Construction and Building Materials 177 (2018) 332–341

Contents lists available at ScienceDirect

Construction and Building Materials

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

Moisture-induced damage resistance of asphalt mixture entirely

composed of gneiss and steel slag
Zongwu Chen a, Yuyong Jiao a,⇑, Shaopeng Wu b, Fubin Tu a
a
Faculty of Engineering, China University of Geosciences (Wuhan), Wuhan 430074, China
b
State Key Laboratory of Silicate Materials for Architectures, Wuhan University of Technology, Wuhan 430070, China

h i g h l i g h t s

Gneiss and re-crushed fine steel slag were simultaneously used.

Gelling of silicate minerals contributed to the solidification of steel slag.

Nature acidity and layered micro-structure of gneiss caused adverse effect.

Steel slag powder improved moisture resistance of asphalt mixture.

Surface status of slag fine aggregate effected moisture stability of mixture.

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

Article history: Gneiss and fine steel slag are seldom used in the construction of asphalt pavement due to their poor per-
Received 29 December 2017 formances. Asphalt mixture entirely composed of gneiss (coarse aggregate) and steel slag (fine aggregate
Received in revised form 27 April 2018 and asphalt filler) was proposed in this research in order to promote their wide use. Limestone fine aggre-
Accepted 9 May 2018
gate and filler were functioned as control groups. The inadequacy of using 100% of gneiss to prepare
asphalt mixture and the performance deterioration mechanism of fine steel slag were discussed based
on some newly revealed material characteristics and previous research results, and corresponding
Keywords:
improvement measures were established first. Then four asphalt mixtures (two fine aggregate and two fil-
Gneiss
Steel slag
ler) were designed by Superpave procedure. Finally, the moisture stabilities of uncompacted aggregate-
Asphalt mastic mastic system and compacted asphalt mixture were fully investigated, and hot water damage mode
Asphalt mixture and freeze-thaw damage mode were adopted. Results suggested that asphalt mixture entirely composed
Moisture stability of coarse gneiss and re-processed fine steel slag possessed satisfactory moisture resistance performance.
Ó 2018 Elsevier Ltd. All rights reserved.

1. Introduction amount of natural resources. Especially in China, the annually con-

struction volume is huge because of the large land area. Only the
Asphalt mixture, cement concrete and semi-rigid materials are total mileage of expressway has reached 136,500 km by the end
widely used in the construction of infrastructure [1–5], especially of 2017 [8], and more than 90% of which is paved with asphalt mix-
in road construction. The road consist of semi-rigid base and ture [9]. Many methods to reduce the dependence of pavement con-
asphalt surface course possesses long lifespan and superior service struction on natural stone resource have been tried. Recycling of
performances (excellent rideability, low noise, high bearing capac- solid wastes (demolition waste [10], smelting waste [1,6,7,9,11–
ity, etc.) simultaneously [2,3]. Hence, the semi-rigid base asphalt 16] etc.) and low-grade natural minerals (acid rock[14,17]) in
pavement is extensively used in the construction of high-grade asphalt mixture is a very important strategy to release the supply
highways in China. Surface course is paved with asphalt mixture, pressure of superior resources. The future pavement needs to be
which mainly composes of aggregate and asphalt mastic (asphalt multifunctional [18–20], resource-saving [14,17,21,22], energy-
binder and filler) [6,7]. The aggregate and filler approximately saving [18,23] and environmentally friendly [19,20,24]. The
account for 95% of volume of dense-graded asphalt mixture. There- exploitation of secondary aggregate can contribute to this goal.
fore, the construction of high-grade pavement consumes a huge Steel slag is a typical smelting waste, generating during steel-
making. Its output accounts for 13% of raw steel [25]. Considering
⇑ Corresponding author. its large output can partly meet the high demand of mineral mate-
E-mail address: [email protected] (Y. Jiao). rials for road construction, the steel slag based asphalt mixture is a

https://doi.org/10.1016/j.conbuildmat.2018.05.097
0950-0618/Ó 2018 Elsevier Ltd. All rights reserved.
 Z. Chen et al. / Construction and Building Materials 177 (2018) 332–341 333

