Thermal Conductivity Evaluation and Road Performance Test of Steel Slag Asphalt Mixture

Abstract

Substituting steel slag for mineral materials in road construction has potential economic and environmental benefits. Due to the excellent thermal conductivity of steel slag, it is often used in functional pavements. However, there are few studies on the thermal conductivity characterization of steel slag asphalt mixture (SSAM). For this reason, the thermal conductivity of SSAM was first qualitatively evaluated by microscopic characterizations. The thermal conductivity was the quantitatively evaluated by the heating wire method. Theoretical calculations were used to verify the reliability of the quantitative characterization. Finally, the effects of steel slag on the volume indices and the road performance of SSAM were studied. Results showed that active minerals such as iron oxides make the steel slag thermally conductive, while a large number of protrusions and micropores on the surface of the steel slag may be detrimental to thermal conductivity. The thermal conductivity first increases and then decreases with the steel slag content. The asphalt mixture with 60% steel slag replacing aggregate of 3–5 mm (6.6% of the mixture) had the highest thermal coefficient of 1.746 W/(m·°C), which is only 4.78% different from the theoretical value. The porosity and water absorption of SSAM gradually increased with the content of steel slag. The road performance test indicated that steel slag increased the high-temperature performance of the asphalt mixture to a certain extent, but weakened the low-temperature performance and moisture resistance. After comprehensive consideration of the thermal conductivity and road performance, it is recommended that the optimum content of steel slag is not more than 60%.

1. Introduction

Clear waters and lush mountains are invaluable assets. Reducing the exploitation of natural resources is of great significance to global environmental protection. Transportation is a resource-intensive industry, in which the construction of transportation infrastructure consumes a lot of mineral resources. Seeking effective substitutes for ore resources has great potential for reducing mine development and promoting the sustainable development of society [1,2,3,4]. Steel slag is a by-product of the steelmaking process. It is a lumpy substance formed by the combination of residual flux and metal oxide in the steelmaking furnace after cooling, iron removal, and crushing. About 0.15 t of steel slag is produced for every 1 t of steel produced [5]. China is the world’s largest steel producer, with annual steel production accounting for about 50% of global steel production capacity. In the past two years, the average annual discharge of steel slag is 100 million tons. The major industrial countries in the world have adopted the comprehensive utilization of steel slag as a strategy for green development [6]. Steel slag has the advantages of high strength, wear resistance, slip resistance, high basicity, and low cost, and has the potential for large-scale utilization in pavements [7,8].

Some research results have been obtained on the application of steel slag in road engineering, mainly involving two aspects: conventional performance and functional performance. Conventional properties are stiffness indicators of SSAMs, such as fatigue properties, expansion properties, aging properties, and other mechanically related properties. For example, Kavussi [9] studied the fatigue behavior of asphalt mixtures with electric arc furnace (EAF) steel slag as coarse aggregate in aged and unaged states. The four-point bending fatigue test showed that although the addition of EAF steel slag into the asphalt mixture improved the fatigue life of the sample, it had no obvious effect on the asphalt mixture in the aging state. To further improve the fatigue performance of the SSAM, Cao [10] modified the surface of the steel slag with cement paste and studied the fatigue life of the SSAM under soaking conditions. The results showed that the fatigue life of the modified SSAM was 23% higher than that of the original SSAM. The dense layer formed on the surface of the steel slag by the cement hydration products prevents the water from invading the steel slag, thereby improving the fatigue performance. Although the physical properties of steel slag meet the application requirements of asphalt mixtures, due to its angular shape, asphalt mixtures using steel slag as aggregates are prone to expansion and are affected by void problems. Mineral aggregates cannot be completely replaced with steel slag [11]. To reduce the volume expansion characteristics of steel slag and make full use of cold abandoned steel slag, Zhang [12] used limestone fine aggregate instead of steel slag fine aggregate to reduce the water swelling of cold abandoned SSAM. Masoudi [13,14] studied the long-term aging performance of the EAF steel slag warm mix asphalt mixture. Steel slag warm mix asphalt has better mechanical properties than limestone hot mix asphalt. The use of steel slag increases the aging index of the asphalt mixture and reduces the anti-aging ability of the asphalt mixture. In addition, the replacement of limestone aggregate with steel slag increases the optimal asphalt aggregate ratio by 0.45%. Zeng [15] studied the nonlinear viscoelastic–plastic deformation characteristics of SSAMs and proposed a new integral viscoelastic–plastic constitutive model combining the Schapery model and the improved Schwartz model. The model can more accurately reflect the elastic, viscoelastic, and viscoplastic deformation of SSAM in the creep process. Ameri et al. [16] studied the performance of steel slag in hot mix asphalt and warm mix asphalt. They showed that the hot mix asphalt mixture with steel slag instead of fine aggregate had the lowest elastic modulus and indirect tensile strength. However, using steel slag as coarse aggregate in warm mix asphalt improved the elastic modulus and indirect tensile strength of the mixture. Steel slag coarse aggregate is recommended for warm mix asphalt. To make full use of the coarse and fine aggregates of slag, Chen [17] discussed the comprehensive utilization technology of steel slag and iron slag, and used steel slag coarse aggregate, iron slag fine aggregate, and steel slag powder to replace 100% limestone. Some performance indicators of the asphalt mixture prepared by combining high-viscosity asphalt and the secondary mixing process are significantly improved. The high-temperature flow times and low-temperature fracture energy were increased by 18% and 23%, respectively. There are many methods for evaluating the rutting resistance of asphalt mixtures, such as rutting test, dynamic modulus test, creep test, etc. [18]. In addition, the rotary compaction process can also reflect the rutting resistance, more compaction times, and strong anti-rutting performance [19]. Huang [20] studied the compaction characteristics of steel slag porous asphalt mixture, and the results of rotary compaction showed that the number of compactions of porous asphalt mixture without steel slag was higher than that of steel slag porous asphalt mixture. It was indicated that the porous asphalt mixture without steel slag has a higher anti-rutting performance. However, Martinho’s [21] research on the high-temperature performance of warm mix SSAM formed the opposite conclusion. In general, steel slag can improve the high-temperature performance of asphalt mixture. Liu [22] evaluated the feasibility of the porous asphalt mixture containing steel slag in the seasonal freezing area, and the results showed that the addition of steel slag can improve the pavement performance of the porous asphalt mixture. When the coarse aggregates are steel slag, the overall performance of the porous asphalt mixture is optimal. Crisman [23] compared the mechanical properties of rubber–asphalt mixture containing steel slag and ordinary asphalt mixture and found that the elastic modulus of steel slag rubber–asphalt mixture was higher than that of ordinary asphalt mixture at high temperature, but lower at low temperature. The permanent deformation resistance of steel slag rubber–asphalt mixture has been improved.

