Construction and Building Materials 128 (2016) 148–153

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier.com/locate/conbuildmat

Engineering properties and performance of asphalt mixtures

incorporating steel slag
Jian-Shiuh Chen ⇑, Shih-Hsiu Wei
Department of Civil Engineering, National Cheng Kung University, Tainan 701, Taiwan

h i g h l i g h t s

We characterize the engineering properties of asphalt mixtures mixed with BOF.

We construct a test road to compare the field performance of BOF test sections.

We discuss construction practices of BOF test sections.

We evaluate performance of BOF test sections.

a r t i c l e i n f o a b s t r a c t

Article history: Lack of natural stone and costly expense of purchasing high-quality aggregate promote the utilization of
Received 18 October 2015 steel slag. Asphalt mixtures incorporating basic oxygen furnace (BOF) steel slag as coarse aggregate were
Received in revised form 17 September prepared and subsequently subjected to laboratory tests to determine the engineering properties of
2016
asphalt concrete. Results showed that the addition of steel slag into asphalt mixtures had high resistance
Accepted 6 October 2016
Available online 21 October 2016
to permanent deformation and moisture-induced damage. Steel slag included angular and rough textured
particles that would enhance the interlocking mechanism and provide good mechanical properties.
According to the laboratory results, a test road was constructed in 2012 by using three different types
Keywords:
Steel slag
of asphalt mixtures as follows: stone mastic asphalt with BOF (SMA-BOF), dense-graded asphalt concrete
Asphalt mix with BOF (DGAC-BOF), and dense-graded asphalt concrete with natural aggregate (DGAC-NA). The rut
Field performance depth of the SMA-BOF section was lowest among the three sites even though it was subject to high stress
induced by braking, accelerating and turning vehicles. Surveys for pavement performance were
conducted at scheduled intervals. The ride quality and the friction characteristics of both BOF sections
performed as well as or better than the section constructed using natural aggregate. Field data indicated
that steel slag could be used as a surface course aggregate in locations where traffic is expected to per-
form heavy braking and cornering maneuvers. No rutting, cracking, or other failures have been observed
to any significant extent on the BOF sections since they opened to traffic throughout the three-year per-
iod. Test results suggest that the use of steel slag as coarse aggregate substitute in surface courses be
technically appropriate.
Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction permanent deformation of the mixture. Lack of good quality of

The properties of the aggregates used in asphalt mixtures have
a direct influence on pavement performance. Aggregates for
asphalt mixtures should also be resistant enough to bear produc-
tion, transportation and construction processes, as well as traffic
loads and climate effects. When an asphalt mixture is designed,
it is normally recommended that angular and rough-surfaced
aggregates are selected, since they tend to improve resistance to
⇑ Corresponding author.
E-mail address: jishchen@mail.ncku.edu.tw (J.-S. Chen).
natural crushed aggregate promotes the use of steel slag. The
industrial by-products from the metallurgical industry have been
applied to base and surface courses in the road structure since
19800 s [1–4]. Steel slag is produced in a basic oxygen furnace
(BOF) or an electric arc furnace. It is estimated that between 7.0
and 7.5 million metric tons of steel slag are produced each year
in the U.S. [3]. BOF steel slag is the largest amount of secondary
resources produced by integrated steel mills. Most of BOF steel slag
has been recycled in developed countries. For instance, about
1.5 million metric tons of BOF steel slag produced in Germany
are used as an aggregate material in construction applications

http://dx.doi.org/10.1016/j.conbuildmat.2016.10.027
0950-0618/Ó 2016 Elsevier Ltd. All rights reserved.
 J.-S. Chen, S.-H. Wei / Construction and Building Materials 128 (2016) 148–153 149

annually [4]. Approximate 97% of BOF steel slag is recycled in the Table 1
same ways in the U.K. [5]. Mechanical properties of BOF steel slag and limestone.

