126 ■ Transportation Research Record 1740

Paper No. 00 - 0414

Asphalt Plug Joints

Refined Material Tests and Design Guidelines

Brian K. Bramel, Charles W. Dolan, Jay A. Puckett, and Khaled Ksaibati

Highway builders and rehabilitators throughout the United States use tified a mismatch between the material and component testing, par-
asphalt plug joints (APJs) in bridge expansion joints according to manu- ticularly in the effective modulus of elasticity and in the relaxation of
facturers’ recommendations. Joint performance varies widely, as indi- stresses. On the basis of these findings, a second series of material
cated in recent surveys. State departments of transportation are installing tests was conducted to evaluate the material relaxation and demon-
joints in unsuitable sites or are overlooking other sites where APJs may strate that a modification of the standard AASHTO thermal specimen,
work well. Rational design guidelines for APJs have been developed. The restrained specimen test (TSRST) (3) can be used to determine the pri-
research suggests suitable applications, material characterization, design mary material characteristics and qualify the material for application
guidelines, and validation procedures. Two critical material properties in APJs.
are required to qualify APJ material: relaxation and glass-transition tem- A one-dimensional model obtained with the modified test data
perature. Both properties may be obtained by using a slight modification demonstrates the time-dependent effects in an APJ. The model
of the AASHTO standard thermal specimen, restrained specimen asphalt aids in the understanding and in the presentation of relaxation as
test. This modified standard test was conducted, and the results were com- it affects joint performance. The test results are compared with the
pared with those of the near-full-scale test. Design guidelines are based on results obtained with the simple model, and finally, design criteria
field observations, material tests, near-full-scale testing, analytical evalu- are presented.
ations, and a survey of the experiences of state departments of transporta-
tion. Design recommendations are provided, and joint design changes are
proposed to help mitigate the present shortcomings. IMPORTANCE OF RELAXATION

Bridge joints are subjected to deformations as bridges expand and

The primary objective of the research described here is to understand contract. The loads or stresses in the joint are displacement-induced
the performance of asphalt plug joints (APJs) and propose tests, de- effects that occur relatively slowly as a function of ambient con-
signs, and construction methods so that APJs perform satisfactorily. ditions as well as the heat-transfer characteristics of the structure.
APJs are bridge expansion joints that use a modified asphalt binder- If the material relaxes nearly as rapidly as the temperature change
aggregate mix to span between the approach slab and the bridge deck. demands, little apparent stress is induced. Hence, material relaxation
This mix is placed in a blockout in the roadway and is bonded to the is very important.
substrate on three sides, as shown in Figure 1. The bridge motions cre- The fundamental APJ behavior for successful in-service perfor-
ate displacements where the gap plate slides on the bottom of the mance is that the joint stress decreases or relaxes at a rate nearly equal
blockout. These displacements concentrate deformations in a local- to the rate at which the temperature-induced motion imposes the stress.
ized region, creating high strain concentrations (1). Analysis based Therefore, only small stresses are built up in the joint and the material
solely on elastic assumptions indicates that high strain (and apparent flow accommodates the deformation. As the temperature drops, the
stress) concentrations should cause these joints to fail. Although fail- ability of the material to flow gradually decreases until an abrupt
ures occur, many joints perform well, even though general linear elas- ductility and viscosity transition occurs. This transition is referred to
tic theory predicts that they should not. That is, simple theories do not as the glass-transition temperature, Tg (see the section with design
describe field observations. In summary, these joints are performing guideline examples). At this temperature, the material viscosity and
well in certain applications and unsatisfactorily in others, and the ductility decrease to virtually zero. Moreover, the material stiffness
impetus of this research is to understand why. increases and the material becomes brittle; a small joint movement or
The research initially used classical engineering material charac- additional temperature drop creates a fracture that typically propa-
terizations of elastic or elastic-perfectly-plastic behavior, whereas it gates through the joint. Such cracks create serviceability problems and
was assumed that the viscoelastic or viscoplastic characteristics were constitute failure.
secondary (2). Temperature dependence was considered a primary A plasticity model assumes an elastic-perfectly-plastic model that
indicator for these characteristics and was determined to be an impor- allows infinite stain in the plasticity range (4). Like a relaxation model,
tant characteristic. The time-dependent material behavior was initially this model also requires that the previous load history of the material
considered a secondary contributor to joint performance. The near- be known. Figure 2 shows a normalized load diagram for a 1-day
full-scale joint validation illustrated the overwhelming importance of movement of a hypothetical bridge located in Cheyenne, Wyoming,
time-dependent characteristics. The near-full-scale test program iden- during the fall or spring season. The following assumptions are used
for this comparison (4):
B. K. Bramel, National Association of Homebuilders, 400 Prince Georges
Avenue, Upper Marlborough, MD 20744. C. W. Dolan, J. A. Puckett, and
K. Ksaibati, Department of Civil and Architectural Engineering, University of • Bridge displacement follows a sinusoidal form,
Wyoming, Laramie, WY 82071. • Material has no residual load at the start of the cycle,
Bramel et al. Paper No. 00 - 0414 127

