Measurement of Silicon Nanoparticles Temperature by Raman Spectroscopy

. The temperature of silicon nanoparticles under laser photo-induced heating is determined from the ratio of the intensities of the Stokes and anti-Stokes components of the Raman scattering. The obtained results of the dependence of nanoparticles temperature on the laser radiation intensity and the temperature dependence of the Raman line position may be used to determine the optimal regimes of photo-hyperthermia enhanced by silicon nanoparticles for cancer therapy. © 2021 Journal of Biomedical Photonics & Engineering.


Introduction
Silicon nanoparticles (Si NPs), including NPs of crystalline and porous silicon, are widely researched for biomedicine applications, in particular, as active elements of biosensors and as part of new drugs for theranostics (simultaneous diagnosis and therapy) of oncological and other socially significant diseases [1][2][3][4][5][6]. If the size of crystalline NPs is less than 10 nm, they exhibit photoluminescence in the biological transparency window (650-850 nm), which can be used for their visualization in cells and tissues [7][8][9][10][11]. The large Si nanoparticles (more than 10 nm) are in turn capable of providing nonlinear biovisualization mechanisms, such as two-photon luminescence and second harmonic generation, the combination of which makes it possible to effectively track both individual silicon NPs and their aggregation in different parts of the cell, while the resolution of such approach is sufficient to obtain 3Dimages [12].
The simultaneous use of the diagnostic functionality of silicon nanoparticles and their therapeutic properties predicts the development of new methods of theranostics, in particular, the theranostics of cancer [13]. One of the fast growing areas of nanobiomedicine is the use of NPs as sensitizers of various therapies. In particular, silicon NPs are used in photohyperthermia techniques as a promising nanomaterial with biocompatible and biodegradable properties [14,15].
The development of hyperthermia techniques raises the problem of heating temperature determining. There are two main techniques to measure temperaturecontact and non-contact. Contact techniques are based on achieving a thermal equilibrium state of the measuring temperature transmitter (temperature sensors) with the studied object during their direct contact. The disadvantages of the contact technique are the distortion of the measured object temperature under introducing a thermometer into it; the temperature of the converter always differs from the actual temperature of the measured object; the temperature measurement range is limited by the properties of the temperature sensors materials. In addition, the problem of temperature measuring the in Si NPs cannot be solved by contact technique. Non-contact temperature measurements are carried out by pyrometers and thermal imagers. Existing non-contact thermal imaging thermometers are based on J of Biomedical Photonics & Eng 7 (1) 30 Mar 2021 © J-BPE 010303-2 measuring the infrared radiation intensity in a nontransparent region for living organisms [15,16].
In the present work, we selected the Raman scattering spectroscopy method to Si NPs temperature monitor. As is known, the intensity of the Stokes and anti-Stokes lines depends on the temperature as a consequence of the Bose-Einstein statistics for the phonon [17]. Furthermore, the Raman scattering lines frequency changes with temperature due to the anharmonicity of lattice oscillations. In particular, this is realized at sufficiently intense photoexcitation causing the sample to heat up [18,19]. We studied samples of nanocrystalline silicon (nc-Si) NPs as well as nanocrystalline mesoporous silicon (mPSi) films. The temperature change of the samples was achieved by changing the exciting laser radiation intensity. The experimental data obtained with nc-Si NPs and mPSi films confirmed the theoretical model developed in Ref. [19].

Experimental technique
Powders of nc-Si were prepared by plasma-chemical synthesis using a plasmatron and pure (99.9 %) microcrystalline Si [1]. Aqueous suspensions were obtained by adding nc-Si to deionized water to achieve a concentration of 1 g/L. In the experiment, 50 µl drop planted on a metal substrate and dried was used. Silicon NPs have a spherical shape and a wide size distribution in the range from 10 to 150 nm. The size distribution has major maxima at 50-120 nm [4]. An argon laser with lex = 514.5 nm and a maximum power of 0.2 W was used as an exciting radiation source. The laser beam was focused on nc-Si sample with 3 μm diameter.
MPSi samples were formed by electrochemical etching substrates of c-Si in an ethanol solution of hydrofluoric acid. The method consists in the formation of mesoporous silicon films with a thickness of 50 to 100 μm, a porosity of 50 to 80%, and a pore size of 2 to 5 nm on the surface of c-Si substrates p-type conductivity with a resistivity of 10 to 50 mOm × cm. The source of exciting radiation for the mPSi heating was a semiconductor laser with a wavelength lex = 633 nm and a maximum power of 70 mW.
The Raman spectra of nc-Si samples were taken on the DFS-52 spectrometer in 90-degree scattering geometry. The receiver served working in the photon counting mode a cooled photomultiplier tube PMT-79. The diffraction grating of the spectrometer was 1800 g/mm. The spectral resolution was 0.1 cm -1 . The mPSi samples were investigated on a Confotec MR-350 setup equipped with an optical microscope, which allowed us to obtain the micro-Raman spectra with a spatial resolution of the order of 1 μm and a spectral resolution of 1 cm -1 .