BOF steel slag, a by-product of the conversion of pig iron to Type Tests BOF Limestone Spec.
steel, has existed since the development of the oxygen converter (%) (%) (%)
in the 1950s. BOF steel slag is one the main by-products of an inte- Coarse LA Abrasion 11 25 <30
grated steel plant with rates amounting to around 100 kg per met- aggregate Flat and elongated 3:1 4 14 <15
ric ton of steel [4,5]. The slag primarily contains calcium silicates Flat and elongated 5:1 1 4 <5
Absorption 1.8 1.1 <2
together with oxides and compounds of iron, manganese, alumina Soundness 4.2 5.6 <12
and other trace elements. It was first used in agriculture as a soil One fracture face 100 100 100
amendment to neutralize soil acidity, stabilize soil structure, and Two fracture faces 100 93 >90
increase plant resistance, but its main valorization has concerned Dry bulk specific 3.41 2.62 –
gravity
road works [6–12].
Nevertheless, two factors negatively affect the widespread use Fine aggregate Soundness na 7.9 <15
Sand equivalent value na 94 >50
of BOF steel slag: its possible expansion and its relatively high
specific gravity (often between 3.2 and 3.6), which is due to a cer- Note: na = not applicable because BOF was only used as coarse aggregate in this
tain amount of iron oxides it contains. This leads to some issues for study.
the transport of the material and thus limits the geographical zone
where it can economically be used. The second problem with BOF
concerns the expansion of the material, which is attributed to the strong because it is primarily responsible for carrying traffic loads
hydration of free lime in the form of calcium and magnesium oxi- in an HMA mix. The Los Angeles abrasion value of coarse aggregate
des. The use of steel slag in asphalt mixtures could generally result should be less than 30% to possess sufficient toughness. BOF steel
in a binder film coating the steel slag and limiting potential expan- slag possessed a better LA abrasion value than natural aggregate.
sion [2,6,7]. Furthermore, few studies of its use may have discour- Flat and elongated proportions must be limited to be a maximum
aged highway engineers from better understanding the value. Both BOF steel slag and limestone met the current criteria of
engineering properties and performance of asphalt mixtures incor- no more than 15 percent 3–1 and no more than 5 percent 5–1 flat
porating BOF steel slag. and elongated particles. Steel slag is highly angular in shape, and
This study was motivated by concerns that asphalt pavements tends to be cube-like with very low percentages of flaky and
mixed with BOF steel slag might not perform well under environ- elongated particles.
mental and traffic conditions. It was concerned that the rut resis- Coarse aggregate is required to provide fracture faces with high
tance, friction, and durability benefits of hot-mix asphalt internal friction to carry the load. If the crushed content is signifi-
mixtures with BOF would be lost due to demanding and heavily cantly less than 100% for the coarse aggregate, an HMA mix is likely
trafficked situation at high risk sites such as junction and crossing to be less resistant to shoving and rutting. Although BOF and
approaches. Prior to this study, little research has been performed natural aggregate were 100 percent crushed, steel slag had more
on the performance of asphalt mixtures incorporating BOF steel fracture and rough surfaces that would be beneficial to provide
slag as a surface course aggregate. The use of less traditional sec- microtexture and interlocking mechanism. The most significant
ondary materials in road construction requires the assessment of difference between BOF and natural aggregate is its high specific
their physical and mechanical performance to ensure that the gravity, which is a consequence of the presence of iron compounds
structures built using them are durable and sustainable. Long- in BOF. The dry bulk specific gravity of BOF and limestone used in
term performance, however, especially the influence of steel slag this study was 3.41 and 2.62, respectively. The binder employed in
under field conditions, is often difficult to predict. All aspects of this study was grade AC-20 according to ASTM D 3381. This mate-
using BOF steel slag in asphalt pavements will be evaluated, rial, supplied by the China Petroleum Cooperation, is the usual
including: bitumen grade used for asphalt pavements in Taiwan. The selec-
tion of AC-20 cement is to emphasize the performance of steel slag

Evaluate the effect of steel slag on the engineering properties of used as coarse aggregate in asphalt mixtures.
hot-mix asphalt mixtures in the laboratory,

Construct a test road to compare the field performance obtained 2.2. Asphalt mixture design
from the use of steel slag and the use of traditional aggregate,
and A total of three mix designs were performed as follows: stone