In terms of the functional utilization of steel slag, Li [24] studied the effect of steel slag on the electrical conductivity of graphite–carbon fiber composite conductive asphalt concrete and found that 100% steel slag content increased the conductivity of the mixture to 0.125 S/m. The contribution of the three materials to the conductivity followed the order carbon fiber, graphite, and steel slag (from highest to lowest). Xiang [25] studied the self-healing performance of SSAM under microwave heating conditions. The addition of steel slag can improve the healing effect of asphalt mixtures. The asphalt mixture of coarse stone and fine steel slag has a more uniform temperature distribution, and the healing effect is 1.11 times higher than that of asphalt mixtures without steel slag. Gao [26] mixed steel slag into asphalt mixture for microwave deicing, measured the temperature distribution and ice melting effect of SSAM, and verified the feasibility of ice melting of SSAM.

To sum up, the application of steel slag in road engineering mainly focuses on the basic properties of SSAM, such as fatigue, expansion, aging, and mechanical properties. This is because the primary task of SSAM is to meet the requirements of engineering construction. There is still insufficient research on the functional aspects of SSAM. Whether it is electrical conductivity, self-healing performance, or ice-melting performance, these properties are all related, to some extent, to the thermal conductivity of SSAM. Existing research rarely considers the basic indicators of its thermal conductivity. Further research into the thermal conductivity of SSAM is required.

Consequently, the goal of this paper is to study the thermal conductivity of SSAM, as well as the road performance of SSAMs with different thermal conductivity. It is expected that SSAM can meet the requirements of road performance while exerting thermal conductivity. Therefore, the thermal conductivity of steel slag was first qualitatively evaluated from a microscopic perspective. Second, the SSAM was prepared according to the principle of equal volume replacement and the heating test was conducted. The thermal conductivity of SSAM was quantitatively evaluated with the thermal coefficient as an index. Afterward, to prove the accuracy of the measured data, the thermal coefficient was estimated by the theoretical calculation method. Finally, the volume indices of the asphalt mixtures with different thermal conductivity were tested, and the road performances of the SSAM was verified.

2. Materials and Methods

The raw materials used to prepare the thermally conductive asphalt mixture mainly include asphalt, fillers, aggregates, and thermally conductive phase steel slag. The technical indicators of the raw materials are tested following the “Standard Test Methods of Bitumen and Bituminous Mixtures for Highway Engineering” (JTG E20-2011) [27]. The index standard comes from “Technical Specifications for Construction of Highway Asphalt Pavements” (JTG F40-2004) [28].

2.1. Materials

2.1.1. Asphalt

The test asphalt is SBS I-C modified asphalt; its technical indicators are shown in

Table 1. Technical indicators of SBS I-C modified asphalt.

Technical Indicators		Unit	Standard	Test Results	Test Method
Penetration (25 °C, 100 g, 5 s)		0.1 mm	60–80	64.7	T0604-2011
Penetration index		-	≥−0.4	0.4	T0604-2011
Ductility (5 °C, cm/min)		cm	≥30	37.8	T0605-2011
Softening point (Ring-and-ball method)		°C	≥55	63.5	T0606-2011
Density		g/cm	-	1.014	T0603-2011
Dynamic viscosity (135 °C)		Pa·s	≤3	1.870	T0619-2011
Flash point (Cleveland Open Cup)		°C	≥230	260	T0611-2011
Solubility (Trichloroethylene)		%	≥99	99.7	T0607-2011
Elastic recovery (25 °C, 10 cm)		%	≥65	80	T0662-2000
Properties after rolling thin film oven test	Quality change	%	≤±1.0	−0.19	T0609-2011
	Penetration ratio (25 °C)	%	≥60	82.9	T0604-2011
	Ductility (5 °C)	cm	≥20	26.3	T0605-2011
	Softening Point (Ring-and-ball method)	°C	≥55	71.8	T0606-2011

.

2.1.2. Aggregates and Filler

The aggregates used in tests are mainly limestone, with a total of 4 fractions of particle sizes. The filler is alkaline ground limestone. The technical indicators of aggregates and filler are shown in

Table 2. Aggregate technical indicators.

