RECYCLED GLASS UTILIZATION IN HIGHWAY CONSTRUCTION

Isaac Finkle, GRA, EIT

Graduate Research Assistant
Department of Civil and Architectural Engineering
University of Wyoming
Dept. 3295, 1000 East University Ave.
Laramie, Wyoming 82071
Tel: (307) 766-3427, Fax: (307) 766-6784
Email: [email protected]

Khaled Ksaibati, Ph.D., P.E.

Professor of Civil Engineering
Department of Civil and Architectural Engineering
University of Wyoming
Dept. 3295, 1000 East University Ave.
Laramie, Wyoming 82071
Tel: (307) 766-6230, Fax: (307) 766-6784
Email: [email protected]

Timothy Robinson, Ph.D.

Department of Statistics
University of Wyoming
P.O. Box 3332
Laramie, Wyoming 82071-3332
Tel: (307) 766-5108, Fax: (307) 766-3927
Email: [email protected]

A Paper Prepared for the Transportation Research Board 86th Annual Meeting
January 21-25, 2007, Washington D.C.

Word count: 4210 + 3 Tables x 250 + 10 Figures x 250 = 7460

August, 2006

TRB 2007 Annual Meeting CD-ROM Original paper submittal - not revised by author.
 Finkle, Ksaibati, and Robinson 1

ABSTRACT
The utilization of recycled glass in highway applications has been occurring over the past
couple of decades. In recent years, the discovery of several economic and environmental
benefits could increase the use of recycled glass in highway construction, making the
evaluation of the engineering properties of glass and aggregate mixes necessary. The uses
of recycled glass have varied widely, depending on the specific application. Crushed
recycled glass, or cullet, has been used independently, and has also been blended with
natural stone construction aggregate at different replacement rates. This research
provides an evaluation of the potential use of glass cullet when used in combination with
natural base course aggregate for roadway construction. This research studied the
strength and moisture/density characteristics of different glass and aggregate blends to
examine the effects of blending glass cullet into base course aggregate. Two sources of
natural aggregate were tested; one being crusher run and very angular in nature and the
other being pit-run and rounded in nature. The glass was introduced into the aggregate at
replacement rates of 10%, 20%, and 30%. Four different maximum glass cullet sizes
were also tested, with maximum sizes ranging from 3/4” to 3/8”. The strength of the
glass-aggregate blends was evaluated using the AASHTO T190-Resistance R-value Test.
Analysis of the data showed that glass cullet mixed with the more angular crusher run
aggregate (LAF) performed more consistently than when the cullet was combined with
the rounded natural aggregate (STAR) for all sizes and replacement rates. The LAF
glass-aggregate blends had average strength values above or slightly lower than the
control mix across all replacement rates and maximum cullet sizes. The STAR blends
exhibited a decrease in strength as both cullet size and replacement rate increased. The
moisture-density relationships were determined in accordance with AASHTO T99-
Standard Proctor Test. The maximum size of glass cullet used was shown to be
insignificant in determining the optimum moisture content and maximum density of the
blends. The replacement rate had a significant effect on both the compaction properties.
As the cullet content increased the optimum moisture content increase and the maximum
density decreased.

TRB 2007 Annual Meeting CD-ROM Original paper submittal - not revised by author.
 Finkle, Ksaibati, and Robinson 2

