Book Series No. 5, SSSA, Madison, WI.
Paszczynski, A., andR.L. Crawford. 1995. Potential for bioremediation of xenobiotic
compounds by the white-rot fungus Phanerochaete chrysosporium. Biotechnology
Progress, 11:368-379.
Pettersen, R.C., V.H. Schwandt, and MJ. Effland. 1984. An analysis of the wood sugar assay
using HPLC: A comparison with paper chromatography. J. Chromatogr. Sci., 22: 478-
484.
8-4
-------
Section 8.0 References
Poincelot, R.P., and P.R. Day. 1960. Rates of cellulose decomposition during the composting of
leaves combined with several municipal and industrial wastes and other additives.
Compost Science, 49:3.
Regan, R.W., and J.S. Jeris. 1970. A review of the decomposition of cellulose and refuse.
Compost Sci. ,46:1.
Reinhart, D.R. 1993. A review of recent studies on the sources of hazardous compounds emitted
from solid waste landfills: A U.S. experience. Waste Management and Research,
11:257-268.
Riffaldi, R., R. Levi-Minzi, A. Pera, and M. de Bertoldi. 1986. Evaluation of compost maturity
by means of chemical and microbial analyses. Waste Management and Research,
4(2):387-396.
Sawyer, C.N., andP.L. McCarty. 1978. Chemistry for Environmental Engineering (3rdEd.) (pp.
343-350). McGraw-Hill, Inc., New York, NY.
Schulze, K.L. 1960. Rate of oxygen consumption and respiratory quotients during the aerobic
decomposition of a synthetic garbage. Compost Science, 36:1.
Schulze, K.L. 1961. Relationship between moisture content and activity of finished compost.
Compost Science, 12:32-34.
Shevchenko, S.M., and G.W. Bailey. 1996. Life after death: Lignin-humic relationships
reexamined. Critical Reviews in Environmental Science and Technology, 26(2):95-153.
Sparks, D.L. 1995. Environmental Soil Chemistry (p. 79). Academic Press, Inc., San Diego,
CA.
Steuteville, R. 1995. MSW composting at the crossroads. Biocycle, 36(11):44-51.
Stevenson, FJ. 1965. Gross chemical fractionation of organic matter. In C. A. Black et al.
(Eds.), Methods of Soil Analysis, Part 2, Chemical and Microbiological Properties.
SSSA Book Series, No. 9, pp. 1409-1421, Madison, WI.
Stevenson, FJ. 1994. Humus Chemistry: Genesis, Composition, Reactions. John Wiley & Sons,
Inc., New York, NY.
Stinson, J.A., and R.K. Ham. 1995. Effect of lignin on the anaerobic decomposition of cellulose
as determined through the use of a biochemical methane potential method. Environ. Sci.
Technol, 29:2305-2310.
Stumm, W., and JJ. Morgan. 1981. Aquatic Chemistry: an Introduction Emphasizing Chemical
Equilibria in Natural Waters (2nd Ed.) (pp. 185-186). John Wiley and Sons, New York,
NY.
8-5
-------
Section 8.0 References
Supelco Chromatography Products Catalog. 1997. Supelco, Inc., Supelco Park, Bellefonte, PA.
Szegi, H. 1988. Cellulose Decomposition and Soil Fertility (pp. 131-132). Translated by
Erzsebet Teleki, Akademiai Kiado, Budapest, Hungary.
Tchobanoglous, G., H. Theisen, and S.A. Vigil. 1993. Integrated Solid Waste Management:
Engineering Principles and Management Issues. McGraw Hill, Inc., New York, NY.
Tenney, F.G., and S. Walksman. 1929. Composition of natural organic materials and their
decomposition in the soil: IV. The nature and rapidity of decomposition of the various
organic complexes in different plant materials, under aerobic conditions. Soil Science,
XXVIIL2.
Tolvanen, O.K., K.I. Hanninen, A.Veijanen, and K.Villberg. 1998.Occupational hygiene in
biowaste composting. Waste Management and Research, 16(6):525-540.
U.S. Environmental Protection Agency (EPA). 1998. Greenhouse Gas Emissions from
Management of Selected Materials in Municipal Solid Waste (Executive Summary).