FIGURE 1 Typical asphalt plug joint cross section.

• APJ is represented as a prismatic joint with no stress concen- a material failure when the stress exceeds the material capacity at the
trations, and specimen temperature. APJ is highly plastic at room temperature but
• A previously calculated elastic-perfectly-plastic model, an becomes brittle at low temperature. When the temperature of the APJ
elastic model, or the relaxation model is used to define the mate- material drops below Tg, the sample fails in tension. The temperature
rial behavior. at failure is the “TSRST temperature” and is considered a lower-
bound temperature for the application of APJs. This lower bound is
Figure 2 shows a sine wave that represents the elastic behavior of a nonconservative estimate because the sample will fail at a temper-
the joint. The plasticity model follows the sine wave until it “flows,” ature above Tg, with a very small displacement. In the TSRST, the
and then the stress reaches a plateau. The relaxation model begins to imposed mechanical strain (bridge movement) is considered to be
reduce the strain-induced load immediately, and therefore, the stresses zero. That is, the entire load effect is due to temperature. Hence, ten-
are lower than those in either the plasticity or the elastic model. These sile failures are ensured at or below this temperature with or without
models demonstrate the lower loads that occur when relaxation occurs any bridge motion. Joint failures due to bridge motion can occur at
rapidly as the bridge moves. The joint is unable to build up loads or temperatures above the TSRST temperature, which may be taken as
stresses and should perform well structurally down to a temperature the Tg for joint evaluation and design.
near Tg. At temperatures near or below Tg, the material becomes The TSRST procedure was modified to evaluate relaxation.
brittle and tensile failures are assured. Relaxation is a time-dependent reduction in load stress while a con-
stant strain is maintained. The material relaxation characteristics are
determined by inducing a small displacement, holding it constant,
MODIFIED MATERIAL TESTING and measuring the load decreases with time. The standard TSRST
was modified to induce an initial displacement while holding the
The AASHTO TSRST uses a prismatic sample of asphaltic material temperature constant. One shortcoming of the TSRST apparatus is
that is bonded to two platens (3). The specimen length is held con- that it cannot record this imposed initial displacement. This issue
stant while the sample temperature is lowered. By holding the sam- was addressed by using the previously determined modulus of elas-
ple length constant, the temperature strain is equal and opposite the ticity, E (4), and the relation ∆l = PLs /AE, where ∆l is the change in
strain due to the tensile restraining stress. The induced stress creates length, P is the applied load, Ls is the original sample length, and A

FIGURE 2 Elastic, elastic-plastic, and relaxation model relative stress comparison.

128 Paper No. 00 - 0414 Transportation Research Record 1740

is the cross-sectional area. Evaluations were all conducted at a con- joint. The first condition is desirable, but the second condition is
stant temperature of 2°C (34°F), a reasonable midrange temperature detrimental to the service life of the joint. This trade-off is discussed
for bridges in cool to cold climates. This research did not examine later in the description of a simple illustrative model that reproduces
the sensitivity of relaxation as the temperature approaches Tg. the behavior observed from the test data.
Samples from the near-full-scale joint were prepared for evaluation To understand the significance of relaxation, a simple and intu-
following completion of the cyclic loadings. High strains and possi- itive one-dimensional rheological model is developed. The simplest
ble degradation of the bond between the aggregate and binder are rheological model is a single Maxwell body that includes a spring
induced and concentrated on the motion side of the joint. The finite- and damper positioned in series. This model consists of an elastic
element analysis models conducted in previous research illustrated element (spring) that can store the energy and a viscous element
these high strain gradients (2). Prisms were cut from the “nonmoving” (damper) that dissipates the energy with time. The solution for the
side of the joint. Samples were sectioned transversely to the direction corresponding linear ordinary differential equation is
of loading in an attempt to obtain specimens that were strained sig-
−t
nificantly and, therefore, that had possible degradation of the binder- P(t ) = U0 Ke τ