BACKGROUND AND INTRODUCTION

As recycling continues to grow as a means of utilizing waste materials in today’s world,
more markets must be established for products containing recycled materials.
Specifically, glass continues to be one of the principal waste products generated by the
public and recovered by the recycling industry. In 2003, 10.7 million tons of waste glass
was generated in the U.S., and 2.35 million tons or 22% was recycled. [1] Recovered
recycled glass is crushed and becomes what is known as glass cullet. Glass cullet is the
basic product that is reused for most recycling applications.
The major market for the glass cullet is the container industry, but the use of glass
cullet in glass container production batches is limited due to color contamination and
transportation costs. [2] Glass cullet which is reused for containers must be well sorted
according to color and must also be relatively free of debris. This required sorting and
cleaning costs time and money. Transportation costs associated with recycled glass are
often high because of the geographic concentration of glass plants near urban areas. [3]
This makes it more difficult to use glass cullet produced in rural locations for the purpose
of making new containers. For example, the glass suppliers for this project send
shipments of cullet hundreds of miles to out of state processors. Transportation costs
range from $20-$35 per ton, while the cullet receives $50 per ton for brown glass and
only $20 per ton for clear when delivered. For an average shipment, the higher price
paid for the brown glass barely covers the shipping cost of the clear glass. If another
market for the clear or mixed glass existed, the cullet could be used more economically.
Over the past few decades, another market for cullet has been developing within
many states, including Texas, Minnesota, California, Washington, Oregon, New York,
and several others. The use of glass cullet as a construction aggregate is becoming more
widespread as the results of research and laboratory testing studies on glass cullet become
better known. When glass is used as construction aggregate, there is less of a need to sort
and separate by color, and less of a need for the cullet to be free of contaminants. The
need for transporting the cullet long distances to container glass manufacturing facilities
can also be eliminated. Glass cullet has the potential to be used in aggregates for
structural or drainage applications. Some of the structural applications include base and
subbase, embankments, structural fills, and utility backfill. Drainage blankets, French
drains, and various waste medias are the possible drainage applications. [4]
Several studies have been conducted regarding the use of glass cullet as
construction aggregate. The largest study to date concerning the use of glass cullet is
probably The Clean Washington Center’s “Glass Feedstock Evaluation Project,”
completed by Dames and Moore, Inc. This study included replacement rates of 15%,
50%, and 100%, and mix gradations of 3/4” minus and 1/4” minus. [6] Several tests were
conducted to determine whether the cullet and aggregate mixtures were suitable from the
engineering and environmental standpoints. The study found that the addition of glass
cullet had no negative effects to the soils it was combined with. [5] Another study
conducted by the Texas Department of Transportation found similar results. TxDOT
concluded that the addition of glass up to 20% didn’t have significant effects on the
engineering properties of the mixture. [7]
Several states have developed specifications for using glass cullet for construction
purposes. Most studies have indicated that glass cullet would be a suitable substitute for
natural aggregate at lower replacement rates. In a survey conducted by the Texas

TRB 2007 Annual Meeting CD-ROM Original paper submittal - not revised by author.
 Finkle, Ksaibati, and Robinson 3

Department of Transportation, most states that indicated using glass cullet, specified
replacement rates of 5%-20%. [7] Minnesota Department of Transportation has had
several counties that have implemented the use of glass cullet in their roadways.
Mn/DOT specifications include the use of 10 percent reclaimed glass in aggregate
material for road base. [8] Otter Tail County used 10% glass in road base on a 4 mile
section of county road. No specialized equipment was needed for placement and
compaction and the county officials were pleased with the results. [8]
The use of glass cullet by state DOT’s is becoming more common, but the lack of
supply or quantity needed for larger projects, may force more city and county agencies to
consider using glass cullet as construction aggregate. This may be more beneficial for a
couple of reasons. Cities and counties are responsible for disposing of the glass and are
therefore less reluctant to consider its use. Also, city and county construction projects
tend to be smaller in nature; therefore the quantity of glass needed could be obtained
without lengthy stockpiling times or without the cost of hauling cullet over long distances
to the project. Whether larger state DOT’s or small city and county agencies adopt
standards for using glass cullet, the engineering properties of the glass and glass blends
must be evaluated.

RESEARCH OBJECTIVES
While use of recycled glass in roadway construction is increasing, many agencies are still
reluctant to consider cullet as a suitable substitute for materials already being used. This
is mainly due to unfamiliarity with the engineering properties of cullet and a lack of
suitable sources that supply glass cullet. This research will examine several properties of
glass cullet through laboratory testing. Specifically, this study will investigate the
strength and moisture-density characteristics of glass cullet blended with natural
aggregates for use as road base material. This study will provide recommendations on the
maximum particle sizes and associated replacement rates at which glass cullet can
effectively be used.

LABORATORY TESTING
Laboratory testing was done at the University of Wyoming and the Wyoming Department
of Transportation (WYDOT) Central Laboratory. All the testing was conducted
according to AASHTO and WYDOT specifications. The Wyoming Department of
Transportation performed the Resistance R-Value testing, and all other testing was
conducted at the University of Wyoming.

Research Variables
The effects of three variables were studied in the testing process. The variables included
were: type of aggregate, size of glass cullet, and glass replacement rate. The test matrix
for this research is shown in Table 1. Two different types of natural aggregate were used
separately in the aggregate and glass blend test samples. The glass cullet was separated
into four different maximum sizes: 3/4”, 5/8”, 1/2”, and 3/8”. A maximum size of 3/4”
indicates that all glass particles pass a 3/4” sieve. These four particle sizes were chosen
based on previous research completed by the Texas Department of Transportation and the
Clean Washington Center and also by visual inspection. There were definite visual
differences in the size of glass cullet particles occurring between the four sizes chosen.
The third variable tested was glass cullet replacement rate. Replacement rates of 10%,

TRB 2007 Annual Meeting CD-ROM Original paper submittal - not revised by author.
 Finkle, Ksaibati, and Robinson 4

20%, and 30% were tested. A replacement rate of 10% indicates that 10%, by dry
weight, of aggregate in the blend is replaced with glass. The properties of all the glass-
aggregate blends were compared to control samples made up of 100% natural base course
aggregate.

