Construction

and Building

Construction and Building Materials 20 (2006) 982–989

MATERIALS
www.elsevier.com/locate/conbuildmat

Laboratory comparison study for the use of stone matrix

asphalt in hot weather climates
Ibrahim M. Asi *

Department of Civil Engineering, Hashemite University, Zarqa 13115, Jordan

Received 17 November 2003; received in revised form 1 May 2005; accepted 30 June 2005
Available online 15 August 2005

Abstract

Stone matrix asphalt (SMA) is a hot mixture asphalt consisting of a coarse aggregate skeleton and a high binder content mortar.
It was developed in Germany during the mid-1960s and it has been used in Europe for more than 20 years to provide better rutting
resistance and to resist studded tyre wear. The main objective of this research study was to compare the performance of the normally
used dense graded asphalt mixtures, named in this research as control mixtures, and SMA mixtures. Samples from both mixtures
were fabricated at their optimum asphalt contents that were 5.3% for control mixtures and 6.9% for SMA mixtures. Comparison
performance tests that included Marshall stability, loss of Marshall stability, split tensile strength, loss of split tensile strength, resil-
ient modulus, fatigue, and rutting testing were performed on both mixtures. Test results showed that although the control mixtures
have higher compressive and tensile strengths, SMA mixtures have higher durability and resilience properties. In addition, although
the research could not prove the superiority of SMA in rutting resistance because of the limited sample sizes, field performance of
SMA mixtures proves its superiority. Therefore, especially in hot weather climates, these properties, (durability, resilience and rut-
ting resistance) give SMA mixtures advantages over dense graded mixtures.
 2005 Elsevier Ltd. All rights reserved.

Keywords: Stone matrix asphalt; Fatigue; Rutting; Indirect tensile strength; Stripping

1. Introduction Ministry of Communications in Saudi Arabia has intro-

Stone matrix asphalt (SMA) is a hot mixture asphalt
consisting of a coarse aggregate skeleton and a high bin-
der content mortar. SMA was developed in Germany
during the mid-1960s and it has been used in Europe
for more than 20 years to provide better rutting resis-
tance and to resist studied tyre wear [1]. Because of its
success in Europe, some States, through the cooperation
of the Federal Highway Administration, constructed
SMA pavements in the United States in 1991 [2]. Since
that time the use of SMA in the US has increased signif-
icantly. Japan has also started to use SMA paving mix-
tures as well with good success [3]. Recently, the
*
Fax: +962 6 551 8867.
E-mail address: asi@hu.edu.jo.
duced SMA in its road specifications. In addition, one
test road was constructed in the Eastern Province of
Saudi Arabia.
According to the SMA Technical Working Group [4],
SMA is a gap graded aggregate–asphalt hot mixture
that maximises the asphalt cement content and coarse
aggregate fraction. This provides a stable stone-on-stone
skeleton that is held together by a rich mixture of as-
phalt cement, filler, and stabilising additive.
The original purpose of SMA was to provide a mix-
ture that offered maximum resistance to studded tyre
wear. SMA has also shown high resistance to plastic
deformation under heavy traffic loads with high tyre
pressures, as well as good low temperature properties
[2,5]. A study conducted in Ontario, Canada, by the
Ministry of Transportation on SMA pavement slabs

0950-0618/$ - see front matter  2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.conbuildmat.2005.06.011
 I.M. Asi / Construction and Building Materials 20 (2006) 982–989 983

