3. Results and discussion
In order to determine the influence of the parameters of the technological process of tin oxide film fabrication on their phase composition the X-ray diffraction analysis was performed.
The XRD spectra of the SnO/SnO2−δ /SnO2 samples deposited in argon and argon–oxygen atmosphere and annealed in air at different temperature are shown in Fig. 1.
Clear characteristic peaks that can be assigned to the (101), (110), (002), (200) and (112) planes of the tetragonal structure of the SnO phase were observed for the samples deposited in argon atmosphere and annealed at 300°C during the 2nd stage of the heat treatment procedure. The thermal oxidation process of SnO to the SnO2 phase with a rutile structure was observed when the annealing temperature increased. As one can see from Fig. 1(a), for the samples annealed at 375 and 450°C, besides characteristic peaks to the SnO phase, peaks that can be assigned to the (101), (211) and (002) planes of the rutile structure of SnO2 are clearly seen. The diffraction peaks that can be assigned to the Sn phase were observed for all samples deposited in argon atmosphere and annealed in the temperature range 300–450°C due to incomplete tin oxidation.
The average SnO2 crystallite size was estimated from the full width at half maximum (FWHM) intensity of the observed peaks using the Debye– Scherrer equation in the form [21]
where K is the shape factor which is usually equal to 0.9, λ is the radiation wavelength of CuKα equal to 0.1542 nm, θ is the Bragg diffraction angle, and β is the full width at half maximum (FWHM) of the diffraction peak.
Estimation of the crystallite sizes for the films deposited in argon atmosphere gives values within the range 10–15 nm. The rise of the crystallite sizes with the annealing temperature was unessential.
A wide band between 20 and 40° without any distinctive characteristic peaks was observed on the diffraction spectra of the samples deposited by reactive DC sputtering (with oxygen content of around 1 vol.%) and annealed at 300°C during the 2nd stage of the heat treatment procedure assuming thus the formation of amorphous tin oxide films (Fig. 1(b)). Annealing at 375 and 450°C induces the formation of a polycrystalline SnO2 phase. The diffraction peaks observed around 26.6°, 33.9°, 38° and 52° can be assigned to the (110), (101), (200) and (211) planes of the tetragonal rutile structure of SnO2. The estimation of the crystallite sizes for these films gives values of about 5–7 nm.
When the oxygen content during reactive magnetron sputtering was increased to the value of around 2 vol.%, only amorphous tin oxide films were synthesized at all annealing temperatures. As a result, only wide bands are visible in the diffraction spectra of these samples shown in Fig. 1(c).
Thus, as a result of the analysis of the XRD spectra of the SnO/SnO2−δ/SnO2 samples, we can conclude: i) tin oxide films of a variable stoichiometric composition with a predominant tin monoxide phase were fabricated by means of DC magnetron sputtering in the argon plasma followed by a 2-stage annealing process; ii) reactive magnetron sputtering with a small content of oxygen (about 1 vol.%) in the argon–oxygen plasma leads to the preferable formation of more thermodynamically stable tin dioxide nanocrystalline films with a tetragonal rutile-type crystalline structure; iii) increasing of oxygen content during the process of reactive magnetron sputtering up to the value of about 2 vol.% induces the formation of amorhous tin oxide films after following a heat treatment procedure regardless of the annealing temperature. It should be noted that the crystalline structure of tin oxide films with a different stoichiometric composition can be improved by means of additional high-temperature annealing in vacuum or inert atmosphere [22].
Assuming that the oxygen vacancies at the interface between different tin oxide phases can work as paramagnetic trapping centres for electrons [23], we used EPR spectroscopy to look for the signal of paramagnetic oxygen vacancies (OV) in the SnO/SnO2-δ /SnO2 samples. This method is known as a sensitive tool for the characterization of the paramagnetic centres on various oxide surfaces [24]. The X-band CW EPR spectra of the SnO/SnO2−δ /SnO2 samples deposited in argon and argon–oxygen atmosphere and annealed in air at different temperatures are shown in Fig. 2. The obtained spectra are typical of the Mn2+ centres [25, 26]. In addition, each spectrum demonstrates an intense line at about 340 mT (g-factor~2). In order to study the origin of this central line, we performed experiments on the substrate without tin oxide. Figure 2(b) represents the EPR signal from the substrate (red online) and the signal from tin oxide on the substrate (black). The spectrum of the empty substrate demonstrates the same EPR signals as observed for the tin oxide samples indicating absence of the OV EPR signals from the tin oxide films.
For the additional characterization of a crystalline structure of the tin oxide samples with a different phase and stoichiometric composition Raman scattering spectroscopy was employed. This method is known as a sensitive tool for the characterization of the local structure, and the crystal surface area of solids and nanomaterials [27]. Information about crystallinity, orientation, composition, presence of structural defects, mechanical stress and doping of materials can be obtained from the Raman spectra. Characterization of tin oxide based structures by Raman spectroscopy is usually performed in order to get detailed information about their crystalline structure, phase and stoichiometric composition, presence of different types of oxygen vacancies etc. [28].
The Raman spectra of the samples synthesized with different content of oxygen during magnetron sputtering and various post-deposition annealing temperature registered with λ = 532 and 473 nm laser excitation sources are shown in Figs. 3 and 4, respectively.
