Introduction

Synthetic dyes have been intensively used during the last few decades. They are the most persistent pollutants, especially in industrial wastewaters. Therefore, the expansion of dye elimination research is an important topic endorsing the search for original methods to protect the environment from the aggressive effect of recalcitrant pollutants. In addition, the photocatalytic degradation of organic pollutants in water using semiconductors has been widely studied [1].

The great versatility of TiO2 is due to its high photocatalytic activity, chemical stability, low cost, and non-toxicity [2,3,4,5,6,7]. It covers various fields, such as water purification, air depollution, and solar energy conversion. In fact, several researchers investigated the fixation of TiO2 for water depollution in order to remove any separation step of the photocatalyst from the treated water.

TiO2 layers on metallic surfaces can be formed employing diverse methods, such as cyclic voltammetry electrodeposition [8], anodization [9], spray plasma technique [10], sol–gel [11], and electrophoretic deposition (EPD), which could provide uniform thin films [12,13,14,15,16,17,18,19,20,21]. Indeed, few studies were reported on the EPD of TiO2 nanoparticles by organic suspensions [22,23,24,25] or mixtures of organic solvents and water [1, 26,27,28,29]. At low voltages, the electrophoresis of TiO2 nanoparticles is generally realized using ethanol–water mixtures without stabilizers addition [1] or by adding compounds such as (C2H5)4NOH and Tiron molecule [26, 27].

Employing aqueous medium without adding organic solvents was studied in few previous works as [16] where EPD was performed by suspensions of commercial TiO2 nano-powder (Degussa P25). The use of aqueous medium can lead to bubble formation in the deposit. To overcome this problem without modifying the EPD process, pulsed voltage was used instead of direct voltage. This method is an interesting process employed to deposit a controlled coating over different conductive substrates by adjusting the deposition parameters, such as the applied voltage, the deposition time, the applied frequency, the duty cycle, and the suspension concentration.

Stainless steel (SS) has also been applied to develop TiO2 films [29,30,31,32] due to its electrical properties. In addition, it has been extensively used in various industrial applications.

Yet, the deposition of coatings over bare SS may present a low adhesion [14, 33]. Thus, several adjustment procedures were introduced to adapt the surface of SS. Because the choice of the interlayer material needs previous knowledge about the application field, a conversion layer (CL) should be developed before the coating deposition.

In this context, this paper presents an innovative approach for the production of TiO2 nanoparticles and their deposition on SS surface. Pulsed electrophoretic deposition of TiO2 coating was applied over the treated SS substrates. The conversion layer has high surface area and a remarkable morphology with microspores and large cavities [34]. These properties help enhance the TiO2 film adherence. Furthermore, the synergy between Fe2O3 and TiO2 enhances the photocatalytic treatment efficiency by acting on charges recombination. Because the TiO2 film crystalline structure, its morphology, and its particle size can affect its photocatalytic efficiency [35], the produced film was studied, in this research work, before and after calcination.

Materials and methods

Preparation of the TiO2 thin films

To have an efficient and moderate hydrolysis, 20 ml titanium tetrachloride (TiCl4, 98%, Sigma-Aldrich) was gradually dropped into 200 ml of ice-deionized water under continuous stirring until the white fog disappeared [36, 37]. Because of the extremely exothermic reaction, the solution temperature became rapidly hot. In order to eliminate the large amount of chloride ions in the system, the prepared solution was dialyzed using a Nadir®-Dialysis Tubing Membrane.

The produced titanium hydroxide/hydrous titanium oxide (TiO2·nH2O) was transparent like soft gel. It was subsequently dried at 120 °C for 6 h. The aggregate dry TiO2 powder was manually ground in a mortar for 15 min until obtaining a fine white powder. TiO2 particles had an average particle size < 80μm as revealed in Fig. 1. The particles of synthesized TiO2 were analyzed by laser diffraction method. The TiO2 grains had a wide range of size distribution ranging from 0.2 to 80 μm. The histogram in Fig. 1 illustrates two peaks corresponding to two average sizes of aggregates. The first one corresponds to 75% of the total volume with a diameter of 7.76 μm, while the second one represents 25% and has a diameter of 0.57 μm.

