EPA/600/R-93/166
September 1993
EVALUATION OF RECYCLED
PLASTIC LUMBER FOR MARINE APPLICATIONS
By
R. W. Beck and Associates
Denver, Colorado 80202-2615
and
The Solid Waste Association of North America
Silver Spring, Maryland 20910
Cooperative Agreement No. 818238
Project Officer:
Lynnann Kitchens
Waste Minimization, Destruction, and Disposal Research Division
Risk Reduction Engineering Laboratory
Cincinnati, Ohio 45268
RISK REDUCTION ENGINEERING LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
Printed on Recycled Paper
-------
DISCLAIMER
The Information in the document has been funded wholly or in part by the
United States Environmental Protection Agency under assistance agreement
CR-818238 to the Solid Waste Association of North America (SWANA). It has
been subjected to the Agency's peer and administrative review and has been
approved for publication as an EPA document. Mention of trade names or
commercial products does not constitute endorsement or recommendation for use.
FOREWORD
Today's rapidly developing and changing technologies and industrial
products and practices frequently carry with them the increased generation of
materials that, if improperly dealt with, can threaten both public health and
the environment. The U.S. Environmental Protection Agency is charged by
Congress with protecting the Nation's land, air, and water resources. Under a
mandate of national environmental laws, the agency strives to formulate and
implement actions leading to a compatible balance between human activities and
the ability of natural systems to support and nurture life. These laws direct
the EPA to perform research to define our environmental problems, measure the
impacts, and search for solutions.
The Risk Reduction Engineering Laboratory is responsible for planning,
implementing, and managing research, development, and demonstration programs
to provide an authoritative, defensible engineering basis in support of the
policies, programs, and regulation of the EPA with respect to drinking water,
wastewater, pesticides, toxic substances, solid and hazardous wastes, and
Superfund-rel ated activities. This publication is one of the products of that
research and provides a vital communication link between the researcher and
the user community.
This publication is part of a series of publications for the Municipal
Solid Waste Innovative Technology Evaluation (MITE) Program. The purpose of
the MITE program is to: 1) accelerate the commercialization and development
of innovative technologies for solid waste management and recycling, and
2) provide objective information on developing technologies to solid waste
managers, the public sector, and the waste management industry.
E. Timothy Oppelt, Director
Risk Reduction Engineering Laboratory
-------
PREFACE
The Municipal Solid Waste Innovative Technology Evaluation (MITE)
Program is managed by the U.S. Environmental Protection Agency (EPA) Office of
Research and Development (ORD). The purpose of the MITE program is to:
1) accelerate the commercialization and development of innovative technologies
and programs for solid waste management and recycling, and 2) provide objec-
tive information on developing technologies and programs to solid waste
managers in the public and private sectors.
These goals are met by selecting, through a competitive process, tech-
nologies and programs that have submitted proposals to EPA through its annual
solicitation. Once selected, EPA, with the cooperation of the technology
developer, formulates a plan which evaluates the costs, effectiveness, and
environmental impacts of the technology. Each project consists of a field
demonstration and an associated evaluation. The MITE program is administered
by the Solid Waste Association of North America (SWANA). SWANA coordinates an
Advisory Committee review as well as assisting in the formulation of the
evaluation plans.
The technology developer, California Recycling Company (CRC), in
conjunction with Morrow Associated Enterprises, manufactured the materials'
tested in this program. The evaluation was conducted at the Florida Institute
of Technology and Rutgers University. The MITE program seeks to gather envi-
ronmental and cost information on new and developing technologies. This
report presents the results of this evaluation.
A limited number of copies of this report will be available at no charge
from EPA's Center for Environmental Research Information, 26 West Martin
Luther King Drive, Cincinnati, Ohio 45268. Requests should include the EPA
document number found on the report's cover. When the supply is exhausted,
additional copies can be purchased from the National Technical Information
Service, Ravensworth Building, Springfield, Virginia 22161, 703/487-4600.
11 i
-------
ABSTRACT
This report presents an evaluation of the recycled plastic materials
(RPM) produced by California Recycling Company (CRC). This evaluation is per-
formed under the Municipal Solid Waste Innovative Technology Evaluation (MITE)
Program of the U.S. Environmental Protection Agency (EPA) Risk Reduction
Laboratory. The lumber is produced from difficult-to-market post consumer
plastic materials which have been recovered from a mixed municipal solid waste
stream at CR Transfer's New Stanton material recovery facility (MRF).
R. W. Beck and Associates assessed the composition of the plastic made
into plastic lumber. This plastic was shredded and granulated followed by air
elutriation to remove paper labels at the New Stanton MRF. The material was
transported to CRC, Los Angeles, where it was formed into 2 in. x 6 in. by
12 ft dimensional plastic lumber at mixtures ranging from 100% commingled
plastic to 50% commingled blended with 50% industrial regrind plastic. The
continuous extrusion process used a foaming agent to impart a lighter core to
the plastic lumber.
A battery of tests was performed on the RPM to determine strength,
creep, serviceability, biological compatibility, and toxicity of the plastic
lumber. These tests were selected to characterize the behavior of the
material for marine applications.
The findings show that the plastic lumber produced by CRC has signifi-
cant creep characteristics which must be adequately addressed by appropriate
architectural design when using this material. Flexural stiffness properties
are less than 1/10 that of wood. The plastic lumber shows excellent nail and
screw retention ability compared to wood. Biological testing has indicated a
far lower toxicity for the plastic than for chromium copper arsenate (CCA)
treated wood.
A limited life cycle analysis was performed to compare CCA-treated wood
with the recycled plastic lumber created. The analysis takes into account
lifetime and environmental impacts of the manufacture of CCA-treated wood and
recycled plastic lumber.
This report was submitted in fulfillment of Cooperative Agreement
CR 818238 with the Solid Waste Association of North America (SWANA). The
report was completed by R. W. Beck and Associates under subcontract to SWANA.
This report covers a period from September/91 to August/93 and work was
completed as of March/93.
-------
CONTENTS
Disci aimer i i
Foreword , i i
Preface.... , i i i
Abstract i v
Table of Contents v
Tables viii
Figures .... ix
Acknowl edgments ; x
I INTRODUCTION.. ....1
Plastics in the Waste Stream 1
Collection Strategies. 2
Recycled Plastic. , 2
Technology Description ... 3
Feedstock 3
Composition Analysts 3
Granul at ion , 6
Extrusion 7
Product Types 7
Applications 7
Other Avail able Technologies , 8
The ET-1 and Superwood Processes 8
The Mobil and AERT Processes 9
Comparison Between Molded and Continuous Extrusion.Systems ....10
II EVALUATION OF THE TECHNOLOGY 13
Physical Testing 15
Creep Tests .. , .15
Sample Preparation .15
Test Conditions 15
Apparatus 15
Verification of Creep Test Jigs with Delrin ....16
Results , 17
Flexure Tests 23
Basic Test Procedure ,23
Resul ts 24
-------
CONTENTS
(continued)
Compression Tests .25
Basic Test Procedure 25
Results '.26
Stati c Fri cti on Tests 27
Basi c Test Procedure ,. 27
Results 28
Mechanical Fasteners in Plastic Lumber 30
Basic Test Procedure .30
Results 31
UV Tests 33
Test Conditions 33
Basic Test Procedure 34
Results 35
Environmental Testing , 35
Biofouling Tests ., 35
Sample Preparation 36
Test Conditions , 36
Test Procedure 36
Resul ts 37
Sea Urchin Fertilization Test...., 39
Sample Preparation , 39
Leaching Procedure , 40
Sea Urchin Preparation 41
QA Procedures 41
Results 42
III MATERIAL COMPARISON , 43
Physical Comparison , 43
Limited Life Cycle Analysis of Recycled Plastic Lumber and
CCA-Treated Lumber 44
Methodol ogy 44
Recycled Plastic Lumber , 45
Plastics Acquisition 45
Transport 45
Manufacture 46
Transport to End Use 47
Construct!on , 47
CCA-Treated Lumber 48
Acqui si ti on 48
vi
-------
CONTENTS
(continued)
Transportation to Lumber Mill 49
Lumber Manufacture , 49
Transportation to End Use 50
Comparison .50
Envi ronmental Impacts 51
Durabil ity/Lifetime. 51
Wood '..'. 51
PI asti c .52
Creep „ 52
UV Testing . 52
IV CONCLUSIONS AND RECOMMENDATIONS 53
Concl us ions 53
Cost 53
Envi ronmental Impacts 54
Lifetime 54
Physi cal Strength .-. 54
Friction > 55
Recommendations 55
Field Testing 55
New Formulati ons 56
Consi stency 56
Other Properties 57
Market Research , 57
Appl i cations ; .57
Barriers 57
References 59
Glossary .61
Appendices are not included in this document, limited quantities are available
from Lynnann Hitchens, US EPA Center Hill Research Facility, 5995 Center Hill
Road, Cincinnati, OH 45224.
-------
TABLES
Number
Paae
1.1 Plastic composition by type of packaging, 5
1.2 Feedstock composition by resin 6
2.1 Testi ng methods.. t , 13
2.2 Secant modulus equations statistics, 24
2.3 Compressive modulus of RPM ,..., 27
2.4 Coefficient of friction using material against itself,
ol d and new tenni s and boat shoes 29
2.5 Nail pullout test results summary , . ..32
2.6 Screw pullout test results summary 33
2.7 Biofouling after 4 months of exposure in Indian River
Lagoon , 38
2.8 Leaching protocol '.., , 41
2.9 Summary of assay test data , 42
3.1 Comparison of properties between, plastic and wood...., 43
3.2 Life cycle costs of plastics.., , 46
3.3 Life cycle costs of wood 48
3.4 LCA analysis of CCA-treated 2 in. by 6 in. wood vs.
piasti c 50
viii
-------
FIGURES
Number
Page
2.1 Individual sample in cut away of dry test jig 16
2.2 Delrin predicted vs. mean test values 17
2.3 Creep test results - commingled plastic at room
temperature....; 18
2.4 Creep test results - commingled plastic at elevated
temperature .. 19
2.5 Creep test results - 4% polypropylene plastic at room
temperature. 19
2.6 Creep test results - 4% polypropylene plastic at elevated
temperature 20
2.7 Creep test results - 10% polypropylene plastic at room
temperature 20
2.8 Creep test results - 10% polypropylene plastic at
elevated temperature 21
2.9 Creep test results - 50% HOPE plastic at room temperature 21
2.10 Creep test results - 50% HOPE plastic at elevated
temperature . 22
2.11 Comparison of four plastic formulations - 100-lb load at
room temperature 22
2.12 Flexure test jig 23
2.13 Plot of flex modulus for the four formulations of RPM 25
2.14 Compression test jig 26
2.15 Inclined plane friction test jig 28
2.16 Nail and screw pullout test jig 31
2.17 UV test chamber cross section 34
2.18 Horizontal sample rotation 35
2.19 Lamp rep! acement and rotati on 35
2.20 Biofouling test frames and sample layout 37
2.21 Top portion of exposure frame after 111 days of exposure 39
2.22 Plastic preparation for leaching testing 40
A.I Sample collection locations 63
ix
-------
ACKNOWLEDGMENTS
This report was produced under the coordination of Lynnann Hitchens,
U.S. Environmental Protection Agency (EPA) Municipal Solid Waste Innovative
Technology Evaluation (MITE) Program Manager, at the Risk Reduction Engi-
neering Laboratory, Cincinnati, Ohio. Contributors and reviewers of this
report include Charlotte Frola, Solid Waste Association of North America
(SWANA), Michael Silva and his staff at CR Transfer, Daryl Morrow and Jeff
Lucas-Morrow of Morrow Associates Enterprises, Inc., and Arie Zuckerman and
Conway Coll is of California Recycling Company, Los Angeles, California.
The report was prepared for EPA's MITE program by Benjamin E. Levie,
Timothy Buwalda, Jonathan Burgiel, Bonnie Taher, E. Larry Beaumont, and
Beverly Bergstrom of R. W. Beck and Associates. Drawings of laboratory
apparatus were produced by Charles R. Lockert of the Florida Institute of
Technology (FIT).
Sea Urchin Fertilization experiments were performed by Pedrick Weiss of
the New Jersey Medical School. All other tests were performed by Charles R.
Lockert under the supervision of Gary Zarillo of FIT.