Technical Indicators	Particle Size/mm				Standard	Test Method
	10–15	5–10	3–5	0–3
Apparent relative density (g/cm3)	2.716	2.727	2.745	2.695	≥2.5	T0304-2005
Gross volume relative density (g/cm3)	2.684	2.685	2.658	2.647	-	T0304-2005
Crush value (%)	22.6	-	-	-	≤26	T0316-2005
Los Angeles abrasion value (%)	15.4	18.5	24.8	-	≤28	T0317-2005
Water absorption (%)	1.2	0.6	0.4	0.4	≤2	T0304-2005
Needle flake content (%)	10.69	8.01	-	-	≤12	T0312-2005

and

Table 3. Performance index of filler.

Technical Indicators	Test Results	Standard	Test Method
Apparent relative density (g/cm3)	≥2.5	2.716	T0352-2000
Hydrophilic coefficient	<1	0.76	T0353-2000
Stability	No discoloration when heated	No color change	T0355-2000

, respectively.

2.1.3. Steel Slag

The steel slags in the research are basic oxygen furnace (BOF) steel slags. The steel slag replaces the aggregate in the asphalt mixture, which can form a coherent heat conduction path in the mixture, speed up the heat transfer rate in the mixture, and improve the thermal conductivity of the mixture. The performance indicators are shown in

Table 4. Performance index of steel slag.

Technical Indicators		Apparent Relative Density (g/cm3)	Gross Volume Relative Density (g/cm3)	Water Absorption (%)	Water Swelling Rate (%)	Crush Value (%)	Los Angeles Abrasion Value (%)
Particle size/mm	0–3	2.875	2.713	1.0	0.6	-	-
	3–5	3.726	3.511	1.6	1.1	-	15.4
	5–10	3.656	3.449	2.3	1.2	11.2	14.2
	10–30	3.445	3.247	2.6	0.9	13.9	20.1
Standard		≥2.5	-	≤3	≤2	≤22	≤26
Test method		T0304-2005	T0304-2005	T0304-2005	T0348-2005	T0316-2005	T0317-2005

. The relative density of the steel slag of 0–3 mm in Table 4 is low, which may be because the fine aggregate of the steel slag is not clean.

2.2. Microscopic Characterization

2.2.1. X-ray Diffraction (XRD) and X-ray Fluorescence (XRF) Tests

As the main by-product of the steel smelting industry, steel slag has a complex mineral composition and a wide variety of elements. Different mineral elements have different activities and different thermal conductivity. To visually distinguish the difference in mineral composition between the steel slag and limestone, XRD and XRF tests were performed to provide a theoretical basis for the thermal conductivity of the steel slag.

2.2.2. Scanning Electron Microscope (SEM) Test

Due to the formation process of steel slag, its structure is not dense. On the one hand, the porous structure affects the water absorption of the material and the amount of asphalt, as well as the thermal conductivity of the steel slag. To further clarify its surface structure, the microscopic morphology of the steel slag surface was characterized by SEM.

2.3. Design of SSAM

2.3.1. Target Grading

The principle of mineral grading design is to meet the road performance; under this premise, the material composition can be changed to meet the requirements of thermal conductivity. A good asphalt mixture gradation should ensure a reasonable proportion of coarse and fine aggregates and can form a good skeleton structure after mixing with asphalt to generate sufficient strength to resist external deformation. Since the mix proportion design of the SSAM is based on the target gradation of raw materials, the selection of the target gradation and the performance verification are the preconditions for the mix proportion design of the SSAM. According to the results of sieving of each fraction of aggregate in

Table 5. Mineral material and target gradation.

Mesh Size (mm)	16	13.2	9.5	4.75	2.36	1.18	0.6	0.3	0.15	0.075	Content (%)
Filler	100	93.3	25.3	0.6	0.4	0.4	0.4	0.4	0.4	0.4	6
0–3	100	100	92.6	2.7	0.4	0.4	0.3	0.3	0.3	0.2	31
3–5	100	100	100	79	2.6	1.2	0.9	0.8	0.7	0.6	11
5–10	100	100	100	100	86.3	62.7	42.7	20.9	11.4	8.5	19
10–15	100	100	100	100	100	100	100	100	95.9	55.3	33
Upper grading limit (%)	100	100	85	68	50	38	28	20	15	8	/
Median value of grading (%)	100	95	76.5	53	37	26.5	19	13.5	10	6	/
Lower grading limit (%)	100	90	68	38	24	15	10	7	5	4	/
Target grading (%)	100.0	97.8	73.9	46.4	33.2	25.8	19.5	12.8	9.6	6.2	/

, an AC-13 (asphalt mixture with a nominal maximum aggregate size of 13 mm) gradation curve close to the center line of gradation was preliminarily selected. The dosages of filler were 0–3, 3–5, 5–10, and 10–15 mm, with aggregate of 6, 31, 11, 19, and 33%, respectively. The Marshall specimen indices of the AC-13 asphalt mixture are shown in

Table 6. Marshall specimen indices of AC-13 under target grading.

Asphalt Aggregate Ratio (%)	Gross Bulk Density (g/cm3)	Void Ratio (%)	VMA (%)	VFA (%)	Stability (kN)	Flow (mm)
3.5	2.375	9.2	16.9	42.2	10.01	2.67
4.0	2.398	6.8	15.7	53.5	10.20	3.02
4.5	2.413	5.2	15.0	63.4	10.97	3.39
5.0	2.425	3.7	14.6	73.2	11.73	3.78
5.5	2.412	3.4	15.1	76.7	9.82	4.52
Test method	T0705-2011	T0705-2011	T0705-2011	T0705-2011	T0709-2011	T0709-2011

; the calculated optimal asphalt aggregate ratio is 4.9%. The performance of the AC-13 asphalt mixture with the optimal asphalt aggregate ratio is shown in

Table 7. Performance of AC-13 with the optimal asphalt aggregate ratio.