Materials
Base course aggregate was used for all testing. The aggregate was donated by two local
sources. Aggregate 1 was obtained from STAR Aggregates of Cheyenne, Wyoming and
will be referred to as STAR. Aggregate 2 was obtained from Lafarge North America,
Inc. and will be referred to as LAF.
Aggregates properties for both aggregates can be found in Table 2. As shown,
STAR was a pit-run aggregate, while LAF was crusher run. The two aggregate were
very similar in nature, except for a couple of properties. STAR aggregate particles were
more rounded in nature, whereas LAF particles were very angular. Both aggregates were
classified as non-plastic according to Atterberg Limit tests, but LAF exhibited
significantly more cohesion than STAR by visual inspection. The WYDOT classification
of the material was base grading ‘W’. The gradation specification for base ‘W’, along
with the gradation of the material used for blending is shown in Table 3.
The recycled glass cullet used for all testing was obtained from three sources
within Wyoming. Two recycling centers donated glass cullet that was crushed by a
crusher. The other source had a unique policy in place at the landfill in which a
designated area was used to store all glass brought in, and was then crushed by a dozer.
The intent of using this source was to compare the effects of different crushing
techniques. The glass was not sorted by color for the testing, because when used for
construction purposes, color sorting would not be necessary. The type of glass accepted
for use was bottles, jars, and other container glass, while cathode-ray tubes (CRT) and
other hazardous glass were discarded. This sorting occurred only at the recycling centers
and was monitored at the landfill operation.

Gradations
Gradation tests were conducted to determine the size distribution of both the natural
aggregate and glass cullet. Testing was in accordance with AASHTO T11 – Materials
Finer than No. 200 Sieve in Mineral Aggregate by Washing and AASHTO T27 – Sieve
Analysis of Fine and Coarse Aggregate. Aggregate from both sources had very similar
gradations; therefore, the aggregate gradation used for all blends was an average of both
source gradations. For the same reason, an average of the glass gradations from the two
glass crushers was used for all blends. The gradations of the aggregate and glass cullet
used for all blends is shown in Table 3.

Resistance R-Value
The strength and moisture susceptibility of the aggregate glass blends were determined in
accordance with AASHTO T190 – Resistance R-Value and Expansion Pressure of
Compacted Soils. This test consists of three separate components: Exudation Test, Swell
Pressure Test, and Stabilometer Test. The range of the results of the R-Value test is 0-
100, with a practical range of 5-85. A material with an R-value of five is a very weak
material, while a value of 85 indicates a high strength material. A minimum R-value of
75 for base and 60 for subbase is required by WYDOT.

TRB 2007 Annual Meeting CD-ROM Original paper submittal - not revised by author.
 Finkle, Ksaibati, and Robinson 5

In addition to strength, an idea about the moisture sensitivity of the material can
gained from conducting the R-Value test. WYDOT classifies a material as moisture
sensitive if on a graph of exudation pressure versus R-Value the curve of R-values at
different moisture contents has a difference in value greater than 5 between the 300 psi
and 200 psi exudation pressure.
The blends for the testing were mixed at the University of Wyoming, while the
Resistance R-Value tests were conducted by the Wyoming Department of Transportation.
The glass was blended with the natural aggregate as a percentage by dry weight of the
total amount needed to conduct the test. As a requirement of the test all material was first
“scalped” on the 3/4” sieve. The material passing the 3/4” sieve was used to prepare the
test blends. All aggregate and glass were separated according to particles size then
blended back according to average gradations. When blended, the proportioning of the
particle sizes for the aggregate and glass separately were held constant throughout testing
according to the average gradations found from the gradation testing. Therefore the
gradations for each blend were slightly different depending on the maximum size of glass
used.

Moisture-Density Relationships
The moisture-density relationships were determined from compaction tests conducted in
accordance with AASHTO T99 – Standard Proctor Test. The optimum moisture content
at which the maximum density can be achieved is obtained from this testing. The
standard Proctor test uses a 5.5 lb rammer dropped from 12 in. A mechanical compactor
was used during testing to ensure uniform compaction of all test blends.
Each blend was prepared in the same fashion that was used for the R-value testing
and then was compacted. Moisture content versus dry density curves were then produced
for each blend.

