I41 w'n,‘

944;:
‘ - ..
‘ ‘3“.

1..wa

J. 1;.
2.2.0.! ::

n. .u ”uh-“4‘" 'htl'u.

tumufi

‘ 5......

4

.u A.

t»

:1 '1'?

 

—2!
. ”can” {I

; $ufi§fit~fi
:. .. ‘1

 

 

I STATE [ES

llljllilillllllillllll

93 01020 0719

my

 

 

 

 

l

“1593

This is to certify that the

thesis entitled

Biodegradability of Polyactide Film in
Simulated Composting Environments

presented by

Joseph Brenton Snook

has been accepted towards fulfillment
of the requirements for

M. 5. degree in Chemical Engr.

 

 

(mm MGM/m

‘ Major professor /

 

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

 

 

 

LIBRARY 1
Michigan State!

University

————

”M01110!“ ywtncotd.

E II RETURN BOXto roman
dd. duo.

FLAG
To AVOID FiNES mum on or baton

E DUE DATE DUE DATE DUE

   
  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

BIODEGRADABILITY OF POLYLACTIDE FILM IN
SIMULATED COMPOSTIN G ENVIRONMENTS

By

Joseph Brenton Snook

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Chemical Engineering

1 994

ABSTRACT

BIODEGRADABILITY OF POLYLACTIDE FILM IN
SIMULATED COMPOSTING ENVIRONMENTS

By

Joseph Brenton Snook

Polylactide is one of many new biodegradable materials being designed for
compatibility with the developing composting infrastructure. Before using any
new material as a part of a compostable waste stream, it is essential to understand
its behavior during composting. Experiments were performed to determine the
rate, extent, and mechanism of polylactide film biodegradation in simulated
composting environments. The effect of polylactide film on the physical,
chemical, and biochemical processes of composting was also addressed.
Polylactide biodegrades in compost by undergoing random non-enzymatic
hydrolysis until the molecular weight average is reduced to approximately Mn
10,000. At this molecular weight average, polylactide oligomers diffuse out of the
polymer bulk and are utilized by microorganisms. Polylactide film rapidly
disintegrated and mineralized in a bench-scale respirometric biodegradability test
and a semi-controlled compost exposure, however, a feedback—controlled pilot-
scale experiment with a lower average temperature resulted in a slower rate of
molecular weight reduction. A temperature dependent model of polylactide
molecular weight reduction was used to accurately predict the extent of polylactide

degradation in the three simulated composting environments.

To my parents, sisters, and brother.

ACKNOWLEDGMENTS

I wish to first thank my advisor, Dr. Ramani Narayan, for providing his uniquely
styled guidance and support. I would also like to express my appreciation to
everyone at the Michigan Biotechnology Institute (particularly the Biomaterials
Group) and the Department of Chemical Engineering for providing an enjoyable
working and learning environment. I am grateful to the Biodegradable Polymers
advisory committee of the American Society of Testing and Materials Institute of
Standards Research for their support of this research and especially Mr. Mike
Chanen of Cargill for his assistance and many helpful discussions. Finally, I thank
my friends in the Department of Agricultural Engineering for their own special

contributions.

iv

TABLE OF CONTENTS

Chapter Page
LIST OF TABLES ............................................................................................ viii
LIST OF FIGURES .............................................................................................. x
1. INTRODUCTION ............................................................................................ 1
1.1 Objectives ........................................................................................... l

1.2 Organization of the Thesis .................................................................. 2

2. BIODEGRADABLE MATERIALS .................................................................. 3
2.1 Designing Materials for Recycling and Disposal ................................. 3

2.2 Biodegradable Materials in Composting ............................................. 4

2.3 Status of the US. Biodegradable Materials Industry ......................... 11

3. POLYLACTIDE ............................................................................................ 15
3.1 Polylactide Synthesis ........................................................................ 15

3.2 Biodegradation Mechanism and Kinetics .......................................... 16

3.2.1 Hydrolytic Degradation ....................................................... 17

3.2.2 Enzymatic Degradation ....................................................... 22

3.3 Life Cycle Assessment ...................................................................... 23

3.4 Applications ...................................................................................... 23

3.5 Polylactide Film Properties ............................................................... 25

\

vi

Chapter Page

4. ANALYTICAL METHODS .......................................................................... 27

5. BIODEGRADATION UNDER EXTERN ALLY CONTROLLED

COMPOSTIN G CONDITIONS ......................................................................... 31
5.1 Overvrew ................................. 31
5.2 Test method ...................................................................................... 31
5.3 Preliminary Experiment 1 ................................................................. 35
5.3.1 Method ................................................................................ 35
5.3.2 Results ................................................................................ 37
5.4 Preliminary Experiment 2 ................................................................. 39
5.4.] Method ................................................................................ 39
5.4.2 Results ................................................................................ 43
5.5 Preliminary Experiment 3 ................................................................. 47
5.5.1 Method ................................................................................ 47
5.5.2 Results ................................................................................ 51
5.6 Biodegradability of Polylactide Film ................................................. 55
5.6.1 Method ................................................................................ 55
5.6.2 Respirometric Biodegradability ........................................... 56
5.6.3 Polylactide Degradation ...................................................... 60
6. BIODEGRADATION UNDER FEEDBACK CONTROLLED
COMPOSTING CONDITIONS ......................................................................... 64
6.1 Overview .......................................................................................... 64
6.2 Pilot-scale Composting System ........................................................ 64
6.2.] Process Control ................................................................... 66
6.2.2 Controller Tuning by the Continuous Cycling Method ........ 70
6.2.3 Controller Tuning by a Process Model-Based Method ........ 72

6.3 Evaluation of Test Matrices .............................................................. 79

vii

Chapter Page
6.3.1 Mixed MSW Compost ........................................................ 79
6.3.2 Sludge Compost .................................................................. 84
6.3.3 Yard Debris Compost .......................................................... 88
6.4 Biodegradation of Polylactide Film ................................................... 97
6.4.1 Experimental Method .......................................................... 98
6.4.2 Effect of Polylactide on Process Dynamics ......................... 99
6.4.3 Effect of Polylactide on Product Quality ........................... 106
6.4.4 Pilot-Scale Polylactide Degradation .................................. 110
7. BIODEGRADATION UNDER SEMI-CONTROLLED
COMPOSTING CONDITIONS ....................................................................... 115
7.1 Overview ........................................................................................ 115
7.2 Test Apparatus and Conditions ....................................................... 116
7.3 Polylactide Degradation .................................................................. 119
8. CONCLUSIONS AND RECOMMENDATIONS ......................................... 121
8.1 Conclusions .................................................................................... 121
8.2 Recommendations for Further Work ............................................... 124

LIST OF REFERENCES .................................................................................. 127

LIST OF TABLES

Table Page
21. Growth in number of composting facilities, U.S. ........................................... 8
2-2. Life cycle requirements for compostable items. ............................................. 9
2-3. Major biodegradable materials producers .................................................... 14
3-1. Rate data used in Arrhenius plot .................................................................. 21
3-2. Characteristics of polylactide film ............................................................... 26
5-1. Results of experiment 1. .............................................................................. 38
5-2. Results of aqueous biodegradability testing ................................................. 39
5-3. Test materials for experiment 2. .................................................................. 40
5-4. Results of experiment 2. .............................................................................. 46
5-5. Results of biodegradability experiment ........................................................ 60
5-6. GPC analysis of films from controlled composting ...................................... 62
6-1. Controller settings from the continuous cycling method. ............................. 72
6-2. Process characteristics for oxygen concentration ......................................... 76

viii

ix

Table Page
6-3. Cohen-Coon settings for oxygen concentration gradient. ............................. 76
6-4. Cohen-Coon process characteristics for temperature ................................... 77
6-5. Settings for temperature control ................................................................... 77
6-6. Analysis of MSW compost feedstock. ......................................................... 80
6-7. Analysis of sludge compost feedstock. ........................................................ 85
6-8. Elemental analysis of Fairfield compost product. ........................................ 86
6-9. Analysis of yard debris materials ................................................................. 88
6-10. Analysis of Kraft paper/yard waste mixture ............................................... 93
6-11. Carbon balance for Kraft paper/yard waste mixture. .................................. 94
6-12. Change in C/N during Kraft/yard waste composting .................................. 94
6-13. Distribution of carbon in humic fractions. ................................................. 97
6-14. Mixtures for pilot-scale composting experiments ....................................... 98
6-15. Compost feedstock characteristics ............................................................. 99
6-16. Carbon balances for pilot-scale composting ............................................. 106
6-17. GPC analysis of films from pilot-scale composting ................................. 1 1 1
7-1. Synthetic compost mixture for semi—controlled composting ....................... l 17

7-2. GPC analysis of films from semi-controlled composting ........................... 119

LIST OF FIGURES

Figure Page
2- 1. Elements of the composting process .............................................................. 5
22 Steps in compost processing .......................................................................... 6
3-1. Synthesis of polylactide ............................................................................... 15
3-2. The polylactide cycle .................................................................................. 16
3-3. Proposed degradation mechanism for polylactide ........................................ 19
3-4. Example of rate constant calculation ........................................................... 20
3-5. Arrhenius plot for polylactide degradation rate constant .............................. 20
5-1. Setup for externally controlled composting experiment 1 ............................ 36

5-2. Biodegradability of acrylic-coated and uncoated paperboard under the

conditions of experiment 1 ........................................................................... 38
5-3. Setup for externally controlled composting experiment 2 ............................ 41
5-4. Biodegradability of materials tested in experiment 2 ................................... 44
5-5. Rates of control material biodegradation measured in experiment 2 ............ 45
5-6. Setup for externally controlled composting experiment 3 ............................ 48

X

xi

Figure Page

5-7. Predicted effect of air flow rate on time required to achieve 97%

purge of gas analysis loop ............................................................................ 50
5-8. Replacement of gas in analytical loop at 60 ml/min ..................................... 50
5-9. Carbon dioxide concentration of the gas exiting each test vessel ................. 52
5-10. Rate of biodegradation of Kraft paper ....................................................... 53
5-1 1. Biodegradability of Kraft paper in experiment 3 ........................................ 54

5-12. Carbon dioxide concentrations in exhaust gas from blank and
polylactide test vessels ................................................................................. 57

5-13. Carbon dioxide concentrations in exhaust gas from Kraft paper and
polyethylene test vessels .............................................................................. 58

5—14. Rates and cumulative biodegradabilities of polylactide film and
Kraft paper .................................................................................................. 59

5-15. Comparison of predicted and measured decreases in Mn with rate
of mineralization during controlled composting ........................................... 61

5-16. Photographs of the mixture of polylactide film and compost inoculum
before and after externally controlled composting ........................................ 63

6-1. Schematic of pilot-scale composting apparatus ............................................ 65

6-2. Typical temperature profile during forced aeration composting (Stentiford
et al., 1985) ................................................................................................. 67

6-3. Primary control algorithm for dual control of temperature and oxygen
concentration ............................................................................................... 69

6-4. Characteristics of the process reaction curve ............................................... 74

xii
Figure Page
6-5. Response of oxygen concentration to a step decrease and increase in the
air flow rate ................................................................................................. 75

66 Response of temperature to a step increase and decrease in the air
flow rate ..................................................................................................... 78

6-7. Process dynamics of MSW compost over the first 32 hours of operation
with no mixing ............................................................................................. 81

6-8. Process dynamics for pilot—scale composting of MSW over the first
nine days ..................................................................................................... 83

6-9. Process dynamics for pilot-scale composting of sludge-wood chips
mixture ..................................................................................................... 87

6-10. Typical process dynamics for pilot-scale yard debris composting trials ..... 89

6-11. Process dynamics for pilot-scale composting of Kraft paper in yard
debris mixture .............................................................................................. 91

6-12. Chemical parameters during pilot- scale composting of Kraft paper

mixture .................................................................................................... 92
6-13. Contribution of each extracted fraction to the total extractable carbon ....... 96
6-14. Process dynamics for compost mixture containing polylactide film ......... 100
6-15. Average temperatures in the vessels during pilot-scale composting ......... 101
6-16. Respirometric data for compost mixture containing polylactide film ....... 103

6-17. Percent mineralization of initial carbon to carbon dioxide during
pilot-scale composting ............................................................................... 104

6-18. Pilot-scale composting product quality parameters .................................. 108

xiii

Figure Page

6—19. Results of germination and root elongation bioassay for fresh and
finished pilot-scale composts ..................................................................... 110

6-20. Comparison of predicted and measured decrease in MD during
pilot-scale composting .............................................................................. 1 12

6-21(a). Photograph of polylactide film prior to pilot-scale composting ............ 113

6-21(b) Photographs of polylactide film after composting two weeks and
four weeks ................................................................................................. 114

7-1. Apparatus for semi-controlled composting experiment .............................. 1 16

7—2. Processing conditions for semi-controlled composting exposure of
polylactide film .......................................................................................... 1 18

7-3. Comparison of predicted and measured decrease in Mn during
semi-controlled composting ...................................................................... 120

CHAPTER I

 

INTRODUCTION

 

The development of inexpensive synthetic plastics has had a dramatic impact on
modern society. Single-use disposable plastic packaging provides both
convenience and public health benefits. However, the proliferation of disposable
products has also resulted in tremendous volumes of solid waste. Although
landfilling is currently the most common method of solid waste management, its
long term economic and environmental costs are becoming apparent. Public
objection to landfilling has impeded the siting and construction of new facilities.
This has resulted in a shortage of landfill space and a need for alternative solid
waste management practices. The use of biodegradable plastics in conjunction
with composting can divert a significant portion of the solid waste stream from

landfills to be reused as a soil amendment.

1.1 OBJECTIVES

The primary objective of the thesis is to evaluate the potential of one material,
polylactide, as a biodegradable replacement for current plastic packaging

materials. This objective is attained by:

- Describing how polylactide can be used in conjunction with
composting to provide a cost-effective and environmentally sound

option for solid waste management.

2
o Detemiining the rate and extent of polylactide biodegradation in

simulated composting environments.

- Studying the physical, chemical, and biochemical interactions in the
composting process that may be affected by the inclusion of

polylactide as a substrate.

0 Modeling polylactide biodegradation in compost.

1.2 ORGANIZATION OF THE THESIS

The thesis is divided into two parts. The first part is a development of the general
and technical background for the work. Chapter 2 reviews the role of
biodegradable materials in solid waste management and presents the status of the
biodegradable material manufacturing industry. Aspects of polylactide production,
use, and biodegradation are reviewed in Chapter 3. The discussion of polylactide
includes a model of polylactide molecular weight reduction and its generalization
to temperatures occurring in compost. Chapter 4 details the experimental methods

used in the research.

The second part of the thesis, Chapter 5 through Chapter 7, describe the
experiments performed to evaluate polylactide biodegradability in compost. The
experiments in each of these chapters use a different apparatus to simulate the
composting environment. Chapters 5, 6, and 7 each begin by relating the rationale
of the test method and describing preliminary experiments used to evaluate and
validate the apparatus and methods. Polylactide molecular weight reduction and
biodegradation using each test method are then compared to the predictions of the
model. Chapter 8 summarizes the findings of the experiments and presents

recommendations for further work.

CHAPTER 2

 

BIODEGRADABLE MATERIALS

 

2.1 DESIGNING MATERIALS FOR RECYCLING AND DISPOSAL

The importance of redesigning single-use packaging for recycling and ultimate
disposal has been discussed extensively (Narayan, 1989; 1991; 1992, 1993a,
1993b; Snook and Narayan 1994). Today's polymeric materials were designed
with little consideration for their ultimate disposability or recyclability. This has
resulted in new environmental regulations and societal concerns regarding the
environmental consequences of such materials when they enter the waste stream
after their intended uses. Of particular concern are polymers used in single use,
disposable plastic applications. Plastics are strong, light-weight, inexpensive,
easily processable and energy efficient. They have excellent barrier properties and
are very durable. However, it is these very attributes of strength and
indestructibility that cause problems when these materials enter the waste stream.
They are resistant to biological degradation because microorganisms do not have
enzymes capable of degrading and utilizing most man-made polymers. In
addition, the hydrophobic character of plastics inhibits enzyme activity, and the
low surface area of plastics with their inherent high molecular weight further

compounds the problem.

Therefore, there is an urgent need to redesign and engineer new polymeric
materials that have the needed performance characteristics of the plastics, but can

be transformed in appropriate waste disposal infrastructures to products that are

4
compatible with the environment or are recyclable to the same or other products.
In addition, after the material's intended use, it must end up in appropriate disposal
infrastructures that use these recyclability or biodegradability attributes. The use
of biodegradable polymeric materials in single use disposable packaging,
nonwovens, and paperboard coatings, in concert with composting infrastructures,
is an ecologically sound approach to managing the resultant waste and diverting it

from landfills.

2.2 BIODEGRADABLE MATERIALS IN COMPOSTING

2.2.1 Composting

As shown in Figure 2-1, composting is a process involving interdependent
biodegradation and polymerization reactions. Composting is defined as
"accelerated degradation of heterogeneous organic matter by a mixed microbial
population in a moist, warm, aerobic environment under controlled conditions"
(Narayan, 1993a). Microorganisms break down cellulosics and other
carbohydrate-based substrates. The breakdown products are used for cellular
metabolism with the formation of biomass, new microorganisms, and the evolution
of carbon dioxide. Lignolic species are broken down into phenolic components.
Some phenolic components are formed during cellular metabolism and released
during lysis of dead microorganisms. It has been suggested that the phenolics
react with nitrogenous compounds resulting in the formation of humic substances
or humic precursors (Flaig, 1975, 1988; Hayes, et a1. 1989). The beneficial effect
of compost as a horticultural material is in part attributed to these humic

substances.

 

Microorganisms Oxygen Moisture

\ l /

Nutrients

 

N.P.K.--- . ORGANIC/COMPOSTABLE MATERIAL
(carbon source)

 

 

 

 

 

 

 

 

———->
Chemical BIOUBQIWWOH HEAT
degradaflol/ ‘ \

Breakdown ]‘_ CELL MASS coz * H2°
Produlcts death
Polymerlntlm l

 

 

 

I . HUM US/
> COM POST

Figure 2-1. Elements of the composting process.

 

 

 

Composting is typically a multi-step process that can be catagorized by the
configuration and control techniques employed in the early stages. Composting
operations can be divided into four types: turned windrow, static pile, in-vessel,
and hybrid processes. Turned windrow processing involves the formation of piles
(windrows) of the material to be composted, typically 2 m high, 3 to 4 m wide, and
30 to 100 m long. Temperature control, oxygenation, and mixing are provided by
periodically turning the windrow using a front-end loader or specialized windrow
turning equipment. In the static pile approach, proper conditions are maintained
by using forced aeration in place of windrow turning. In-vessel systems are

designed to provide optimal conditions by combining forced aeration, intensive

6
mixing, and complete enclosure of the compost. A variety of hybrid composting
schemes are used that combine elements of turned windrow, static pile. and in—

vessel composting.