Performance	Dynamic Stability (Times/mm)	Flexural Tensile Strain (με)	Residual Stability (%)	Freeze–Thaw Splitting Strength Ratio (%)
Test result	3947	5109	96.7	91.5
Standard	≥2800	≥2500	≥80	≥75
Test method	T0719-2011	T0715-2011	T0709-2011	T0729-2011

. The test results meet the requirements of the specification [28].

2.3.2. Steel Slag Particle Size and Replacement Principle

• Steel slag size selection

The optional 0–3, 3–5, and 5–10 mm steel slags (numbered S-1, S-2, and S-3, respectively) were screened. The sieving results were compared with the sieving results of various fractions of limestone aggregates and used as the basis for the selection of steel slag particle size. The sieving curves are shown in Figure 1. The pass rates of S-1 and S-2 are similar to the fractions of 0–3 and 3–5 mm, respectively, while the pass rate of S-3 is quite different from that of limestone of 5–10 mm. Therefore, it may be easier to use S-1 to replace 0–3 mm limestone aggregate or S-2 to replace 3–5 mm limestone aggregate to prepare SSAM with a gradation close to the target gradation. As S-1 may be impure, S-2 was used to replace the 3–5 mm aggregate. The replacement ratios are 20%, 40%, 60%, 80%, and 100%. The aggregate proportions of different SSAMs are shown in

Table 8. Material proportions of SSAM (volume fraction).

Replacement Rate	The Proportion of Various Materials (%)
	Filler	0–3 mm	3–5 mm	S-2 (Steel Slag)	5–10 mm	10–15 mm
20%	6	31	8.8	2.2	19	33
40%	6	31	6.6	4.4	19	33
60%	6	31	4.4	6.6	19	33
80%	6	31	2.2	8.8	19	33
100%	6	31	0	11	19	33

, and the final gradation curves are shown in Figure 2.

2.

Replacement principle

The density and water absorption of steel slag are 20% higher than that of natural aggregates, so the design method of mixtures dominated by quality control is not suitable for the gradation design of SSAM. In this study, based on the principle of equal volume replacement, the design of SSAM was conducted by the method of density conversion [29]. After the steel slag replaces the aggregate, the aggregate volume that passes through each sieve hole remains the same. The volume of steel slag and the mass of steel slag were converted by using the relative density of the gross volume of aggregate as a medium. Density conversion is performed according to Equation (1).

$$P_s=P_i·\eta ·\frac{{\rho }_s}{{\rho }_i}$$

(1)

where P_s is the quality of steel slag to be used, P_i is the total weight of a certain type of limestone aggregate to be replaced, η is the volume percentage of aggregate replaced by steel slag, ρ_s is the gross volume relative density of steel slag, and ρ_i is the gross volume relative density of the limestone aggregate to be replaced.

2.4. Determination of Thermal Coefficient of SSAMs

The thermal coefficient of the steel slag mixture was determined concerning the standard “Refractory materials—Determination of thermal conductivity—Hot-wire method” (GB/T5990-2006) [30]. The schematic diagram of the principle of measuring the thermal coefficient by the heating wire method is shown in Figure 3a. The specimens used for the determination of thermal coefficient were small Marshall specimens formed by standard compaction methods. The size of the specimen is φ 101.6 × 63.5 mm³. The treated specimen is shown in Figure 3b. The H hole is located at the center of the upper surface of the Marshall specimen. The hole diameter is 6 mm and the depth is 50 mm for embedding the heating wire. Holes P1, P2, and P3 are used to embed temperature sensors. The three holes are arranged equidistant from the center hole H, and the spacing meets the requirements of 20 ± 5 mm. After the hole is drilled, the debris inside the specimen should be cleaned. Subsequently, the embedding of the heating wire and the temperature sensor is performed, as shown in Figure 3c. After the heating wire and the temperature sensor are inserted into the corresponding holes, they are sealed with thermally conductive silicone. The heating wire used is a Ni–Cr alloy with a diameter of 6 mm, a length of 50 mm, and a heating power of 60 W. It has the characteristics of fast heating and good heat resistance. The temperature sensor is a PT100 platinum resistance thermometer with a diameter of 4 mm and a length of 30 mm. The working temperature of the temperature sensor is −50–200 °C, and the temperature error is 0.1 °C.

Figure 3d shows the framework of the thermal coefficient test system, of which the main parts are a computer, a data acquisition instrument, a heating wire, and a temperature sensor. When the ambient temperature fluctuation during the test is lower than 0.3 °C within 2 min, the power supply is started to heat the heating wire. The data acquisition instrument records data every 5 s. In the processes of power-on and heating, a polystyrene foam board is used for thermal insulation of the specimen. In this way, the heat exchange between the side surface of the specimen and the outside air is prevented, thereby reducing the test error. The temperature rise curve of the asphalt mixture under each steel slag content can be drawn, and the thermal coefficient of the SSAM can be obtained by substituting the temperature data into Equation (2).

$$\mathsf{\lambda }=\frac{\mathrm{UI}}{4\mathsf{\pi }\mathrm{L}}· \frac{-{\mathrm{E}}_{\mathrm{i}}(\frac{-{\mathrm{r}}^2}{4\mathsf{\alpha }\mathrm{t}})}{∆{\mathsf{\theta }}_2-∆{\mathsf{\theta }}_1}$$

