X-ray diffraction measurements were carried out on all prepared samples. The spectra did not show any characteristic Bragg peaks for silicon oxide, indicating the amorphous structure of this oxide. Figure 1 contains a sample of the XRD spectra that we obtained, which is related to the sample annealed at 1200°C (sample F). In all measured XRD spectra, the sharp peak is related to Si (111) (CSM card no. 65-1060).
Figure 2 shows the (1000–1300) cm− 1 band in FT-IR spectra of the prepared samples (detailed characterization of the spectra was covered in a previous work [21]). This broad and intense band is attributed to asymmetric and symmetric stretching vibrations of Si-O-Si bonds [22]. The shoulder at 1190 cm− 1 appears due to the splitting of longitudinal optical and transverse optical stretching motions [23]. The presence of this shoulder is an indication that each asymmetric broad peak noticed in the range 1000–1300 cm− 1 in FTIR spectra may be due to the overlapping of two peaks.
In order to investigate the overlapping peaks, each band in the region 1000–1300 cm-1 was deconvoluted into two Gaussian-Lorentzian line shapes (peak1 and peak2). Figure 3 shows example of the deconvolution processes that concern the case of sample F that treated at 1200 ○C. The peak1 is attributed to asymmetric and symmetric stretching vibrations of Si-O-Si bonds and the peak2 is due to the splitting of longitudinal optical and transverse optical stretching vibrations.
Figure 4 shows the intensities of the peaks 1 and 2 as functions of annealing temperature. We notice that, as the annealing temperature increases, the peak2 intensity increases at the expense of peak1 intensity with the tendency for the two peaks to have the same intensity at higher temperatures.
On the other hand, Fig. 5 shows that the ratio of the intensities of the two peaks (2:1) reaches its maxima at the annealing temperature 1100 ○C (sample E). In our previous work, we found that, at this annealing temperature the highest concentration of crystal defects is obtained. The decrease in the ratio value observed in the case of the sample annealed at 1200 ○C may be due to the regression of the splitting process.
Figures 6 and 7 show the position as a function of the annealing temperature for peaks 1 and 2. We notice that the position of peak2 is more affected by temperature than the position of peak1. In the case of peak1 (Fig. 6), we notice that when the temperature increases within the range 800–1100 ○C, the peak shifts towards the higher wave numbers. This behavior can be attributed to the changes in the structural composition of the surrounding environment because of silicon oxidation. However, the peak1 shifts back toward lower wavelengths upon annealing at T = 1200 ○C (sample E). This is explained by the decrease in the intensity of the forces acting on the bonds because of the regression of the splitting process.
In the case of peak1 (Fig. 7), we notice that, when annealing at 900 ○C, the shift towards higher wave numbers reaches its maximum due to the enhancement of the splitting process. As the temperature is raised above 900 ○C, the peak shifts towards lower wave numbers due to the attractive forces to which the Si-O-Si bonds are exposed. Upon annealing at T = 1200 ○C (sample E), the peak2 shifts back toward higher wavelengths due to the decrease in the intensity of the attractive forces because of the regression of the splitting process.
Figure 8 shows the peaks 1 and 2 width as functions of annealing temperature. It can be seen that while the width of peak1 does not appear to be significantly affected by the change in the annealing temperature, the width of peak2 is clearly affected by the changes in the structural composition of the surrounding environment imposed by the change in the oxidation temperature.
In our previous work [21], we found that thermal oxidation affects the intensity of the silicon peak in the XRD spectra. Here, we study the effect of oxidation on another parameter, the silicon peak position, by investigating the relationship between the ratio I2/I1 and the silicon XRD peak position (Fig. 9).
We notice that, as the ratio increases, the silicon peak shifts to the right and that the shift peaks in the case of sample F (annealed at 1200 ○C). This result is evidence that peak2 growth causes defects in the silicon crystal structure.
This conclusion can be supported by optical reflectivity measurements, were in our previous work [21], we observed the formation of a plasma edge of silicon nanoparticles in the reflectivity spectra in the range 400–750 nm (Fig. 11). The plasma edge is not appearing in the spectra of the samples Aand B. The formation of silicon nanoparticles is evidence of crystal defects [21].
We notice that, the plasma edge is proportional to the silicon peak position. The relationship between them is linear except for the point corresponding to the sample F at which the regression of the splitting process happened. This result supports the relationship between the peak shift and the formation of crystal defects, which we deduced from Fig. 9. This is because the shift of the plasma edge towards short wavelengths is associated with an increase in the concentration of conductive electrons [24,25]. These electrons belong to silicon nanoparticles resulting from breaking the long-range arrangement of the silicon lattice as a result of the formation of crystal defects [23].