aggregate adhesion. This preconditioning was a characteristic of the

research owing to the limited samples of material and is not generally where
recommended for a standard process. Samples of virgin material are
U0 = initial imposed displacement (constant),
preferred for conducting the modified TSRST evaluation.
K = spring constant from the Maxwell body,
The samples were placed in the TSRST equipment and were
τ = normalized variable defined as r/K, where r is the viscous
cooled to a temperature of 2°C (34°F). The specimens were loaded
constant, and
to approximately 356 N (80 lb), and their lengths and temperatures
t = time after application of the load.
were then held constant. The specimen load was recorded with respect
to time. The recording was stopped when the load appeared to stabi- The Maxwell body gives an initial instantaneous load of P(t = 0)
lize at a constant value. All samples stabilized within 15 min of the equal to U0 K, and the load decays with time as the viscous ele-
initial displacement (Figure 3). In summary, the TSRST and the mod- ment deforms and the system is unloaded. A single Maxwell body
ified TSRST are simple to conduct, yield useful information, and provides a simple approximation of relaxation, but it does not rep-
work well for qualifying APJ materials. Further work is required to resent the test materials in the present study because the load
determine parametric relaxation characteristics as the temperature relaxes too rapidly. Relaxation tests show that the load relaxed
approaches Tg. rapidly to an intermediate level, at which point the decay rate was
significantly lower. To better represent this type of relaxation, a
more refined model consisting of two Maxwell bodies in parallel
MODELING RELAXATION BEHAVIOR was used. The use of two bodies in parallel allows one of the
bodies to represent the relatively fast initial relaxation, whereas
A design trade-off exists between proportioning of the mixture and the second body represents the slower rate associated with the
blending of the binder. The joint must be viscous so that it accom- secondary relaxation. Note that more bodies could be used, if nec-
modates deformation from the temperature-induced motions. If it is essary. The model used is shown in Figure 4. This model is analo-
overly viscous, there is unwanted flow, aggregate separation, and gous to the Burger rheological model, which is often used to describe
rutting and it will flow out under the weight of the tires crossing the creep (5).

FIGURE 3 Plug joint relaxation (1 lb = 0.454 kg).

Bramel et al. Paper No. 00 - 0414 129

Intermediate Condition 2 is as follows: 95 percent of the long-term

decay is completed, and time t equal to t2 is much greater than t1:
− t2
τ1
e = 0.05
− t2
τ2
e ≈ 0. 0

P(t2 ) = U0 (0.05 K1 ) (3)

The equation parameters shown in Table 1 were derived by curve

fitting these equations to the data derived from the modified TSRST.
A comparison of the data from the analytical model with the test data
is illustrated in Figure 5. The two-element model aligns well. When
these data from the model are superimposed with the initial data, as
shown in Figure 5, the degree of correlation is acceptable. The APJ
materials from Watson Bowman Acme and Pavetech are well repre-
sented by this model. The curve responds slower than that for the
Koch/LDI APJ material in the first few minutes but correlates well
by 10 min.
FIGURE 4 Two-element relaxation
rheological model.
APPLICATION TO BRIDGE MOVEMENT

