Abstract

Quantitative analysis of silicon tetrachloride, carbon disulfide, and dichloroethane concentrations to obtain vapor-liquid equilibrium data of the SiCl4-CS2 and SiCl4-C2H4Cl2 binary systems was established by Raman spectroscopy. The cheap glass sampling pipe was used as a carrier for Raman spectroscopy measurements. The Raman peak height of the internal standard was used to remove interference factors such as sampling pipe diameter, temperature, laser power, and other effects from the instrument. The peak height ratio between the Raman characteristic peak of the analyte and that of the internal standard was proportional to the analyte concentration. During the measuring process of vapor-liquid equilibrium data for the SiCl4-C2H4Cl2 binary system, the linear equation of y= 0.0068 + 0.75x with R2 of 0.9939 was used for the determination of SiCl4 concentration at the 422 cm−1 band. The linear equation of y= 0.0019 + 0.2266x with R2 of 0.9966 was used for the determination of C2H4Cl2 concentration at the 754 cm−1 band. For the SiCl4-CS2 binary system, the linear equation of y= 0.0494 + 4.7535x with R2 of 0.9962 was used for the determination of SiCl4 concentration at the 422 cm−1 band. The linear equation of y= 0.8139 + 8.7366x with R2 of 0.9973 was used for the determination of CS2 concentration at the 654 cm−1 band. The concentration of standard samples calculated by these standard curves was compared with the actual value to verify the accuracy of this method. The reproducibility is good when determining silicon tetrachloride and dichloroethane concentrations for the SiCl4-C2H4Cl2 binary system, with RSEP values of 2.81% and 2.17%, respectively. Meanwhile, the RSEP values are 3.55% and 4.16%, respectively, when determining silicon tetrachloride and carbon disulfide concentrations for the SiCl4-CS2 binary system.

1. Introduction

Titanium-containing liquid waste coming from the rectifying process is the major waste in the titanium metallurgy industry. Generally speaking, the concentration of SiCl4 and TiCl4 in the titanium-containing liquid waste is 15∼20% and 60∼70%, respectively [1]. The liquid waste containing TiCl4 and SiCl4 is a valuable resource, but it is hydrolyzed easily in a moist environment, producing a large amount of HCl, which is a toxic gas [24]. Therefore, improper disposal of titanium-containing liquid waste would lead to waste of resources and destruction of the eco-environment [5, 6]. Resource utilization of the titanium-containing liquid waste is always of great importance in China, and effort has been made to extract SiCl4 from the liquid waste for a long time [79]. Besides SiCl4 and TiCl4, there are also a small amount of impurities in the liquid waste such as FeCl3, MgCl2, CS2, C2H4Cl2, and so on. In order to extract purified SiCl4, these impurities must be removed. The method to obtain high-grade SiCl4 includes adsorption, rectification, partial hydrolysis, and photochlorination [1013]. Among all these methods, rectification may be the best way to extract purified SiCl4 from the liquid waste because no additional impurities are introduced. Unfortunately, due to the little difference in SiCl4 and CS2 or C2H4Cl2 in boiling point, the rectification process is energy-sucking. In order to reduce energy consumption, the accurate vapor-liquid equilibrium data for the SiCl4-CS2 and SiCl4-C2H4Cl2 binary systems must be measured to optimize the design of the rectifying column. During the measurement, the content of SiCl4, CS2, and C2H4Cl2 in the sample must be analyzed to obtain accurate gas-liquid equilibrium data.

Generally, inorganic substances could be detected by inductively coupled plasma-atomic emission spectrometry (ICP-AES) by dissolving in acid solution. Organic compounds could be analyzed by gas or liquid chromatography and infrared (IR) spectroscopy [1417]. However, these methods could not be used for simultaneous analysis of inorganic and organic compounds. Recently, quantitative analysis of components in solution by Raman spectroscopy has been reported, and there are several advantages to using Raman spectra for quantitative analysis [1822]. For example, it is a nondestructive technology that is usually applicable without any sample preparation process, which is convenient for online quantitative determination [23, 24]. This analytical method may also be applied to samples containing SiCl4, CS2, and C2H4Cl2. Creighton and Sinclair studied the Raman spectra of liquid SiCl4 and CS2 as early as 1973 [25], Gong studied the Raman scattering from liquid CS2 for different concentrations of benzene [26], and Steffen gave information about the 2D Raman response of liquid CS2 [27]. Meanwhile, the spectral information of C2H4Cl2 could also be found in literature [2830], but there is nearly no study on the quantitative determination of CS2, C2H4Cl2, and SiCl4 in liquid samples by using Raman spectra.

