Inﬂuence of the Curing and Annealing Temperatures on the Properties of Solution Processed Tin Oxide Thin Films

: We report the effect of the curing (T curing ) and annealing (T anneal ) temperatures on the structural, electrical, and optical properties of solution processed tin oxide. T anneal was varied from 300 to 500 ◦ C, and T curing from 200 ◦ C to T anneal . All T anneal lead to a polycrystalline phase, but the amorphous phase was observed at T anneal = 300 ◦ C and T curing ranging from 250 to 300 ◦ C. This could be explained by the melting point of the precursor (SnCl 2 ), occurring at 250 ◦ C. The crystallinity can be effectively controlled by the annealing temperature, but the curing temperature dramatically affects the grain size. We can reach grain sizes from 5–10 nm (T curing = 200 ◦ C and T anneal = 300 ◦ C) to 30–50 nm (T curing = 500 ◦ C and T anneal = 500 ◦ C). At a ﬁxed T anneal , Hall mobilities, carrier concentration, and conductivity increased with the curing temperature. The Hall mobility was in the range of 1 to 9.4 cm 2 /Vs, the carrier concentration was 10 18 to 10 19 cm − 3 , and the conductivity could reach ~20 S/cm when the grain size was 30–50 nm. The optical transmittance, the optical bandgap, the refractive index, and the extinction coefﬁcient were also analyzed and they show a correlation with the annealing process.

For large area electronics, vacuum processing is usually preferred due to the high yield production. In vacuum processing, the substrate temperature, the partial pressure, and/or the total pressure are processing parameters before the annealing step [22][23][24]. However, vacuum processing requires high investments and maintenance costs. Non-vacuum processing like spin-coating [6,25], inkjet printing [26] have demonstrated competitive results with the vacuum processing materials and devices.
In the solution processing, like spin-coating, thin film formation follows the following steps: the preparation/fabrication of the precursor solution, the coating followed by the curing (steps (repeated until a target thickness is obtained [6]), and finally the annealing. Yet, if the annealing temperature [25] has usually an influence on the physical, chemical, electrical, and optical properties, there are too few investigations of the impact of the curing temperature on the properties of the thin films.
Various studies have demonstrated a substantial correlation between the grain size and the electrical, the optical properties of SnO 2, and the maximum process temperature.
The increase in electrical conductivity, carrier concentration, and mobility is understood to be correlated to the increase in crystallinity and crystal size of SnO 2 [27]. The crystal size also depends on the film thickness [28].
S.-S. Lin et al. showed the effect of the substrate temperature when fabricating SnO 2 thin films by sputtering. At a substrate temperature of 400 • C, the carrier concentration was 3.3 × 10 18 cm −3 , the mobility was 7.26 cm 2 /Vs and the resistivity was 2.58 × 10 −1 Ωcm. A substrate temperature of 500 • C leads to a carrier concentration of 1.7 × 10 18 cm −3 , an electrical resistivity of 6.78 × 10 −2 Ωcm and a mobility of 54.3 cm 2 /Vs. The films had a preferred (101) orientation and an optical bandgap of 3.5 eV [29].
Hydrophilic porous SnO 2 :H 2 O thin films deposited by the sequential ionic layer adsorption and reaction (SILAR) method demonstrated a high roughness (over 200 nm) with crystals of~15 nm [30] and a high resistivity (10 7 Ωcm at room temperature). On the other hand, it is possible to obtain films with mobility of 8.6 cm 2 /Vs after annealing at 500 • C from SnCl 2 ·H 2 O precursors [31].