DATA ANALYSIS AND RESULTS

Gradation
There were two objectives for conducting the gradation testing. The first objective was
to determine the particle size distributions for all the sources of aggregate and glass. In
doing this, a common or average gradation for the aggregate and a common or average
gradation for the glass could be established. These average gradations were used for
proportioning all of the aggregate-glass blends and remained constant throughout the
research. The decision was made to use the average of the glass gradations that were
already being produced from the glass sources, rather than specifying a new gradation. In
doing this, readily available recycled glass cullet could be used in practice, without
having to implement major changes in the glass crushing procedure. The gradations for
the aggregate and glass that were used for blending are shown in Table 3.
The glass gradation resulted in a relatively even distribution of particle sizes over
the range of sieve sizes 1” (25mm) to 3/8” (9.5mm). This provides constant shifts in the
particle size distribution when blended with other materials. Also notice the very low
amount of glass cullet passing the #4(4.75mm) sieve. An average of only 12% of the
glass cullet from crushers passed the #4(4.75mm) sieve; therefore the addition of glass
cullet from a crusher will have the most effect on the larger sieve sizes. It is possible to
increase the amount of fine material passing the #4 sieve by using glass pulverizers rather

TRB 2007 Annual Meeting CD-ROM Original paper submittal - not revised by author.
 Finkle, Ksaibati, and Robinson 6

than glass crushers. Glass pulverizers can produce cullet 3/8” and smaller, while glass
crushers produce cullet 2” and smaller. The amount of fine particle size cullet is minimal
if obtained from a glass crusher. A glass pulverizer should be considered if it is desired
to supplement the fine aggregate properties such as plasticity.
The second objective was to determine whether the glass gradation produced from
glass crushers was different from the gradation produced by the dozer crushing. This
comparison was made mainly for smaller agencies, to determine the feasibility of using
dozer crushed glass if commercial crusher cullet is not available. Using the particle size
distribution curves in Figure 1, it can be shown that gradations for one of the crushers and
the dozer were very similar. In fact, there was a greater difference between the two
crushers than between the crushers and the dozer. Between the crushers, there was up to
a 15% difference in the amount passing on the 1/2” (12.5mm) sieve and 10% difference
passing the 5/8” (15.9mm) and 3/8” (9.5mm) sieves. On the other hand, the one crusher
and the dozer differed 4% on material passing the 1/2” sieve (12.5mm) and had 0% or
1% difference in the percent passing all other sieves. The difference in gradations
between the crushers is most likely due to an adjustment that can be made to the steel
roller that crushes the glass. It also should be noted that, when the dozer was used to
crush the glass, whole bottles were visible in the pile, even after the dozer had passed
over it. The unbroken bottles were excluded from the gradation testing.

Resistance R-Value
The objective of conducting the Resistance R-Value test was to study the effects that
adding glass cullet to base course material had on strength and moisture susceptibility.
The strength and moisture susceptibility of a base course can be directly related to the
performance of the overlaying pavement, therefore knowledge of the effects of the
addition of glass on these two engineering properties is essential. The test was performed
three times for each blend, except for the two control mixes with 0% glass. These were
only conducted twice because numerous existing data on the material matched the two
test results for each aggregate. The R-Value was averaged across the three tests for each
blend.
The first aggregate tested was the STAR material. The STAR base course
material had a control R-Value (0% glass) of 78. This is above the WYDOT
specifications of 75 for base course material. The results of the R-value test for the
STAR aggregate are shown in Figure 2. The addition of glass cullet at a 10%
replacement rate show no significant effect for the maximum glass sizes of 3/8” and 1/2”,
but significantly lowered the strength at sizes of 5/8” and 3/4” to an average of 72. At a
20% replacement rate, the 3/8” size material still performed above the required
specification with a value of 76. The 1/2” size dropped significantly from 79 to 71 at 20%
replacement. At the larger sizes of glass the samples began falling apart in the
stabilometer which produced varying results. The larger glass particles were most likely
breaking as the load was applied, causing the layers inside the sample to move and break
apart. The same behavior was observed when 30% glass cullet was used. The 3/8” size
again slightly decreased and the larger sizes showed unusual testing behavior. Testing
with the rounded STAR particles was very inconsistent at higher replacement rates and
larger glass sizes.
The second aggregate tested was the LAF material. The LAF base course had a
control R-value of 80. This is again above WYDOT specifications for base course

