Construction and Building Materials 178 (2018) 542–550 Contents lists available at ScienceDirect Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat Study on the effective composition of steel slag for asphalt mixture induction heating purpose Jiuming Wan a, Shaopeng Wu a,⇑,1, Yue Xiao a,⇑,1, Zongwu Chen b, Dong Zhang a a b State Key Laboratory of Silicate Materials for Architectures, Wuhan University of Technology, 430070, China Faculty of Engineering, China University of Geosciences (Wuhan), Wuhan 430074, China h i g h l i g h t s  Effect of iron and Fe3O4 on induction heating of steel slag were investigated.  Size effect on steel slag’s induction heating had been studied.  How iron element proportion affect the induction heating were tested.  The induction heating mechanisms of Fe3O4 and iron are analyzed. a r t i c l e i n f o Article history: Received 11 December 2017 Received in revised form 15 May 2018 Accepted 22 May 2018 Keywords: Steel slag Induction heating Effective component Iron element proportion Aggregate size a b s t r a c t Reuse steel slag as aggregate for induction heating has grown up to an important subject of asphalt pavement. However, how the composition and size of steel slag affect its induction heating efficiency still requires further study. This paper investigated steel slags’ induction heating efficiency by components, size and iron element proportion dependency analysis. Steel slags and powders of iron, Fe3O4, natural mineral as well as steel slags were employed to fabricate samples that simulate steel slags as various composition. Their corresponding heating efficiencies were acknowledged through infrared camera. Moreover, X Ray Fluorescence and Electron Probe Micro Analyzer are used to characterize the element compositions of the samples and elements distribution within steel slags. Results showed that, the effective components of steel slag’ induction heating were Fe3O4 and iron. Fe3O4 contributed the dominant part to steel slags’ induction heating efficiency at micron size (smaller than 75 lm). In addition, the size increase of steel slags showed negative effect on their induction heating efficiency, and steel slags of higher iron element proportion generally had higher heating efficiency. Alternatively, the induction heating mechanisms of Fe3O4 and iron are distinct. The dominant mechanism of iron is eddy current loss, while hysteresis loss and residual loss contribute the main part to Fe3O4’s induction heating. Through the superposition of the mechanisms, Fe3O4 consequently shows better induction heating efficiency than iron. This study provides laboratory foundation for selection and utilization of steel slags which is used as the heated material in further steel slag based asphalt mixture. Ó 2018 Elsevier Ltd. All rights reserved. 1. Introduction Rapid developing construction all over the world sharply consumes engineering materials. Excessive extraction of natural mineral resources such as basalt and limestone has already become a serious threat to ecological environment. To address this issue, researchers tried to use industrial solid waste [1] and marginal aggregate [2] to replace natural aggregate. ⇑ Corresponding authors. 1 E-mail addresses: wusp@whut.edu.cn (S. Wu), xiaoy@whut.edu.cn (Y. Xiao). S.P. Wu and Y. Xiao equally contributed to this work. https://doi.org/10.1016/j.conbuildmat.2018.05.170 0950-0618/Ó 2018 Elsevier Ltd. All rights reserved. Steel slag [3–5] is a kind of solid waste that has been utilized widely in civil engineering, and its utilization has benefit the environment as well as the entire industry. Researchers also found that steel slag also can be used as a material for induction heating. Li et al. [6] claimed that conductive asphalt concrete consists of steel slag has a higher temperature rise compared to the natural aggregates through induction heating. Addition of steel slag could improve the self-healing efficiency of asphalt pavement. Induction heating [7] has been widely used in industry and engineering. In pavement engineering, it had been employed as an approach to heat asphalt pavement that consist of magnetic materials [8–10]. J. Wan et al. / Construction and Building Materials 178 (2018) 542–550 The damages in pavement can be therefore healed by the flow of asphalt [11] due to the increment of temperature. The main elements of steel slag are Ca, Fe, Si, C and O. Researchers conclude that, metallic oxides such as MgO, CaO, Fe3O4 and nonmetallic oxides like SiO2 and C2S are the main components in steel slag [12–14]. To realize the induction heating on an object, it is indispensable that the heated object should contain magnetic components [15,16], such as iron and carbon steel. Without which induction heating will consequently fail. In this paper, the components that assist in the steel slag’s induction heating are