The characteristic equation for the four-element model shown in Bridge motion due to temperature is either continuous for bridges
Figure 4 is on elastomeric pads or discontinuous for bridges on metal bearings.
Both cases are examined. Many in-service bridges do not expand
P(t ) = U0 K1 䡠 e ( −t
τ1
+ K2 䡠 e
−t
τ2
) and contract smoothly. They store strain energy due to temperature
changes until the frictional resistance in the bearings is overcome.
where the forces in the two elements add to balance the reactions. They then move in discrete increments, and this is generally referred
The spring constants K1 and K2 and the normalized variables τ1 and to as the stick-slip phenomenon. The stick-slip phenomenon may
τ2 were developed from the relaxation data recorded from the modi- represent a joint condition in the bridge motion more critical than a
fied TSRST, the modulus of elasticity determined in previous research continuous deformation function.
(4), and the long-term decay time obtained from the near-full-scale Conceptually, a significant relaxation implies that the joint nearly
joint validation program (2). By using these values and the follow- completely relaxes faster with each stick-slip event that occurs with
ing initial conditions and assumptions, characteristic equations were thermal bridge deformation. The stick-slip behavior imparts a high
solved. The initial condition was as follows: level of stress on the joint. Therefore, the stick-slip condition is exam-
ined first in this discussion. (The behavior of the joint material is anal-
P0 = U0 ( K1 + K2 ) (1) ogous to that of “silly putty.” If it is pulled slowly, it will deform with
large strains, yet no critical stress will develop. If it is pulled quickly,
U0 is backcalculated from the initial conditions such that U0 = PL/AE, it will fracture.)
where P is the induced load, L is the effective length of the joint, A For comparative analysis, these discrete movements of the hypo-
is the cross-sectional area, and E is the joint modulus of elasticity. thetical bridge are assumed to occur in equal displacements of 1.27 mm
Intermediate Condition 1 is as follows: 99 percent decay of the (0.05 in.). This represents approximately 1⁄20 of the total bridge move-
short-term relaxation is completed, because time t1 is much less than ment. If the APJ “completely” relaxes between discrete movements,
t2 and the difference in load is approximately due only to the short- then the stress is zero when the next slip occurs; that is, the stress is
term relaxation; therefore, at time t equal to t1, independent of the next 1.27-mm (0.05-in.) movement. If full relax-
− t1
ation has not occurred, stresses accumulate. Figure 6 shows the load
τ2
e = 0.01 histories for six discrete movements, assuming complete relaxation.
To develop a design rationale, two time constants, tB75 and t75, were
∆ P = ( P0 − P1 ) ≅ U0 K2 (2) defined. tB75 is the amount of time between discrete bridge motions

TABLE 1 Relaxation Equations

NOTE: 1 lbf = 4.54 N; 1 psi (poundforce per square inch) = 6.89 kPa; 1 in. = 25.4 mm; 1 lbf/in. = 0.18 N/mm.
130 Paper No. 00 - 0414 Transportation Research Record 1740

FIGURE 5 Plug joint relaxation model versus experimental data at 2°C (34°F) (1 lb = 0.454 kg).

(a demand). It is approximated by linearizing the maximum daily the increment is left to the discretion of the designer; however, a
motion over a 12-h period and then dividing it by the assumed stick- methodology is indicated in the example that follows.
slip incremental movement per motion. t75 is the time required for the For continuous joint movement, as may occur in an elastomeric
material to relax to 25 percent of its initial load, that is, a 75 percent bearing, the solution is slightly more complex. If Ts is the seasonal
reduction in load. These constants have been calculated from data temperature range for a bridge, the joint movement (δ) is equal to
from the modified TSRST and are given in Table 2. These two time αTsl, where α is the coefficient of thermal expands and l is the length
constants allow direct comparison of the demand and the material’s of the bridge. If the temperature varies sinusoidally over a 24-h period,
capacity. the deformation may be written as δ(t) = Ts sin[(π/720)t]αl, where
The high t75 for the Pavetech material reflects the characteristic of t is in minutes. The rate of change in the bridge movement is [dδ(t)]/dt
the material in which it stabilizes at a stress level approximately 60 per- = (π/720) Ts cos[(π/720)t]αl, and the maximum rate of change occurs
cent of the initial stress level. Therefore, unlike the other two ma- when the cosine function is equal to 1. Thus, maximum [dδ(t)/dt] =
terials, Pavetech joints oscillate between 60 and 100 percent of the (πTsαl)/720. Noting that Tsαl is the maximum seasonal bridge joint
induced stress and do not relax to the 75 percent target. motion, the maximum translation rate of change is dδ(t)/dt = (maxi-
For this discussion approximations of the discrete increments due mum πδ)/720 = maximum δ/230. Thus, the maximum rate of move-
to stick-slip need to be made. Increments should be larger for steel ment is approximately 1 ⁄ 200 of the daily motion or 1⁄10 of the stick-slip
roller bearings and smaller for elastomeric bearings. For this exam- movement. Thus, if a stick-slip analysis is satisfactory, the relaxation
ple, a 1.2-mm (0.05-in.) increment has been used. The selection of is likely sufficient to relax the continuous load application.