TRB 2007 Annual Meeting CD-ROM Original paper submittal - not revised by author.
 Finkle, Ksaibati, and Robinson 7

material. LAF material showed very consistent behavior across all testing. The results
for the LAF R-value test are shown in Figure 3. Nine of the 12 blends had average R-
values above the control of 80. The three blends that did not increase in value had an
average of 79. This consistency could be due to the fact that the aggregate material is
very angular in nature, as is the glass cullet at all the sizes. The glass cullet behaved
better when used with a more angular aggregate possibly because interlocking of the
particles can still easily occur. With the rounded particles from the STAR material, the
glass cullet can slip when it come in contact with the rounded particles.
Statistical analysis of the R-value data indicates that of the 3 variables tested only
aggregate type and cullet size are statistically significant factors in predicting the strength
of the glass-aggregate blends. Replacement rate was not shown to significantly affect the
strength of the blends. This may be true because of the inconsistent results at the higher
replacement rates for the STAR aggregate as strength is expected to decrease at higher
cullet contents. The scatter plot shown in Figure 4 shows the relationship between R-
value or strength and cullet size. The squares indicate the LAF source of aggregate,
while the circles represent the STAR aggregate. The predicted regression indicates a
downward trend in R-value as cullet size increases for both the STAR and LAF
aggregate.

Moisture-Density Relationships
To determine the moisture-density relationships of the glass-aggregate blends the
standard proctor test was performed. Moisture-density curves were developed to
determine the optimum moisture content and maximum dry density for each glass-
aggregate blend. The moisture-density curves for each source are shown in Figures 5 and
6. Each figure shows the cullet size and the corresponding curves for the control and
10%, 20%, and 30% replacement rates.
The maximum density of the blends will be discussed first. Statistical analysis of
the curves show that type of aggregate and replacement rate are significant factors in
determining the maximum density of a blend. As shown in Figure 7, the maximum
density decreases as more cullet is introduced into the blend. This is true for both sources
tested. The decrease in density was slightly greater for the rounded STAR aggregate
versus LAF aggregate. At a 10% replacement rate, STAR aggregate remained at 99% of
the density found for the control mix. Approximately a 2% additional decrease occurred
at replacement rates of 20% and again at 30%. At the highest replacement rate tested the
density was still found to be a 95% of the control mix. Since most specifications allow
compaction at approximately 95% of maximum, this implies that cullet used at the
highest replacement rate tested would still meet this requirement. LAF aggregate
performed very similar, but at higher maximum densities. LAF exhibited only a slight
decrease (0.3%) when combined with 10% glass and only a 3% decrease in density from
the control at 30% glass.
The average maximum densities versus size are shown in Figure 8. The 95%
confidence intervals overlap across all sizes tested. This shows that size of cullet used
does not significantly affect the density that can be achieved. This is true for both
sources of natural aggregate tested. The confidence intervals are much larger for the size
factor compared to the replacement rate factor because it encompasses values of
maximum density across the three rates tested. Since the replacement rates of glass
significantly affects density, the three values for one size of glass are significantly

TRB 2007 Annual Meeting CD-ROM Original paper submittal - not revised by author.
 Finkle, Ksaibati, and Robinson 8

different leading to larger intervals. The same is true for the confidence intervals of
optimum moisture content.
The addition of glass cullet affected the optimum moisture content as well. Type
of aggregate and replacement rate are significant factors in determining optimum
moisture content, and again, cullet size is not. Figure 9 illustrates the relationship of
moisture content and replacement rate and Figure 10 shows the relationship between
moisture content and glass particle size. It was expected that as the glass content
increased the optimum moisture would decrease, but this was not observed. As cullet
content increased the moisture content at which maximum density could be achieve also
increased. This means that by adding glass, compaction or maximum density of mix is
slightly harder to achieve.
Again, the addition of glass with the rounded aggregate had greater affect on the
behavior of the material than it did when added to the angular aggregate. The optimum
moisture content for STAR aggregate increased approximately 7% from the moisture
content of the control when 10% glass was blended with the natural aggregate. The
increase was approximately 13% for 20% glass and 19% for 30% glass. The moisture
content for LAF aggregate increased approximately half as much as STAR for the same
replacement rates. The increase was approximately 3% when 10% cullet was added, 6%
when 20% cullet was added, and 9% when 30% cullet was added. The increase in
optimum moisture content as a function of replacement rate is shown in Figure 9.
The average optimum moisture contents versus size are shown in Figure 10.
Again, the 95% confidence intervals overlap across all sizes tested. This shows that size
of cullet used does not significantly affect the optimum moisture content that is required
to achieve maximum density. This is true for both sources of natural aggregate tested.