defined as the ‘‘effective component”. Among these components of steel slag, the most metallic oxide, nonmetallic oxides, covalent compounds and metallic salt are not magnetic materials. Hence they do not contribute to induction heating. Nevertheless, after analysis and elimination, there are still two possible components, Fe3O4 and iron. They are commonly used as induction heating materials, their effect on steel slags still require further study. Iron is considered as the residual product after steelmaking, magnetic separation and aging process. The formation and existing form of Fe3O4 are more complicated, and the possible source is steel slag’s long-term outdoor complex aging. It should be pointed out that there may be other objects that can also be induction heated. However, considering their low mass proportion and few contributions to induction heating, they are not discussed. This study also investigated how the size and iron element proportion of steel slag affect steel slags’ induction heating efficiency, which are important issues in steel slag asphalt pavement engineering [17] for induction heating purpose. A components dependency experiment was developed to assess induction heating efficiency of the combination samples which are affected by iron and Fe3O4. Powders of Fe3O4, iron, mineral and steel slags were separately mixed to fabricate combination samples, which were then induction heated. Size effect on steel slags’ induction heating efficiency had also been tested in this study. Steel slags were classified as three size ranges, and their corresponding induction heating efficiencies were tested. Furthermore, iron element dependency experiment was also included to investigate the correlation between steel slags’ induction heating efficiency and its iron element proportion. Steel slags were classified according to their distinct induction heating efficiencies. Total element analysis [18] of these steel slags was conducted by X Ray Fluorescence (XRF) to characterize their corresponding element composition. Electron Probe Micro Analyzer (EPMA) [19,20] was also used to conduct EDS [21] map-scanning analysis on steel slags. This study provides an experimental foundation to researchers and engineers on steel slags selection and utilization in order to improve the induction heating efficiency of steel slag asphalt pavement. 2. Materials and methodologies 2.1. Materials Powders of Fe3O4, iron, traditional mineral and steel slag were used in component dependency experiment for induction heating purpose. Steel slag powder throughout this study, with essentially the same element composition with steel slag particles, was prepared in Jiangxi province. All powders were screened to particle size of smaller than 75 lm. Detailed information about the powders is given in Table 1. Steel slags throughout this study were also provided from Jiangxi province. These steel slags were cleaned and divided into grading ranges of 1.18 mm–2.36 mm, 2.36 mm–4.75 mm and 4.75 mm–9.5 mm. The specific heat capacities of these steel slags are almost the same. 543 2.2. Research methodologies Fig. 1 demonstrates the research outline of components, size and iron element proportion dependency experiments. Induction heating equipment is introduced to heat samples. Total element analysis and map-scanning analysis were conducted to characterize samples’ element compositions and element distributions within steel slag. Powders were used to fabricate the combination samples for components dependency experiment. Steel slags were used in size and iron element proportion dependency experiment. Then the mechanism of the effective composition for induction heating had been analyzed. 2.2.1. Equipment Fig. 2 presents the induction heating equipment, infrared camera and infrared picture of sample. Induction heating equipment is designed for laboratory use and offers an alternating magnetic field. An infrared camera is used to record samples’ resulted induction heating temperatures. The parameters of induction heating are shown in Table 2. Every sample was heated according to the parameters. Total element analysis is conducted on steel slag powder and steel slags for component characterization by using XRF. Map-scanning analysis of steel slags is also conducted by using EPMA, in order to examine the distribution of the main elements of steel slags. 