-------
SECTION I
INTRODUCTION
The Environmental Protection Agency CEPA) through the Municipal Solid
Waste Innovative Technology Evaluation Program (MITE) is funding the evalu-
ation of a new process to produce "Recycled Plastic Materials" (RPM). The
plastics targeted are from residential and commercial waste streams. Special
emphasis is placed on using "lower value" plastic waste, or plastic waste in
which the more marketable plastics have been removed. The RPM formed from
these plastics is envisioned as a substitute for wood dimensionable lumber,
such as 2 in. x 6 in. and 4 in. x 4 in. cross section lumber.
The company manufacturing the RPM used in this evaluation is California
Recycling Company (CRC). The CRC manufacturing plant is located in Los
Angeles, California. Process and equipment design was performed by Morrow
Associates Enterprises, Inc.
PLASTICS IN THE WASTE STREAM
Plastic is defined as "any of various nonmetallic compounds, syntheti-
cally produced, usually from organic compounds by polymerization, which can be
molded into various forms and hardened, or formed into pliable sheets or
films, fibers, flexible or hard foams, etc., for commercial use." From this
nroad definition comes plastic's inherent recycling dilemma - there are many
plastic resins, sometimes different types for the same end use (which are
incompatible), making it difficult to economically recover and recycle
plastic. Sorting the plastic by resin type is difficult to do automatically,
and currently the markets for secondary plastic are only strong for soft drink
bottles (polyethylene terephthalate, or PET) and milk jugs (natural high
density polyethylene, or HOPE). Other resins have less recycling value. This
leaves a large portion of the plastic waste stream without an economic alter-
native to landfil'ling or incinerating.
According to a recent EPA study (Franklin and Associates, 1992), plastic
discards contributed 9.1% of solid waste by weight in 1988. This represents
about 14 million tons per year of plastic of which roughly 70% are found in
nondurable and packaging applications. Concentrating on and recycling this
portion of the plastic waste stream is the focus of this project.
-------
COLLECTION STRATEGIES
Two main methods are followed for collecting plastics from the waste
stream. Commingled solid waste can be collected as a single stream and
processed at a centralized facility to remove the recyclable material. In
this scenario, the centralized processing facility can be designed and
operated to extract and sort the plastic which is marketable. Non-marketable
plastic is not recovered, but could be at a later date as markets improve for
these materials.
The other method of collecting plastic is through source separation.
Here, the waste generator is instructed what to separate from the waste
stream. The recyclable materials can be curbside collected or taken by the
waste generator to a drop-off location. Recyclables can be kept separate or
commingled in various-combinations, for subsequent sorting at a material
recovery facility (MRF). For plastics^, source separation programs generally
accept only specific products, such as PET and natural HOPE. However, if
markets exist for other resins, the program could be modified to accept other
plastic containers.
RECYCLED PLASTIC
Thermoplastics, the plastics which comprise all of the common consumer
packaging and nondurables, can be remelted and reformed for a wide range of
non-food contact applications. If proper separation, washing, and processing
of the resins are accomplished, many products can be made. The amount of
recycled plastic in the new products can range up to 100% depending on the
explication. Products currently manufactured with recycled plastic include
bags, detergent and motor oil bottles, carpet fiber, and slip sheets
(substitute for pallets).
Plastic lumber is the generic term given for making dimensional wood and
steel substitutes from post consumer, or industrial, scrap plastic. One or
more plastic resins can be used, along with various additives for color,
strength, or density reduction. Color additives can add to aesthetic quality.
Strength additives can include fiberglass, particular plastic resins, and
wood/fiber. Density reduction can be attained using foaming agents which
create air spaces in the middle of a cross section of plastic lumber. Density
can also be reduced by forming plastic with hollow, or ribbed, internal cross
sections.
Plastic lumber cannot compete with conventional building materials on a
strength basis, but it has certain characteristics which can make it
competitive for specific applications. Included in this list are:
-------
• Weather resistance
• Biological resistance
• Salt water resistance
• Environmental compatibility
• Reuse of waste materials .
These characteristics can make plastic lumber especially competitive as
a substitute for chromium copper arsenate (CCA) treated lumber for marine
applications. This evaluation focused on marine applications as a market
niche for the plastic lumber created from a specific process and manufacturer
of plastic lumber. Marine applications include docks or piers, bulkheads, and
boardwalks. Laboratory and field testing can determine values of the proper-
ties which are relevant in these applications. Short-term testing can also be
used to infer long-term performance of the plastics.
TECHNOLOGY DESCRIPTION
There are three main steps in producing plastic lumber: collecting the
material or feedstock, shredding and granulating, and forming into the end
product. There are aspects to the CRC process which make it somewhat unique.
These will be detailed in the individual operation descriptions.
Feedstock _
The feedstock for CRC comes from the New Stanton, California, mixed
waste MRF. This facility receives approximately 750 tons/day of municipal
solid waste (MSW). The material is sent through a large trommel which sorts
the waste by size onto conveyors. Plashes are hand picked from the larger
.-,ize fraction of MSW. The PET and natural HOPE are recovered for separate
recycling. The hand pickers are instructed to take out all other plastic
containers, plates, and objects. These are sent by conveyor to a shred-
der/granulator. CRC requires at least 70fof the plastic delivered to them be
from HOPE containers. This can be ensured through selective picking from the
conveyor.
Composition Analysis
The plastic was not washed before,it was granulated. The bottles, for
the most part, did not contain liquids^. Visually, there was only a minimal
amount of non-plastic material going to We granulator. Occasionally an
aluminum can would be transported to the granulator. After granulation, the
plastic appeared clean, even though no washing was performed. Paper labels
and contamination were removed in an air elutriation system explained below.
-------
The composition of plastics from the New Stanton facility going to CRC
was determined by a composition audit. The Quality Assurance Project Plan
(QAPjP) for this project stipulated taking ten 50-lb grab samples of the mixed
plastic destined for CRC; 11 samples were taken with a total weight of 633.8
Ibs. The samples were taken from the conveyor leading up to the plastics
granulator. All the plastics on this conveyor were picked over the course of
about 10 hours distributed over 2 days.
Plastics from the conveyor were placed in 30-gal trash cans. Each can
weighed from 5-10 Ibs when full. Five to ten trash cans were required for a
single waste composition sample. The trash cans were dumped onto a tarp where
they were sorted into 16 separate bins representing the seven resin codes
commonly used in packaging and their major product types. The bins were tared
and weighed on a digital platform balance.
Table 1.1 shows the average composition of the plastics based on the
11 samples. The relative standard deviation (RSD) is calculated according to
Section 8.0 of the QAPjP. The 90% confidence interval is calculated from the
standard deviation and the sample size. As can be seen, the RSD is greater
than 50% for most of the components. This is expected since the number of
components is large and the amount of each type is generally small. This
illustrates the heterogeneity of the MSW stream and the variability of the
plastic composition with respect to the components as sorted.
The sample size was selected in order to obtain a reasonable accuracy of
the composition with respect to HOPE, the resin with the largest overall frac-
tion. CRC specifies at least 70% of the plastic from the New Stanton MRF be
HOPE. Table 1.2 shows the composition condensed into the seven resin
categories. The RSD and 90% confidence intervals are shown. As is indicated.
the -RSD for HOPE is 12.7%. well below the QA objective of 50%. As seen/ the
RSD for the other resins is greater than 50%. This is again due to the
relatively large variation in the plastic waste stream and the small percent-
ages of these resins.
-------
TABLE 1.1 PLASTIC COMPOSITION BY TYPE OF PACKAGING
Component
PET, other
bottles and jars
PET rigid
containers
PET soda bottles
HOPE colored
blow-molded
bottles
HOPE i ejection -
molded containers
HOPE opaque blow
molded bottles
Polyvinyl chlo-
ride bottles
Polyvinyl chlo-
ride containers
LDPE composite
containers
LDPE bottles
LDPE containers
Polypropyl ene
containers
Polystyrene
containers
Polystyrene foam,
other packaging
Polystyrene food
products
Uncoded plastic
or plastic
labeled #7
Resin
1
1
1
2
2
2 '
3
3
4
4
4
5
6
6
6
7
Average
Composition
1.5%
1.2%
0.1%
44.5%
24.9%:
2.3%
0.6%
0.1%
0.6%
1.0%
0.6%
8.0%
1.0%
1.8%
1.3%
10.7%
RSD
174.2%
68.0%
262.6%
17.4%
30.9%"
165.2%
85.4%
214.3%
94.4%
258.9%
96.8%
75.4%
. 66.1%
198.7%
•45.8%
46.1%
90%
Confidence
Interval
(+\-)
1.5%
0.5%
0.1%
4.4%
4 . 4.%
2.1% "
0.3%
0.1%
0.3%
1.4%
0.3%
3.5%
0.4%
2.1%
0.3%
2.8%
Relative %
Error
(+\-)
100%
. 39%
150%
10%
18%
95%
49%
123%
54% '
148%
55%
43%
38%
114%
26%
26%
-------
TABLE 1.2 FEEDSTOCK COMPOSITION BY RESIN
Component
PET
HOPE
PVC
LDPE
Polypropylene
Polystyrene
Uncoded
Resin
1
2
3
4
5
6
7
Average
2.9%
72.1%
0.7%
2,2%
7.7%
4.2%
10.4%
90% Confidence
RSD Interval (+\-)
86.1%
12.7%
62.1%
128.6%
81.2%
97.1%
48.6%
1.4%
5.3%
0.2%
1.6%
3.6%
2.3%
2.9%
Relative %
Error (+\-)
49.3%
7.3%
35.6%
73.7%
46.5%
55.7%
27.9%
The 90% confidence interval for HOPE is 5.3%, which indicates that the
mean of the entire plastic waste stream from this MRF should be from 66.8 to
77.4% with 90% confidence. Stridtly speaking, therefore, the composition
could have less than the 70% HOPE specified. This can have a significant
Influence on the runability of this feedstock in the extruder, and the
resulting strength of the plastic lumber which is made.
Approximately 1,000 Ibs'of plastic was shipped to CRC, of which
633.8 Ibs was sampled. Of the 750 tons/day of MSW received at the MRF.
approximately 1 to 2 tons of plastic is collected and shipped to CRC. This is
less than the theoretical composition of plastic tailings in the waste.
Granulation
Shredding and granulation of the feedstock is done at the MRF. The
shredder is an Allegheny using a 50-hp motor. The granulator is made by
Formost and has a pair of 5.0-hp motors. The granulator has a nameplate capac-
ity of 1 ton/hr but reportedly has been proven to only accept 0.6 ton/hr or
approximately 14 tons/day. The plasti'c is first shredded and then cut up in a
high-speed granulator to a size of about 10 mm. The air elutriation system
uses a stream of air to separate the lighter material from heavier plastic
pieces. It forces lighter weight granulated material into a separate
container for disposal.
_
-------
Extrusion
The continuous extrusion "line" consists of a Reifenhauser twin screw
extruder; a die to form the dimensions of the plastic lumber; a vacuum
"calibrator" which uses a vacuum to pull the extruded plastic against the
sides of a perforated mold; and a wet calibrator, in which water flows over
the die to cool the plastic down, before it. finally leaves the extrusion line.
This system is unique in that it has been designed specifically for a post-
consumer feedstock. Due to the proprietary nature of this design, specific
features of the extrusion line will not be detailed.
Product Types
Products can b.e varied according to the properties and shapes required
by the marketplace. CRC can manufacture 2 in. x 6 in. and 4 in. x 4 in.
dimensional lumber with a hollow core or.a foam core. The hollow core has a
ribbed cross section w'hich provides some extra strength, without addirtg
greatly'to the weight. The core is made by using a foaming agent which essen-
tially produces a gas in the interior of the plastic lumber as it is formed.
This creates a rigid foam-like interior which lightens the overall weight
without causing a substantial loss in compressive or flexural strength.
')
The foam core lumber was chosen for this evaluation because the market
for foam core was stronger for CRC, and the development of the foaming tech-
nique was more reliable at the time of producing the plastic lumber for this
study. The hollow core is lighter but is seen to be more difficult to fasten,
as the wall thickness is about 3/16 in. The foam core achieves some weight
saving over a solid cross section, but still has enough structure to allow for
screws and nails to be used for fastening during construction. This is, seen
to be an advantage when substituting for wood without having to change
building practices.
