TIRE CHIPS AS

LIGHTWEIGHT SUBGRADE FILL

AND RETAINING WALL BACKFILL

Dana N. Humphrey
and
Thomas C. Sandford

Associate Professors
Department of Civil Engineering
5711 Boardman Hall
University of Maine
Orono, Maine 04469-5711
Phone: (207) 581-2176
FAX: (207) 581-2202

Submitted to:
Symposium on Recovery and Effective Reuse of
Discarded Materials and By-Products for
Construction of Highway Facilities
Denver, Colorado
October 19-22, 1993
 TIRE CHIPS AS LIGHTWEIGHT SUBGRADE FILL
AND RETAINING WALL BACKFILL

DANAN.HUMPHREYANDTHOMASC.SANDFORD
DEPARTMENT OF CIVIL ENGINEERING
UNIVERSITY OF MAINE
ORONO, MAINE

ABSTRACT
Scrap tires that have been cut into chips are coarse grained, free draining, and have a
low compacted density, thus offering significant advantages for use as lightweight
sub grade fill and retaining wall backfill. This paper presents the engineering properties
that are needed use tire chips for these purposes. The following properties were
determined for tire chips from three suppliers: gradation, specific gravity, compacted
density, shear strength, compressibility, coefficient of lateral earth pressure at rest, and
permeability. For most of these properties tests were also performed on tire chips from a
fourth supplier. The maximum size of the tire chips tested was 76 mm (3 in.). This large
size, in combination with the high compressibility of the tire chips, made it necessary to
design and fabricate custom testing equipment.
The tests showed that the tire chips are composed of uniformly graded, gravel sized
particles that absorb only a small amount of water. Their compacted dry density is 0.618
to 0.642 Mglm 3 (38.6 to 40.1 pcf) which is about 113 that of compacted soils. The shear
strength was measured in a large scale direct shear apparatus. The friction angle and
cohesion intercept ranged from 19° to 25° and 8 to 11 kPa (160 to 240 psf), respectively.
The compressibility tests showed that tire chips are highly compressible on initial loading
but that the compressibility on subsequent unloadinglreloading cycles is less. The
horizontal stress was measured during these tests and showed that the coefficient oflateral
earth pressure at rest varied from 0.26 for tire chips with a large amount of steel belt
exposed at the cut edges to 0.47 for tire chips composed of entirely of glass belted tires.
The permeability of tire chips was measured in a 305-mm (l2-in.) diameter permeameter.
The permeabilities ranged from 1.5 to IS cm/s.
 TIRE CHIPS ASLIGHTWEIGHT SUBGRADE FILL
AND RETAINING WALL BACKFILL

Dana N. Humphrey and Thomas C. Sandford

INTRODUCTION

An estimated 2 billion scrap tires have been discarded in huge open piles across the
United States. Furthermore, an additional 189 million are added to these piles each year
(EP A, 1991). These piles are a serious fire hazard, a prolific breeding ground for
mosquitoes, and an ugly scar on our landscape. The transportation industry may be able
to make an important contribution to solving this problem especially by using waste tires
as fill. In this application the tires are cut into chips that are durable, coarse grained, free
draining, a good thermal insulator, and most importantly, have a low compacted density.
Tire chips have already been used as lightweight fill for highway embankments (e.g.,
Dodson, 1990; Edil, et aI., 1990; Geisler, et aI., 1989) and as an insulating layer beneath
an unpaved road in a northern climate (Humphrey and Eaton, 1993a,b). Another use
would be as backfill behind retaining walls and bridge abutments where the low density of
tire chips would potentially result in lower pressures on the wall and lower foundation
settlement.
The potential for tire chip fill to make an important contribution to disposal of waste
tires is clear when it is considered that each cubic meter of fill contains about 100 waste
tires (75 tires/yd 3 ). As an illustration, a recent 180-m (600-ft) long test section in Maine
used over 20,000 waste tires (Humphrey and Eaton, 1993a,b). In contrast, the Intermodal
Surface Transportation Efficiency Act (ISTEA) required that, by 1997, one-fifth of all
road projects must include 10 kg of recycled tire rubber per megagram (20 lb/ton) of hot
mix (IS TEA, 1991). This translates into a mere 1. 5 tires per megagram (1.4 tires/ton) of
hot mix.
Use of tire chip fill also has the potential to reduce the cost and improve the
performance of some transportation facilities. It is sometimes necessary to build highway
embankments over weak or compressible soils. An example would be an approach fill for
a bridge abutment founded on a weak clay slope where the weight of soil fill would cause
slope instability. One solution would be to construct the embankment from lightweight
fill. However, traditional lightweight fill such as expanded shale are very expensive
($57/m 3; $44/yd 3 in Maine; Manion and Humphrey, 1992). In contrast, tire chip fill can
be obtained for less than $12/m 3 ($9.50/yd 3). The low unit weight of tire chips would
reduce long term settlement, thereby improving performance and reducing maintenance
costs. When used as retaining wall backfill, the low compacted density of tire chips would
potentially result in low horizontal pressures on the wall. Thus, a lighter, less expensive
wall could be used. Furthermore, their low compacted density would reduce the
settlement of underlying compressible soils and would increase the global stability of the
wall. In some cases, this would allow the wall to be placed on a spread footing rather than
a pile foundation, which would significantly reduce construction costs. Since tire chips are
free draining, there would be no need for clean granular backfill.
 ~ -
Humphrey and Sandford Page 2

