*3.3. Influence of Number of Layers on Thin Film Morphology*

As previously discussed, by increasing the number of consecutive dips and thereby the number of layers within the TiO2 thin film samples, the average thickness was increased as expected. It has been found that by increasing the number of layers within TiO2 sol-gel, the crystallite size decreases, as the individual single layer thickness increases, owing to the mechanism of growth as additional layers adhere to previous layers [22]. The addition of surfactant produces individual thicker layers, due to the increased viscosity of the sol as the surfactant coordinates around titania particles. As a result, the samples produced with surfactant are on average much thicker, with the increased concentrations of surfactant producing thicker. The addition of surfactant also made individual layers thicker, due to the increased viscosity within the thin film aiding the "sticking" of subsequent layers to the initial layers, which in turn resulted in thicker films overall. Thickness measurements were made using side-on SEM images, and are listed in Table 2. For example, sample A1 produced without surfactant and one layer, has a thickness of 42 nm, and can be compared to the one layer samples produced with addition of surfactant; B1, C1, D1 and E1 which have average thin film thicknesses of 180, 200, 110 and 120 nm respectively. In addition, as the thin film thicknesses increases with the number of layers, the particle size is also found to increase, whereby the smaller crystallites formed in the increasingly thicker layers bind together to form agglomerated particles, as specifically observed in samples C3 and D3, whereby the average particle size is 46 and 40 nm respectively. Agglomeration is also found to occur between the TiO2 layers, whereby the initial layers can act as particle nucleation sites, for example sample B3 shows increased agglomeration owing to the higher concentration of Tween® 40 surfactant used, as seen in Figure 2.

The effect of number of layers on the root mean square surface roughness of the thin films does not have a consistent trend between samples, as seen in Table 2. Typically increasing the number of layers would be expected to increase the surface roughness of samples due to irregular adhesion between layers where different sized particles are placed on top of one another. However, most samples show a decrease in surface roughness from one layer to three layers, e.g. Sample B1 (318 nm), sample B2 (196 nm), and sample B3 (209 nm). This is likely to be the result of particles agglomerating between layers as they adhere to one another, leading to a smoother top surface overall.
