Determination of chromium valence state in the CaO–SiO2–FeO–MgO–CrO x system by X-ray photoelectron spectroscopy

Abstract The chromium valence states in the CaO–SiO2–FeO–MgO–CrOx system were investigated by X-ray photoelectron spectroscopy (XPS). The results indicated that the XPS peaks of Cr 2p3/2 and Cr 2p1/2 locate at the binding energy of ∼577 and ∼586 eV, respectively. There are three kinds of chromium ions such as bivalent Cr(ii), trivalent Cr(iii), and hexavalent Cr(vi) in the CaO–SiO2–FeO–MgO–CrOx slag. Cr(iii) is the dominant valence state, and more than 77.99% Cr is trivalent Cr(iii). The fraction of Cr(ii)/Cr is in the range of 11.24–17.22%. The fraction of Cr(vi)/Cr is below 4.80%. The fraction of Cr(ii)/Cr decreases with increasing slag basicity, Cr2O3 content, temperature, or oxygen pressure log(PO2), while the fraction of Cr(iii)/Cr increases with increasing basicity, Cr2O3 content, temperature, or oxygen pressure. The trend of change is opposite. Low log(PO2), high Cr2O3 content, and high temperature are beneficial to reduce the toxic hexavalent Cr(vi). The slag basicity has little influence on the fraction of Cr(vi)/Cr.


Introduction
Steel slag is a by-product of steelmaking industry. In the steelmaking process, the amount of steel slag is about 10%-15% of the crude steel output [1,2]. Since the domestic steel slag comprehensive utilization level is lower than the developed countries [3,4], although the yield of Chinese crude steel is huge, still a large amount of steel slag is untreated and only stockpiled for disposal [5]. Some steel slags such as stainless steel slag, electric arc furnace steel slag, and chromium-containing hot metal steelmaking slag contain a certain amount of chromium oxides. It is possible that part of the chromium may exist in the form of toxic hexavalent chromium Cr(VI) in steel slag. Steel slag stockpiled in the open-air slag yard would pollute air, soil, and groundwater. The chromium element may exist in the form of bivalent Cr(II), trivalent Cr(III), and hexavalent Cr(VI) in silicate melts or metallurgical slag [6,7]. Thus, the chromium oxides can be expressed as CrO x . Therefore, study on the valence state of chromium in the metallurgical slag and its influencing factors has a very important significance for the clarification of the existing forms of chromium-containing phases and prevention of the pollution of toxic hexavalent chromium.
Considering the testing cost and the accuracy of testing results, a simple method for rapid detection of chromium valence states in slag is still sparse. Some studies also investigated the chromium valence states through thermodynamic calculation. Xiao and Mirzayousef-Jadid et al. [8,9] calculated the chromium valence states in the CaO-SiO 2 -CrO x system. It was found that both divalent and trivalent chromium coexist in these slags. Low slag basicity and low oxygen potential lead to the increase in divalent chromium oxide, instead of trivalent chromium oxide. Pretorius et al. [10] calculated the chromium valence states in CrO x -SiO 2 , CaO-CrO x -SiO 2 , and CaO-SiO 2 -Al 2 O 3 -CrO x systems and found that the valence states of chromium were mostly affected by the oxygen partial pressure. In recent years, X-ray photoelectron spectroscopy (XPS) was applied to valence state detection in metallurgical slag [11][12][13]. XPS can carry out not only qualitative analysis of the valence states of elements in metallurgical slag but also quantitative analysis.
In this article, the research object is the CaO-SiO 2 -FeO-MgO-CrO x system, which is the main component of chromium-containing steel slag. The slag samples were prepared under different experimental conditions and XPS was used to determine the chromium valence states [14,15] in the slag through qualitative analysis and semiquantitative analysis. Through the investigation of influence factors affecting chromium valence states, we expect to provide a theoretical reference for reducing the toxic hexavalent chromium.

Materials
The chemical compositions of chromium-containing steel slag used in the present study are given in Table 1