Therefore, the purpose of the present study is to establish a rapid and accurate method to determine the concentration of CS2, C2H4Cl2, and SiCl4 simultaneously in liquid samples. During the analysis procedure, the standard solutions with internal standards were sealed in cheap glass capillaries for the Raman spectrometer. Then, the peak height ratio between the Raman characteristic peak of the analyte and that of the internal standard was calculated to establish the calibration curves of CS2, C2H4Cl2, and SiCl4. After that, the Raman spectrometries of standard samples were measured, and the concentration of the analyte was calculated using the standard curves to confirm the availability of this method.

2. Materials and Methods

2.1. Raman Spectrometer

Raman spectrometry of all samples was detected by the LabRAM HR Evolution spectrometer, equipped with an Olympus BX41 microscope with a 10X objective lens (NA = 0.25), which was used for focusing the laser beam on the sample. A 532 nm Ar ion laser (Spectra Physics 2017) was used for excitation with power of ∼50 mW. The back scattering light was passed into a monochromator and detected with a charge coupled device (CCD) detector. The Raman spectrum between 100 and 3200 cm−1 was collected with an exposure time of 30 s and 2 accumulations for each spectrum.

2.2. Sample Preparation for Raman Analysis

The sample containing SiCl4 is easily hydrolyzed to release HCl in the air, which is not convenient for Raman detecting. Therefore, the sample containing SiCl4 will be sealed before it was submitted for Raman detecting according to our previous work [31]. The liquid sample stored in a sealed colorimetric tube was first moved into a drying oven. The colorimetric tube was then opened, and a little glass sampling pipe was inserted into the colorimetric tube. After the liquid sample entered the little glass sampling pipe by capillary action, the little glass sampling pipe was taken out and moved to the alcohol lamp. The both ends of the glass sampling pipe were heated and melted to allow the liquid sample to get sealed in it. The preparation process is convenient, and the material is inexpensive.

2.3. Preparation of Standard Samples

SiCl4, C2H4Cl2, and CS2 were purchased from Coron Chemical Industry Co., Ltd. There is often little water remaining in the purchased C2H4Cl2 and CS2, which may lead to the hydrolysis of SiCl4. Hence, the purchased C2H4Cl2 and CS2 were dehydrated first. The dehydration process is also similar to our previous work [31], just with a different distillation temperature: firstly, anhydrous calcium chloride (stored in a vacuum drying oven at 105°C for 24 h before use) and C2H4Cl2 or CS2 were put into a desiccative conical flask with a stopper. Then, the mixture was magnetically stirred at room temperature for 12 h. Finally, the filtrate from the mixture was distilled at certain temperature (90°C for C2H4Cl2 and 50°C for CS2), and the distillate was collected for the preparation of standard samples. In the measurement of vapor-liquid equilibrium data for the SiCl4-CS2 and SiCl4-C2H4Cl2 binary systems, the concentrations of SiCl4, CS2, and C2H4Cl2 in liquid samples should be analyzed simultaneously. Thus, CS2 was used as an internal standard during the concentration determination of SiCl4 and C2H4Cl2 for the SiCl4-C2H4Cl2 binary system, while C2H4Cl2 was used as an internal standard during the concentration determination of SiCl4 and CS2 for the SiCl4-CS2 binary system. Then, the working standard solutions at several concentration levels of SiCl4, C2H4Cl2, and CS2 were prepared.

2.4. Establishment of Calibration Curves

Working standard solutions with different concentrations of SiCl4, C2H4Cl2, and CS2 were prepared for Raman detecting. The baselines of the obtained Raman spectra were taken out by using Origin 8.5 to avoid noise and fluorescence effects. After that, the characteristic peak heights of SiCl4, C2H4Cl2, and CS2 were extracted from the revised Raman spectra, respectively. The peak height ratio between the Raman characteristic peak of the analyte and the internal standard was subsequently calculated according to the following equation:where IA is the Raman intensity of analyte and IS is the Raman intensity of internal standard. The calibration curve was generated by plotting the characteristic peak height ratio against the concentration of analytes.