Compare the field performance of asphalt pavements con- mastic asphalt with BOF (SMA-BOF), dense-graded asphalt con-
structed with steel slag with that of a test section constructed crete with BOF (DGAC-BOF), and dense-graded asphalt concrete
with conventional aggregate. with natural aggregate (DGAC-NA). The master aggregate grada-
tion bands listed in Table 2 were based on the field test road
2. Materials selected. The 19-mm maximum aggregate size gradation is gapped
on the 4.75-mm sieve for SMA mixtures. The percentage of mineral
2.1. Aggregate and binder fillers used for SMA mixtures was 9.5 percent. The job mix formula
was decided using the Marshall mix design method to determine
Two types of aggregate were used as follows: limestone the volumetric properties of the asphalt mixtures. The asphalt con-
obtained from the Kao-Ping River, and BOF steel slag supplied by tent of all mixtures was determined as that resulting in 4% air
the CHC Resources Corporation. Steel slag was stockpiled outdoors voids.
for three months to expose the material to moisture from natural Both the control mix (i.e., DGAC-NA) and the slag mix (i.e.,
precipitation. The purpose of aging was to allow potential hydra- DGAC-BOF and SMA-BOF) were fabricated by the Marshall mix
tion and its associated expansion to take place prior to use of design procedure according to ASTM D1559. The Marshall com-
BOF steel slag in asphalt mixtures. The mechanical properties of pactor was used to compact the DGAC samples 75 times each side,
BOF steel slag and natural aggregate are summarized in Table 1. but 50 times each side for the SMA samples. The optimum binder
Steel slag was employed in this study to replace the coarse faction content was determined to be 5.1%, 4.3% and 5.2% for DGAC-NA,
of hot-mix asphalt mixtures (HMA). Coarse aggregate must be DGAC-BOF and SMA-BOF mixtures, respectively. The bulk specific
150 J.-S. Chen, S.-H. Wei / Construction and Building Materials 128 (2016) 148–153

Table 2 2.5
Asphalt mixture design. DGAC-NA DGAC-BOF SMA-BOF

Strength or Modulus
Characteristics DGAC-NA DGAC-BOF SMA-BOF 2
Sieve (mm)
100 (25) 100 100 100.0 1.5
3/400 (19) 95.1 97.4 94.9
1/200 (12.7) 82.5 78.3 57.9 1
3/800 (9.5) 74.2 70.6 45.4
#4 (4.75) 51.7 47.0 24.9
#8 (2.36) 37.2 31.5 18.6 0.5
#16 (1.18) 31.8 25.4 16.6
#30 (0.6) 22.3 18.7 14.7 0
#50 (0.3) 12.5 11.7 12.5 IDT (MPa) M r (GPa )
#100 (0.15) 7.9 7.3 11.0
#200 (0.075) 5.0 4.9 9.5 Fig. 1. Indirect tensile strength and modulus of BOF steel slag and control mixes.
Binder content (%) 5.1 4.3 5.2
Bulk gravity 2.36 2.81 2.88
Voids (%) 4.0 4.0 4.0 100
90

Tensile strength ratio (%)

VMA (%) 15.6 14.3 16.9 86
VFA (%) 74.3 72 76.3 78 Min > 75
75

gravity of slag mixtures is higher than that of traditional mixtures 50

due to the fact that BOF has higher density. Note that the SMA-BOF
mix had the highest asphalt content among three mixtures after its
bulk specific gravity was taken into consideration.
Relatively low voids in the mineral aggregate (VMA) were
observed for the DGAC-BOF mix as compared to the DGAC-NA
mix. Since BOF steel slag shapes like a cube, cubic particles are
more likely to stack up after compaction. The shape of BOF steel
slag may contribute to a reduction in VMA due to limited void
space inside asphalt mixtures. Nevertheless, all mixtures met min-
imal VMA values recommend by highway agencies.
3. Laboratory study
Laboratory tests were performed to evaluate the following engi-
neering properties of asphalt mixes: strength, moisture suscepti-
bility, and rutting resistance. Testing data presented in this study
are the average values of at least three replicate specimens. The
laboratory study centered on the effect produced by steel slag on
the following properties:

Resistance to indirect tension,

Resistance to water damage, and

Resistance to permanent deformation.
3.1. Strength test
The resilient modulus and the indirect tensile tests were con-
ducted to determine the strength of the slag mixtures. The
repeated-load indirect tension test for determining the resilient
modulus was conducted by applying compressive loads with a
haversine waveform according to ASTM D 4123. The load was
applied vertically in the vertical diameter plane of a cylindrical
specimen of asphalt concrete through a curved loading strip. The
resulting horizontal deformation was measured and used to calcu-
late the resilient modulus (Mr) and the indirect tensile strength
(IDT). The indirect tensile strength is defined as the maximum
stress from a diametric vertical force that a Marshall sample can
withstand. Fig. 1 shows the average and range of one standard
deviation of the test results for each mix tested at 25 °C. Although
SMA is not nearly as strong in tension as it is in compression, SMA
tensile strength is important in pavement applications. The indi-
rect tensile strength test is used to determine the tensile properties
of the asphalt mixture that can further be related to the cracking
properties of the pavement.
25
0
DGAC-NA DGAC-BOF SMA-BOF
Fig. 2. Moisture-induced damage results of BOF steel slag and control mixes.
DGAC-BOF mixes exhibited higher indirect tensile strength and
resilient modulus values than DGAC-NA mixes as shown in Fig. 1.
Analysis of variance (ANOVA) was employed in making inferences
about different treatment population means. The null hypothesis is
that no strength differences among specimens of different mix
types could be achieved. At the 0.05 significance level, the test
statistics could reject the null hypothesis. Significant differences
were observed in the indirect tensile strength and the resilient
modulus for asphalt mixtures prepared using different aggregates.
For the dense-graded asphalt mixtures, the indirect tensile
strength of the DGAC-BOF mixes was about 2 times higher than
that of the DGAC-NA mixes as demonstrated in Fig. 1. It appears
that the addition of BOF contributes to an increase in the indirect
tensile strength and the resilient modulus of dense-graded asphalt
mixtures, indicating that slag had a good adhesion to bitumen.
Paving mixtures with low tensile strength values may have a ten-
dency to develop fatigue cracking in the field, especially when used
for surface mixtures. An indirect tensile strength of 1 MPa at 25 °C
is usually considered an average value for asphalt surface mixtures
[7]. The DGAC-BOF mix possessed values of close to 2 MPa, and
reflected better cracking resistance as compared to the DGAC-NA
mix.
Fig. 1 indicates that the resilient modulus and the indirect ten-
sile strength values of SMA mixes were lower than those of dense-
graded asphalt mixtures. Since SMA is a gap-graded asphalt mix,
the amount of stone-on-stone contact is much greater in a SMA
mix and more of the traffic load is carried on the coarse aggregate
particles instead of on the fine aggregate-asphalt cement mortar
[13]. The test results shown in Fig. 1 may not really represent
the true nature of the stone skeleton in the SMA mixes. The key
features of SMA mixes are focused on their resistance to rutting
as well as water damage in this study.
3.2. Moisture-induced damage
The moisture-induced damage was assessed using the Lottman
test in accordance with AASHTO T 283. In this test, three samples
 J.-S. Chen, S.-H. Wei / Construction and Building Materials 128 (2016) 148–153 151

5
Port Area
DGAC-NA DGAC-BOF SMA-BOF
4
Rutdepth (mm)

SMA-BOF

250 m
2 N

0
0 1000 2000 3000 4000 5000 6000 DGAC-BOF DGAC-NA
LoadRepetitions 250 m 250 m

Fig. 3. Rut depth results of BOF steel slag and control mixes.
Fig. 4. Layout of test road (not to scale).