Appli cations
The applications for this plastic, .lumber will depend on its properties,
which this study attempts to characterize. The testing protocol attempts to
give information both for general properties as well as specific ones
important in marine applications. Marine applications can include:
• Decking for boardwalks and boat docks
• Stationary and movable piers
• Bulkheads which can control beach erosion
• Sea walls
-------
• Pilings
• Pier impact protectors
• Railings for various platforms
Of these applications, some will demand higher strength material than
others. For instance, railings and deckings will require less strength and
creep resistance than will pilings and bulkheads -which will have a more
constant load placed upon them.
OTHER AVAILABLE TECHNOLOGIES
There are two major methods for making plastic lumber. The first
method is a continuous extrusion process where the molten plastic is passed
through a die and then cooled by a water spray,or bath, as used by CRC. The
second method is a molded process where the molten plastic is extruded into a
mold and then cooled in the mold. The following sections describe a few of
the available processes.
The ET-1 and Suoerwood Processes
These processes are derived from a plastic lumber system developed and
patented by Lankhorst Recycling LTD of the Netherlands. Granulated plastics
(and in some instances densified film plastics) are fed into an extruder where
they are melted. The extruder then pumps the molten plastic into a sealed
mold under low pressures, until the mold is filled and pressurized. The
filled mold then rotates down into a cooling water bath, and later back out as
other molds are sequentially filled. The finished product is air-ejected from
the molds.
An Irish company called Superwood International, LTD, paid Lankhorst
Recycling for the right to build and sell plastic lumber systems using the
concept of rotating molds. Although Superwood International went out of
business in 1991, plastic lumber companies still use the process and systems
sold to them by Superwood.
Advanced Recycling Technology, LTD (ART) of Belgium also began selling a
system with rotating molds which moved from a filling station to a submerged
cooling station to an ejection station, ART called its system the ET-1
system. After a patent' battle between ART and Lankhorst Recycling, ART signed
a licensing agreement in 1992 giving it the right to sell its ET-1 system in
several North American and European countries including the United States,
Canada, Germany, France, and England.
-------
Several generations of ET-1 systems have been designed and sold by ART.
The systems vary in their output depending on the model and processing condi-
tions, .but may range from 300 "Ibs/hr per line for older models to 700 Ibs/hr
for newer models. The Superwood systems will have outputs similar to the
older ET-1 models. This compares with an observed rate of approximately
500 Ibs/hr for the CRC foamed core material, which is made by a continuous
extrusion process.
Low-pressure molding systems such as Superwood, ET-1, and other systems
without rotating molds all produce similar solid profiles. The molding
process does not have the ability to produce hollow profiles as a continuous
extrusion system can, nor can a molded process produce very long profiles.
Internal voids are common, but can be controlled by keeping wet plastic or
foamed plastics such as expanded polystyrene out of the feedstock. Other
techniques such as pressurizing the molds with the molten plastic will
minimize the creation of voids while helping to maintain dimensional stability
upon cooling. The slow nature of the mold filling process, however, usually"
results in lumber profiles which have at least some variation in density from
one end of the profile to the other. This happens because the first plastic
to enter the mold is partially cooled by the time the final plastic enters the
.mold, so it is not easily pressurized. The variation in density will result
in profiles where one end may have more voids and less stiffness than the
other end of the profile.
Low pressure molding systems also usually produce a coarse fibrous
surface texture on their profiles. This fibrous texture is visually pleasing
since it gives the plastic lumber a wood-like appearance. The texture is a
result of the slow molding process. When molten plastic enters the mold, it
begins to solidify, beginning first with the low melting point plastics such
as low-density polyethylene (LDPE). Other plastic continues to enter the
mold, pushing the partially solidified LDPE along. As it is pushed along,
friction with the mold wall elongates it into strings, resulting in the
fibrous texture. The CRC plastic lumber formed has a smooth surface due in
part to the higher speed of the sliding motion in the molds.
The Mobil and AERT Processes
The Mobil Chemical Company and Advanced Environmental Recycling Tech-
nologies, Inc. (AERT) both make continuously extruded plastic lumber products
that contain wood fibers. Mobil purchased a company called Riverhead Milling,
acquiring the plant, process, and the product trademark Timbrex™. The
company is now known as the Composite Products Division of Mobil Chemical
Company, although the product is still referred to as Timbrex™. Details
about either the Mobil or AERT processes are difficult to obtain at thi's time.
-------
Mobil refers to its product as a wood-polymer composite. Mobil uses
mostly hardwood fiber wastes from the furniture industry for the wood fraction
in its product. The plastic fraction comes from film plastics, mostly LDPE
from grocery sacks or other sources, and may include some HOPE film as well.
Mobil does not use plastic that has been collected from the public in curbside
programs in its product. Timbrex™ contains up to 50% wood, is a very
pleasing brown color, and looks similar to wood, only without knots or grain.
Timbrex™ appears to be very solid, with good dimensional uniformity and no
voids. The controlled plastic feedstock, wood fibers, and low processing
temperature allowed by using LDPE all contribute to its good quality.
AERT also makes its product from both plastic and wood wastes, and like
Mobil, prefers to call it a composite material. AERT calls its product
MoistureShield™, and sells it to the building products industry, where it is
made into window and door sills, and drawer tracks for furniture. AERT uses
finely ground cedar wood waste fiber for the wood fraction of its product.
The plastic fraction comes from post-consumer grocery bags, HOPE containers,
ana LDPE film from recycled LDPE-coated milk cartons. MoistureShield™
appears to have a higher fraction of wood in it than Mobil's product does.
MoistureShield™ is a very pleasing brown color, and looks similar to wood,
only without knots or grain. It appears to be very solid, with good .dimen-
sional uniformity and no voids. It smells like cedar wood.
AERT has submitted six patent applications for its product processing
and film reclamation technologies. The films are compressed and forced
through a die to form dense pellets. The film plastics have not been melted
in forming the pellets, so the pellets can be broken apart. Film densifica-
tion systems used by other plastic lumber producers include heat to melt and
densify film plastics before they enter the extruder.
Comparison Between Molded and Continuous Extrusion Systems
Continuous extrusion systems take up more space than molded lumber
systems, because t-he continuous extrusion systems must have a long, straight
cooling section. As the length of the cooling section on a continuous extru-
sion system is reduced, the extruder output must also be reduced to allow time
for the extruded plastic to cool. This can reduce the greater theoretical
capacity that a continuous extrusion system can have down to the capacities of
molded systems. The time spent in cooling the extruded profile also will
affect its dimensional uniformity. Although the exterior of a continuously
extruded profile can be quickly cooled, the interior may still be hot. This
temperature gradient will result in the creation of voids, and profile faces
which contract inward toward the center of the profile as it cools. The
slower cooling rates of molded lumber systems usually help those systems
achieve greater dimensional uniformity.
10
-------
In summary, continuous extrusion systems lend themselves to producing
plastic lumber profiles which are of a wide range of lengths and-are hollow or
foamed in their interior. This is very advantageous from a weight-saving
standpoint. Molded lumber systems can produce thicker and heavier profiles,
and may be more forgiving of dirty or inconsistent feedstock. Molded systems
have in the past been generally less expensive to purchase and easier to
operate. These cost advantages may be erased as advances are made in continu-
ous extrusion technology.
11
-------
-------
SECTION II
EVALUATION OF THE TECHNOLOGY
Five separate types of plastics manufactured by CRC were tested. All
were made with a blowing agent to give them a foam core. The blowing agent is
composed of a powder which is mixed in the plastic in small quantities. As
the plastic is extruded, the blowing agent produces gas bubbles which achieve
two results: 1) the bubbles displace plastic in the core of the material,
which increases strength to weight ratio and 2) the gas forces the plastic
against the mold which then precludes the need for a vacuum in the calibrator.
The five different types of plastic manufactured for this evaluation
were 100% commingled 2 in. x 6 in., 4% polypropylene industrial regrind added,
10% polypropylene industrial regrind added, 50% HOPE industrial regrind added.
and 4 in. x 4 in. with 50% HOPE industrial regrind added. The manufacture of
the 2 in. x 6 in. material was observed.
The reason for adding polypropylene is to evaluate its effect on
improving the modulus of elasticity. It has been found that adding up to 35%
polystyrene can improve the modulus dramatically. Polystyrene, however, is
more difficult to collect since it is usually in foam form and thus not
thought to be-as ;practical an additive as polypropylene. Adding 50% HOPE can
assure a more consistent and marketable product, as this material is clean and
consistent.
Table 2.1/fescribes testing methods.
TABLE 2.1 TESTING METHODS
Parameter
Holding
strength of
nails and
screws
Method
tfethod Method Title Type
^ASTM D1761 Standard Test Methods for Mechanical
f Mechanical Fasteners in
t Wood
*
Source
Reference
ASTMd) .
i
13
-------
TABLE 2.1 TESTING METHODS (cont.)
Parameter
Method
Method Title
Method
Type
Source
Reference
Resistance-to
deformation by
flexural
stress
Strength of
the profiles
when force is
applied in
compression
Strength of
profiles in
bending
Friction of
the profiles,
while wet and
dry
Loss of
surface layers
due to UV
Bioassay of
th'e profiles
Biofouling of
the profiles
ASTM D2990 Standard Test Methods for
Tensile, Compressive and
Flexural Creep and Creep-
Rupture of Plastics
ASTM D695 Standard Test Method for
Compressive Properties of
Rigid Plastics
ASTM D790 Standard Test Methods for
Flexural Properties of
Unreinforced and
Reinforced Plastics and
Electrical Insulating
Materials
ASTM D4521 Standard Test Method for
Coefficient of Static
Friction of Corrugated
and Solid Fiberboard
ASTM G53 Standard Practice for
Operating Light- and
Water-Exposure Apparatus
(Fluorescent UV-Condensa-
tion Type) for Exposure
of Nonmetallic Materials
EPA 1008 Sea Urchin Fertilization
and Embryolarval Test
ASTM D3623 Standard Method for
Testing Antifoul ing
Panels in Shallow
Submergence
Mechanical
Mechanical
Mechanical ASTM(1)
Mechanical
Light
Biological
Biological ASTM
-------
A series of tests as set forth in the QAPjP was performed on the plastic
made at CRC. The tests can be divided into two groups: physical testing and
environmental testing. Results from these tests are given below.
PHYSICAL TESTING
The physical tests used were standard ASTM tests. Charles R. Lockert,
P.E., of the Florida Institute of Technology Department of Oceanographic Engi-
neering performed the tests under the supervision of Gary Zarillo. Ph.D. A
summary of each test is given along with results of the tests.
Creep Tests
Creep tests are designed specifically for measuring long-term movement
in plastic materials due to a continuous force placed on the materials. If
one places a weight on the center of a span of plastic lumber supported only
on two ends, the material will flex somewhat initially. Over time this flex
will increase due to the unraveling of the polymeric entanglements of the'
plastic. This phenomenon is called creep and can be Important in certain
applications of plastic lumber.
ASTM D 2990, "Standard Test Methods for Tensile. Compressive, and
Flexural Creep and Creep-Rupture of Plastics." was used and the tests measured
the change in deflection of the plastic lumber over a period of 1,000 hrs.
Sample Preparation
Thirty-two-inch-long samples were cut from nominal 2 in, x 6 in.'s using
an electric miter saw. Samples were placed in test jigs under no load, but at
test temperature, for a minimum of 72 hours prior to loading.
Test Conditions
Room temperature tests were conducted with the temperature varying
between 74-76°F and humidity between 65-67%. Elevated temperature of 140°F
was conducted between 139-141°F and a humidity of 100% (submerged in heated
water).
Apparatus--
The creep test jigs were constructed out of readily available materials.
Figure 2.1 shows a schematic of a single test jig. The 2 in. x 3 in. steel
tubing is fastened to a fixed stand, and the middle 1/2 in. x 30 in. section
can move freely through this tubing. The basic idea is to apply a force to
15
-------
the center of a 24 in. span of 2 in. x 6 in. plastic. The deflection is then
measured with a digital caliper at set intervals of time. For the higher
temperature tests, the plastic is immersed in a hot water bath which is
controlled to 140°F. A more extensive description is given in the ASTM
procedure.
Verification of Creep Test Jias with Delrin
Verification of the creep test jig was accomplished by testing a virgin
material and comparing test results with those published by the manufacturer.
