A novel and potentially scalable CVD-based route towards SnO2:Mo thin films as transparent conducting oxides

In this report, we prepared transparent and conducting undoped and molybdenum-doped tin oxide (Mo–SnO2) thin films by aerosol-assisted chemical vapour deposition (AACVD). The relationship between the precursor concentration in the feed and in the resulting films was studied by energy-dispersive X-ray spectroscopy, suggesting that the efficiency of doping is quantitative and that this method could potentially impart exquisite control over dopant levels. All SnO2 films were in tetragonal structure as confirmed by powder X-ray diffraction measurements. X-ray photoelectron spectroscopy characterisation indicated for the first time that Mo ions were in mixed valence states of Mo(VI) and Mo(V) on the surface. Incorporation of Mo6+ resulted in the lowest resistivity of 7.3×10-3Ωcm\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$7.3 \times 10^{{ - 3}} \Omega \,{\text{cm}}$$\end{document}, compared to pure SnO2 films with resistivities of 4.30×10-2Ωcm\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$4.3\left( 0 \right) \times 10^{{ - 2}} \Omega \,{\text{cm}}$$\end{document}. Meanwhile, a high transmittance of 83% in the visible light range was also acquired. This work presents a comprehensive investigation into impact of Mo doping on SnO2 films synthesised by AACVD for the first time and establishes the potential for scalable deposition of SnO2:Mo thin films in TCO manufacturing.


GRAPHICAL ABSTRACT Introduction
Transparent conducting oxides (TCOs) are a type of material based on wide-bandgap metal oxide semiconductors that are simultaneously optically transparent in the visible-NIR light range and electrically conductive as thin films ðthickness \ 1lmÞ. They have been used in a large number of applications, such as flat-panel displays [1], light emitting diodes (LEDs) [2], solar cells [3,4] and as low-emittance window materials [5]. The most widely used TCO material both in the laboratory and industry is tindoped indium oxide (ITO), which has a resistivity as low as 1 À 2 Â 10 À4 X cm and a transmittance over 85% across the visible light range [6]. As electronics industry has been developing rapidly since the 1970s, the demand for ITO, in particular in flat-panel displays, has been increasing significantly [1]. Currently, ITO represents 60% of the transparent conductor market and approximately 60% of indium consumption globally [7]. However, the reserves of indium in the Earth's crust are relatively scarce and it is refractory to purify [8]. Therefore, alternatives to ITO with similar or even better optoelectronic performance must be found.
Semiconductor thin films based on tin oxide (SnO 2 ) have a large bandgap ([ 3.6 eV) that is higher than the visible transparency requirement for TCOs ([ 3.1 eV) and possess a resistivity value in the order of 10 À2 À 10 À3 X cm primarily caused by oxygen vacancies [9]. Incorporation of ions with a distinct valence state into host SnO 2 , i.e. aliovalent ion doping, is able to enhance the electrical conductivity via increasing the mobility and carrier concentration [10,11]. Commonly used doped variants of SnO 2 contain Fat the anion site which is a good candidate to replace ITO and possesses a resistivity of $ 4 Â 10 À4 X cm and an over 80% transmittance [12][13][14]. Additionally, antimony (Sb) with a resistivity of around 6 Â 10 À4 X cm and approximately 80% transmittance [15][16][17] and tantalum (Ta) [11,18] at the cation site can also be used as dopants to improve conductivity. Consequently, research on doped SnO 2 is crucial for discovering alternative TCO materials.
To date, a wide variety of synthetic pathways towards doped SnO 2 thin films have been reported including magnetron sputtering, [19] pulsed laser deposition (PLD) [20,21], spin coating [22][23][24], spray pyrolysis [15,25,26] and aerosol-assisted chemical vapour deposition (AACVD) [11,12,27]. AACVD is a variant of conventional CVD which depends on carrier gas transporting precursor aerosols to the deposition zone with elevated temperature. In this modified CVD technique, a precursor or precursors containing the requisite atoms to produce the solid-state materials is dissolved in a solvent and the solution is then placed above a vibrating piezoelectric plate to generate mist. The precursor-infused mist is carried by an inert gas to a reaction chamber containing substrates, heated at the required temperature for thin-film deposition. Volatility of precursor is not a requirement in AACVD, which considerably expands the choice range of precursors possible. This technique allows precise control over dopant stoichiometry in resultant products as dopant contents in products are directly proportional to the dopant concentration in the precursor solution [28]. Another advantage is that it simplifies the vaporisation and delivery of precursor solution via use of a nebuliser and carrier gas, as compared to conventional CVD, which depends on evaporating volatile precursors to transport precursors to the reaction zone. As a consequence, cost of the deposition process can be lowered [29,30]. However, good solubility of precursors in certain solvents becomes a vital requirement for this technique, which makes sure the generation of precursor mist. Solvent viscosity should be also considered whilst selecting organic solvents as average small-sized aerosols can be generated more easily from solvents with a lower viscosity; therefore, a successive and homogeneous supply of precursor droplets can be realised and hence improves film growth [31]. Nevertheless, AACVD has several advantages over common thin-film deposition techniques, such as an extensive range of possible precursors that can be utilised owing to the requirement for volatile precursors found in conventional CVD being removed, simple operation, relatively precise control over doping levels due to the bottom-up nature of the process and the potential for scale-up [32]. To be specific, when compared to deposition of thin films by spin coating, which involves a thermal treatment postdeposition for hours, there is no need of such a treatment in AACVD technique, which saves on both processing cost and time.
As a potential candidate of TCOs, research has been directed towards investigation of the effect of Mo on SnO 2 thin films, in particular focusing on the doping impact on optoelectronic performance [33][34][35][36]. In this paper, we now report the deposition of SnO 2 :Mo thin films via AACVD and the impact of Mo doping on crystal growth mechanics and optoelectronic properties of SnO 2 films prepared were systematically studied by a wide range of characterisation techniques including powder XRD, SEM, Raman spectroscopy and UV-Vis spectroscopy. To the best of the authors' knowledge, synthesis of SnO 2 :Mo thin films by AACVD and the effect of Mo doping on the film growth and optoelectronic performance have not been reported yet.