FIGURE 6 Zoom plot of relaxation model (1 lbf = 4.45 N).

Bramel et al. Paper No. 00 - 0414 131

TABLE 2 Determined Material Time Constant, t 75

JOINT GEOMETRY CONSIDERATION There is a difference in joint motion of 8 mm (0.35 in.) between
the joints with thicknesses of 50 mm (2 in.) and 100 mm (4 in.). The
A finite-element analysis indicated that the strain resulting from joint joint capacity decreases with additional thickness; therefore, the
movement was concentrated within a triangular cross section radiat- designer should consider typically using a joint 50 mm (2 in.) thick.
ing out from the edge of the gap plate (2, 4). Slip planes at approxi-
mately 60 degrees from the blockout bottom define the high-strain
region (Figure 7). The typical APJ blockout is constructed symmet- OTHER SERVICE ISSUES
rically simply because the gap plate motion is undefined. By con-
straining the gap plate to be fixed to one side of the blockout, the One major disadvantage of relaxation is that the joint is susceptible to
motion is forced to occur on the other side. On the basis of these geo- rutting. The traffic induces a load that forces the material to flow out
metric criteria, the blockout can be constructed as shown in Figure 7 of the wheelpath. These ruts can be reduced by limiting joint applica-
with a corresponding reduction in the required joint material. tion to locations where the traffic is moving relatively fast, such as on
Some debonding will occur at the gap plate. The minimum dimen- highways. These joints should not be installed in locations where the
sion for the plug joint is a wedge extending upward at 60 degrees likelihood of slowly moving or stationary traffic exists, such as at
from the edge of the gap plate and the debond zone. A 25-mm (1-in.) intersections. The reason for this is twofold. The slowly moving or
debond zone is shown for illustrative purposes (Figure 7). The fixed stationary traffic places the loading on the joint for a longer time inter-
side of the gap plate must be secured to force the joint motion into val, thereby increasing the flow of the joint material under the wheel.
the preferred side. This may be accomplished with APJ binder or Additionally, the possibility of steering and wheel scrub accelerates
with mechanical fasteners. It is recommended that the gap plate be the rutting and track-out conditions.
fixed on the upstream traffic side so that traffic will assist with the The skew angle may also affect tracking out of the joint because of
healing of any debonding on the downstream edge. Traffic tends traffic loading. APJs are isotropic, and theoretically, their behavior is
to reconsolidate the joint. independent of whether the traffic is perpendicular to the joint cross
It is the material viscosity that is paramount for successful appli- section or is situated at an oblique angle. As the skew angle increases,
cation. If the structures are moving at or about the same rate as the the joint could have a longer effective cross section. The longer length
relaxation, the joint loads are low and the joint structurally has no increases the rutting due to traffic. A reasonable limit on the skew is
theoretical limit of motion. In practice, a functional limit exists to 30 degrees because this would effectively limit the load along the joint
provide a smooth transition from the pavement to the bridge deck (6). to one-half that across the joint. This recommendation is consistent
A smooth transition onto the bridge requires that a volumetric joint with the limits imposed by Oregon and Connecticut (6).
deformation be imposed on the motion. This geometric relationship
requires that the material inside the 60-degree failure envelope not
depress or rise by more than the specified limit under maximum DESIGN GUIDELINE EXAMPLES
joint displacement (horizontal). By limiting the depression or bump
to 19 mm ( 3⁄4 in.), a volumetric relationship based on the failure zone The sections below illustrate design sheets used to determine the
defined by the finite-element analysis is established. Table 3 gives applicability and size of an APJ. The intent is to determine if a site is
these motion limits on the basis of the results obtained for a 19-mm suitable for APJs. On the basis of the anticipated temperature range
( 3⁄4-in.) limit and various plug joint thicknesses. The imposition of a and structural fixity, the maximum daily motion of a bridge is esti-
smaller depression results in smaller maximum motions. mated. From this maximum daily bridge motion, compute the bridge

FIGURE 7 Optimized joint geometry.