hot research field in recent years. Previous research results indi- 2.2.2.1. Uncompacted aggregate-asphalt mastic system.
cated that steel slag coarse aggregate (SSCA) can improve many 2.2.2.1.1. Hot water damage mode. The water boiling test is originally developed to
performances of asphalt mixture such as high-temperature stabil- investigate the adhesive level between coarse aggregate and pure asphalt binder
[30]. The adhesive level was determined based on the stripping percent of asphalt
ity [12,15,25,26], fatigue life and fatigue durability [11,27], binder from the surface of asphalt coated aggregate particle after suffering water
moisture-induced damage resistance [11,13] and skid resistance boiling damage for 3 min. While the stripping percent is determined by visual
[25,26]. While fine steel slag is not as attractive as SSCA because inspection, which is very subjective, and the boiling time of 3 min might not be
of its poor angularity and substandard surface status [12,14]. enough to pick out the winner among different aggregate-mastic systems. Hence,
the traditional water boiling test shows some disadvantages.
What’s more, the introduction of the whole gradation of steel slag
Some modifications have been made on water boiling test in this research. It
(coarse and fine aggregate) in asphalt mixture will give a great risk was used to evaluate the adhesive level between aggregate (both coarse and fine
in raising asphalt content. The whole gradation of steel slag is not particles) and asphalt mastic. The volume loss percent of asphalt mastic desqua-
acceptable. Therefore, the application of fine steel slag is limited. mating from the surface of asphalt mastic coated particle was proposed as a new
Gneiss is a typical low-grade natural mineral when referring to index to determine the aggregate-mastic system’s adhesive level. The maximum
boiling water damage time can reach 30 min. The volume loss percent was com-
asphalt mixture because of its acidity [17]. The bonding perfor- puted according to Eq. (1):
mance between acid aggregate and asphalt binder is inferior. The
Mam M0am
poor bonding behavior would worsen many performance of asphalt qm M am  M 0am
Vl ¼ MamMa
 100% ¼  100% ð1Þ
mixture such as moisture-induced damage resistance. Therefore, M amMa
qm
the project decision makers prefer to use more secure natural min-
where Vl is the volume loss percent, %; Ma is the original dry mass of aggregate par-
erals such as limestone and basalt. The use of gneiss in asphalt ticle, Mam is the dry mass of asphalt mastic coated particle before subjected to boiling
pavement requires serious consideration in China at current situa- water damage, g; M0am is the residual dry mass of mastic coated particle after boiling
tion. In order to reduce the consumption of high-grade mineral, uti- water damage, g; qm is the density of asphalt mastic, g/cm3.
lizing the alternative pavement materials to the greatest extent is The bonding characteristics of coarse aggregate-mastic system under water
boiling damage were evaluated first. Two types of asphalt mastic were involved.
anticipated. Therefore, it will have a broad application prospect if
One consist of asphalt and SSP, and the other one was composed of asphalt and
the combined use of gneiss and steel slag becomes a reality. LSP. Therefore, there were two gneiss-mastic systems in this research. Considering
Preparing asphalt mixture entirely composed of coarse gneiss the viscosity of modified asphalt is often very large. Base asphalt AH-70 with a
and fine steel slag (fine aggregate and filler) was considered in this much lower viscosity was adopted in order to determine the winner of different
research. Based above, the key factor to realize the synchronous aggregate-mastic system efficiently. The water boiling test of coarse aggregate-
asphalt mastic system was conducted as following: (1) The approximating cube
utilization of coarse gneiss and fine steel slag is enhancing the
gneiss coarse particles with size of 19 mm were selected. (2) The suitable blending
bonding performance of asphalt mixture system. The moisture- proportion of asphalt binder and filler (SSP or LSP) was determined based on the
induced damage resistance of asphalt mixture is very sensitive to distribution uniformity of filler particles in mastic. (3) Hanging the dry gneiss par-
the bonding performance. Hence, two main works were con- ticles into hot asphalt mastic and ensuring the particle surfaces were fully covered,
and then moving the mastic coated particles to an oven of 45–60 °C, in which the
ducted: (1) The inadequacy of directly using gneiss and fine steel
redundant asphalt mastic naturally dropped down. (4) Finally subjecting the suffi-
slag was revealed, and corresponding improvement measures were ciently cooled mastic coated particles to water boiling damage, and the volume loss
proposed first. (2) And then the bonding characteristics of uncom- percent of asphalt mastic was quantized according to Eq. (1).
pacted aggregate-mastic system and compacted asphalt mixture There were four fine aggregate-mastic systems (two types of fine aggregate and
under multiple moisture damage modes were deeply investigated. asphalt mastic) needed to be investigated. Considering the volume of particles in
fine aggregate is much smaller, hanging the mastic coated fine particles is not con-
The research flow chat of this research was shown in Fig. 1. Lime-
venient, and the test results of different particles may also fluctuates significantly
stone fine aggregate and filler were functioned as control group. due to the hard control of particle’s geometry. Therefore, the water boiling test of
fine aggregate-asphalt mastic system was a little different: (1) 100 g steel slag fine
2. Materials and methods particles with size of 2.36 mm were prepared, and the mass of used limestone fine
particles was 100/qslag  qlimestone g in order to keep the volume of steel slag fine
2.1. Raw materials particles and limestone fine particles the same. (2) Then they were mixed with
the same volume’s asphalt mastic, respectively. The dosage of used asphalt mastic
In this research, base asphalt binder AH-70 with penetration of 66, ductility of can fully coat the surface of fine particles was enough, and ensuring no free asphalt
159 cm (5 cm/min, 15 °C), softening point of 46.1 °C and Styrene-Butadiene-Styrene mastic existed. (3) Considering there existed mass loss of asphalt mastic during
(SBS) modified asphalt binder with penetration of 68, ductility of 40 cm (5 cm/min, mixing process, taking the mass of mixed particles moved into the heating con-
5 °C) and softening point of 64 °C were used. The used coarse aggregate was gneiss tainer as Mam. Water boiling damage was applied and the volume loss percent of
(＞2.36 mm), and fine aggregates were steel slag (SSFA) and limestone. Steel slag asphalt mastic was also measured based on Eq. (1).
powder (SSP) and limestone powder (LSP) were also functioned as fillers. The tech- 2.2.2.1.2. Freeze-thaw cycle damage mode. The standard treatment process of single
nical properties of aggregate and filler were measured according to Chinese stan- freeze-thaw damage used in the field of asphalt mixture is that freezing samples at
dard methods [28], and the test results were listed in Tables 1 and 2, a freeze of -18 °C for 16 h and then thawing them at a hot water bath of 60 °C for 24
respectively. Results suggested that all technical properties of used aggregates h. Considering the asphalt mastic coating on the particle surface may be not strong
and fillers met the requirements of Chinese standard [29]. enough to resist the longtime hot water damage, and the flow of asphalt mastic may
happened. Therefore, the freeze-thaw condition has made some modifications. The
thaw process was conducted by placing frozen sample at room temperature with-
2.2. Experimental methods
out water environment to avoid the flow of asphalt mastic. The single freeze-thaw
damage process of aggregate-mastic system was conducted as following: (1) The
2.2.1. Inadequacy analysis
asphalt mastic coated aggregate particles were immersed in room temperature
The inadequacy of directly using gneiss and fine steel slag in asphalt mixture
water for more than 15 min in advance. (2) Then the particles were subjected to
were discussed based on their material characteristics. The mineral phase, surface
freeze-haw damage. Freeze-thaw cycle mode was applied until obvious stripping
texture and average chemical component of gneiss and steel slag were detected
phenomena of surface coating layer can be observed. The freeze-thaw damage
by X-ray diffraction (XRD) and Polarizing Optical Microscopy (POM), Scanning Elec-
was different from hot water damage. The former caused failure by generating vol-
tron Microscope (SEM) and X-ray fluorescence (XRF), respectively.
ume expansion and shrinkage, and the later caused failure by stripping mastic
directly. The freeze-thaw damage may resulted in the loss of loose section on aggre-
2.2.2. Moisture stability evaluation gate surface during volume expansion and shrinkage process, especially for slag fine
Aggregate and asphalt mastic are the main components of asphalt mixture. aggregate, which easily generates new components on the surface of slag particles
Essentially, moisture damage works on the aggregate-asphalt mastic system. There- during weathering process. The weight difference of mastic coated particles before
fore, the bonding characteristics of uncompacted aggregate-asphalt mastic system and after freeze-thaw damage may not be equal to the mass of stripped mastic.
were evaluated first, and then the traditional moisture stability indicators of com- Therefore, measuring the volume loss of mastic coated particles according to Eq.
pacted asphalt mixture including Retained Marshall Stability (RMS) and Tensile (1) was not suitable. The volume loss was computed as following:
Strength Ratio (TSR) were measured. Two moisture damage modes were adopted,
Vs
namely hot water damage and freeze-thaw damage. Four replicates for each test Vl ¼  100% ð2Þ
were considered. V am
334 Z. Chen et al. / Construction and Building Materials 177 (2018) 332–341

Traditional fine steel slag Gneiss aggregate

Performance deterioration Disadvantage analysis of
mechanism using in asphalt mixture
Improved processing method
Steel slag powder
(for large-scale production)
Limestone powder Control method
Steel slag fine aggregate (for engineering application purpose)
Fillers
Limestone fine aggregate Asphalt

Fine aggregates Asphalt mastic Gneiss coarse aggregate

Fine aggregate-mastic Coarse aggregate-mastic Asphalt mixtures

Hot water damage mode Hot water damage mode

Volume loss of mastic layer RMS test

Freeze-thaw damage mode

Freeze-thaw damage mode
Volume loss of mastic coated
aggregate particles TSR test

Fig. 1. The research flow chat of this research.