Synthesis of SnO 2 thin films by AACVD
Float glass substrates were cut into 1:5 cm Â 1:5 cm pieces and were ultrasonically cleansed in aqueous detergent solution, acetone, ethanol and then deionised water for 10 min each step. After cleaning, they were dried in air at room temperature and then put into an ozone cleaner to thoroughly remove organic residues on the surface. The experimental apparatus has been previously described in detail [37,38]. In a typical experiment, 1.2 mmol of total precursors (i.e. mol (butyltin trichloride) ? mol (ammonium orthomolybdate)) was dissolved in 20 mL methanol and the mixture was allowed to magnetically stir for around 15 min to obtain precursor solution with varying molar ratios of butyltin trichloride and ammonium orthomolybdate ð Mo MoþSn ¼ 0; 1; 1:5; 2; 2:5; 3; and 5 mol%Þ, respectively. A flask containing the precursor solution was placed over a piezoelectric nebuliser (the ultrasonic aerosol generator), and hence, aerosols generated were transported to a quartz tube where substrates were horizontally placed. The tube was housed in a tubular furnace. The deposition temperature was 450 C, the carrier gas flow rate was fixed at 220 cm 3 /min, and the reaction time was 50 min. After the reaction was complete, the gas flow was not cut out until the furnace temperature dropped below 100 C. The glass slides were then removed from the tube and stored in the ambient atmosphere for subsequent characterisations.
and Ka 2 ¼ 1:54443Å) with an intensity ratio of 2:1 at 40 kV and 40 mA. Lattice parameters (a, b and c) and unit cell volumes were calculated by the CELREF software using powder XRD data. Scanning electron microscopy (SEM) images were collected by using a FEI Quanta 250 electron microscope with secondary electron mode at an accelerating voltage of 15 kV. Energy-dispersive X-ray spectroscopy (EDX) was performed in the SEM at an accelerating voltage of 25 kV. Raman spectroscopy was performed on a Renishaw 1000 Raman system with an Ar-laser excitation wavelength of 514 nm, and spectra were acquired over the range of 200-900 cm -1 with a 50 Â objective at 50% laser power. XPS spectra were measured using a Thermo Scientific Kratos Axis Ultra instrument equipped with a monochromatic Al K a Xray source (hm = 1486.6 eV). Survey scans were collected in the range 0-1200 eV, and higher-resolution scans were recorded for the main core lines. A charge neutraliser was used to minimise charging and spectra are aligned on the binding energy scale relative to the hydrocarbon C-C/C-H peak at 284.8 eV. Spectra are fitted using Voigt-like peak shapes in the CasaXPS software. Spin-orbit splitting ratios and splitting energies are constrained to obtain physically meaningful fits. The sheet resistance was measured by an Ossila four-point probe station with 25 repeats at room temperature, thereby determining the resistivity q combining the thickness values collected by SEM cross-sectional images. Other electrical properties including mobility (l) and carrier concentration (n) were determined by a home-built Hall effect equipment in the Van der Pauw geometry. A magnetic field of 0.088 T and an input current of 1 mA were applied for the measurement. A Perkin Elmer Lambda 1050 UV-Vis-NIR spectrometer was employed to record transmittance profiles across a wavelength range of 300-1500 nm against a blank glass reference.