132 Paper No. 00 - 0414 Transportation Research Record 1740

TABLE 3 Volumetric Relationship-Based TABLE 4 Material Characteristics

Motion Limits

NOTE: 1 in. = 25.4 mm.

time requirement (tB75) as illustrated. The examples use the Asphalt a low temperature of −28°C (−21°F) with a 94.8 percent reliability;
Institute Superpave program to determine the pavement temperature seasonal average temperature, −7°C (19°F); maximum seasonal move-
range (7 ). The required material characteristics can be obtained from ment, 0.55 mm/°C × (−7 to −28°C), which is equal to 11.55 mm
this paper, Table 4, testing, or the manufacturer. The material time (0.5 in.) of movement; and maximum daily motion two times the
capacity, t75, is verified against the bridge time requirement to ensure maximum seasonal motion, which is equal to 2 × 11.5 mm, which
that the relaxation is adequate for the location. The lowest anticipated equals 23.0 mm (0.9 in.). To determine the bridge time requirement,
temperature is checked against Tg. If all checks are satisfied, the assume that the maximum daily motion occurs over a 12-h period
maximum joint thickness is selected, typically 50 mm (2 in.). and that movements are discrete 1.27-mm (0.05-in.) movements.
The bridge time requirement tB75 is

Example 1, Cheyenne, Wyoming

t B75 =   (60 min 1h )
12 h
 [23 mm (1.27 mm increment )] 
This procedure is illustrated by using a hypothetical bridge located in
Cheyenne, Wyoming. Given a bridge in Cheyenne with a temperature = 40 min increment
motion of 0.55 mm/°C (0.04 in./°F), the solution is as follows: pave-
ment temperature range from Superpave model data for bridges (7 ); This information is summarized in Figure 8 and concludes that the
high temperature of 58°C (142°F) with a 99.9 percent reliability and Koch /LDI material is suitable for this location.

FIGURE 8 Example 1 design guideline worksheet.

Bramel et al. Paper No. 00 - 0414 133

Example 2, Denver, Colorado CONCLUSIONS

Example 2 uses a bridge that is similar to that used in Example 1 and This research focused on the structural and serviceability issues that
that is located in Denver, Colorado. Given a bridge in Denver, Col- directly affect plug joints. The research intentionally did not address
orado, with a temperature motion of 0.40 mm/°C (0.03 in./°F), the the material chemistry and long-term issues such as fatigue and age
solution is as follows: pavement temperature range from Superpave embrittlement due to the loss of volatile material used to plasticize
the joint materials. The information from this research can be used to
model data for bridges (7 ); high temperature of 58°C (142°F) with
focus on the development of better binders that can produce better
a 99.9 percent reliability and a low temperature of −22°C (−8°F)
joints. Tg and relaxation should be qualified. The modified TSRST
with a 78 percent reliability; a seasonal average temperature of 0°C
illustrates a possible approach through a temperature range of 0°C
(32°F); and a maximum seasonal movement of 0.40 mm/°C × (0 to
(32°F) to Tg; however, software for the TSRST equipment must be
−22°C), which is equal to 8.8 mm (0.35 in.) of movement; and max- modified for commercial and production testing. Tests for long-term
imum daily motion two times maximum seasonal motion, which is aging might be pursued because the binder is asphalt based and it will
equal to 2 × 8.8 mm, which equals 18.0 mm (0.69 in.). To determine lose its ductility over time. The loss of ductility accelerates fatigue
the bridge time requirement, assume that the maximum daily motion fractures, which will likely produce tensile failures over time.
occurs over a 12-h period and that movements are discrete 1.27-mm The combinations of the long-term fatigue susceptibility, the
(0.05-in.) movements. Then the bridge time requirement is loss of ductility, and the serviceability issues of rutting imply that
these joints should not be viewed as service-free joints. The main-
t B75 =   (60 min 1h )
12 h tenance required is relatively straightforward when a joint exhibits
 [23 mm (1.27 mm increment )]  either cracking or rutting, but it will be required. Frequent main-
tenance repairs by the addition of binder is a reasonable approach
= 32 min increment but may damage the base material because of the required heat,
thereby driving out the volatile materials that provide the necessary
This information is summarized in Figure 9. The Koch / LDI and viscoelasticity.
Watson-Bowman-Acme joints work in this location. The Pavetech This research provides guidelines for the qualification of APJ
joint will likely accumulate stress because of the high t75 value. materials and the selection of sites suitable for APJ installation. APJ
Whether it works in the field will depend on how quickly the stresses installations will involve some maintenance, and locations where
accumulate. additional joint fractures may occur are identified.