Table 1
The technical properties of aggregates.

Parameter measured Coarse aggregate Fine aggregate Requirements in China

Gneiss Limestone SSFA
19–9.5 9.5–4.75 4.75–2.36 2.36–0
Apparent specific gravity 2.742 2.736 2.738 2.697 3.361 2.5
Flakiness and elongation (%) 11.6 14.9 NA NA NA 18
Los Angeles abrasion (%) 22.1 22.1 20.9 NA NA 30
Crush value (%) 20.5 NA NA NA NA 28
Fine aggregate angularity (%) NA NA NA 49 50 30
Sand equivalent (%) NA NA NA 60 63 60

Table 2 were tested by HM-3000 master loader with a constant deformation rate of 50
The technical properties of fillers. mm/min. The indicator RMS, demonstrating the hot water damage resistance abil-
ity of compacted asphalt mixture, was determined according to Eq. (3):
Properties LSP SSP Requirements in China
3
Specific gravity (g/cm ) 2.701 3.321 2.5 MSi
Percent passing (%) 0.6 mm 100 100 100 RMS ¼  100% ð3Þ
MS0
0.15 mm 97.1 93.7 90–100
0.075 mm 89.3 84.3 75–100
where MS0 is the average Marshall stability of control group, kN; MSi is the average
Marshall stability of conditioned group after immersing treatment for i h (i = 24, 48,
where Vl is the volume loss percent of mastic coated particles, %; Vs is the volume of 72 in this research), kN.
stripped part, ml; Vam is the original volume of mastic coated particles, ml. Vs and Vam
are determined by drainage method.
2.2.2.2.2. Freeze-thaw cycle damage mode. The preparation process of
tested specimens was the same as the version used in hot water damage test.
The only difference was that the air void of SGC compacted specimens was 6–7%.
2.2.2.2. Compacted asphalt mixture system. Four asphalt mixtures with the nominal
It can be realized by reducing the gyratory numbers of SGC. The conditioned groups
maximum size of 12.5 mm were designed by Superpave procedure. The mineral
were subject to standard freeze-thaw damage for n cycles (freeze at 18 °C for 16 h
compositions for each mixture were listed in Table 3. The used volume proportions
and thaw at 60 °C for 24 h as a cycle). The Indirect Tensile Strength (ITS) of all spec-
of coarse aggregate, fine aggregate and asphalt filler were 66%, 30% and 4%, respec-
imens were measured by MTS equipment, also applying 50 mm/min deformation
tively. The design process was conducted according to standard Superpave proce-
speed. The indicator TSR, indicating the resistance ability of compacted asphalt
dure and SBS modified asphalt binder was used.
mixture under freeze-thaw cycle damage, was determined according to Eq. (4):

2.2.2.2.1. Hot water damage mode. Cylindrical specimens with 63.5 mm thick and
ITSn
100 mm in diameter were used. They were prepared by coring and cutting the large TSR ¼  100% ð4Þ
ITS0
size specimens with air voids of 4–5% provided by Superpave Gyratory Compactor
(SGC). The prepared specimens were divided into four groups. One was functioned
as control group, and the others were conditioned groups. Conditioned groups were Where ITS0 is the control group’s average ITS, MPa; ITSn is the average ITS of condi-
immersed in water bath of 60 °C for i h. The Marshall stabilities of all specimens tioned group after freeze-thaw damage for n cycles (n = 1, 2, 3 in this research), MPa.
 Z. Chen et al. / Construction and Building Materials 177 (2018) 332–341 335

Table 3
The mineral composition of designed asphalt mixtures.

Coarse aggregate Fine aggregate Filler

Size range (mm) 19–9.5 9.5–4.75 4.75–2.36 2.36–0
Volume proportion (%) 24 20 22 30 4
GSS Gneiss SSFA SSP
GSL SSFA LSP
GLS Limestone SSP
GLL Limestone LSP