Methods
A horizontal tube MoSi 2 furnace (SGL-1700; Shanghai Jvjing Precision Instrument Manufacturing Co., Ltd) with a proportional-integral-differential controller was used to prepare the slag samples. Fine powders of the raw materials were carefully weighed and well-mixed together in an agate mortar. A 40 g mixture was charged into a MgO crucible. The MgO crucibles containing the slag samples were positioned inside a square alumina holder. The alumina holder was pushed into the eventemperature zone of the tube MoSi 2 furnace.
According to the equilibrium of C-O reaction and the process conditions of steelmaking end point, the calculated oxygen potential log(PO 2 ) values are around −5.6 to −3.0. Therefore, the designed log(PO 2 ) values in the present work are −5.0 to −3.5 and are listed in Table 1. Gas mixtures of CO-CO 2 were employed to obtain the targeted oxygen potential. The gas compositions of CO-CO 2 at equilibrium states were gained by equation (1) [16]. The calculated volume ratios of CO/CO 2 are listed in Table 2. (1) In view of the sensitivity of chromium to oxygen, the experiments were started by evacuating and filling the furnace chamber with purified argon before the CO-CO 2 gas mixture was introduced. Figure 1 shows the temperature control curves in experiments. When the target temperature of each experiment was reached, the samples were held for 60 min in the furnace chamber. After experiments, the MgO  crucibles were quickly pulled out of the reaction tube and quenched in liquid nitrogen. The slag samples were separated from the crucibles and milled to below 0.074µm, then subjected to XPS analyses. The XPS measurements on the slag sample prepared under various conditions were carried out by means of a spectrometer (Thermo Scientific Escalab 250Xi). In the testing, Al K Alpha was used as the source gun type, and the energy type size was 0.1 eV. The measurements were conducted with a spot size of 500 µm. In order to ensure the experimental results, the vacuum degree was controlled below 5 × 10 −9 mbar.
3 Results and discussion 3.1 Chromium spectra Figure 2 illustrates the typical XPS wide scan spectrum of the CaO-SiO 2 -FeO-MgO-CrO x slag. The auger peaks of the constituent elements in the slag were marked on the spectrum. Two minor peaks around the binding energies of 573-593 eV represent Cr 2p. A major peak at a binding energy of ∼284.6 eV [17] represents C 1s, which appeared due to the contamination of hydrocarbon.
For qualitative and quantitative analyses, the valence state of chromium element in the slag samples and the broad peak covering several peaks were analyzed. The Cr 2p spectra were divided into Cr 2p 1/2 and Cr 2p 3/2 by spin-obit interaction. As shown in Figure 3, the peaks of Cr 2p 3/2 and Cr 2p 1/2 locate at the binding energy of ∼577 and ∼586 eV, respectively. The broad peak was deconvoluted into several separate peaks to determine the individual area of each peak. Generally, it can be deconvoluted into three individual valence states, i.e., Cr(II), Cr(III), and Cr(VI). The proportions of above three kinds of ions in the slag samples were deduced from the area under the computer-resolved peaks. As can be seen in Figure 3, Cr(III) is the main existing form of Cr element, Cr(II) comes second, while only trace amount of Cr(VI) in the slag. The result is consistent with that of Mittal [18]. The reason for the existence of Cr(II) and Cr(VI) is that the trivalent chromium can be reduced by the FeO in the slag and oxidized by the small amount of oxygen in the atmosphere.

Effect of log(PO 2 ) on the valence state of chromium
In view of the absence of Cr(II) and Cr(VI) in the starting materials, the redox reactions of Cr(III) can be described as follows:

(5)
Therefore, x can be calculated by the following equation: The value of x can be obtained by relating Cr(III)/ Cr(II) and Cr(VI)/Cr(II) to the oxygen pressure using the law of mass action. As it is very difficult to detect the fractions of Cr(II)/Cr, Cr(III)/Cr, and Cr(VI)/Cr, we applied the ion and molecule coexistence theory [19][20][21] to obtain these fractions. The calculation process is detailed in one of our previous study [22]. Figure 4 shows both the calculated x and experimental x in the present study. It can be found that the x in CrO x increases with increasing oxygen pressure log (PO 2 ). The result is consistent with Jadid's finding in the CaO-SiO 2 -CrO x slag and the CaO-SiO 2 -Al 2 O 3 -CrO x slag [16]. Statistics of the various valence states of chromium in the CaO-SiO 2 -FeO-MgO-CrO x slag are shown in Figure 5. The fraction of Cr(II)/Cr decreases with increasing oxygen pressure log(PO 2 ), while Cr(III)/Cr and Cr(VI)/Cr increase. However, the increase tendency of Cr(VI)/Cr is not obvious, and the fraction of Cr(VI)/Cr is always below 4.5% under the present experimental conditions.   with increasing CaO content [23,24]. Meanwhile, it is also found that the fraction of Cr(VI)/Cr is almost constant with the increase in slag basicity.

Effect of Cr 2 O 3 on the valence state of chromium
Compared with ordinary steel slag, the chromiumcontaining steel slag contains a higher Cr 2 O 3 content. Therefore, it is necessary to make clear whether the increase in Cr 2 O 3 content has an effect on the valence distribution of chromium. The influence of Cr 2 O 3 content on the fractions of Cr(II)/Cr, Cr(III)/Cr, and Cr(VI)/Cr in the slag can be seen in Figure 8. 3.5 Effect of temperature on the valence state of chromium   temperature, or oxygen pressure. The trend of change is just the opposite.
(3) Low log(PO 2 ), high Cr 2 O 3 content, and high temperature are beneficial to reduce the toxic hexavalent Cr(VI). The slag basicity has little influence on the fraction of Cr(VI)/Cr.