2.2.2. Components dependency experiment Since iron and Fe3O4 are assumed as the main effective components in steel slags’ induction heating. How the effective components affect induction heating efficiency of steel slag is discussed. Total element analysis of steel slag powder and map-scanning analysis of steel slag were conducted. Components dependency experiment is designed to investigate the contributions of iron and Fe3O4 to induction heating efficiency of steel slag. Fig. 3 illustrates the test program for components dependency analysis. Iron, Fe3O4, mineral and steel slag powders listed in Table 1 were mixed as combinations in round plastic containers, which were then being induction heated. Combinations of these powders are assumed to simulate steel slags and aggregates that consist of various compositions, which can be viewed as the ‘‘simulated aggregate and steel slags”. By detecting the induction heating efficiency of the combination samples, influence of the effective components can be investigated. Table 3 shows the component and their mass proportion in combinations, and the total mass of every sample is 30 g. Components effect on induction heating efficiency of steel slag was then evaluated based on corresponding induction heating efficiency. Mineral powder, steel slag powder, iron powder and Fe3O4 powder are abbreviated as M-powder, S-powder, I-powder and F-powder respectively. Both induced temperature increment and heat quantity produced during induction heating process of samples were used to determine the samples’ induction heating efficiency. Eq. (1) was used to determine induced temperature increment (ITI) of samples in this study. ITI ¼ Final average temperature - Initial average temperature ð1Þ Temperature field and analysis method for ITI analysis is explained in Fig. 4. Since the upper surface of the container is circular, a same detection circle is set to calculate the overall temperature of samples. Radius of the detection circle is 80% of sample’s surface circle, and the centers of the two circles are in same position. The average temperature within detection circle is calculated 544 J. Wan et al. / Construction and Building Materials 178 (2018) 542–550 Table 1 Powder materials property at 20 °C. Materials Mineral powder Steel slag powder Iron powder Fe3O4 powder Component Particle size (lm) Purity Production location Specific heat capacity (KJ/(°C  Kg)) Magnetic Resistivity (X  m) Aging time (Year) Limestone 75 – Wuhan city 0.59 X Insulator 1 Steel slag 75 – Jiangxi province 1.2 X Insulator 5 a-Iron 75 >99.99% Beijing city 0.45 X 1  107 1 Fe3O4 75 >99.9% Beijing city 0.92 p Mineral powder Slag powder Iron powder Induction heating Sample preparation Fe3O4 powder Steel slag Components dependency Mechanism analysis on effective compositions for induction heating Size dependency Screening Iron element proportion dependency EPMA 102–103 1 in software, and was defined as the final average temperature of the samples. The initial average temperature of the samples are the same with environmental temperature. In addition, the temperature of samples at every 60 s was also recorded. The quantities of heat produced by samples were also taken into account, which is related to specific heat capacity (SHC) [22] of each components. Eq. (2) shows the calculation method of heat quantity. Q ¼ C  m  DT ð2Þ Where Q (J) is the produced heat quantity, C(J/(°C  Kg)) is the SHC, m (Kg) is the mass and DT(°C) is the temperature increment. Meanwhile, since the combinations consist of different powders, the overall SHC of samples’ should be calculated according to Eq. (3). It expressed the SHC of a combination that consists of different objects. Element distribution Fig. 1. Outline of research methodologies. Heating machine Sample Fig. 2. Induction heating equipment, infrared camera and sample’s infrared images. Table 2 Experimental parameters for induction heating. Parameters Heating distance (mm) Heating time (s) Alternating frequency (KHz) Power (Kw) Rated voltage (V) Environment temperature (°C) Value 46 240 123 7.1 650 24.3–25.0 Iron powder Fe3O4 powder Mineral powder Steel slag powder Mixing Sample of specific combination Under induction heating Fig. 3. Process of components dependency analysis. J. Wan et al. / Construction and Building Materials 178 (2018) 542–550 Table 3 Sample’s component combinations. Composition description Component combinations and mass proportion (%) M-powder S-powder I-powder F-powder M-powder and S-powder 100 83.3 66.6 50.0 33.3 16.6 0 0 16.6 33.3 50.0 66.6 83.3 100 0 0 0 0 0 0 0 0 0 0 0 0 0 0 M-powder and I-powder 83.3 66.6 50 33.3 0 0 0 0 16.6 33.3 50 66.6 0 0 0 0 M-powder and F-powder 83.3 66.6 50 33.3 0 0 0 0 0 0 0 0 16.6 33.3 50 66.6 S-powder and I-powder 0 0 0 0 0 90 80 70 60 50 10 20 30 40 50 0 0 0 0 0 S-powder and F-powder 0 0 0 0 0 90 80 70 60 50 0 0 0 0 0 10 20 30 40 50 S-powder, I-powder and F-powder 0 0 0 0 0 0 0 80 70 60 50 70 60 50 10 20 30 40 10 10 10 10 10 10 10 20 30 40 545 how the sizes of steel slags affect its induction heating efficiency, and the size ranges are 1.18 mm–2.36 mm, 2.36 mm–4.75 mm and 4.75 mm–9.5 mm. These screened steel slags were put into the containers that were used in components dependency experiment. The mass of steel slags in each sample are 50 g. Then the samples were induction heated and corresponding