EPA530-S-98-013. Office of Solid Waste and Emergency Response, Washington, DC.
September.
Van Durme, G.P., B.F. McNamara, and C.M. McGinley. 1992. Bench-scale removal of odor and
volatile organic compounds at a composting facility. Water Environment Research,
64(1): 19-27.
Vaughan, D., and R.E. Malcolm. 1987. Influence of humic substances on growth and
physiological processes. Humus and effect on plant growth. Dordrecht, Boston, MA.
Vicuna, R. 1988. Bacterial degradation of lignin. EnzymeMicrob. Technol, 10:646-655.
Vigon, B.W., D.A. Tolle, B.W. Cornaby, C.L. Harrison, and T.L. Boguski. 1993. Life-Cycle
Assessment: Inventory Guidelines and Principles. EPA/600/ R-92-245 (NTIS PB93-
139681). Risk Reduction Engineering Laboratory. Cincinnati, OH. January.
Wilber, C., and C. Murray. 1990. Odor source evaluation. Biocycle, March, pp. 68-72.
Wilkins, K. 1994. Volatile organic compounds from household waste. Chemosphere, 29(1):47-
53.
Yadav, J.S., and C.A. Reddy. 1993. Degradation of benzene, toluene, ethylbenzene and xylenes
(BTEX) by the lignin-degrading basidiomycete Phanerochaete chrysosporium. Applied
and Environmental Microbiology, 5 9(3 ): 75 6-762.
Young, P.J., and A. Parker. 1983. The identification and possible environmental impact of trace
gases and vapors in landfill gas. Waste Management and Research, 1:213-226.
8-6
-------
Section 8.0 References
Young, R. 1998. Personal communication. Department of Forestry, University of Wisconsin-
Madison, Madison, WI.
Zibilske, L.M. 1994. Carbon mineralization. In R. W. Weaver et al. (eds.), Methods of Soil
Analysis, Part 2, Microbiological and Biochemical Properties (pp. 835-863). SSSA
Book Series No.5, Madison, WI.
8-7
-------
Appendix A Audit Report
Appendix A
Audit Report
A-l
-------
Appendix A Audit Report
Audit Report
National Risk Management Research Laboratory
Air Pollution Prevention and Control Division
Technical Services Branch
Quality Assurance
Tracking Number 94017/III
Audit Type Technical Systems and Performance Audit
Audit Dates June 2-3,1997
Project Life Cycle Inventory of Municipal Solid Waste Composting
Project Officer Susan Thorneloe
Atmospheric Protection Branch (MD-63)
Auditors Richard C. Shores, U.S. EPA
Technical Services Branch (MD-49)
Audit Site University of Wisconsin-Madison, WI
Site Contacts Robert K. Ham, professor, University of Wisconsin-Madison.
Dimitris P. Komilis, research assistant, University of Wisconsin-Madison
A-2
-------
Appendix A Audit Report
Table of Contents
Section Page
1.0 Introduction A-4
1.1 Background A-4
1.2 Purpose A-4
1.3 Audit Summary A-5
2.0 Audit Findings A-5
2.1 Technical Systems Audit Results A-5
2.2 Performance Audit Results A-5
3.0 Audit Activities A-6
3.1 Audit Preparation A-6
3.2 Supporting Documentation A-6
3.3 On-Site Activities A-6
3.4 Auditing Activities A-7
A-3
-------
Appendix A Audit Report
Section 1
Introduction
1.1 Background
Treatment and disposal of solid waste are major concerns for most municipalities. Land
filling has been the major method to dispose of solid waste; relatively few treatment systems are
operational in the United States. Even with a treatment system, a landfill is still required for the
disposal of residues, untreated materials, and so on. The lack of nearby landfill nearby landfill
capacity has driven municipalities to find ways to divert several waste streams from landfills.
One way is to divert the organic fraction of solid waste that can be treated by aerobic biological
decomposition. This management option is called composting.
The composting process must be well-understood to be controlled and optimized.
Unfortunately, the lack of such understanding has resulted in the construction of several
inefficiently designed municipal solid waste composting facilities and the accumulation of the
product due to a limited market. Because MSW is so heterogeneous, it is important to
understand the physical and biochemical processes that govern the composting process with
respect to individual organic components (e.g., food, paper, and yard waste ).