Results and discussion
Compositional analysis of thin films The elemental compositions of these SnO 2 :Mo coatings were analysed by energy-dispersive X-ray (EDX) spectroscopy. The observed Mo amount ( Mo MoþSn at%) in prepared films is 1.9, 2.1, 2.4, 2.7, 4.4 and 6.9 at.%, respectively. The comparison between theoretical amount and experimentally discovered amount of Mo is displayed in Fig. 1, and the original EDX spectroscopy with calculated component results by atomic percentage is shown in Figure S1 (see Supporting Information). The EDX spectroscopy results confirmed again the successful doping of Mo into SnO 2 films by AACVD. In the plot, an increasing trend and a quasi-linear relationship between theoretical and experimentally found contents of molybdenum are observed, which indicates that control over doping levels within SnO 2 :Mo films by this technique is possible. To date, such a precise manipulation upon doping levels for SnO 2 :Mo has not been achieved as far as we are aware.

Structural characterisation of thin films by diffraction
Undoped and Mo-doped SnO 2 thin films were characterised by powder XRD, and the resulting diffraction patterns are displayed in Fig. 2. All thin films analysed present the well-defined single-phase cassiterite SnO 2 with a tetragonal rutile structure, which is in good agreement with the standard card ICDD: 04-008-4169 (rutile, P4 2 /mnm).
It is worth noticing that no reflections of any secondary phases are observed in these patterns, which differs from previously reported SnO 2 films synthesised by AACVD, [39] where diffraction peaks belonging to Sn and SnO were observed. Hence, any crystalline impurities in our films lie below the limit of detection for pXRD (ca. \ 5% w/w). For the undoped thin films, a preferred orientation along the (110) plane is apparent and this preferential crystallite alignment has been reported a number of times previous to this work [14,40,41] and is believed to be the most stable crystallographic orientation due to the lowest formation energy [41,42]. At lower Mo at%, the relative intensities of the other major reflections corresponding to the (101), (200) and (211) planes are increased, suggesting a more random crystallite growth. The (200) plane is thought to be a less closely packed plane; namely, a low atomic density plane and a high intensity of the (200) plane correlate with a high concentration of oxygen vacancies [43]. As we further increase the dopant level, the thin films represent a preferred direction along the (110) plane, as per those doped with 2.7 and 6.9 at% Mo whilst the coatings of 4.4 at% dopant level exhibit a favoured orientation along the (101) plane, which has also been previously reported by Ponja et al. [12], Choi et al. [44] and Kwon et al. [45]. However, no reasonable explanations have been proposed yet. We suggest tentatively that it may be because crystal distortion occurs at high dopant levels, and to maintain crystallographic stability, SnO 2 nuclei will preferentially grow along the plane with a relatively low surface energy. It is generally believed that the (110) plane is of the lowest surface energy, i.e. the most stable crystal plane, and hence, films doped with 2.7 and 6.9 at% Mo grew along with this direction. The (101) plane also has a low surface energy in the rutile SnO 2 structure in line with the first-principles calculations of Oviedo [42]; therefore, the plane (101) in coatings doped with 4.4 at% is the preferred orientation.
The plots of variation in lattice parameters (a = b, c) and unit cell volumes against Mo contents in the prepared films are depicted in Fig. 3. An overall declining trend in both lattice parameters and cell volumes is able to be observed, which illustrates that Mo 6? ions successfully substituted Sn 4? in the matrix as the Shannon-Prewitt crystal radius of Mo 6? (60 pm) is smaller than that of Sn 4? (71 pm) [46]. It is noteworthy that when Mo concentration was at 6.9 at% in the precursor, the volume of the films somewhat increased compared to 4.4 at.% films, which was most probably ascribed to the substitution of Mo 5? (63 pm) [46] (which was detected by XPS and is shown later) as well as Mo 6? for Sn 4? in the unit cell as Mo ions at the highest valence state were likely to  be reduced to lower states during the deposition procedure [47].
The crystallite size of these coatings was approximated by the Debye-Scherrer equation: where k is the wavelength of the incident X-rays emitted by the diffractometer, b is the full-width in half-maximum (FWHM) in radians, h is Bragg diffraction angle and K is the shape factor. Here K was assigned as 0.9 since SnO 2 crystallites in the system could be considered as pseudo-spherical. The planes (110), (101), (200) and (211) were selected to calculate the crystallite size as the intensity of these diffraction peaks is relatively high and the broadened degree is relatively low; therefore, the calculated outcomes are more accurate with less errors. The calculated results (