Results and discussion
The Stokes and anti-Stokes Raman spectra of nc-Si samples at different powers are shown in Fig. 1. With the increase of the excitation power level, a shift in the spectra to lower frequencies region is observed, which is caused by the increase of sample temperature during photoinduced heating [19]. Simultaneously with the increase of the excitation power, a change in the ratio between the Stokes and anti-Stokes scattering components intensities is observed. From the ratio of the Stokes and anti-Stokes components intensities, the NPs temperature can be estimated using the following formula [17,20]: where IS и IA are Stokes and anti-Stokes lines intensities, ω is phonon frequency, k = 1.38 × 10 −23 J/К, ℏ = 1.05 × 10 −34 J×s. Eq. (1) follows from the Bose-Einstein statistics for phonons, neglecting the difference in the absolute values of frequencies and absorption coefficients of the Raman components [17]. Fig. 2 shows the dependence of the IS / IA value on the laser radiation intensity, which is almost linear. The decrease of the IS / IA value with the increase of the excitation intensity is explained by the photoinduced heating of nc-Si. From Eq. (1), the nc-Si temperature in the excitation region can be estimated by the formula:   . It can be seen that the experimental error is maximum in the low temperatures region, which corresponds to a very weak photoinduced heating of nc-Si, while the error decreases with increasing temperature. This is occurring due to the increase of the absolute the Raman signals intensities and the decrease of the measurement error of the IS / IA ratio. It should be noted that with the increase of the signal accumulation time, the measurement error of the latter value can be reduced to 1%, which will give an accuracy of the temperature determining of the order of 10-15% in the temperature range of 310-320 K.
The dependence of the Raman line shift on the photoinduced heating temperature obtained from the analysis of the spectra in Fig. 1 is shown in Fig. 4. The Raman line temperature shift to the lower frequencies region is in good agreement with the literature data for crystalline and nanocrystalline silicon [18,19]. The observed linear temperature dependence is convenient to analyze the Si NPs photoinduced heating. Fig. 4 Raman shift of the nc-Si sample as a function of its temperature at lex = 514.5 nm.
The Stokes Raman spectra of mPSi samples at different powers are shown in Fig. 5. Fig. 6 shows the position of the Raman line maximum, taking into account the data in Fig. 4, as a function of the photoinduced heating temperature. It can be seen that, regardless of the type of nanostructures and the method of their preparation, an almost linear temperature dependence of the Raman peak position is observed. It should be noticed that the absolute positions of the Raman peak frequencies for nc-Si and PSi are 0.5-0.7 cm -1 less than those for the c-Si sample taken from the literature, which can be associated with both the accuracy of the Raman frequency determination and with the size effect for phonons in silicon nanocrystals [19].

Conclusions
The temperature effect of laser radiation on the Raman spectra of nc-Si and mPSi is studied. Measuring the ratio of the Stokes and anti-Stokes components intensities allows us to determine the temperature with a minimum error at temperatures above 350-400 K. It is found that with temperature increase during photoexcitation, the Raman line frequency decreases linearly, which makes it possible to use this dependence to accurately determine the temperature even at low heating. The results obtained can be used to determine the optimal regimes of photo-hyperthermia upon sensitization with silicon nanocrystals in cancer therapy. The optimal regime of hyperthermia in general, and photo-hyperthermia in particular, is defined by the advisable local temperature, which must be achieved to obtain a therapeutic effect in the selected area, on the one hand, and not leading to irreversible changes in healthy tissues, on the other hand. The temperature shift of the Raman spectrum components observed in silicon nanocrystals allows us to determine the photoinduced heating value with high precision, thus selecting the required radiation power individually in each case.

Disclosures
All authors declare that there is no conflict of interests in this paper.