The objective of this study is to characterize the composting process for certain MSW
components under controlled aerobic conditions. This characterization will be achieved by
conducting mass balances, primarily with respect to carbon. The components selected for
characterization are:
• food waste
• paper waste (i.e., newsprint, office paper, corrugated cardboard), and
• yard waste (i.e., branches, leaves, grass clippings).
These waste components will be degraded either individually of in combinations, while
air is forced into a vessel containing the wastes to maintain aerobic conditions. The air flow will
be controlled so that the oxygen concentration in the exhaust gas is above 15% on a per volume
basis. The air exiting the vessel is analyzed for CO2, O2, VOC, and NH3.
Research Triangle Institute (RTI) has a cooperative agreement with the U.S. EPA-Air
Pollution Prevention and Control Division (APPCD) and the University of Wisconsin at Madison
is under contract with RTI to conduct these experiments. These experiments and the audit were
conducted in Madison, Wisconsin.
1.2 Purpose
The results of these experiments will be used to manage and operate waste treatment
facilities, and estimate the best alternative for waste management. The purpose of this audit was
to evaluate the implementation of the Quality Assurance Project Plan (QAPP), prepared by The
A-4
-------
Appendix A
Audit Report
University of Wisconsin at Madison, and to assure that the data from the field study satisfied the criteria
specified in the QAPP. To accomplish this, a technical systems audit and a performance
audit were conducted. The overall objectives of these audits, were as follows:
• Evaluate the implementation of the approved QAPP.
• Conduct a performance audit to evaluate the accuracy of gaseous measurement.
1.3 Audit Summary
The audit results indicated excellent agreement with the performance audit standards and
that the method by which the experiment was being conducted was acceptable. This conclusion
is based on performance audit results, where the experimental laboratory determined the
concentration of audit gases.
Section 2.0
Audit findings
2.1 Technical Systems Audit Results
Technical Systems audits are intended to assess how well a QAPP was followed by field
personnel and identify activities that will have an adverse effect on the data collected in the field.
Observations made during the technical systems audit are listed below.
No specific problems were identified during the technical systems audit.
2.2 Performance Audit Results
Performance audits are intended to assess the accuracy of measurements made in the
laboratory. The performance audit standards were two compressed gas cylinders, one containing
carbon dioxide (CO2) and oxygen (O2) and the second containing ammonia (NH4).
The results are presented in the following table. The columns identified as GC/TCD and
Bubble Solution are the laboratory results.
Audit Gas
Carbon Dioxide,CO2
Audit Concentration1
10.1% or 6.907 gms2
GC/TCD3
10.19,12.3, 10.13 or
Average @ 10.87%
7.6% Diff.6
Bubble Solution4
6.78, 6.78, 6.88 or
Average @ 6.81 gms
or 1.4%Diff5'
A-5
-------
Appendix A
Audit Report
Oxygen,O2
Ammonia, NH3
15.16%
26. 9 ppm or 3. 256
mg5
15.29, 14.8, 15.19%
or Average @
15. 09% or 0.5%
Diff.6
None7
None7
3.026,2.929,3.263
or Average @ 3.073
mg or 2.4% Diff.6
1. Cylinder gas concentration as reported by Scott Speciality Gases, NIST traceable to <+l- 5
percent.
2. Bubbling audit gas for 202.5 minutes at 189.5 accm and calculating the actual mass of carbon
dioxide bubbled through the solution.
3. GC/TCD analysis of gas is used to ensure that the vessels are maintaining 15% O2 and also as
a check on the CO2 concentration.
4. The mass balance calculations and waste product emissions are calculated using the bubbling
solution concentrations.
5. Bubbling the audit gas for 946 minutes at 187.1 accm and calculating the actual mass of
ammonia bubbled through the solution.
6. Difference calculated from {[(Laboratory-Audit)/Audit]x 100}t Laboratory values are
GC/TCD and Bubbling Solution.
7. Oxygen concentrations are only measured using the GC/TCD and Ammonia is only measured
using the bubbling technique.
Section 3.0
Audit Activities
3.1 Audit Preparation
Preparation for this audit included planning meetings with the Project Officer, review of
the QAPP, prepared by the University of Wisconsin, discussions with both the RTI and the University
of Wisconsin personnel.