Characterisation of thin films by electron microscopy
Thin-film morphologies Morphological analysis of the surface of SnO 2 thin films was performed by secondary electron SEM (Fig. 4) using Photoshop software, with an average of 450 nm for all. The large-scale cross-sectional images are depicted in the Supporting Information (see Figure S2).

Dependency of size of aggregates on Mo doping
The evolution of the size of aggregates of all coatings is depicted in Fig. 5. It should be noted those anomalously huge aggregates were not included when calculating the average aggregate size. It can be seen that the average aggregate size presents an overall declining trend (293:9 AE 40:3 nm for pure SnO 2 , 169:6 AE 20:2 nm for 2.4 at% doping level and 103:1 AE 9:7 nm for 6.9 at% doping level), which is likely resulted from lattice distortion and stress induced by the continuously increased doping levels.

Structural analysis of thin films from Raman spectral data
Raman spectra of SnO 2 coatings were measured within the range of 200-850 cm -1 at room temperature (Fig. 6). The Raman peaks centred at 475, 637 and 775 cm -1 in these systems are assigned to Raman active modes for tetragonal rutile SnO 2 : E g , A 1g and B 2g , respectively, of which the former is related vibrations in the plane vertical to c-axis whilst the latter two represent vibrations in the plane of the c-axis [51][52][53]. It is observed that all the aforementioned bands are broadened with decreased intensity in the synthesised films as compared with those of bulk SnO 2 , which has been interpreted previously to be induced by the nanocrystalline structure of thin films [41,54]. The A 1g band, which appears in all Modoped films studied, is shifted to the lower frequency, in particular for 4.4 and 6.9 at% Mo-doped films, compared to the undoped counterpart. This variation is possibly due to reduction in the grain size of these films caused by the Mo incorporation [51], which is consistent with the grain size calculation presented above. We also note that there is a shoulder-like band next to A 1g peak centred at ca. 575 cm -1 in all film samples, whereas it is not observed in bulk reference. The appearance of this peak has been attributed to nanoscale SnO 2 and is strongly associated with the atom vibrations in the superficial layers of the nanocrystallites [55,56]. In addition, the smaller the particle size is, the higher wave numbers the band will shift to and the higher the intensity will be,  which could be confirmed by the contrast between the undoped sample and the samples doped with lower Mo concentration. The peak at 425 cm -1 in films doped with high Mo concentration is assigned to an A 2g silent phonon mode for bulk polycrystalline SnO 2 . This vibration mode has been reported in SnO 2 nanoribbons [57], but the causes of this mode are not well understood. In low-frequency range, there are three bands with weak intensities centred at 245, 320 and 350 cm -1 , which are ascribed to infrared radiation active modes, E u (TO, the former one) and E u (LO, the latter two). TO and LO are transverse optical and longitudinal optical, respectively. These bands that are not detected in bulk materials are correlative with the transformation of infrared active modes to Raman active modes resulted from the grain size reduction and particle aggregation [51,52,58].