Figure 1
figure 1

Grain size frequency histogram for a synthesized TiO2 analyzed by laser granulometer

Full size image

A suspension of 10 g/l concentration was made by mixing TiO2 nanoparticles and deionized water in a glass beaker. It was magnetically stirred for 30 min. This suspension was used to deposit TiO2 over both bare SS 316L and a pre-functionalized steel with a CL. The working electrode disks of 15 mm diameter and 0.8 mm thickness were mechanically polished using SiC papers in a sequence of 400, 600, 800, and 1200. They were cleaned using deionized water. Then, they were dried at ambient temperature.

The conversion coatings were prepared on SS disks by an oxidation in a conversion bath composed of 5 ml of sulfuric acid mixed with 0.126 g of hydrated sodium thiosulfate and 0.6 ml of propargylic alcohol in a 100-ml volumetric flask [34]. The immersion period of the sample in the conversion bath was 35 min at a temperature between 60 and 70 °C. The prepared samples were rinsed in demineralized water. Then, they were dried at 120 °C for 60 min. After that, the layers were subject to heat treatment at 450 °C for 2 h to achieve their structural stabilization.

The SS electrode was immersed in HCl for 10 min and rinsed with deionized water. Acetone was used to further clean the substrates. TiO2 thin films were produced using pulsed electrophoretic deposition [15,16,17,18,19,20,21]. A platinum electrode was used as the counter electrode. The working electrode and the counter electrode were mounted at a distance of 20 mm from each other, perpendicular to the liquid surface in the glass beaker at ambient temperature.

Pulsed electrophoretic deposition was performed using the setup illustrated in Fig. S1. The electrodes were connected to a home-made generator (HP 3314 A) employed in order to produce an AC signal and a digital oscilloscope (Tektronix) used to control the pulsed voltage. The frequency of the pulse cycle was fixed to 10 kHz. It remained steady for each immersion. The pulse time intervals were measured by a field-effect transistor (FET).

The operating parameters were chosen according to references [15, 20, 38]. Thus, the deposition was conducted up to 10 min at a voltage of 20 V and with a duty cycle of 50%. These mixtures between the employed electrophoretic bath and the circuit were qualified to control the definite speed and height of each pulse through the whole deposition process. To obtain a homogeneous and smooth deposition, it is essential that the particles stay totally dispersed and firm. Large particles always have a tendency to deposit due to gravity. To avoid this deposition, the mobility of particles resulting from the application of an electrical field of the pulsed EPD must be higher than that caused by gravity. Therefore, it is challenging to get uniform deposition from the sedimentation of suspension with large particles. Consequently, particles with low weight will have high mobility, which leads to nano-film deposition.

After deposition, the thin films were dried in air for 1 day. To test the effect of heat treatment on crystallinity and related photocatalytic activity in a later stage [35], two working electrodes (one over SS and the other over a CL) were dried in a box-type furnace. The calcination temperature was 550 °C for 4 h 30 min with a heating rate of 2.5 °C/min [20].

Characterization of the films

Microtrac s3500 was used to examine the size distribution of the synthesized particles. It used three precisely placed red laser diodes to characterize particles. The S3500 measures particle size ranging from 0.02 to 2800 microns.

The crystalline structures of thin films were examined by X-ray diffractometer (DRX model X’Pert Pro, PANalytical) by means of Cu Kα (λ = 1.5405980). The presence and ratio of the different crystalline phases in the film were verified by XRD [39].

SEM–EDAX spectra of nano-TiO2 were carried out with a LaB6 Scanning Electron Microscope (JEOL JSM-5400LV) operating at 0–30 kV. The coatings were protected before observations with a conductive carbon layer to prevent any charging effects.

