**3. Results**

#### *3.1. TiO2 Raman Spectra at 488.0 nm*

The Stokes Raman spectrum of anatase powder, recorded at room temperature, using a laser at 488.0 nm with an input power of 1.72 mW, is reported in Figure 2. The spectrum clearly shows an intense peak centered at 143 cm<sup>−</sup><sup>1</sup> and four peaks, at 197, 397, 515 and 640 cm<sup>−</sup>1, with lower intensity.

Anatase crystals are characterized by 15 optical modes at the Γ point of the Brillouin zone, described with the following irreducible representation of the normal vibrational modes [28,35]:

$$1A\_{1\mathcal{K}} + 1A\_{2\mathcal{u}} + 2B\_{1\mathcal{K}} + 1B\_{2\mathcal{u}} + 3E\_{\mathcal{K}} + 2E\_{\mathcal{U}}$$

Among these, only six modes, *<sup>A</sup>*1*g*, <sup>2</sup>*B*1*g* and 3*Eg* are the Raman-active ones (reported in Figure 2 and in Table 1); the Raman spectrum shows only five peaks since the *<sup>B</sup>*1*g* and *<sup>A</sup>*1*g* modes, at 512 and 518 cm<sup>−</sup><sup>1</sup> respectively, are not distinguishable at the experimental conditions, due to their intrinsic amplitude [28,36].


**Table 1.** Experimental and literature [28] TiO2 anatase Raman-active modes, excited at 514.5 nm.

The experimental frequencies, reported in the central column of Table 1, corrected by using the cyclohexane frequencies as calibration frequency, are comparable to data reported in literature [28].

The Stokes (positive Raman shift) and anti-Stokes (negative Raman shift) Raman spectra are reported in Figure 3, where it is possible to observe also the zooms of the low intensity peaks. By observing Stokes and anti-Stokes data, it turned out that the best peak for temperature monitoring is the Eg mode at 143 cm<sup>−</sup>1. It is well defined, very intense even at low laser powers and highly sensitive to temperature, as will be demonstrated later, thanks to its low Raman shift.

#### *3.2. Raman Spectra of TiO2 as A Function of Temperature, Excitation Wavelength and Input Power*

Stokes and anti-Stokes Raman spectra of anatase have been collected in the temperature range of 283–323 K, by exciting at 488.0, 514.5, 568.2 and 647.1 nm, using an input power of 1.47, 1.20, 2.20 and 5.86 mW, respectively. The Raman spectra collected at room tempera-

ture, at different excitation wavelengths, are reported in Figure 4a, while the 143 cm<sup>−</sup><sup>1</sup> Eg mode, excited at 514.5 nm, at different temperatures, is illustrated in Figure 4b.

**Figure 3.** Stokes (positive Raman shift) and anti-Stokes (negative Raman shift) Raman spectrum of TiO2 anatase recorded at 488.0 nm, with input power of 1.6 mW. In the right panel the zooms of the three less intense peaks are displayed, Stokes in the upper part, anti-Stokes in the lower one.

**Figure 4.** (**a**) Experimental anti-Stokes and Stokes spectra of TiO2 collected at room temperature and different excitation wavelengths (488.0 nm blue line, 514.5 nm green line, 568.2 nm yellow line and 647.1 nm red line); (**b**) Stokes and anti-Stokes spectra of the 143 cm<sup>−</sup><sup>1</sup> Eg mode, collected at 488.0 nm as function of temperature (from 283 to 323 K, with 5 K increment step).

All Raman spectra have been analyzed with Matlab, using a Lorentz fitting, to obtain the Raman spectrum parameters, such as frequency position, width, intensity and area of the peak (see Appendix A).

An example of the data, obtained from the analysis of the Raman spectra centered on the 143 cm<sup>−</sup><sup>1</sup> peak, by exciting at 514.5 nm at different temperatures, is reported in Table 2, where data from the Stokes spectra are reported together with the area of the anti-Stokes peaks, which allows to calculate the anti-Stokes/Stokes ratio, a parameter important for the determination of the local temperature. The errors of the peak positions and widths are experimental errors, directly derived from the characteristics of the experimental set-up; the errors on the area have been derived from the fitting (see the example in Appendix A); for the anti-Stokes/Stokes ratio data, the error propagation has been considered [12].


**Table 2.** Frequency position and width of the Stokes peak at 143 cm<sup>−</sup>1, Stokes and anti-Stokes areas and anti-Stokes/Stokes ratios for the temperature range 283–323 K, explored during the calibration procedure at 514.5 nm.

Corresponding to the increasing in temperature, both the peak frequency position and the anti-Stokes/Stokes ratio clearly show an increment, from 142.5 to 145 cm<sup>−</sup><sup>1</sup> and 0.48 and 0.54, respectively, while the peak width varies in between 10.7 and 11.4 cm<sup>−</sup>1.

In order to test whether the Raman spectra, and the corresponding parameters, are perturbed by the laser irradiation, it is also necessary to investigate the effect of increasing incident laser power by keeping constant all other variables, like excitation wavelength. The laser power was changed in a range between 0.1 and 18 mW (depending on the laser wavelength); the experimental results (peak position, peak width (FWHM, i.e., full width half maximum), peak area and anti-Stokes/ Stokes ratio), for the 143 cm<sup>−</sup><sup>1</sup> Eg mode, at 514.5 nm, are shown in Figure 5.

**Figure 5.** Peak position, peak width, anti-Stokes/ Stokes ratio and peak area for the 143 cm<sup>−</sup><sup>1</sup> Eg mode of anatase as function of the laser power incident on the sample; *λexc* = 514.5 nm. The dashed lines are only for eyes guidance. All data are reported against the laser Power (mW) in a logarithmic scale.

It is evident that parameters are not increasing with the laser power only in the small range, 0–2 mW, while at higher input power they increase with the increase of the laser power. In order to assume that the temperature of the sample is not influenced by the presence of the laser irradiating the sample, it is necessary to keep the laser power at values lower than 2 mW.

#### *3.3. Test Sample Measurement*

A validation of the method can be done by evaluating anti-Stokes and Stokes intensities in various positions of the *Test Sample*, thus verifying if the parameters measured all over the sample are uniform or not, and comparable with the results previously obtained. The calculated anti-Stokes/Stokes ratios are reported in Table 3.

**Table 3.** Repeated measurements of the anti-Stokes/Stokes ratio at room temperature collected at 488.0, 514.5, 568.2 and 647.1 nm using a laser power of 1–2 mW depending on the excitation wavelength used. In the final row, mean values of the anti-Stokes/Stokes ratios and the standard deviations are reported.