within each specimen group were used as a control. Another three

were vacuum saturated with water without any freeze/thaw cycle,
because the temperature in Taiwan rarely falls below freezing
point. The stripping potential was further quantified by comparing
the average tensile strength of the conditioned samples with that
of the control samples, within each group. The tensile strength
ratio (TSR) is defined as the ratio of the average tensile strength
of the conditioned specimens to that of the control specimens.
As shown in Fig. 2, the DGAC-NA mix showed the highest water
damage while the SMA-BOF mix had the lowest water damage. The
Taiwan Highway Bureau specifies a minimum TSR value of 75% for
asphalt mixes. The addition of BOF into asphalt mixes was shown
to improve moisture resistance, implying that BOF had an excellent
affinity to bitumen. The TSR value of SMA-BOF was approximate
90%, indicating that this mixture is the most robust among the
three mixtures in spite of gap gradation. The SMA mix had a higher
asphalt content resulting in a relatively thick asphalt film on BOF
Fig. 5. Rutting resistance of BOF steel slag and control sites [mo = month].
steel slag that could lead to a better protection against water
damage as well as volumetric expansion. The test results suggest
that in addition to the common emphasis of improving 6000 cycles were less than 5 mm for all mixes, and were consid-
aggregate-binder interfacial properties, it is possible to enhance ered acceptable. BOF steel slag appeared to deliver a high resis-
the resistance to moisture-induced damage of an asphalt mixture tance to permanent deformation under traffic when used for
by changing its gradation as well as aggregate type. asphalt surface layers. BOF steel slag is highly angular in shape,
and has rough surface texture. These characteristics could make
3.3. Rut depth BOF a potentially suitable aggregate in wearing course hot mix
asphalt and in surface treatments.
A wheel-tracking test was performed to evaluate the rutting Fig. 3 shows that DGAC-NA had a rut depth curve similar to that
characteristics of slag mixtures. The asphalt slab was rigidly of DGAC-BOF, indicating that the dense gradation did not signifi-
restrained in a 300 mm  300 mm  50 mm steel mould. The cantly improve the resistance of asphalt mixtures to rutting. The
wheel driven by a motor and a reciprocating device loaded the slab most significant reduction in rutting occurs at the use of the SMA
bi-directionally. The wheel track consisted of a steel wheel gradation. The SMA-BOF skeleton was shown to provide a good
(200 mm in diameter and 50 mm in width) applying a pressure stone-on-stone interlocking mechanism. The dynamic stability
of 689.4 kPa (100 psi) on top of a slab specimen. At a rate of (DS) value was 1788, 2111 and 3553 cycle/mm for DGAC-NA,
42 cycles/min, the wheel covered a loading distance of DGAC-BOF and SMA-BOF mixtures, respectively. The dynamic
230 ± 10 mm. The vertical deformation at the middle of the slab stability greater than 1500 cycle/mm is recommended for an
was recorded. For the standard test condition, loading pressure asphalt mix. The SMA results obtained are in good agreement with
was 700 kPa at 60 °C under dry conditions. The dynamic stability those of other researchers [13,14]. According to the laboratory test
(DS) used to characterize the rutting resistance of each mix is results, BOF steel slag had the same or better engineering proper-
calculated as follows: ties of natural aggregate, and it could be incorporated in an asphalt
t2  t1 mix.
DS ¼ N ð1Þ
d2  d1
where d1 = deformation at t1 minutes (mm), d2 = deformation at t2 4. Field evaluation
minutes (mm), and N = speed of the wheel, 42 cycle/minute in this
study. Default values of t1 and t2 were used in this study, which are 4.1. Test road
at 45 and 60 min, respectively.
Fig. 3 represents the average rut depth curve of each mix from It is important to verify that the engineering properties of a
the wheel-tracking test. The rut depth of the asphalt mix appears material obtained from the laboratory could be validated on a field
to be sensitive to gradation and aggregate type. The rut depths at project. After obtaining laboratory test results, the Tainan City
152 J.-S. Chen, S.-H. Wei / Construction and Building Materials 128 (2016) 148–153