trafficked with a wheel-tracking machine gave less rut design procedure was used in this research to optimise
depths in comparison to that occurring in a dense fric- the asphalt content for both types of mixtures. With
tion coarse [3]. In the United States, the Georgia SMA mixtures optimum asphalt contents are normally
Department of Transportation has also performed a sig- selected as that which produces 3.0–4.0% air voids [6].
nificant amount of wheel tracking tests on SMA mix- Table 1 lists the suggested requirements for SMA mix-
tures that gave positive results. In addition, SMA has tures developed by the FHWA sponsored SMA Techni-
a rough surface texture, which provides good friction cal Working Group (TWG) in publication IS 118 [4].
properties after the surface film of asphalt cement is These requirements were followed in this study for the
removed by traffic. Other essential factors that enhance design of SMA mixtures. The experimental research
the feasibility of SMA in contrast to conventional hot programme followed in this study consisted of three
mixture asphalt are increased durability, improved aging tasks (Fig. 1). In the first task, a preliminary investiga-
properties and reduced traffic noise [4]. tion was conducted to characterise the used asphalt
SMA is a hot mixture with a relatively large propor- and aggregate. In the second task, optimisation of the
tion of stones and a substantial quantity of asphalt and two mixtures was performed and enough samples from
filler. The main concept of having a gap gradation of both mixtures were fabricated at their obtained
100% crushed aggregates is to increase a pavementÕs sta- optimum asphalt contents. The third task was devoted
bility through interlock and stone-to-stone contact. This
mixture is designed to have 3–4% air voids, and a rela- Table 1
tively high asphalt content due to the high amount of Suggested requirements for SMA mixtures developed by TWG [4]
voids in the mineral aggregate. The mixture contains Property Criteria
high filler content (10% passing the 0.075-mm sieve),
Coarse aggregate
and typically contains a polymer in the asphalt cement, L.A. Abrasion (AASHTO T 96) 30 Max
or fibre (cellulose or mineral) in the mixture to prevent Flat and elongated particles (ASTM D 4791) 3:1, 20% Max
drainage of the asphalt cement. This mixture has a sur- 5:1, 5% Max
face appearance similar to that of an open graded fric- Sodium sulfate soundness (AASHTO T 104) 15% Max
Percent fractured faces
tion course, however it has low in-place air voids
One or more 100% Min
similar to that of a dense graded HMA. Two or more 90% Min
Brown et al. [2] carried out a study to evaluate the per- Absorption (AASHTO T 85) 2% Max
formance of SMA in the United States by evaluating 86 Coarse and fine durability index 40 Min
SMA projects. Data was collected on material and mix- (AASHTO T 210)
ture properties, and performance was evaluated on the Fine aggregate 100% Crushed
basis of rutting, cracking, ravelling, and fat spots. The Sodium sulfate soundness (AASHTO T 104) 15% Max
major conclusions from their study were: (1) 85% of the Liquid limit (AASHTO T 89) 25% Max
surveyed projects had an aggregate Los Angeles abrasion Total aggregate – gradation
value more than 30%; (2) SMA mixtures were produced 19.0 mm 100
12.5 mm 85–95
90% of the time with 25–35% of the material passing
9.5 mm 75 Max
the 4.75-mm sieve and 80% of the time with 7–11% of 4.75 mm 20–28
the material passing the 0.075-mm sieve; (3) 30% of the 2.36 mm 16–24
surveyed projects had average air voids during construc- 600 lm 12–16
tion less than 3%; (4) 60% of the projects exceeded 6.0% 300 lm 12–15
75 lm 8–10
asphalt content; (5) over 90% of the SMA projects had
20 lm 3 Max
rutting measurements less than 4 mm; (6) SMA mixtures
appeared to be more resistant to cracking than dense Asphalt cement AASHTO M 226
mixtures; (7) there was no evidence of raveling on the Mineral filler
SMA projects; (8) fat spots appeared to be the biggest Plasticity index 4 Max
Percent passing 20 mm 20%
performance problem in SMA mixtures.
Stabiliser
Cellulose 0.3%
Mineral fibre 0.4%
2. Experimental program Polymer –