A necessary first step to fully implement the applications described above is to

determine the engineering properties of tire chips. The gradation, specific gravity,
compacted density, and compressibility of tire chips from Sawyer Environmental Recovery
in Hampden, Maine, were determined in a preliminary study (Manion and Humphrey,
1992; Humphrey and Manion, 1992). In a subsequent study (Humphrey, et aI., 1992,
1993) these properties were determined for tire chips from the following three additional
suppliers: F&B Enterprises, New Bedford, Massachusetts; Palmer Shredding, North
Ferrisburg, Vermont; and Pine State Recycling, Nobleboro, Maine. In addition, the shear
strength and permeability of the tire chips from the three additional suppliers was
detennined. These studies provided the basis for full-scale field trials that will be
conducted by the University of Maine in 1993 using tire chips as fill beneath paved roads
and as retaining wall backfill.
The composition of the tire chips from each supplier spanned a wide range. The F&B
chips were composed entirely of glass belted tires and were less than 38 mm (1-112 in.) in
size while the Sawyer, Palmer, and Pine State chips were composed of a mixture of glass
and steel belted tires. The Palmer chips had a significant amount of steel belt exposed at
the cut edges of the tire chips. The Palmer chips were 76-mm (3-in.) maximum size while
the Sawyer and Pine State chips were 51-mm (2-in.) maximum size. The Sawyer, Palmer
and Pine State chips tended to be long relative to their thickness where as the F&B chips
tended to be more equidimensional.
The large size and high compressibility of the tire chips necessitated that conventional
test procedures be modified, and that custom testing apparatus be designed and fabricated.
The test procedures, apparatus, and results for each property will be discussed below.
Further details are given in Manion and Humphrey (1992), Humphrey and Manion (1992),
and Humphrey, et al. (1992).

GRADA TION, SPECIFIC GRAVITY, AND COMPACTED DENSITY

The gradation of the tire chips from the four suppliers was determined using
AASHTO T 27-87 (AASHTO, 1986). Fig. I shows that the tire chips are uniformly
graded and composed of gravel sized particles. The Palmer chips were the coarsest and
the F&B chips were the finest.
The specific gravity of the tire chips was determined, using AASHTO T 85-85
(AASHTO, 1986), except that the samples were air dried rather than oven dried at the
start of the tests. The apparent specific gravities based on the average of two tests were
1.14 for F&B Enterprises, 1.27 for Palmer Shredding, 1.24 for Pine State Recycling, and
1.23 for Sawyer Environmental. These specific gravities are less than half of those typical
of soils. The specific gravity of the F&B chips is lower than the other three since the
former are entirely glass belted.
The test procedure used to determine the compacted density of air dried tire chips
was adapted from AASHTO T 180-86 (AASHTO, 1986). A 254-mm (IO-in.) diameter
by 254-mm (IO-in.) high mold with volume of 0.012 m3 (0.44 ft3) was used. The tire
chips were compacted in three layers with a 4.536 kg (I0-lb) hammer falling 0.457 m (18
 ~ -
Humphrey and Sandford Page 3

I
80
c..? GRAVEL I SAND
Z
(f)
(f)
« 130
[L

I-
Z
w
u 40
oc
w
[L

20

~OO 10 Q..Q.!JJJJl Palmer Shredding Inc, vr

QQ.Q.Q.O F &. B Enterprises, MA
GRAIN SIZE (mm) ~ Pine Stale Recycling, ME
~ Sawyer Environment Rec. Foc .. ME

Fig. 1 Gradations of tire chips from four suppliers (Humphrey, et aI., 1992).

in.). The initial study (Manion and Humphrey, 1992; Humphrey and Manion, 1992)
showed that decreasing the compaction energy from modified Proctor to 60% of standard
Proctor reduced the density by only 0.03 Mglm 3 (2 pcf) and that compaction of wet
versus air dried tire chips made only a 0.016 Mglm 3 (1 pcf) difference in the density.
Since the compaction energy and wet versus air dried tire chips had only a small effect,
60% of standard Proctor energy and air dried tire chips were used for this study. The
compacted density of air dried tire chips from the four suppliers fell within a fairly narrow
range. The compacted density based on the average of three tests was 0.618 Mglm 3 (38.6
pcf) for F&B Enterprises, 0.619 Mglm 3 (38.7 pcf) for Palmer Shredding, 0.642 Mglm 3
(40.1 pcf) for Pine State Recycling, and 0.625 Mglm 3 (39.0 pcf) for Sawyer
Environmental. These values are about ]/3 of those typical for compacted soils showing
the potential of tire chips to be used as lightweight fill.