ITI values were acknowledged. Additionally, since the steel slags’ specific heat capacities of different size ranges are similar and samples’ mass are equal, the heat quantities produced by samples can be determined by their ITI. 2.2.4. Iron element proportion dependency experiment Since the mass proportion of iron and Fe3O4 have a certain correlation with the iron element proportion of steel slag. It is assumed that the iron element proportion of steel slags possibly affect its induction heating efficiency. In other word, if iron and Fe3O4 definitely play an essential role in the induction heating of steel slag, the iron element proportion of steel slag also should affect the induction heating efficiency. To examine the influence of iron element proportion on steel slag’s induction heating efficiency, iron element proportion dependency experiment was then discussed. Firstly, steel slags of 4.75 mm–9.5 mm size were induction heated. Then steel slags whose overall temperature rises were over 60 °C were picked out and labeled as High ITI steel slags. Steel slags whose overall temperature rises were below 20 °C were labeled as Low ITI steel slags. Total elements analysis was conducted on these steel slags by XRF for characterizing their corresponding element mass proportion. Therefore, the correlation between steel slag’s iron element proportion and its induction heating efficiency is then established. 3. Results and discussion 3.1. Component dependency experiment Fig. 4. Sample’s temperature field and temperature detection areas. PC i C o ¼ Pm i ð3Þ Where C o (J/(°C  Kg)) is the overall SHC of a combination sample, C i (J/(°C  Kg)) is the SHC of an object and mi (%) is the object’s corresponding mass proportion in the combination. 2.2.3. Size dependency experiment Size selection of steel slags in engineering is also an important issue. Size dependency experiment was introduced to investigate Main elements and their mass proportions in steel slag powder are presented in Table 4. Mass proportion of these elements are represented through the form of their corresponding typical oxides. Table 4 illustrates that, Fe, Ca and Si are the main elements in steel slag powder, and the mass proportion of Fe2O3 is 29.04%. Results of main elements distribution on the internal face of steel slags is shown in Fig. 5, and the magnification is 200. Corresponding elements are marked and distinguished by points of different colors. Fig. 5 suggests that distribution of iron element (Fe) is inhomogeneous and discontinues, and iron element substances are generally of micron size. Precisely based on this result, in this experiment, iron powder and Fe3O4 powder were designed to be mixed and scattered among the powder combinations, which are also inhomogeneous, discontinues and micron size. Therefore, these combinations approximately simulate the existence form of iron element substances within steel slags. Their induction heating efficiencies can serve as a reference to steel slag that of same composition. Steel slag powder and mineral powder are mixed as combinations, which were then induction heated to verify steel slag powders’ induction heating efficiency. Fig. 6 illustrates that, ITI values of combination samples show a gradual rising tendency along with increase in mass proportion of S-powder. It proves that the steel slag powder is able to be induction heated, and its ITI value within 240 s is 9.5 °C. Hence, steel slag powder also can be used as an effective induction heating object. Alternatively, combinations of M-powder with F-powder or I-powder were also induction heated to examine the induction heating efficiency of iron or Fe3O4. Their 546 J. Wan et al. / Construction and Building Materials 178 (2018) 542–550 Table 4 Main components and their mass proportion in steel slag powder. Components Fe2O3 CaO SiO2 MgO Al2O3 P2O5 MnO TiO2 Mass proportion (%) 29.04 29.08 13.84 10.09 5.02 0.64 1.88 0.65 ITI results are summarized in Fig. 7. Both ITI values of two combinations show rising tendency as mass proportion of F-powder or Ipowder increase. However, when F-powder and I-powder occupy a same mass proportion of combinations, the F-powder’s ITI are generally much higher than I-powder’s ITI. It can be noted that, combinations containing Fe3O4 have much greater ITI than combinations containing iron in same mass proportion. The disparity between their ITI rises as their mass proportion increase. In addition, since the induction heating on steel slag is a heat production process, thus the heat produced by combinations were discussed. Due to short heating time and few temperature difference between combination and environment, their heat dissipations to the atmosphere were not included in this experiment. Heat quantities produced during induction heating process is approximately defined as the heat quantities that raised