The main objective of this research study was to com- Mix design
Stone on stone contact –
pare the performance of densely graded hot mixture as- Voids in total mix 3–4
phalt mixtures and stone matrix asphalt mixtures. In VMA 17
this research, the densely graded mixtures are referred Asphalt content 6.0% Min
to as control mixtures, and the stone matrix mixtures Compactive effort 50 Blows
are referred to as SMA mixtures. The Marshall mix Draindown 0.3% Max
984 I.M. Asi / Construction and Building Materials 20 (2006) 982–989
Fig. 1. Research experimental program.
to the comparison of the performance of both mixtures.
The performance tests were Marshall stability, loss of
Marshall stability, split tensile strength, loss of split ten-
sile strength, resilient modulus, fatigue and rutting.
3. Materials used
3.1. Aggregate
The used aggregate in this study was obtained from
Abu Hadriyah located in the Eastern Province of Saudi
Arabia. Fig. 2 shows the recommended gradation limits
for the dense graded wearing coarse gradation and the
selected gradation for the control mixture. The selected
gradation was in the middle of both limits. Fig. 2 also
shows the recommended gradation limits by the TWG
for SMA mixtures and the selected gradation in this re-
search which was in the middle of the limits.
In SMA recommended mixtures, 8–10% of the total
amount of aggregate in the mixture passes the
0.075-mm sieve. This large amount of filler plays an
important role in the properties of SMA mixture partic-
ularly in terms of air voids, voids in the mineral aggre-
gate, and optimum asphalt content [7]. Since the
amount of material passing the 0.075-mm sieve is rela-
tively large, the SMA mixtures perform very differently
from other HMA mixtures.
3.2. Asphalt cement
The asphalt cement used in this investigation was ob-
tained from Riyadh oil refinery. The asphalt was 60/70
Penetration Grade with a softening point of 51 C.
3.3. Fibre stabilisation
One of the major problems usually encountered in
SMA mixtures is the draindown of the binder during
mixing, transporting and compaction. To overcome this
problem, fibres are usually added to SMA mixtures.
Loose organic fibres, such as cellulose, and mineral
 I.M. Asi / Construction and Building Materials 20 (2006) 982–989
#500
#200
100
Middle
80
Limit
PERCENT PASSING, %
60
40
20
0.01
Fig. 2. Grain size distribution of both control and SMA mixtures.
fibres are used as stabilising agents in SMA mixtures, at
the rate of 0.3% and 0.4% by weight of mixture, respec-
tively [8,9].
In this study, 0.3% mineral fibres, by weight of mix-
ture, were uniformly combined with the dry aggregate
before the asphalt cement was added.
3.4. Polymer stabilisation
In addition to fibre stabilisation, polymer stabilisa-
tion is also used in SMA mixtures. In SMA mixtures,
polymer stabilisation is mainly used to minimise the
asphalt cement draindown. In addition, it is used to
increase the stiffness of the AC at high in service temper-
atures and/or to improve the low temperature properties
of the binder material. Polymers are typically added to
the mixture at a rate of 3.0–8.0% by weight of the as-
phalt cement [10].
In this research, no polymers were added to the SMA
mixture, to exclude the effect of polymers when compar-
ing the performance of the SMA and control mixtures.
3.5. Optimisation of the mixtures
Marshall mix design (ASTM D 1559) procedure is
normally used to optimise the HMA mixtures in Saudi
Arabia. The Marshall procedure was used in this re-
search to optimise the control mixtures. Six percentages
of asphalt cement (3.5%, 4.0%, 4.5%, 5.0%, 5.5%, and
6.0%) were used to optimise the asphalt content. Three
samples at each percentage were prepared using a stan-
dard Marshall Compactor. Seventy-five blows on each
side of the specimens were applied, as per Ministry of
985
SIEVE SIZE
1/2
3/4
3/8
#50
#30
#8
#4
Control
SMA
0.10 1.00 10.00 100.00
SIEVE OPENING, mm
Communication (MOC) design requirement. Standard
Marshall specimens (63.5-mm height · 101.6-mm diam-
eter) were used in this optimisation process and the rest
of this study. According to the Ministry of Communica-
tion (MOC) design procedure, the optimum asphalt con-
tent was selected to have maximum stability, maximum
unit weight, and median of allowable limits for percent
air voids (AV limits for wearing coarse is 4.0–7.0%).
The average asphalt cement (AC) content at these three
values is selected and checked to satisfy the AV, VMA,
stability and flow specification limits. The obtained opti-
mum asphalt content for the control mixtures was 5.3%.
In SMA mix design, usually the Marshall method of
mix design is used to verify satisfactory voids in SMA
mixtures. Laboratory specimens were prepared using
fifty blows of the Marshall hammer per side. Seventy-
five compaction blows were not used since they would
tend to break down the aggregate more and would not
result in a significant increase in density over that pro-
vided by 50 blows. SMA mixtures have been more easily
compacted on the roadway to the desired density than
the effort required for conventional HMA mixtures
[11]. The optimum AC content for SMA mixtures is usu-
ally selected to produce 3.0–4.0% air voids. Marshall
stability and flow values are generally measured for
information but not used for acceptance [4]. In this
research, compaction of all the SMA samples was
performed using fifty blows of the Marshall hammer
per side. Fig. 3 shows the effect of varying the asphalt
content on the values of Marshall stability, flow, specific
gravity, voids in mineral aggregate (VMA), and air
voids (AV) for the SMA mixtures. Four asphalt percent-
ages (5.0%, 6.0%, 7.0% and 8.0%) were used in the
986 I.M. Asi / Construction and Building Materials 20 (2006) 982–989