(2)

where λ is the thermal coefficient of the specimen (W/(m·°C)), U is the voltage of the heating wire (V), I is the current of the heating wire (A), L is the length of the heating wire embedded in the specimen (m), r is the distance between the thermocouple and the heating wire (m), ɑ is the thermal diffusivity (m²/s), t₁ and t₂ are the measurement time in the stable state of temperature rise during the measurement process (s), ∆θ₁ and ∆θ₂ are the temperature rise at the measurement time t₁, t₂ (°C), and $-{\mathrm{E}}_{\mathrm{i}}(\frac{-{\mathrm{r}}^2}{4\mathsf{\alpha }\mathrm{t}})$ is the exponential integral of ${{\displaystyle \int }}_{\mathrm{x}}^{\mathrm{u}}\frac{-{\mathrm{e}}^{\mathrm{u}}\mathrm{du}}{\mathrm{u}}$, and its value can be found through the ratio table of $\frac{∆{\mathsf{\theta }}_2}{∆{\mathsf{\theta }}_1}$ [30].

2.5. Volume Indices of SSAM

Due to the porosity of steel slag, the diversity of chemical elements, and the complexity of composition, the content of steel slag affects the performance of asphalt mixtures. Six substitution levels were designed to replace 0, 20, 40, 60, 80, and 100% of aggregates 3–5 by at equal volume. Under the condition that the optimal amount of asphalt remains unchanged, Marshall specimens of SSAMs were prepared.

Due to the significant density difference between steel slag and natural aggregate of the same size, replacing limestone with an equal volume of steel slag increases the quality of the asphalt mixture. The asphalt aggregate ratio is expressed as the percentage of asphalt mass to mineral aggregate. Due to the addition of steel slag, the original asphalt aggregate ratio has changed under the condition of the same amount of asphalt. Therefore, it is necessary to redefine the asphalt aggregate ratio after adding steel slag. In this paper, the asphalt aggregate ratio in the SSAM is defined as the nominal asphalt aggregate ratio, which reflects the actual amount of asphalt in the mixture. The asphalt aggregate ratio, the nominal asphalt aggregate ratio, and the conversion relationship are calculated according to Equations (3)–(5) respectively.

$${\mathrm{P}}_{\mathrm{a}}=\frac{{\mathrm{m}}_{\mathrm{a}}}{\mathrm{M}}\times 100\%$$

(3)

$${\mathrm{P}}_{\mathrm{a}}^{\prime }=\frac{{\mathrm{m}}_{\mathrm{a}}}{\mathrm{M}+{\mathrm{m}}_{\mathrm{i}}·\mathsf{\eta }·(\frac{{\mathsf{\rho }}_{\mathrm{s}}}{{\mathsf{\rho }}_{\mathrm{i}}}-1)}\times 100\%$$

(4)

$${\mathrm{P}}_{\mathrm{a}}^{\prime }={\mathrm{P}}_{\mathrm{a}}·\frac{1}{1+{\mathrm{P}}_{\mathrm{i}}·\mathsf{\eta }·(\frac{{\mathsf{\rho }}_{\mathrm{s}}}{{\mathsf{\rho }}_{\mathrm{i}}}-1)}\times 100\%$$

(5)

where P_a is the asphalt aggregate ratio of the asphalt mixture (%), ${\mathrm{P}}_{\mathrm{a}}^{\prime }$ is the nominal asphalt aggregate ratio of the SSAM (%), m_a is the asphalt mass (g), M is the total mass of aggregate without adding steel slag (g), m_i is the aggregate mass replaced by steel slag (g), η is the volume percentage of aggregate replaced by steel slag (%), ρ_s is the gross volume relative density of steel slag (g/cm³), ρ_s is the relative density of the gross volume of the aggregate to be replaced (g/cm³), and P_i is the total weight of a certain type of limestone aggregate to be replaced (kg).

2.6. Road Performance Test of SSAM

For the SSAM to have the potential to melt ice and snow, it should have good road performance as well as good thermal conductivity. Steel slag was added to the asphalt mixture to replace the aggregate, and the SSAM was prepared under the condition that the optimal amount of asphalt remained unchanged. Through high-temperature rutting tests, low-temperature trabecular bending tests, and freeze–thaw splitting tests, the effects of different amounts of steel slag on the road performance of asphalt mixture were studied.

2.6.1. Rutting Test

At present, the test method used to evaluate the high-temperature stability of asphalt mixture is mainly the rutting test. The rutting test is used to evaluate the high-temperature deformation resistance of the asphalt mixture according to the standard test method (T0719-2011) [27]. The testing effect is intuitive and clear. The size of the rutting plate specimen is 300 × 300 × 50 mm³. The specimens were formed by the wheel rolling method (T0703-2011) [27], and the compaction temperature was 160 °C. The temperature during the rutting test was 60 °C and the rubber wheel pressure was 0.7 MPa. The correlation between high-temperature performance and rutting depth is characterized by dynamic stability. Dynamic stability is the ratio of the number of rolling runs to the depth of the rut. The greater the dynamic stability, the stronger the deformation resistance of the specimen, and the better the high-temperature stability of the asphalt mixture.