G.K. Deyu et al. showed that SnO 2 films made by spray pyrolysis and using SnCl 4 ·H 2 O precursors could lead to films with mobility of~23 cm 2 /Vs and a carrier concentration of~10 20 cm −3 . They could achieve grains over 142 ± 27 nm [32]. S.-S. Lin et al. used SnCl 4 as precursors for SnO 2 and reported thin films with a grain size of~31.8 nm. They used spray pyrolysis at a substrate temperature of 400 • C. The preferred crystal orientation was (211), but the electrical properties were not reported. The films had a band gap of 4 eV [33]. The nebulized spray pyrolysis method using a precursor solution prepared with SnCl 4 ·H 2 O at a substrate temperature of 450 • C lead to SnO 2 thin films with a (110) preferred orientation, crystal sizes of~38 nm, a Hall mobility of~18 cm 2 /Vs, a carrier concentration of 5 × 10 19 cm −3 , and an optical bandgap of 3.75 eV [34]. Spray pyrolysis at 430 • C of solutions made with SnCl 4 lead to SnO 2 films with a (211) preferred orientation. The films demonstrated a carrier concentration of 2.343 × 10 18 cm −3 and a mobility of 57 cm 2 /Vs [35]. The SnCl 2 ·H 2 O precursor led to ultrasonic sprayed pyrolyzed SnO 2 that had a (110) preferred orientation [36].
For perovskite solar cells, plasma assisted atomic layer deposition of SnO 2 led to crystallite sizes of~25 nm, a carrier concentration in the 10 19 cm −3 range, and mobilities of 35 cm 2 /Vs [20]. SnO 2 fabricated by PEALD had crystals in the~50 nm range size, bandgaps of 3.95 eV, a Hall mobility of 27 cm 2 /Vs, and a carrier concentration of 4 × 10 19 cm −3 [24].
Doping with F has been a successful strategy to obtain low resistivity. With F doping of 15%, Peale et al. reported solution processed thin films with resistivity of~1 mΩcm, a mobility~10-15 cm 2 /Vs, and a carrier concentration above 10 20 cm −3 [41]. We note also the preferred (200) orientation of the films. Without doping, their films had a resistivity of~0.6 Ωcm. Using tin hydroxide nitrate based precursors, Nadarajah et al. reported undoped SnO 2 thin films crystallizing at 300 • C, but the films did not show any electrical properties even at higher annealing temperatures. F incorporation was necessary to reach sufficiently high carrier concentrations [42]. When SnO 2 is doped with 10% of F, an annealing temperature of 350 • C leads to the resistivity in the mΩcm range, a carrier concentration of 4 × 10 19 cm −3, and a mobility of~1 cm 2 /Vs [42].
In reported TFTs, we note that the (110) preferred orientation [7] leads to TFTs with mobilities of 40 cm 2 /Vs. The preferred (101) leads to ambipolar behavior [10], while the oxygen flow rate could control the p-or n-type character of tin oxide [11]. High field-effect mobility of 147 cm 2 /Vs was obtained with the (110) preferred orientation [9]. SnCl 4 ·H 2 O used as the precursors for SnO 2 TFT and led to crystals of 10.5 nm in size, an optical band gap of 3.66 eV, and a (110) preferred crystal orientation. The annealing temperature was 300 • C. The TFTs demonstrated a mobility of~5.96 cm 2 /Vs [8].
From our previous study on amorphous tin oxide [6], we wanted to know if there was a range in the process temperature to obtain an amorphous phase. Also, as in the case of IGZO, it was important to compare the properties of the amorphous phase and the polycrystalline phase for further applications. In this study, we demonstrate the role of the curing (T curing ) and the annealing (T anneal ) temperatures on the optical (transmittance and optical bandgap), electrical (Hall mobility, carrier concentration, and conductivity), and physical properties (crystallinity, grain size, and morphology) of SnO 2 thin films made by spin-coating. We show that the curing step has a high impact on the final properties at a fixed annealing temperature. Also, we focus on the thin film properties and leave the device fabrication and analysis for another study.