Nano-indentation was performed by the Nano-indenter NHT2 (CSM Instruments, Switzerland) using a Berkovich three-sided pyramidal diamond indenter with a nominal angle of 65.3° and a nominal radius curvature of 20 nm at a normal load of 100 mN and a sliding velocity of 1 µm min−1. The measurements were taken with the load control mode. Load–displacement curves were attained employing the Oliver and Pharr method [40]. From the load–penetration curves, surface hardness (H) and Young’s modulus (E) were determined.

Scratch test was realized at ambient temperature via the same apparatus using a Rockwell spherical diamond tip with a radius of 50 µm. The tests were achieved by increasing the load from 30 mN to 15 N (load rate 30 N/min) with a table speed of 6 mm min−1 over a distance equal to 3 mm.

Adhesion strength of the coating, confirmed by the optical observations of the scratch track, was determined from the loads, at which cracking and delamination occurred, and from the variation of friction force.

The hydrophilic property was calculated by examining the water contact angle of TiO2 thin films using a contact angle meter (ATTENTION, THETA).

The photoluminescence (PL) was measured at room temperature employing a PerkinElmer LS 55 fluorescence spectrophotometer. The excitation wavelength was 270 nm. The analysis of the spectrum was achieved with Gaussian function fitting in Origin 8.5. UV–visible diffuse reflectance spectra were realized using PerkinElmer Lambda 950 spectrophotometer over the wavelength range of 375–600 nm.

Photocatalytic performance

The photocatalytic activity was evaluated using a typical textile dye amido black 10B (C22H14N6Na2O9S2, M = 616.5). The decolourization of amido black (AB) in aqueous solution was monitored at ambient temperature. In each experiment, the coated electrode was immersed into quartz vessel containing 100 ml of amido black 10B (AB-10B) solution with a concentration of 10 mg/l. Before the irradiation, the electrode was immersed in the dark for 30 min to achieve the adsorption equilibrium of AB onto the catalysts. Then, we stirred the solution exposed to UV light. A high-pressure mercury lamp (HPK 125 W, Cathodeon) was applied as a UV excitation source (λ = 365.5 nm). A noticeable peak at 617 nm and two small peaks at 226 and 318 nm characterized the absorbance spectra of AB-10B. The concentration of the AB was monitored using a UV–Vis spectrophotometer by examining the main absorbance at 617 nm. Chemical oxygen demand (COD) analysis was performed as mineralization test. The COD elimination was calculated employing the following formula [41]:

$$ {\text{COD}}\; {\text{elimination}}\;\% = \left( {\frac{{{\text{COD}}_{0} - {\text{COD}}_{t} }}{{{\text{COD}}_{0} }}} \right) \times 100, $$
(1)

where COD0 and COD t indicate the COD before and after decolourization, respectively.

Results and discussion

Structural characterization of titania catalysts films

Figure 2A represents the XRD patterns of the SS substrate before and after the conversion bath.

Figure 2
figure 2

A XRD patterns of (i) conversion layer and (ii) bare stainless steel; B XRD patterns of different thin films: (a) bare SS, (b) annealed TiO2/SS, (c) non-annealed TiO2/SS

Full size image

The two peaks at 2ϴ = 43.61° (hematite) and 2ϴ = 50.78° (magnetite) refer to the substrate made of SS 316L (Fig. 2A-ii). After immerging our substrate into the conversion bath, there appeared other peaks of hematite (Fe2O3) structure at 2ϴ = 24.12° (012) and 33.2° (104) as well as an increase in the 2ϴ = 43.47° (202) peak intensity. As a result of the aggressive nature of the conversion bath, there was direct chemical attack of metal surfaces to produce oxide metal. The sample showed the presence of magnetite (Fe3O4) at 2ϴ = 35.6° (311) and 50.7° (003).