 

 

 

Feed-
Product
stock: Post-
Processing
rejects rejects

 

 

 

Figure 2-2. Steps in compost processing.

Figure 2-2 shows the four steps that are common to many composting operations.
Pre-processing involves the collection and preparation of the compost feedstock
materials, which typically includes some type of shredding or other size reduction.
The high-rate composting phase is characterized by high oxygen demand and
thermophilic temperatures. Frequent windrow turning, forced aeration, or in-
vessel operation are required to control the process and to prevent odors during
this stage, which typically lasts several weeks. The curing stage is normally
performed by storage in a windrow or non-aerated pile over several months with

minimal processing. Temperature and oxygen demand decrease during this phase,

7
which continues until a fully-stabilized product is obtained. Post-processing
usually consists of passing the mature compost through a 1/4- or 1/2— inch screen

to remove any non-degraded items and ensure a fine and uniform product.

A major logistical advantage of composting over other types of recycling is that of
it requires a only a semi-segregated waste stream. Conventional recycling requires
the separate collection, transportation, and processing of each component of the
waste stream. Processing a mixture of materials is not only possible by
composting, it is often essential to the production of quality compost product.
Composting can simultaneously handle any mix of materials that are compatible
with the composting infrastructure. When operated properly using suitable
feedstocks, low-technology composting offers an economically efficient and

environmentally sound alternative for managing certain waste streams.

2.2.2 Growth of composting infrastructures

After many years in obscurity, composting has returned as an important approach
to managing waste, especially the biodegradable component of our waste stream.
Over 3,000 facilities compost yard waste, about 180 compost sludge, 30 compost
food and food processing waste, and 25 compost mixed municipal solid waste
(MSW). Table 2—1 shows the dramatic growth in the number of facilities,
especially for yard waste composting. Since 1988, an average of nearly 500 new
yard waste composting facilities have opened each year due to state laws

forbidding landfilling, incineration, and open burning of yard wastes.

A number of factors have contributed to the growth of composting infrastructures:

1. Nearly every state has either established or raised recycling goals within

the past five years. Based on the composition of MSW, the only way to

8
reach goals of 30, 40, or 50% recycling is to have a substantial composting
element. The US. EPA and many states identify composting as a form of
recycling.
2. Relative costs for disposal are much higher. The unpopularity of landfills
and incinerators and the strict regulations governing them have pushed
disposal costs sharply upward. In many parts of the country composting is
becoming competitive with other waste management approaches.
3. Separation technologies have improved and contaminants can be
effectively separated. Also, community separation programs have had
excellent participation.
4. Legislative mandates have been the biggest factor. Twenty six states plus
the District of Columbia now have some kind of provision to separate and/or
ban yard waste from landfills. Some are outright bans, others are incentives

and endorsements.

TABLE 2-1. Growth in number of composting facilities, U.S.

 

 

 

 

Operating composting facilities at year end
YEAR YARD WASTE SLUDGE MIXED MSW
1988 651 79 --
1989 986 1 l9 7
1990 1,407 133 13
1991 2,201 149 18
1992 2,980 159 21
1993 3,014 182 25

 

 

 

 

Sources: Biocycle, U.S. Composting Council

 

9
Another major incentive driving composting is that many states have instituted
compost procurement policies. Thus, markets for the finished compost have been
created by government fiat. In many states, the state agencies and local
governments are required to procure compost product for land maintenance
activities including highway construction, landscaping, re-cultivation, and soil

erosion control.

2.2.3 Designing materials for compostability

Life Cycle Assesment (LCA) has been used since the 1960’s to analyze the
environmental impacts associated with a specific product (Curran, 1993). The
Society of Environmental Toxicology and Chemistry has defined LCA as “looking
holistically at the environmental consequences associated with the cradle-to-grave
life cycle of a process or product (Fava et al., 1990). For a compostable item to be
properly designed using a biodegradable material, each aspect of the product’s life
cycle needs to be addressed as shown in Table 2-2. This table lists the necessary
attributes in the order that each is required. Issues related to composting and

biodegradability (items 5 through 7) are discussed in detail below.

TABLE 2-2. Life cycle requirements for compostable items.

 

1. Environmentally sound raw materials and production.
Processability.

Price acceptable to markets or legislation requiring use.
Functionality acceptable to end-users.

Post-use routing to compost infrastructure.

Compostability .

395"???)

Complete biodegradability.

 

 

 

10
For a waste stream to be useful as a feedstock for composting, all of its
constituents must be biodegradable, have no negative impact on compost
processing, and contribute to formation of a quality compost product. To meet
these requirements, biodegradable plastic items must be made of materials that are
both inherently biodegradable and compostable. A material that is inherently
biodegradable can, under the proper conditions, be metabolized by
microorganisms within a reasonable time frame so that the material does not
accumulate in the environment. Compostability requires that the material can be

processed in a composting system without ending up in the reject fractions.

Polymeric materials, when designed for compostability using renewable resources
(agricultural products) as the major raw material, can be inherently biodegradable
and be incorporated into the compost product. Biodegradation of such natural
materials will produce high quality compost, along with water and carbon dioxide
as by-products. Compost amended soil can have beneficial effects by increasing
soil organic carbon, increasing water and nutrient retention, reducing the need for
additional chemical inputs, and suppressing plant disease. By using the compost
product in agriculture, the carbon in the polymeric materials, compost, and carbon
dioxide become a part of the biological carbon cycle. The problem of waste

disposal could become the solution for low-input sustainable agriculture.

Studies by the Michigan Biotechnology Institute have demonstrated that
conventional plastics are inherently recalcitrant to biodegradation and interfere
with the composting process (Snook and Narayan, 1994). For composting to play
an ecologically sound role in the biological carbon cycle, the compost product

must be of high quality. The presence of inert contaminants such as fragments of

11
glass and plastic have a negative affect on compost appearance and marketability.
Because of the volume and mass reductions that occur during composting, a few
percentages of recalcitrant, inert synthetics in the feed material results in a
substantial concentration in the compost product. When a compost containing a
small amount of inert material is applied to soil repeatedly, a build up of the inert
material will occur as compost organic carbon is utilized. In the end, the
processability and product quality of compost depends on the quality of every

feedstock stream.

2.3 STATUS OF THE US. BIODEGRADABLE MATERIALS
INDUSTRY

The US. biodegradable materials industry made a false start by introducing starch-
filled polyethylene and photodegradable plastic bags as biodegradable materials in
the 1980’s. Films produced from polyethylene with 6-15% starch were designed
to disintegrate as the starch component was biodegraded. However, only the
starch particles at the surface of the film biodegraded, leaving the embedded starch
particles and the polyethylene matrix intact. Photodegradable films were also
designed for disintegration by incorporating photosensitive groups or additives into
the plastics. Unfortunately, when bags made of photodegradable plastics ended up
in landfills or compost piles, they were not exposed to the ultraviolet radiation
(sunlight) necessary to initiate radical induced degradation. These initial failures

made the public suspicious of biodegradable plastics.

Biodegradable products being introduced today are fully biodegradable or

compostable. US. Federal Trade Commission guidelines state that:

12
“Unqualified degradability claims should be substantiated by evidence
that the product will completely break down and return to nature, that is,
decompose into elements found in nature within a reasonably short

period of time after the customer disposes of it in the customary way.”

( U .S . Federal Trade Commission, I 992 )

Because of this requirement, the industry is active in the development of standard
methods for determining biodegradability to substantiate biodegradability claims.
ASTM standards have been developed for determining intrinsic biodegradability,
biodegradability under composting conditions, and biodegradability under other
environmental conditions. Biodegradable material producers also recognize the
need for simultaneous development of appropriate waste management
infrastructure, such as composting, so that the material’s biodegradability
attributes are used when customers dispose of the products in the customary way.
The biodegradable plastic industry’s attention to scientific substantiation of
degradability claims and the public’s appreciation of the role for biodegradable
materials in solid waste management is leading to greater acceptance of

biodegradable materials.

A broad range of biodegradable materials are being produced today or are planned
for production, as shown in Table 2-3. Several of these fall under the category of
natural biopolymer based materials, such as cellophane, cellulose and starch esters,
and starch blends. A second set of technologies are based on polyesters produced
from fermentation products. These include the polylactide polymers and poly-

(hydroxybutyrate-co-hydroxyvalerate) polymers. The remaining biodegradable

l3
polymers (polyvinyl alcohols. polyethylene oxides, polycaprolactones) are

chemically synthesized from petrochemical feedstocks.

The major target markets for biodegradable polymeric materials are single-use
disposable packaging, disposable serviceware (cutlery, plates, cups), disposable
nonwovens (diapers, personal care products), and paper and film coatings. More
than 50 billion pounds of material are used and disposed of annually in these
markets (Narayan, 1993b). The planned capacities of the existing biodegradable
plastics from Table 2-3 is less than 1 billion pounds, leaving a large potential

market share yet to be claimed by biodegradable materials.

14

TABLE 2-3. Major biodegradable materials producers (Narayan, 1993).

 

 

 

 

 

 

 

 

 

 

 

 

 

 

COMPANY BASE POLYMER FEEDSTOCK COST, CAPACITY, MM
Mb LB/Y R
Cargill, Minn' capolis, MN Polylactide (EcoPLA) Renewable Resources, 1.00 -3.00 10 ('94 scaleup): 250
(.‘om (mid-1996)
Ecochem. Wilmington. DE Polylactide copolymers Renewable Resources. < 2.00 0.15 (‘94 scaleup)
Cheese whey, Com proj'd
Flexel. Atlanta GA Cellophane. (Regenerated Renewable resources 2.15 100
cellulose)
Zeneca. (business unit of [Cl Polyfhydroxybutyrate-co- Renewable resources -- 8.00 - 0.66. additional
Americas) hydroxyvalerate). (PHBV) carbohydrates (glucose), 10.00, 4.00 capacity slated for'96
organic acids proj'd is 11 - 22
Novamont. Montedison, Italy Starch—synthetic polymer blend Renewable resources + 1.60 - 2.50 50. in Tumi. Italy
containing approx. 60% starch petrochemical
Novon Products (Warner- Thermoplastic starch polymer Renewable resources, 2.00 - 3.“) 100
Lambert). Morris Plains. NJ‘ compounded with 5-25% additives Starch
Union Carbide, Danbury. CT Polycaprolactone (TONE polymer) Petrochemical 2.7 < 10
Air Products 8:. Chemicals, Polyvinyl alcohol (PVOH) & Petrochemical 1.0 - l .25 150 — 200 (water sol.
Allentown, PA Thermoplastic PVOH alloys (PVOH); PVOH ); 5 (V INEX)
(VINEX) 250-3.“)
(V INEX)
National Starch & Chemical. Low ds starch ester Renewable resources, 2.00 - 3.(X) Not available
Bridgewater, NJ Starch
MEI/Japan Com Starch Co., Water repellent. thermoplastic Renewable Resources, 1.0 - 1.50 0.1(pilot scale); 150
Lansing. MI modified starches (AMYPOL) Starch slated for eariy '96
Planet Packaging Technologies, Polyethylene oxide blends Petrochemical 3.00 10
San Diego, CA (Envimplastic)
Showa Highpolymer C 0.. Ltd. condensation polymer of glycols Petrochemical approx 0.2 (pilot); 7(semi-
with aliphatic dicarboxylic acids 3.00 commercial. end '94)

 

(BIONELLE)

 

 

 

 

 

CHAPTER 3

 

POLYLACTIDE

 

3.1 POLYLACTIDE SYNTHESIS

Lactic acid polymers prepared by ring-opening polymerization of the lactide dimer
are known as polylactides (see Figure 3-1). The six-membered cyclic diester
L-lactide is prepared by dehydrating the naturally occurring form, L-lactic acid.
High molecular weight poly—L-lactides can be synthesized in a controlled manner
using a variety of metal catalyst initiators, preferably tin catalysts (Dijkstra et al.,
1990). The poly-L-lactide polymer (PLLA) is an optically active, isotactic,

semicrystalline polymer, with a glass transition temperature (Tg) around 58°C and

 

cu3 CH3
2n Lactic Acid "04'”. m2" m_cfl m2"
* catalyst
0 ° cu3
n Lactide + 2n H20 I 1
CH, 0 o
i catalyst
ca, 0 CH, 0
Polylactide (_o-.',.._£_o_é.._ll,_) n

 

 

 

Figure 3-1. Synthesis of polylactide.

15

16

melting temperature (Tm) around 178°C. This material has a very narrow window
of processibility, since molecular weight degradation occurs at around 180°C.
Polymerization with a small amount of lactide containing D-lactic chiral centers
disrupts the crystallinity and produces an amorphous polylactide. This decreases
the melting point from 178°C to 124°C as the composition is changed from pure

L-lactide to 8% of the D-lactide (Spinu, 1993) and improves processability.

3.2 BIODEGRADATION MECHANISM AND KINETICS

Polylactide degradation occurs in two distinct stages. First, random non-
enzymatic chain scission of the ester groups leads to a loss in molecular weight.
The molecular weight of the polymer eventually decreases to the point where the
oligamers can diffuse out of the polymer bulk and be utilized by microorganisms.
The pathways for chemical and biological cycling of polylactide are shown in
Figure 3-2.

 

 

glucose
n C H,2O ‘5 lactic acid

V fermentation 3’7 C3H so:
enzymatic

starch hydrolysis _..>2n H20
( Col-1.005),. cyclization
5" 02 Biological l (1
photos nthesis Ch I acti e
5n H,O-_> y an.“ n 06H ,0,
carbon dioxide
pol erizaa'on
6n CO2 yrn
microbial
metabolism

hydrolysis POIYIactldO

(CSHOO2)n
\ 2n H20

 

n H20

   
   
 
 
 
 

 

   
    
  

 

   
  
  
 

 

   

 

6n H20

     

lactic acid
2n 03H603

  

6n 0,

  

 

 

 

Figure 3-2. The polylactide cycle (adapted from Ozeki and Ohara, 1993).

17
3.2.1 Hydrolytic degradation

Polylactide chain scission is autocatalyzed by the carboxy endgroups as described
by the following rate equation (Schindler et al., 1977; Pitt et al., 1981; Zhu et al.,
1991)

d [COOH] _
--——--dt — k [COOH] [H20] [ester] (3_1)

where [COOH] is the concentration of carboxy end groups in the polymer, [H20]
is the concentration of water in the polymer, and [ester] is the concentration of the
ester groups in the polymer. The carboxy group concentration can be related to
ester concentration, degree of polymerization (DP), weight, volume, and density of

the polymer sample (W,V,p), and the number average molecular weight (Mn):

 

 

 

 

_ [ester]
[COOH] — DP- 1 (3 2)
W P
[COOH] — an — Mn (3-3)
Using a lumped rate constant, k’:
. _ k [Hzor p
k — Mn (3-4)
the rate equation for the change in degree of polymerization is:
d(1/DP) ,DP-l
dt ‘ k or2 (3'5)
Integrating this equation gives:
1- DP
1n l-DPo — -k t (3-6)

with DP and DP() being the degree of polymerization at time t and t = 0. For DP
>> 1 this expression becomes:

DP Mn
DPo = In Me = -k t (3-7)

 

 

In

18
Equation 3-7 applies until the molecular weight is reduced to around Mn = 10,000
when the diffusivity of the oligomer through the polymer matrix is high enough for

weight loss to begin.

Figure 3—3 shows a proposed schematic of polylactide degradation via a two-step
mechanism. Initially molecular weight reduction occurs by non-enzymatic
hydrolysis without significant diffusion of oligomer out of the polymer bulk. As
the polylactide molecular weight decreases, oligomers are increasingly able to

diffuse out of the polymer bulk and be utilized by microorganisms.

In massive polylactide objects a concentration gradient of carboxy groups develops
with the highest concentration at the center. This leads to more rapid weight loss
from the center, leaving a hollow core (Li et al., 1990a). In thin films, weight loss
occurs at a nearly constant rate through the film, leaving a porous structure.
Crystalline regions are inaccesible to water and oligomer diffusion, so are
degraded from the edges. Higher chain mobility in amorphous regions is believed
to reduce the energy required to achieve the transition state of the hydrolysis
reaction. Since hydrolysis occurs at a higher rate in amorphous regions, the

overall crystallinity initially increases as degradation begins (Li et al., 1990b).

The rate of hydrolysis of polylactide and other poly(ct-hydroxy acids) increases
with temperature and shows an increase in rate at the glass transition temperature
(Reed and Gilding, 1980). The sources in Table 3-1 provided data for polylactide
molecular weight loss with time at constant temperatures under the conditions
listed. Rate constants at the various temperatures were calculated as shown in

Figure 3-4 using these data and the rearranged form of Equation 3-7:

 

l9

 

STEP 1

 

= 1 00,
00° Mn >1o,ooo
molecular weight
reduction by 0
non-enzymatic
hydrolysis ’
'7.
.t. j.
polymer matrix
compost matrix No significant diffusion of oligomers
or lactic acid through polymer
n> 10 000 Mn < 10,000
further molecular weight loss r
allows lactic acid to
d: diffuse out of polymer c I
L. j
e
oligomer diffusion

01 mar matrix 0 out of Imer
compost matrix . 0

002 + H20 I microbial

 

utilization
_ . polylactide with
carboxy endgrcup
. lactic acid or low
MW oligomer

 

 

 

 

Figure 3-3. Proposed degradation mechanism for polylactide.

 

20

 

4.7
4.6
4.5 ~
4.4 .
4.3 «
4.2 -
4.1 ~
4 .-
3.9 J. :
o 50 100 150

    

lnMn=lnMn.o-k't
k‘ = slope = 0.0063/day

ln(Mn)

I

I

 

 

 

Time (days)

Figure 3-4. Example of rate constant calculation using data from
Reed and Gilding, 1980.

T (°K)"1
-0.0034 -0.0033 -0.0032 -0.0031 -0.0030 -0.0029
1 r r 1

l

 

 

 

 

 

Figure 3-5. Arrhenius plot for polylactide degradation rate constant.

21

In Mn = 1n Mm -k' t (3-8)
The resulting rate constants were used to generate an Arrhenius type rate
expression for the rate constant k’ using a least squares regression to the ln(k) vs.