3. Results and discussions XRD analysis and optical microscopy observation results both
proved that steel slag was rich in silicate minerals and iron related
3.1. Inadequacy analysis of fine steel slag and gneiss phases. Although the content of f-CaO was much lower compared
to them, its effect on the use of steel slag was more obvious. The
In this section, the inadequacy of directly using gneiss and steel f-CaO consisted of two parts: residual lime from steelmaking raw
slag was discussed based on their material characteristics, and materials and precipitated lime from liquid steel slag [32]. The
which mainly included mineral phase, average chemical compo- generation temperature of liquid steel slag was beyond 1300 °C.
nent and surface texture. Some previous research results were also The activity of f-CaO was still good although slight over-burning
cited here to explain the inadequacy. happened at such high temperature. The f-CaO can easily cause
volume expansion when contacting with water. Weathering treat-
ment, placing steel slag particles in natural condition to eliminate
3.1.1. Fine steel slag the activity of f-CaO before their utilization, was commonly
The gradual performance deterioration of fine steel slag is a adopted to improve the stability of steel slag. This method was also
main factor that limiting its wide use in asphalt mixture. The per- economical and convenient to process large-scale original slag. The
formance deterioration is closely related to its material character- f-CaO transformed into stable calcium carbonate during weather-
istics, especially the surface traits. The representative mineral ing process by reacting with water and carbon dioxide. While the
phases results of fresh steel slag detected by XRD were shown in silicate minerals similar to cement clinker also showed certain gel-
Fig. 2. It can be seen that the diffraction peaks were intricate and ling property during the weathering process. The transformation of
seriously overlapped. It indicated that the components of steel slag silicate minerals has been confirmed in previous research [31].
were complex. Fig. 2 showed that almost all strong peaks belonged Therefore, weathering process essentially was the simultaneous
to silicate minerals and iron related phases. The previous research transformation of f-CaO and silicate minerals. Their transformation
also support it [31]. The former included tricalcium silicate process was as following [14]:
(3CaOSiO2, C3S) and dicalcium silicate (2CaOSiO2, C2S), and the
later involved more phase types: metal iron (MFe), bivalent and f  CaO þ H2 O ! CaðOHÞ2 ð5Þ
ferric iron (FeO, Fe2O3) and complex RO phases (mainly FeO-
MgO-CaO-P2O5 system). The rich iron was responsible for the high 3CaO  SiO2 ð2CaO  SiO2 Þ þ H2 O
specific gravity of steel slag. Free lime (f-CaO) was the main harm-
! ðCaOÞx  ðSiO2 Þy  ðH2 OÞz þ CaðOHÞ2 ð6Þ
ful substances that limit the wide use of steel slag. It was not easily
detected during XRD analysis due to its low content in steel slag.
The POM were further adopted to examine the XRD analysis CaðOHÞ2 þ CO2 ! CaCO3 þ H2 O ð7Þ
results. Fig. 3 showed the observed mineral morphology. These pic- A great amount of powder generated during the production of
tures suggested that silicate minerals and iron related minerals can steel slag aggregate. Therefore the proportion of powder in fine
be easily observed and the distribution traits of these minerals steel slag was very large. The surfaces of aggregate particles in
were revealed. Red hematite particles distribute around SSFA were seriously coated by the powder. This was because the
xenomorphic-hypidiomorphic C3S and spherical C2S particles, used raw materials (blocky steel slag) was usually hydrous and
and metal iron connected to each other as a whole piece in sepa- humid. The solidification of coating layer on the particle surface
rate regions. happened during the weathering process because of the gelling
property of silicate minerals (see picture a in Fig. 4), and the parti-
cles also cemented together as shown in Fig. 4(c).
: 3CaO SiO2 : FeO Except for coating the surface of fine particles, the hardening of
: 2CaO SiO2 : Fe2O3 surplus powder also happened. It made the direct use of fine aggre-
gate almost impossible. The gelling behavior also effected the per-
Relative Intensity(CPS)

O: Fe : RO : f-CaO

(a) (b)
C2S
O
C3S
O
Hematite

10 20 30 40 50 60 70
2 () 200 200
Fig. 2. Mineral phases of fresh steel slag. Fig. 3. Mineral morphology of fresh steel slag.
336 Z. Chen et al. / Construction and Building Materials 177 (2018) 332–341

(a) Slag particle Powder (b) 80

70
Acid rock

60
Alkaline rock

Contents (%)
Coating layer 50
40
30
Original particle 10 mm 20
10
(c) (d)
0
SiO2 Al2O3
SiO2 Al2O3 CaO Fe2O3 K2O
CaO Fe2O3 K2O Na2O
Na2O

Fig. 6. Main chemical components of gneiss.

10 mm 10 mm below 2.36 mm. It exactly meet the maximum size for fine aggre-
gate particles (<2.36 mm) specified in Chinese standard [28]. The
Fig. 4. (a) Solidification process of fine particles and powder on the surface of coated particles with relative large size such as 4.75 mm can be
relatively large-size particles in SSFA, (b) morphology of SSCA with coating layer, (c) fully crushed by doing this. The hardened powder blocks were also
morphology of cemented particles in SSFA, and (d) morphology of SSCA with clean
crushed adequately and some new powder generated. (2) Large
surface.
proportion of powder made the gradation of re-crushed fine steel
slag not satisfied the requirement of technical specification [29].
Therefore, the residual powder were separated from the re-
formance of steel slag coarse aggregate (SSCA). It covered the rich crushed fine steel slag by screening, and it was used as asphalt fil-
surface texture and lowered the angularity of coarse particles in ler. Therefore, fine aggregate and filler were simultaneously pre-
SSCA (see picture b and d in Fig. 4), and which was not beneficial pared in this research as shown in Fig. 5.
for the use in asphalt mixture. Therefore, the performance deterio-
ration of slag aggregate should mainly attribute to the gelling
behavior of silicate minerals. 3.1.2. Gneiss aggregate
In order to solve the adverse effect caused by silicate minerals, The acid-alkali characteristics of rocks are usually simply classi-
modification has been made on the performance control of SSCA fied according to their SiO2 content [34]. The SiO2 content of gneiss
such as placing water cleaning process. The first used key equip- was higher than 70% as shown in Fig. 6, which was beyond the
ment was drum screen [33], being later replaced by vibrating minimum value for acid rocks (SiO2 > 65%). So the used gneiss
screen as shown in Fig. 5. It worked as follows: (1) Transferring was a kind of acid rock. Fig. 6 also showed that except for SiO2,
the particles to vibrating-dewatering screen with a certain tilt gneiss contained some percent of Al2O3, CaO, Fe2O3, K2O and Na2O.
angle (ensuring the free movement of coarse particles under the The different mineral phase types of some material can be seen
vibration of screen), (2) and then the coated powder was washed as the various combination of its chemical components. The min-
away by the vertical high-pressure water jet and the sewage con- eral phase observation results of gneiss were shown in Fig. 7. The
taining powder flowed into the storage pool through the sieve figures indicated that gneiss was mainly composed of quartz
pores. (SiO2), albite (K2O-Al2O3-SiO2), plagioclase (Na2O-CaO-Al2O3-SiO2)
While this method was not suitable for treating fine steel slag, and biotite (K2O-Al2O3-SiO2-Fe2O3). They agreed with the main
because the fine aggregate itself required some parts of powder chemical components (Fig. 6). The acidity characteristic of gneiss
and it was also impractical to treat such huge amount of powder made it uncompetitive when selecting raw materials for asphalt
by water washing. Hence, a new method suitable for processing mixture because of poor bonding performance with asphalt binder
large-scale hardened fine steel slag was proposed in this research. or mastic.
It worked as following: (1) Subjecting the hardened slag block to Except for the acidity, the micro-texture of gneiss was also obvi-
secondary crushing, and ensuring the sizes of all particles were ously different from that of other common aggregates such as

Fig. 5. The production process of steel slag aggregate.

 Z. Chen et al. / Construction and Building Materials 177 (2018) 332–341 337

asphalt mixture system [17]. Steel slag was reported to be alkaline

(a) (b) and it has great potential to improve the bonding performance of
asphalt mixture system [31]. In this research, using the re-
processed SSFA and steel slag powder (SSP) to replace the fine part
and filler in gneiss mineral mixture, respectively, can combine
these two strategies. Therefore, making asphalt mixture entirely
composed of coarse gneiss and fine steel slag was theoretically ten-
Quartz X100 Albite X100 able and very worthwhile trying.