For quantitative phase analyses, the preferential orientation of the CL films was studied by calculating the texture coefficient [TC (hkl)] using the following formula [42, 43]:

$$ {\text{TC}}_{{\left( {{\text{hkl}}} \right)}} = \frac{{I_{{({\text{hkl}})}} {\text{/}}I_{0} }}{{\left( {1{\text{/}}N} \right)\sum {I_{{({\text{hkl}})}} } {\text{/}}I_{0} }}, $$
(2)

where TC (hkl) is the texture coefficient of the plane specified by Miller indices (hkl), I represents the measured intensity, I 0 corresponds to the standard intensity, and N refers to the number of planes observed in the X-ray diffraction pattern.

The deviation of the texture coefficient from unity implied the preferred orientation of the growth of the thin film. TC values of CLs are given in Table 1 showing that the film is highly oriented along (202) plane due to the high crystallinity of hematite.

Table 1 Different texture coefficient [TC (hkl)] values of the CL
Full size table

Figure 2B-b-c shows the intensities of diffraction peaks of TiO2 over SS with and without calcination. The reflections at the angles 2ϴ = 25.3° (101) and 37.98° (004) confirm the presence of anatase phase. The peaks at 30.8° (121), 36.2° (012), 44.6° (114), and 48.12° (231) reveal the existence of brookite, whereas that at 54.1° (211) corresponds to rutile phase. The percentages of the different crystalline phases are represented in Table S1. The crystallite size was approximated using the full width at half maximum (FWHM) of the intense diffraction peaks of crystal phase according to the Scherrer formula [44]:

$$ D = \frac{K }{{\beta \;{ \cos }\;\theta }}, $$
(3)

where D is the mean size of the crystalline; K stands for a dimensionless shape factor of about 0.9; λ corresponds to the X-ray wavelength; β represents the line broadening at half the maximum intensity (FWHM) in radians; and θ refers to the Bragg angle (in °).

Figure 3 represents the XRD patterns of TiO2 film over the CL before and after annealing the thin film. The peaks of the non-annealed film over CL (Fig. 3c) are the same as those of the non-annealed film over bare SS (Fig. 2c). The absence of the peaks corresponding to the substrate in both cases refers to the presence of TiO2 thick film. Nevertheless, the annealed film (Fig. 3b) has peaks linked to the CL (Fig. 2A-i) at 2ϴ = 24.12°, 33.2°, and 35.6°.

Figure 3
figure 3

X-ray diffraction patterns of different thin films: (a) conversion layer; (b) annealed TiO2 deposited over CL; (c) non-annealed TiO2 deposited over CL

Full size image

The hematite (diameter of 16.8 nm) appeared at the surface of the SS substrate after the treatment in a conversion bath. Once TiO2 was deposited over the CL and the new film was calcined, the hematite structure was observed again (Fig. 2b) with a diameter of 83.14 nm. This could be explained by the presence of cracks and the output of Fe2O3 to the surface under temperature effect.

Table S1 shows the crystallite size of the different phases calculated by the Scherrer formula as mentioned before. (Anatase diameter is D A, rutile diameter is D R and brookite diameter is D B.) The mass fractions of anatase W A, rutile W R, and brookite W B were calculated as shown in Eqs. (4), (5), and (6). K A and K B are coefficients with values equal to 0.886 and 2.271, correspondingly as in Ref. [39].

$$ W_{\text{A}} = \frac{{K_{\text{A}} A_{\text{A}} }}{{K_{\text{A}} A_{\text{A}} + A_{\text{R}} + K_{\text{B}} A_{\text{B}} }}; $$
(4)
$$ W_{\text{R}} = \frac{{K_{\text{R}} }}{{K_{\text{A}} A_{\text{A}} + A_{\text{R}} + K_{\text{B}} A_{\text{B}} }}; $$
(5)
$$ W_{\text{B}} = \frac{{K_{\text{B}} A_{\text{B}} }}{{K_{\text{A}} A_{\text{A}} + A_{\text{R}} + K_{\text{B}} A_{\text{B}} }}, $$
(6)

where anatase intensity is A A, rutile intensity is A R, and brookite intensity is AB.

