On the use of surfactant for precursor solution processing and influence on the properties of tin dioxide thin films

Tin dioxide (SnO 2 ) thin films have several sorts of applications, including as gas sensors and in photocatalysis, where the surface area plays a relevant role. The use of the Triton X-100 non-ionic surfactant in the precursor solution processing has revealed as a fundamental aid to increase adhesion of the SnO 2 film on the substrate. Moreover, the surfactant presence leads to a significant increase in the surface area of deposited thin films, and the roughness (rms) increases from 57.6 nm to 275 nm. The influence can also be observed in the XRD profiles, with a higher degree of crystallinity for Triton deposited films. It also increases the defect density and the electron scattering, leading to rather resistive films which are thermally excited only above room temperature. The activation energy for the defect ionization is rather high, 800 meV, but the thermal excitation takes place at a lower temperature range, compared to films prepared without Triton in the solution processing. The performance on photocatalysis is improved for films prepared with Triton X-100, since most of the methylene blue dye is degraded in the first 90 minutes when interacting with the sample.


Introduction
Tin dioxide (SnO2) has received increasing interest for use in several sorts of applications, ranging from gas sensors, catalysts, transparent optoelectronic devices to solar cells [1]. One of its interesting properties is its high transparency in the visible range of the electromagnetic spectrum, which combined with a tunable electrical conductivity makes it particularly attractive for application in optoelectronic devices. This conductivity depends on the deposition method, and may be rather high even without doping [2], taking into account that SnO2 is a naturally n-type semiconductor oxide, occurring due to the presence of punctual defects in its lattice, such as oxygen vacancies and interstitial tin atoms, which act as donors in SnO2 [3]. As a consequence of these intrinsic donor levels, the electron density is rather high [4] being the conductivity modulated by the electron mobility.
The sol-gel dip-coating technique has been chosen in this paper to deposit samples (films), for being quite simple, inexpensive and with possibility of coating large areas [5,6].
It leads to films with interesting characteristics, such as larger contact surfaces and greater porosity compared to other deposition methods [7]. On the other hand, some properties arising from this deposition technique may not be efficient in relation to the requested applications, such as the adhesion of the thin film on the desired substrate, thus, the use of a surfactant in the colloidal suspension may improve this condition.
The introduction of surfactants to the synthesis of films and nanoparticles has been used recently, either as a capping and stabilizing agent, or as a controlling mechanism to manipulate the size of crystallites. Afzal et al. [8] developed a SnO2-surfactant composite film through pyrolysis that resulted in higher sensitivity for gas-sensing properties for compositions containing 4% of Triton X-100 and a smaller particle size for highly doped SnO2 films. Bhattacharjee and coworkers [9] adds triton X-100 to the synthesis process also as a capping agent and for manipulating the formation and size of nanoparticles, besides comparing the results with a cationic surfactant (CPC). Rac et al. [10] shows that the interrelationship between the precursor solution's pH and the stability of complex structures (SnO2 agglomerates) and complex molecules (polymers) correlates with the properties of resulting nanoparticles. From these relationships between the synthesis path and obtained nanoparticle's properties we may try to establish parallels between thin-film samples and applications for photocatalysis. Tin dioxide has been reportedly efficient with regard to photocatalytic properties [11], moreover, the coupling of SnO2 with other semiconductor oxides such as TiO2, ZnO and α -Fe2O3 has originated heterostructures with improvements in the electron-hole recombination rate, as well as compatible crystallographic directions [12][13][14] In this work, direct application of the surfactant Triton X-100 leads to improvement of adhesion of thin films through homogenization of deposition by the formation of micelles, which prevents particle agglomeration and promotes the distribution of metallic Sn particles throughout the colloid, and the homogenization of morphology [15][16][17]. In addition, the influence of this nonionic organic compound on electrical, optical, structural and morphological characteristics of deposited thin films of SnO2 is analyzed, aiming for the obtainment of thin films with larger contact surface area and higher porosity, which are interesting characteristics for application in photocatalytic devices and gas sensors [18]. It is observed that the addition of the chosen surfactant affected the samples in different ways: better adhesion and homogeneity of the deposited film on different substrates, increased roughness and increased pore volume.
The Triton X-100 molecule is represented in figure 1, where the division between hydrophilic and hydrophobic parts is also shown. The improvement in adhesion observed with the addition of this surfactant as a reagent in the colloidal suspension is a result of different interactions. While still in the precursor solution, with films not yet deposited, the surfactant leads to stability and aggregation of particles through the formation of micelles, thus acting as dispersants and stabilizers in the solution. Through this dispersing behavior, the surfactant prevents the formation of particles above a critical volume, avoiding precipitation and maintaining homogeneity in the solution, thus, helping in the homogeneity of the prepared sample. When deposited, the surfactant assists in the interaction of the film with the substrate, establishing connections with the substrate regardless of hydrophilic or hydrophobic characteristics, varying only the binding mechanism. In the first case, the extensive polymer chain binds to the surface, in the second case the aromatics and non-polar chains bind and keep the chains apart [19]. These configurations maximize contact area between hydrophilic regions while minimizing contact areas between hydrophobic regions with the aqueous solution. As the deposition proceeds, drainage removes part of the solvents and increases the interaction surface among the deposited polymer, the substrate and the particles of the material being deposited. The used substrate (soda-lime glass) is hydrophilic (wetting angle about 15° [20]), with a larger surface area generated by the structure formed by the polymer, in conjunction with the solution viscosity and the interaction between solution and substrate, which also becomes higher. The overall result is a better adhesion and greater possibility of branching and growth of the deposited material.