SHEAR STRENGTH

Testing apparatus
The shear strength of tire chips was determined using a direct shear apparatus that
was custom designed to accommodate the large size and high compressibility of the tire
chips. In addition, special provisions were made to eliminate friction between the two
halves of the shear box.
 ~ -
Humphrey and Sandford Page 4

A 305-mm (l2-in.) square shear box (nominal dimension) made from 9.5-mm (3/8-
in.) thick steel was chosen for the initial design. This was felt to be adequate since the
largest size tire chips to be tested were minus 76-mm (3-in.). Thus, the shear box would
be four times larger than the largest tire chip. The lower half of the shear box was 76 mm
(3 in.) high and was bolted to a supporting bench. The top half of the shear box was 152
mm (6 in.) high. This height was needed to accommodate the large compressibility of the
tire chips. To determine if the area of the shear box influenced the test results, a 406-mm
(16-in.) square (nominal dimension) shear box was also fabricated.
It was essential to maintain a gap between the two halves of the box to prevent
introduction of additional horizontal stresses due to friction. During sample preparation a
6-mm (l14-in.) gap was created by placing spacers at each corner between the halves of
the box. Then to maintain the gap during testing, two 51-mm (2-in.) diameter steel wheels
with low friction ball bearing hubs (similar to old fashioned roller skate wheels) mounted
in a steel frame were clamped to each end of the box. The spacers were then removed.
During testing the wheels rode along the top of the supporting bench, carrying the top of
the box to maintain the gap between the box halves.
The normal stress was applied using dead weights hung from a hanger suspended
beneath the sample. A maximum dead load of5570 kN (1250 Ib) could be applied to the
sample. This resulted in a maximum normal stress of68 kPa (9.9 psi) for the 305-mm box
which is equivalent to approximately 3 m (lOft) of soil fill.
The horizontal shearing force was provided by a 1I8-hp electric motor acting through
a gear box which allowed the rate of horizontal deformation to be adjusted. A rate of
approximately 7.6 mmlmin (0.3 in/min) was used. The horizontal shearing force was
measured with a 4450-kN (lOOO-lb) capacity load cell. Two linear variable differential
transformers (L VDTs) were used to measure horizontal and vertical displacements. A
schematic of the apparatus is shown in Fig. 2.

Sample Preparation
The inside of the upper half of the shear box was greased to minimize the portion of
the applied vertical load that would be transmitted to the sides of the box by friction.
Then, the samples were compacted with 60% of standard Proctor energy. The box was
filled in three 64 to 76-mm (2.5 to 3-in.) layers to approximately 25 mm (I in.) from the
top. To make sure that there was no effect on the shearing plane from a smooth surface
between the first two layers, care was taken that the top of the first layer did not coincide
with the gap between the halves of the box.

Results
Direct shear tests were run using the 305-mm box at three normal stresses. Three
tests were done for each normal stress for tire chips from each of the following suppliers:
F&B Enterprises, Palmer Shredding, and Pine State. A total of 27 tests were done with
the 305-mm box. In addition, tests were done using the 406-mm box with Pine State tire
chips with three normal stresses.
 -:Ji -

Humphrey and Sandford Page 5

ru'IUU R:R YffiTlCN.. lYDT

~ CROSS ruR'-,
~~==~~------~~
Sf€M fICl(
(top t...Ul

/-11----- I..CI'.D cal

' - I - / - - Sf€M fICl(

Il:oHon hoUl

-10--- SJ'ftRT FlI'JE

Fig. 2 Schematic oflarge scale direct shear apparatus.

In direct shear tests, failure is considered to be the peak shear stress or, if no peak is
reached, failure is generally taken as the shear stress at a horizontal displacement equal to
10% of the length of box (AASHTO, 1986). The latter criterion controlled for tire chips.
Thus, for the 305-mm box, which had an inside dimension of 286 mm (11.25 in.), failure
was taken as the shear stress at a deformation of28.6 mm (1.1 in.). For the 406-mm box,
which had an inside dimension of387 mm (15.25 in.), failure would be at a deformation of
38.7 mm (1.5 in.). However, the travel of the L VDT used to measure horizontal
displacement was limited to 35.6 mm (1.4 in.), so failure for tests with this box were taken
to be the stress at this displacement.
The shear stress versus horizontal deformation for Pine State tire chips with the 305-
mm box is shown in Fig. 3. This shows that the shear stress continues to increase past a
horizontal deformation equivalent to 10% of the length of the box. The curves for the
F&B and Palmer chips in the 305-mm box and the Pine State chips in the 406-mm box
were similar (Humphrey, et ai., 1992).
Humphrey and Sandford Page 6

40,----.----.-----.----.----.-----.----.----,
APPROXIMATE
NORMAL STRESS:
GOO8EJ 68 kPo
Gl3I3BEJ 34 kPo
~ 17 kPo

Vl
Vl
W
g: 20
Vl

HORIZONTAL DEFORMATION (mm)

Fig. 3 Shear stress versus horizontal displacement for Pine State chips.

The average shear stress versus the average normal stress at each of the three loading
increments for each of the samples is shown Fig. 4. Each point is the average of two or
three trials at a given normal stress. All of these lines plot slightly concave down. Also,
Fig. 3 shows that for the Pine State tire chips, the 305-mm and 406-mm boxes give nearly
identical results. Thus, the 305-mm box is large enough for the size tire chips investigated.
Comparing the failure envelopes in Fig. 4, the F&B chips are stronger than the others.
This may be because these tire chips were smaller and more equidimensional. During
shearing these tire chips would tend to lock together more instead of having the chance to
slide past one another on the shearing plane as with the larger, flatter pieces. This is
particularly true since the large flat pieces tended to be oriented parallel to the horizontal
shear plane.