(l/T) data, as shown in Figure 3-5. The expression for the rate constant:

-14543
T

could then be used in Equation 3-9 to calculate the rate of hydrolysis at

 

-E
k, = a ' “NR—1'1.) = 3.664 x10” -exp( (3-9)

temperatures near or below Tg and to predict the rate of decrease in number

average molecular weight:

-14543
T

 

1n Mn = lnMn‘o —3.664x10'8 -exp( )-t (3-10)

The activation energy for hydrolytic molecular weight reduction, E3, was
calculated at 29.8 kcal/mol. Fitting a logarithmic curve to k vs. T data yielded an

expression equivalent to that found using the Arrhenius plot.

Table 3-1. Rate data used in the Arrhenius plot.

 

 

 

 

Label Source PLA Type Conditions of Experiment
(1) Li et a1, 1980 semi-crystalline PLLA 37°C in buffered saline, pH 7.4
(2) Schindler et al., 1977 amorphous P-D.L-LA 30°C and 60°C in water

(3) Nakamura etal., 1989 crystalline PLLA in vivo (rabbit, 37°C)

(4) Reed and Guilding, 1980 37% crystalline PLLA 37°C in buffer. pH 7.0

(5) Mason et al., 1980 amorphous P-DL-LA 25°C and 50°C in deionized water
(6) Mason et al., 1980 amorphous P-Dr-LA 37°C in pH 7 buffer

(7) Mason et al., 1980 amorphous P-D.L-LA 25°C and 37°C in porcine plasma

 

 

 

 

22

The ratio Mw/Mn gives a measure of the polydispersity of the polylactide. A wider
distribution of molecular weights gives a larger polydispersity since the
contribution of each molecule to the number average molecular weight, Mn, is
proportional to its mass, and its contribution to the weight average molecular

weight, MW, is proportional to the square of its mass as shown in Equations 3-11

\
and 3'12. \ Mi, Ni)
.1") N1 (d °°
M \ \. ZNi Mi
" v '1, , M :: —-—-—l=l 3'11
\}\.J) .- ' f) I] II n °° ( )
\ \J\f "‘4 (T's 1:1

M =—=‘—— (3-12)

where N, is the number of molecules of type i and M, is the molecular weight of
molecule type i. Since the hydrolysis of polylactide occurs randomly, longer
polymer chains are more susceptible to cleavage than short chains. This will result

in a narrowing of the distribution of molecular weights as degradation proceeds.

3.2.2 Enzymatic degradation

Several enzymes, including proteinase K, pronase, and bromelain, have been
shown to bring about hydrolysis of polylactide in vitro (Williams, 1981).
However, enzymes are large molecules and are unable to diffuse through
crystalline regions. Little enzymatic degradation occurs during the early stages of
degradation when the polymer crystallinity is high (Holland et al., 1986). In the
later stages enzymatic involvement can become more pronounced as pores
develop, fragmentation occurs, and more polymer regions are accessible to

enzymes.

23
3.3 LIFE CYCLE ASSESSMENT

The raw materials for lactic acid production are annually renewable agricultural
products that are produced in surplus. The lactic acid fermentation itself makes
especially efficient use of the feedstock (glucose) carbon. Unlike the
fermentations used to produce other agriculturally-derived products, such as
ethanol, the lactic acid fermentation does not result in a loss of carbon as carbon

dioxide. This results in a larger yield of product per unit of feedstock consumed.

Polylactide is a thermoplastic polyester that can be processed by melt flow or by
solution processing. The price of polylactide is decreasing as production scales
increase and processing technologies improve. Resin mechanical properties and
shelf life under moderate temperature storage make it applicable for a number of

single-use disposable products.

The mechanism of polylactide biodegradation makes it especially suited to
disposal in composting infrastructures. High compost temperatures and moisture
aid in rapid hydrolysis, and compost microorganisms use the lactic acid as an
energy source. The heat generated accelerates decomposition of other compost
substrates, and biomass sustained by the lactic acid eventually becomes a part of

the compost product.

3.4 APPLICATIONS

Implantable Prosthetic Devices

The potential of lactic acid polymers as a material for resorbable prosthetic devices
was first investigated by the US. Army (Kulkarni et al., 1966). Massive

implantable devices of poly(lactic acid) provide temporary structural support or

24
fixation until natural healing occurs and do not require a second surgical procedure
for device removal. Extensive in vivo testing in laboratory animals has been
necessary to precisely balance the rate of erosion with the rate of tissue

regeneration, and such devices are only now entering clinical trials.

Resorbable Sutures

Lactic acid polymers found their first commercial success as a replacement for
collagen fiber in resorbable sutures (Cutright et al., 1971). Polymers of lactic acid
and glycolic acid are now widely used to produce bioresorbable sutures that are
sold under the trade names Dexon, Medifit, and Vicryl. These sutures have a
predictable rate of mechanical property reduction and are resorbed by the body
within 2 or 3 months (Craig etal., 1975).

Controlled-Release Matrices

Polymers of lactic acid were among the first to be applied in the pharmaceutical
industry as matrices for the controlled-release of drug products (Yolles, 1971).
Microspheres, capsules, and pellets are designed so that zero-order drug release is
achieved by balancing the rate of diffusive release with the rate of release by
polymer erosion. Similarly, pesticides and other agricultural chemicals could be
incorporated into poly(lactic acid) matrices for slow release action (Sinclair,
1973). Incorporation into a polymer matrix has been proposed as a way to provide

safer handling of pesticides and reduce chemical runoff.

Biodegradable Packaging
Surgical and pharmaceutical applications for poly(lactic acid) are low volume
markets able to tolerate high resin costs. In contrast, penetration of the 50 billion

pound per year markets for single-use disposable packaging depends on low resin

25
price. Considerable research and development effort is currently aimed at
commercial production of lactic acid polymers (Narayan, 1993b). Argonne
National Laboratories is developing a technology for polymerizing lactic acid
produced by fermentation of potato waste. Low molecular weight poly(lactic
acids) are synthesized by condensation polymerization and spliced together using
coupling agents. Batelle and Golden Technologies are jointly developing lactic
acid polymer technology for packaging applications. Cargill and Ecochem (a joint
venture of DuPont and ConAgra) are producing lactic acid from corn and cheese
whey. This is followed by dimerization to the lactide and ring-opening
polymerization to high molecular weight polylactide polymers suitable for
packaging, as described below. Cargill brought a 10 million pound per year pilot

lactic acid production facility on-line in early spring of 1994.

3.5 POLYLACTIDE FILM PROPERTIES

The polylactide film used in this work was produced by Cargill Inc., Minneapolis,
MN (Gruber et al., 1993). Table 3-2 gives some physical and chemical

characteristics of this film.

TABLE 3-2. Characteristics of polylactide film.

 

 

Film Thickness

Appearance

Elemental Analysis:
Carbon
Hydrogen
Nitrogen
Oxygen
Phosphorus
Sulfur

Mn

Mw

Mw/Mn

Degree of crystallinity

T g

Tm

 

2.91 mil

transparent

49.8%
5.8%
0.004%
44.8%
0.016%
< 10 ppm
95,971
215,034
2.24
43%
57°C
138°C

 

 

CHAPTER 4

 

ANALYTICAL METHODS

 

Methods that were used in the studies of polylactide biodegradability are outlined
in this chapter. The weights and volumes used for characterization of test
materials and feedstock materials that were more homogeneous were sometimes

lower than those listed for compost analysis.

Weight Determination

Three balances were used to measure sample mass. Large samples were weighed
using an Ohaus I-10 platform balance, which had a range from 0.00 to 50.00 kg.
The weight of medium size samples, 0.000 to 2000.00 g, were measured using a
Fisher Scientific XD-2KD balance. A Perkin-Elmer AD-4 Autobalance was used

to measure the mass of samples in the 0.000 to 20.000 mg range.

Measurement of pH
A slurry of 10 g wet weight compost was made by mixing with 50 ml deO for 30

seconds. Slurries were prepared in triplicate and pH was measured using an

Omega PHB-70X water analyzer.

Moisture Content
Triplicate samples of approx. 150 g w.w. compost were dried to constant weight in

a vacuum oven at 105°C and 15 in. Hg. Weights of samples are referred to on the

27

28

basis of their wet weight (w.w.) or dry weight (d.w.), and the moisture content is

calculated as the percent of the wet weight removed by drying.

Milling
Dried composts were milled in a Fitzpatrick Model JT-6 homoloid mill to pass

through a 25 mesh (0.7 mm) screen. This material was used for determining ash

content, for elemental analyses, and for humic substance fractionation.

Ash Content
Approximately 2 g d.w. of dried, milled compost was ignited at 550°C for 2 h to
determine the ash content of the composts. Ash content was reported as percent of

dry weight remaining after ignition.

Carbon and Nitrogen Contents (CHN)
Samples of 5-10 mg of dried, milled compost were analyzed for carbon and
nitrogen content using a Perkin-Elmer 2400 CHN Elemental Analyzer.

Humic Substances Analvsis

The degree of humification of compost samples was determined using the method
of Ciavatta et a1. (1990). Dried, milled compost samples (0.8 g) were mixed with
40 ml of a solution of 0.1 M NaszO7 and 0.1 M NaOH and shaken at room
temperature for 2-3 hours. After centrifuging 20 min. at 13,000 x g the
supernatant was removed and filtered through a 0.45 pm syringe filter and referred
to as total extract (TE). Humic acids (HA) were then precipitated by adding 6N
H2804 to 20 ml of TE to acidify it to pH 2.0, and humic acids were collected by
centrifuging 20 min. at 13,000 x g and redissolving the pellet in 0.1 N NaOH. The
20 ml of supernatant from this centrifugation was loaded onto 2 g of solid
polyvinylpyrrolidone (PVP) equilibrated with 0.005 N H2804. The PVP was

29

washed with an additional 20 ml of 0.005 N H2804, and the two non-retained
fraction were combined and referred to as the non-humified (NH) fraction. The
retained fulvic acids (FA) were eluted from the PVP using 40 ml of 0.1 N NaOH.
The carbon contents of the TB, HA, FA, and NH fractions were determined by

dichromate oxidation using 0—150 ppm Hach COD vials.

Cress Gemrip;a:tion Bioassay

Compost maturity and phytotoxicity were detemrined using the cress germination
bioassay based on the method of Zucconi et al. (1981). Compost extracts were
prepared by adding 100 g D.W. of moist compost to distilled water to a final
moisture content of 1000 g and mixing in a blender on lowest speed for 30
seconds. The slurry was then strained through a 1 mm mesh and centrifuged for
30 min. at 1000 x g to remove solids. The compost extract was used as a
germination medium for curly cress (L. sativum). Eight replicate 15 x 95 mm petri
dishes were prepared with 2 ml of compost extract and 10 cress seeds on a 9.0 cm
Whatrnan #1 filter paper. The dishes were incubated at 28°C in the dark for 48
hours when the seeds were removed and the number of seeds germinated and their
weight were determined and compared to the germination of seeds in a control of

water only.

Gas Chromatoggaphy (GC)

Samples of compost vessel gases were collected in a syringe, and composition was
determined using a Gow-Mac 580 Gas Chromatograph. Separation of 200 rtl
sample was performed on a CarboSphere column at 100°C and detected at 150°C

with helium as the carrier gas at 30 ml/min.

30

Gas Analyzer Calibration

The Ultramat 22F and Ultramat 21P carbon dioxide analyzers and the Model 755A
oxygen analyzer were calibrated using four cylinder gases from AGA Industries.
Medical air and nitrogen were used to span and zero the oxygen analyzer. The
carbon dioxide analyzers were zeroed using nitrogen and spanned using certified
3.85% carbon dioxide/balance nitrogen for the 0-4% range channels, or using

100% carbon dioxide for the 0-100% range channels.

Differential Scanning Calorimetry (DSC)
Polylactide glass transition temperature and melting temperature were measured
using a Perkin-Elmer DSC7 differential scanning calorimeter at a temperature

ramp of 10°C/min.

Gel Permeation Chromatography (GPC)

A solution of 0.5% polylactide in chloroform was separated at 1 ml/min on a
Waters K-800 Ultrastyragel column at 35°C. A Waters model 410 differential
refractometer was used to detect polylactide and results were reported based on

calibration with polystyrene standards.

CHAPTER 5

 

BIODEGRADATION UNDER
EXTERNALLY CONTROLLED
COMPOSTING CONDITIONS

 

5.1 OVERVIEW

As discussed in Section 2.2, the inherent biodegradability of new materials must
be established if they are destined for disposal in composting infrastructures. This
chapter presents the development of an apparatus for measuring the
biodegradability of materials under controlled composting conditions and its

application in measuring the biodegradability of polylactide film.

5.2 TEST METHOD

Evaluating the biodegradability of any material requires some parameter or set of
parameters that can be measured and related to the rate and extent of
biodegradation of the material being tested. Under aerobic conditions,

biodegradation of an organic substrate can be expressed as:

C(substrate) + 02 """" > CO2 + H20 + C(cell biomass) (5-1)

If the experiment to measure biodegradability continues over many cell

generations, biomass is metabolized by predator species:

31

32

C(cell biomass) + 02 -------- > C02 + H20 + C(cell biomass) (5-2)

and as the substrate is depleted and predation continues, the carbon originally

present as substrate will be largely mineralized to carbon dioxide.

This view of the microbial processes occuning during biodegradation has been
used to develop test methods which use the amount of carbon released as carbon
dioxide to approximate the amount of carbon consumed from the material being
tested. One of many such methods, American Society for Testing and Materials
(ASTM) method D 5338-92, has been designed specifically to be used for
laboratory measurement of the degree and rate of aerobic biodegradation of plastic
materials on exposure to a controlled composting environment. This method
involves exposing the test material to an inoculum of mature compost in a two-liter
vessel and either measuring the carbon dioxide evolved directly or by trapping
evolved carbon dioxide in an alkaline scrubbing solution and quantifying by

titration.

The conditions of the experiment mimic the physical and biochemical environment
present during composting. By housing the two-liter vessels containing the test
material and compost matrix in a controlled temperature room, temperatures
characteristic of optimal composting can be maintained. During the 45 day
experiment the temperature is set at 35°C for the first day, 58°C for four days,
50°C until day 28, and at 35°C for the remaining 17 days. Although high
temperatures improve the kinetics of decomposition, excessive composting

temperatures kill or inactivate many microbial species and result in a less diverse

33
microbial population. At temperatures in excess of 55-60°C, fungi are strongly
inhibited, and bacteria and actinomycetes are reduced to a few therrnotolerent
species (McKinley and Vestal, 1984; Strohm, 1985). High temperatures also
result in conditions which favor the formation of ammonia and the volatilization of
other malodorous compounds. The temperature regimen enforced on the
composting mixture is representative of the temperatures occurring in full—scale
processes tuned for maximum decomposition and minimum odor generation

(MacGregor et al., 1981; Sikora and Sowers, 1985; Stentiford et al., 1985).

Aerobic decomposition results in less odor production and proceeds at a higher
rate than anaerobic metabolism. Since aerobic metabolism requires oxygen for
oxidative phosphorylation, maintaining a healthy population of aerobes depends
on providing adequate aeration. A higher concentration of oxygen in the
interstices of the compost matrix leads to a greater depth of oxygen penetration
into the biofilrn layer on the surface of a compost particle. Increasing the layer of
non-oxygen limited conditions maximizes substrate utilization (Suler and Finstein,
1977; Hamelers, 1993). The air flow rates through the test vessels are set so that
the oxygen content in the effluent gas remains above 6% by volume while the
carbon dioxide concennation stays at a concentration high enough to be accurately

detected or efficiently trapped in solution.

Moisture content is another parameter that is set to maximize microbial activity.
The optimum depends on the moisture holding capacity of the compost, which is a
function of the particle size distribution, porosity, water solubility, and structural
integrity of the compost particles. Excessively dry conditions limit microbial

growth and motility, but a water content higher than the moisture holding capacity

34
will cause clumping or will block interparticle spaces and hinder oxygen transfer.
Moisture levels below 30% and above 70% by weight severely limit aerobic
activity in composts. For most composts the optimal moisture content lies
between 50% and 60% (Jeris and Regan, 1973). The initial moisture content of
the test material and compost matrix is set in this range. Drying is prevented by

humidifying the air before it enters the test vessels.

Other chemical parameters are set within the optimal ranges for composting. The
initial hydrogen ion concentration of the test mixture is checked to make sure that
it falls between a pH of 7 and 8. The ratio of carbon to nitrogen in the initial

mixture is set between 25 and 40.

In addition to the sample mixtures containing the test material, three sets of vessels
provide control data during the experiment. These include a set of "blank" vessels
containing the compost matrix alone, a set containing polyethylene as a negative
control, and another using cellulose as a positive control. The amount of carbon
dioxide which can be attributed to the test material is calculated as the difference
between the amount of carbon dioxide evolved from the test material/compost
mixture and the amount of carbon dioxide evolved from the blank vessels
containing compost only. Biodegradability is expressed as the cumulative
percentage of the carbon from the test material which has been mineralized to
carbon dioxide. Experimental error is calculated as the standard deviation of the

biodegradability.

35
5.3 PRELIMINARY EXPERIMENT 1

The first controlled composting experiment evaluated the biodegradability of
paperboard used in frozen food packaging. An acrylic copolymer coating on this
paperboard provides the moisture resistance necessary in frozen food packaging
applications. If the acrylic coating is sufficiently biodegradable it will be
consumed during composting and will allow normal biodegradation of the

paperboard component.

5.3.1 Method
The experimental set-up was based on the ASTM D 5338 procedure described

above. Each two-liter test vessel contained 180 g dry weight of compost and 30 g
dry weight of the test material, adjusted to an overall water content of 50% by
weight. Polyethylene and cellophane films 25 pm (1 mil) thick were cut into 6
cm x 6 cm squares, and paperboards 300 pm (12 mil) thick were cut into 3 cm x 3
cm squares. Compost used in the experiment resulted from a mixture of restaurant
waste (55%), leaves (10%), turkey manure (10%), pine chips (3%), and onion tops
(22%) windrowed for 10 weeks. The finished product had the following
characteristics: carbon content, 22.6%; nitrogen content, 1.92%; pH of 7.5; and
volatile solids, 63%. This compost formulation provided an appropriate nutrient
balance for test material exposure and did not result in a background carbon

dioxide evolution rate which would obscure that from the test material.

An incubator room, maintained at 37° C, housed the test apparatus (Figure 5-1).
Air from an aquarium pump bubbled through two 20-liter preconditioning vessels
in series for carbon dioxide removal and humidification before passing to the 15

sample vessels. The first preconditioner contained 10 N NaOH solution and the

36
second contained water so that carbon dioxide-free, humidified air entered the
sample vessels. A manually adjusted needle valve at the exit of the sample vessel
controlled the air flow rate. After passing through the compost and test material
mixture, the air bubbled through two 50 ml trapping tubes containing 30 ml each
of 5 N NaOH hydroxide. Periodic titration of the trapping solutions provided data

for calculating the rates of carbon dioxide evolution.