Government began construction of a full-scale test road in January, and left wheel paths. The IRI guideline of 3.5 m/km for acceptable
2012. Three test sections were constructed as follows: stone mastic ride quality is recommended by highway agencies in Taiwan. This
asphalt with BOF (SMA-BOF), dense-graded asphalt concrete with IRI level is in good agreement with drivers’ acceptability [16].
BOF (DGAC-BOF), and dense-graded asphalt concrete with natural Fig. 6 shows that roughness of the three test sections remained
aggregate (DGAC-NA). These test sections are located adjacent to relatively constant over the first five months. Then, after ten
each other on an urban roadway with each section approximately months, the DGAC-NA section had an increase in the IRI values.
250 m in length as shown in Fig. 4. The BOF sections showed little change in roughness over time.
The SMA-BOF test section is situated in an intersection near a The BOF sections appear to be smooth while the DGAC-NA section
commercial port, and connected with the other two test sections. rough.
The SMA-BOF section is located at a high risk site with heavily- A closer examination of the data shown in Fig. 6 indicates that
trafficked and left-turn movements. Each section consists of a variability in the time-sequence roughness on the test road occurs
10-cm asphalt surface course over a 30-cm coarse aggregate base due to the following factors: (1) variation in the profiled path, and
course. The surface course was constructed with two 5-cm lifts (2) seasonal effects. There is considerable transverse variability,
thickness including tack coat and prime coat sprayed in between which can cause variation in roughness, depending on the wheel
to achieve smooth requirements. Following construction, distress path that was followed. Changes in roughness also occur in pave-
surveys were conducted on each section during trafficking. Surveys ments due to changes in profile caused by seasonal effects. Each
for rutting, ride quality, and skid resistance were performed at section tends to follows a cyclic pattern through the year, with
scheduled intervals. rough values in the summer (such as 15 mo and 28 mo) and
smooth values in winter (such as 10 mo and 22 mo). The DGAC-
4.2. Permanent deformation NA section with roughness in excess of 3 m/km exhibited larger
increases in roughness over time as compared to the BOF sections.
Permanent deformation in the wheel path was measured
according to ASTM E1703. This test road has an average traffic vol-
4.4. Friction
ume of 27,500 vehicles per day with approximately 23% truck traf-
fic. Rutting is a common phenomenon on the test sections because
In order to assess whether any changes in skid resistance
they are subjected to frequent passes of traffic and heavy loadings
measured on the BOF sections were related to variation that would
of vehicles. Each bar in Fig. 5 represents the average rut depth of
have occurred on the test road, the three test sections were period-
three replicates, and the error bars indicate one standard deviation
ically surveyed for friction. Pavement skid resistance was mea-
from the average value. The graph in Fig. 5 shows average rut
sured by the British Pendulum Tester according to ASTM E303
depth measured on in-service mixtures of SMA-BOF, DGAC-BOF
and expressed by a British Pendulum number (BPN). This test mea-
and DGAC-NA sections. The rut depth of the DGAC-NA section
sures frictional properties by determining the loss in kinetic energy
was highest among the three sections, that of SMA-BOF the lowest,
of a sliding rubber shoe when it is in contact with the pavement
and that of DGAC-BOF in between. The severity level is considered
surface. All the measurements were taken when the air tempera-
to be low when the mean rut depth is less than 12.5 mm, moderate
ture reached about 25 °C. Changes in BPN values over time are
when rutting is between 12.5 and 25 mm, and high when rutting is
shown in Fig. 7. Test data were averaged for each section and for
higher than 25 mm.
a longitudinal position (left and right wheelpaths) for the British
Test results indicate that SMA-BOF and DGAC-BOF test sections
pendulum test device.
possesses good resistance to plastic deformation since both sec-
Fig. 7 shows an initial BPN value of 46–51. Skid resistance was
tions had rut depth lower than 12.5 mm. In particular, the SMA-
relatively low just after construction because of an asphalt binder
BOF section provides a good example of performance in locations
film coating aggregate at pavement surfaces. As a consequence of
where frequent braking, accelerating and turning occurs. Rutting
the disappearance of the binder film covering the surface of the
ranged from 4 mm for the SMA-BOF section to 20 mm for the
aggregate, skid resistance improved after test sections opened to
DGAC-NA section after three years in service as shown in Fig. 5.
traffic. After the initial period of wearing in, road surfaces eventu-
The SMA-BOF section has a coarse gap-graded gradation that could
ally reach an equilibrium state of polishing. Since the traffic level is
result in stone-on-stone contact. The presence of BOF steel slag on
almost constant for all test sections, skid resistance fluctuated
the approach to the traffic light junction could be seen as a sharp
through seasonal weathering and polishing cycles, but remained
reduction in permanent deformation. The angularity and tough-
at about a constant level. ANOVA at a significance level of 0.05
ness of BOF is shown to improve the rutting resistance of SMA mix-
was used to test the significance of the aggregate types on the
tures. According to field observations, BOF is well encapsulated by
a rich mortar binder in SMA. The SMA-BOF section with a greater
proportion of mastics would be expected to be durable. The use
of BOF steel slag as coarse aggregate is shown to present appropri-
ately technical solutions for road applications, which could provide
an excellent level of resistance to permanent deformation.