combinations’ overall temperature to corresponding degree. Eq. (2) expresses the calculation method of combinations’ heat quantities produced during induction heating. SHC of these combinations can be calculated according to Eq. (3). Fig. 8 indicates the heat quantity and SHC of combinations of Mpowder with F-powder or I-powder. Both the heat quantity of F- Ca Si Ka1 Ka1 Mg Fe Ka1 Ka1 powder combinations and I-powder combinations increase as their mass proportion in combinations increase. Combinations containing F-powder had produced much higher heat quantities during the process. On the other hand, the SHC values of F-powder combinations increase along with the rise of F-powder’s mass proportion as well. However, I-powder combinations show an opposite tendency, and generally the SHC values of F-powder combinations are higher. Because the SHC value of F-powder is higher than M-powder, and I-powder has the lowest SHC value. Thus, the SHCs of Fpowder combinations are accordingly higher than I-powder combinations. The SHC of F-powder combination increase as its mass proportion rise while I-powder combinations decrease at the same time. Because F-powder combinations has higher temperature rise (DT) and higher SHC (C) than I-powder combinations. Consequently, F-powder combinations of same mass (m) are able to produces much higher quantity of heat than I-powder combinations during induction heating process. Therefore both the ITI and heat quantities produced by F-powder combinations are much higher than I-powder combinations in equal mass proportion. Fe3O4 has greater induction heating efficiency than iron at this size (smaller than 75 lm). Fig. 9 presents the infrared temperature field of a combination during the induction heating process. The infrared temperature Fig. 7. ITI result of combinations of M-powder with F-powder or I-powder. Fig. 5. Main elements distribution within steel slags. Fig. 6. ITI result of combinations of M-powder with S-powder. Fig. 8. Heat quantity and SHC of combinations of M-powder with F-powder or Ipowder. J. Wan et al. / Construction and Building Materials 178 (2018) 542–550 field pictures of combination at every 60 s are recorded. As the picture illustrated, the temperature fields of a combination gradually become uniform as the induction heating time extend. It indicates that there is also a heat transferring process within the powders. Fig. 10 shows the ITI rising tendency of combinations of Mpowder with I-powder or F-powder respectively during induction heating process. It suggests that ITI values of combinations at every 60 s are positively correlated with their final ITI. Combinations that contains Fe3O4 powder prefer to show higher temperature rising rate during the entire induction heating process. How Fe3O4 and iron affect steel slag combinations’ induction heating efficiency is demonstrated in Fig. 11. F-powder and Ipowder combinations show definitely distinct induction heating efficiency. Mass proportion change of I-powder in the combinations doesn’t show a remarkable influence on ITI. It suggests that iron powder of same mass has parallel ITI with steel slag powder at this size. On the other hand, ITI of combinations that contain Fe3O4 powder show a gradual rise along with mass proportion increase, which is consistent with combinations of M-powder with F-powder. Furthermore, results of combinations of S-powder with Ipowder or F-powder are shown in Fig. 12, respectively. Part (a) illustrates that, mass proportion change of I-powder in the combinations also has no remarkable impact on ITI values. Part (b) suggests that increasing mass proportion of F-powder lead to the rise of combinations’ ITI. Hence, the co-existence of Fe3O4 and iron in steel slag powder has mutual influence, their contributions to induction heating are independent. 60s 120s 547 Fig. 11. ITI result of combination of S-powder with I-powder or F-powder. Heat quantity and SHC results of combinations of S-powder with F-powder or I-powder are presented in Fig. 13. Both SHC values of the I-powder and F-powder combinations show decreasing tendency along with the mass proportion rise. The descent rate of I-powder combinations is faster. Because of S-powder’s high SHC, the addition of the two powders reduce the overall SHC of combination, and the more the powders are inserted, the more of SHC will decrease. Besides, heat quantity of I-powder combinations also decrease as the mass proportion increase, while F-powder combinations’ heat quantities illustrate a definitely increasing tendency. Considering the fact that ITI of I-powder combinations are similar and the mass are equal, the lower SHC surely result in lower heat quantity produced according to Eq. (2). Although the SHCs of F-powder combinations presents a decreasing tendency, the remarkable ITI rises of F-powder combinations also lead to the