............................................................................................................................
..v.- - noon-f --------------------------------------------------------------------------------------
.-....,.'..'........'...-.....,.....-.;. ...............................................................................

needle valve

manifold

   
 
    

 

 

 

 

 

diaphragm
pump .
sample vessel
controlled
temperature
incubator "

..--u.....'......................_.’......._.;._.......... 4 ~ "-....*o‘...'...'-u ...........................................
.....................................

...........
.....................................................................................................................................................

Figure 5-1. Setup for externally controlled composting experiment 1.

,6
At high pH, carbon dioxide reacts with sodium hydroxide to form sodium
carbonate:

C02 + 2 NaOH ---> N32CO3 + H20 (53)

37
Addition of barium chloride to this mixture precipitates the carbonate as barium
carbonate. Titration of the initial and precipitated sodium hydroxide trapping
solutions with standardized 0.05 N HCl provides the concentrations of sodium
hydroxide in the trapping solution before and after carbon dioxide absorption. This
titration allows the calculation of the amount of carbon dioxide evolved from each

vessel and from the test materials individually.

5.3.2 Results

Difficulties in maintaining consistent air flow rates through the apparatus led to
excessive variation within triplicates and disruptions in the data collection. The

pump delivered low pressure air to the 15 vessels, and a small adjustment to one
vessel’s valve affected the air flow rates in all of the vessels. The long time
required to regain system pressure after any valve adjustment or loss of pressure
made flow rate adjustment a painstakingly slow procedure. However, the
tendency for a leak in any vessel to stop air flow through all of the trapping

solutions presented the largest problem.

The results of this preliminary experiment (Figure 5-2) show that cellophane and
both paperboards biodegraded significantly in one month, but the high variation
within triplicates indicates the limitations of the data (Table 5-1). Sample vessels
containing polyethylene film consistently produced less carbon dioxide than the
blanks. The flexible squares of polyethylene film tended to envelope portions of
the compost matrix, which may have created anaerobic pockets and reduced
microbial activity, while the rigidity of the cellophane and paperboards did not

cause this behavior.

38
The constant temperature of 37°C provided less than optimal conditions for
biodegradation. Exposure to the typical pattern of temperatures found in
composting gives rise to periods in which different sets of microorganisms
dominate. Furthermore, higher temperatures increase the rates of hydrolysis and

other chemical and enzymatic reactions occurring.

TABLE 5-1. Results of Experiment 1.

 

 

 

 

%C in sample mineralized to carbon dioxide
TEST MATERIAL MEAN Standard deviation
Cellophane 44.2 1 1.9
Uncoated paperboard 36.3 17.7
Acrylic coated paperboard 33.0 11.4
Polyethylene -2.5 3.9

 

 

 

A separate study investigated the biodegradability of the acrylic copolymer coating
material in its liquid form (David, 1994). This experiment, based on the method
ASTM D 5209-91, determined the aerobic biodegradation of plastic materials in
the presence of municipal sewage sludge. Instead of using compost as a test
matrix and inoculum source, this method uses the return activated sludge from a

municipal wastewater treatment plant.

The ASTM D 5209-91 apparatus was nearly identical to the composting apparatus
described previously, using a higher pressure air source and needle valves placed
at the inlet to the sample vessels. This alleviated the air flow problems and

resulted in much smoother rate data. Results from the polymer and cellulose after

 

39
38 days of exposure in duplicate are included in Table 5-2. Biodegradation of the
acrylic coating material reached only 18% while cellulose achieved 85%,
indicating that the acrylic polymer is either only partially biodegradable or

biodegrades slowly under these conditions.

TABLE 5-2. Results of Aqueous Biodegradability Testing (David, 1994).

 

 

 

TEST MATERIAL %C mineralized to C02
Cellulose powder 85
Acrylic coating (liquid) 18

 

 

 

5.4 PRELIMINARY EXPERIMENT 2

5.4.1 Method

The second controlled composting experiment used an improved apparatus to
measure the biodegradability of six test materials and five control materials. These
included an MBI modified starch formulation (Amypol), polycaprolactone (PCL),
polyhydroxybutyrate—co—hydroxyvalerate (PHB/V), two polyvinyl alcohols
(PVOH), and a degradable film (Material A). Cellulose and starch powders were
the positive controls, and Kraft paper provided an additional positive control that
would give bulking effects similar to rigid films. Polyethylene film was used as
the negative control material along with an additional negative control
polypropylene film (Material B) that was used because it had the same
geometry as Material A. As in the previous experiment, 180 g dry weight of

compost provided the exposure matrix for 30 g dry weight of each test material.

40
Table 5-3 gives descriptions, sources, and geometries of the six test materials and

five control materials.

Several modifications to the test method and apparatus improved the quality of the
resulting data (Figure 5-3). A change from using an aquarium pump to using a
pressure-vacuum air handling system allowed much easier control of air flow

rates. The needle valve at the sample vessel outlet regulated the air flow rate as a

TABLE 5-3. Test Materials for Experiment 2.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MATERIAL DESCRIPTION SOURCE GEOMETRY
Amypol thermoplastic starch MBI 180-710 uM
Cellulose analytical grade Sigma 20 uM powder
Kraft paper 40# Kraft paper Tape. Inc. 5.6 mil sheet
PE transparent film Webster Industries 1.3 mil film
PCL Tone 767 Union Carbide 180-710 uM
PHB/V 12%HV Zeneca <425 uM
PVOH hot water soluble Air Products & Chemicals 2.2 mil film
PVOH c.w. cold water soluble Air Products & Chemicals 1.7 mil film
Starch hylon VII National Starch & Chemical 180-7 10 uM
Material A degradable film - 0.67 mil film
Material B polypropylene film - 0.67 mil film

 

 

..............................................................................................................................................................................................................................................
................................................................................................................................................................................................................................................
..............................................................................................................................................................................

    

house
compressed

    

 

 

 

 

 

 

 

 

 

 

house
--————— vacuum
sample vessel NaOI-I traps
controlled
temperature
incubator

...................................................................
.v. .-. . . ..-.-'. . .-.-. .-.-. . . . .-. . . .v.-.-. . . . . . . . . .;._...'..._.:...I.,. . 5...: . .........................
.....................................

Figure 5-3. Setup for externally controlled composting experiment 2.

vacuum pulled air through the trapping solutions. Air supplied to the sample
vessels sparged through two 20 liter bottles containing sodium hydroxide solution
and distilled water. An exhaust port in the second preconditioner ensured that air
was being delivered to the sample vessels at a low pressure and at a rate not

exceeding the vacuum take-off rate.

A method for automatic titration of the trapping solutions reduced the time
required for analysis and increased the accuracy of the titrations. Instead of
measuring the depletion of sodium hydroxide during the trapping, the new titration
measured the amount of carbon dioxide trapped. An Orion 960 Autochemistry
System was programmed to add 0.5 N H2804 to a sample of the trapping solution

42
until the pH reached 8.3, which takes the carbonate system to the endpoint of the

reaction:

0

CO3; + H+ ----> HCO3 (5-4)

Further addition of titrant until reaching the pH 4.5 endpoint of the reaction:

HCOg' + H+ ---—> H2C03 (5-5)

is used to calculate the amount of carbon dioxide trapped based on the volume of

titrant required to convert bicarbonate ion to carbonic acid.

Problems with liquid slugging up the narrow diameter trapping tubes at higher air
flow rates were resolved by switching from 50 ml tubes to 500 ml bottles.
Investigations into the efficiency of the new traps using 300 ml of 1 N NaOH
solution per bottle at an air flow rate of 30 mI/min showed that over 90% of the
carbon dioxide was trapped by a single bottle. The remainder of the detectable
carbon dioxide could be trapped using a second bottle in series. This trapping
method proved to be more reliable than the previous method and was adopted for
the new set-up. The air flow rate of 30 ml/min resulted in a maximum carbon

dioxide concentration of around 12% in the sample vessels.

A controlled composting temperature succession of 35-58-50-35°C, as described
earlier, was used in this experiment instead of the constant 37°C used for
Experiment 1. Although this succession gave a better approximation of the
temperatures in practical composting, the higher temperatures caused the tubing in

the trapping solutions to plug and also caused the compost mixtures to dry out.

43
After each bubble pushed out of the stainless steel tubing into the trapping solution
a small amount of the solution backed into the tubing. This continual rewetting of
the tubing wall formed a deposit of sodium hydroxide that eventually restricted or
even blocked air flow. At the higher temperatures this plugging occurred more

frequently than at the lower temperatures.

Evaporative cooling in the preconditioning bottles presented another temperature
related problem. The water temperature in the final humidification bottle
remained 5 to 10°C below the temperature of the air entering and leaving this
bottle, which was at the temperature of the incubator room. Drying of the
composting samples resulted, since the air passed to the sample vessels was
saturated with water at the lower temperature. Water added to the samples on

days 23 and 27 of the experiment made up for moisture lost by this drying.

5.4.2 Results

Biodegradabilties of the eleven test materials tested fell into two groups as shown
in Figure 5-4 and Table 5-4. Starch, cellulose, PCL, PHB/V, Kraft paper, and
Amypol gave 55-80% conversion to carbon dioxide while PVOH, Material A, and
the negative controls remained below 20%. The cold water soluble PVOH
triplicates were aborted after the first day because the film dissolved and filled the
air spaces in the compost matrix. Standard deviation from the mean of the
triplicates was less than 8% for all test materials, an improvement over the results
of Experiment 1. Additionally, the improved control of air flow rates through the
vessels resulted in the generation of meaningful carbon dioxide rate data as shown

in Figure 5-5.

44

 

90
3° ‘ .o—-"’°' """"" 0- ----- o
.z
a ’fi- ------ .
_ . .. I -" " ' T'
50 q 1'.-
“WWW/H

50 « mewwwro‘“

%C converted to carbon dioxide
8 8

B

 

 

 

 

 

 

M

10 . {M *3.

,. , -_:.; .Q’.*_..-.x—--—X' ....... x...-.'*

O -‘ . 1‘ ' I l r l 1 r

i 10 20 30 40 50 60

Time (days)
-10

--o—caurooo +STARCH ------ a ------ PCL --o---PHer -I — roan Paper

WAMYPOL -~D-~MatetldA +PVOH #3?me —*-PE

Figure 5-4. Biodegradability of materials tested in experiment 2.

%C converted to 002 per day

45

 

 

 

—0— Cellulose

- a - Kraft paper
—0— Material B
“re—- Polyethylene

I
‘g ‘
' ‘ _“ :5, -—fl-..__ ,

-MA'M.I.

 

'Il
\
1
\
1
II
| \
\
l \
l a)?” i\
' \
.3”
10 20

...~.—

30 4O 50 60
“are (days)

 

Figure 5-5. Rates of control material biodegradation measured

in experiment 2.

 

46

TABLE 5-4. Results of Experiment 2.

 

 

 

 

%C in sample mineralized to carbon dioxide
TEST MATERIAL MEAN Standard deviation
Cellulose 79.0 7.4
Starch 74.0 6.1
PCL 73.4 5.7
PHB/V 65.8 4.2
Kraft Paper 65.0 4.3
Amypol 56.8 6.5
Material A 18.5 4.9
Material B 11.9 6.7
PVOH 12.8 7.8
PE 4.2 2.2

 

 

 

 

The low conversion to carbon dioxide measured for PVOH and the data from the
negative control Material B highlight some of the physical effects of test materials
that either influence actual biodegradation or affect experimental results by
altering the rate of background carbon dioxide generation from the compost
matrix. During the initial four days at 58°C, the hot water soluble PVOH film
became gummy and formed aggregates with compost particles when the
temperature was lowered to 50°C. This may have lowered the overall rate of
microbial activity in the compost matrix, decreased the difference between the
blank triplicates and the PVOH sample triplicates, and resulted in an erroneously

low calculation of the extent of mineralization of PVOH.

Conversely, Material A and its negative control, Material B, were very thin films

that had a high bulk volume as used. When mixed with these materials the

47
compost matrix spread into thin layers on the film surface, which improved
aeration. This resulted in carbon dioxide evolution greater than the blank for both
Material A and its negative control. The calculation of test material mineralization
attributed the consequential increase in microbial activity in the compost matrix to
the test material itself. A similar effect was seen for polyethylene film, but this
was less dramatic since the thicker polyethylene film did not increase the bulk

volume to the same extent as the thinner films.

5.5 PRELIMINARY EXPERIMENT 3

5 .5.1 Method

A final overhaul of the controlled composting system design (Figure 5-6)
addressed the problems encountered with the previous set-ups and raised the
experiment’s reliability another level. The use of constant flow rate restrictors
(Mott Mettallurgical Model 25) provided the major improvement by setting a
stable rate of air delivery of through the system. Instead of using one set of air
preconditioning vessels for the entire system, one small humidification bottle
treated the air entering each sample vessel. Due to the increased surface area per
volume, the smaller bottles remained within 0.5°C of the room temperature and
minimized drying in the sample vessels. A manifold of Clippard ES Series
miniature three-way solenoid valves directed the air flow from one vessel at a time
through a carbon dioxide analyzer. The exhaust gas purged the sample loop and
Siemens Ultramat 22P infrared carbon dioxide analyzer for 15 minutes; after this
purge the carbon dioxide concentration was measured 30 times at one second
intervals and averaged. The Ultramat 22P has two channels for detecting low
(0.000 to 4.000%) and high (0.00 to 100.00%) carbon dioxide concentrations. A

Microsoft Quickbasic program running on an IBM personal computer was

48
interfaced with DuTec I/OPlexer input-output boards and facilitated data

acquisition and valve switching.

..............................................................................................................................................................................................................................
..................................................................................................................

 

(45 sample vessels solenoid

in parallel) valve
manrf old

    
 

 

 

 

restrictor

 

 

humidifier
controlled

temperature
incubator

 

computer I/O board

Figure 5-6. Setup for externally controlled composting experiment 3.

The gas sampling loop was designed to allow accurate and frequent measurement
of gas composition. A mathematical model of the analytical loop as a mixed unit
was used to predict the time required to replace the gas from one sampling with
the exhaust gas from the next sample vessel after switching. Sampling the effluent
air composition from each of the 45 sample vessels twice per day required that
each purge and measurement could be completed in 16 minutes. Setting the purge
completion criterion at 97% replacement after a step change resulted in accuracy

within 1% for a sample reading because of the small difference in carbon dioxide

tom
limi
con

the

Th

3?

cap

49
concentration between samples. The pressure of the in-house compressed air
limited the air flow rate to a maximum of 70 ml/min. The need to maintain a
concentration of carbon dioxide sufficient for accurate measurement also limited

the air flow rate through the vessels.

The response of the carbon dioxide concentration at the detector to a step change
in the input air composition was used to determine the volume of the loop, V, by
applying the equation

V = F(CO-C)/(dC/dt) (5-6)

In Equation 5-6 F is the air flow rate, C is the carbon dioxide concentration, and
Co is the carbon dioxide concentration at time t = 0. At an air flow rate of 20
ml/min the carbon dioxide concentration required 50 minutes to reach 97% of its
final concentration. This response revealed that the volume of the analytical loop
was approximately 300 ml. Using this value, the response of carbon dioxide

concentration at different air flow rates was predicted using the equation

C(th) = C(tn) + F(C0'C(tn))/V (5’7)

with C0,.) and C(tnrl) being the carbon dioxide concentrations during successive
time intervals. Figure 5-7 shows the time required for purge verses air flow rate.
From these results it was anticipated that an air flow rate of 60 ml/min would
provide a purge within 15 minutes; measurement of the system response at this air

flow rate verified the model’s prediction (Figure 5-8).

A short controlled composting experiment using 40# Kraft paper revealed the

capabilities of the new apparatus. Strips of paper approximately 0.5 cm x 4 cm

Time for 97% purge

%Replaced

50

 

 

 

 

50
40 a
30 a
20 a
10 . 1 . I
20 30 40 50 60 70

Air flow rate (ml/min)

Figure 5-7. Predicted effect of air flow rate on time
required to achieve 97 % purge of gas analysis loop.

 

 

 

 

 

100
80 -
60 -
4O -
20 _ . —model
a measured
0 I j I I
0 5 10 15 20 25

Time (minutes)

Figure 5-8. Replacement of gas in analytical loop
at 60 ml/min.

51
were prepared using an office paper shredder. The customary mixture of 30 g dry
weight test material and 180 g dry weight compost matrix was adjusted to the
moisture holding capacity (MHC) of the mixture. MHC was calculated as a
weighted average of the compost matrix MHC of 43% and the Kraft paper MHC
of 49%, each detennine by application of ASTM D 425. The mixture was
exposed to a temperature succession of 2 days at 35°C, 4 days at 58°C, and 11

days at 50°C at an air flow rate of 60 ml/min.

5.5.2 Results

Figure 5-9 compares the carbon dioxide concentrations from the vessels containing
Kraft paper in compost to the concentrations from the blank vessels containing
only the compost. Rate data observed over shorter time intervals provided insight
into the biodegradation process not available using the previous method. The small
peak in carbon dioxide generation for the Kraft paper mixtures during the first day,
when the paper still possessed some structural integrity, was most likely the result
of the bulking effect seen in previous experiments. A short burst of carbon dioxide
production occurred in both the test and blank vessels when the temperature was
raised to 58°C. This apparently resulted from the mesophilic populations which
built up during the 35°C phase. Before they were inactivated by the increased
temperature, the improved chemical and enzymatic kinetics caused a brief burst of
activity. Approximately one day later carbon dioxide evolution from the Kraft
paper peaked dramatically (Figure 5-10). Increased activity of thermophilic
cellulose-degrading species was seen before the temperature was reduced to 50°C
on the sixth day. Microbial activity decreased in all of the vessels when the

temperature was lowered.

52

 

 

 

 

5 » 10
Time (days)

20

Figure 5-9. Carbon dioxide concentration of the gas exiting each

test vessel.