Sample preparation
The substrates used for thin film deposition were soda-lime glass and silica, the second being used only to determine SnO2 bandgap, since soda-lime glass and tin dioxide have the fundamental absorption edge in the same range, what could mask a more precise determination. A solution of 30% volume of Extran neutral detergent and 70% of deionized water was used for substrate cleaning, being kept for 20 min. in a USC-1800 ultrasonic cleaner. After being removed from this solution, substrates were washed with deionized water, acetone and isopropyl alcohol, consecutively, and dried using a Black and Decker HG2000-B2 thermal blower. Precursor solutions of SnO2 were prepared in aqueous media (0.5 mol/L), obtained by dissolving tin tetrachloride pentahydrate (SnCl4.5H2O). Hydrolysis of Sn 4+ ion was accelerated by addition of ammonium hydroxide (NH4OH) under constant magnetic stirring until reaching pH 11. Then, the resulting volume is separated into semipermeable cellulose membranes that are submerged in deionized water so that hydrolysis takes place and leading to Clion ejection from the solution. The water is kept in circulation to maintain the hydrolysis process over approximately two weeks. After this period, the pH is about 7, and the contents of the membrane bags are transferred to a beaker, for evaporation of part of the solvent and increase in viscosity, and subsequent film deposition [21].
SnO2 thin films were prepared from two colloidal suspensions with 0.27M of SnO2, however 1% of the volume of Triton X-100 surfactant was added to one of them. The technique used for thin films deposition was dip-coating, using a controlled speed of 10 cm/min. For each prepared sample, 10 layers of film were deposited, with heating between each immersion at 250°C by 10 min at room atmosphere, in order to obtain a better layer compacting, resulting in more dense thin films. Samples were submitted to a final thermal annealing at 450 °C by 1h, using an oven, with heating rate of 1 °C / min.  The samples resulting from this overall procedure are summarized in table 1, where the dimensions of tin dioxide layers are listed. The film thickness is obtained by the following procedure: a kapton tape was placed at the upper end of the substrate so that film was not deposited in this region, thus, with the step formed between the region with and without film, it was possible to estimate the sample thickness by means of a morphological analysis using confocal microscopy. Beside the thickness, table 1 lists the electrical contact width and distance between contacts. SWT means colloidal suspension with surfactant Triton X-100, and SWOT without the presence of it during processing. Samples were also analyzed in form of powder. In this case, all solvent from the tin dioxide suspension was evaporated. Thus, after smashing the resulting crystals, followed by thermal annealing, powders with and without the surfactant in the precursor solution were obtained, as shown in table 2.