(c) (d)
3.2. Moisture damage resistance evaluation

In asphalt mixture, the aggregate particles were coated by

asphalt mastic consist of asphalt binder, filler and the powder from
fine aggregate. Therefore, the moisture mainly worked on the mas-
tic and the interface between aggregate and asphalt mastic. In
Plagioclase X100 Biotite X100 accordance with it, generating two types of failure: the internal
failure of asphalt mastic and the interface failure of aggregate-
Fig. 7. Mineral morphology of gneiss.
mastic. While traditional moisture damage evaluation was focus
on mechanics index changes of compacted asphalt mixture such
as Marshall stability and strength, and which did not directly
limestone and basalt. As shown in Fig. 8, gneiss presented obvious reveal the changes of asphalt mastic and aggregate-mastic inter-
layer accumulation structure and the surface of each layer was face. Therefore, except for traditional moisture evaluation for com-
smooth. This structure also resulted in some disadvantages when pacted asphalt mixture, moisture damage on uncompacted
referring to be used in asphalt mixture: (1) The smooth texture aggregate-mastic system were also evaluated to fully reveal the
was not good for the physical bonding performance between changes of asphalt mastic and aggregate-mastic interface in this
aggregate and asphalt binder or mastic due to smaller specific sur- research.
face area. (2) Smooth texture showed adverse effect on the stability
of aggregate interlock although the macrostructure of aggregate 3.2.1. Aggregate-asphalt mastic system
played the major role. The suitable blending proportion of asphalt binder and filler in
Based above, preparing asphalt mixture with 100% of gneiss the used mastic in this section was determined first. As shown in
was not desirable. Using other types of raw materials with superior Fig. 9, mastic samples from the top (section A) along the pensile
performance to replace some parts of gneiss was a common direction and the side (section B) especially at the corner were
method to utilize gneiss aggregate. Currently there are two strate- selected for SEM observation. Considering the specific gravity of fil-
gies: (1) Replacing the fine part of gneiss based mineral mixture by ler (SSP and LSP) was much larger than that of asphalt binder, the
high-grade fine aggregates such as alkaline minerals [14,35]. (2) filler particles in mastic coating layer may gradually gather on the
Using some new types of fillers presenting very strong antistrip- vertical direction (see Fig. 9) during the process of releasing redun-
ping ability to enhance the bonding performance of gneiss based dant mastic by hanging mastic coated coarse particles in an oven of

(a) (b) (c)

25 kv × 2,000 10 μm JSM-5610LV 25 kv × 2,000 10 μm JSM-5610LV 25 kv × 2,000 10 μm JSM-5610LV

Fig. 8. SEM figures: (a) gneiss, (b) limestone and (c) basalt.

Aggregate Slicing Mastic layer Filler particles distribution

characteristic in mastic
A
A

B B

Asphalt mastic Aggregate Gathered filler particles

Fig. 9. Observation of the filler particle distribution in mastic layer.

338 Z. Chen et al. / Construction and Building Materials 177 (2018) 332–341

(a) Mastic (b)

Aggregate
50 μm 40%SSP Section A 50 μm 40%SSP Section B

(c) (d)

Fig. 11. Water boiling test results of mastic coated coarse aggregate particles.

50 μm 30%SSP Section A 50 μm 30%SSP Section B 

9ROXPHORVVSHUFHQW
/LPHVWRQH/63PDVWLF
 6WHHOVODJ/63PDVWLF
Fig. 10. SEM figures of SSP particles in mastic: (a) Mastic from section A containing
/LPHVWRQH663PDVWLF
40% SSP by volume, (b) mastic from section B containing 40% SSP by volume, (c)
 6WHHOVODJ663PDVWLF
mastic from section A containing 30% SSP by volume, (d) mastic from section B
containing 30% SSP by volume.



45–60 °C. The distribution density of filler particles in section A
will be lower than that in section B. The gathering phenomenon 
would be much obvious if larger amount of filler was used in the
mastic. Therefore, the maximum proportion of filler used in mastic 
  
were determined based on the distribution uniformity of filler par-
:DWHUERLOLQJWLPHPLQ
ticles in mastic coating layer on the surface of coarse aggregate.
The specific gravity of SSP was also quite different from that of Fig. 12. Water boiling test results of mastic coated fine aggregate particles.
LSP, so the volume proportion of filler in mastic was used, includ-
ing 20%, 30% and 40%.
The SEM images of observation areas consist of SSP mastics and than LSP in improving the stripping resistance of asphalt mastic
gneiss coarse aggregate were shown in Fig. 10. It can be seen that under hot water damage, and the superiority were strengthened
the distribution of SSP particle in section A was obvious different when subjected to longtime hot water damage.
from that in section B when the used volume proportion of SSP The water boiling test results of mastic coated fine aggregate
was 40%, and the later was much denser (see picture a and b). It particles were displayed in Fig. 12. It can be seen that the effect
indicated that the flow of SSP particles happened during the of filler types on stripping resistance of mastic on the surface of
release of free mastic by hanging mastic coated gneiss coarse par- fine aggregate was more remarkable than that of fine aggregate
ticles. As a result, 40% of SSP by volume was not acceptable. The types. Specifically, when referring to filler types, the difference of
distribution uniformity of SSP particles in mastic was much better the values of volume loss percent between SSP mastic and LSP
when referring to 30% of SSP by volume (see picture c and d). mastic coating on limestone particles can measure up to 12% after
Meanwhile, using SSP as much as possible was also desirable in water boiling damage for 30 min. The situation for mastic coated
order to reveal the effect of filler on the moisture damage resis- SSFA particles was similar, and the difference of volume loss per-
tance of aggregate-mastic system obviously. Therefore, 30% of cent also exceeded 10%. SSP always taken the upper hand in
SSP by volume in mastic was selected. Considering the specific enhancing the stripping resistance of mastic. Converted to the side
gravity of LSP was much lower than that of SSP, the movement of fine aggregate types, the difference of volume loss percent
of LSP particles in mastic would be not stronger than that of SSP between LSP mastic coated slag particles and LSP mastic coated
particles. Hence, volume proportion of 30% also suitable for LSP. limestone particles were only about 4% even after water boiling
Hanging mastic coated particle to release the residual mastic was damage for 30 min. The difference was also no more than 4% when
not designed for fine aggregate-mastic system. Therefore, the vol- SSP mastic was used. The advantage of SSFA in improving moisture
ume proportion of used fillers will not affect the distribution uni- damage resistance of mastic was revealed under longtime hot
formity of filler particles in mastic theoretically. The same water damage.
volume proportion of fillers namely 30% was used in the investiga- Based above, the water boiling test results of mastic coated
tion of fine aggregate-mastic system for convenience. coarse particles and fine particles, respectively, suggested that
SSP and SSFA played positive roles in strengthening the ability of
3.2.1.1. Hot water damage mode. Fig. 11 gave the water boiling test mastic to resist hot water damage.
result of mastic coated gneiss coarse aggregate. It showed that the
volume loss percent of SSP mastic and LSP mastic on the surface of 3.2.1.2. Freeze-thaw damage mode. Fig. 13 showed the volume loss
gneiss coarse particles were not obviously different when short- results of mastic coated gneiss coarse particles under cyclic freeze-
term water boiling damage was applied. While the difference thaw damage. It displayed that the volume loss percent of mastic
became significant with the increase of water boiling damage time. coated particles rapidly rise with the increase of freeze-thaw dam-
LSP mastic coated gneiss particle showed larger volume loss per- age cycles. The volume loss percent of SSP mastic coated particles
cent of surface mastic. The exceeding value can be reach 12% after was a little higher than that of LSP mastic coated particles at small
water boiling damage for 30 min. It indicated that SSP did better numbers of freeze-thaw cycles, while the difference was quite
 Z. Chen et al. / Construction and Building Materials 177 (2018) 332–341 339