60,------,------,-------,------,
00000 PINE STATE, 30S-mm BOX
t:,.t;0.t;0. PINE STATE, 406-mm BOX
00000 PALMER, 30S-mm BOX

UJ 40
***** F&B, 30S-mm BOX
UJ
w
a::
f-
UJ

a::
« 20
w
I
UJ

°0~----~2~0----~4~0----~6~O----~80
NORMAL STRESS (kPo)

Fig. 4 Failure envelopes for tire chips from three suppliers.

Humphrey and Sandford Page 7

The friction angles <p and cohesion intercepts c were determined using best fit straight
lines through the data and are summarized in Table 1. This shows that the Palmer chips
had the highest cohesion even though their friction angle was low. This may be because
they have a large amount of exposed steel belts which interlock and do not rely on normal
stress to develop their strength.

Table I Values of <p and c from the direct

shear tests.
Supplier <p c
(kPa)
F&B Enterprises 25° 8.6
Palmer Shredding 19° 11.5
Pine State (305-mm box) 21° 7.7
Pine State (406-mm box) 26° 4.3
1 kPa = 20.89 psf

Choosing failure as the shear stress at a horizontal deformation equal to 10% of the
length of the box (i.e., 28.6 mm for the 305-mm box) is rather arbitrary. To investigate
what effect the failure criteria could have, the <p and c were also determined for shear
stresses at 15.2 mm (0.6 in.) and 35.6 mm (1.4 in.) of horizontal deformation. The results
are summarized in Figs. 5 and 6. Fig. 5 shows that the friction angle decreases somewhat
at the deformation chosen as failure decreases. However, Fig. 6 shows that the cohesion
intercept decreases significantly as the deformation decreases. This suggests that large
deformations are required to develop the cohesion intercept but that the friction angle is
less sensitive to the level. of deformation.

30

25
U)
w
w
a: 20
D
W
e.
w
..J 15
D
Z
«
z
0 10 Ii!i!!I F&B ENTERPRISES
;::
'-'
5" ~ PALMER SHREDDING
u..
5 .f>. PI NE STATE

0
0 5 10 15 20 25 30 35 40
FAILURE CRITERION (mm)

Fig. 5 Friction angle versus horizontal deformation chosen as failure criterion.

Humphrey and Sandford Page 8
15,---.----.----.---,----,---.----om__,

r0-
o..
C.
f- 10
0..
W
~
W
f-
:;;:;
Z
0
05 5
w
:r:
0 Iii!!! F&B ENTERPRISES
'-'
@ PALMER SHREDDING

A PINE STATE

0
0 5 10 15 20 25 30 35 40
FAILURE CRITERION (mm)

Fig. 6 Cohesion intercept versus horizontal deformation chosen as failure criterion.

There are two items that should be given special consideration when choosing the
strength parameters of tire chips for use in stability analyses. The first is that it takes
significant deformation to develop the cohesion. These deformations may be unacceptable
for some applications. In these cases it would be prudent to take the cohesion intercept to
be zero. The second item is strain compatibility between the tire chips and the foundation
soil. If the foundation soil develops its peak strength at small deformations and then
undergoes strain softening, as would occur with sensitive clays, failure would occur before
the tire chips had undergone sufficient deformations to develop their strength. Thus, for
failure surfaces that pass through the tire chips into strain softening foundation soils, the
tire chips should be assigned a very low or zero strength.

COMPRESSIBILITY

Testing apparatus

Sample Container
The container used for the compression tests consisted of a piece of 305-mm (l2-in.)
diameter (nominal), schedule 40 PVC pipe that was 318 mm (12.5 in.) long and had a wall
thickness of 8.1 mm (0.32 in.). Four strain gages were placed with a horizontal
orientation 89 mm (3.5 in.) above the base. They were calibrated to give the horizontal
stress exerted on the inside of the container by the tire chips. An additional two strain
gages were placed vertically. They were calibrated to measure the portion of the applied
load transferred from the tire chips to the container by friction (Manion and Humphrey,
1992; Humphrey and Manion, 1992).
Humphrey and Sandford Page 9

Loading and dala acqllisilion syslem

An Instron 4204 universal testing machine controlled by an 80286 computer was used
to apply the vertical load. The computer controlled the rate of deformation and obtained
measurements of the vertical load and vertical displacement. A wheatstone bridge was
used to take readings from the strain gages. The output voltages from the bridge were
read by an analog to digital converter with an accuracy of 16 bits. The readings were
taken at ten second intervals. To help offset electronic noise and imbalance at the time of
a reading, the computer would take ten readings from each strain gage, which were
averaged for the final reading (Manion and Humphrey, 1992; Humphrey and Manion,
1992).

Testing Methodology

Sample preparalion
Compacted samples were prepared by clamping the container to the steel base plate.
Grease was brushed on the inside of the container to reduce the friction between the tire
chips and the wall of the container. The tire chips were compacted in five layers with 60%
of standard Proctor energy (Humphrey and Manion, 1992; Manion and Humphrey, 1992).
The sample was then placed in the Instron and the clamps were removed.