3. Results and Discussion

3.1. Raman Spectra of the Mixed Solution Containing SiCl4, C2H4Cl2, and CS2

In order to obtain the characteristic peaks of SiCl4, C2H4Cl2, and CS2 for quantitative analysis, the Raman spectra of a mixed solution containing SiCl4, C2H4Cl2, and CS2 are detected, as shown in Figure 1. Figure 1(a) shows the full spectrum of the mixed solution from 200 cm−1 to 3200 cm−1. The full spectrum is partly displayed in Figures 1(b) and 1(c) so that the characteristic peaks of SiCl4, C2H4Cl2, and CS2 can be clearly observed. Figure 1(b) shows the spectrum of the mixed solution from 200 cm−1 to 1000 cm−1 band, and Figure 1(c) shows the spectrum from 1000 cm−1 to 1600 cm−1 band. In order to rapidly and conveniently find the Raman band of SiCl4, C2H4Cl2, and CS2, the details of the Raman peaks in Figure 1 are shown in Table 1. The C-S vibrating mode of carbon disulfide [25] is found at the 654 cm−1 band in Figure 1(b), and the intensity reaches up to 5368. Two obvious bands with the maximum intensity at 422 and 480 cm−1 appeared, respectively, in Figure 1(b), which is attributed to the Si-Cl vibrating modes of silicon tetrachloride [32]. The vibrating mode of liquid dichloroethane is found in Figures 1(a)–1(c). The band at 2968 cm−1 shown in Figure 1(a) belongs to the symmetrical stretching vibration of C-H2. The bands at 300 and 754 cm−1 in Figure 1(b) are ascribed to deformation vibration of C-C-Cl and stretching vibration of C-Cl, respectively [33]. The peak at 1053 cm−1 band in Figure 1(c) is assigned to stretching vibration of C-C, and the peaks at 1204, 1301, and 1431 cm−1 band are due to the twisting vibration, wagging vibration, and scissoring vibration of C-H2, respectively [33].

In order to reduce the background noise interference, a characteristic peak with high intensity should be used for quantitative analysis. Thus, the characteristic peak of dichloroethane at 300, 1053, 1204, 1301, and 1431 cm−1 with weak intensity could be excluded during the analysis while the characteristic bands of dichloroethane at 754 and 2968 cm−1 with high intensity could be used for quantitative analysis. Meanwhile, C-S vibrating mode of carbon disulfide at 654 cm−1 band and Si-Cl vibrating modes of silicon tetrachloride at 422 and 480 cm−1 band are clearly observed with high intensity. Therefore, these characteristic bands may be used for quantitative determination.

3.2. Standard Curves for the SiCl4-C2H4Cl2 Binary System
3.2.1. Raman Spectra of the Standard Solutions for the SiCl4-C2H4Cl2 Binary System

In order to obtain the standard curves for the SiCl4-C2H4Cl2 binary system, the Raman spectrums of the standard solutions with different SiCl4 and C2H4Cl2 concentrations were detected when carbon disulfide was added as an internal standard, and the results are shown in Figure 2. The Raman spectrum around the characteristic peak of carbon disulfide at 654 cm−1 in the standard solutions is shown in Figure 2(a), which could be used to calculate the peak height ratio for the standard curve. It can be seen from Figure 2(a) that the intensity of carbon disulfide at 654 cm−1 is not changed with the variation of silicon tetrachloride and dichloroethane concentration. The slight increase or decrease in intensity may be due to the change in laser power.

The Raman spectra of standard solutions on the characteristic peak of silicon tetrachloride at 422 and 480 cm−1 are presented in Figure 2(b). The Raman spectra of standard solutions around the characteristic peak of dichloroethane at 754 and 2968 cm−1 are shown in Figures 2(c) and 2(d), respectively.