CONCLUSIONS AND RECOMMENDATIONS

The objectives of this study were to evaluate the strength and moisture susceptibility and
also the moisture-density relationships of natural aggregate and glass cullet blends. Two
different sources of aggregate, three cullet replacement rates, and four maximum cullet
sizes were evaluated. Analysis of the data indicated the following conclusions:
A dozer is an acceptable alternative to a glass crusher in terms of producing
similar glass gradations.
Aggregate type has a significant effect on engineering properties of glass-
aggregate blends. Specifically, the angularity of the natural aggregate will
influence the behavior of the mix. More angular aggregate performs better
when combined with glass cullet.
Strength of the glass-aggregate blend is dependent on the size of glass cullet
used in the mix and the aggregate type. The strength of the blend will decrease
as the cullet size increases when combined with either aggregate.
Cullet replacement rate is a significant factor in determining the maximum
density and optimum moisture content of glass-aggregate blends.
The addition of glass cullet decreases the maximum density that can be
achieved by the blend. Ninety-five percent of the control maximum density is
still obtained at all replacement rates tested.
The optimum moisture content is significantly affected by the addition of glass
cullet. An increase in optimum moisture content occurs as total cullet content in
the blend increases.

TRB 2007 Annual Meeting CD-ROM Original paper submittal - not revised by author.
 Finkle, Ksaibati, and Robinson 9

After conducting this evaluation of the glass-aggregate blends, it can be

concluded that crushed recycled glass, or cullet, can be used as a supplement to natural
road base material. This study recommends that the use of cullet be limited to a 20%
maximum replacement rate and a maximum size of 1/2”. It is more feasible to use glass
cullet with crushed, angular aggregate since it performs more consistently than when
blended with rounded natural aggregate.
Utilization of cullet mixed with natural aggregate could prove to be very
economically beneficial especially for rural locations. The feasibility and exact economic
benefits of using glass cullet for construction purposes should be considered on an
individual basis by agencies that are considering its use. Nonetheless, based on this
study, the strength and compaction properties of glass-aggregate blends are comparable
to those of natural road base material. The addition of glass to road base, at the
recommended levels, does not compromise the properties of the aggregate it is combined
with while providing additional environmental and economic benefits.

ACKNOWLEDGEMENTS
This study was funded by the U.S. Department of Transportation University
Transportation Program through the Mountain-Plains Consortium and the University of
Wyoming. The donations of material by ARK of Laramie, Wyoming, Jackson
Community Recycling, Fremont County, STAR Aggregates, and LaFarge Incorporated
were greatly appreciated. The authors would also like to express their appreciation to Jim
Kladianos of the Wyoming Department of Transportation for providing support in
conducting the research and to Jim Keenan, Basil Brookhouser, and Robert Gilmore of
the Wyoming Department of Transportation Central Laboratory for helping with the
testing of the samples.

TRB 2007 Annual Meeting CD-ROM Original paper submittal - not revised by author.
 Finkle, Ksaibati, and Robinson 10

REFERENCES

1. United States Environmental Protection Agency (USEPA). “Municipal solid

waste generation, recycling, and disposal in the United States: Facts and
Figures for 2003,” Report No. EPA530-F-05-003, Washington, D.C.,
http://www.epa.gov/epaoswer/non-hw/muncpl/pubs/msw05rpt.pdf, April,
2005.
2. Nash, P.T., Jayawickrama, P., Tock, R., Senadheera, S., Viswanathan, K. and
Woolverton, B., “Use of Glass Cullet in Roadway Construction: Laboratory
Testing and Specification Development,” Research Study No. 0-1331-2F,
Texas Department of Transportation, July 1996.
3. Shin, C. J., and Sonntag, V. “Using Recovered Glass as Construction Aggregate
Feedstock,” Transportation Research Record 1437, Transportation Research
Board, Washington, D.C., 1994.
4. Wartman, J., Grubb, D.G., Nasim, ASM, “Select Engineering Characteristics of
Crushed Glass,” Journal of Materials in Civil Engineering, American Society
of Civil Engineers, Vol. 16, Issue 6, Dec. 2004.
5. Dames and Moore, Inc., “Glass Feedstock Evaluation Project,” Reports 1-5, for
Clean Washington Center, Washington State Department of Trade and
Economic Development, March 1993.
6. Clean Washington Center, “A Tool Kit for the Use of Post-Consumer Glass as a
Construction Aggregate,” prepared by Soil & Environmental Engineers, Inc
and Re-Sourcing Associates, Inc., January 1998.
7. Nash, P.T., Jayawickrama, P., Tock, R., Senadheera, S., Viswanathan, K. and
Woolverton, B., “Use of Glass Cullet in Roadway Construction: Phase I.
Literature Review and Identification of Sources and Suppliers,” Research
Study No. 0-1331-1, Texas Department of Transportation, March 1995.
8. Minnesota Technology Transfer Local Technical Assistance Program,
“Reclaimed Glass,” Information Kit.
9. Wyoming Department of Transportation, “WYDOT Material Testing Manual,”
updated 2004.