Fig. 15. (a) Newly crushed slag particles with two types of appearance, (b) and (c)
the failure process of mastic coated old surface.

asphalt mastic slip. Because the powder-like particles were

uncoated by mastic, they cannot come from LSP. Therefore, they
belonged to the surface of newly produced SSFA particles. Which
indicated that the coating layer corresponding to the old surface
Fig. 13. Freeze-thaw damage results of mastic coated coarse aggregate particles.
shown in Fig. 15(a) was damaged during cyclic freeze-thaw pro-
cess. Considering the free powder-like particles cannot be obvi-
ously observed in other situation, the strong levels of coating
small. Lower volume loss percent of SSP mastic was showed layer may be quite different when referring to different SSFA par-
although the volume loss of these two mastics were both higher ticles. It can explain why no consistent regularity can be obtained
than 10% after freeze-thaw damage for 20 cycles. It indicated that when investigating the effect of fine aggregate types on the freeze-
serious freeze-thaw process enabled SSP to play its anti-stripping thaw resistance of mastic coated particles. The failure process of
role better. poorly solidified coating layer was also shown in Fig. 15 (see pic-
The freeze-thaw damage results of mastic coated fine aggregate ture b and c). In simple terms, the volume change of infiltrated
particles were displayed in Fig. 14. The figure showed that the water caused expansion and cracks in coating layer. It accelerated
improvement of SSP on the ability to maintain the volume integrity the stripping of asphalt mastic during freeze-thaw damage. It was
of mastic coated fine particles was obvious. Specifically, for lime- similar to the damage process of weathering product layer coating
stone fine aggregate, the volume loss percent of SSP mastic coated on the surface of coarse steel slag particles, which has been proved
particles was always lower than that of LSP mastic coated particles in previous research [9].
when freeze-thaw cycles ranging from 5 to 20. The decrement of Concluding this section, the evaluation results suggested that
volume loss percent was about 0.7–4.0%. For SSFA, the situation SSP did better in maintaining the integrity of mastic coated fine
was similar, and the decrement of volume loss percent was about aggregate particles when facing serious freeze-thaw cycle damage.
1.2–7.5%. No consistent regularity can be observed when referring While if SSFA had some advantage than limestone fine aggregate in
to the effect of fine aggregate types on the freeze-thaw resistance increasing the stability of mastic coated particles mainly depended
of mastic coated particles. LSP coated slag particles possessed on the surface status of SSFA particles.
higher volume loss percent than LSP coated limestone particles
after serious freeze-thaw damage, while the situation for SSP 3.2.2. Compacted asphalt mixture
coated particles were just the opposite. It was due to the contribu- RMS and TSR were commonly used to evaluate the moisture
tion of inconsistent surface status of some slag particles. damage resistance of compacted asphalt mixture, which were cor-
As shown in Fig. 15(a), some newly produced particles by re- responding to hot water damage mode and freeze-thaw damage
crushing the solidified fine steel slag block possessed two types mode, respectively. Asphalt mixtures were designed according to
of morphology. It was due to the contribution of these relative Superpave procedure, and two indicators (RMS and TSR) of com-
large-size particles (4.75 mm), and which were seriously coated pacted asphalt mixtures were measured in this section.
by fine powder in original fine steel slag. It must result in the gen-
eration of some small-size particles (2.36 mm) partly containing 3.2.2.1. Design of asphalt mixture. Mixing some kinds of aggregate
old surface (fine powder coating layer) when these large-size par- with different size ranges according to a certain mass proportion
ticles were re-crushed. Fig. 14 showed that the volume loss percent is a normal practice during the design of mineral mixture. While
of LSP mastic coated slag particles were very big after 10 cycles or it may result in error when making crosswise comparison among
20 cycles of freeze-thaw damage. It can observe small amount of asphalt mixtures composed of different aggregate types. Taking
free powder-like particles on the fracture surface of stripped GSS and GLS shown in Table 3 for example. The only difference
between GSS and GLS is the fine aggregate. The specific gravity of
 SSFA was 24.6% larger than that of limestone fine aggregate. It
9ROXPHORVVSHUFHQWRIPDVWLF

/LPHVWRQH/63PDVWLF
meant that the volume compositions of GSS and GLS will be differ-
6WHHOVODJ/63PDVWLF
FRDWHGSDUWLFOHV

/LPHVWRQH663PDVWLF
ent if the same mass proportions are used.