4.3. Pavement smoothness

Pavement smoothness is widely regarded as the most signifi-

cant measure of pavement performance because it is the measure
most evident to the traveling public [15,16]. The international
roughness index (IRI) is generally used as an indicator of pavement
surface conditions, and expressed by m/km. The ICC Surface Pro-
filer used in this study is a multi-wheeled inclinometer-based sys-
tem that is pushed by an operator at a walking speed of 1.2 km/h.
Fig. 6 demonstrates a plot of ride quality collected over time. Each
IRI value represents an average across at least two runs of the right Fig. 6. Ride quality of BOF steel slag and control sites [mo = month].
 J.-S. Chen, S.-H. Wei / Construction and Building Materials 128 (2016) 148–153
Fig. 7. Skid resistance of BOF steel slag and control sites [mo = month].
BPN value. The test results showed the aggregate type was not a
statistically significant factor, with a p-value of more than 0.05.
In addition, the mix type was determined to be not statistically sig-
nificant at the 95% confidence level. It seems that the aggregate
and mix types were not different enough to influence the mea-
sured BPN value.
The influence of weather fluctuation and the evolution of
frictional characteristics over time are presented in Fig. 7. During
seasons with warmer temperatures (such as 15 mo and 28 mo),
skid resistance decreased, and during seasons with cold tempera-
tures (such as 10 and 22 mo), skid resistance increased. It is likely
that runoff of rainwater tends to bring surface contaminants such
as dust and sand on the road surface. In the summer when the
rainfall in Taiwan is high, runoff is at its highest and deposition
of contaminants tends to increase; hence, at this time, skid resis-
tance would be low. On the other hand, during the winter months,
detritus roughens aggregate surfaces; as a result, skid resistance
would be high at that time of year.
Skid resistance on the test sections not only varies from year to
year, it also occurs during different periods. The pattern for both
BOF steel slag and the control sites is similar, and this indicates
that BOF steel slag coarse aggregate performs similarly to natural
aggregate in terms of skid resistance. A BPN value higher than 45
is considered sufficient and safe for roadway pavements [17]. The
BPN value of test sections built with BOF appears to provide good
skid resistance after three years in service.
No raveling, cracking or other failures have been observed on
the BOF sections to any significant extent since opening to traffic
in 2012. According to the results obtained from field surveys, the
positive features of BOF steel slag in hot-mix asphalt mixtures
included good frictional properties, high stability, and resistance
to rutting.
5. Conclusions
On the basis of the test results and the performance survey con-
ducted in this study, the following conclusions can be drawn. Basic
oxygen furnace (BOF) steel slag could meet the specification
requirements set by highway agencies as road construction aggre-
gate. The use of steel slag as a coarse aggregate substitute could
enhance the engineering properties of asphalt mixtures. BOF steel
slag included angular and rough textured particles that would
increase the interlocking mechanism. The tensile strength ratio
153
showed the superior moisture resistance of steel-slag-based
asphalt mixtures to natural-aggregate-based asphalt mixtures.
The slag mixture had better rutting resistance when compared
with the conventional one. In-service pavements of three different
types of asphalt mixtures were constructed and surveyed periodi-