increasing heat quantity tendency. Therefore, F-powder has higher ITI and produces higher heat quantity than I-powder in combinations’ induction heating. It indicates that Fe3O4 plays a dominant role when combined with S-powder, while iron does not contribute much at this size. 180s 3.2. Size dependency experiment 180s 240s Fig. 9. Infrared temperature field of combination sample during induction heating. Average ITI results of steel slag’s induction heating by different size ranges are presented in Table 5. Steel slags show a decreasing ITI tendency as the size ranges decrease. Since the steel slags are of similar SHC values and their mass are equal, it implies that the steel slags has higher ITI will surely produce higher quantity of heat. Thus, size increment may have a negative effect on steel slags’ induction heating efficiency. 3.3. Iron element proportion dependency experiment Results in Table 6 indicate that, the Fe2O3 mass proportion of high ITI steel slags is 19.57%, which is over 5 percent higher than that of low ITI steel slags. Except iron element, other elements’ mass proportion almost stays the same, hence the variable in this experiment is the iron element mass proportion. It proves that, steel slags of higher iron element proportion are likely to has greater induction heating efficiency. Iron element substances play an dominant role in steel slags’ induction heating. 3.4. Mechanism analysis Fig. 10. Temperature rising tendency of combination of M-powder with I-powder. Results of component dependency experiment indicate that, as effective components, Fe3O4 powder produces much more quantity of heat than iron powder during induction heating. To explain this 548 J. Wan et al. / Construction and Building Materials 178 (2018) 542–550 Fig. 12. ITI results of combination of S-powder, I-powder and F-powder. We ¼ afd B2m 2=q 2 Fig. 13. Heat quantity and SHC of combinations of S-powder with F-powder or Ipowder. phenomenon, mechanisms which contribute to steel slags induction heating are analyzed. Induction heating of steel slags is actually magnetic loss, which means the magnetic materials produce energy losses in alternating magnetic fields. The magnetic loss for magnetic materials contains three parts, eddy current loss [23,24], hysteresis loss [25] and residual loss [26,27]. It can be concluded as Eq. (4). W ¼ We þ Wh þ Wr ð4Þ Where, W(J) is the total magnetic loss, W e (J) is the eddy current loss, W h (J) is the hysteresis loss and W r (J) is the residual loss. 3.4.1. Eddy current loss Eddy current is an induction current within an object. It is produced in an alternating magnetic field and determined by several parameters. The We can be summarized as Eq. (5) [28]. Table 5 Overall ITI and heat quantity of steel slags by various size ranges. Component (%) 1.18–2.36 mm 2.36–4.75 mm 4.75–9.5 mm Overall ITI (°C) Heat quantity (J) 29.9 1798.21 25.2 1513.85 17.1 1021.21 ð5Þ Where, a is the fixed number, f (Hz) is the frequency of the alternating magnetic field, d(m) is the thickness of the material, Bm (A/m) is the magnetic induction intensity and q(X  m) is the resistivity of material. In this study, a, f, d and Bm are considered to be consistent for each test. The decisive factor is their resistivity. Seen from Table 1, resistivity of iron is even about 109 times lower than Fe3O4. According to Eq. (5), We of iron is therefore almost 109 times higher than that of Fe3O4. In another word, the heat produced by eddy current loss of Fe3O4 is very slight, which even can be ignored comparing to the heat quantity produced by iron. Alternatively, the thickness of iron materials introduced in this study is also very small (smaller than 75 lm). According to Eq. (5), iron powder surely has much lower induction heating efficiency comparing to iron materials of larger size. Fig. 5 illustrates that the distribution of iron element in steel slag is generally of micron size. Hence the induction heating efficiency of residual iron exists in the steel slag is restrained by its size. 3.4.2. Hysteresis loss In the process of repeated magnetization, the induction magnetic of magnetic materials always lags behind outer magnetic field strength. This phenomenon is called hysteresis, and power loss due to hysteresis phenomenon is the hysteresis loss [29]. This part of loss is irreversibly converted to the heat, calculation of hysteresis loss of unit volume in a circle can be expressed as Eq. (6) [30]. I Wh ¼ ð6Þ HdB Where H (A/m) is the magnetic field intensity applied on the materials, B (A/m) is the magnetic induction intensity. An important factor that determines the intensity of hysteresis loss is the coercive force (Hc) of magnetic material. Materials of greater coercive force usually have stronger ability to resist being magnetized, which also results in higher hysteresis loss [29]. Coercive force of pure