E. I. DuPont's Delrin 150 was chosen for the test. A Delrin rod 1-1/2 in. in
diameter was cut to a length of 32 in. Three rods under an identical load
were tested simultaneously. Figure 2.2 shows a comparison between the Delrin
mean test results and- the published results at 1,250 psi. The expected Delrin
results were taken from Modern Plastics, "Guide to Plastics." 1985.
1/2" elbow
digital calipers
1/2" x 8" nippje
£|
sand bae . L . J ' sand has .
1/2" PVC snap-on tee
1/2" x 6" niple
4" PVC concrete filled
." x 3" steel tubing
1/2" x 30" nipple
2" x 4" test sample
END VIEW
Figure 2.1 Individual sample in cut away of dry test jig.
16
-------
Stress 1250 psi
Test Mean
Predicted
400 600 800
DURATION (MRS)
Figure 2.2 Delrin predicted vs. mean test values.
1,000
1,200
The small variation between test samples, as well as the 5% or less
error between predicted and test values, provided assurance the test jig and
accompanying procedure were reliable.
Results
The results of these tests are shown in plots of the apparent modulus of
elasticity versus time for each plastic type, loading, and temperature. The
apparent modulus in pounds per square inch (psi), is given in the following
equation:
£ =
PL3
48A/
where:
E
P
L
A
is the apparent modulus in psi
is the constant load in Ibs
is the distance in in. between supports
is the deflection of the sample in in.
is the moment of inertia in in^
By plotting the log of the modulus versus the log of time, the data
points form a relatively straight line in most cases. From these plots
comparisons can be made between the different feedstock "recipes" used to make
the plastic lumber. The creep behavior can also be predicted over longer
17
-------
periods of time by extrapolating the plots beyond the time examined in the
tests.
Figures 2.3 to 2.10 show the results of the creep tests under room and
elevated temperatures. The elevated temperature runs were conducted at 140°F.
Due to the variability of the materials it is shown that the creep is not as
strongly correlated with weight as it should be. That is, the material vari-
ability from one piece to another is likely greater than the change in creep
due to a different force being applied to the material. Figure 2.11 compares
the four plastic formulations for a 100-lb load at room temperature.
i
8
100% COMMINGLED
ROOM TEMPERATURE
5 T
60 pound load
100 pound load
140 pound load
180 pound load
0 1
LOG TIME (HRS)
F'igure 2.3 Creep test results - commingled plastic at room temperature.
18
-------
o
o
-2
-1
100% COMMINGLED
ELEVATED TEMPERATURE
5 -•
4.9 •
4.8
4.7 •
4.6 •
4.5-
4.4
0 1
LOG TIME (HRS)
60 pound load
100 pound load
"* 140 pound load
180 pound load
Figure 2.4 Creep test results - commingled plastic at elevated temperature,
o
O
-2 -1.5
4% POLV^OPYLENE
ROOM TEMPERATURE
5T
-0.5
0 0.5 1 1.5
LOG TIME (HRS)
60 pound load
100 pound load
"* 140 pound load
ISOooundload
2.5
Figure 2.5 Cr% test results - 4% polypropylene plastic at room
frature.
3.5
19
-------
8
4% POLYPROPYLENE
ELEVATED TEMPERATURE
5 --
4.9 • •
4.8 --
4.7 ••
4.6 -•
4.5--
4.4
-2 -1.5 -1 -0.5 0 0.5 . 1
LOG TIME (MRS)
~" 60 pound load
100 pound load
"* 140 pound load
•° 180 pound load
1.5
2.5
Figure 2.6 Creep test results - 4% polypropylene plastic at elevated
temperature.
CO
8
10 % POLYPROPYLENE
ROOM TEMPERATURE
5-r
4.9 -
4.8 ••
"" 60 pound load
100 pound load
140 pound load
180 pound load
-1.5 -1 -0.5
0.5 1 1.5
LOG TIME (MRS)
2.5
3.5
Figure 2.7 Creep test results - 10% polypropylene plastic at room
temperature.
20
-------
I
-1
10% POLYPROPYLENE
ELEVATED TEMPERATURE
5y
4.9-
4.8
4.7 •
4.6-
4.5-
4.4-
0 1
LOG TIME (HRS)
60 pound load
100 pound load
140 pound load
180 pound load
Figure 2.8 Creep test results - 10% polypropylene plastic at elevated
temperature.
to
2
in
o
O
2
-2
-1
50% HOPE
ROOM TEMPERATURE
60 pound load
lOOpouna load
140 pound load
180 pound load
1
LOG TIME (HRS)
Figure 2.9 Creep test results - 50% HOPE plastic at room temperature.
21
-------
53
8>
tn
I
-1
50% HOPE
ELEVATED TEMPERATURE
5j
4.9-
4.8 ••
0 1
LOG TIME (HRS)
' 60 pound load
100 pound load
140 pound load
180 pound load
Figure 2.10 Creep test results - 50% HOPE plastic at elevated temperature.
LOG MODULUS (PSI)
-1
FORMULATION EFFECT ON CREEP
ROOM TEMPERATURE, 100 LB LOAD
4.9 j
4.8 -.
0 1
LOG TIME (HRS)
"• COMMINGLED
1 4% POLYPROPYLENE
'* 10% POLYPROPYLENE
50% HOPE
Figure 2.11- Comparison of four plastic formulations - 100-lb load at room
temperature.
22
-------
Flexure Tests
Flexural modulus is a measure of the stiffness of a material. It is
important in that it will determine the span length of supports underneath
decking among other things. This length is generally about 24 in. for 2 in. x
6 in. wood but will be expected to be less for the more flexible plastic
studied here. Flexural testing followed the ASTM method D 790, "Standard Test
Methods for Flexural Properties of Unreinforced and Reinforced Plastics and
Electrical Insulating Materials." Tests were performed on 2 in. x 6 in.
lumber only.
Basic test procedure--
Flexure tests were performed using the. Baldwin universal test machine in
simply supported three point bending as shown- in Figure 2.12. This machine is
designed to apply a load on a material by gradually compressing it between a
moving .plate and a fixed plate. A maximum deflection of 3 in. is allowed with
this test apparatus. None of the samples failed prior to bottoming out on the
jig. As such, no ultimate strength could be calculated. Increasing the
strain rate from 0.01 in./in./minute (used in testing) to 0.1 in./in./minute
still did not cause failure. As a 3 in. deflection is greater than the 5%
maximum strain in the outer fibers that ASTM allows, increasing the -maximum
allowable deflection would be of no use. The Baldwin load cell, an instrument
which measures the force applied to the material, measured load while a
digital extension indicator measured deflection. A corrected load and deflec-
tion were calculated to account for the preloading to ensure the sample was
tight against the supports and loading nose. The secant modulus of elasticity
was calculated between each data point and the beginning data point to best
reflect a comparable value with wood. The secant modulus is the load divided
by the deflection.
Loading Nose
Support
Figure 2.12 Flexure test jig.
23
-------
Results--
The flex test, as indicated, did not show the ultimate strength of the
material, as the material was elastic enough not to fail within the specified
range of deflection of 3 in. Flexural modulus was calculated. Generally the
tangent modulus calculated is at the first part of the stress versus strain
curve, where the slope is constant. For the plastic, the initial portion of
the curve was not linear. Therefore the secant modulus for each part of the
curve was calculated and found to vary linearly with stress. Secant modulus
is the absolute stress divided by the strain, instead of the change in stress
divided by the change in strain. The secant modulus versus load was linearly
regressed to yield an equation for each test.
The secant modulus equations allow calculation of strains and deflec-
tions for loads in the expected service environment of up to 500 Ibs. The
coefficient and constant statistics of these equations are summarized for the
four 2 in. x 6 in. plastic compositions in Table 2.2.
TABLE 2.2 SECANT MODULUS EQUATIONS STATISTICS
Sample
RSD RSD
Coefficient Constant Coefficient Constant Completeness
100% Commingled
4% Polypropylene
10% Polypropylene
50% HOPE
-88
-106
-73
-85
81.717
86,951
79.319
92.636
6.8%
2.5%
13.5%
9.8%
7.1%
3.6%
3.8%
4.6%
95%
90%
90%
90%
A graph of the four average equations is given in Figure 2.13. This
demonstrates the higher modulus of the 50% HOPE formulation. Also to be noted
is the greater stiffness of the 10% polypropylene plastic lumber at higher
loadings. A single factor analysis of variance (ANOVA) performed on the data
demonstrated the distinct statistical difference in flexural modulus of the
four formulations.
24
-------
FLEX MODULUS VERSUS LOAD
• 100% COMMINGLED
• 4% POLYPROPYLENE
• 10* POLYPROPYLENE
•50% HOPE
100
ISO
200
250
LOAD (IBS)
300
350
400
450
500
Figure 2.13 Plot of flex modulus for the four formulations of RPM.
Compression Tests
Compression testing is important in determining the structural charac-
teristics of the plastic lumber. A structure composed entirely of plastic
lumber will be under compression, tension, and flexural force at various
points due to the loads caused by the weight of the structure and loads from
external forces and weights. Therefore, a reasonable structural design will
require values for these parameters, including variability of the material.
The tensional strength was not measured in this study; however, as it is going
to play a smaller role in design of marine applications, it can be estimated
using flex and compression data if required. The ASTM test D 695*.
"Compressive Properties of Rigid Plastics," was followed.
Basic test procedure--
Samples were placed between the bed and lower cross arm of the Baldwin
universal test machine. The bed and cross arm provided flat and parallel
surfaces. No provisions for pivoting of the sample were made. The sample was
compressed at a cross head rate of 0.01 in. per minute. Deflection of the
sample was recorded using a Mitutoyo Digimatic Indicator. The extension
indicator and load indicator were interfaced with an IBM PC that recorded load
and deflection data at approximately 500-lb intervals throughout the test.
The cross head movement was continuous until sample failure.
The maximum load prior to failure was used to calculate ultimate
.strength. A corrected load and displacement were calculated to account for
the 200- to 300-lb preload placed on the sample to ensure the sample was tight
with the bed and cross arm. Secant modulus of elasticity was calculated
25
-------
between adjacent data points throughout the test. An average modulus of
elasticity was calculated for all like samples by averaging the maximum indi
vidual "secant modulus of each sample. Figure 2.14 below Indicates the
arrangement of the test jig used.
Digimatic
Indicator
oo
/ V
Sample
Figure 2.14 Compression test jig.
Results
Results from the tests are summarized in Table 2.3. It appears that the
samples with 50% industrial regrind HOPE show the smallest relative standard
deviation from the average. This is expected, since the HOPE should be very
consistent in its characteristics. It also looks like polypropylene addition
does not result in significant increases in compressive strength. This is
verified by ANOVA which does not show a statistically significant difference
between the 100% commingled and 10% polypropylene formulations.
26
-------
TABLE 2.3 COMPRESSIVE MODULUS OF RPM
Sample Type
100% Commingled
4% Polypropylene
10% Polypropylene
50% HOPE
50% HOPE 4 in. x
4 in.
Number
19
8
10
• 10
10
Complete-
ness
95%
89%
100%
100%
100%
Average
Compressive Standard
Modulus
(psi)
44,942
42,163
48,713
55,150
33,697
Deviation
(psi)
7,315
4,788
7,874
3,743
3,449
RSD
16%
11%
16%
7%
11%
90% Confi-
dence
Interval
. (+.V-)
2,910
3,208
4,882
2,508
3,288
Static Friction Tests
Static friction testing ca'n give an indication of how well the plastic
lumber will function as a decking material. ASTM D.4521, "Coefficient of
Static Friction of Corrugated and Solid Fiberfaoard." was performed on both
plastic and CCA-treated wood for comparison. Wet testing was performed as
well as dry to eval uate appl.i cati.ons where surfaces may be periodically wet.
Basic test procedure--
Two -inch x 6 inch x 6 inch samples were clamped to the inclined plane as
-shown in Figure 2.15 below. The inclined plane is raised at a rate of
0.5 degree per second until the shoe or block placed on the sample begins to
slide. The inclined plane is raised by the cross-arm of the Baldwin test
machine. This provided a very smooth and controlled rate of lift. When the
block or shoe begins to slide, inclining is stopped and the angle of incline
is noted. The angle of incline is provided by the protractor placed at the
pivot point of the incline plane. The test is repeated three times and the
angle of incline on the third lift is used to calculate the coefficient. The
coefficient of static friction is equal to the tangent of the incline angle a-t
which sliding begins.