Audit gases were obtained from Scott Speciality to evaluate the accuracy of
concentrations measured in the University of Wisconsin Laboratory. These gases are described
in the following table.
Cylinder Number
BO02415
ALM058883
Component Gas
Carbon Dioxide
Oxygen Nitrogen
Ammonia
Nitrogen
Certified
Concentration
10.10 percent/molar
15.16 percent/molar
balance gas
26.9 ppm-moles
balance gas
Certified Analytical
Accuracy
+/- 5%
+/- 5%
+/- 5%
A-6
-------
Appendix A Audit Report
3.2 Supporting Documentation
Documentation referred to during this audit included the Quality Assurance Project Plan,
prepared by the University of Wisconsin, dated October, 1996. This Quality Assurance Project.
Plan was reviewed and approved by the APPCD QA Staff. The specified date identifies the final
revision, in response the APPCD QA Staff review comments.
3.3 On-Site Activities
Monday, June 2,1997
Traveled from RTF to Madison, Wisconsin. After arriving in Madison, drove to the
Federal Express office and picked-up the audit gas cylinders. Met a University of Wisconsin
representative at the hotel and transported the cylinders to the laboratory building.
The audit began with an organizational meeting, to give everyone a chance to explain
where roles Auditors arrived at the University, the US EPA project officer, Susan Thorneloe gave
a presentation of the Life Cycle project. This presentation included a discussion of what data
was being collected and what organizations are interested in the results or the project. Richard
Shores, the EPA auditor, gave an introduction on the activities associated with the audit and what
was expected or the University Laboratory.
Went into the laboratory, reviewed the measurement systems and started the ammonia
audit. It was important to start the ammonia gas bubbling to ensure that enough mass would be
captured in the bubble solution. Ended the audit activities and traveled to the hotel.
Tuesday, June 3,1997
Met at the Laboratory and began auditing activities immediately. The Ammonia bubbler
was stopped and began the Carbon Dioxide bubbler. Reviewed the calibration and QC activities
associated with the GC/TCD used for gas analysis. Observed ambient air and span gas analysis
to ensure that the GC/TCD was functioning properly. Reviewed Ammonia and Carbon Dioxide
analysis procedure. Completed the auditing activities with an exit meeting/discussion meeting
with the University of Wisconsin personnel. Shipped the audit gases via Federal Express and
flew back to RTF.
3.4 Auditing Activities
The concentration of the performance audit gases were evaluated using both the Varian
3300 GC/TCD gas analysis system and the bubbling system. Gas samples were analyzed using
the GC/TCD by injecting a small volume of gas, with an air tight syringe into the GC/TCD. These gas
samples included oxygen and carbon dioxide.
A-7
-------
Appendix A
Audit Report
To evaluate the consistency of analysis and the consistency or gas concentration analysis,
the audit cylinder gas was analyzed three times with the following results.
Date
June 2, 1997
June 2, 1997
June 2, 1997
Time
15:36
16:05
16:30
%O2
ConcentrationA
15.29
14.8
15.19
% CO2
Concentration8
10.19
12.3
10.13
ATrue concentration of 15.16 %.
BTrue concentration of 10.1 %.
These results would indicate that there is the potential for variability in the gas analysis
concentrations determined using the GC/TCD and that any anomalous result should be
reanalyzed. These results also emphasize the importance of calculating out the actual
concentration from the GC/TCD results before proceeding with other work. Based on these
results, the consistency of sample analysis when analyzing the vessel gases numerous times is
acceptable and meets the project goals. Vessel # 5 was also analyzed numerous times to evaluate
analysis variability when analyzing gases containing moisture. These results are as follows.
Date
June 2, 1997a
June 3, 1997
June 3, 1997
June 3, 1997
Time
14:48
10:37
12:16
13:06
%O2
Concentration
8.2
13.25
14.38
14.9
% CO2
Concentration
10.96
8.06
7.84
7.14
laboratory operator changed the vessel # 5 flow rate after this measurement because of the low
oxygen concentration. Project goals were to maintain 15% oxygen concentration within the
vessels.
The three samples collected on June 3 seem to indicate a trend where the oxygen concentration
is approaching 15 %, in response to having the flow rate adjusted.
A-8
------- |