2.6.2. Low-Temperature Bending Test

When the temperature is lower in winter, the volume of the asphalt surface tends to shrink. However, the asphalt surface layer cannot shrink normally under the restraint of the surrounding materials, so thermal stress is generated inside the structure. When the generated temperature stress is greater than the allowable tensile stress of the asphalt mixture, the asphalt mixture is pulled and cracked, resulting in cracks in the asphalt pavement. Therefore, the asphalt mixture must have good resistance to low temperature cracking to ensure the normal use of the road surface. The low-temperature deformation resistance of asphalt mixture is usually evaluated by low-temperature bending tests according to the standard test method (T0715-2011) [27]. The specimens were formed in the same way as in Section 2.6.1, but the specimens were cut into beams after forming. The evaluation index is the maximum bending tensile strain of the specimen at low-temperature failure. The larger the bending tensile strain, the better the low-temperature flexibility of the asphalt mixture; that is, the better the crack resistance.

2.6.3. Retained Marshall Stability Test and Freeze–Thaw Split Test

The insufficient moisture resistance of asphalt mixtures is manifested in the reduction of the cohesion of asphalt mixtures under the action of water. As a result, the asphalt peels off the surface of the aggregate particles, and the pavement has potholes. To ensure the service level and service performance of the road, the water damage resistance of the asphalt mixture should be improved. The moisture resistance evaluation methods include the retained Marshall stability test and freeze–thaw split test. The corresponding indicators are the residual stability ratio and the freeze–thaw splitting strength ratio. These two methods were used according to specifications (T0709-2011 and T0729-2011) [27] to evaluate the moisture resistance of the SSAM. The specimens were formed by the compaction method (T0702-2011) [27], and the compaction temperature was 160 °C.

3. Results and Discussion

3.1. Microscopic Characterization of Thermal Conductive Phase

3.1.1. Mineral Composition

It can be seen from the XRD pattern of the steel slag in Figure 4 that the mineral components in the steel slag include iron-containing oxides (FeO), magnesium-containing oxides (MgO), aluminum-containing oxides (AlO), etc. The mineral component of limestone is relatively simple; most of the peaks point to calcite (CaCO₃). The main mineral components of steel slag and limestone are shown in

Table 9. Main mineral composition of steel slag and limestone.

Materials	Mineral Composition (%)
	CaO	Fe2O3	SiO2	MgO	MnO	Al2O3
Steel slag	37.15	31.28	14.40	5.38	4.28	2.77
Limestone	54.37	0.23	0.74	0.86	0.01	0.37

. It is worth noting that since the XRD analysis software does not show the peak positions of SiO₂ and MnO, the mineral composition displayed by the XRD test is somewhat different from the XRF test.

Studies have shown [31,32] that the heating rate of minerals is related to the activity of their composition. The better the activity of the material, the faster the heating rate; that is, the stronger the thermal conductivity. According to the different heating rates of mineral components, the components can be classified as more active (Fe₃O₄), active (Fe₂O₃, FeS), less active (MgO, Al₂O₃), and inactive (SiO₂, CaO). Minerals with higher iron content are more active. The classification results of the ingredients are shown in Figure 5.

There are many kinds of minerals in steel slag, notably the presence of iron-containing active minerals, which not only increases the activity of steel slag but also greatly improves the thermal conductivity of steel slag itself. On the contrary, limestone only contains calcite, an inactive mineral, which has poor thermal conductivity.

3.1.2. Surface Topography

Steel slag is a typical porous material with higher porosity than limestone, so steel slag has a stronger adsorption capacity for water and asphalt. In addition, the thermal conductivity of the solid components in the steel slag is significantly higher than that of the air in the pores. Therefore, the larger the slag voids, the poorer the thermal conductivity may be. Figure 6 shows the SEM images of steel slag at 5000× and 40,000× magnification. In Figure 6a,b, large and small protrusions can be seen on the surface of the steel slag, which makes the surface very rough. In Figure 6b, the micropores on the surface of the steel slag can be clearly seen. The protrusions and micropores of steel slag may increase the porosity of SSAM [29], which may affect the overall thermal conductivity of the SSAM.

3.2. Thermal Coefficient of SSAM

3.2.1. Measured Thermal Coefficient

During the heating process, the current in the heating wire is constant, and the heating power remains unchanged. Therefore, the rate of change of temperature rise measured by the thermocouple can directly reflect the thermal conductivity of the specimen. The temperature rise curves of asphalt mixtures with different steel slag contents are shown in Figure 7. During a heating period of 30–75 s, the mixture between the temperature sensor and the heating wire is preheating. The heat transfer to the sensor is lower and the temperature rise is smaller. The recorded data are the ambient temperature in the initial state. With the prolongation of heating time, the mixture gradually began to heat up. The heating wire temperature is transmitted to the sensor through the mixture. Elevated temperature and time are not linear in the transition phase due to heat transfer instabilities. When heated to 90 s, the heat transfer is stable, the mixture heats up sharply, and the temperature rise is almost linearly related to time. This obvious phenomenon can be observed in the temperature rise curves of asphalt mixtures under all slag contents. Theoretically, the slope of the temperature rise curve can reflect the heating rate of the material. Although the data collected and recorded by the three temperature sensors in the same specimen are slightly different, the slopes of the three curves are the same. This indicates that the distribution of steel slag in the asphalt mixture is relatively uniform, and the results of the test method are reliable.