Plastic samples and CCA-treated wood samples were tested. The first
test of each sample was determining a coefficient of friction between the
sample and itself. This involved sliding a 2 in. x 6 in. x 6 in. block of the
same sample material and determining the coefficient of static friction. Then
a tennis shoe was placed on the sample and a coefficient between the sample
and tennis shoe was determined. The procedure is repeated for a rubber-soled
27
-------
boat shoe. All three tests, sample block, tennis shoe, and boat shoe are
repeated under wet or moist conditions. Water from a spray bottle is squirted
on the sample face and the bottom soles of the shoes. Sufficient water is
used to ensure saturation, evident by standing water on the sample surfaces.
The plane is inclined three times as before with fresh water being applied to
the sample between lifts. Fresh water is used to replace the water that has
run off as the result of being inclined. The angle of the third lift is used
in calculations.
The same procedure was used with the wooden samples using a wooden
block, and the same tennis and boat shoe. After all the dry runs, water was
squirted on the samples to simulate wet, rainy conditions. It should be
pointed out, no water.was squirted on the sample during the actual inclination
process. It was placed 'only o'n the horizontal sample before each lift.
Inclined plane
Figure 2.15 Inclined plane friction test jig.
Results
Table 2.4 shows a summary of the test results for plastic and wood
samples. The coefficient of friction statistics for the material tested
against itself, a used and a new tennis shoe, and a used and a new boating
shoe were determined. A new Sperry Topsider™ boat shoe and a new Converse
Chuck Taylor™ tennis shoe were used. Only 100% commingled plastic was
28
-------
tested, since a substantial difference was not expected for the other plastic
compositions.
The results show that the plastic is significantly more slippery than
wood, especially under wet conditions. It appears that the old tennis and
boat shoes show an increase in the coefficient of friction with a wet surface.
However, this phenomenon does not occur with new tennis and boat shoes. This
has been verified statistically by ANOVA. Also for plastic, the tennis shoe
coefficient is always greater than the boat shoe, while for wood, this is only
true for the new shoe tests.
TABLE 2,4 COEFFICIENT OF FRICTION USING MATERIAL AGAINST ITSELF, OLD AND NEW
TENNIS AND BOAT SHOES
Sample
Dry Plastic
Plastic '
01 d tennis shoe
New tennis shoe
Old boat shoe •
New boat shoe
Wet Plastic
Plastic
01 d tenni s- shoe
New tennis shoe
Old boat shoe
Mew boat shoe
Dry Wood
Wood
Old tennis shoe
New tennis shoe
Old boat shoe
New boat shoe
Wet Wood
Wood
Old tennis shoe
New tennis shoe
Old boat shoe
New boat shoe
.'. Average Coeff,
0.222
0.583
0.8
0.541
Or6
0.241 ..
0.600
0.7
0,563
'•0.545
0.631
0.700
0.933
0.862
0.876
0.948
0.872
0.804
0.994
0.785
STD Deviation
0.012
0.014
0.019
0.014
0.031
0.018
0.023
0.011
0.576
0.022
0.138
0.092
0.015
0.064
0.016
0,113
0.021
0.015
0.035
0.029
RSD
=>%
2%
2%
2%
5%
7%
4%
2%-
4%
4%
22%
13%
2%
7%
2%
12%
2%
2%
3%
4%
29
-------
Mechanical Fasteners in Plastic Lumber
Testing the ability of the plastic lumber to hold nails and screws is
important in ascertaining whether the plastic can be fastened using conven-
tional nails and screws. ASTM D 1761, "Mechanical Fasteners in Wood," was
followed. • - .
Nail: #16d galvanized deck nails were driven manually using a carpen-
ter's claw hammer. One nail was driven into each sample. Nails were driven
in the geometric centriod of the 3.5 in. x 5.5 in. face and the 1.5 in. x
3.5 in. face. All nails were driven into to an uncut manufactured surface.
The nails were driven to a penetration depth of 2.2 in. The 2.2 in. penetra-
tion in the 5.5 in. face therefore penetrated both the top and bottom faces of
the sample. Nails remained in untested samples for 15 to 22 days to allow
relaxation of the plastic around the nails and screws.
Screw: One #10 galvanized deck screw was placed in each sample. Screws
were placed in the center of the 5.5 in. face and the 1.5 in. face. Pilot
holes were drilled to one-half the screw penetration depth and with a diameter
70% of the screw root diameter. Pilot holes were drilled and the screws were
inserted using an electric drill press. Screws were driven to a penetration
depth of 2.4 in. The 5.5 in. face screw penetrated both the top and bottom
faces. Screws remained in the untested samples for 15 to 22 days.
Basic test procedure--
The maximum load in pounds required to pull out the nails and screws
vertically was determined with the use of a Baldwin universal test machine.
Figure 2.16 below describes the test jig and setup used in both cases.
The bottom cross arm remained stationary while the top cross arnrmoved
upward. The rate of cross arm movement was 0.1 in. per minute. The maximum
load during removal of the fastener was recorded.
30
-------
Angle Iron
Nail/Screw
Sample
Figure 2.16 Nail and screw pullout test jig.
Results ........... .
The results indicate that the resistance to nail pullout, as shown in
Table 2.5, i.s somewhat variable and not well correlated with the composition.
In fact the 100% commingled and the 50% industrial regrind HOPE added have
better average pullout resistance than the polypropylene samples. The
standard deviations of data suggest, however, that the samples may not be
.significantly different in pullout resistance. This has been verified by a
single factor ANOVA on the edge-driven nails. However, the face pullout tests
were significantly different according to the statistical ANOVA test.
It is also noted that the nails going through the 5.5 in. face appear to
be more difficult to remove than through the edge. This is not corroborated
statistically by ANOVA.
31
-------
TABLE 2.5 NAIL PULLOUT TEST RESULTS SUMMARY
Sample
100%
Commingled
4%
Polypropylene
10%
Polypropylene
50%
HOPE
Side
Edge
Face
Edge
Face
Edge
Face
Edge
Face
Average
Nail
Pullout
(Ibs)
264.6
241.8
209.6
257.6
. 213.6
234.0
276.4
334.6
STD
Deviation
(Ibs)
53.7
56.4
42.8
28.3
65.4
43.0
89.0
81.8
RSD
20%
23%
20%
11%
31%
18%
32%
24%
90% Confidence
Interval Completeness
(Ibs)
295.7
274.5
236.1
274.0
251.5
260.7
328.0
382.0
233.5 '
209.1
183.0
241.2
175.7
207.3
224.8
287.2
100%
100%
90%
100%
100%
90%
100%
100%
The screw pullout test results are shown in Table 2.6. In these tests,
the face-directed screws are less strongly held than the edge screws for three
of the four composition types. This is corroborated by 2 factor ANOVA which
indicates significant differences between the different formulations and the
different angles of penetration. It could mean that more of the threads see
plastic with the edge-driven screws, since the screws do not penetrate all the
way through from this angle.
32
-------
TABLE 2.6 SCREW PULLOUT TEST RESULTS SUMMARY
Average
Screw STD 90% Confidence
Sample Side Pullout Deviation RSD. Interval Completeness
(Ibs) (ibs) (Ibs)
100%
Commingled
4%
Polypropylene
10%
Polypropyl ene
50%
HOPE
Edge
Face
Edge
Face
Edge
Face
Edge
Face
629.4
511.4
' 478.4
524.4
625.4
472.0
694.2
605.8
81.7
38.3
89.7
60.9
76.4
80.3
190.4
72.1
13%
7%
19%
12%
12%
17%
27%
12%
676.8
533.6
530.4
559.7
669.7
518.6
804.6
647.6
582.0
489.2
426.4
489.1
581.1
425.4
583.8
564.0
100%
100%
100%
100%
100%
100%
100%
100%
UV TESTS
ASTM G-53, "Operating Light and Water Exposure Apparatus (Fluorescent UV
Condensation Type) for Exposure of Nonmetallic Materials," was performed to
test the effect of UV light on the plastic. The effect studied was loss of
thickness which would be caused by further polymerization and oxidation of the
plastic lumber surface. The experiment was designed to accelerate ambient
conditions by using UV lamps.
Forty samples were prepared with ten samples coming from types 2A, 2C.
2D, and 3A materi al .
Conditions
Samples were exposed to a UVCON Ultra Violet/Condensation Screening
Device utilizing eight UVA-340 sun lamps with peak emission at 340 nm and
range between 365 nm and 295 nm (Figure 2.17). New sunlamps were installed at
the beginning of the tests. After 400 hrs, one lamp was replaced in each bank
of four and the others were rotated as outlined in the ASTM test and shown in
33
-------
Figure 2.19. Samples were exposed in a cycle of 8 hrs of exposure followed by
4 hrs off non-exposure. This resulted in a sample exposure of 16 hrs a day.
Samples were rotated every 2 to 3 days using the horizontal rotation method
recommended by ASTM (Figure 2.18). During the exposure cycle, the test
chamber temperature was maintained at 50°C, while it returned to room tempera-
ture during off cycles. No condensation testing was performed. Samples were
exposed to the UV/Elevated Temperature combination for a total of 544 hrs.
Specimen
Fluorescent UV
Lamp
Figure 2.17 UV test chamber cross section.
Basic test procedure--
Samples were cut and allowed to sit in the air-conditioned lab for
48 hrs. The samples were then measured using digital calipers. The width and
thickness of each sample was recorded to the nearest thousandth of an inch.
This measurement was actually an average of several readings along the sample
length. Although the samples were cut square, the release of internal
stresses upon cutting coupled with the non-square manufactured face produced
samples with varying width and thickness. The variation was +0.007 in.
Samples were then placed into the UV chamber and exposed to UV and a tempera-
ture of 50°C for 8 hrs and then 4 hrs of no UV and room temperature.
Every 2 to 3 days the samples were rotated horizontally as outlined in
the ASTM test and shown in Figure 2.18 below. After 400 hrs, two of the sun-
lamps were replaced and the others were rotated as outlined in the ASTM test
(Figure 2.19). After a total UV/Elevated Temperature combination of 544 hrs,
the samples were removed and allowed to cool to room temperature for 72 hrs.
At the end of 72 hrs the samples were measured using the same digital calipers
to determine any change of dimensions.
34
-------
Mounted Sampl
Figure 2.18 Horizontal sample rotation.
Discard
Discard
Figure 2.19 Lamp replacement and rotation.
Results
No change in dimensions was detected after exposure that was outside of
the ±0.007 in. error. The internal stress release straining coupled with the
curved manufactured surfaces makes precise measurements of the samples
impossible.
ENVIRONMENTAL TESTING
Biofouling Tests
Biofouling of panels is generally directed at coatings which resist the
attack of biological organisms in the sea. In this evaluation, the test was
used to demonstrate the compatibility of the plastics with marine life.'
35
-------
Btofouling is thus a measure of the absence of toxicity in the plastic lumber
as compared with untreated lumber controls. ASTM D 3623, "Antifouling Panels
in Shallow Submergence," was followed. The test was performed from 8/28/92 to
12/17/92 in the Florida Institute of Technology (FIT) Exposure Platform,
Indian River Lagoon, Grant, Florida.
Sample preparation--
Two inch x 6 in. x 18 in. long samples were cut using an electric miter
saw. A 1/4 in. hole was drilled in each corner of the samples to attach the
sample to exposure frames with the use of nylon tie strips.
Test conditions--
Samples were attached to a PVC exposure frame which was, in turn,
submerged to a minimum depth of 10 in. to 12 in. The exposure platform is in
the middle of the Indian River Lagoon with a surrounding water depth of 8 to
10 ft. The lagoon has direct access to the ocean through Sabastian Inlet
approximately 10 miles south. The lagoon contains brackish water with all the
biofouling organisms normally associated with coastal biofouling in this area.
Test procedure--
As shown in Figure 2.20, PVC frames were constructed and plastic lumber
samples were attached to these frames with the use of nylon tie strips. The
samples were arranged so that at least a 1/2 in. gap remained between adjacent
samples and between the PVC frame. The frames had two levels of samples. The
top level held the top on the sample 10 in. to 12 in. below the water surface
while the bottom level had the top of the samples approximately 3 ft below' the
surface. Four different types of plastic lumber were tested as well as
untreated southern pine. Five samples of each type were exposed. The five
samples were arranged on four exposure frames, such that every type had three
samples in the top level and two samples in the lower level. Photographs of
the samples and frames were made prior to exposure and approximately every 30
days for a total exposure time of 90 days. In addition to the photographs,
estimations as to the type, percent coverage, and number of the various
biofouling organisms were recorded for the same 5.5 in. x 18 in. face on each
inspection date.