cally. On the sites tested, steel slag sections performed as well as or
better than the equivalent control site constructed using natural
aggregate. The steel slag section was shown to have good ride qual-
ity and provide sufficient skid resistance. Furthermore, the field
work demonstrated that steel slag could be used as coarse aggre-
gate in locations where traffic is expected to perform heavy brak-
ing or turning maneuvers. Test results from the laboratory and
the test road suggest that the use of steel slag as aggregate be a
viable option for hot-mix asphalt mixtures.
Acknowledgements
The authors are grateful for the National Science Council project
(NSC 101-2221-E-006-214-MY3), the Tainan City Government, and
the CHC Resources Corporation for providing financial and field
supports to complete this study.
References
[1] J.J. Emery, Slag Utilization in Pavement Construction. Extending Aggregate
Resources, ASTM STP 774, 1982. pp. 95–11.
[2] P. Kandhal, G. Hoffman, Evaluation of steel slag fine aggregate in hot-mix
asphalt mixtures, Transp. Res. Rec. 1583 (1997) 28–36.
[3] Federal Highway Administration, User Guidelines for Waste and By-Products
in Pavement Construction, FHWA-RD-97-148, 1998.
[4] M. Legret, P. Chaurand, A. Benard, Y. Capowiez, D. Deneele, J. Reynard, L.
Lassabatere, D. Yilmaz, J. Rose, J. Domas, B. Bechet, D. Richard, Y.J. Bottero, A
multidisciplinary approach for the assessment of the environment behavior of
basic oxygen furnace slag used in road construction, in: 6th European Slag
Conference, 2010, pp. 77–88.
[5] P.G. Roe, Basic Oxygen Steel Slag as Surface Course Aggregate: An Investigation
of Skid Resistance, TRL566, 2003.
[6] G.D. Aiey, A.C. Collop, N.H. Thom, S.E. Zoorob, A. Shiratori, Laboratory
evaluation of secondary aggregates in bituminous mixtures, J. Assoc. Asphalt
Paving Technol. 73 (2004) 731–769.
[7] A.S. Noureldin, R.S. MacDaniel, Performance evaluation of steel furnace slag-
natural sand asphalt surface mixtures, J. Assoc. Asphalt Paving Technol. 59
(1990) 276–303.
[8] U. Bagampadde, H.I. Al-Abdul Wahhab, S.A. Alban, Optimization of steel slag
aggregates for bituminous mixes in Saudi Arabia, J. Mater. Civ. Eng. 11 (1999)
30–35.
[9] P.Y. Mahieux, J.E. Aubert, G. Escadeillas, M. Measson, Quantification of
hydraulic phase contained in a basic oxygen furnace slag, J. Mater. Civ. Eng.
26 (2014) 593–598.
[10] D. Deniz, E. Tutumluer, J.S. Popovics, Evaluation of expansive characteristics of
reclaimed asphalt pavement and virgin aggregate used as base materials,
Transp. Res. Rec. 2167 (2010) 10–17.
[11] A. Coomarasamy, T.L. Walzak, Effect of moisture on surface chemistry of steel
slags and steel slag-asphalt paving mixes, Transp. Res. Rec. 1492 (1995) 85–95.
[12] B. Fronek, P. Bosela, D. Delatte, Steel slag aggregate used in Portland cement
concrete – U.S. and international perspectives, Transp. Res. Rec. 2267 (2012)
37–42.
[13] B. Prowell, L.A. Cooley, R. Schreck, Virginia’s experience with 9.5-mm nominal-
maximum-aggregate-size stone matrix asphalt, Transp. Res. Rec. 1813 (2002)
133–141.
[14] E. Brown, J. Haddock, Method to ensure stone-on-stone contact in stone matrix
asphalt paving mixtures, Transp. Res. Rec. 1583 (1997) 11–18.
[15] B. Choubane, R.L. McNamara, G.C. Page, Evaluation of high-speed profiles for
measurement of asphalt pavement smoothness in Florida, Transp. Res. Rec.
2002 (1813) 62–67.
[16] K. Shafizadeh, F. Mannering, Acceptability of pavement roughness on urban
highways by driving public, Transp. Res. Rec. 2003 (1860) 187–193.
[17] J. Donbavand, Skidding Resistance of Road Surfaces – Implication for New
Zealand, 1989. Transit New Zealand.