Materials and Methods
We fabricated the 0.2 M SnO 2 precursor solution by mixing SnCl 2 with acetonitrile and ethylene glycol (35% and 65% in volume). The solution was then stirred overnight. The glass substrates were cleaned with acetone, and ultrasonication in isopropanol for 10 min, and dried under flowing N 2 . The solution was then spin-coated at 2000 rpm, cured at 100 • C for 5 min, and again cured at a T curing for 5 min. The coating and curing steps were repeated once. The sample was then annealed in air at T anneal for 2 h. When T anneal was 300 • C, T curing was 200, 220, 250, 280, or 300 • C. When T anneal was 400 • C, T curing was 200, 300, or 400 • C. When T anneal was 500 • C, T curing was 200, 300, 400, or 500 • C.
We evaluated the surface morphology by scanning electron microscopy (SEM) with a Hitachi S-4700. The surface roughness was evaluated by atomic force measurement (AFM) with a Park System Xe-7 with the tapping method on a 1 µm × 1 µm surface. The optical properties (transmittance, n, and k) and the thicknesses of the thin films were evaluated with a Jobin YVON -Uvisel ellipsometer in the 1.5-5 eV range. The Hall effect was measured with an Ecopia HMS-3000. At least 10 points were evaluated for averaging. We used the Van der Pauw geometry for the measurements. We measured the crystallinity by X-ray diffraction (XRD) by using the Cu Kα radiation at a wavelength of 1.54 Å. Thermogravimetric analysis was measured with a TG-DTA STD Q600 under N 2 atmosphere. The temperature was raised at a rate of 10 • C/min. Figure 1 shows the thermogravimetric and the differential thermal analysis of the precursor SnCl 2 . Three routes were analyzed: when the temperature increased (a) once from room temperature (RT) to 500 • C, (b) from RT to 220 • C, then decrease to RT, then increase to 300 • C, (c) from RT to 280 • C, then decrease to RT, then increase to 300 • C. In route (a) the weight percentage decreases monolithically down to 20% [43], while the temperature difference shows an endothermic peak at 254 • C. The peak shows the melting of SnCl 2 . As shown in route (b), we also observe the melting related peak at 253 • C. On the other hand, when following route (c), the peak at 253 • C is still observed, but on the cooling, we observe an exothermic peak by 225 • C, which we understand demonstrates the recrystallization of the powder. Then, when increasing to 300 • C, the endothermic peak appears at a slightly lower temperature of 251 • C. We note that up to 400 • C, the weight variation is less than 10%.   36.93, and 51.93°, respectively. We observe that for Tanneal at 300 °C, a Tcuring of 200 or 220 °C would lead to the polycrystalline SnO2, while when Tcuring is in the range 250-300 °C, an amorphous phase is formed (see Figure 2b). We previously reported the amorphous phase when Tcuring was 280 °C and Tannealing was 300 °C [6]. Therefore, there is a Tcuring range to obtain the amorphous phase of SnO2. On the other hand, independent of Tcuring, annealing at 400 (Figure 2c), or 500 °C (Figure 2d), polycrystalline SnO2 is formed. We note that at Tcuring = Tanneal = 500 °C, we observe a (110) preferred orientation, which was reported to lead to high mobility TFTs [9]. The presence of a process window leading to an amorphous phase for a polycrystalline material has been previously observed. Tin oxyhydroxide led to an amorphous tin oxide phase for annealing temperature between 500 and 700 K [44]. In the present study, the process window is for the Tcuring and not the Tanneal. Also, as introduced before, the preferred orientation depends on the fabrication process and the process temperature. We note that spray pyrolysis could lead to (211) orientation [33,35], or (110) [34,36]. Also, we note that the TFTs leading to the highest mobility had a (110) preferred orientation [7,9]. We extracted the crystalline size from the Scherrer equation:

Results and Discussions