It can be seen from Figure 2(b) that the Raman intensity of silicon tetrachloride at 422 cm−1 increases with the rising SiCl4 concentration in standard solutions, while the Raman intensity of silicon tetrachloride at 480 cm−1 is irregular with the SiCl4 concentration. Thus, the characteristic peak of silicon tetrachloride at 480 cm−1 is not suitable for quantitative analysis. In addition, although the Raman intensity of silicon tetrachloride at 422 cm−1 increases with the increase in SiCl4 concentration, the linear relationship between Raman intensity and SiCl4 concentration is not good. It can be verified from Figure 3 that the equation is y= 32.74 + 23883.45x with R2 of 0.9719, where y and x denote the Raman intensity and the mole fraction of silicon tetrachloride, respectively. This indicates a bad linear relationship. Therefore, Raman absolute intensity cannot be used to calculate the SiCl4 concentration accurately.

3.2.2. Standard Curves for the SiCl4-C2H4Cl2 Binary System

Generally, the Raman intensity of SiCl4 at characteristic peak (422 cm−1)indicated as IA can be given by the following equation [34]:where CA is the SiCl4 concentration, PL is the density of laser power, and KA is the Raman signal constant of SiCl4 at 422 cm−1, which can be affected by instrumental throughput and the apparent Raman-scattering efficiency. V is the volume of sample illuminated by the laser and viewed by the spectrometer. According to equation (2), the Raman intensity of SiCl4 (IA) will change with the variation of KA and PL. There is often inevitable fluctuation of laser power, test temperature, and instrument during the measurement of the Raman spectrum. Thus, it is easy to understand the bad linearity between the Raman intensity and the mole fraction of SiCl4. In order to remove the effect of laser power, test temperature, and instrument, the characteristic peak height ratio was introduced:where RA is the peak height ratio between SiCl4 and the internal standard CS2, CIS is the concentration of CS2, which is considered as invariable in all standard solutions, and KIS is the Raman signal constant of CS2 at 654 cm−1. Thus, the characteristic peak height ratio is proportional to the concentration of SiCl4, which is verified in Figure 3. The result in Figure 3 indicates that the linearity between the mole fraction of SiCl4 and the Raman peak height ratio at 422 cm−1 is good, and the equation is y= 0.0068 + 0.75x with R2 of 0.9939, where y and x denote the Raman peak height ratio and the mole fraction of SiCl4, respectively. Therefore, the characteristic peak height ratio will be used for the following standard curves.

The Raman intensity of C2H4Cl2 at 754 and 2968 cm−1 also increases with the increase in C2H4Cl2 concentration in standard solutions, which could be clearly observed in Figures 2(c) and 2(d). It can be seen from Figure 2(d) that the characteristic peak of dichloroethane at 2968 cm−1 divides into two peaks, so the mean Raman intensity of the two peaks was used for quantitative determination of dichloroethane. Similarly, the characteristic peak height ratio at 754 and 2968 cm−1 was used to establish a standard curve for the determination of C2H4Cl2 concentration, and the result is presented in Figure 4. It can be seen from Figure 4(a) that the linearity between the mole fraction of dichloroethane and the Raman peak height ratio at 754 cm−1 is good, and the equation is y= 0.0019 + 0.2266x with R2 of 0.9966, where y and x denote the Raman peak height ratio and the mole fraction of dichloroethane, respectively. The result in Figure 4(b) indicates that the characteristic peak height ratio at 2968 cm−1 is also proportional to the concentration of dichloroethane, and the equation is y= 0.00046 + 0.2334x with R2 of 0.9995.

3.2.3. The Veracity of the Standard Curves

To verify the accuracy of these standard curves, the Raman spectra of standard samples with known concentrations of silicon tetrachloride and dichloroethane were measured. The peak height ratios calculated from the spectra were obtained and are shown by hollow squares in Figures 3 and 4. These hollow squares lie on the standard curve, though the measurement condition and target concentration are different from the standard curve. These results suggest that the characteristic peak height ratio method is credible for the determination of silicon tetrachloride and dichloroethane.

To compare the different characteristic peaks applied, the relative standard error of prediction, RSEP, is calculated according to the following equation:where CA is the actual concentration of silicon tetrachloride or dichloroethane, C is the concentration calculated from the Raman spectrum, and n is the number of samples. The result in Table 2 shows that the RSEP value for the concentration of SiCl4 is 2.81% when the 422 cm−1 band is used. The RSEP value for the concentration of C2H4Cl2 is 2.17% at 754 cm−1 band and 7.32% at 2968 cm−1 band. This means the band of dichloroethane at 754 cm−1 may be the better choice for quantitative analysis.