Sample characterization
To analyze the surface homogeneity and roughness, as well as the presence of pores, the samples surfaces were analyzed through confocal microscopy, using the Leica DCM 3D equipment, equipped with high power white (emission centered at 530 nm) and blue LED (460 nm). X-ray diffraction measurements were carried out on a Rigaku D/MAX-2100 / PC diffractometer, using Cu Kα radiation (1.5405 Å), and equipped with a Ni filter to attenuate the Kβ radiation. The used scanning rate was 1° / min, in the thin film configuration (2θ) with an incidence angle of 1.5°, and in the powder configuration (2θ / θ). From the obtained diffractograms, the average crystallite size (t) was obtained from the Scherrer equation [22], where θ is the Bragg angle resulting in constructive interference, B is the width of the diffraction peak measured at half of its maximum intensity (in radians), K is a proportionality constant that depends on the geometry of the particles (in this case, it was used 0.9, considering that crystallites have a spherical shape) and λ is the wavelength of the incident X-rays (1.5405 Å, CuKα radiation).
Transmittance and optical reflectance measurements were performed using a Perkin Elmer spectrometer, model Lambda 1050 Uv/Vis/Nir. The measurements were performed in the range of 1800 -250 nm (scanning in the direction of increasing energy), using an integrating sphere. With the transmittance, it was possible to evaluate the optical bandgap using the Tauc plot [23]. Although the SnO2 band structure is rather complex, and there is a great controversy on the nature of its bandgap transition concerning direct [24,25] or indirect [26,27] nature, our best results are obtained considering indirect transition [28], where the square root of optical absorption coefficient (α 1/2 ) is plotted against the energy (hν).
The optical absorption coefficient may be obtained from [29]: where R is the reflectance, T is the transmittance and d is the film thickness.
When the absorption coefficient is well determined, it is also possible to estimate the Urbach tail width for the samples. The analysis is carried out near the fundamental absorption edge, where the absorption coefficient can be approximated by equation 2. This tail generally appears in highly degenerate, low-crystalline or amorphous materials, due to localized states that extend inside the bandgap (intra-bandgap states). When obtaining the Urbach energy, it is possible to estimate the material disorder, concerning defects near the conduction band (for p-type semiconductors) or the valence band (for n-type semiconductors). Urbach's energy can be determined by [30]: where is a constant and Ee is the width of the Urbach tail. Higher values of Ee indicate lower crystallinity and higher disorder in nanomaterials.
Electrical characterization of the films was done through measurements of electrical current as function of the applied voltage and temperature. Measuring current for a varying potential as function of temperature, and using the sample dimensions (distance between contacts and width, film thickness) the resistivity can be evaluated. The biasing and collecting of electrical current was performed with the aid of Keithley electrometers models 2400-c and 6517A. Current measurements as a function of temperature were also performed on samples in a cryostat of APD Cryogenics, operating between 200 and 420K, connected to a temperature controller from Lake Shore Cryotronics model A330 that controls temperature within 0.05 degree of precision. A compressor is also coupled to the system, forming a He gas closed cycle, using cold water for cooling.

Application to Photocatalysis
The use of the surfactant Triton X-100 in the colloidal suspension leads to modifications on the surface structural and morphological characteristics, such as the surface area of thin films. To take advantage of that, photocatalysis measurements were performed taking into account that similar studies on titanium dioxide (TiO2) showed a direct relation of photosensitivity and surface area of thin films [31,32]. For this goal, methylene blue (MB) was used as photocatalysis agent, being a dye with a well-known absorption spectrum, commonly used in the investigation of the photocatalytic properties of a material [33]. For use in the photocatalysis experiments in this work, MB is diluted in a volume of water and then the sample is dipped. While submerged, the sample is irradiated with ultraviolet light.
The maximum of methylene blue absorption spectrum is found about 670 nm, whereas the ultraviolet irradiation generally occurs below 350 nm [34]. In the present work an Osram mercury lamp (11 W) was used, with emission peak at 254nm.
The photocatalysis measurements were performed on samples that were submerged in the solution of water and MB for two hours, for adsorption of MB molecules on its surface.
After these two hours had elapsed, an aliquot of the solution was removed to determine the basis (for each sample) from which the photocatalysis process could be observed. After this first measurement, ultraviolet light was directed to the sample surface still submerged for 90 min and new measurements of transmittance in volumes removed from the solution were performed. This is repeated one last time after 180 minutes of sample irradiation. These aliquots of MB were analyzed through optical transmittance spectra once more.

Structural and morphological properties
Confocal microscopy images for films with and without the addition of the surfactant Triton X-100 are shown in Fig. 3. Images shown in fig. 3            showing a much less homogeneous sample and quite amorphous.

Optical characterization
Transmittance (T) and reflectance (R) spectra for a film deposited on silica substrate from a precursor solution containing triton (sample SWT), are shown in figure 8(a). Film exhibits high transparency and low reflectance for visible-NIR region. The formation of interference fringes is not observed, probably due to a lack in homogeneity on the topography of surface and at the film/substrate interface for the phenomenon to occur, as the high roughness and possibly high porosity inhibits the appearance of fringes. With the values of R and T it was possible to calculate the optical absorption coefficient according to equation (1), and thus estimate the value of the indirect bandgap through the Tauc plot ( Fig. 8(b)), obtaining the value of 3.7 eV, in good agreement with the literature [37][38].
Taking into account the transmittance obtained for the sample containing the surfactant, it was possible to calculate the Urbach energy using equation (2), yielding the value of 263 meV. Studies point that for polycrystalline tin dioxide in a temperature range of 340 to 360 K, the Urbach energy ranges from 200 to 250 meV [30,39,40]. We believe that our value is close enough to the found in the literature, when taking into account that the deposition method is not the same.
The Urbach energy of the film obtained with the addition of Triton indicates the high polycrystallinity of the film, that is, with a high density of grain boundaries. It leads to a high electron scattering rate, and a high resistivity, which is temperature dependent. Besides, when the material is analyzed as function of temperature the free carrier statistics is highly dependent on the electron trapping at grain boundaries and ionized intrabandgap energy defects, suggesting high activation energy, as will be shown.