where D is the crystalline size, the wavelength of the incident CuK line, is the full width at half maximum (FWHM), and is the Bragg angle. The crystallites are ~7 nm at Tanneal =300 °C, ~9 to ~17 nm at Tanneal = 400 °C, and ~8 to ~18 nm at Tanneal = 500 °C.   36.93, and 51.93 • , respectively. We observe that for T anneal at 300 • C, a T curing of 200 or 220 • C would lead to the polycrystalline SnO 2 , while when T curing is in the range 250-300 • C, an amorphous phase is formed (see Figure 2b). We previously reported the amorphous phase when T curing was 280 • C and T annealing was 300 • C [6]. Therefore, there is a T curing range to obtain the amorphous phase of SnO 2 . On the other hand, independent of T curing , annealing at 400 (Figure 2c), or 500 • C (Figure 2d), polycrystalline SnO 2 is formed. We note that at T curing = T anneal = 500 • C, we observe a (110) preferred orientation, which was reported to lead to high mobility TFTs [9]. The presence of a process window leading to an amorphous phase for a polycrystalline material has been previously observed. Tin oxyhydroxide led to an amorphous tin oxide phase for annealing temperature between 500 and 700 K [44]. In the present study, the process window is for the T curing and not the T anneal . Also, as introduced before, the preferred orientation depends on the fabrication process and the process temperature. We note that spray pyrolysis could lead to (211) orientation [33,35], or (110) [34,36]. Also, we note that the TFTs leading to the highest mobility had a (110) preferred orientation [7,9]. We extracted the crystalline size from the Scherrer equation: where D is the crystalline size, λ the wavelength of the incident CuKα line, β is the full width at half maximum (FWHM), and θ is the Bragg angle. The crystallites are~7 nm at T anneal =300 • C,~9 to~17 nm at T anneal = 400 • C, and~8 to~18 nm at T anneal = 500 • C.  Figure 3 shows the transmittance of the thin films. First, we observe that the thin film annealed at 300 °C and cured in the 250-300 °C range show a higher transmittance than their 200 and 220 °C annealed sample counterpart (see Figure 3a). At 550 nm, only the thin film with Tcuring = 200 °C and Tanneal = 300 °C has a transmittance of 78.10%, lower than 80% Second, at Tanneal of 400 ( Figure 3b) and 500 °C (Figure 3c), the transmittance is over 80%.  Figure 3 shows the transmittance of the thin films. First, we observe that the thin films annealed at 300 • C and cured in the 250-300 • C range show a higher transmittance than their 200 and 220 • C annealed sample counterpart (see Figure 3a). At 550 nm, only the thin film with T curing = 200 • C and T anneal = 300 • C has a transmittance of 78.10%, lower than 80%. Second, at T anneal of 400 ( Figure 3b) and 500 • C (Figure 3c), the transmittance is over 80%. Figure 3 shows the transmittance of the thin films. First, we observe that the thin films annealed at 300 °C and cured in the 250-300 °C range show a higher transmittance than their 200 and 220 °C annealed sample counterpart (see Figure 3a). At 550 nm, only the thin film with Tcuring = 200 °C and Tanneal = 300 °C has a transmittance of 78.10%, lower than 80%. Second, at Tanneal of 400 ( Figure 3b) and 500 °C (Figure 3c), the transmittance is over 80%. The inset of each figure shows the refractive index and the extinction coefficient as a function of the energy. The maximum refractive index in the visible region is 2 to 2.8 for polycrystalline SnO2, and above 3 for amorphous tin oxide. The refractive index taken at 1.95 eV (corresponding to a wavelength of 633 nm) [24] has a value of 1.5 to 1.7, which is smaller than the reference handbook value of 2.06 eV [45]. A lower value was previously attributed to lesser dense films [24]. We previously reported the density of a-SnOx to be 5.29 g/cm 3 [6], a value smaller than the handbook value of 6.45 g/cm 3 [24,45]. Relating to the TGA data, up to 400 °C the material mass loss is less than 10%, but at 492 °C, ~78% of the total mass is lost, denoting the poor reactivity of the material at a temperature below 400 °C [43,46].