36
-------
Bottom
Water
Bottom
Water Level
FRAME4
Figure 2.20 Biofouling test frames and sample layout.
Results
Results from this test are summarized in Table 2.7.
Figure 2.21 shows the results of Exposure Frame 1 which contains all of
the plastic compositions and one 2 in. x 6 in. untreated wood length.
The wood length is on the left portion of the frame. It is seen from
the table and figure that the wood panel has a similar degree of fouling, but
does not have the barnacles which cover the plastic samples. This may be due
to the surface characteristics of the plastic, extractives from the wood which
irritate the barnacles, or other phenomena. In any event, it appears that the
plastic is easily biofouled and does not deteriorate in the short span of this
test.
37
-------
o
o.
o
o
C9
C£
O
z
in
01
c
CO
V)
0»
c in M
O 3 O
O t-
C CO
O)
<
C3
o
CO
a> en
«- i—
o <
(1)
Ol
01
—
<:
0>
O!
o
o
CO
cr>
o>
s
0.
o
Q.
38
-------
Figure 2.21 Top portion of exposure frame after 111 days of exposure.
SEA URCHIN FERTILIZATION TEST
This test is used to characterize the toxicity of plastic with respect
to a sensitive sea animal, the Sea Urchin. The bioassay followed was EPA Test
Method 1008, Sea Urchin Fertilization. In addition, longer term monitoring of
the larval growth of the fertilized gametes was also performed.
Sample Preparation
A piece of RPL 1-1/2 in. x 5-1/2 in. x 12 in. (representing a standard
piece of "dimensional" 2 in. x 6 in. lumber) was received in June. The
material submitted for testing was prepared on Nov. 2 as follows (the end
product included the original surface):
• Four 1 x 1 cm strips were cut (by radial table saw) from one edge
along the 12 in. length.
• Each strip was cut into 1 x 1 x 0.5 cm blocks using a band saw.
• One 1 x 1 x 10 mm strip was cut from each block with a stainless
steel razor blade.
39
-------
• The 1 x 1 x 10 mm strips were cut down to:
-- Size A: Ixlxl, yielding 6 mm2 surface area
-- Size B.: 1x1x2, yielding 1.0 mm2 surface area
-- Size C: 1 x 1 x 4.5, yielding 20 mm2 surface area
Thus, each piece has its longest surface from the original surface 'of
the plank, each piece consists of 1 mm depth from that surface, and the
sampling is spread out over the entire length of the plank. See Figure 2.22.
These prepared pieces were stored in glass vials at room temperature
(24°C) and under fluorescent lamps (=* 9 hrs/day).
OBTAINING SAMPLES
FOB LEACHING TESTS
,5cm
1x1x1
Original surface
>• 01x1x2
\ OJ 1x1x4.5
Scrap
Figure 2.22 Plastic preparation for leaching testing.
Leaching Procedure
Sea water was collected from Gardiner's Bay (Eastern Long Island, New
York) on the incoming tide. The water was transported .*'n seasoned (well-used)
40
-------
polyethylene carboys to the laboratory and stored cold (4°C) and aerated. For
48 hrs prior to use, this water was stirred with activated charcoal in a glass
carboy at 24°C. The leaching was performed for one week (it has been demon-
strated that the results of the assays were similar with leaching for 1, 2,
and 3 weeks - Weis, et al ., 1992). The leaching was done in standard vials
(liquid scintillation vials, borosilicate glass, Kimball Corp.) at 20° ±0.1°C,
dark, in an incubator, with twice daily hand agitation. The materials, water
volumes, and ratios are presented in Table 2.8.
TABLE 2.8 LEACHING PROTOCOL
RPL
Size
Size
Size
Size
Size
Piece
A
A
B
C
C
Dimensions
(mm)
1
1
1
1
2 of 1
X 1 X
X 1 X
X 1 X
X 1 X
X 1 X
1
1
2
4.5
4.5
Surface Area
(mm2)
6
6
10
20
20 x 2
Water
Vol ume
(L)
20
10
10
10
10
•Surface/
Volume Ratio
(cm2/L)
3
6
10
20
40
The lowest surface/volume ratio used was the result of having the
smallest replicable unit of plastic that could be used, and the maximum volume
of water that could be contained in the standard test vials required by the
EPA method. This ratio was two to three times that of the maximum environ-
mental exposure found in completely developed residential canals on the
Atlantic and Gulf Coasts.
SEA URCHIN PREPARATION
QA Procedures
The quality assurance procedure involves testing fertilization with
0-4.0 mg/L of detergent dodecyl sulfate (SDS) added to the test vials. Probit
analysis, using a BASIC program included in the EPA manual, indicated the EC5Q
(50% fertilization rate for a determined concentration of SDS) and spontaneous
response rates (SRR) (level of fertilization below which a toxic effect is
measured) both with 95% confidence intervals (CIg5).
41
-------
Results
For the first assay, the fertilization success was 74.2% ±8.41 for the
controls. Probit analysis demonstrated an £€50 for SDS at 1.55 (1.12 to
1.99 CIgs) and an SRR of 70-85%. All the leachates had fertilization
successes of 82-91% better than controls and clustered about the predicted
upper limit of the SRR. It is noted that the plastic leachate causes a small
rise in fertilization over the controls. This indicates hornesis, a low level
of toxicity increasing metabolic rates. These data suggest no effect of
plastic on fertilization.
For the second assay, the fertilization success was 90.4 ±5.32% for the
controls. Probit analysis demonstrated an EC$Q of 1.46 (1.16-1.79 0195) and
an SRR of 91.4 (86.1 to 96.6 CIgs). The RPL-exposed larvae were again
unaffected, all the fertilization rates were within the SRR.
On the thir'd assay, since the lowest SDS reference exposure and all of
the RPL exposures yielded >90% fertilization, the fertilization results of
control replicates A. B, and C were interpreted as due to contamination of the
vials. With this massaging of data, probit analysis gave an EC$Q of 2.25 mg/L
SDS (1.91 to 2.63 CIgs) and an SRR of 88-9 to 96-3%- A11 the RPL exposures
were within this SRR, indicating no effect.
In three assays of RPL, there were no acute effects (fertilization
success), even at the highest ratio of surface area to water volume. A
summary of the assay test data is presented in Table 2.9.
TABLE 2.9 SUMMARY OF ASSAY TEST DATA
Surface/Volume
SRR (cm2/L)
Control (0)
3
6
10
20
40
Assay 1
77.2
74.2
82.3
84.0
84.3
89.3
91.0
Assay 2
91
90.4
91.8
9.0
91.6
94.8
93.0
Assay 3
92.6
94.5
95.3
93.8
94.3
95.3
95.8
42
-------
SECTION III
MATERIAL COMPARISON
PHYSICAL COMPARISON
Comparisons with wood can be made by using values for wood properties
found in the literature. A summary of various properties is given in
Table 3.1. Included are comparisons for flexural modulus, compressive
ultimate strength, and nail and screw pullout resistance.
TABLE 3.1 COMPARISON OF PROPERTIES BETWEEN PLASTIC AND WOOD
Property
Plastic
Wood
Flexural-Modulus
Compressive Modulus
Ultimate Strength (psi)
Nail Pullout Strength (Ibs)
Screw Pullout Strength (Ibs!
80,000-90,000
,• 1,000-1,200
200-300
500-600
800,000-1,700,000(1)
800-1,900 (parallel to grain)
600 (perpendicular to grain)(1^
=100(2)
(1) Western Wood Products'Association.
(2) Uniform Building Code, 1SJ58.
From this comparison, it is apparent that the flexural modulus is the
limiting property of the CRC plastic lumber. The ultimate strength of the
plastic is similar to wood. Surprisingly, the nail and screw pulTout strength
is greater for plastic than wood. This does not, however, take into consid-
eration elevated teferatures and the effect of creep over time which affects
plastic more severely than wood.
43
-------
LIMITED LIFE CYCLE ANALYSIS OF RECYCLED PLASTIC LUMBER
AND CCA-TREATED LUMBER
Methodology
A life cycle analysis (LCA) examines all the costs associated with a
specific material or product through its entire useful life. These "cradle to
grave" analyses attempt to quantify the energy used and environmental impacts
resulting from:.
• Acquiring the raw material.
• Transporting it to the point of manufacture'.
• Processing the raw material into a product.
• Transporting the product to the point of use.
• Disposal of the product after its useful life is complete.
The difficulty in doing these analyses stems from a number of factors,
including:
• Too much detail - the analysis can suffer due to trying to incor-
porate too much data, especially data which are unreliable or
i'naccurate.
• Not enough detail - leaving out important energy uses or environ-
mental impacts.
• Unfair methodology - the methodology may unfairly sway the results
due to assumptions used or the level of detail agreed upon.
• Inability to accurately convert environmental impacts into
economic costs, such as loss of natural resources (like fossil
fuel or minerals) and land use (opportunity cost).
i
The limited LCA attempted in this analysistavoided some of these
pitfalls, but at the expense of yielding an absolute result. That is, the
analysis which follows is meant to look at the major energy and environmental
costs associated with plastic lumber and CCA-treated lumber, but will not be
all encompassing, and does not assign costs to environmental impacts. The two
materials were compared, looking at the energy required by them in their life-
times, how long they will last, and the air emissions associated with their
life cycle.
This LCA does not go "one step back" in the analysis. Many LCAs include
in the calculation of impacts the environmental damage caused by producing the
fuel used to produce energy for manufacturing or transportation. Assume that
44
-------
these will be small compared to the impacts of burning this fuel. Further-
more, no consideration was given to the impacts of building facilities or
equipment to process and transport the products. These impacts are also
generally assumed to be small per ton of material produced.
Recycled Plastic Lumber
Inherent in the analysis of plastic lumber is the assumption that the
recycled plastic has already cycled through its originally intended lifetime.
This means that the energy and impacts associated with producing the original
plastic feedstock are not included. By doing this, the plastic's "first"
lifetime as a container or other object bears the costs associated with the
original plastic feedstock acquisition, transportation, and manufacturing.
Once this lifetime is complete, the plastic which is not currently recyclable
is essentially a waste, which has no inherent value. Further, the costs of
disposal of this waste, including garbage collection, landfilling, or inciner-
ating, and the environmental impacts associated with this disposal, are also
borne by the plastic in its first lifetime. There will be a cost associated
with separating out the nonrecyclable plastic from the waste streatn, which is
where the analysis begins.
Plastics Acquisition
Acquisition of the plastic will be by either' curbside collection of the
separated plastics, or by separation of the plastics from MSW at a MRF. In
the case of curbside collection of separated plastics, it is possible to
.separate the contribution to energy and associated truck emissions which •
result from the plastic portion of the recyclables collected. For standard
garbage collection and subsequent s-eparation of the plastics, energy and envi-
ronmental impacts from transportation are assumed to be negligible as the
waste would be picked up and taken to a landfill or incinerator if it-were not
taken to the MRF. The energy and environmental impacts at the MRF for
plastics are assumed to be small.
For a
energy used
col lection.
Transport
curbside collection approach, the top of Table 3.2 gives the
and environmental impacts resulting from the trucks used in
Transporting the plastic from the MRF to manufacturer is not calculated
since the plastic would normally go directly to a landfill if it were not
separated at a MRF or collected-by curbside collection.
45
-------
Manufacture
The manufacture of the plastic includes granulation of the plastic, and
extrusion. Table 3.2 shows the various economic, energy, and environmental
costs attributed to this process.