Surface analysis by XPS
XPS was utilised to characterise the valance states and compositional information at the surface of these prepared coatings. The survey scan spectra normalised to the intensity of Sn 3d peaks are displayed in Figure S3 (see Supporting Information). All three representatives have Sn, O, Mo and C at the surface, which is consistent with XRD and EDX analysis. The C element is believed to the residual of organic compounds in the precursor solution. The narrow Mo 3d core-level scan is depicted in Fig. 7. The Mo 3d spectra are fitted with two spin orbit split doublets (again corresponding to 3d 5/2 and 3d 3/2 splitting) with splitting energy of 3.2 eV. All three spectra are dominated by a spin-orbit doublet with peaks at binding energies 232.2 ± 0.1 and 235.5 ± 0.1 eV, which is associated with Mo in the formal (VI) oxidation state. The other spin-split doublet with peaks at 230.9 ± 0.1 and 234.2 ± 0.1 eV are assigned to the valence state of Mo(V) [59]. This demonstrates that part of Mo 6? ions was reduced during the film deposition. MoO 2 can be ruled out due to the complex Mo 3d line shape associated with Mo(IV) in MoO 2 [60].
In all cases, the Mo(VI):Mo(V) ratio is roughly 2:1, which is indicative of dominance of Mo(VI) in derived peak of the spectra. It can be reasonably deduced that free carriers in the doped SnO 2 films are generated by mixed types of donors, i.e. Mo 6? and Mo 5? , which is, to the best knowledge of authors, reported for the first time.
The Sn 3d spectra in Figure S3b consist of a spin orbit split doublet with 3d 5/2 and 3d 3/2 peaks at binding energies of 486.3 eV and 494.7 eV, consistent with SnO 2 [61,62]. O 1 s spectra shown in Figure S3c are similar in that they consist of a narrow feature at a binding energy of 530.2 eV which arises from the metal oxide [62]. The broad peak is fitted to account for contamination which will consist of adsorbed hydrocarbons, hydroxides and adsorbed water.