Dala acqllisilion and slress compulalions

The load was applied to the sample at a constant rate of deformation of 13 mm/min
(0.5 in.lmin). Readings from the strain gages, vertical load, and vertical deformation were
taken every 10 seconds;. From these readings the following were calculated: average
vertical stress in the sample (O"avg), the vertical strain (lOy), vertical stress in the sample at
the strain gage height (O"gage), and the horizontal stress at the gage height (O"h). The
relationship between O"avg and the known stresses at the top of the sample (O"top) and
O"gage is illustrated in Fig. 7. The vertical stress at gage height (O"gage) is found by
subtracting the load transmitted by friction to the container as measured at the gage height
(PfricU from the load applied at the top of the sample (P applied) and then dividing by the
area of the sample. The average vertical stress (O"avg) is the vertical stress at mid height of
the sample. It was computed by assuming that the load transmitted by friction to the
container varies linearly from zero at the top of the sample to a maximum at the bottom.
Since the strain gages are located at 89 mm from the bottom, the load carried at mid
height (P avg) is given by
P avg = Papplied - [(HJ2) * Pfrict/(H-89)] ................................... (1)

where H is the current height of the sample in mm. The average vertical stress (O"avg) is
found by dividing P avg by the area of the sample.
Humphrey and Sandford Page IO

APPLIED VERTICAL STRESS

PVC AT TOP 0TOP
CYLINDER
VERTICAL STRESS IN SAMPLE
SIDE
FRICTION OTOP (APPLIED)
TIRE
CHIPS H/2
VERTICAL
STRAIN
GAGE
VERTICAL
STRAIN
GAGE
__I
1H(2T!L- -OAVG (ESTIMATED)

S( mt
0GAGE (MEASURED)

, 0SOTTOM (ESTIMA TED)

Fig. 7 Effect of friction on vertical stress in sample.

Loading and unloading cycles

Most samples were subjected to 3 cycles of loading and unloading. The
loading/unloading cycles' are of particular importance for highway applications since they
indicate the deformation behavior that would occur under repetitive vehicle loading. To
apply the first loading cycle, the vertical load (P applied) was increased until it reached 40
kN (9000 Ib). This was chosen as the upper limit ofloading since it is near the maximum
capacity of the Instron. The clamps that held the container to the base were put in place
and the sample was then unloaded until the average vertical load in the middle of the
sample (P avg) was reduced to 3.0 kN (670 Ib) or about 41 kPa (6 psi). This process
continued until 3 cycles of loading and unloading had been performed. The clamps were
left in place for the second and third loading/unloading cycles.

Results
For each supplier, 3 tests were performed on samples compacted with 60% of stan-
dard Proctor energy. Most of these tests consisted of 3 loading and unloading cycles.
Vertical compressibility and horizontal stresses are discussed below in separate sections.
Selected results are presented to illustrate the general compressibility behavior. Summa-
ries are made to permit a comparison of the compressibility of tire chips from the four
suppliers. Then elastic parameters that were computed from the combined measurements
of vertical compressibility and horizontal stresses are presented. Complete compressibility
results are given in Manion and Humphrey (1992) and Humphrey, et al. (1992).
Humphrey and Sandford Page II

Vertical compressibility
Results from one test on Palmer Shredding chips are shown in Fig. 8 to illustrate a
typical graph of vertical strain (tv) versus average vertical stress (cravg)' The initial
portion of the first loading curve is very steep indicating high compressibility. The first
loading curve then flattens out at higher stresses. The slope of subsequent unloading and
reloading curves are similar to the flatter part of the first loading curve. The reloading
curves lie slightly above the unloading curves. Tests on tire chips from the other suppliers
showed similar behavior.
To permit a comparison of the initial compressibility, the vertical strain for the first
loading cycle at average vertical stresses of 69 kPa (10 psi) and 276 kPa (40 psi) are
shown in Table 2. Each value is the average of three to five tests. Ordering the results
from least to most compressible (i.e., Sawyer, F&B, Pine State, Palmer) shows that there
is a general trend of increasing compressibility with increasing amounts of exposed steel
belts. However, from a practical viewpoint, the differences in compressibility between tire
chips from the four suppliers is small.

Horizontal stress
The horizontal strain gages were used to measure the increase in horizontal stress as the
sample was loaded. A typical graph of horizontal stress (crh) versus vertical stress at gage
height (crgage ) for compacted Pine State tire chips is shown in Fig. 9. For the initial
loading the graphs show a flatter slope up to a horizontal stress of approximately 69 kPa
(10 psi). After this point the line is steeper. This change in slope coincides with the point
where the calibration curve for the horizontal strain gages changes from a straight line for
stresses less than 69 kPa 'to a second order polynomial for higher stresses. This causes the
distinct transition at 69 kPa. Nonetheless, the initial portion of the curve has a flatter
slope. It has been theorized that the flatter initial slope is due to the compression of the
voids and the steeper upper portion is due mqinly to deformation of the rubber particles
(Manion and Humphrey; 1992). Tests on tire chips from the other suppliers showed a
similar behavior (Manion and Humphrey, 1992; Humphrey, et aI., 1992).