TABLE 3.2 LIFE CYCLE COSTS OF PLASTICS
Plastics Acquisition (Curbside)
Energy (Truck Fuel) 1,182,000
Environmental (Truck Emissions)^)
NOX
Particulates
HC
CO
Manufacturing of Plastic Profiles
Granulation Of Plastics^3)
Shredder
Grinder
Capacity
Electricity Rate
EQUIVALENT BTU RATE
Extrusion Of Plastics
Throughput
Lines
Availability
Lineal density
Capacity
Electrical rate
358Btu/foot
0.00
0.10
0.01
0.02
0.07 g/foot
38 kW
75 kW
188 kWh/ton
0.5 kWh/foot
1,962,323 Btu/ton
1,619 Btu/foot
500Ibs/hr/line
2
0.9
1.65Ibs/foot
3942tons/yr Equivalent Btu Rate
1104 kWh/ton 11,554.160 Btu/ton
0.91kWh/foot 9.532 Btu/foot
0
344
25
82
243 g/ton
50 hp
'100 hp
0.6 tph
(continued)
46
-------
TABLE
Envi
SOX
NOX
ronmental
Particulates
HC
CO
(el
3.2 LIFE:
ectricity
CYCLE COSTS OF
from coal power
(PM10)
0.
0
0.037
0.0112
0.0427
000053
.00032
PLASTICS
Pi
ants)<
Ibs/kWh
(continued)
2)
16
5
19
0
0
.80
.08
.39
.02
.15
g/kWh
(1) Koch. 1992.
(2) Compilation of Air Pollution Emission Factors, 1985.
(3) Energy use based on design.
Transport to End Use
It is assumed that this transportation is local for the most part and
therefore small compared to transportation from a wood lumber mill to the end
use which can- be hundreds of miles, depending on the source of the lumber.
Construction ' .
Construction costs are a function of not only the cost of building but
also the cost of maintaining or renewing a structure over the time frame of
interest. One-time construction costs and impacts will be more expensive than
wood since plastic used for decking will require more support underneath
(joists) to prevent excessive deflection. From the evaluation it was deter-
mined that the joist spacing should be less than wood construction dictates.
However a major factor in life cycle analysis is the lifetime of the plastic
versus that of wood under equivalent environmental conditions. If a marine
structure must be renewed every 5 years for wood and only every 10 years for
plastic, all of the life cycle costs for wood will-be multiplied by two to
compare with an equivalent life for plastic.
Unfortunate-ly the information needed to determine absolutely the life-
time of the plastic is not available from the testing program. The testing
program was limited to short-term laboratory tests which do not adequately
predict long-term performance in the field. If the assumption is made that
creep is the important factor in terms of lifetime, the creep testing data ,can
be used to determine the amount of deflection expected for the plastic lumber
over time. By then assuming a maximum deflection which can be tolerated, the
time to reach this deflection can be determined,, and hence the plastics
life.time.
Cost of construction is estimated to be more expensive per square foot
for a plastic boardwalk versus a wood, one. It is assumed that environmental
47
-------
impacts and energy use during construction are small compared to the other
activities in the life cycle.
CCA-Treated Lumber
Acquisition
Impacts during acquisition of lumber are due to use of gasoline and
diesel-powered engines for cutting trees and loading onto trucks. Proper
forest management is assumed, which minimizes environmental impacts.
Table 3.3 gives the energy and estimated air emissions resulting from this
activity. Impacts are given on the basis of tons of oven dried lumber.
TABLE 3.3 LIFE CYCLE COSTS OF WOOD
Tree Acquisition
Environmental (2)
SOX
NOX
Parti culates
HC
CO
Tree Transport^)
943,OOOBtu/ton<1)
0 g/gallon
5.26
1.8
713
1,618
250,000 Btu/ton
4 gal gas
2 gal diesel
0 g/ton
21.04
7.2
2,852
6,472
100 miles
0 g/foot
. 0.016306
0.00558
2.2103
5.0158
15 ODT/truck
Environmental
SOX
NOX
Particulates
HC
CO
0 g/mile
11.44
0.7
2.52
8.53
Linear Density
1.551bs/ft
Lumber Manufacture^1^
El ectri ci ty 786,000 Btu/ton
0 g/ton
76
5
17
57
0.00 g/foot
0.06
0.00
0.01
0.04
74 kWh/ton
600.21 Btu/ft
Environmental
sox
NOX
Parti culates
HC
CO
0.00021b/kWh
0.004
0.0067
0.0022
0.0333
7 g/ton
134
225
74
1,119
0.01 g/ft
0.10
0.17
0.06
0.87
(continued)
48
-------
TABLE 3.3 (continued)
Steam Heating
4,060,000 Btu/ton
560 Ibs wood/ton lumber
Envi ronmental
SOX
NOX
Particulates
HC
CO
3,147 Btu/foot
0.151b/ton
chips
2.8
5
1.7
25
19 g/ton
lumber
356
636
216
3178.
0.01 g/foot
0.28
0.49
0.17
2.46
Lumber
Transport^)
125,000 Btu/ton
500 miles
30.00 ODT/truck
Environmental ^
SOX
NOX
Parti cul ates
HC
CO
0 g/mile
11.4
0.7
2.5 ' •••
8.5
0 g/ton
38.1
2.3
8.4
28.4
0.00 g/foot
0.03
0.00
0.01
0.02
(1) Boyd, 1976.
(2) EPA, 1985.
Transportation to Lumber Mill
The distance between the cutting area and the lumber mill is assumed to
be 100 miles. It is assumed that the wood is green, roughly 50% by weight
moisture. Therefore, moisture is accounted for in the shipping of wood to the
mill. Although a large portion of the wood from the tree will be used for
energy, pulp, or other wood products, it is assumed that impacts will accrue
to these uses, and not to the lumber.
Lumber Manufacture-- -
Manufacture of lumber at the lumber mill demands steam and electricity.
It is assumed that the energy required for this process comes from wood
byproducts such as bark and hog fuel (waste wood). Thus the environmental
emissions are derived from wood-fired power plant data, which tend to have
relatively low sulfur, NOX, and particulate emissions compared with a coal-
fired plant. The other advantage to burning wood, which is not accounted for
by this analysis, is that the impacts due to harvesting the trees for lumber
are shared by the wood used to create energy for the lumber mill. In fact,
there would be a net energy surplus if all of this non-lumber wood went to
49
-------
create energy. This simplified analysis does not credit wood for this
environmental advantage.
Transportation to End Use
A distance of 500 miles from the lumber mill to the community where the
lumber will be bought and used was assumed.
Compari son
Table 3.4 shows the comparison of energy and environmental impacts asso-
ciated with life cycles for plastic and CCA-treated wood lumber. Since the
lifetimes of the two materials in a marine application are different, these
values need to be adjusted to make a fair comparison. The economic costs will
also need to be adjusted.
TABL& 3.4 LCA ANALYSIS OF CCA-TREATED 2 IN. BY 6 IN. WOOD VS. PLASTIC
Lumber
SOX
NOX
Particulates (PM10)
HC
CO
Energy (Btu/foot)
Plastic
SOX
NOX
Particulates (PM10)
HC
CO
Energy (Btu/foot)
Acquisition
G/foot
0.02
2.21
0.01
0.00
730
0.00
0.10
0.01
0.02
0.07
358
Transport Manufacturing Total
G/foot G/foot G/foot
0.02
0.00 0.38
0.00 0.67
0.06 0.22
0.00 3.33
193 3,755
23.29
7.05
26.88
0.03
0.20
11.151
0.02
0.40
2.88
0.29
3.33
4.680
23.29
7.16
26.89
0.06
0.28
11.509
50
-------
.Environmental Impacts
The bioassays and biofouling experiments that were run can be compared
with similar experiments performed on CCA-treated wood and recycled plastic.
In previous sea urchin fertilization experiments, Weis, et al. ("Toxicity of
Construction Materials in the Marine Environment: A Comparison of Chromated-
Copper Arsenate-Treated Wood and Recycled Plastic," Archives of Environmental
Contamination and Toxicology, 1992) showed that CCA-treated lumber 1/10-the
area of plastic test samples was far more toxic to fertilization and develop-
ment. Weis and Weis ("Construction materials in estuaries: reduction in the
epibiotic community on chromated copper arsenate (CCA) treated wood," Marine
Ecology Progress Series, vol. 83: 45-53, 1992) found that CCA-treated wood
during biofouling experiments showed significant decreases in number and
biodiversity of species adhering or fouling compared to recycled plastic and
untreated wood. 'The biofouling on CCA-treated wood had many times the levels
of copper, chromium, and arsenate than did that on recycled plastic or
untreated wood.
The results indicate that the CRC recycled plastic lumber behaves
reasonably close to previous studies of recycled plastic for biofouling and
sea urchin testing. Sea urchin.experiments for plastic showed no impact on
fertilization and no correlation between plastic concentration on development
of the embryos. The biofouling experiments showed the plastic to be fouled
with a greater diversity of species than even untreated wood. These results
cannot be compared directly with previous work in different locales, but they
point to the general biological compatibility of the plastic in'marine
envi ronments,.
DURABILITY/LIFETIME
Wood
Tests have been performed on CCA-treated lumber used in marine environ-
ments (cite literature). These tests are generally multi-year tests which are
designed to show the resistance of treated wood to wood-borers. Lifetimes of
the wood are dependent on the site of the construction. Different species of
wood-borers live in different regions and degrade specific types of wood at
different rates. Lifetimes for exposed wood can be as little as 1 year and
greater than 20 years, depending on the specific environment.
The lifetime of freeze-thaw cycles may have a large effect on longevity
of CCA-treated lumber in northern climates. Boardwalks have been found to .
have a life span of as little as 2 years in marine environments.
51
-------
Plastic
For recycled plastic lumber, the limited testing performed here cannot
accurately predict its lifetime. Only extensive field testing of actual
applications can demonstrate the expected longer wear of plastic over wood.
The two tests performed to demonstrate longevity are the creep tests and-UV
test.
Creep
Creep is a phenomenon of plastic which causes the plastic to deform over
time under a continuous load. The rate of deflection with a continuous load
occurs at an exponential rate; therefore, the load will cause a larger change
in deflection in the'first month than in the next 12. The deflection may not
be permanent as the material can recover pver time after the load is removed.
Creep must be adequately addressed in design of construction in order to
avoid excessive deflection or failure of the plastic under a load. There are
no construction guidelines for plastic lumber, and.it would be recommended
that continuous loads- be kept minimal, unless deflection 'is not an important
consideration. For instance, a bulkhead will have a continuous load, but may
not be objected to if the plastic lumber bows outward over time, as long as
the soil is adequately retained. In the case of decking, as long as heavy
objects are not placed on the plastic lumber for long periods of time, creep
should not be a problem.
UV Testing
The UV resistance of the plastic was shown to be good based on the short
amount of testing performed. This test simulates and accelerates the exposure
of plastic to UV. UV has a tendency to cause the surface of plastic to become
embrittled by promoting additional polymerization of the plastic. This can
result in a loss of the surface layers which absorb the UV radiation. The
laboratory simulations showed that the rate of loss is negligible, and should
not have a substantial influence on the lifetime of the recycled plastic.
It should be noted that the UV testing performed did not monitor other
properties in the plastic which,may have changed or remained constant due to
the UV exposure. These could include flex testing, creep, and compressive
strength. These were not performed due to 'the expectation that these proper-
ties would also not change significantly and the fact that the plastic lumber
would be of such small dimensions to fit in the exposure tester as to not
represent real properties of the full size lumber.
52
-------
SECTION IV
CONCLUSIONS AND RECOMMENDATIONS
Based on the results of this study, some preliminary conclusions
concerning the usefulness of CRC's RPM can be made. It is important to note
that tests were not performed on actual field applications, which greatly
restricts-the ability to make conclusions about the RPM's practical perform-
ance. Using engineering judgment and laboratory results, inferences can be
drawn about the expected performance of this material in real world applica-
tions. Engineering estimates are no replacement for long-term testing of RPM
in field use. Only field testing can produce results which will confirm a new
material's ability to withstand environmental conditions, various forces, and
other difficult to simulate effects.
The benefit of the MITE evaluation is to generate an independent
analysis of key properties which can then be compared to other building
materials. As such, the ASTM and EPA tests provide a benchmark for this
material, as well as other plastic materials to be produced in the future.
When linked with results from field testing, the laboratory tests will have
even greater usefulness.
CONCLUSIONS
The conclusions are divided into five categories: Cost, Environmental
Impacts. Lifetime, Physical Strength, and Friction. Most results fit in one
particular category, although there is some overlap. For instance, strength
characteristics are related to time due to the creep phenomenon. These then
will impact the ultimate lifetime of the material'.