3.3. Standard Curves for the SiCl4-CS2 Binary System

Similarly, C2H4Cl2 was used as an internal standard during the quantitative determination of SiCl4 and CS2 for the SiCl4-CS2 binary system. The Raman spectrum of the standard solutions around the characteristic peak of C2H4Cl2 (754 cm−1) is shown in Figure 5, which could be used to calculate peak height ratio. The Raman spectra of standard solutions around the characteristic peak of silicon tetrachloride and carbon disulfide (422 cm−1 and 654 cm−1, respectively) are presented in Figures 6 and 7. First, the characteristic peak heights of silicon tetrachloride, carbon disulfide, and dichloroethane were taken from these Raman spectrums, which are shown in Figures 57. Then, the peak height ratio was calculated according to the peak height of the solute and the internal standard. Finally, the peak height ratio was plotted against the mole fraction of silicon tetrachloride and carbon disulfide, which is also shown in Figures 6 and 7.

The results in Figure 6 show that the linearity between the mole fraction of silicon tetrachloride and the Raman peak height ratio at 422 cm−1 is good, and the equation is y= 0.0494 + 4.7535x with R2 of 0.9962, where y and x denote the Raman peak height ratio and the mole fraction of silicon tetrachloride, respectively. Similarly, there is also a good linearity between the mole fraction of carbon disulfide and the Raman peak height ratio at 654 cm−1 The equation is y= 0.8139 + 8.7366x with R2 of 0.9973, which is shown in Figure 7. In order to verify the accuracy of these standard curves, the Raman spectra of standard samples with known concentrations of silicon tetrachloride and carbon disulfide were also measured. The peak height ratios calculated from these spectra are shown by hollow squares in Figures 6 and 7. These hollow squares lie on the standard curve, which also confirms that the characteristic peak height ratio method is credible for the determination of silicon tetrachloride and carbon disulfide concentrations for the SiCl4-CS2 binary system.

The relative standard error of prediction (RSEP) of the standard curve for the SiCl4-CS2 binary system is also calculated according to equation (4), and the results are shown in Table 3. The result in Table 3 shows that the RSEP value of the SiCl4 concentration for the SiCl4-CS2 binary system is 3.55% when the peak height ratio at the 422 cm−1 band is used. The RSEP value of the CS2 concentration for the SiCl4-CS2 binary system is 4.16% at the 654 cm−1 band. These results also indicate that these standard curves are reliable during the determination of SiCl4 and CS2 concentrations for the SiCl4-CS2 binary system.

4. Conclusion

Raman spectroscopy was applied to quantitatively determine the concentration of silicon tetrachloride, carbon disulfide, and dichloroethane during the resource utilization of the titanium-containing liquid waste. The liquid sample was sealed in a glass sampling pipe with an internal standard, and the whole preparation process is convenient and inexpensive. The peak height ratio between the Raman intensities of the target substance and those of the internal standard is proportional to the concentration of the target substance with a good linear relationship. This method shows good accuracy when the standard samples are detected and the calculated concentration is compared with the actual concentration. The relative standard error of prediction (RSEP) is 2.81% and 2.17%, respectively, when determining silicon tetrachloride and dichloroethane concentrations for the SiCl4-C2H4Cl2 binary system. The RSEP values are 3.55% and 4.16%, respectively, when determining silicon tetrachloride and carbon disulfide concentrations for the SiCl4-CS2 binary system. These results indicate that Raman spectroscopy can quantitatively determine the concentration of silicon tetrachloride, carbon disulfide, and dichloroethane when measuring vapor-liquid equilibrium data for the SiCl4-CS2 and SiCl4-C2H4Cl2 binary systems.

Data Availability

All the data are included in the manuscript.

Conflicts of Interest

The authors declare that they have no conflicts of interest.

Acknowledgments

This study was financially supported by the National Natural Science Foundation of China (grant nos. 51604055 and 51674057), the National Science Foundation for Post-doctoral Scientists of China (grant no. 2018M643409), and the “China Scholarship Council” fellowship (grant no. CSC201802075009).