The optical band gap (Eg) can be extracted by the Tauc plot [47][48][49]: where α is the absorption coefficient, h the Planck constant, ν the frequency of the incident light, A is a proportional coefficient, and n is an exponential coefficient. Usually, n = 2 (for possible direct transitions), and n = 1/2 (for possible indirect transitions) are used for the evaluation of the optical bandgaps [50,51]. Various methods exist to extract the absorption coefficient. One method consists of using the extinction coefficient k because [50,52]: where λ is the wavelength. Nevertheless, the energy at which the absorption increases, (i.e., when k becomes positive), can help identify the optical bandgap [53]. The extraction of the bandgap from the extinction coefficient has been used for semiconductors and insulators [46,[54][55][56][57]. So; we compared the values of the optical bandgap (Eg) extracted either at k = 0 or from the Tauc plot. The Tauc plots are shown in Figure 4. At Tanneal = 300 The inset of each figure shows the refractive index and the extinction coefficient as a function of the energy. The maximum refractive index in the visible region is 2 to 2.8 for polycrystalline SnO 2 , and above 3 for amorphous tin oxide. The refractive index taken at 1.95 eV (corresponding to a wavelength of 633 nm) [24] has a value of 1.5 to 1.7, which is smaller than the reference handbook value of 2.06 eV [45]. A lower value was previously attributed to lesser dense films [24]. We previously reported the density of a-SnO x to be 5.29 g/cm 3 [6], a value smaller than the handbook value of 6.45 g/cm 3 [24,45]. Relating to the TGA data, up to 400 • C the material mass loss is less than 10%, but at 492 • C,~78% of the total mass is lost, denoting the poor reactivity of the material at a temperature below 400 • C [43,46].
The optical band gap (E g ) can be extracted by the Tauc plot [47][48][49]: where α is the absorption coefficient, h the Planck constant, ν the frequency of the incident light, A is a proportional coefficient, and n is an exponential coefficient. Usually, n = 2 (for possible direct transitions), and n = 1/2 (for possible indirect transitions) are used for the evaluation of the optical bandgaps [50,51]. Various methods exist to extract the absorption coefficient. One method consists of using the extinction coefficient k because [50,52]: where λ is the wavelength. Nevertheless, the energy at which the absorption increases, (i.e., when k becomes positive), can help identify the optical bandgap [53]. The extraction of the bandgap from the extinction coefficient has been used for semiconductors and insulators [46,[54][55][56][57]. So; we compared the values of the optical bandgap (E g ) extracted either at k = 0 or from the Tauc plot. The Tauc plots are shown in Figure 4. At T anneal = 300 • C, the E g extracted from the k = 0 method (the Tauc plot) showed a value of~3.9 (~4.2) eV for the polycrystalline SnO 2 , while the amorphous tin oxide thin films have a smaller bandgap of 3.6 (3.78 to 3.91) eV. At T anneal = 400 • C, the extraction from the k = 0 method (the Tauc plot) led to values of E g of 4.25 (4.42), 4.4 (4.59), and 4.45 (4.69)eV, when T curing was 200, 300, and 400 • C, respectively. At T anneal = 500 • C, the values were 4.45 (4.64), 4.3 (4.56), 4.45 (4.62) and 4.4 (4.66)eV, when T curing was 200, 300, 400, and 500 • C, respectively. The k = 0 method leads to relatively smaller values of the bandgap. The difference lies in the extraction methodology: in the Tauc plot method, the value is extracted from a linear part of the (αhν) 2 plot, while the value extracted from the k = 0 method is taken at k = 0. The difference in the values has also been reported before [53]. We note that J. Gong et al. showed that the optical bandgaps increased from 1.2 to 3.7 eV when the annealing temperature increased from room temperature to 600 • C for thermally evaporated SnO 2 [58]. The difference in the values has also been reported before [53]. We note that J. Gong et al. showed that the optical bandgaps increased from 1.2 to 3.7 eV when the annealing temperature increased from room temperature to 600 °C for thermally evaporated SnO2 [58].  