2.40 7.0

2.38

Marshall Stability, kN
6.8

2.36

Gmb
6.6
2.34

6.4
2.32

2.30 6.2
4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5
Asphalt Content, % Asphalt Content, %

5.0

4.5

Determination of Optimum
Air Voids, %

4.0
Asphalt Content
3.5

3.0 Asphalt Content at 3.5 %

Air Voids = 6.90 %
2.5

2.0
4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5
Asphalt Content, %

8.0 20.0
Voids in Mineral Aggregate,%

7.5
7.0
Marshall Flow, mm

19.5
6.5
6.0
5.5 19.0
5.0
4.5
18.5
4.0
3.5
3.0 18.0
4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5

Asphalt Content, % Asphalt Content, %

Fig. 3. Mix design of SMA mixtures.

design of the SMA mixtures. It was found that at 3.5% compacted using the 50 Marshall compaction blows
air voids, the required asphalt cement was 6.9%. each side at an asphalt content of 6.9%. The prepared
samples were subjected to the following tests: Mar-
shall stability, loss of Marshall stability, split tensile
4. Comparison of the performance of control and SMA strength, loss of split tensile strength, resilient modulus,
mixtures and fatigue and rutting evaluation. Following is a dis-
cussion of these tests and their results.
In order to compare both the control and SMA
mixtures, 30 samples of the control mixture using the 4.1. Marshall stability and loss of Marshall stability
Marshall mixing procedure and the selected gradation
at an asphalt content of 5.3% were fabricated. An In order to find the difference in the Marshall
additional 30 samples of the SMA mixtures were stability and loss of Marshall stability between the
 I.M. Asi / Construction and Building Materials 20 (2006) 982–989 987

Table 2 1000
Marshall test results
SMA@ 25 C

Initial Tensile Strain, microns

Mix type Marshall stability at Marshall stability at % Loss Control@ 50 C
SMA@ 50 C
60 C, 35 min water 60 C, 24 h water Control@ 25 C
immersion (kN) immersion (kN)
Control 10.8 8.4 29
SMA 7.1 6.3 20 100