FIGURE 9 Example 2 design guideline worksheet.

134 Paper No. 00 - 0414 Transportation Research Record 1740

1. APJ material can be qualified by using the TSRST and a mod- FUTURE WORK
ified TSRST procedure to obtain Tg and t75. The equipment for these
tests was developed under the Strategic Highway Research Program Future work with APJs should focus on the following topics:
and is readily available to departments of transportation. Software
modifications will be required to perform t75 tests. • Modification of TSRST equipment software to define the t75 data,
2. APJs should not be installed in areas where the lowest antic- • Laboratory verification of the optimized joint geometry, and
ipated temperature is below Tg. • Investigation of the long-term fatigue and aging characteristics
3. Time-dependent material properties are extremely impor- of the joint.
tant to joint performance. The relaxation of the APJ material
should be sufficient to relieve the stress due to the applied thermal
displacement. ACKNOWLEDGMENTS
4. The total serviceable deformation available in an APJ is based
on volumetric considerations of the joint material. Recommenda- The authors thank the Wyoming Department of Transportation and
tions for thickness and movement limits are provided in Table 3. the Colorado Department of Transportation for financial support
The small differences in Table 3 suggest that a 50-mm (2-in.) thick- and Koch/LDI, Watson Bowman Acme, and Pavetech for providing
ness is optimal. APJs should not be placed with a thickness of less materials.
than 50 mm (2 in.), and a thicker joint has little benefit.
5. APJs should be installed only where traffic moves at highway
speeds. REFERENCES
6. APJs exposed to rapid or large thermal shocks will experience
1. Bramel, B. K., J. Kostage, C. W. Dolan, and J. A. Puckett. Experimental
more fractures than APJs in structures with gradual joint movements. Evaluation of Asphaltic Plug Joints. Fourth World Congress on Joint
7. Securing the gap plate on one side of the joint and forcing the Sealants and Bearing Systems for Concrete Structures. Report ACI SP-164.
displacement to only one side may reduce the quantity of material American Concrete Institute, Farmington Hills, Mich., 1997.
required in the joint. 2. Bramel, B. K. Asphaltic Plug Joints: Characterization and Specification.
Ph.D. dissertation. University of Wyoming, Laramie, 1999.
3. Standard Tests Methods for Thermal Stress Restrained Specimen Tensile
Strength, 1st ed. AASHTO, Washington D.C., 1993.
RECOMMENDATIONS 4. Bramel, B. K., C. W. Dolan, K. Ksaibati, and J. A. Puckett. Asphalt Plug
Joints: Material Characterization and Specifications. Proc., Fifth Inter-
APJs are a viable joint if they are properly qualified and appropriately national Conference on Short and Medium Span Bridges, Calgary, Alberta,
used. All bridges are not candidates for the application of APJs, how- Canada, 1998.
5. Bodig, J., and B. A. Jayne. Mechanics of Wood and Wood Products. Van
ever. The lowest anticipated temperature should be above Tg. Ther- Nostrand Reinhold, New York, 1982.
mal shocking should be avoided. Bridge joint motions should be less 6. Bramel, B. K., C. W. Dolan, K. Ksaibati, and J. A. Puckett. Asphalt Plug
than those shown in Table 2, and service-free life expectancy is likely Joint Usage and Perception in the United States. TRB, National Research
to be less than 5 years because of fatigue considerations and gradual Council, Washington, D.C., 1998.
7. Superpave Level 1 Mix Design. Superpave Series No. 2 (SP-2). Asphalt
stiffening of the APJ due to the loss of volatile plasticizers. Traffic Institute, Lexington, Ky., 1995.
helps to “heal” joints. Geometric modifications made to improve joint
performance are illustrated. Publication of this paper sponsored by Committee on General Structures.