The thermal coefficient of the asphalt mixture under different steel slag contents is shown in Figure 8. With the increase in steel slag content, the thermal coefficient of asphalt mixtures first increases and then decreases. When the content of steel slag is about 60%, the thermal coefficient of SSAM is the largest, which is about 120.3% of the thermal coefficient without steel slag. The thermal coefficient of asphalt mixture with 20% and 40% steel slag content also increased to a certain extent compared with 0% content, increasing by 11.4% and 14.6%, respectively. When the content of steel slag exceeds 60%, the thermal coefficient obviously decreases. When the content is 100%, the steel slag completely replaces the aggregate of 3–5 mm, and the thermal coefficient of the asphalt mixture is even lower than that of the original mixture (11.1%). This phenomenon occurs because steel slag is a typical porous material, and its porosity is significantly higher than that of the limestone aggregate. As the steel slag content increases, the porosity of the SSAM increases, which is verified in Section 3.3. Solids conduct heat better than air. Therefore, the higher slag content hinders the improvement of the thermal conductivity of SSAM. Although more slag means more iron-containing oxides in SSAM, when the slag content is high, the effect of thermal conductivity reduction caused by increased porosity may be more prominent. When the content of steel slag is moderate, the highly active metal elements in the steel slag may play a major role relative to the porosity, so the thermal coefficient of the SSAM is higher than that of the asphalt mixture without steel slag.

3.2.2. Calculated Thermal Coefficient

There are many test methods for the thermal coefficient of asphalt mixture, and different test methods have different test results due to different test principles and test conditions. Theoretical calculations determine the theoretical reference value according to the material composition characteristics. The theoretical value and the actual measurement value can be mutually verified to verify the feasibility of the test method to a certain extent.

In 1972, Williamson [33] proposed the theoretical calculation Equation (6) of the thermal coefficient prediction model of asphalt mixture based on the research on the thermal properties of asphalt mixture.

$${\mathsf{\lambda }}_{\mathrm{ac}}={({\mathsf{\lambda }}_{\mathrm{b}})}^{{\mathrm{V}}_{\mathrm{b}}}·{\left({\mathsf{\lambda }}_{\mathrm{m}}\right)}^{{\mathrm{V}}_{\mathrm{m}}}·{\left({\mathsf{\lambda }}_{\mathrm{a}}\right)}^{{\mathrm{V}}_{\mathrm{a}}}·{\left({\mathsf{\lambda }}_{\mathrm{w}}\right)}^{{\mathrm{V}}_{\mathrm{w}}}$$

(6)

where λ_ac is the thermal coefficient of the asphalt mixture; λ_b, λ_m, λ_a, and λ_w are the thermal coefficient of asphalt, aggregate, air, and water, respectively; and V_b, V_m, V_a, and V_w are the volume percentages of asphalt, aggregate, air, and water in the asphalt mixture, respectively.

According to the test data in

Table 10. Test results for Marshall specimens.

Asphalt Aggregate Ratio (%)	Steel Slag Content (%)	Nominal Asphalt Aggregate Ratio (%)	Gross Volume Relative Density	Water Absorption (%)	Void Ratio (%)
4.90	0	4.90	2.420	0.4	3.71
	20	4.87	2.447	0.6	3.70
	40	4.83	2.460	0.8	3.78
	60	4.80	2.469	1.0	3.82
	80	4.77	2.478	1.1	4.13
	100	4.73	2.489	1.3	4.37

, when the content of steel slag is 0%, the density of the asphalt mixture is 2.420 g/cm³, and the porosity is 3.71%. Combined with the density of each raw material in Table 1 and Table 2, it can be calculated that the volume percentages of asphalt, mineral material, and water in the asphalt mixture are 11.11, 85.12, and 0.06%, respectively. The thermal coefficients of asphalt, limestone aggregate, air, and water were taken as 0.7, 2.04, 0.024, and 0.6 W/(m·°C), respectively. By substituting the above data into Equation (6), λ_ac can be obtained as 1.3821 W/(m·°C).

The theoretically calculated thermal coefficient is 1.3821 W/cm³, which is only 4.78% different from the measured value of 1.4515 W/cm³ measured by the parallel heating wire method. The thermal coefficient of asphalt mixtures measured by the heating wire method is relatively reliable.

3.3. Volume Indices of SSAM

The test results of the SSAM are shown in Table 10. With the increase of steel slag content, the nominal asphalt aggregate ratio and gross volume relative density increase, because the density of steel slag aggregate is higher than that of limestone aggregate. At the same time, the water absorption and porosity of SSAM also increase with an increase in steel slag content. Because the gradation curves of each SSAM are similar, the reason for the increase in porosity may be the increase in the porosity of the steel slag itself. In addition, the roughness of the steel slag surface reduces the workability of the SSAM, which may also be a factor affecting the porosity of the SSAM [29].

3.4. Road Performance Verification of SSAM

3.4.1. High-Temperature Performance

The test results of the dynamic stability of the asphalt mixture with different steel slag contents are shown in Figure 9 below. The dynamic stability of asphalt mixtures first increases and then decreases with the increase of steel slag content. When the steel slag completely replaces the aggregate, the dynamic stability is even smaller than that without steel slag. When the slag content increased from 0% to 60%, the dynamic stability value increased from 3947 to 5732, an increase of 45.2%. An appropriate amount of steel slag can significantly increase the high-temperature deformation resistance of the asphalt mixture, which is due to the large specific gravity and high strength of the steel slag itself. The close contact between the steel slag and the aggregate enhances the integrity of the structure, improving the resistance to high-temperature deformation. However, with the further increase in steel slag content, the pores of the mixture increase. Pore adsorption of asphalt reduces the free asphalt content in the mixture and reduces the adhesion between aggregates. The kneading effect at high-temperature conditions causes the low-adhesion aggregates to loosen and fall off, which eventually leads to a decrease in the strength and high-temperature stability of the asphalt mixture.