53

 

12

CD
I

%C as 002 per day
a:

h
l

 

O

 

 

5‘ ’
f i'! :: as:
ates oa it
“I

I I

5 10
Tlme (days)

15

Figure 5-10. Rate of biodegradation of Kraft paper. (Error
bars show the standard deviation from the mean of triplicates)

 

20

Cumulative %0 as 002

54

 

70

60-

50-

40‘

30 ~ .5

20-

10-

 

 

 

O l I l
0 5 1 0 1 5
Time (days)

Figure 5-11. Biodegradability of Kraft paper in Experiment 3.
(Error bars show the standard deviation from the mean of
triplicates)

20

55
The percent of the carbon in the Kraft paper mineralized to carbon dioxide accrued
to nearly 60% over 17 days as shown in Figure 5-11. In the previous experiment,
Kraft paper achieved less than 45% cumulative conversion in the same period. The
increased rate of biodegradation was attributed to a combination of several factors
including the smaller size of the shredded paper pieces, the decrease in the
moisture content of the mixture to its MHC, and the higher oxygen concentration
in the compost interstices due to the higher air flow rate. Less information was
available from the cumulative conversion data than the rate data, so few
conclusions could be drawn from these data alone. This underscored the
importance of rate data as a source of insight into the behavior of test materials
under controlled composting conditions. If collected on a sufficiently frequent
basis, rate data also provide a quality control check and monitor of system

performance.

5.6 BIODEGRADABILITY OF POLYLACTIDE FILM

5.6.1 Method

The controlled composting apparatus developed and validated in the three
preliminary experiments was used to measure the biodegradability of polylactide
film. Unbleached virgin softwood Kraft paper 165 pm (6.5 mil) thick was the
positive control and transparent polyethylene film 100 pm (4 mil) thick was the
negative control. All of the test materials were cut into 2 cm x 2 cm inch squares,
a mixture was prepared using the standard 30 g d.w. test material and 180 g d.w.
compost inoculum, and the mixtures were adjusted to their moisture holding
capacities. The 70 day experiment used the 35-58-50-35°C temperature
succession with an air flow rate of 60 ml/min to each test vessel. Purging of the

sample loop for 17 minutes after switching effluent gas samples allowed each

56
bottle’s effluent to be measured twice per day. The vessels were mixed by shaking

three times per week.

5.6.2 Respirometric Biodegradability

Carbon dioxide concentrations in the test vessel effluents over the first 45 days are
shown in Figures 5-12 and 5-13. These figures show the carbon dioxide
concentrations measured from the triplicates for each test material, in the sequence
that they were measured. Except for three gas samples resulting from the Kraft
paper test vessels, all carbon dioxide concentrations were measured in the low
range (0.000 to 4.000%) cell of the carbon dioxide analyzer. The rates of carbon
dioxide generation displayed variability between the triplicate samples, but the rate
for each bottle remained relatively smooth (see Fig 5-12 inset). This demonstrated
that the main source of variability was the actual difference in the rate of carbon
dioxide generation between triplicate vessels and did not result from the carbon

dioxide measurement technique.

Figure 5-14 shows the rate and cumulative biodegradation of the polylactide film
and Kraft paper. Table 5-5 gives their 70 day cumulative biodegradabilities.
These values were calculated using the average of the triplicates containing the test
materials and the average of the blank triplicates. Carbon dioxide generation from
the polyethylene negative control was slightly above that of the blank vessels, as
observed in previous experiments. The Kraft paper rate of biodegradation peaked
during the 58°C period from days 2 to 5. During the following period at 50°C the
carbon dioxide generation rate decreased steadily. By the time of the decrease to
35°C the Kraft paper was asymptotically approaching complete biodegradation,

and the rate of mineralization was affected only slightly.

%002

%C02

57

 

 

Blank

 

 

 

 

 

 

 

 

f

fl

 

 

 

 

0 1O 20 30 40
Tlme (days)
5
Polylactide
4 .
3 4
2 ..
1 ..
O F I I I
O 10 20 30 40
Time (days)

Figure 5-12. Carbon dioxide concentrations in exhaust
gas from blank (top) and polylactide (bottom) test vessels.
(Inset: detail of the concentrations from the blank vessels)

96002

%002

58

 

 

 

 

 

 

 

 

 

 

5
Kraft paper
4 .
3 1
2 .
1 l
o I I I I
0 10 20 30 40
Time (days)
5
Polyethylene
4 ..
3 ..
2
1 ..
- w. """M AAAAAAA
o . . 2' " W“ . r
O 10 20 30 4O
Tlme (days)

Figure 5-13. Carbon dioxide concentrations in exhaust
gas from Kraft paper (top) and polyethylene (bottom) test
vessels.

Cumulative %0 as 002

59

 

10

94

0’
r

 

“sq:

%0 as 002 per day
a) at

 

-o-40# Kraft paper
—I— Polylactide film

   

 

""""""

 

 

 

 

 

 

I

lllllllllllllllllllll
rrrrrrrrrrrrrrrrrrr
rrrrr
iiiiiiiiii
a“
it“
rrrrrr
II
‘C
‘i
\t
t“
‘0
‘a
0

 

-0-40# Kraft paper
+ Polylactide film

   

 

 

I I I I r

10 20 30 4o 50 60
Time (days)

Figure 5-14. Rates and cumulative biodegradabilities
of polylactide film and Kraft paper.

70

60
Polylactide displayed a lag in biodegradation until day five, when the rate of
mineralization began to steadily increase to a maximum and steady rate of 2.8%
per day. The decrease in temperature on day 28 slowed the rate of mineralization

to less than 0.7% per day for the remainder of the experiment.

TABLE 5-5. Results of Biodegradability Experiment.

 

 

 

 

%C in sample mineralized to carbon dioxide
TEST MATERIAL Mean Standard deviation
Kraft Paper 91.3 5.0
Polylactide film 85.2 2.5
Polyethylene film 2.6 0.9

 

 

 

5.6.3 Polylactide degradation

Samples of polylactide film residue were collected on days 10, 25, and 45, and the
molecular weight distributions were determined by GPC at Cargill Inc.,
Minneapolis, MN. Approximately 100 mg of film fragments were collected from
the triplicate test vessels at each sampling time and air dried at room temperature.

Table 5-6 shows the results of the GPC analysis.

Using Equation 3-10 for molecular weight loss verses time:

-14543
T

and the temperature profile of the externally controlled composting experiment

 

InMn =lnMM —3.6o4><10‘8 -exp( )-t (3-10)

resulted in a prediction of the molecular weight loss of polylactide. Figure 5-15

 

61

 

 

 

 

 

 

100000 5
90000 -
---- Mn (model)
80000 ‘ I Mn (GPC) * 4
70000 - -O-°/oC/day
60000 ' . 3
5‘ 50000
40000 . l. 2
30000 ~
20000 J - - 1
. . . I
10000 , a .
I
o l I I I 0
0 10 20 30 40 50

Time (days)

Figure 5-15. Comparison of predicted and measured
decreases in M. with rate of mineralization during
controlled composting (ASTM D 5338).

lisp .Ied zoo se 3%

62
shows the model’s predictions for Mn over the course of the experiment compared
to the Mn measured by GPC and the rate of mineralization. Most of the molecular
weight reduction occurred during the 58°C period between the first and fifth days,
bringing the polylactide below Mn 10,000 soon after the tenth day. The
polydispersity decreased from 2.24 to 1.57 in the first 45 days, indicating that the
hydrolytic cleavage of the polylactide occurred randomly during composting.
Between days 5 and 15 the rate of carbon dioxide evolution from the polylactide
climbed to its maximum. The behavior in compost was consistent with the
biodegradation mechanisms described earlier, with the initial lag time required for
hydrolysis of the polymer to diffusable oligamers before the onset of mass loss

from the polymer and microbial utilization.

TABLE 5-6. GPC analysis of films from controlled composting.

 

 

 

 

 

 

 

Time (days) Mn MW Mw/Mn
0 96,000 215,000 2.24
10 12,500 25,000 2.03
25 8,300 15,600 1.89
45 5,600 8,800 1 .57

 

By the end of the 70-day composting experiment the polylactide film had been
reduced to tiny, opaque, brittle fragments of low molecular weight, and
biodegradation continued at a steady rate toward completion. The photographs in
Figure 5-16 illustrate the disappearance of the polylactide film by hydrolytic

degradation and microbial utilization.

63

 

Figure 5-16. Photographs of the mixture of polylactide film and
compost inoculum before (top) and after (bottom) externally
controlled composting.

CHAPTER 6

 

BIODEGRADATION UNDER
FEEDBACK CONTROLLED
COMPOSTING CONDITIONS

 

6.1 OVERVIEW

The experiment in Section 5.6 demonstrated the inherent biodegradability of the
polylactide film under externally controlled composting conditions. Although
inherent biodegradability is a necessary attribute of a compostable material, it is
not sufficient. As explained earlier, the material must not hinder compost
processing, reduce product quality, or end up in the reject fractions. This chapter
addresses these issues by determining the behavior of the polylactide film in a
pilot-scale feedback controlled compost system. Section 6.2 describes the system
design and implementation of optimal process control for the pilot-scale units.
The process dynamics and compost quality resulting from different compost
feedstocks are explored in Section 6.3 to establish normal composting behavior.
Finally, in Section 6.4 the fate and effects of the polylactide film in pilot-scale

composting are determined.

6.2 PILOT-SCALE COMPOSTING SYSTEM

The pilot-scale composting system (Figure 6-1) consisted of two identical 60-liter
vessels, each constructed from a 3 foot long section of 12-inch PVC pipe.
Thermal mass flow controllers (MKS Type 1159B) regulated the flow rates of

64

65

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

02 <
analyzer exhaust
flowmeter C02
anal r
um
condenser
air chiller
‘
exhaust
calibration ases
exhaust Emm
<
solenoid )l alves I 5:283:34
controlled 1 I
RTD temperature I 2
jacket : '
J... 0
60 L I
reactor .
w I
ers
RI‘D .
- air
hurnidification : L
bypass T '- - -
v water
humidification
column

 

Figure 6-1. Schematic of pilot-scale composting apparatus.

66
influent air, which bubbled through a humidification column before passing into
each composting vessel through its perforated base. After the air flowed upwards
through the composting mass, Asco 8320 Series 3-way solenoid valves directed
the effluent air either to exhaust or through a condenser and into the gas analysis
loop. Effluent gas composition was measured using a Beckman Model 755A
paramagnetic oxygen analyzer and Siemens Ultramat 21P infrared carbon dioxide
analyzer. Stainless steel RTD probes inserted into the test vessels measured the
temperatures approximately 4 inches into the top and bottom of the composting
mass. An electrically heated insulating jacket, which was maintained at the lower

of the temperatures in the composting mass, minimized conductive heat losses.

Humidification of the cooling air supply and condensation at the outlet prevented
drying of the compost during processing. Air entered the vessel saturated with
water at 25°C, or approximately 0.02 g water/g air. In reaching saturation at 55°C
(0.11 g water/g air), the air removed heat from the compost equivalent to the of
heat of vaporization of 0.09 g water/g air plus the sensible heat required to raise
the temperature of the moist air by 30°C. The 0.09 g water/g air carried out of the
compost condensed at the vessel outlet and returned to the compost, and periodic

mixing redistributed this moisture.

6.2.1 Process Control

DuTec l/OPlexer input-output boards facilitated communication between the
system instrumentation and FIX D/MACS personal computer-based data
acquisition and control software. Control algorithms implemented via this
software maintained set oxygen levels and temperatures in the compost by

manipulating the aeration rate. A dual-control PID algorithm was used to maintain

67
a minimum oxygen concentration of 15% in the effluent gas and a maximum

compost temperature of 55°C.

Heat and mass transfer limitations quickly result in temperature and oxygen
concentration gradients in a composting mass (Finger et al., 1976). Forced
aeration results in temperature and oxygen concentration gradients through the
compost pile with lower temperature and higher oxygen concentration where the
air enters the compost. Figure 6-2 shows how these temperature gradients develop
in full-scale aerated static pile composting (Stentiford et al., 1985). Similarly, the
pilot—scale composting vessels developed an axial temperature gradient from the
inlet to outlet of the composting vessels. The largest effect occurred during
periods of peak aeration demand, when temperatures in the vessel were as low as

30°C near the inlet and 55°C at the outlet. Since composting at the optimum

 

 

 

Figure 6-2. Typical temperature profile during
forced aeration composting (Stentiford et al., 1985).

cor
air
cm

110

68

conditions used in the pilot-scale apparatus requires as much as nine times more
air for heat removal than is necessary for maintaining an adequate oxygen
concentration (MacGregor et al., 1981) a steep oxygen concentration gradient was
not established. Air entered the vessel at approximately 21% oxygen and left at a
minimum of 15% oxygen during the compost warm-up phase. However, the
effluent oxygen concentration remained even higher than 15% during most of the

composting, since excess air was supplied based on the need for heat removal.

The computer control algorithms collected process data, performed valve
switching, manipulated the air flow rates to the vessels, and maintained the vessel
jacket temperatures. Every ten minutes the primary control algorithm (Figure 6-3)

used process data to calculate a new air flow rate for both vessels as follows:

1. Switch solenoid valves to direct flow from desired vessel through
analytical loops.

2. Wait 5 minutes to purge instruments.

3. Record carbon dioxide and oxygen concentrations.

4. Calculate PID output (pT) based on oxygen concentration setpoint.

5. Record compost temperatures.

6. Calculate PID output (p0) based on temperature setpoint.

7. Select minimum air flow rate or higher PID output as new air flow rate.

8. Repeat 1-7 for the other vessel.

Preliminary experiments detennined that, after switching flows in step 1 above, an

analytical loop purge time of 5 minutes was sufficient to accurately measure gas

69

 

 

 

 

 

 

 

 

 

 
 
  

 

 

 

 

 

 

 

 

 

T(t) 16mph)
selector temperature n———~
e detectors ‘__.

temperature min air
setpoint (T...) 1 flow rate

temperature Drill

PID controller__.
oxygen CONC- selector ——> process —---—>
setpoint (0..) 90(1) air flow

 

 

 

 

 

 

 

 

 

ox en conc.‘ controlled
I f ' Piggcontroller po(t) rate(t) variables (t)

00) oxygen
detector

 

 

 

 

 

 

 

 

Figure 6-3. Primary control algorithm for dual control of
temperature and oxygen concentration.

composition. The data acquisition program collected oxygen and carbon dioxide
concentrations every second and used a smoothed value in Step 3 of the primary
control algorithm. Temperatures were collected every 10 seconds and smoothed to
be used in Step 5 and in controlling the jacket temperatures. On-off control of the
heat element in the insulating jackets kept the exterior of the vessels at the lower

of the temperatures measured in the compost vessel.

The digital control algorithm shown in Figure 6-3 used proportional, integral, and
derivative (PID) feedback control to drive process variables to their setpoints. The
deviations er and co were calculated as the difference between the actual values of
temperature and oxygen concentration, T and O, and the setpoints, TR and OR. By
changing the controlled variable (air flow rate) in response to the deviations from
the setpoints, the PID blocks attempted to reduce the deviations to zero and

maintain desired process conditions. The PID blocks calculated the appropriate

70
control output signals, pT and pa, for temperature and oxygen concentration control
using the discrete form of the PID equation:
100

T.
Apn = E [en + %§CR + ‘A—:(Cn'Cn-l)] (6—])

where: Apn change in controller output at time n
en = process variable deviation at time n

PBc = proportional band

T1 = integral time constant
Td = derivative time constant
At = scan time of the digital PID block.

Proportional, integral, and derivative control actions are provided by the three
terms in brackets on the right hand side of Equation 6-1. The first term imparts a
change to the controller output that is proportional to the deviation from the
setpoint. The next term provides the integral action by bringing the controlled
variable back to the setpoint in the presence of a sustained deviation. The
derivative action is supplied by the third term, which uses the rate of change in the
error to make a predictive adjustment to the controlled variable. Equation 6-1
shows that, for a given deviation, small values of PBc and T1 and large values of TD

cause greater changes to the controlled variable.

6.2.2 Controller Tuning by the Continuous Cycling Method

Selecting values of the parameters PBC, T1, and To to give optimal control of
process variables is known as controller tuning. Several methods of controller

tuning are available ranging from trial-and-error to model-based controller design.

71
Controller tuning parameters used in the PID algorithms were selected based on a

combination of these methods.

These tuning techniques were developed for processes that are linear and
stationary and have dynamics that approximate the simple models these techniques
are based on. Biological processes are usually non—linear, that is, the process
dynamics change as the process values change. In addition, biological systems are
highly non-stationary, meaning that different responses result from the same
stimulus depending on the state of the system (Wilson, 1991). The following
discussion describes the selection of controller settings that were applied during

the various pilot-scale composting experiments.

Tuning parameters used in the preliminary pilot-scale composting experiments of
Section 6.3 were selected using the closed loop continuous cycling method
(Ziegler and Nichols, 1942). The continuous cycling method was performed by
tuning the controller for proportional action only, i.e. setting T1 at its maximum
value and TD at zero. By trial-and-error the lowest value of PBc was found that
resulted in a stable oscillation after a step change to the setpoint. At this ultimate
proportional band, PBcu , the time for a single cycle in the oscillation is known as
the ultimate period of the controller, P... The controller settings were then

calculated based on the values of PBcu and Pu as:

PBc = l .7-PBcu (6-2)
T; = 0.5-Pu (6-3)
TD = 0.125-Pu (6-4)

72
This tuning method generally provides a good set of starting values that must be

further adjusted by trial-and-error. Using the feedstock mixture described in

TABLE 6-1. Controller settings from the continuous cycling tuning method.

 

 

 

 

Oxygen Concentration Temperature
Parameter Range Z-N Setting Setting used Z-N Setting Setting used
PBc 1- 10,000 425 500 20 20
T. (min) 0-99 50 50 99 90
TD (min) 0-20 12.5 0 20 O

 

 

 

 

 

 

Section 6.3.1, the continuous cycling tuning method gave PBcu of 250 and P“ of
100 minutes for the oxygen concentration PID controller. Initial tuning of the
temperature PID controller led to the use of overly tight control parameters, but
subsequent tuning gave better values of PB,“ at 10 and Pu at 1200 minutes.
Adjustments to these resulting control parameter settings were made based on
subsequent trial-and-error tuning trials. Table 6-1 shows the best controller
settings found during the preliminary pilot-scale composting experiments in

Section 6.3, which were based on these adjusted Ziegler-Nichols settings.

6.2.3 Controller Tuning by a Process Model-Based Method

As described in Section 6.3, compost process dynamics varied depending on the
feedstock mixture. A second round of controller tuning was required before
proceeding from the preliminary composting trials to the pilot-scale tests of

polylactide biodegradation. However, due to the long period of oscillation for the

 

73
temperature controller, several days were required for each iteration by the closed
loop continuous cycling method. Model—based controller tuning methods, such as
the process reaction curve method, can give controller settings after a single trial
and are therefore desirable in tuning controllers for slow processes. A method for
extracting tuning parameters from the open loop response of the process based on
an empirical model (Ziegler and Nichols, 1942; Cohen and Coon, 1953) was
applied to find controller settings suited to the pilot-scale composting feedstock

mixture for the polylactide biodegradation experiments.