Elastic parameters
Elastic parameters were calculated using the measurements of vertical compressibility
and horizontal stress for unloading/reloading at low stress levels. Calculation procedures
are given in Manion and Humphrey (1992), Humphrey and Manion (1992), and
Humphrey, et al. (1992, 1993). The elastic parameters are summarized in Table 3. Each
value is the average of the results of two to five tests.
Examining the coefficient oflateral earth pressure at rest (Ko) and Poisson's ratio (~),
and recalling the amount of exposed steel belt in the tire chips from the different suppliers,
it is seen that these parameters decrease with increasing exposed steel belt. The
implication is that tire chips with a significant amount of exposed steel belt would produce
lower horizontal stresses on retaining walls. It is instructive to compare the Ko and ~ for
 ~ -
Humphrey and Sandford Page 12

-10
~
z
<i' -20
(l'
~
<Jl

-'
C5 -30
f=
(l'
w
>
-40

-5~--~~~b-~~~~~
~ 100 200 300 400 500
AVERAGE VERTICAL STRESS (kPo)

Fig. 8 Deformation behavior of Pine State tire chips.

Table 2 Vertical strain at average vertical stresses of69 kPa

and 276 kPa.
Vertical Strain* At Average
Vertical Stress =
Supplier 69 kPa 276kPa
F &B Enterprises 23.1 37.6
Palmer Shredding 29.0 42.5
Pine State Recycling 27.3 38.4
Sawyer Environmental 21.9 35.9
1 kPa = 0.1450 psi
*Each value is the average of the results of three to five tests.

Ul 300
Ul
w
(l'
~

Ul 200
--'
;::
z
o
'::' 100
(l'
o
:r::

0~~~1~ObO~-02*OO~~3UObO~~4+0~0--~5~00
AVERAGE VERTICAL STRESS (kPa)

Fig. 9 Horizontal stress versus vertical stress for Pine State tire chips.
Humphrey and Sandford Page 13

Table 3 Summary of elastic parameters* for compacted samples.

Supplier Ko ~ D (kPa) E (kPa)
F&B Enterprises OA7 0.32 1270 770
Palmer Shredding 0.26 0.20 1680 1120
Pine State Recycling 0 Al 0.28 1470 1130
Sawyer Environmental OA4 0.30 1730 1250
I kPa = 0.1450 psi
*Each value is the average of the results from two to five tests.

tire chips to values typical for granular soils. The average Ko values for tire chips ranged
from 0.26 to OA7 as compared to typical Ko of normally consolidated granular soils of
0.35 to 0.50 (Holtz and Kovacs, 1981). Thus, only the Ko for the Palmer sample falls
below the typical range for granular soils. Typical ~ for granular soils range from 0.15 to
OA5 (Holtz and Kovacs, 1981). The average values for tire chips (0.20 to 0.32) fall in the
lower half of this range. For comparison the ~ of solid tire rubber is 0.5 (Beatty, 1981).
The constrained modulus (D) of the tire chips ranged from 1270 kPa (184 psi) for the
F&B tire chips to 1730 kPa (251 psi) for the Sawyer tire chips. Young's modulus (E)
ranged from 770 kPa (112 psi) for the F&B chips to 1250 kPa (181 psi) for the Sawyer
chips. This suggests that small, glass belted tire chips have lower unloading/reloading
modulus than mixtures of larger, mixtures of glass and steel belted tire chips. For
comparison, the Young's modulus of the tire rubber itself ranges from 1240 to 5170 kPa
(180 to 750 psi) (Beatty, 1981) and for granular soils it typically ranges from 10,000 to
170,000 kPa (1,500 to 25,000 psi; Das, 1990). Thus, the Young's modulus of tire chips is
2 to 3 orders of magnit~de less than the modulus of granular soils typically used as a base
beneath paved roads. The implication of this is that 0.6 to 1.8 m (2 to 6 ft) of
conventional soil fill is needed on top of the tire chip layer to prevent excessive deflections
of the overlying pavement. Additional discussion of this statement are given in previously
published works (Manion and Humphrey, 1992; Humphrey and Manion, 1992). These
parameters were also used as input to numerical analyses of tire chips as fill beneath paved
roads and as retaining wall backfill (Gharegrat, 1993).

PERMEABILITY

The permeability of tire chips from F&B, Palmer, and Pine State at several different
void ratios was determined in a custom built apparatus. The apparatus, test procedure,
and results are discussed in the following sections. Further details are given in Humphrey,
et al. (1992).

Testing apparatus
The permeability of tire chips was determined using a constant head apparatus whose
design was based on one give by the California Department of Transportation (Bressette,
1984). For the present study, the body of the permeameter consisted ofa 0.96-m (38-in.)
 .;J} -

Humphrey and Sandford Page 14

long PVC pipe with an inside diameter of 295 mm (11.6 in.). The bottom of the pipe was
sealed with an end cap. A 38-mm (l.S-in.) diameter water inlet was fixed to the center of
the end cap. A 100-mm wide by SO-mm deep (4-in. by 2-in.) slot was cut into the top of
the PVC pipe to allow flow of water out of the top of the apparatus. The general design
of the apparatus is shown in Fig. 10.
The sample was supported by a perforated plate constructed of two thicknesses of
19-mm (3/4-in.) plywood glued together. A similar perforated plate was placed on top of
the sample. The initial length of the sample was about 600 mm (24 in.). A 90-kN
(10-ton) hydraulic jack, pressing on the crosspiece of a steel frame, was used to compress
the tire chips by the desired amount as shown in Fig. 10.
Head loss as the water flowed through the tire chips was determined by the difference
between piezometeric heads measured with four standpipes fixed to the side of the PVC
container (Fig. 10). To determine the corresponding quantity of flow, the overflow was
collected for a measured time interval. The quantity of flow was determined by weighing
the collected water.