TRB 2007 Annual Meeting CD-ROM Original paper submittal - not revised by author.
 Finkle, Ksaibati, and Robinson 11

LIST OF TABLES AND FIGURES

TABLE 1 Test Matrix for One of the Sources of Aggregate
TABLE 2 Aggregate Properties of the Two Sources Tested
TABLE 3 Aggregate and Glass Gradations and Specifications
FIGURE 1 Glass Crusher and Dozer Gradation Comparisons
FIGURE 2 STAR R-value Results
FIGURE 3 LAF R-value Results
FIGURE 4 Scatter Plot of median R-value vs. Size
FIGURE 5 STAR Aggregate Moisture Density Curves
FIGURE 6 LAF Aggregate Moisture Density Curves
FIGURE 7 Interval Plot of Maximum Density vs. Replacement Rate
FIGURE 8 Interval Plot of Maximum Density vs. Cullet Size
FIGURE 9 Interval Plot of Optimum Moisture Content vs. Replacement Rate
FIGURE 10 Interval Plot of Optimum Moisture Content vs. Cullet Size

TRB 2007 Annual Meeting CD-ROM Original paper submittal - not revised by author.
 Finkle, Ksaibati, and Robinson 12

TABLE 1 Test Matrix for Each Source of Aggregate

Replacement Rate

10% 20% 30%

Maximum Glass Size

3/4" 3 3 3

5/8" 3 3 3

1/2" 3 3 3

3/8" 3 3 3

TRB 2007 Annual Meeting CD-ROM Original paper submittal - not revised by author.
 Finkle, Ksaibati, and Robinson 13

TABLE 2 Aggregate Properties of the Two Sources Tested.

Coarse Bulk Specific

Fine Agg. Absorption
Aggregate Liquid Plasticity Agg. Gravity
Type Angularity
Sources Limit Index Angularity
(%)
(%)
FA CA FA CA

Pit-Run Non-
STAR 0 47.3 74 / 35 2.601 2.621 0.806 0.776
Base Plastic

Crushed Non-
LAF 0 46.7 99 / 96 2.57 2.616 0.969 0.715
Base Plastic

TRB 2007 Annual Meeting CD-ROM Original paper submittal - not revised by author.
 Finkle, Ksaibati, and Robinson 14

TABLE 3 Aggregate and Glass Gradations

AGGREGATE AND GLASS GRADATIONS USED FOR BLENDING

WYDOT Base
Aggregate Gradation Glass Gradation
Sieve Size Grading 'W'

Percent Passing
1 1/2" 100 100 100
1" 98 88 90 - 100
3/4" 91 74 --
5/8" -- 63 --
1/2" 73 48 60 - 85
3/8" 63 32 --
#4 46 12 45 - 65
#8 36 6 33 - 53
#30 20 2 10 - 30
#200 9.0 0.5 3 - 12

TRB 2007 Annual Meeting CD-ROM Original paper submittal - not revised by author.
 Finkle, Ksaibati, and Robinson 15

100

90

80

70

60
% Passing

50

40

30

20

10

0
0.1 #200 #100 #50 #30 1 #16 #8 #4 3/8"
10 5/8" 1" 1 1/2" 100
1/2" 3/4"
Sieve Size
ARK GRADATION JACKSON GRADATION FREMONT GRADATION(dozer)

FIGURE 1 Glass Crusher and Dozer Gradation Comparisons

TRB 2007 Annual Meeting CD-ROM Original paper submittal - not revised by author.
 Finkle, Ksaibati, and Robinson 16

84
82
80
78
76
R-Value

74
72
70
68
66
64
62
60
Control 10% 20% 30%
Glass Size

Control 3/8 in. 1/2 in. 5/8 in. 3/4 in.

FIGURE 2 STAR R-value Results

TRB 2007 Annual Meeting CD-ROM Original paper submittal - not revised by author.
 Finkle, Ksaibati, and Robinson 17

84
82
80
78
76
R-Value

74
72
70
68
66
64
62
60
Control 10% 20% 30%
Glass Size

Control 3/8 in. 1/2 in. 5/8 in. 3/4 in.