The process reaction curve method uses an empirical model of the process based
on the response, Ay, to a step change in the controller output, Ap. For many
processes the response of the system has the characteristics of the curve shown in
Figure 6-4. This type of response, first-order-plus time delay, can be modeled

using the transfer function:

K - exp(-6s)

G(s) = --—-— (6-5)
ts + l
where:
K = the process gain
I = dominant time constant
9 = time delay constant
0,; = 6 + At/2 (for digital controller)

The process gain is the ratio of the change in the steady-state value of the
controlled variable to the change in controller output, Ay/Ap. The time constants 9
and t can graphically extracted from the process reaction curve as shown in Figure
6-4. A set of correlations developed by Cohen and Coon (1953) can then be used

to find the controller settings PBC, TI, and TD :

 

output y(t)

AY

 

.................................................................................................................

   

\

process
reacflon

to step
change

in controller
output

 
 

 

 

 

 

final steady-state

inital steady-state

Figure 6-4. Characteristics of the process reaction curve.

PB

       

“—1)

100 =de 3 +4“:

l-d

D:

32+69, /1:
134-894/1

___i4__
“11+29d/r

)

(6-6)

(6-7)

(6-8)

Figure 6-5 shows how oxygen concentration responded to step changes in the air

flow rate.

increased and decreased air flow rate.

Process reaction curves were generated using step changes as both

These step changes were made at

conditions near the target concentration of 15% oxygen. Process characteristics

extracted graphically from the process reaction curves were similar for increased

and decreased air flow rates, as shown in Table 6-2.

%Oxygen

%Oxygen

 

75

 

 

 

 

 

 

 

14.5 I l I I I I I
-20 0 20 40 60 80 100 120
11me (minutes)

 

17

16.5 -

16‘

15.5 -

15-

 

14.5 -

 

 

 

 

1 4 I. . I . I I I
-20 0 20 4O 60 80
Time (minutes)

Figure 6-5. Response of oxygen concentration to a
step decrease (top) and increase (bottom) in the
air flow rate.

76

TABLE 6-2. Process characteristics for oxygen concentration

 

 

 

 

 

 

Parameter Decreased air flow rate Increased air flow rate
K 6.8 7.4
I (min) 50 30
9,, (min) 12 8.5

 

The process characteristics were used to calculate the controller settings based on
the Cohen-Coon correlations. These settings were developed to give a setpoint
overshoot with a decay ration of 1:4, meaning that the amplitude of the initial
setpoint overshoot is four times that of the next oscillation. By relaxing the
Cohen-Coon controller settings to the final settings shown in Table 6-3, tight
control of oxygen concentration was maintained with less oscillation of the

process variables.

TABLE 6-3. Cohen-Coon settings for oxygen concentration control.

 

 

 

Parameter Range Decreased air flow Increasing air flow Setting
PBc 1-10,000 117 150 300
T; (min) 0-99 27 18.8 80
TD (min) 0-20 4.2 2.9 0.3

 

 

 

 

 

 

Figure 6-6 shows how temperature responded to step changes in the air flow rate.
Again, process reaction curves were generated using step changes as both

increased and decreased air flow rate. Due to the long settling time of temperature

after a process perturbation, steady-states of temperature were extremely difficult
to maintain. Oxygen concentration reached a new steady-state approximately one
hour after a change in air flow rate, and in this time the system was relatively
unchanged. Temperature, on the other hand, required one to two days to reach a

new steady-state. Radical changes to the system could occur in the time for

77

temperature to settle at the new steady-state.

graphically from the process reaction curves revealed this non-stationary behavior,

as shown in Table 6-4.

Process characteristics extracted

TABLE 6-4. Cohen-Coon process characteristics for temperature.

 

 

 

 

 

Parameter Decreased air flow rate Increased air flow rate
K 0.19 0.54

1 (min) 900 1740

9d (min) 600 5

 

 

TABLE 6-5. Settings for temperature control.

 

 

 

 

 

 

 

Parameter Range Decreased air flow Increased air flow Setting
PBc l-10,000 8.41 0.12 30
T. (min) 099 1 178 12 99
TD (min) 0-20 194 1 .8 10

 

 

Temperature (°C)

Temperature (°C)

 

 

 

 

 

E\

 

 

 

25 I I I W I I
-10 0 10 20 30 4O 50 60
Time (hours)

 

 

 

 

 

 

 

 

 

55

-
d
.4

I I

O 5 10 1 5 20 25 30
Time (hours)

Figure 6-6. Response of temperature to a step
increase (top) and decrease (bottom) in the air
flow rate.

79

The process characteristics were then used to calculate the controller settings for
temperature control based on the Cohen-Coon correlations. Based on these
settings, the final values were selected as shown in Table 6-5. Controller setting
were selected primarily on the basis of the results from the response of the system
to the decreased air flow rate. The increased air flow rate response was
disregarded, since the step change drove the temperature out of the thermophilic
range and the process reaction curve did not exhibit a time delay. The settings
selected used primarily proportional and derivative (PD) controller action, which

is common in temperature control.

6.3 PRELIMINARY EVALUATION OF COMPOST F EEDSTOCKS

Municipal solid waste (MSW), sludge, and yard debris are common feedstocks for
composting. Pilot-scale composting of mixtures from each of these three
categories provided process data that was used to evaluate the processing

requirements, dynamics, and product quality using each of these feedstocks.

6.3.1 MSW Compost

Pretreated MSW from the Recomp of Minnesota composting plant in St. Cloud,
MN was selected as a representative MSW feedstock, collected and shipped to
Lansing, MI, and stored refrigerated at 2-4°C until used. This material had been
homogenized for three days in a rotating drum digester and passed through a 1.5 in
screen. At the Recomp facility this pretreatment was followed by 7 weeks of
composting in an aerated, agitated trench system. The raw compost was extremely
heterogeneous, and had a high level of plastic, glass, and metal contamination.

Analysis of this fraction by researchers at the Ohio State University revealed that

80
these inert materials accounted for 62% of the solids in the raw compost (lonatti et

al., 1993). Other characteristics of the feedstock are shown in Table 6-6.

TABLE 6-6. Analysis of MSW compost feedstock.

 

 

 

Moisture content (‘70) 47
Ash content (‘70) 33
pH 6.6
%C 38
%N 0.72
C/N ratio 53

 

 

The first batch of MSW compost to be tested performed as shown in Figure 67.
Within 15 h of loading the oxygen concentration dropped below 15%, and an air
flow rate increase followed. A local maximum in microbial activity occurred at 19
h, as evidenced by the carbon dioxide production rate (CPR) and inflection of the
compost temperature verses time. This occurred when the average temperature of

the compost was 35-40°C, the optimum temperature for many mesophilic species.

As the compost temperature increased, overall activity decreased temporarily until
the average temperature exceeded 50°C. At this time a surge of activity attributed
to thermophilic species brought about another increase in the air flow rate. During
the period of peak air flow rate the temperatures at the top and bottom of the

composting vessel began to drift apart. Operation was discontinued after 32 h

81

 

 

60
50 1
Top T (°C)
Bottom T (°C)
40 a
30 a
20 ' ygen in effluent (%)
... Cr '
10 - CPR
Ida °k DW
$90 Q Airflow rate (LPM) Jr
0 -——r- r i ' T r
0 5 1O 15 20 25 30

 

 

 

Figure 6-7. Process dynamics of MSW compost over first
32 hours of operation with no mixing.

Time (hours)

35

82
when the aeration demand exceeded the 10 L/min limit of the mass flow

controllers.

A second trial with the MSW compost successfully passed through periods of peak
activity after the mass flow controller limits were reset to 30 L/min. Results are
shown in Figure 6-8. Mixing after the first day prevented the distinction of
mesophilic and thermophilic maximums apparent in the first MSW trial. Tight
temperature control was possible because of the high level of inert material in the

compost. The energy balance for a compost can be written as:

(Heat removal) = (losses by conduction) + (losses to cooling air)+

+ (losses to evaporation) - (heat generation) (6-9)

Energy removal through the cooling air will make conductive losses negligible if
the system is well insulated, and if inert materials do not result in heat generation.
Increasing the air flow rate to an inert material results in the classical cooling
response that is a function of the heat capacity of the material and heat transfer
rate to the cooling air. An increase in air flow rate to an active biomass, however,
can have many effects. The rate of heat generation will increase if the air flow
rate change relieves an oxygen or temperature limited condition. If an increased
air flow rate takes the system away from optimal conditions the heat generation

rate will decrease.

Control was effective due to the high level of inerts, but controller settings were
extreme and resulted in instability. PBc and T1 were set at 10 and 50, respectively
for the MSW trials. This resulted in far too much integral action in the

temperature controller. Evidence of this instability showed up at around 3 days as

83

 

60 "——5 f fl

50a / TopT(C)

4° ‘ ' Bottom T (°C)

30- '

Oxygen in

20 - effluent (%)

 

10 ~
Air flow rate

(LPM)

0 l I I I I I I I I

0 1 2 3 4 5 6 7 8 9

 

 

 

 

18 - CPR
(gC/day-ngW)

    

%C as 002

 

 

 

 

 

 

4 5
Time (days)

Figure 6-8. Process dynamics for pilot-scale composting of MSW over
the first nine days.

84
integral windup. Further controller tuning led to the application of the more
relaxed tuning parameters in Table 6-1 for subsequent trials, giving less rigorous

but more robust control.

The loss of activity and subsequent resurrection on mixing at 3.2 days and 5 days
was due to failure of the moisture control mechanism described earlier. At high
flow rates the air entrained water as it condensed at the vessel outlet and prevented
it from returning to the compost. Drying was severe at the air inlet side of the

vessels, and water was added before mixing the compost.

6.3.2 Sludge Compost

A mixture of dewatered anaerobic digester sludge and shredded wood from the
water pollution control facility (WPCF) of Fairfield, CT was selected as a
representative sludge composting feedstock, collected and shipped to Lansing, MI,
and stored refrigerated at 2-4°C until used. At the Fairfield facility this sludge was
dewatered to 15-20% solids, aided by the addition of 10 1b/ton Percol 763
acrylamidezDMAEA copolymer. The dewatered sludge was then mixed with
shredded wood at a 1:4 ratio (sludgezwood, dry weight basis) and composted for 3
weeks in an aerated, agitated trench system. Chemical characteristics of the

feedstock are shown in Table 6-7.

Sludge compost is relatively homogeneous and free of plastic film, but there is
some concern regarding the possibility of high metal concentrations in municipal
sludge. A sample of finished compost product from the Fairfield WPCF was
analyzed by the Michigan State University Animal Health Diagnostics Laboratory

using Ion Coupled Plasma (ICP) Spectroscopy. The concentrations of 25 metals,

85

TABLE 6-7. Analysis of sludge compost feedstock.

 

Moisture content (‘72) 57.6
Ash content (%) 19.2
pH 7.3
%C 43.3
%N 1.56
C/N ratio 28

 

 

 

 

shown in Table 6-8, are reported in ppm. ICP spectroscopy did not detect any

metal concentrations in excess of the EPA 503 limits for compost.

Pilot—scale composting was operated for 19 days at an oxygen concentration
setpoint of 15% and with the temperature setpoint at 55°C from day 0-10, 50°C
from day 10-15, and 40°C from day 15-19. Results are shown in Figure 6-9.
Oxygen demand determined the air flow rate for the first 5 days until the compost
temperature reached 55°C. Temperature control was not as tight for pilot-scale
sludge composting as it was for MSW composting, but performed adequately. The
cooling air demand and carbon dioxide production rates for the sludge mixture
were lower and less variable than for MSW. Even though the MSW compost used
had a high level of inerts, the organic fraction consisted mostly of high energy
food waste and cellulosics. Fairfield WPCF sludge, on the other hand, had
already been through aerobic and anaerobic treatment steps. When composted at
the pilot-scale the sludge and shredded wood mixture gave a maximum carbon

dioxide production rate half that of the MSW mixture.

86

TABLE 6-8. Elemental analysis of Fairfield compost product.

 

 

 

 

 

 

Element Detection Limit Concentration stande EPA
(ppm) (ppm) deviation 503 limits

A1 5590 1 178

As 12.5 41
B 25.0 27.6 *

Ba 282 64

Ca 13533 1877

Cd 2.50 2.89 0.17 39
Co 3.79 0.37

Cr 40.2 5.2 1200
Cu 483 61 1500
Fe 12200 5336

Hg 50 17
K 4867 316

Mg 2603 439

Mn 584 84

Mo 5.48 0.23 18
Na 250

Ni 20. 1 0.9 420
P 6067 1070

Pb 121 16 300
S 5687 524

Sb 25

Se 100

T1 62.5
V 8.13 2.75

Zn 1 123 162 2800

 

 

*only one of the triplicate samples was above the DL for boron

 

87

 

60 O... .0000 6.0.0

 

f“ TopT (°C)
50 o Mixing

 

 

 

/\r/\/\/

40

l

Bottom T (°C) 77‘"

7

Oxygen in effluent (°/o)

'MFJMNM\W\ rV\L\.

Airflow rate (LPM)

20-

 

 

10

 

 

 

 

 

 

 

 

  
    
 

 

 

 

15 20
30 0000 00000 00000
25 - oMixing
20 " 790 as 002
15 -
10 .. CPR
(go/day ngW)
0 ~1.:‘/
0 5 20
Time (days)

Figure 6-9. Process dynamics for pilot-scale composting of
sludge-wood chips mixture.

88

6.3.3 Yard Debris Compost

A mixture of yard debris, obtained from the Michigan State University grounds
department, was selected as a representative yard debris composting feedstock.
Leaves, grass, and shredded wood were mixed at a ratio of 3:1:1. Water was
added to raise the moisture content of the mixture to 60%. Chemical
characteristics of the yard debris components and the mixture are shown in

Table 6-9.

TABLE 6-9. Analysis of yard debris materials.

 

 

 

 

 

 

 

LEAVES GRASS WOOD MIXTURE
Moisture (%) 52.8 81.1 52.9 65
pH 6.6 8.0 8.3 8.4
%Carbon 43.0 44.4 48.5 44.4
%Nitrogen 1.16 3.92 0.14 1.51
C/N ratio 37.1 11.3 >100 29.4

 

Several short pilot-scale composting trials were performed using this yard debris
formulation. The resulting process data was used in fine tuning PID parameters
suitable for pilot-scale composting of this mixture. Typical behavior of yard
debris compost is shown in Figure 6—10. As discussed earlier, control of oxygen
concentration at the setpoint was easily achieved since the response of oxygen
concentration to a change in air flow rate was relatively predictable and occurred
rapidly in this system. On the other hand, temperature response was non-

stationary and had a natural period on the order of one day, making tight control

 

89

 

 

 

 

 

4—

 

 

 

 

 

 

 

 

 

60 ‘ \ Top T (00) r\
50 - \
40 d V\, BottomT (°C) L
30 -
20 M Oxvoen in effluent (%) ,fi __
W LA
10 -
Ml Air flow rate (LPM) _.\
\ f g_ _,
o . , 74L .
0 2 4 6 8 to
30 . .
25-
20 _ CPR (gC/dayngW)
%C as 002
15 -
10 T j
5 ..
0 , . . .
o 2 4 5 8 10
Time (days)

Figure 6-10. Typical process dynamics

for pilot-scale

yard debris composting trials.

90
very difficult. These process characteristics require that PID tuning parameters
reflect a compromise between the desire to keep the temperature from exceeding
the setpoint and the desire to prevent overcooling. Based on the results of several
such trials using yard debris, a process control strategy using relaxed temperature

control parameters was selected for subsequent pilot-scale composting.

Composting of a mixture of yard debris and 40# Kraft paper was performed in the
pilot-scale system using the process control strategy developed for yard debris
composting and the PID tuning parameters in Table 6-3 and Table 6-5. The
experiment used a mixture of 40% leaves, 30% grass, 20% wood mulch, and 10%
Kraft paper adjusted to a moisture content of 60%. Initial characteristics of this

mixture are shown in Table 6-10.

Results of the 38-day composting experiment are shown in Figure 6-11. A period
of peak activity occurred during the initial week, when carbon dioxide generation
rates were high and air flow rate was controlled based on compost temperature.
Over the first four days of the experiment the air flow rate exceeded 10 L/min, and
a steep temperature gradient developed in the composting vessel. Water was
added during mixing on day 4, since the overall moisture content had fallen to
37%. From day 7 to 20 the compost oxygen demand detennined the air flow rate.
After this period the minimum air flow rate of 1 L/min was sufficient for
oxygenation. The compost had an overall moisture content of 75% on day 26. In
an attempt to lower the moisture content the humidification column was bypassed
and the air flow rate was increased to 2 L/min. However, the compost was at

ambient temperature by this time and very little evaporative drying followed.

91

70-

 

Top T (°C)

   
  
 

60-

50~

 

40 omT (°C)

 

30

Oxygen in effluent (%)

 

20

10

 

 

 

 

 

 

 

 

 

Air flow rate (LPM)
0 I I g —_ I r I I
0 10 20 30 40
7o - *_
60 a
%C IOC'
50 -
40
l
l
30 l.
20 . CPR (gC/day-ngW)
10
o . NF\L . - - :-—
0 10 20 30 40
Tlme (days)

Figure 6-11. Process dynamics for pilot-scale composting of
Kraft paper in yard debris mixture.

92

1 0 1 00
+ pH
+ Moisture (%) _. go

 

-80

 

L70

 

~60

-50

 

r40

 

30

 

 

40

30 90
A C/N -10.Gexp(-O.15t)+16

0 DH . 33(1-exp(-O.121))+48

 

 

. . ’80

- 70
DH'IS
- 60

 

L50

 

 

 

o 10 20 30 40
Time (days)

Figure 6-12. Chemical parameters during pilot-scale
composting of Kraft paper mixture.

93
Tracking of chemical parameters provided insight into the transformation of the
yard debris mixture into a stabilized compost (Figure 6-12). Initially the yard
debris mixture was at pH 6.1 and had risen to pH 7.9 within four days. The low
initial pH results from the formation of organic acid intermediates in the yard
debris mixture, and the subsequent rise to pH 8.7 is due to the rapid thermophilic
growth of microorganisms capable of metabolizing these intermediates. As the

composting temperature fell below 40°C the pH lowered to near neutral.

TABLE 6-10. Analysis of Kraft paper/yard waste mixture.

 

 

 

 

Moisture content(%) 60
pH 6.1
%Carbon 44.6
%Nitrogen 1 .68
C/N ratio 26.6

 

Table 6-11 shows the carbon balance for the composting experiment. Solid phase
carbon was measured by CHN elemental analysis and carbon dioxide evolution
was calculated using the air flow rate and effluent gas composition. This
respirometric calculation of carbon evolved as carbon dioxide was less reliable
than the solid phase analysis. Based on the solid phase analysis 51% of the initial

carbon was converted to carbon dioxide as opposed to 65% based on respirometry.