(ALL DIMENSIONS IN INCHES)

I THREADED ROD.

BLACK IRON
PIPE

t------+.7t--__j_ HYDRAULIC JACK

OVERFLOW SLOT

PLYWOOD
TOP cAP
3
I
57.5

1 B.I!7I--",,=-I-~~-+- TIRE CHIPS

t--f.ft----t- PVC PIPE

,,--,- SPACER

[
!.J::~~~~~~~~Ffr~ BANDCOUPLING
rL""'tlLe CLAMP
BAND CLAMP

OWJNEL
SECTION
INLET

Fig. 10 Apparatus used to measure the permeability of tire chips (1 in=2S.4 mm).
 ~ -
Humphrey and Sandford Page IS

Testing methodology
The PVC pipe was filled with a known weight of tire chips in five 120-mm (4.8-in.)
thick layers. Each layer was compacted with 60% of standard Proctor energy. This
required 146 blows per layer with a modified Proctor hammer. The total compacted
height of each sample was 600 mm (24 in.). The compacted density and void ratio were
determined from the known weight, volume, and specific gravity of the tire chips.
The experimental set-up was then arranged as shown in Fig. 10. Water was passed
through the sample from bottom to top at rates between 50 and 100 IImin (I4 and 28
gpm). After the water levels stabilized, the water overflowing from the top of the
apparatus was collected for between 30 and 60 sec. During this time, the average
difference in water level between standpipes I and 2, 1 and 3, and I and 4 (see Fig. 7 for
numbering) was noted. Using the known lengths between these points, the hydraulic
gradient (i) was calculated. Thus, knowing the discharge quantity (Q), elapsed time (t),
hydraulic gradient (i), and cross sectional area of permeameter (A), the permeability (k)
could be found from Darcy's law
k = Q/(iAt) ...................................................... (2)

This procedure was used to test 3 samples of tire chips from 3 suppliers. InitiallY,
three observations were taken without compressing the sample. After that, the sample
was compressed in increments of 50 mm (2 in.) and three observations were taken at each
increment. The maximum compression that could be applied with the jack varied with
each type of tire chip and ranged from 140 to ISO mm (5.5 to 6 in.).

Results
The average measured permeabilities for tire chips from three suppliers are
summarized in Table 4. For each supplier tests were run on tire chips compressed to four

Table 4 Summary of permeability results.

Compression Density Permeability
Supplier (%) (Mglm 3 ) Void Ratio (cm/s)
Pine State Recycling 00 0.644 0.925 7.7
8.3 0.704 0.761 6.0
16.6 0.774 0.601 3.4
22.4 0.833 0.488 2.1
Palmer Shredding 0.0 0.601 1.114 15.4
8.3 0.657 0.935 12.7
16.6 0.723 0.758 8.2
24.9 0.803 0.583 4.8
F&B Enterprises 0.0 0.622 0.833 6.9
8.3 0.681 0.676 5.0
16.7 0.748 0.523 2.8
22.9 0.808 0.414 1.5
1 Mglm 3 = 62.4 pcf
 ~ -
Humphrey and Sandford Page 16

different void ratios. The corresponding values of dry density and void ratio are also
shown. Each value is the average of three trials. Examination of Table 4 shows that
Palmer tire chips have the highest void ratio and the highest permeability at any given
compression while the F&B tire chips have the lowest void ratio and permeability.
Nonetheless, the difference in permeability between the tire chips from the three suppliers
is relatively small. Overall, the permeabilities are similar to those that would be expected
for a clean gravel.
Compressing the samples decreased their permeability. This is shown by plotting
permeability versus void ratio in Fig. 11. The results from each test are plotted along with
the best fit straight line through the data points from each supplier. It is seen that the
decrease in permeability with void ratio is roughly linear. The tire chips from Palmer
Shredding experienced the largest decrease in permeability with void ratio.

CONCLUSIONS

Several conclusions can be drawn from this research:

1. Gradations of the tire chips from the three suppliers show that they were uniformly
graded and ranged in size from 13 to 76 mm (0.5 to 3 in.).
2. The specific gravity of the tire chips was slightly greater than that of water and ranged
from 1.14 to 1.27. Tire chips composed entirely of glass belted tires have a lower
specific gravity than those composed of a mixture of glass and steel belted tires.
3. The compacted dry densities of the tire chips were in a narrow range of 0.618 to 0.642
20.0 rrr~~~~~~~~rrr~~~~~~rr,

m PINE
....J,M. STATE
._'!.~.'!!'
F & B
oumD PALMER SHREDDING
___ 15.0 -
U1
'-.... m •
E
u
•
~

..q 10.0 m

-D
o
---'"
QJ

E
L
Q)
(L 5.0

o. 0 UJ.-'T',.u-=~'*'nC"-'-~-'-'-;;-\;'-~~,",*'~
1.1 0.9
'~~~
0.7 0.5 0.3

Void ratio

Fig. 11 Permeability versus void ratio for tire chips from three suppliers.
Humphrey and Sandford Page 17