For the samples deposited in argon atmosphere the peaks corresponding to the vibration modes inherent both in SnO and SnO2 crystalline structure were observed. For example, peaks at 111 and 750 cm−1 in the vicinity of the classical vibration modes B1g and B2g for single crystal SnO2 were observed for all annealing temperatures [2]. Besides that, for the samples deposited in argon atmosphere the peak at 692 cm−1 that can be attributed to the vibration mode A2u of longitudinal optical phonons was detected [2]. The peak at 211 cm−1 corresponds to the Sn–O vibrations in the tin oxide phase [29]. The intensity of this peak was decreased for the samples annealed at 375°C and was unessential for the samples annealed at 450°C. On the other hand, the amplitudes of inherent in SnO2 structure peaks near 692 and 750 cm−1 increased as the annealing temperature was raised. These modifications of the Raman spectra can be explained by oxidation of tin oxide and formation of the tin dioxide phase.
In contrast to the Raman spectra measured for the samples synthesized using magnetron sputtering in the argon plasma, no any distinctive peaks inherent either in SnO or in SnO2 crystalline structure were observed for the films fabricated by reactive magnetron sputtering. Only two broad bands in the range of about 50–200 and 400–700 cm−1 were detected for these samples. Existence of broad bands instead of narrow lines in Raman spectra is reported for nanocrystalline tin oxide samples and their attribution is closely connected with the grain size [30]. For example, a broad band in the low-frequency region observed for powders with the particles smaller than 7 nm is attributed to acoustic modes associated with the vibration of an individual nanoparticle as a whole [30]. Broadening of the classical modes for monocrystalline tin dioxide and formation of a broad band in the high-frequency range is activated by surface disorder for nanosized grains [30]. Influence of the nanocrystaline sizes is not the only reason for modifications of the Raman spectra of tin dioxide samples in a high-frequency range. Shift and broadening of the classical vibrations modes, the appearance of additional peaks and shoulders on the Raman spectra of nonstoichiometric tin dioxide samples are attributed to the existence of different types (subbridging, bridging, in-plane) of oxygen vacancies [8]. The existence of the peaks near 140 and 170 cm−1 for the samples obtained by reactive magnetron sputtering can be explained by formation of phases Sn2O3 and Sn3O4 [4].
Differences in the intensity of the Raman spectra of the same samples measured at different excitation wavelengths can be explained by the resonant Raman scattering effect. It is known that in the case when the excitation wavelength coincides with the energy of electronic transition or photon absorption in the material, strong enhancement of the Raman active vibrational modes is observed [31]. The band gap width of our samples obtained by deposition in argon is in a range of 1.9–2.5 eV in dependence on post-deposition annealing temperature. According to the X-ray diffraction analysis the tin oxide phase prevails in these samples (a band gap of tin monoxide is in a range of 2.5–3 eV [32]). Nonstoichiometric tin oxide films SnO/SnO2−δ /SnO2 obtained by reactive magnetron sputtering are characterized by a band gap in a range of 2.6–3.2 and 2.7–2.9 eV for the samples deposited in the argon–oxygen plasma with 1 and 2% oxygen content, respectively. The band gap for our samples was calculated from the results of the measurements of their adsorption spectra [20]. The deviation of the band gap from the value typical of monocrystalline tin dioxide is induced by nanocrystallinity and existence of defects (including oxygen vacancies) in our films. The dependence of the band gap width of the SnO/SnO2−δ /SnO2 samples on the synthesis conditions is presented in Table 1.
| Sputtering method and oxygen content | Second stage annealing temperature, °C | Band gap, eV |
|---|---|---|
| DC magnetron sputtering in argon plasma | 300 | 1.9 |
| 375 | 2.4 | |
| 450 | 2.5 | |
| Reactive magnetron sputtering (1 vol.% oxygen content) | 300 | 2.6 |
| 375 | 3.3 | |
| 450 | 3.2 | |
| Reactive magnetron sputtering (2 vol.% oxygen content) | 300 | 2.7 |
| 375 | 2.8 | |
| 450 | 2.9 |
For the measurements of the Raman spectra of our films we used lasers with an excitation wavelength of λ = 532 and 473 nm corresponding to the energy of 2.33 and 2.62 eV, respectively. Therefore the intensity of the Raman active modes for the samples obtained under the same synthesis conditions is higher in the case when we used a laser with a λ = 473 nm excitation wavelength. However, the samples characterized approximately by the same band gap have strongly different Raman spectra. For example, the films deposited in the argon plasma and annealed at 450°C have a band gap of 2.5 eV and are characterized by the presence in the Raman spectra of vibration modes inherent in SnO2. Meanwhile, for the films deposited in the argon–oxygen plasma with 1% oxygen content plasma and annealed at 300°C with a band gap of 2.6 eV we did not observe in the Raman spectra the vibration modes inherent in SnO2. According to the X-ray diffraction analysis both tin monoxide and tin dioxide phases are present in the films obtained by deposition in the argon plasma and annealed at 450°C. The tin monoxide phase is absent for the films deposited in the argon–oxygen plasma with 1% oxygen content and annealed at 300°C.
We can conclude that the presence of the tin monoxide phase in our samples is important for the fulfilment of resonant Raman scattering conditions. Increasing of the intensity peaks inherent in tin dioxide vibrational modes can be attributed to the dissipative transition of electronic excitation from tin monoxide atoms to tin dioxide atoms as an intermediate state in light scattering.