In all graphs, the black, red, blue, and green line represents a curing temperature of 200, 300, 400, and 500 °C, respectively. In (a) the yellow, pink, and grey line represents a curing temperature of 220, 250, and 280 °C, respectively. Figure 5 shows the results of the measurements of the electrical properties by the Hall effect. Figure 5a shows the variation of the carrier concentration as a function of the curing temperature at various annealing temperatures. At a fixed Tanneal, the carrier concentration increases with Tcuring. At Tanneal = 300 °C the carrier concentration was 3.88 ± 0.73 × 10 18   In all graphs, the black, red, blue, and green line represents a curing temperature of 200, 300, 400, and 500 • C, respectively. In (a) the yellow, pink, and grey line represents a curing temperature of 220, 250, and 280 • C, respectively. Figure 5 shows the results of the measurements of the electrical properties by the Hall effect. Figure 5a shows the variation of the carrier concentration as a function of the curing temperature at various annealing temperatures. At a fixed T anneal , the carrier concentration increases with T curing . At T anneal = 300 • C the carrier concentration was 3.88 ± 0.73 × 10 18 , 7.22 ± 0.42 × 10 18 , and 8.46 ± 0.30 × 10 18 cm −3 when T curing was 250, 280, 300 • C. At T anneal = 400 • C the carrier concentration was 4.05 ± 1.49 × 10 18 , 4.97 ± 0.85 × 10 18 , 5.46 ± 1.34 × 10 18 cm −3 when T curing was 200, 300, 400 • C At T anneal = 500 • C the carrier concentration was 5.02 ± 3.09 × 10 18 , 7.50 ± 0.31 × 10 18 , 9.38 ± 0.26 × 10 18 , and 14 ± 0.94 × 10 18 cm −3 when T curing was 200, 300, 400, and 500 • C. Similarly, the mobility and the conductivity increased with Tcuring at a fixed Tannealing as shown in Figure 5b,c. At Tanneal = 300 °C the mobility was 1.00 ± 0.12, 3.61 ± 0.19, and 5.02 ± 0.49 cm 2 /(Vs) when Tcuring was 250, 280, 300 °C, respectively. At Tanneal = 400 °C, the mobility was 1.08 ± 0.55, 1.96 ± 0.27, 4.49 ± 0.88 and cm 2 /(Vs) when Tcuring was 200, 300, 400 °C, respectively. At Tanneal = 500 °C, the mobility was 1.14 ± 0.69, 2.49 ± 0.16, 5.76 ± 0.19, 9.4 ± 0.1 cm 2 /(Vs) when Tcuring was 200, 300, 400, and 500 °C. At Tanneal = 300 °C, the conductivity was 4.07 ± 1.65, 4.19 ± 0.44, and 6.78 ± 0.44 Scm −1 when Tcuring was 250, 280, 300 °C. At Tanneal = 400 °C, the conductivity was 0.62 ± 0.23, 1.53 ± 0.01, and 3.36 ± 0.12 Scm −1 when Tcuring was 200, 300, 400 °C At Tanneal = 500 °C, the conductivity was 0.85 ± 0.17, 2.99 ± 0.19, 8.65 ± 0.09, and 21.10 ± 1.21 Scm −1 when Tcuring was 200, 300, 400, and 500 °C.
Interestingly, the amorphous SnO2 demonstrates higher electrical performances than its polycrystalline counterparts at a curing temperature of 300 °C, as was reported for IGZO [1], where the amorphous phase demonstrated higher Hall mobility than the crystalline phase. As presented in the introduction, the precursor and the method of fabrication have an impact on the electrical properties. Spray pyrolysis has been shown to provide high Hall mobilities (over 10 cm 2 /Vs) with a preferred (211) orientation [32,35]. Our results are similar to the SnO2 fabricated by the SILAR method [31]. We note nonetheless that our films are less than 50 nm, (compared to the >100 nm layers used by spray pyrolysis) and with thicker layers, the electrical conductivity may increase [28].