iron is less than 96 A/m, and Fe3O4’s coercive force is generally over 4700 A/m. As a consequence, Fe3O4 usually show a much stronger resistance to magnetization, higher hysteresis loss is accordingly Table 6 Main element and their mass proportion of high and low ITI steel slag. Component (%) Fe2O3 CaO SiO2 MgO Al2O3 P2O5 MnO TiO2 High ITI Low ITI 19.57 14.56 40.98 39.78 20.49 23.36 5.58 5.62 4.53 4.60 1.98 1.75 2.42 3.46 1.10 1.41 J. Wan et al. / Construction and Building Materials 178 (2018) 542–550 549 3. The induction heating mechanisms of Fe3O4 and iron are distinct. Eddy current loss contributes greatly to iron’s induction heating. However, iron’s induction heating efficiency is constrained by its micron size in steel slag. Eddy current loss has almost no impact on Fe3O4 since the conductivity of Fe3O4 is low. On the other hand, residual loss and hysteresis loss rather than eddy current loss contribute the dominant part to Fe3O4’s heat production in high-frequency magnetic field. Through the superposition of the mechanisms, Fe3O4 consequently shows better induction heating efficiency than iron. 5. Conflict of interest The authors declare no conflict of interest. Fig. 14. Mechanisms’ contribution to the heat production of Iron and Fe3O4 during induction heating. Acknowledgement produced during its induction heating. Therefore, the heat quantity due to hysteresis loss of Fe3O4 powder is higher than iron powder. 3.4.3. Residual loss Residual loss refers to the other losses except eddy current loss and hysteresis loss. The residual loss at high frequency field is mainly caused by the size resonance [26], domain wall resonance [31] and natural resonance [32]. Moreover, residual loss is the predominant mechanism in ferrite’s induction heating [26,27,33]. The frequency of magnetic field introduced in this study is over 104 Hz, it belongs to the category of high frequency. Also because Fe3O4 is a ferrite, residual loss contributes a major part to Fe3O4’s induction heating owing to the complex resonance effect. Fig. 14 simply concludes the contributions of different mechanisms of iron and Fe3O4, induction heating mechanisms’ contributions of the two effective components are different. Eddy current loss caused by magnetization process contributes greatly to the induction heating of iron powder, which has virtually no effect on Fe3O4, while the main magnetic loss mechanism of Fe3O4 is residual loss and hysteresis loss. Through superposition of these mechanisms, Fe3O4 consequently has better induction heating efficiency than iron at this size. 4. Conclusion To investigate how the effective components, sizes and iron element proportion affect steel slags’ induction heating efficiency, components, size as well as iron element proportion dependency experiments were conducted. Based on the corresponding results, the following conclusions can be acknowledged. 1. The steel slags were able to be induction heated, and the effective components for induction heating were supposed to be iron and Fe3O4. Temperature increment and heat quantity of Fe3O4 powder was much higher than iron powder when mixed with mineral powder and steel slag powder by same mass and size. Their contributions to induction heating were independent when they were co-existed in the samples. It proves that the Fe3O4 plays the dominant part to steel slags’ induction heating when their sizes are smaller than 75 lm. 2. Size dependency experiment result indicated that the size increment had a negative effect on steel slags’ induction heating efficiency. On the other hand, results of iron element proportion experiment suggested that the steel slags of higher iron element proportion are likely to show higher temperature rise during induction heating. It means that iron element proportion of steel slags is positively correlated with steel slags’ induction heating efficiency. The founding from the National Natural Science Foundations of China (51778515 and 51708437) are gratefully acknowledged. References [1] G. Tchobanoglous, H. Theisen, S. Vigil, Integrated solid waste management: engineering principles and management issues, Water Sci. Technol. Lib. 8 (1) (1993) 63–90. [2] G.B.R. Kumar, U.K. Sharma, Abrasion resistance of concrete containing marginal aggregates, Constr. Build. Mater. 66 (1) (2014) 712–722. [3] P. Ahmedzade, B. Sengoz, Evaluation of steel slag coarse aggregate in hot mix asphalt concrete, J. Hazard. Mater. 165 (1–3) (2009) 300–305. [4] H. Motz, J. Geiseler, Products of steel slags an opportunity to save natural resources, Waste Manage. (Oxford) 21 (3) (2001) 285–293. [5] S. Hu, H. Wang, G. Zhang, Q. Ding, Bonding and abrasion resistance of geopolymeric repair material made with steel slag, Cem. Concr. Compos. 