The effect of this carbon loss on the carbon to nitrogen ratio is shown in Table
612. High temperature and alkalinity during the first half of the experiment

provided conditions favorable for the release of nitrogen as ammonia, and total

TABLE 6-11. Carbon balance for Kraft paper/yard waste composting.

94

 

 

 

 

 

 

Carbon species Designation g Carbon
Initial mixture C, 2227
Final compost Cf 1099
Mineralization to CO; Ccog 1397
Carbon balance closure (Cf + CCO2)/Ci x 100 (112%)

 

TABLE 6-12. Change in C/N during Kraft/yard waste composting.

 

 

 

 

 

Initial Final
Dry weight (g) 5000 2701
%Carbon 44.6 40.7
%Nitrogen 1.68 2.41
Total carbon (g) 2227 1099
Total nitrogen (g) 84 65
C/N 26.6 16.9

 

 

 

nitrogen content decreased. Since the total nitrogen content decreased by only
23% while the total carbon content decreased by over 50%, the carbon to nitrogen
ratio decreased from 26.6 to 16.9 over the course of the experiment as shown in

Figure 6-12.

95
Fractionation of humic substances provided a means of following the changes in
the carbon species present (Ciavatta et al., 1990). As described in Chapter 4,
compost organic matter was separated into the alkaline extractable carbon (TE)
and the non-extractable carbon, or humin (HU). The portion of the TB fraction
that precipitated below pH 2 is called humic acid (HA). Chromatographic
fractionation of the pH 2 soluble material on polyvinylpyrrolidone isolated the
soluble humified fraction, fulvic acid (FA), from the non-hmnified (NH)

components.

The degree of humification (DH), i.e. the ratio of humified carbon (HA + FA) to
total extractable carbon (TE), increased rapidly from 48% at the beginning of
composting and converged on 80% at the end of composting (Figure 6-12). The
contribution of each extractable carbon fraction to the TB is shown in Figure 6-13.
Increasing DH during composting was due to an increase in the relative amount of
HA carbon, a stable level of FA carbon, and a decrease in NH carbon. The
compost sample collected on day 32 gave a C/N ratio and humic substance profile
that was characteristic of a less mature compost than the prevalent trend, and was

treated as the result of poor sampling.

Table 6-13 shows the initial and final carbon species distributions both on the
basis of carbon species weight and as percentage of the total carbon in the solid
phase. There was no net change in the amount of HA during composting.
However, the mineralization of HU, NH, and FA to carbon dioxide resulted in a
final HA concentration twice that of the initial mixture. This large increase in HA
and small decrease in FA relative to the large decrease in NH caused the high DH

of the composted material.

Percent of TE

96

 

70-

60*

50 ~

40‘

30

201

10‘

 

-o—ha
+nh
+fa

 

I‘l-

 

10

I

20
Time (days)

30

 

40

Figure 6-13. Contribution of each extracted fraction to the

total extractable carbon.

97
Kraft paper rapidly disintegrated in the pilot-scale yard debris composting trial.
The high moisture content through most of the trial aided in the loss of structural
integrity. In addition, the Kraft paper discolored and by day 22 visual distinction
of the Kraft paper became difficult. At the end of the 38 day composting trial the
compost product was a homogeneous brown-black color with no distinguishable

paper residue.

TABLE 6-13. Distribution of carbon in humic fractions.

 

 

 

Species Initial g C Final g C Initial % of Final % of
solid phase C solid phase C
Carbon dioxide 0 1128 0 0
Humin (HU) 1858 943 83.4 85.8
Humic acid (HA) 106 104 4.8 9.4
Fulvic acid (FA) 71 19 3.2 1.8
Non-humified (NH) 192 33 8.6 3.0

 

 

 

 

 

 

6.4 BIODEGRADATION OF POLYLACTIDE FILM

After the composting experiments in Section 6.3 established process dynamics and
chemical transformations that typically occur during pilot-scale composting, the
final round of experiments were performed to determine the effects of polylactide
film on process dynamics and compost product quality. Polylactide molecular
weights were measured after two and four weeks, and compared to the values

predicted by the model of polylactide hydrolytic degradation.

98
6.4.1 Experimental Method

The four pilot—scale composting experiments lasted 28 days each, using the
mixtures outlined in Table 6-14. Polylactide film, Kraft paper, and polyethylene
film, all in 6 x 6 inch sheets, were loaded at 12.2% on a dry weight basis into a
mixture of shredded green yard debris and anaerobic digester sludge from the
Fairfield WPCF. Table 6-15 gives the initial characteristics of this mixture. Since
the mixture consisted primarily of green yard debris, the process control setpoints
and PID tuning parameters were set as for the yard debris composting experiment
in Section 6.3.3. On days 4, 7, 11, 14, 18, 21, and 25 the composts were mixed by
transfening the contents of each pilot vessel into five pails, mixing the contents in
the pails, and reloading the vessel in reverse order. The compost pH was
measured six times over the course of the experiment. Samples of the compost
and sheets of the test materials were removed after two weeks and four weeks for

chemical analysis.

TABLE 6-14. Mixtures for pilot-scale composting experiments.

 

 

 

 

 

 

 

 

Compost Test Matl. Total Thickness

(kg d.w.) (kg d.w.) (kg d.w.) (mil)
Blank 6.80 - 6.80 —
polyethylene 6.80 0.942 7.74 l .0
40# Kraft paper 6.80 0.942 7.74 5.6
polylactide 6.80 0.942 7.74 2.9

 

99
TABLE 6-15. Compost feedstock characteristics

 

Moisture content 66 i 2%
pH 73:05
Ash content 19.2 i 2.4%
Carbon (dry weight basis) 43.3 :1: 2.1%
Nitrogen (dry weight basis) 1.56 :l: 0.19%
C/N ratio 28.1 i 4.1

 

 

 

 

6.4.2 Effect of Polylactide on Process Dynamics

The four compost test batches displayed process dynamics resembling those of the
polylactide compost mixture shown in Figure 6—14. These mixtures behaved
similarly to the sludge-wood chip mixture composting described in Section 6.3.2,
but the use of green brush and yard debris material in place of wood chips resulted
in less process stability and increased setpoint overshoots. For the first half of the
experiment temperature-based process control dominated. Compost temperatures
exceeded the 55°C setpoint only briefly during this period, and at these times only
in the top portion of the vessels. The bulk of the compost remained within
temperature ranges desired for minimizing odor production. Figure 6-15 shows
the average compost temperatures of the four mixtures. During the first half of the
experiments the average temperatures of the four batches stayed approximately
equal. However, toward the end of the experiments the compost containing Kraft

paper maintained higher temperatures than the other three batches. The Kraft

100

 

 

 

 

 

 

 

 

 

 

 

 

70

o o o e o e 0

fl fl OMixing
60 -
Top T(°C)

50 -

l
40 -

Bottom T(°C) F
30 ‘
Oxygen in effluent (%)
20 d } WJ‘Wtf‘Il-v
l

 

 

 

1 0
Air flow rate (LPM)
0 1 "T I ‘ I b
0 5 1 O 1 5 20 25 30
Time (days)

Figure 6-14. Process dynamics for compost mixture
containing polylactide film.

Compost Average Temperature (°C)

101

 

 

  

— Kraft-average
— Blank-average
--- EooPLA-average
W... PE-average

 

 

fl I I T I

5 1o 15 20 25
Time (days)

Figure 6-15. Average temperatures in the vessels during
pilot-scale composting.

3O

102
paper added an extra energy source to the compost mixture; addition of

polyethylene and polylactide films did not affect the compost temperatures.

Mineralization of the carbon in the mixtures to carbon dioxide followed the trends
of the polylactide compost mixture shown in Figure 6-16. The highest rate of
carbon dioxide generation coincided with the high temperatures that occurred
during the first two weeks of the experiments. Peak rates of carbon dioxide
evolution during the first five days exceeded 10 gC/day'kg d.w., but generally
remained at levels below 5 gC/day-kg d.w. This carbon dioxide production rate
followed the trends of the sludge-woodchip mixture composting described in
Section 6.3.2. However, as with the temperature, the green yard debris used in
place of wood chips resulted in an oscillatory process (and rate of carbon dioxide

production) during the early stages of the experiments.

The cumulative mineralization of the carbon in the mixture to carbon dioxide was
calculated on the basis of the compost materials alone, or on the basis of the total
carbon in the compost plus the test material. Figure 6-16 shows the percent
mineralization to carbon dioxide evolution from the polylactide on these two
bases. A comparison of the different rates of mineralization for the four batches is
made in Figure 6-17. The percentages shown were calculated based on the
assumption that the Kraft paper contributed to the carbon dioxide production and
the polyethylene and polylactide films did not. Even with the Kraft paper batch
conversion calculated on this basis, the cumulative percentage was greater than for
the other batches. Part of this could be because the 12.2% additional Kraft paper
mineralized at a rate greater than the compost material itself. However, most of

this excess conversion can be attributed to the higher overall microbial activity in

103

 

 

 

 

 

 

      
   
 

 

 

 

 

 

 

 

30
e o o o e o e
o Mixing
25 -
%C from compost
20 r
%C including PLA carbon
1 5 -
1 0 a
/ CPR (gC/daykg DW)
5 - PM
Mr—W—I—M
0 I l I I I
0 5 10 1 5 20 25 30
Time (days)

Figure 6-16. Respirometric data for compost mixture
containing polylactide film.

%C converted to 002

35

104

 

30‘

25-

20-

 

 

 

 

Kraft (+)
Blank
PE (-)
PLA (-)
5 1 0 15 20 25 30
Time (days)

Figure 6-17. Percent mineralization of initial carbon to carbon dioxide

during pilot-scale composting.

35

105
the mixture, since the excess carbon dioxide generation from the Kraft paper
mixture corresponds to the period when this batch maintained a higher

temperature.

The addition of polylactide film did not affect the temperature of the compost or
have a large impact on the rate of carbon dioxide generation from the compost
mixture. The cumulative carbon dioxide production from the mixture containing
polylactide film was slightly lower than the production from the blank and
polyethylene mixtures. Most of this difference resulted during a period of lower
temperature in the polylactide-compost mixture at day 15. The difference could be
simply an effect of this lower temperature, or could be the result of microbial
activity inhibition by the high concentration (12.2% loading) of polylactide film.
Polylactide underwent little reduction of molecular weight, so inhibition is
probably not due to diffusion of lactic acid or oligamers into the compost.

Without further data to support this explanation, no firm conclusion can be drawn.

The carbon dioxide rate, initial and final compost carbon contents, and total dry
weights gave the carbon balances shown in Table 6-16. The first three rows give
the total grams of carbon present initially, and show how the initial carbon
partitioned between the carbon dioxide and the final product. The next two rows
describe the partition as the percentage of initial carbon that ended up in each of
the final fractions. For the polyethylene mixture a lower percent of the original
carbon was recovered as carbon dioxide and end product, possibly because the
separation of the polyethylene film and compost into two phases resulted in the

collection of a non-representative sample of compost.

TABLE 6-16. Carbon balances for Pilot-Scale Composting Experiments.

 

 

 

 

 

Carbon species (grams) Blank PE Kraft PLA
Initial mixture (C,) 2938 3711 3353 3390
Final compost (Cr) 2118 2309 2469 2363
Mineralization (Ccog) 894 858 1108 789
%Initial C as compost 72% 62% 74% 70%
%Initial C as C02 30% 23% 33% 23%
Carbon balance closure (102%) (85%) (107%) (93%)

 

 

 

 

 

 

6.4.3 Effect of Polylactide on Product Quality

Characterization of the four composts after two weeks and four weeks was
performed to determine whether the addition of polylactide film at 10% loading
affected the quality of the compost product. Measurements of compost moisture
content, total weight, pH, carbon to nitrogen ratio, ash content, volatile solids
reduction, degree of humification, and cress seed germination index provided a
wide range ofcompost monitoring and product quality parameters. These
measurements do not reveal any changes to the compost product quality due to the
addition of polylactide film, but do give a record of the chemical environment to

which the polylactide was exposed.

Moisture content

The moisture content of the raw compost feedstock started at 66%. Addition of

12.2% dry test materials to three of the batches brought their moisture contents to

107
63%. Over the course of the 28 day experiments, the moisture contents remained

in the range of 63-68%.

Total weight
After two and four weeks the contents of the vessels were removed and weighed.
The composts underwent a 21-26% reduction in dry weight of the four week

experiments (Figure 6-18a).

Compost pH

The pH of the compost was measured six times during the experiments. At the
end of one week all of the composts reached their maximums of pH 8.3 to 8.5.
After this the pHs decreased to final values of pH 5.9 to 6.6 (Figure 6-18b). This
is nearly identical to the pH during the yard debris composting experiment in

Section 6.3.3.

Carbon to nitrogen ratio

Initially the carbon to nitrogen ratios of the composts were 26 to 27. By the
second week they were very close to their final values of 17 to 19 (Figure 6-18c).
This mirrored the initial and final carbon to nitrogen ratios of the yard debris

composting experiment.

Ash content

The ash content increased from 17% initially to 21.5-23.0% over the first two
weeks. Each batch also increased during the second half of the experiment, to

final values of 23.1-24.6% ash (Figure 6-18d).

108

 

 

 

 

 

 

 

 

 

start 2 wk 4 wk
(a) %original d.w. vs. time

 

 

 

 

 

 

 

 

 

 

start 2wk dwk start 2wk 4wk

 

 

 

 

 

 

 

 

 

(c) C/N ratio vs. time (d) %Ash vs. time
100
m -l
so 4
70 l ..
so 1
so . . - . . . .
start 2m 4m start 2m 4m
(e) %original vol.solids vs. time (t) degree of humitication vs. time
—0— blank - a - polyethylene ---°--- kraft - - t - - polylactide

Figure 6-18. Pilot-scale composting product quality parameters.

109

Volatile solids reduction

The change in mass of volatile solids was calculated for the total dry weights of
the composts. By the end of four weeks the total volatile solids of the initial

compost mixes had been reduced by 27 to 33% (Figure 6-18e).

Degree of humification

Modest humification of the compost mixes occurred compared to the yard debris
humic substance increases seen during yard waste composting in Section 6.3.3.
The initial degree of humification of 43% was only slightly lower than the initial
degree of humification of the yard debris mixture, but only increased by about
10% to final values of 53.6-57.0%. This is much less than the over 30% increase

in degree of humification during the yard debris composting experiments.

Cress bioassa y

The results of cress germination and root elongation bioassays of the raw and
finished composts are shown in Figure 6-19. Results are shown as the total weight
of seeds genninated in full strength compost extract as a percentage of the total
weight of seeds germinated in the water control. Fresh composts prevent
germination and inhibit root growth due to the presence of transient species , such
as organic acids. As composts mature this toxicity decreases. This type of
calculation accounts for both the lack of germination due to high concentration
toxicity and the inhibition of root growth caused by low concentration toxicity.
Error bars show the 95% confidence interval for the ten replicate dishes of
ten seeds each. The initial compost strongly inhibited germination and root
elongation, giving only 49% of the genrrination/elongation of water. By the end of
four weeks the blank, Kraft paper, polylactide, and polyethylene composts gave
95, 92, 89, and 76%, respectively. The final values show less inhibition than the

110
fresh compost, but the level of gennination of the four batches are not different

from each other at the 95% confidence level.

 

_L .5

CD 0 N

O O O
l l

%germination relative to water
0)
O
l

 

 

 

.cxa:
«accuse
assert
LLID

Figure 6-19. Results of germination and root elongation
bioassay for fresh (0 day) and finished (28 day) pilot-scale
composts. (Error bars show the 95% confidence intervals for
n=10 dishes with 10 seeds each)

6.4.4 Pilot-Scale Polylactide Degradation
Samples of polylactide film were recovered from the pilot-scale vessels after two
weeks and four weeks of composting, and the molecular weight distributions were

detemrined by GPC. Pieces of film equivalent to approximately two 6 x 6 inch

111
sheets were removed and air dried at room temperature. Table 6-17 shows the

results of the GPC analysis.

TABLE 6-17. GPC analysis of films from pilot-scale composting.

 

 

Time (days) Mn Mw Mw/Mn
0 96,000 215,000 2.24
14 55,000 115,500 2.10
28 45,800 107,000 2.33

 

 

 

 

 

 

The decrease in molecular weight over 10 minute time increments was calculated

using Equation 3-10:

-14543
T

 

lnMn =lnMM—3.664><1018 'exp( )-t (3-10)

and the temperature average of the compost in the pilot-scale vessel during each
time increment. Figure 6-20 shows the model’s predictions for Mn over the course
of the experiment compared to the Mn measured by GPC. Most of the molecular
weight reduction occurred during the periods of peak temperature during the first
two weeks of the experminet. The molecular weight average reached Mn 55,000
after two weeks and decreased slightly to Mn 45,800 after two more weeks.
During the second half of the experiment the compost average temperatures were
much lower so very little hydrolysis occurred. With the small reduction in
molecular weight, the polydispersity of the polylactide molecular weights

remained nearly constant during the experiment.

112

1 00000

 

—-nxuel
I GPC

90000 -
80000 .
70000 a
60000 -

§ 50000 -

 

 

40000 - T“
30000 W
20000 J

10000 -

 

O I I I I I

 

 

0 5 1 0 1 5 20 25
Time (days)

Figure 6-20. Comparison of predicted and measured
decrease in M, during pilot-scale composting.

113
The photographs in Figure 6-21 show the appearance of the film initially and after
two and four weeks. By day three the transparent film was opaque white as a
result of increased crystallinity. The film became brittle early in the experiment
as the molecular weight decreased, but at the end of composting the sheets of film

remained essentially intact.

 

Figure 6-21(a). Photograph of polylactide film prior
to pilot-scale composting.

114

 

Figure 6-21(b). Photographs of polylactide film after
composting two weeks (top) and four weeks (bottom).

CHAPTER 7

 

BIODEGRADATION UNDER
SEMI-CONTROLLED
COMPOSTING CONDITIONS

 

7.1 OVERVIEW

The detailed analysis of a material’s fate and effects in a feedback-controlled
compost environment, as presented in Chapter 6, may not be necessary after
inherent biodegradability has been established using the methods in Chapter 5.
Composting processes that start with fresh feedstocks are more biochemically and
microbiologically active than test environments using mature compost as an
inoculum. If ASTM Method D 5338 results in rapid and complete mineralization
of a test material, then biodegradation can be expected to proceed at least as
rapidly in a dynamic compost. However, this still does not eliminate the need for
demonstrating compostability of items made of inherently biodegradable materials.
The simple method described in this chapter for exposing test materials to
composting conditions provides another look at polylactide molecular weight loss
and disintegration. The accommodating size of this type of apparatus also makes
it ideal for determining degradation behavior of biodegradable resins in end-user

form.