In Figure 6, we show the SEM and AFM images taken for Tcuring = Tanneal = 300 °C, (a and b), 400 °C (c and d), and 500 °C (e and f), respectively. All samples show some porosity, typical of solution processing [59]. The surface of the thin film annealed at 300 °C is amorphous, while the two others are polycrystalline. The surface of the 400 °C annealed film shows a better uniformity in grain size than the other polycrystalline film. Interestingly, on the surface of the Tcuring = Tanneal = 500 °C, several grains seem to reach ~50 nm, yet the majority of the grains are ~10 nm. However, as shown in the inset of Figure 6e, some of the ~50 nm grains are actually made of smaller grains being ~10 nm in size. This confirms the grain size value extracted by the calculation from the Scherrer equation. The AFM images confirm the amorphous (Figure 6b) and polycrystalline phase (Figure 6d,f) of SnO2. Both polycrystalline SnO2 show a variation in height, which could be evaluated with the peak to valley roughness (Rpv). They were 10.266, and 36.022 nm for the film annealed at 400, and 500 °C, respectively. The surface smoothness was evaluated with the RMS roughness (RRMS). The RRMS were 0.154, 1.020, and 1.566 nm for the films annealed at 300, 400, and 500 °C, respectively. Similarly, the mobility and the conductivity increased with T curing at a fixed T annealing as shown in Figure 5b,c. At T anneal = 300 • C the mobility was 1.00 ± 0.12, 3.61 ± 0.19, and 5.02 ± 0.49 cm 2 /(Vs) when T curing was 250, 280, 300 • C, respectively. At T anneal = 400 • C, the mobility was 1.08 ± 0.55, 1.96 ± 0.27, 4.49 ± 0.88 and cm 2 /(Vs) when T curing was 200, 300, 400 • C, respectively. At T anneal = 500 • C, the mobility was 1.14 ± 0.69, 2.49 ± 0.16, 5.76 ± 0. 19, 9.4 ± 0.1 cm 2 /(Vs) when T curing was 200, 300, 400, and 500 • C. At T anneal = 300 • C, the conductivity was 4.07 ± 1.65, 4.19 ± 0.44, and 6.78 ± 0.44 Scm −1 when T curing was 250, 280, 300 • C. At T anneal = 400 • C, the conductivity was 0.62 ± 0.23, 1.53 ± 0.01, and 3.36 ± 0.12 Scm −1 when T curing was 200, 300, 400 • C At T anneal = 500 • C, the conductivity was 0.85 ± 0.17, 2.99 ± 0.19, 8.65 ± 0.09, and 21.10 ± 1.21 Scm −1 when T curing was 200, 300, 400, and 500 • C.
Interestingly, the amorphous SnO 2 demonstrates higher electrical performances than its polycrystalline counterparts at a curing temperature of 300 • C, as was reported for IGZO [1], where the amorphous phase demonstrated higher Hall mobility than the crystalline phase. As presented in the introduction, the precursor and the method of fabrication have an impact on the electrical properties. Spray pyrolysis has been shown to provide high Hall mobilities (over 10 cm 2 /Vs) with a preferred (211) orientation [32,35]. Our results are similar to the SnO 2 fabricated by the SILAR method [31]. We note nonetheless that our films are less than 50 nm, (compared to the >100 nm layers used by spray pyrolysis) and with thicker layers, the electrical conductivity may increase [28].