Electrical properties of thin films
The sheet resistance (R sh ), electrical resistivity (q), and results of the Hall effect measurement including Hall coefficient values, mobility (l) and carrier concentration (n) are shown in Table 2, and the latter three parameters are plotted in Fig. 8. All SnO 2 thin films doped with Mo behaved as n-type semiconductors, illustrated by the negative Hall coefficient (R H ) values. For pure SnO 2 thin films, the sheet resistance and resistivity are 953.1 X=h and 4:3 0 ð Þ Â 10 À2 X cm, respectively, which is in general comparable to that of samples synthesised by other techniques [18,63], including AACVD [11]. For doped SnO 2 thin films, the resistivity was continuously decreased, and the mobility along with carrier concentration simultaneously grew up before the dopant saturation limit was reached. This can be envisaged as substitution of Sn 4? ions with Mo 6? / Mo 5? in doped samples, generating free electrons in the conduction band [33,47]. The presence of Mo 5? has been confirmed by XPS analysis previously. However, the fact that XPS measurements merely probe a limited thickness of the films (6-10 nm) should be noted. At the doping level of 2.4 at% Mo, the resistivity was substantially reduced to 7:3 Â 10 À3 X cm, the lowest value in all the systems. Mobility and carrier concentration also gained considerably high values at 11.85 cm 2 V -1 s -1 and 2:26 Â 10 20 cm À3 , as compared with those of other synthesised samples. These values of electrical properties are comparable to those of doped SnO 2 coatings reported in the literature [41,63,64]. Additionally, the continuous layers occurring in the SEM images of 2.4 at% Mo-doped thin films (see Fig. 3d) would reduce to a large extent the grain boundaries and hence improve the electron transport performance within the films. This is another potential reason for achieving the lowest resistivity at 2.4 at% Mo doping level. The resistivity generated from the optimal doping level is comparable and even lower than that of SnO 2 :Mo films synthesised by other techniques [35,65].
Upon increasing the doping amount above 2.4 at%, the conductivity deteriorated, such that an evident growth of resistivity and a declining trend in both mobility and carrier densities were able to be observed. The significant increase in carrier densities would give rise to the falling formation energy of compensating defects such as cation vacancies (V Sn ), resulting in the reduction in carrier concentration after the dopant limit [66,67]. In addition, heavy doping suggests increase in the number of ionised impurities in the crystallites, which would lead to the increase in ionised impurity scattering and hence reduce the mobility. When dopant concentration in resulting films was high (4.4-6.9 at%), the SnO 2 aggregates became less continuous and connective, and the aggregate size declined compared to that of samples with lower dopant concentration, which has an adverse impact upon mass electron transport in the host matrix. As a consequence, the mobility experienced a plummet at higher doping contents.   Fig. 9. The average transmittance of all thin films is approximately 80%, and the highest value is 83% for coatings doped of 2.4 at% Mo content (Table 2), which satisfies the transparency requirement for TCOs (over 80% in the visible range). Interference fringes are observed in the spectrum, which is indicative of the thickness variation consistent with the visible wavelengths of light. The dark colouration can be observed in the films doped with higher Mo amount, in particular for the 6.9 at% sample, and therefore the transparency declined slightly. The optical bandgap energies for the thin films (E g ) were calculated by the following equation [68]: where a is the absorption coefficient and hv is the photon energy, A is the constant and E g is the optical bandgap. SnO 2 is a direct bandgap semiconductor, and therefore, n should be 2. In the plot of ahv ð Þ 2 versus hv (eV), the estimated bandgap value was generated via extrapolating the linear part at high photon energy to x-axis (see Supporting Information Figure S4). The calculated values are presented in table, and the changing plot is depicted in the inset of Fig. 9. As can be seen, the bandgap is all ca. 4 eV and the deviation is only ± 0.05 eV. The Mo doping did not induce the expansion of bandgap in SnO 2 thin films, i.e. the Moss-Burstein effect (shift of band gap to higher values owing to higher levels of the conduction band being occupied by extra electrons), as usually observed in other doped SnO 2 systems [64,69]. It is likely due to the limited replacement by Mo 6? for Sn 4? in SnO 2 unit cells; therefore, finite higher levels in the conductions band were filled with electrons provided by doping ions. Similar reports of this phenomenon have appeared in the literature where the bandgap energy shift did not appear after the incorporation of Mo ions into matrix SnO 2 films [65].
The appearance of Urbach-type band tails results from the thermal and structural disorder in thin films, and in doped films where structural disorder caused by dopants is predominant [70]. The absorption coefficient empirically follows an exponential dependence on photon energy in the sub-bandgap region, that is [71]: where a 0 is the constant, hm is the photon energy and E g is the bandgap. The a follows a dependence on transmittance T [72]: where d is the film thickness. As a consequence, variation in a versus photon energy can be acquired. The Urbach energy (E u ) describes the width of band tails and can be calculated from the slope of the linear  Table 2. An evident increase in Urbach energy values can be observed in films with high doping levels (4.4 and 6.9 at%), compared to those of undoped and doped thin films with lower doping levels, which indicates that an excess quantity of dopants would lead to more lattice disorder. To compare the performance of different TCOs, the figure of merit (U) proposed by Haacke [73] is a useful value to evaluate optoelectronic performance of nascent TCOs: where T is the average transmittance in the visible range (400-700 nm) and R sh is the sheet resistance.
The calculated results of all coatings are included in Table 2. The highest figure of merit was achieved at films of 2.4 at% dopants. In principle, the higher the figure of merit of a TCO film is, the more superior optoelectronic property it will have. Consequently, the SnO 2 films synthesised from the solution with 2.4 at% Mo possess the greatest performance. As mentioned above, TCOs with an excellent electronic performance have a high figure of merit. Industrially produced ITO films basically have a figure of merit at the order of 1:2 Â 10 À2 X À1 . The performance comparison through figure of merit between the present work and reported work in the literature is displayed in Table 3, which suggests that the present performance is comparable to other produced transparent conducting thin films, and consequently, SnO 2 :Mo is a promising complement to ITO although there is other research worth launching, such as finding the optimal tin and molybdenum sources and the corresponding dosage.

Conclusions
In summary, pure and SnO 2 :Mo thin films were prepared by AACVD at 450 C as an alternative to ITO in various applications. Powder XRD patterns of as-grown films showed the films are crystalline and possess the cassiterite SnO 2 tetragonal structure with a preferred growth orientation along the (110) plane relative to the substrate in most films. Topdown SEM images presented a closely packed and uniformly distributed nanocrystals on the surface, and cross-sectional images displayed the smoothness of these films with *400 nm thickness. EDX results exhibited good agreement between theoretical Mo concentration in the solution and experimentally found amount of Mo in films, indicating an exquisite control over doping levels. Analysis on Raman spectra demonstrated several unique active modes resulting from the nanoscale nature of the coatings in addition to main peaks generally observed in the bulk material. With regard to optoelectronic properties, the optimal doping level (2.4 at% Mo) realised the best performing film with an electrical resistivity of 7:5 Â 10 À3 X cm and a transmittance of 83% in the visible light range, which is comparable to transparent conductive films prepared in the literature in accordance with the figure of merit comparison. The results illustrate that SnO 2 :Mo thin films synthesised by AACVD are potential candidate TCOs even though other studies need to be developed to complete systematic investigations.

Declaration
Conflict of interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licen ses/by/4.0/.