control and SMA mixtures, six samples from each

mixture were immersed in the water bath at 60 C.
The Marshall stability values for three samples from
10
each mixture were obtained after 35 min of water
1000 10000 100000
immersion. In addition, the Marshall stability values Number of Load Repetitions to Failure
after 24 h water immersion were also obtained. Table
2 shows the test results and the percent loss in Mar- Fig. 4. Fatigue curves for control and SMA mixtures at different test
temperatures.
shall stability of both mixtures. It can be observed
that although the stability of control samples is higher
than SMA samples, but the loss in Marshall stability
modulus of asphalt mixtures compared to the control
is higher in the control mixtures. The reasons behind
mixtures. This might be attributed to the higher asphalt
these properties are that control mixtures have a dense
content and mineral filler in the SMA mixtures than the
graded aggregate gradation giving them higher stabil-
control mixtures giving the SMA mixtures better resil-
ity, but the SMA mixtures have a higher asphalt con-
ience properties.
tent resulting in a thicker asphalt film thickness on the
aggregate leading to a higher protection against water
4.4. Fatigue performance
damage.
Diametral fatigue test results for control and SMA
4.2. Water sensitivity test (Lottman test AASHTO
mixtures at testing temperatures of 25 and 50 C are
T-283)
shown in Fig. 4. These results show a normal linear rela-
tionship between the logarithm of applied initial tensile
This test was carried out in order to find the water
strain and the logarithm of fatigue life (number of ap-
susceptibility (stripping resistance) of control and
plied load repetitions until failure). The fatigue data
SMA mixtures utilising indirect tensile strength (ITS).
were analysed by running a regression analysis to deter-
Table 3 shows the percentage loss in the indirect tensile
mine the fatigue relationship parameters in the following
strength after Lottman conditioning of the samples. The
form:
obtained results indicate that the average percentage loss
S
in strength due to water damage is effectively reduced in et ¼ I  ðN f Þ ; ð1Þ
SMA mixtures. This is attributed to the higher asphalt
where et is the initial tensile strain, Nf is the number of
film thickness around the aggregate.
load repetitions to failure, I is the anti-log of the inter-
cept of the logarithmic relationship, and S is the slope
4.3. Resilient modulus test, MR (ASTM D 4123)
of the logarithmic relationship. Regression parameters
for Eq. (1) are shown in Table 4. The results shown in
Resilient modulus (MR) is the most important variable
Fig. 4 and Table 4 show that the SMA mixtures have
to mechanistic design approaches for pavement struc-
lower fatigue life than the control mixtures. This is
tures. It is the measure of pavement response in terms
attributed to the lack of the mechanical locking of the
of dynamic stresses and corresponding strains. At 25 C
aggregate.
test temperature, SMA mixtures showed higher MR val-
ues (1894.7 MPa) than the control mixtures (936.3 MPa).
Therefore, SMA has improved the diametral resilient
Table 4
Regression factors for fatigue test
Table 3 Mix type Regression factors at Regression factors at
I.T.S test results 25 C 50 C
Mix type Initial ITS (kN) Final ITS (kN) % Loss I S I S
Control 10.8 8.4 22 Control 175842 0.7295 2788459 1.212
SMA 7.1 6.3 11 SMA 42524 0.6282 4935231 1.295
988 I.M. Asi / Construction and Building Materials 20 (2006) 982–989

10 straight line. Parameters I and S in the above equation

were obtained by using permanent deformation experi-
mental data in a regression analysis.
Permanent Deformation, microns

Results indicate that the control mixtures have better

permanent deformation resistance as compared to the
1 SMA mixtures. In spite of the fact that the research
could not prove the superiority of SMA in rutting resis-
tance, but field performance of SMA mixtures proved
this superiority [1,2,7,11,12]. The lack of the resem-
blance of the laboratory samples to field performance
Control @ 50 ms
in rutting resistance can be attributed to the tested sam-
0.1
SMA @ 50 ms ple size (100 · 65 mm) which might not have facilitated
Control @ 75 ms the stone to stone contact which is the main advantage
SMA @ 75 ms
of the SMA mixtures that gives it the superiority in rut-
Control @ 100 ms
SMA @ 100 ms ting resistance.