In Figure 6, we show the SEM and AFM images taken for T curing = T anneal = 300 • C, (a and b), 400 • C (c and d), and 500 • C (e and f), respectively. All samples show some porosity, typical of solution processing [59]. The surface of the thin film annealed at 300 • C is amorphous, while the two others are polycrystalline. The surface of the 400 • C annealed film shows a better uniformity in grain size than the other polycrystalline film. Interestingly, on the surface of the T curing = T anneal = 500 • C, several grains seem to reach~50 nm, yet the majority of the grains are~10 nm. However, as shown in the inset of Figure 6e, some of the~50 nm grains are actually made of smaller grains being~10 nm in size. This confirms the grain size value extracted by the calculation from the Scherrer equation. The AFM images confirm the amorphous (Figure 6b) and polycrystalline phase (Figure 6d,f) of SnO 2 . Both polycrystalline SnO 2 show a variation in height, which could be evaluated with the peak to valley roughness (R pv ). They were 10.266, and 36.022 nm for the film annealed at 400, and 500 • C, respectively. The surface smoothness was evaluated with the RMS roughness (R RMS ). The R RMS were 0.154, 1.020, and 1.566 nm for the films annealed at 300, 400, and 500 • C, respectively.  Considering all conditions, the surface RMS roughness (R rms ) was between 0.154 and 1.566 nm. Interestingly, for the amorphous SnO 2 (T anneal = 300 • C and T curing from 250 • C to 300 • C), the peak to valley roughness (R pv ) was below 2.3 nm, while the polycrystalline SnO 2 had peak-to-valley from 4.9 to 36 nm. The various values are gathered in Table 1. We note that the grain sizes were in the range 5 to 10, 10 to 30, and 15 to 50 nm at T anneal of 300, 400, and 500 • C, respectively. Nevertheless, the AFM images could not reveal the morphology of the bigger grains. This could lead to overestimation in the size of the grains, and also, because of the difference in height (R pv ), there could also be a misinterpretation in the assessment of the various grain sizes. As a comparison, only the spray pyrolyzed SnO 2 at 420 • C could lead to a bigger grain size of 142 ± 27 nm, but had a (211) preferred orientation [32]. Therefore, we can understand that the smooth surfaces of the amorphous tin oxide thin films can explain in part the high mobility of the thin films, while the bigger grains can control the higher mobility and conductivity of the polycrystalline tin oxide.
Therefore, we understand that T curing is a dominant factor when annealing at 300 • C. When the T curing is smaller than the melting point, then there is no reorganization of the material, and the crystallization of SnO 2 is favored. When T curing is at least equal to the melting temperature, then the multiple melting and recrystallizations could favor a random disposition of the atoms further leading to the amorphous phase of SnO 2 . At higher T anneal , T curing influences all properties, and the higher the T curing , the higher the electrical properties and the grain sizes.

Conclusions
We report the impact of the curing and annealing temperature on the electrical, physical and optical properties of solution processed SnO 2 . We report that due to the melting point of~250 • C of SnCl 2 , SnO 2 can be amorphous when the curing temperature is between 250 and 300 • C and the anneal temperature was 300 • C. This is due to the melting point of the precursor SnCl 2 and the multiple curing steps, which favors the random disposition of the atoms. The amorphous phase had a mobility of 5.02 ± 0.49 cm 2 /(Vs), carrier concentration of 8.46 ± 0.30 × 10 18 cm −3 . SnO 2 crystal grain increases with increasing both T curing and T anneal , ranging from 5-10 nm for T curing = 200 • C and T anneal = 300 • C, to 35-50 nm when T curing = T anneal = 500 • C. Under that condition the grain preferred orientation is (110), the mobility is 9.4 ± 0.1 cm 2 /(Vs) and the carrier concentration is 14 ± 0.94 × 10 18 cm −3 . We compared the k = 0 and the Tauc plot extraction methods for the optical band gap. The former leads to smaller values than the latter. The bandgap of polycrystalline SnO 2 slightly increases with the annealing temperature (3.9-4.45 eV), whereas the amorphous phase is~3.6 eV in the aforementioned T curing range. Considering the impact of T curing , the present study can also give a path to making new precursors for low temperature processed amorphous oxide semiconductors.