6WHHOVODJ663PDVWLF Therefore, the coarse aggregate, fine aggregate and filler were
blended according to their volume proportions in this research.
 In order to keep the volume composition of each asphalt mixture
almost the same, the particle gradation of limestone fine aggregate
was adjusted to be consistent with that of SSFA considering the
 used dosage of fine aggregate is large. Considering some newly
produced SSFA particles partly included the old surface (seriously
coated by fine powder), and the strong levels of coating powder

   layer were changeable when referring to different SSFA particles.
)UHH]HWKDZGDPDJHF\FOHV A more viscous SBS modified asphalt binder was used to enhance
the bonding performance of asphalt mixture. Also, the SBS modi-
Fig. 14. Freeze-thaw damage results of mastic coated fine aggregate particles. fied asphalt were often used in the surface layer construction of
340 Z. Chen et al. / Construction and Building Materials 177 (2018) 332–341

entity asphalt pavement engineering with nominal maximum size the mechanism performance of gneiss based asphalt mixture well.
of 12.5 or 13.2 mm. The TSR value of GSS asphalt mixture exceeded 80% even after
freeze-thaw damage for 2 cycles, and 80% of TSR is the minimum
3.2.2.2. RMS and TSR. The original average Marshall stabilities of value specified in Chinese standard after one cycle freeze-thaw
GLL, GSL, GLS and GSS were 11.6, 13.5, 12.0 and 13.1 kN, respectively. damage [29]. Which meant that the combination of SSFA and SSP
All the values met the requirement of Chinese specification (mini- also gave the best result. It also suggested that the use of SBS mod-
mum value = 8 kN), and SSFA showed much more stable and posi- ified asphalt binder can offset the adverse impact of some SSFA
tive effect on the Marshall stability of asphalt mixtures. The particles partly containing old surface on the freeze-thaw damage
computed RMS values of the four asphalt mixtures were shown resistance of asphalt mixtures.
in Fig. 16. It clearly displayed that the decay speed of the Marshall
stability of GLL asphalt mixture was the fastest, then GSL and GLS, 4. Conclusions
and GSS was the last. Which meant that the SSFA and SSP both
showed good ability for enhancing the resistance of hot water The feasibility of simultaneously using coarse gneiss and fine
damage on the mechanism performance of asphalt mixture, and steel slag (fine aggregate and filler) in asphalt mixture were dis-
the combination of SSFA and SSP performed best among them. cussed in this research. The inadequacy of using gneiss and fine
The RMS value of modified asphalt mixture should not be less than steel slag were concluded first, and then steel slag fine aggregate
85% after immersing in 60° hot bath for 48 h according to Chinese (SSFA) and steel slag powder filler (SSP) were prepared by sec-
specification [29]. It can be seen from Fig. 16 that the RMS of GSS ondary crushing of the hardened fine slag block, and finally the
asphalt mixture was still higher than 85% even after hot water moisture sensitivity of uncompacted aggregate-mastic system
damage for 72 h. and compacted asphalt mixture under hot water damage and
The ITS and TSR change rules of the four asphalt mixtures were freeze-thaw damage were fully investigated, respectively. Based
similar to that of Marshall stability and RMS. The original average on the results discussed above, the following items can be
ITS of GLL, GSL, GLS and GSS were 1.45, 1.57, 1.48 and 1.67 MPa, concluded:
respectively. SSFA also played more positive role in increasing
the strength of asphalt mixture than SSP. It may be related to the (1) The silicate minerals should be responsible for the perfor-
large amount of fine powder in SSFA. Fig. 17 gave the TSR values mance deterioration of fine steel slag, and the nature acidity
of asphalt mixtures, it also showed that GLL performed worst in and layered micro-structure of gneiss rock make it unsuit-
resisting the freeze-thaw damage, GSS worked best, and GLS, GSS able for directly using in asphalt mixture theoretically.
were somewhere between them. It indicated that replacing lime- (2) SSP and SSFA played more positive roles in strengthening
stone fine aggregate by SSFA or replacing LSP by SSP can retain the ability of mastic to resist longtime hot water damage
compared to limestone powder (LSP) and fine aggregate,
respectively. SSP also did better in maintaining the integrity
5HWDLQHG0DUVKDOOVWDELOLW\506

 of mastic coated particles when facing serious freeze-thaw

*// *6/ */6 *66 cycle damage. While the contribution of SSFA was mainly
 Minimum value depended on its surface status.
me=48 h (3) RMS and TSR results suggested that SSFA and SSP both can
improve the ability of compacted SBS modified asphalt mix-

ture to resist the hot water damage and freeze-thaw dam-
age, and their combination performed even better than one
 using alone. Therefore, the SBS asphalt binder can offset
the adverse impact of some SSFA particles with poor surface
 status on the freeze-thaw damage of asphalt mixtures.

 Conflict of interest
  
:DWHUGDPDJHWLPHK
The authors have declared that no conflict of interest exists.
Fig. 16. RMS results of various compacted asphalt mixtures.
Acknowledgments

 The Project was Supported by the Fundamental Research Funds

for the Central Universities, China University of Geosciences,
7HQVLOHVWUHQJWKUDWLR765

*// *6/ */6 *66

 Wuhan (Nos. CUG170685 & CUG170682) and the Natural Science
Minimum value Foundation of China (Nos. 41731284 & 11672360).
cycle number=1

References

[1] Y. Huang, R.N. Bird, O. Heidrich, A review of the use of recycled solid waste
 materials in asphalt pavements, Resour. Conserv. Recycl. 52 (1) (2007) 58–73.
[2] D. Merrill, A.V. Dommelen, L. Gáspár, A review of practical experience
throughout Europe on deterioration in fully-flexible and semi-rigid long-life
 pavements, Int. J. Pavement Eng. 7 (2) (2006) 101–109.
[3] J. Zheng, Design guide for semirigid pavements in China based on critical state
 of asphalt mixture, J. Mater. Civ. Eng. 25 (7) (2013) 899–906.
   [4] Q. Chen, H. Zhu, Z. Yan, J.W. Ju, Z. Jiang, Y. Wang, A multiphase
micromechanical model for hybrid fiber reinforced concrete considering the
)UHH]HWKDZGDPDJHF\FOHV aggregate and ITZ effects, Constr. Build. Mater. 114 (2016) 839–850.
[5] Z. Yan, Q. Chen, H. Zhu, J.W. Ju, S. Zhou, Z. Jiang, A multi-phase
Fig. 17. TSR results of various compacted asphalt mixtures. micromechanical model for unsaturated concrete repaired using the
 Z. Chen et al. / Construction and Building Materials 177 (2018) 332–341 341