FIGURE 3 LAF R-value Results

TRB 2007 Annual Meeting CD-ROM Original paper submittal - not revised by author.
 Finkle, Ksaibati, and Robinson 18

84

82

80
R-value

78

76

74

72

70
0.375 0.500 0.625 0.750
Size

FIGURE 4 Scatter plot of median R-value vs. Size

TRB 2007 Annual Meeting CD-ROM Original paper submittal - not revised by author.
 Finkle, Ksaibati, and Robinson 19

STAR 3/8" Maxim um Glass Size STAR 1/2" Maxim um Glass Size

140.0 140.0

135.0 135.0

130.0 130.0

125.0 125.0

120.0 120.0
3.0% 4.0% 5.0% 6.0% 7.0% 8.0% 9.0% 10.0% 3.0% 4.0% 5.0% 6.0% 7.0% 8.0% 9.0% 10.0%
M o ist ur e C o nt ent ( %) M o ist ur e C o nt ent ( %)

Cont rol 10% 20% 30% Cont rol 10% 20% 30%

STAR 5/8" Maxim um Glass Size STAR 3/4" Maxim um Glass Size

140.0 140.0

135.0 135.0

130.0 130.0

125.0 125.0

120.0 120.0
3.0% 4.0% 5.0% 6.0% 7.0% 8.0% 9.0% 10.0% 11.0% 3.0% 4.0% 5.0% 6.0% 7.0% 8.0% 9.0% 10.0% 11.0%
M o ist ur e C o nt ent ( %) M o i st ur e C o nt ent ( %)

Cont rol 10% 20% 30% Control 10% 20% 30%

FIGURE 5 STAR Aggregate Moisture Density Curves

TRB 2007 Annual Meeting CD-ROM Original paper submittal - not revised by author.
 Finkle, Ksaibati, and Robinson 20

LAF 1/2" Maxim um Glass Size LAF 3/8" Maxim um Glass Size

145.0 145.0

Dry Density (pcf)

140.0
Dry Density (pcf)

140.0

135.0
135.0
130.0

130.0
125.0
3.0% 4.0% 5.0% 6.0% 7.0% 8.0% 9.0% 10.0%

125.0 Moisture Content (%)

3.0% 4.0% 5.0% 6.0% 7.0% 8.0% 9.0% 10.0%

Moisture10%
Control Content
20%(%) 30% Control 10% 20% 30%

LAF 5/8" Maxim um Glass Size LAF 3/4" Maxim um Glass Size

145.0 150.0

145.0
140.0
140.0

135.0 135.0

130.0
130.0
125.0

125.0 120.0
3.0% 4.0% 5.0% 6.0% 7.0% 8.0% 9.0% 10.0% 3.0% 4.0% 5.0% 6.0% 7.0% 8.0% 9.0% 10.0% 11.0%
M o i st ure C o nt ent ( %) M o i st ur e C o nt ent ( %)

Control 10% 20% 30% Cont rol 10% 20% 30%

FIGURE 6 LAF Aggregate Moisture Density Curves

TRB 2007 Annual Meeting CD-ROM Original paper submittal - not revised by author.
 Finkle, Ksaibati, and Robinson 21

Interval Plot of max density vs type, rate

95% CI for the Mean
144

142

140
max density (pcf)

138

136

134

132

130
rate 0.0 0.1 0.2 0.3 0.0 0.1 0.2 0.3
type STAR LAF

FIGURE 7 Interval Plot of Maximum Density vs. Replacement Rate

TRB 2007 Annual Meeting CD-ROM Original paper submittal - not revised by author.
 Finkle, Ksaibati, and Robinson 22

Interval Plot of max density vs type, size

95% CI for the Mean

145

140
max density (pcf)

135

130

125
size 0.000 0.375 0.500 0.625 0.750 0.000 0.375 0.500 0.625 0.750
type STAR LAF

FIGURE 8 Interval Plot of Maximum Density vs. Cullet Size

TRB 2007 Annual Meeting CD-ROM Original paper submittal - not revised by author.
 Finkle, Ksaibati, and Robinson 23

Interval Plot of optimum moisture vs type, rate

95% CI for the Mean

8.2

8.0

7.8
optimum moisture

7.6

7.4

7.2

7.0

6.8

rate 0.0 0.1 0.2 0.3 0.0 0.1 0.2 0.3

type STAR LAF

FIGURE 9 Interval Plot of Optimum Moisture Content vs. Replacement Rate

TRB 2007 Annual Meeting CD-ROM Original paper submittal - not revised by author.
 Finkle, Ksaibati, and Robinson 24

Interval Plot of optimum moisture vs type, size

95% CI for the Mean
9.0

8.5
optimum moisture

8.0

7.5

7.0

6.5
size 0.000 0.375 0.500 0.625 0.750 0.000 0.375 0.500 0.625 0.750
type STAR LAF

FIGURE 10 Interval Plot of Optimum Moisture Content vs. Cullet Size

TRB 2007 Annual Meeting CD-ROM Original paper submittal - not revised by author.