30 (3) (2008) 239–244. [6] B. Li, Y. Sun, Q. Liu, H. Fang, S. Wu, J. Tang, Q. Ye, Ice melting properties of steel slag asphalt concrete with microwave heating, Materials Science and Engineering Conference Series, 2017, p. 012041. [7] F. Bay, V. Labbe, Y. Favennec, J.L. Chenot, A numerical model for induction heating processes coupling electromagnetism and thermomechanics, Int. J. Numer. Meth. Eng. 58 (6) (2003) 839–867. [8] S. Wu, L. Mo, Z. Shui, Z. Chen, Investigation of the conductivity of asphalt concrete containing conductive fillers, Carbon 43 (7) (2005) 1358–1363. [9] H. Wang, J. Yang, H. Liao, X. Chen, Electrical and mechanical properties of asphalt concrete containing conductive fibers and fillers, Constr. Build. Mater. 122 (2016) 184–190. [10] X. Liu, W. Liu, S. Wu, C. Wang, Effect of carbon fillers on electrical and road properties of conductive asphalt materials, Constr. Build. Mater. 68 (15) (2014) 301–306. [11] Á. García, Self-healing of open cracks in asphalt mastic, Fuel 93 (1) (2012) 264– 272. [12] Y.X. Lun, K. Li, Characteristics of component of steel slag sand, Adv. Mater. Res. 598 (2012) 488–491. [13] Y.J. Zhang, L.C. Liu, Y. Xu, Y.C. Wang, D.L. Xu, A new alkali-activated steel slagbased cementitious material for photocatalytic degradation of organic pollutant from waste water, J. Hazard. Mater. 209–210 (4) (2012) 146–150. [14] J.K. Kang, C.G. Lee, J.A. Park, S.B. Kim, S. Yu, T.H. Kim, Use of converter furnace steel slag for bacteria removal in flow-through columns, Desalin. Water Treat. 51 (40–42) (2013) 7681–7689. [15] D.W. Kerst, The acceleration of electrons by magnetic induction, Phys. Rev. 58 (9) (1941). 841–841. [16] H. Huang, D.C. Fraser, Inversion of helicopter electromagnetic data to a magnetic conductive layered earth, Geophysics 68 (4) (2003) 1211–1223. [17] T.S. He, X.G. Zhao, S.Y. Zhao, Z.B. Li, Gray system study on the influence of particle size distribution of steel-slag powder on performance of cementitious materials, Key Eng. Mater. 531–532 (2013) 684–688. [18] T.I. Mclaren, C.N. Guppy, M.K. Tighe, N. Forster, P. Grave, L.M. Lisle, J.W. Bennett, Rapid, nondestructive total elemental analysis of vertisol soils using portable X-ray fluorescence, Soil Sci. Soc. Am. J. 76 (4) (2012) 1436–1445. [19] R.K. Sahoo, S.K. Singh, B.K. Mishra, Surface and bulk 3D analysis of natural and processed ruby using electron probe micro analyzer and X-ray micro CT scan, J. Electron. Spectrosc. Relat. Phenom. 211 (2016) 55–63. [20] D.E. Newbury, R.B. Marinenko, D.S. Bright, R.L. Myklebust, Computer-aided imaging: quantitative compositional mapping with the electron probe microanalyzer, Scanning 10 (6) (1988) 213–225. [21] F. Christien, J.F. Pierson, A. Hassini, F. Capon, R.L. Gall, T. Brousse, EPMA–EDS surface measurements of interdiffusion coefficients between miscible metals in thin films, Appl. Surf. Sci. 256 (6) (2010) 1855–1860. 550 J. Wan et al. / Construction and Building Materials 178 (2018) 542–550 [22] K. Xu, J. Song, F. Zhao, H. Ma, H. Gao, C. Chang, Y. Ren, R. Hu, Thermal behavior, specific heat capacity and adiabatic time-to-explosion of G(FOX-7), J. Hazard. Mater. 158 (2–3) (2008) 333. [23] H. Toda, Z. Xia, J. Wang, K. Atallah, D. Howe, Rotor eddy-current loss in permanent magnet brushless machines, IEEE Trans. Magn. 40 (4) (2004) 2104– 2106. [24] K. Yamazaki, F. Yu, M. Sato, Loss analysis of permanent-magnet motors with concentrated windings—variation of magnet Eddy-current loss due to stator and rotor shapes, IEEE Trans. Ind. Appl. 45 (4) (2008) 1334–1342. [25] M. Enokizono, H. Shimoji, T. Horibe, Loss evaluation of induction motor by using magnetic hysteresis E&S 2 model, IEEE Trans. Magn. 38 (5) (2002) 2379– 2381. [26] K. Takadate, Y. Yamamoto, A. Makino, Effect of grain size on residual loss of Mn-Zn ferrites, J. Magn. Soc. Jpn. 22 (2011) 381–383. [27] H. Kobiki, A. Fujita, S. Gotoh, Relationship between residual loss and microstructure of Mn-Zn ferrites, J. Phys. IV 7 (1997) 103–104. [28] F. Fiorillo, Characterization and measurement of magnetic materials, J. Virol. 65 (10) (2004) 5448–5456. [29] J.L. Woodward, Influence of coercive force on iron losses at low frequencies, Proc. Inst. Electr. Eng. 113 (4) (2010) 710–716. [30] J. Carrey, B. Mehdaoui, M. Respaud, Simple models for dynamic hysteresis loop calculations of magnetic single-domain nanoparticles: application to magnetic hyperthermia optimization, J. Appl. Phys. 109 (8) (2011). 083921-083921-17. [31] J.F.D. Jr, H.E.E. Jr, Domain wall motion and ferrimagnetic resonance in a manganese ferrite, J. Appl. Phys. 30 (2) (1959) 202–213. [32] S. Sugimoto, K. Okayama, H. Ota, M. Kimura, Y. Yoshida, H. Nakamura, D. Book, T. Kagotani, M. Homma, Effect of substitutional elements on the natural resonance frequency of barium M-type ferrite, J. Magn. Soc. Jpn. 23 (1) (1999) 611–613. [33] T. Chiba, S. Yamada, E. Otsuki, Temperature dependence of Eddy current loss and residual loss in Mn-Zn ferrites, J. Magn. Soc. Jpn. 22 (1998) 301–304.