electrochemical deposition method, Int. J. Solids Struct. 50 (24) (2013) 3875– [20] Y. Xiao, M.F.C.V.D. Ven, A.A.A. Molenaar, S.P. Wu, Possibility of using epoxy
3885. modified bitumen to replace tar-containing binder for pavement antiskid
[6] P. Ahmedzade, B. Sengoz, Evaluation of steel slag coarse aggregate in hot mix surfaces, Constr. Build. Mater. 48 (19) (2013) 59–66.
asphalt concrete, J. Hazard. Mater. 165 (1) (2009) 300–305. [21] F. Xiao, S. Amirkhanian, C.H. Juang, Rutting resistance of rubberized asphalt
[7] Z. Chen, S. Wu, Y. Xiao, M. Zhao, J. Xie, Feasibility study of BOF slag containing concrete pavements containing reclaimed asphalt pavement mixtures, J.
honeycomb particles in asphalt mixture, Constr. Build. Mater. 124 (2016) 550– Mater. Civ. Eng. 19 (6) (2007) 475–483.
557. [22] F. Xiao, S.N. Amirkhanian, J. Shen, B. Putman, Influences of crumb rubber size
[8] Ministry of Transport of the People’s Republic of China. The annual statistics of and type on reclaimed asphalt pavement (RAP) mixtures, Constr. Build. Mater.
traffic and transportation, Available on line: <http://zizhan.mot.gov.cn/zfxxgk/ 23 (2) (2009) 1028–1034.
bnssj/zhghs/201803/t20180329_3005087.html>. Cited on April 20, 2018. [23] F. Xiao, P.E.W. Zhao, S.N. Amirkhanian, Fatigue behavior of rubberized asphalt
[9] Z. Chen, S. Wu, Y. Xiao, W. Zeng, M. Yi, J. Wan, Effect of hydration and silicone concrete mixtures containing warm asphalt additives, Constr. Build. Mater. 23
resin on Basic Oxygen Furnace slag and its asphalt mixture, J. Cleaner Prod. 112 (10) (2009) 3144–3151.
(2016) 392–400. [24] P. Cui, S. Wu, Y. Xiao, M. Wan, P. Cui, Inhibiting effect of Layered Double
[10] J. Zhu, S. Wu, J. Zhong, D. Wang, Investigation of asphalt mixture containing Hydroxides on the emissions of volatile organic compounds from bituminous
demolition waste obtained from earthquake-damaged buildings, Constr. Build. materials, J. Cleaner Prod. 108 (12) (2015) 987–991.
Mater. 29 (2012) 466–475. [25] D. Shen, C. Wu, J. Du, Laboratory investigation of basic oxygen furnace slag for
[11] Z. Chen, Y. Xiao, L. Pang, W. Zeng, S. Wu, Experimental assessment of flue gas substitution of aggregate in porous asphalt mixture, Constr. Build. Mater. 23
desulfurization residues and basic oxygen furnace slag on fatigue and (1) (2009) 453–461.
moisture resistance of HMA, Fatigue Fract. Eng. Mater. Struct. 37 (11) (2014) [26] Z. Chen, S. Wu, L. Pang, J. Xie, Function investigation of stone mastic asphalt
1242–1253. (SMA) mixture partly containing basic oxygen furnace (BOF) slag, J. Appl.
[12] Y. Xue, S. Wu, H. Hou, J. Zha, Experimental investigation of basic oxygen Biomater. Funct. Mater. 14 (2016) S68–S72.
furnace slag used as aggregate in asphalt mixture, J. Hazard. Mater. 138 (2) [27] H. Ziari, S. Nowbakht, S. Rezaei, A. Mahboob, Laboratory investigation of
(2006) 261–268. fatigue characteristics of asphalt mixtures with steel slag aggregates, Adv.
[13] J. Xie, S. Wu, J. Lin, J. Cai, Z. Chen, W. Wei, Recycling of basic oxygen furnace Mater. Sci. Eng. 2015 (1583) (2015) 1–5.
slag in asphalt mixture: material characterization & moisture damage [28] Ministry of Transport of the People’s Republic of China. Test methods of
investigation, Constr. Build. Mater. 36 (2012) 467–474. aggregate for highway engineering. JTG E42, 2005 (in Chinese).
[14] Z. Chen, S. Wu, J. Wen, M. Zhao, M. Yi, J. Wan, Utilization of gneiss coarse [29] Ministry of Transport of the People’s Republic of China. Technical
aggregate and steel slag fine aggregate in asphalt mixture, Constr. Build. specifications for construction highway asphalt pavements. JTG F40, 2004
Mater. 93 (2015) 911–918. (in Chinese).
[15] S. Wu, Y. Xue, Q. Ye, Y. Chen, Utilization of steel slag as aggregates for stone [30] Ministry of Transport of the People’s Republic of China. Standard test methods
mastic asphalt (SMA) mixtures, Build. Environ. 42 (7) (2007) 2580–2585. of bitumen and bituminous mixtures for highway engineering. JTG E20, 2011
[16] J. Xie, S. Wu, L. Zhang, Y. Xiao, W. Ding, Evaluation the deleterious potential (in Chinese).
and heating characteristics of basic oxygen furnace slag based on laboratory [31] Z. Chen, J. Xie, Y. Xiao, J. Chen, S. Wu, Characteristics of bonding behavior
and in-place investigation during large-scale reutilization, J. Cleaner Prod. 133 between basic oxygen furnace slag and asphalt binder, Constr. Build. Mater. 64
(2016) 78–87. (2014) 60–66.
[17] Z. Chen, S. Wu, F. Li, J. Chen, Z. Qin, L. Pang, Recycling of Flue Gas [32] C. Shi, Steel slag-its production, processing, characteristics, and cementitious
Desulfurization residues in gneiss based hot mix asphalt: materials properties, J. Mater. Civ. Eng. 16 (3) (2004) 230–236.
characterization and performances evaluation, Constr. Build. Mater. 73 [33] M. Zhao, S. Wu, Z. Chen, C. Li, Production and application of steel slag coarse
(2014) 137–144. aggregate in asphalt mixture, Emerg. Mater. Res. 6 (1) (2017) 219–222.
[18] P. Pan, S. Wu, Y. Xiao, G. Liu, A review on hydronic asphalt pavement for energy [34] D.R. Groome, The geochemistry of the devonian lavas of the Northern Lorne
harvesting and snow melting, Renewable Sustainable Energy Rev. 48 (2015) plateau, Scotland, Mineral. Mag. 39 (306) (1974) 621–640.
624–634. [35] S. Wu, D. Hu, L. Pang, H. Wang, Utilization of gneiss in asphalt concrete
[19] Y. Xiao, M.F.C.V.D. Ven, A.A.A. Molenaar, Z. Su, K. Chang, Design approach for mixtures, Mater. Sci. Forum 620–622 (1) (2009) 1–4.
epoxy modified bitumen to be used in antiskid surfaces on asphalt pavement,
Constr. Build. Mater. 41 (41) (2013) 516–525.