3.4.2. Low-Temperature Performance

The bending test results of asphalt mixture trabeculae with different steel slag contents are shown in Figure 10 below. With the increase in steel slag content, the flexural and tensile strain of asphalt mixture at low-temperature failure decreases linearly, and the range of decrease is large. When the content exceeds 80%, it does not meet the requirements of the specification, indicating that the addition of steel slag has a great influence on the low-temperature crack resistance of the asphalt mixture. This is because the steel slag has more pores and a larger amount of adsorbed asphalt per unit volume, which reduces the adhesion of asphalt to aggregates. The cohesion between the aggregate particles is reduced, and it is not sufficient to resist the shrinkage deformation caused by the temperature stress. In addition, with the increase in steel slag content, the porosity of asphalt mixtures increases. The steel slag adsorbs the asphalt, resulting in a significant stress relaxation effect near the steel slag. However, the asphalt content in the SSAM is fixed, which makes the asphalt content in some non-steel slag areas insufficient. Overall, the SSAM lacks the cementation of asphalt, making it more susceptible to cracking from thermal stress. In general, steel slag reduces the low-temperature performance of the asphalt mixture and the replacement ratio of steel slag should not exceed 60 %.

3.4.3. Moisture Resistance

As shown in Figure 11, as the steel slag increases, the freeze–thaw splitting strength ratio and residual stability of SSAMs decreases. The incorporation of steel slag reduces the moisture resistance of the asphalt mixture. For the freeze–thaw splitting strength index, the addition of steel slag reduces the water damage resistance of the asphalt mixture, resulting in a maximum reduction of the freeze–thaw splitting strength ratio by 9%. However, the freeze–thaw split index still meets the specification value of 75%. For the residual stability index, when the slag content is greater than 80%, the residual stability drops sharply. The maximum decrease in residual stability is 17.5%, which does not meet the specification value of 80%. The steel slag affects the moisture resistance of the asphalt mixture because the steel slag absorbs a large amount of asphalt, which reduces the free asphalt in the asphalt mixture. The asphalt film on the aggregate surface becomes thinner. Water has a stronger adsorption force to the aggregate, causing the water to pass through the asphalt membrane to separate the aggregate from the asphalt. As a result, the asphalt peels off the surface of the aggregate, the mixture becomes loose, and the strength is reduced. In addition, with the increase in the amount of steel slag, the porosity of the asphalt mixture increases, so the moisture can more easily enter the interior of the asphalt mixture structure. Under the action of hydrodynamic pressure, the asphalt migrates or even peels off from the aggregate surface, which reduces the moisture resistance of the asphalt mixture. In addition, incomplete slag aging may also be a factor in reducing moisture resistance [34]. Similar conclusions are also reflected in some literature [35,36]. Of course, some scholars have pointed out that with the increase in steel slag content, the moisture resistance of SSAM increases [29,37,38]. This is because the adhesion of steel slag to asphalt is higher than that of ordinary aggregate. Therefore, in conclusion, the good adhesion performance of steel slag and asphalt has a dual effect, which may lead to increased asphalt content near the steel slag and decreased asphalt content in the non-steel slag area. The effect of steel slag on the moisture resistance of SSAM is still open for discussion.

In terms of road performance, the addition of steel slag can significantly improve the high-temperature rutting resistance of asphalt mixtures, but it can also reduce the low-temperature crack resistance and moisture resistance in some cases. Combined with the thermal conductivity and road performance of asphalt mixture under different steel slag content, when the steel slag content is 60%, the asphalt mixture has the best high-temperature stability and the highest thermal conductivity. Under these conditions, the low-temperature crack resistance and moisture resistance meet the requirements of the technical specification [28]. Therefore, the optimum content of steel slag content should not exceed 60% (6.6% of the total mixture).

4. Conclusions

To better utilize the functional properties of steel slag in asphalt mixtures, such as ice and snow melting, self-healing, etc., the microscopic properties and thermal conductivity of steel slag and SSAM were studied in this paper. Further, to ensure the SSAM meets the engineering requirements, the basic road performance of SSAMs with different thermal conductivity is verified. The main conclusions are as follows:

(1)

The XRD and XRF tests of steel slag showed that, compared with limestone, iron-containing active oxides are the main factor driving the heat conduction of steel slag. The SEM test showed that the surface of the steel slag was rough and microporous, which may be detrimental to the thermal conductivity of the SSAM.

(2)

The thermal conductivity of SSAM first increases and then decreases with the content of steel slag. When the content of slag is less than 60%, the beneficial effect of iron-containing oxides on thermal conductivity may be dominant. When the content is greater than 60%, the main factor affecting the thermal conductivity of steel slag may be the increased voids in the mixture. The porosity of SSAM increases with the increase of steel slag content, which supports this conjecture.

(3)

When the steel slag content is 60%, the highest thermal coefficient of 1.746 W/(m·°C) is reached. The difference between the test value and the theoretical value is 4.78%, which verifies the reliability of the test result.

(4)

The road performance tests show that there is an optimal content to make the high-temperature performance optimal, while the low-temperature performance and moisture resistance decrease with increasing slag content.

(5)

After considering the thermal conductivity and road performance of SSAM for functional applications, the optimum amount of steel slag should not exceed 60% of the aggregate volume of 3–5 mm (6.6% of the total SSAM). In this paper, a preliminary evaluation of the thermal conductivity of SSAM is presented, and the long-term performance of the mixture needs to be further studied.