Mg/m3 (38.6 to 40.1 pct) which clearly shows the potential to use tire chips as
lightweight fill.
4. Compression tests indicate that the tire chips are highly compressible during the initial
portion of the first loading cycle but that the compressibility is significantly less during
subsequent unloading/reloading cycles.
5. The friction angle of the tire chips ranges between 19° and 25° and the cohesion
ranges between 8 to II kPa (160 and 240 pst).
6. The amount of exposed steel belt seems to have a systematic effect on some of the
engineering properties of tire chips. Large amounts of exposed steel belts tends to
cause higher compressibility during the first loading cycle, higher Young's modulus
during unloading/reloading cycles, lower coefficient of earth pressure at rest Ko, and
lower shear strength.
7. These laboratory results suggest that there may be some advantage to using tire chips
with large amounts of exposed steel belt as retaining wall backfill since they have a
lower K o.
8. The permeability of tire chips ranged from 1.5 to 15 cm/s. These values are about the
same as would be obtained from clean gravel.

ACKNOWLEDGMENTS

The authors wish to thank the New England Transportation Consortium, Maine
Department of Transportation, and National Science Foundation (Award No. 9100646)
who provided the funding for this research. The testing was performed by graduate
students William Manion, Michelle Cribbs and Haresh Gharegrat, and undergraduate
Shelley Pressley. Their hard work is greatly appreciated.

REFERENCES

I. AASHTO (1986), Standard Specifications for Transportation Materials and

Methods of Sampling and Testing, Part II: Methods of Sampling and Testing, 14th
ed., American Association of State Highway and Transportation Officials,
Washington, DC.
2. Beatty, J.R. (1991), "Physical Properties of Rubber Compounds," Mechanics of
Pneumatic Tires, S.K. Clark, ed. National Highway Traffic Administration, U. S.
Department of Transportation, , Washington, DC.
3. Bressette, T. (1984), "Use Tire Material as an Alternate Permeable Aggregate,"
Office of Transportation Laboratory, California Department of Transportation,
Sacramento, California, 23 pp.
4. Das, D.M. (1990), Principles of Foundation Engineering, 2nd ed., PWS-Kent
Publishing Company, Boston.
Humphrey and Sandford Page 18

5. Dodson, T. (1990), "Camas Valley-Muns Creek, Highway 35 Landslide: Plans,

Construction Sequencing and Staging," Geotechnical Group, Oregon State
Highway Division.
6. Edil, T.B., Bosscher, J.P., and Eldin, N.N (1990), "Development of Engineering
Criteria for Shredded Whole Tires in Highway Applications," Interim report to the
Wisconsin Department of Transportation, University of Wisconsin-Madison.
7. EPA (1991), "Markets for Scrap Tires," Report EPN530-SW-90-047B, U.S.
Environmental Protection Agency, Washington, D.C.
8. Geisler, E., Cody, W.K., and Niemi, M.K. (1989), "Tires for Subgrade Support,"
Annual Conference on Forest Engineering, Coeur D'Alene, ID.
9. Gharegrat, H. (1993), "Numerical Analysis of Tire Chips as Fill Beneath Paved
Roads and as Retaining Wall Backfill", Master of Science Thesis, Department of
Civil Engineering, University of Maine, Orono, Maine.
10. Holtz, R.D., and Kovacs, W.D. (1981), An Introduction to Geotechnical
Engineering, Prentice-Hall, New Jersey.
11. Humphrey, D.N., and Eaton, R.A. (1993a), "Tire Chips as Insulation Beneath
Gravel Surfaced Roads," Proceedings of the Second International Symposium on
Frost in Geotechnical Engineering, A. Phukan, Ed., Anchorage, Alaska, pp. 137-
149.
12. Humphrey, D.N., and Eaton, R.A. (1993b), "Tire Chips as Subgrade Insulation -
Field Trial," Symposium on Recovery and Effective Reuse of Discarded Materials
and By-Products for Construction of Highway Facilities, Denver, Colorado,
October 19-22.
13. Humphrey, D.N., and Manion, W.P. (1992), "Properties of Tire Chips for
Lightweight Fill," Grouting, Soil Improvement and Geosynthetics, R.H. Borden, et
aI., eds. ASCE, Vol. 2, pp. 1344-1355.
14. Humphrey, D.N., 'Sandford, T.e., Cribbs, M.M., Gharegrat, H., and Manion, W.P.
(1992), "Tire Chips as Lightweight Backfill for Retaining Walls - Phase I," Dept.
of Civil Engineering, University of Maine, Orono, Maine.
15. Humphrey, D.N., Sandford, T.e., Cribbs, M.M., and Manion, W.P. (1993), "Shear
Strength and Compressibility of Tire Chips for Use as Retaining Wall Backfill,"
Paper No. 930925, Transportation Research Record, Transportation Research
Board, Washington, D.C.
16. IS TEA (1991), "Intermodal Surface Transportation Efficiency Act of 1991," P.L.
102-240.
17. Manion, W.P., and Humphrey, D.N. (1992), "Use of Tire Chips as Lightweight
and Conventional Embankment Fill, Phase I - Laboratory," Technical Paper 91-1,
Technical Services Division, Maine Department of Transportation, Augusta, Maine.