0.01
10 100 1000 10000 100000
Number of Load Repetitions 5. Conclusion
Fig. 5. Rutting curves for control and SMA mixtures at 25 C.
This research focused on a laboratory evaluation of
the performance of stone matrix asphalt by comparing
its performance with dense graded mixtures. The follow-
10.00
ing points can be concluded from the research:
Control @ 50 ms
Permanent Deformation, microns

SMA @ 50 ms 1. Stone matrix asphalt (SMA) is hot mixture asphalt

Control @ 75 ms
SMA @ 75 ms consisting of a coarse aggregate skeleton and a high
Control @ 100 ms binder content mortar.
1.00 SMA @ 100 m
2. Although the Marshall stability of control mixtures is
higher than SMA mixtures, but the loss in Marshall
stability is higher in the control mixtures.
0.10 3. SMA mixtures have better durability (resistance to
water damage) than control mixtures. This is attrib-
uted to the higher asphalt film thickness around the
aggregate.
0.01 4. SMA has improved the diametral resilience proper-
10 100 1000 10000 ties of the asphalt mixtures as compared to the con-
Number of Load Repetitions
trol mixtures.
Fig. 6. Rutting curves for control and SMA mixtures at 50 C. 5. SMA mixtures have a lower fatigue life than control
mixtures.
6. Although the research could not prove the superiority
4.5. Permanent deformation of SMA in rutting resistance due to the limited sam-
ple sizes, the field performance of SMA mixtures
The permanent deformation was simultaneously re- proved this superiority.
corded while running the fatigue test at both testing tem- 7. For hot weather climates, durability, resilience and
peratures of 25 and 50 C. Results are presented in Figs. rutting resistance give SMA mixtures advantage over
5 and 6. These results indicate that a straight-line rela- dense graded mixtures.
tionship exists between the logarithm of number of load
repetitions and the logarithm of permanent strain. The
permanent deformation properties were determined
using the following equation form: References
S
ep ¼ I  ðN Þ ; ð2Þ [1] Scherocman JA. SMA reduces rutting. Better Roads 1991;61(11).
[2] Brown ER, Mallick RB, Haddock JE, Bukowski J. Performance
where ep is the accumulated permanent strain, N is the of stone matrix asphalt (SMA) mixtures in the United States.
number of load repetitions, I is the anti-log of the inter- National Center for Asphalt Technology, NCAT Report No.
cept of the logarithmic relationship, and S is the slope of 97-1, Auburn University, Alabama; 1997.
 I.M. Asi / Construction and Building Materials 20 (2006) 982–989
[3] Brown ER, Haddock JE, Mallick RB, Lynn TA. Development of
a mixture design procedure for stone matrix asphalt (SMA).
National Center for Asphalt Technology, NCAT Report No. 97-
3, Auburn University, Alabama; 1997.
[4] National Asphalt Pavement Association. Guidelines for mate-
rials, production, and placement of stone matrix asphalt
(SMA). Technical Working Group (TWG), Publication No. IS
118; 1994.
[5] Cooley LA, Brown ER, Potential of using stone matrix asphalt
(SMA) for thin overlays. National Center for Asphalt Technol-
ogy, NCAT Report No. 03-01, Auburn University, Alabama;
2003.
[6] Brown ER, Mallick RB. Stone matrix asphalt – properties
related to mixture design. National Center for Asphalt Tech-
nology, NCAT Report No. 94-2, Auburn University, Alabama;
1997.
989
[7] Davidson JK, Kennepohl GJ. Introduction of stone mastic
asphalt in Ontario. Report for AAPT meeting in Charleston,
South Carolina, February 24–26; 1992.
[8] Brown ER, Manglorkar H. Evaluation of laboratory properties of
SMA mixtures. National Center for Asphalt Technology NCAT
Report No. 93-5, Auburn University, Alabama; 1993.
[9] Cooley LA, Brown ER, Watson DE. Evaluation of OGFC
mixtures containing cellulose fibers. National Center for Asphalt
Technology, NCAT Report No. 2000-05, Auburn University,
Alabama; 2000.
[10] An introduction to stone mastic asphalt (SMA), introductory
handout. Scanroad, Sweden, Nobel Industry; 1991.
[11] Scherocman JA. Construction of SMA test sites in the US. AAPT
meeting; February 24–26, 1992.
[12] Brown ER. Experience with stone mastic asphalt in the United
States. Alabama: NCAT Publication, Auburn University; 1992.