Electrical characterization
The temperature dependent electrical behavior of the SnO2 films was evaluated through measurement of electrical current for SWT and SWOT films with a fixed applied voltage of 20 V. Based on these data and the sample dimensions (table 1), the resistivity was evaluated and it is shown in figure 9. The inset in figure 9 is the evaluation of the main activation energy in each case, for the higher temperature range It is interesting to note the typical behavior of a semiconductor material for both samples in fig. 9, that is, the resistivity decreasing with the temperature rise, which becomes more pronounced after 360K, where the electrons trapped in the defects have sufficient thermal energy so that they can be excited to the conduction band. Overall, the resistivity of the film deposited from the colloidal suspension containing Triton is higher. This is expected, due to the higher electron scattering caused by the porosity of this sample. The drastic drop in the resistivity of SnO2 as the temperature is increased may also be related to the removal of water and O2 ionic species (at approximately 80 °C [41]) from the film surface. When the temperature required for this removal is reached, both samples behave in a similar way, differing slightly in the excitation rate, as can be seen from the inset of figure 9.
Using data of current as function of temperature, it was possible to determine the main activation energy of both films deposited with and without surfactant. The inset of figure 9 brings the Arrhenius plot from the electrical current data, leading to a behavior typical of multi-level ionized defects [42], characteristic of the continuous variation of the curve slope.
This Arrhenius plot allows evaluating the activation energy for the deepest energy level, ionized at higher temperature range, which is above 360 K (1000/T about 2.7 K -1 ). The addition of surfactant to the colloidal suspension of SnO2 decreases the activation energy for the deepest defect level, being 800meV for the film with Triton and 906 meV without Triton, which may be an indication that the deeper intrabandgap state requires slightly less energy to be ionized, a phenomenon that may be related to the increase in the density of defects present in the sample obtained by the solution containing the Triton X-100. It is interesting to note the similarity of these values with the grain boundary potential barrier [43], obtained for samples deposited under similar conditions, which evidences the great predominance of electron scattering at the grain boundary for these films, that contain very small nanocrystallites, as shown in table 3.

Application to Photocatalysis
Photocatalysis measurements were performed on films obtained from colloidal suspension with and without the addition of the surfactant Triton, which were cut from the whole sample. Figure 10 1) the use of the surfactant causes more pores to show up on the surface, thus increasing the contact area in relation to the films that do not contain the surfactant in its preparation [9], which has already been confirmed through confocal microscopy images, which denoted a significant increase in film roughness; 2) The second factor that could lead to the obtained results concerns the surface dimensions of the thin films deposited, as inferred from SEM images (figure 4) since the sample deposited from the solution with Triton has much larger deposited areas, caused by better adhesion, and leading the photocatalytic capacity to be affected due to these effects generated by the addition of the surfactant.
To make the comparison on the photocatalytic activity of these samples more effective, transmittance variation by unit area is plotted, as observed in figure 10(b). After 90 min a higher photocatalytic activity is observed for SWT, denoted by a more pronounced variation in transmittance, whereas for SWOT no significant changes were observed. During the next 90 min, totaling 180 min of experiment, the variation increased with a lower rate for SWT, whereas for SWOT it takes place with increased rate. This phenomenon can be explained by the increased reactivity of the semiconductor oxide surface with the addition of the surfactant, since in the first 90 min the photocatalytic activity was greater for SWT, but after the first 90 min there was no longer a high concentration of available OHions to degrade the MB solution. Different behavior occurs with SWOT, obtained without the presence of the surfactant, which does to not have great reactivity in the first stage of the photocatalysis process, and no marked variations on the transmittance were observed. Only at the second stage the degradation rate increases significantly. The overall result, after 180 min of MB degradation process for both samples, is a higher reactivity for the sample obtained with the surfactant, which highlights its effectiveness on tin dioxide processing for photocatalytic applications.

Conclusions
The use of the Triton X-100 non-ionic surfactant plays a fundamental role in the fixation of the SnO2 film on soda-lime glass substrate, as evidenced by images produced through confocal and electron scanning microscopy. The surfactant's influence can also be observed in energy dispersive X-ray spectroscopy and XRD profiles, in the former case with a higher degree of crystallinity for Triton deposited films, with clear peaks of (110), (101) and (211)  The introduction of this surfactant also increases the resistivity of SnO2, caused by the increase in the defect density and also in the surface area, which increases the electron scattering in the naturally n-type tin dioxide. Moreover, the addition of surfactant to the colloidal suspension decreases the activation energy for the defect with the deepest intrabanbgap energy level in the film.
Photocatalysis measurements revealed an increase in the performance when comparing films prepared without and with Triton X-100 non-ionic surfactant. From the performed experiments' data, it can be observed that most of the dye is degraded in the first the increase in surface area and formation of pores, which were not observed in the sample not containing surfactant.