Cost
The price charged by CRC for their 2 in.
approximately $1.00/ft in larger quantities.
x 6 in. foamed core RPM is
Compared with CCA-treated wood, the price is greater for the plastic.
However, wood prices have been increasing rapidly. It is difficult to make an
absolute judgment as to whether the RPM is more cost effective than wood -
without designing, constructing, and testing a real world application where
the RPM would serve as a substitute for wood.
53
-------
Environmental Impacts
From the sea urchin experiments, it can be concluded that the RPM would
not hinder fertilization at practical surface to volume ratios. CCA-treated
wood has been shown to reduce fertilization of sea urchins. In environmen-
tally sensitive areas, RPM could have a distinct environmental, and thus
marketing, advantage.
The biofouling experiments demonstrated the compatibility of the RPM
with the ecosystem in the marine environment of Melbourne, Florida. There did
not appear to b.e any indication of toxic behavior of the RPM, even though the
RPM was produced from unwashed post consumer plastic, suggesting that the
extrusion process may sterilize the plastics and vaporize detergents and oils.
The biodiversity was greater for RPM plastic than even untreated wood.
Lifetime
.The lifetime of plastics would be difficult to predict from the labora-
tory testing alone. Proper design will be necessary to ensure that creep does
not become a limiting factor. Resistance to UV and marine environments should
be better than wood. High temperatures which could be generated in the
plastic from sunlight will increase the rate of creep as seen in the results
of the elevated temperature creep experiments. Reflective coatings may reduce
temperature adequately. The relatively low conductivity of plastic will
produce a significant temperature gradient in the plastic. This temperature
gradient was not duplicated in the laboratory, but its effect on creep could
perhaps be predicted by modeling. Field testing is needed on actual plastic
lumber applications before a final judgment can be made with respect to life-
time of the plastic.
Physical Strength
Strength may be the key to RPM acceptance in most marine applications.
While RPM has many advantages over CCA-treated wood, its strength will deter-
mine in what application it will ultimately be used. Strength over time will
determine its lifetime in a given application.
The compressive and flexural strength of the RPM was found to be far
lower than that of CCA-treated wood. This makes its use for structural appli-
cations impractical without some type of modification of the plastic recipe.
or construction design. A glass fiber reinforced RPM was an originally
planned material sample, but this was not manufactured by CRC in time for
testing.
Non-structural uses such as decking material on docks, piers, and board-
walks; railings: pier impact protection; benches and tables; and movable piers
54
-------
seem to be best suited to the RPM tested. These applications generally have
small and infrequent loads and thus strength requirements which can be
attained through proper design.
The plastic property known as creep is an additional characteristic
which requires attention in designing for a specific application. Creep
causes the plastic to move over time in the direction in which constant force
is applied. The result is a deflection of the plastic. Thus proper design
using plastic lumber must take creep into account.
Friction
Friction is an important property when designing a platform such as a
boardwalk or dock. The tests demonstrated that the CRC plastic lumber does
not have as high a coefficient of friction as wood. A surface treatment can
be applied to the plastic to enhance this coefficient. While at CRC, a
process to apply an epoxy which had sand mixed in for "traction" was observed.
Paint does not adhere to the plastic without some type of surface preparation.
RECOMMENDATIONS
Further exploration may lead to increased acceptance and use of recycled
plastic lumber in general and the tested material in particular. First, field
testing of the RPM is needed to demonstrate its performance under real appli-
cations. This testing must be done over a long period of time in a variety of
typical applications and locations. In conjunction with field and continued
laboratory testing is development and refinement of the manufacturing process.
There are many variables including additives, washing of the plastic, control
over the mix of plastics, processing temperatures and speeds, and shape
configurations which could potentially lead to enhanced properties. Finally,
market research could be performed to increase market penetration.
Field Testing
Performance of the RPM in the field was not done in this study. Field
performance testing would demonstrate the material's effectiveness under real
world applications. Certain problems can be discovered through longer term
use in a combination of environmental conditions that cannot be foreseen and
modeled in the laboratory.
Widespread acceptance of the RPM as a building material can be further
obtained with demonstrations of its use over long periods of time. While
scientific reports such as this one can provide useful data for purposes of
comparing materials, practical demonstrations of the use of the materials will
help gain public acceptance.
55
-------
From a construction standpoint, the material must be accepted by the
construction industry as a viable alternative. This will require properties
which are both reasonable and consistent. Construction codes may need to be
altered to allow substitution of plastic for conventional building materials
in some cases. Since the architects and building contractors are, in most
cases, responsible for a structure's performance, they must be able to rely on
the performance of a building material. RPM acceptance by architects and
contractors can be -bolstered by a demonstration program using RPM for various
construction projects.
Environmental effects can be verified with field testing. Especially of
interest here is the effect of various animals on the plastic's lifetime.
Woodborers are among the chief causes of deterioration of treated wood in
marine applications. The plastic should be highly resistant to such
organisms. Environmental impacts can also be further quantified through a
field testing program.
New Formulations
As the material manufactured here represents a relatively new process
and product, there is room for engineering improvements. Depending on the
application, strength can be improved through the use of various additives, or
by altering the feedstock composition. It should be noted, however, that the
100% commingled plastic formulation tested in this study represents probably
the least expensive and most recyclable option. Additives, control of the
feedstock composition, washing of the feedstock, etc., will add to the cost of
manufacturing the plastic.
At the time of taking samples for testing, the hollow-core RPMs were not
being produced. Samples of these observed from a trial in Austria looked
intriguing. The melting of the plastics appeared more uniform, although
perhaps due to the fact that the feedstock was from source-separated post-
consumer waste. The hollow core retains much of the strength of the foam core
with the added advantage of potential for adding concrete or steel rebar to
give structural integrity.
Other shapes could be produced using different dies and molds. The
Florida Institute of Technology has done some research on the strength of
various shapes, each of the same cross sectional area. Shapes like I-beams
have more flexural strength than simple hollow cores.
Consistency
Consistency of properties, as mentioned before with respect to construc-
tion, is important to sales. Consistency can perhaps better be maintained
once the process is optimized, and the feedstock is controlled to some
56
-------
reasonable specifications of composition and cleanliness. A minimal testing
program could be put in place to maintain feedstock and product quality. At a
minimum, the flexure and compression strength should be tested each time the
product is switched and at regular intervals thereafter.
Other Properties
Properties such as friction, and others which are found to be important
after a thorough field testing program, should be examined to see if modifica-
tions or enhancements can be made. In particular for friction, a research and
testing program designed to try various surface treatments and their wear
resi-stance and friction characteristics should be performed.
MARKET RESEARCH
For a relatively new product such as RPM, marketing must work side by
side with customer feedback, and research and development. There are three
areas where marketing can play a key role: finding applications for the
plastic, determining barriers to market penetration, and strategic planning to
overcome these barriers.
Appli cations
The marine applications mentioned in this study are general and can be
better defined and specified through market research. For instance, plastic
lumber may find a niche in areas which have more concentrated use of CCA-.
treated wood in relatively still water. In these cases, plastic's negligible
toxicity will be a clear advantage over wood, as leaching of metals from wood
may have an environmental impact which is unacceptable. It would be worth-
while to study the advantages and disadvantages of plastic to identify market
niches.
Barriers
Market research can identify barriers to the proliferation of plastic
lumber. These could include perceptions about plastic's characteristics based
on poorly made or inappropriately designed products by other manufacturers.
It is important to determine why perceptions exist in order to try to change
them.
57
-------
-------
REFERENCES
American Society for Testing and Materials (ASTM), Annual Book of Standards,
Philadelphia, Pennsylvania, Sections 4, 6, 8, 14, 15.
C. W. Boyd, et al., "Wood For Structural and Architectural Purposes," Journal
of the Society of Mood Science and Technology, 8:1, 1976.
A. H. Buchanan, "Bending Strength of Lumber," Journal of Structural
Engineering, 116: 5, 1990.
Compilation of Air Pollutant Emission Factors, "Vol. II: Mobile Sources," 4th
ed, Sept. 1985, and "Vol. I: Stationary Sources."
J. S. Graham, "Pressure-Treated Wood Effect on Marina Environment," In
Proceeding of the. First International Conference on World Marina, Long Beach,
California, Sept. 4-8, 1991.
R. D. Graham and D. J. Miller, "Marine Exposure Tests of Pressure-Treated
Douglas Fir and Southern Pine," Presented at the Third Pacific Area National
Meeting, American Society for Testing Materials, San Francisco, California,
October 13, 1959.
B. Johnson and R. Jackson, "Durability of Heartwood in Treated Southern Pine
Bulkheads," Forest Products Journal, 40: 7/8. 1990.
P. Koch, "Wood versus non-wood materials in U.S. residential construction:
some energy-related global implications," Forest Products Journal, 42:5, 1992.
C. R. Southwell and J. D. Bultman, "Biological Deterioration of Woods in
Tropical Environments, Part 3 - Chemical Wood Treatments for Long-Term Marine-
Borer Protection," NRL Report 7345, 1971.
Uniform Building Code, International Conference of Building Officials, 1988.
United States Environmental Protection Agency, "Short-Term Methods for Estab-
lishing the Chronic Toxicity of Effluents and Receiving Waters to Marine and
Estuarine Organisms," EPA 600/4-87/028, Environmental and Support Lab,
Cincinnati, Ohio, p. 239-272, 1988.
59
-------
0. S. Weis and P. Weis, "Transfer of Contaminants from CCA-Treated Lumber to
Aquatic Biota," Journal of Marine Biology -and Ecology, 1:11, 1992.
Ibid, "Construction-Materials in Estuaries: Reduction in the Epibiotic Commu-
nity on the Chromated Copper Arsenate (CCA) Treated Wood," Marine Ecology
Progress Series, 80: 45-53, 1992.
J. S. Weis, P. Weis, Greenberg, and T. J. Nosker.. "Toxicity of Construction
Materials in the Marine Environment: A Comparison of Chromated-Copper
Arsenate-Treated Wood and Recycled Plastic," Archives of Environmental
Contamination and Toxicology, In press.
P. Weis. J. S. Weis, and L. M. Coohill. "Toxicity to Estuarine Organisms of
Leachates from Chromated Copper Arsenate Treated Wood," Arch. Environ. Contam.
Toxicol, 20:118-124. 1991.
Western Wood Products Association, Western Lumber Product Use Manual, 0709/A/
Rev. 11-91/40M.
6. A. Zarillo, C. R. Lockert, G. W. Swain, and L. E. Harris. "Feasibility of
Using Recycled Plastic for Marine Construction," prepared for the Florida
Center for Solid and Hazardous Waste Management, April 1991.
60
-------
GLOSSARY
Acute toxiclty
Baldwin Universal
Machine
Blow-molded plastic
Biofouling
Chronic toxicity
Compressive strength
Creep
Flex
Injection molded
Life cycle analysis
Modulus of elasticity
Secant modulus of
elasticity
Toxicity of a material which causes reproductive
failure in the sea urchin fertilization test.
Apparatus for performing physical tests on the
material.
Forming process generally used to create plastic
bottles; plastic is blown out against a two-piece
mold which is pulled off after cooling, leaving a
seam around the length of the bottle.
Propensity of different species of marine or§anisms
to adhere to the material.
Toxicity which affects the growth of the sea urchin
after it has been fertilized.
The ability of a material to resist deflection
causeti by a force that pushes the material inward.
Movement of a material over time due to a continuous
force applied to the material.
The bending of a material; in this study caused by a
force at the middle of a span of plastic lumber that
is supported on its ends.
Forming process for plastic containers; plastic is
injected into a mold; a small raised point on the
center of the bottom of the container is observed.
Evaluation which accounts for all costs and impacts
of a material or product from its raw material
extraction to its ultimate disposal.
Slope of the stress - strain curve at low stress;
measures the strength of a material in the regime
where elastic behavior occurs (no permanent
deformation).
Similar to modulus of elasticity except measures the
slope of the stress-strain curve from the origin to
the load of interest (about 500 Ibs for this study).
61
-------
Static friction
Strain
Stress
A measure of resistance to movement <5f the surface
of a material fitted against either the same
material or a different one such as a shoe sole.
A deformation, or change of dimension in the direc-
tion perpendicular to a stress placed on a material.
Force applied to a material which tends to strain or
deform it.
62
&U.S. GOVERNMENT PRINTING OFFICE: 1993 - 750-002/80270
------- |