115

116
7.2 TEST APPARATUS AND CONDITIONS

The vessel for the semi-controlled compost exposure consisted of a twenty liter
pail fitted with a perforated false floor for air distribution (Figure 7-1).
Humidified air entered the vessel at 150 ml/min from the bottom an exited from
the top. A synthetic compost prepared from eight substrates (Table 7-1) provided
the matrix for exposure. This mixture (1 kg dry weight) had a carbon to nitrogen
ratio of 25:1 and a moisture content of 60%, and filled over half of the vessel.
Three squares of polylactide film 5 cm x 5 cm were placed in plastic berry baskets
with compost, and three baskets were loaded into the center of the vessel. These
baskets facilitated the recovery of partially disintegrated film fragments. The
entire set-up was subjected to the ASTM D 5338-92 temperature profile.

t-Izt:v;o'c-I:l'l:llb'a.u‘l:.'l.1:)».lzi'q '-'c'v' u'o'-'p'n'o‘u'-'.'o'o 'u'. ''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''
''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''

flowmeter

 

 

valve

 

 

 

humidifiction . _
bottle

 

 

controlled
temperatmc

 

 

 

 

 

 

.;.;.;.-.;.;.;.;.;.-.;.;.;.;.;.;.;.;.j.;.;.;.;.;.;.3.1.3.;.;.;.;.;.-.:.;.:.;.;.-.;.-.-.-.;.-.'.-.-.-.-.'.-.-.-.-.-.-.:.-.:.:.;
.......................................................................................

 

Figure 7-1. Apparatus for semi-controlled composting experiment.

1 17
Temperature was measured nine times over the course of the experiment using a
handheld thermocouple unit. Carbon dioxide concentration of the exhaust gas was
measured by gas chromatography. Every four days the baskets were removed
from the vessel, pieces of film were retrieved, the compost pH was measured, and

the compost was mixed and returned to the vessel.

TABLE 7-1. Synthetic compost mixture for semi-controlled composting.

 

 

 

 

 

FEEDSTOCK Type %C %N %M totalC totalN g DW g WW
cypress mulch C 48.5 0.14 53 728 2 150 319
Kraft paper C 44.1 0.05 1 441 1 100 101
corn cob grits C 59.0 0.60 5 885 9 150 158
ground carrots C,N 42.7 1.06 86 214 5 50 357
shredded leaves C,N 43.0 1.16 5 645 17 150 158
alfalfa N 45.0 3.30 5 1 125 83 250 263
cat food bits N 50.6 5.80 32 506 58 100 147
mature compost inoc. 23.0 2.00 28 115 10 50 69
water 925
TOTALS 4658 1 84.8 1000 2500

 

 

 

 

 

 

 

 

 

Figure 7-2 shows the processing conditions of the semi-controlled composting
experiment. The compost self-heated above the temperature of the incubator; this
behavior is the reason for refening to the conditions as semi-controlled. Compost
carbon dioxide evolution reached a peak during the mesophilic warm—up period

and another peak after one week. During these peaks of activity the carbon

 

118

 

70

W Compost temp (°C)

so. M ¥+

50-

 

 

 

 

‘— lncubator temp (°C)

40- -20

 

_, %COz in effluent

30- P15

 

.b10

 

 

 

Tlme (days)

Figure 7-2. Processing conditions for semi-controlled
composting exposure of polylactide film.

119
dioxide concentrations exceeded 15%. The compost pH increased from an initial

value of pH 6.0 to a final value of pH 8.8.

7.3 POLYLACTIDE DEGRADATION

Samples of polylactide film were recovered after 4, 8, and 12 days of composting.
On day 16 no distinguishable fragments of film could be found. The molecular
weight reduction of the polylactide again followed the kinetic model for
polylactide degradation (Figure 7-3). By day 12 the number average molecular
weight had decreased below MH 5000 and the polydispersity index had fallen to
Mw/Mn 1.40 as shown in Table 7-2. Due to the higher temperatures, polylactide
degradation proceeded more rapidly than in either the externally controlled or

feedback-controlled composting experiments.

TABLE 7-2. GPC analysis of films from semi-controlled composting.

 

 

 

 

 

 

 

Time (days) Mn Mw Mw/Mn
0 96,000 215,000 2.24
4 28,700 58,000 2.03
8 10,400 19,000 1 .86
12 4,800 6,700 1 .40

 

120

 

1 00000

90000 -

80000 a

70000 -

60000 a

50000 a

40000 a

30000 ‘

20000 -

10000 -*

 

 

 

 

Time (days)

Figure 7-3. Comparison of predicted and measured decrease
in M, during semi-controlled composting.

CHAPTER 8

 

CONCLUSIONS AND RECOMMENDATIONS

 

8.] CONCLUSIONS

Composting is increasingly being recognized as a cost-effective and
environmentally sound method for managing certain solid waste streams. Unlike
other methods of recycling, composting does not require segregation of individual
material types. Any mix of compostable materials can be collected and processed

simultaneously.

The thesis provided an evaluation of the potential of polylactide as a compostable
replacement for current plastic packaging. The following conclusions can be

made:

0 Polylactide and other biodegradable plastics can replace current
non-biodegradable plastics in certain applications to result in completely
compostable waste streams. For a material to be compostable it must not end
up in the reject fractions, must be completely biodegradable, and must not be

detrimental to compost product quality or processing.

0 A new respirometric biodegradability testing apparatus allows accurate and
reliable measurement of carbon dioxide evolution from biodegradable materials
under externally controlled composting conditions as per ASTM Method
D 5338.

121

122

Upon exposure to the ASTM D 5338 conditions, the molecular weight average
of polylactide film rapidly decreases by non-enzymatic hydrolysis. Microbial
utilization begins when the molecular weight average approaches Mn 10,000
and continues at a rate that is nearly first order with respect to temperature.
The ASTM D 5338 temperature profile results in 85% conversion of
polylactide carbon to carbon dioxide after 70 days. The majority of the
biodegradation occurs before day 28, when the temperature is decreased from

50°C to 35°C.

The pilot-scale composting system using PID feedback control provides
conditions similar to those found in full-scale forced aeration composting
processes. Tight control at temperature and oxygen concentration setpoints
requires linear and stationary process dynamics. Even though biologically
active composts display non-ideal temperature dynamics, control is possible

with only brief periods of temperature setpoint overshoot.

As substrate carbon is mineralized to carbon dioxide, the carbon to nitrogen
ratio of the compost decreases, the ash content increases, the degree of
humification increases, and the phytoxicity decreases. Feedstock
transformation into compost proceeds rapidly during the initial stages of
composting when the temperatures are high. As feedstock materials stabilize,
the temperature drops and chemical and biochemical character of the compost

changes slowly.

Under forced aeration a temperature gradient is established in the compost that
ranges from the cooling air temperature to the temperature control setpoint.

Since the average compost temperature is low the rate of polylactide molecular

123

weight reduction is slow. Polylactide film exposed to the pilot-scale
temperatures decreases in average molecular weight from Mn 96,000 to
Mn 48,600 after four weeks of composting. Most of this molecular weight
decrease occurs during the first two weeks, when the temperature is higher.
Since the polylactide film does not degrade under these conditions, its presence

does not affect compost process dynamics or product quality.

The semi-controlled composting apparatus provides a simple method for
exposing biodegradable materials to a dynamic matrix of synthetic compost.
The compostability of larger items in their end-user form can be testing in this

apparatus after their inherent biodegradability has been established.

The semi-controlled compost maintains temperatures higher than the externally
controlled or feedback-controlled composts. Polylactide film exposed by this
method degrades rapidly. After only twelve days of exposure the molecular
weight average drops to Mn 4,800 and after sixteen days fragments of

polylactide film are indistinguishable in the compost.

A first-order model of polylactide molecular weight reduction exists which was
developed based on the kinetics of polylactide degradation when implanted
into animal tissues. Some data on polylactide degradation is available for
temperatures out of the physiological range. The rate constant as a function of
temperature can be developed by applying Arrhenius kinetics to this molecular

weight reduction data. The resulting rate constant expression:

-14543
T

 

lnMl1 =lnMn‘o—3.664x10‘8-exp( )-t (3-10)

124

can then be applied in the model to predict the degree of polylactide
degradation during a composting exposure with time—variant temperature.
Molecular weight reductions measured by GPC confirm that the temperature

dependent model is applicable to polylactide degradation in compost.

o The molecular weight reductions predicted by the kinetic model and measured
by GPC, and the rate of carbon dioxide evolution give evidence for the
biodegradation mechanism of polylactide during composting. Carbon dioxide
evolution from the polylactide ensues only after the average molecular weight
has been dramatically reduced. This reveals that utilization by compost
microorganisms occurs after non-enzymatic degradation of polylactide
produces low molecular weight oligamers that are able to diffuse out of the

polymer bulk.

8.2 RECOMMENDATIONS FOR FURTHER WORK

Based on this research, a strategy for evaluating compostability of biodegradable
materials can be recommended. Knowledge of chemical and enzymatic
degradation mechanisms for a particular material can aid in the design of
biodegradability and compostability experiments. Inherent biodegradability under
composting conditions can be detennined using standard sized pellets, powders, or
films in the new test apparatus for ASTM D 5338. If a material reaches 100%
biodegradation in a reasonable time, this is evidence that the material and its
degradation products are not strongly inhibiting microbial activity in the compost
inoculum. For some materials, ASTM D 5338 biodegradability may be rapid but

not reach 100%. The process dynamics of pilot-scale composting may reveal

125

inhibition effects from such materials. The semi-controlled exposure method
described in Chapter 7 can replace the pilot-scale system as a method for exposing
large items to a dynamic compost. However, certain attributes of the pilot-scale
system that translate to full-scale processes, particularly the temperature gradients,

are not present in the semi-controlled exposure method.

As described in Section 3.2.1, there may be an increase in the slope of the
degradation rate constant vs. temperature for polylactide at the glass transition
temperature. During most of the polylactide exposures in this research the
temperature remained near or below Tg. Exposures of polylactide to compost at
constant temperatures near and above Tg could be used to generate a separate
temperature dependent rate expression for temperatures above Tg. This
modification would result in dependable predictions of degradation behavior at

higher temperatures.

The kinetics of polylactide degradation make temperature the most critical
parameter for achieving rapid degradation, as long as other process characteristics
are in their typical ranges. However, odor management is the primary problem of
most composting facilities, and compost researchers have determined that aeration
and low temperatures minimize odor generation. As compost facilities
successfully control odor production by temperature control the potential for
polylactide degradation is reduced, as in the pilot-scale experiments of Section 6.4.
New process control strategies that maintain an even temperature profile could
benefit composting by ensuring low odor formation potential and rapid

degradation of polylactide and other substrates. Composting facilities tuned to

126

provide favorable conditions for both processing and biodegradation will allow
widespread use of biodegradable materials in a rational approach to solid waste

management.

LIST OF REFERENCES

LIST OF REFERENCES

American Society of Testing and Materials, ASTM D 425-88 Standard Test
Method for Centrifuge Moisture Equivalent of Soils, Annual Book of ASTM
Standards, Vol. 04.08, 1988.

American Society of Testing and Materials, ASTM D 5209-91 Standard Test
Method for Determining the Aerobic Biodegradation of Plastic Materials in
the Presence of Municipal Sewage Sludge, Annual Book of ASTM Standards,
1991.

American Society of Testing and Materials, ASTM D 5338-92 Standard Test
Method for Determining Aerobic Biodegradation of Plastic Materials Under
Controlled Composting Conditions, Annual Book of ASTM Standards, 1992.

Carroll, B.A., Caunt, P., and G. Cunliffe, J. IWEM, 7:175-181, 1993.

Ciavatta, C., Govi, M., Vittori Antisari, L., and P. Sequi, J. Chromatog, 509:141-
146, 1990.

Cohen, G.H., and GA. Coon, Trans. ASME, 75:827-839, 1953.

Craig, P.J., Williams, J.A., Davis, K.W., Magoun, A.D., Levy, J .A., Bogdansky,
S., and JP. Jones, Surg. Gynecol. Obstet., 141:1-10, 1975.

Curran, M.A., Environ. Sci. Technol., 27(3):430-436, 1993.
Cutright, D.E., Beasley, J.D., and B. Perez, Oral Surg., 32:165-173, 1971.

David, 1., Master’s Thesis, Dept. of Chemical Engineering, Michigan State
University, 1994.

Dijkstra, P.J., Eeninck, M.J.D., and J. Feijen in Degradable Materials: Prospects,
Issues, and Opportunities, CRC Press, 1990.

127

128
Dreinhoefer, L.H., IEEE C ontrol Systems Mag, 56—60, April 1988.

Fava, J.A. et al., Soc. Environ. Toxicol. Chem. Smuggler’s Notch workshop, 1990.

Flaig, W. in Humic Substances and Their Role in the Environment, F. H. Prime]
and R. F. Christman, Eds., Wiley, Chichester, p 75—92 (1988).

Flaig, W., Beutelspacher, H., and E. Rietz, in Soil Components Vol I, J.E.
Gieseking Ed., Springer-Verlag, New York, p 1-211 (1975).

Finger, S.M., Hatch, R.T., and T. M. Regan, Biotech. Bioeng, 17:1193-1218,
1976.

Fordyce, A.P., Rawlings, 1.8., and TE Edgar, in Computer Control of
Fermentation Processes, Ed. by DR. Omstead, 165-220, CRC Press, 1990.

Gruber, P.R., Hall, E.S., Kolstad, I.J., Iwen, M.L., Benson, R.D., Borchardt, R.L.
U .S . Patent Number 5,274,073, Dec. 1993.

Hamelers, H.V.M., in Science and Engineering of Composting: Design,
Environmental, Microbiological, and Utilization Aspects, H. A. I. Hoitink
and H. M. Keener, Ed., p. 36-58, 1993.

Hayes, M.H.B., MacCarthy, P., Malcolm, R.L., and RS. Swift, Ed., Humic
Substances II, In Search of Structure, p. 6-7, 1989.

Holland, S.I., Tighe, B..l., and PC. Gould, J. Control Rel, 4:144-180, 1986.

Iannotti, D.A., Pang, T., Toth, B.L., Elwell, D.L., Keener, H.M., and H.A.J.
Hoitink, Compost Sci. & Util., 1(3):52-65, 1993.

Jeris, 1.8., and R.W. Regan, Compost Science, 8-15, March-April 1973.

Kulkarni, R.K., Parri, K.C., Neuman, C., and F. Leonard, Arch. Surg., 93:839-843,
1966.

Li, S.M., Garreau, H., Vert, M., J. Materials Sci.:Materials in Medicine, 1:123-
130,1990a.

Li, S.M., Garreau, H., Vert, M., J. Materials Sci.:Materials in Medicine, 1:131-
139, 1990b.

129

Luyben, W.L., Process Modelling, Simulation, and Control for Chemical
Engineers, 2nd Ed., McGraw-Hill, 1990.

MacGregor, S.T., Miller, F.C., Psarianos, K.M., and MS. Finstein, Appl. Environ.
Microbiol, 41(6). P 1321-1330, 1981.

Mason, N.S., Miles, CS, and RE. Sparks, Polymer Science and Technology,
American Chemical Society, 279-291, 1980.

McKinley, V.L., and JR. Vestal, Appl. Environ. Microbiol, 47(5):933-94l , 1984.

Narayan, R., K unststofi‘e, 79(10): 1022-1027, 1989.

Narayan, R., INDA Nonwoven Res., 3: 1, 1991.

Narayan, R., ACS Symp. Ser. 476, l-lO, 1992.

Narayan, R. in Science and Engineering of Composting: Design, Environmental,
Microbiological, and Utilization Aspects, H. A. J. Hoitink and H. M. Keener,
Ed., p. 339-362, 1993a.

Narayan, R. in Opportunities for Innovation: Biotechnology. US. Dept. of
Commerce NIST GCR 93-633, R.M. Busche, Ed., p. 135-150, 1993b.

Ozeki, E., and H. Ohara, Third International Workshop on Biodegradable Plastics
and Polymers, Osaka, 1993.

Pitt, C.G., Chasalow, F.I., Hibionada, Y.M., Klimas, D.M, and A. Schindler, J.
Appl. Polym. Sci., 26:3779, 1981.

Pitt, C.G., in Biodegradable Polymers and Plastics, ed. M. Vert, J. Feijen, A.
Albertsson, G. Scott, and E. Chiellini, Royal Society of Chemistry, 1992.

Reed, A.M., and D.K. Gilding, Polymer, 22:487-498, 1980.

Seborg, D.E., Edgar, TE, and DA. Mellichamp, Process Dynamics and Control,
Wiley Publ., 1989.

Schindler, A., Jeffcoat, R., Kirnmel, G.L., Pitt, C.G., Wall, ME, and R.
Zweidinger, C ontemp. Top. Polym. Sci., 2:251-286, 1977.

Sikora, L.J., and MA. Sowers, J. Environ.Qual. 14(3):434-439, 1985.

 

130

Sinclair, R.G., Environ. Sci. Techn, 7:955, 1973.

Snook, J .B., and R. Narayan, in Waste Management and Recycling International
I 994, MA. Johnson and Y.Samiullah, Ed., Sterling Publ. Ltd., London, 40-
43, 1994.

Spinu, M., US. Patent Number 5,270,400 Dec. 1993.

Stentiford, E.I., Taylor, P.I., Leton, T.G., and DD. Mara, Wat. Pollut. Control,
23-32, 1985.

Strom, P.F., Appl. Environ. Microbial, 50(4):899-905, 1985.
Suler, DJ, and Finstein, M.S., Appl. Environ. Microbiol, 33(2):345-350, 1977.

United States Federal Trade Commission, Guides for the Use of Environmental
Marketing Claims, July, 1992.

Vestal, J .R., and DC. White, BioScience, 39(8):535-541, 1989.
Wedel, A.W. and CR. Daubert, J. Ferm. Qual Control, 13:45-64, 1994.
Williams, D.F., Eng. Med, 10(1):5-7, 1981.

Wilson, J.A., in Measurement and Control in Bioprocessing, Ed. by KC. Carr-
Brion, Elsevier, London, p 1-35, 1991.

Yolles, S., Acadamy of Pharm. Sci., San Francisco, 1971.
Zeigler, J.G. and NB. Nichols, Trans. ASME, 64:759-768, 1942.

Zhu, K.J., Hendren, R.W., Jensen, K., and CG. Pitt, Macromolecules, 24:1736,
1991.