The anatase crystallites become bigger in size with a subsequent decrease in the mass fraction of their grain under the influence of the annealing temperature. Brookite is the most metastable phase, and its crystal size was generally between 11 and 35 nm. Therefore, it would convert straight to rutile. However, anatase may be transformed to rutile or brookite and then to rutile. Anatase to rutile phase transformation is claimed to rise either at the grain edge [45] or inside the grains [46, 47]. The rutile nuclei form and rise in size by consuming the surrounding anatase. If the rutile nuclei form inside the anatase crystallites, the rutile grain growth goes exponentially with temperature. However, in our case, there was shrinkage in the rutile crystallite size, which can be explained by the formation of rutile nuclei on the surface of anatase particles [46]. The volume fraction of the two phases approved the intensity of their XRD peaks.

A typical micrograph of the CL revealing the attacked surface containing many cracks is presented in Fig. 4. This cracked morphology is similar to the morphology of treated substrates in a conversion bath, as reported in the literature [34].

Figure 4
figure 4

SEM micrograph and EDAX spectrum of the conversion layer

Full size image

Once the SS substrates were immersed in the conversion bath, they were attacked by the aggressive medium, which created cracks in the substrate. Moreover, the energy-dispersive X-ray (EDAX) spectrum depicted in Fig. 4 indicates that the obtained product is mainly composed of Fe2O3 as shown before by XRD analysis. The innumerable peaks are related to additional elements composing the SS, namely Cr, Ni, Mn, and Mo.

From the scanning electron microscopic images (Fig. 5), it is clear that TiO2 thin films were made, before calcination, of smaller aggregated particles, resulting in a high porous volume due to aggregation between tiny TiO2 particles. Equally, after heat treatment, the films were composed of bigger agglomerated particles which may result from the phase transformation from anatase to rutile (as shown in Table S1). In addition, it could be attributed to smaller particles interaction that causes their aggregation into particles with bigger sizes.

Figure 5
figure 5figure 5

SEM images and EDAX of: a annealed TiO2 over a CL; b non-annealed TiO2 over a CL; c annealed TiO2 over SS; and d non-annealed TiO2 over SS

Full size image

The average sizes of the primary particles predictable from the SEM images are presented earlier in Table S1. These sizes are in good arrangement with those calculated from the XRD pattern using the Scherrer equation. Additional reflection shows that the large number of mesopores is caused by the aggregation of primary particles or crystallites.

The surface revealed different grain sizes with the presence of some cracks, especially in the case of annealed TiO2 over SS (Fig. 5c). These cracks could be explained by the quick evaporation of water during annealing, thickness of TiO2 film, and the large variance of coefficient of thermal expansion between the substrate and TiO2 thin film.

EDAX spectra of TiO2 over SS showed well-definite peaks corresponding to Ti along with a peak of oxygen, while the other metals in the alloy did not appear.

After heat treatment, Fe and Cr appeared. These elements descended from the passive film on top of the pitted zone (cracks) caused by calcination. Fe was derived from iron oxide as presented before in XRD analysis (Fig. 2).

Nonetheless, Cr appearance resulting only from calcination was due to its small trace as initial material in the substrate surface. Table S2 shows the composition of both ordinary SS 316L and CL. In our study, we obtained a mixture of new austenitic stainless steel by reducing the nickel content and increasing both molybdenum and manganese. Some metal species impurities can act as recombination centers for charge carriers during photocatalysis [48,49,50,51]. However, in case of TiO2 over a CL, EDAX spectra showed the presence of Ti, O, and Fe elements. After annealing the TiO2/CL film, there appeared some trace of C.

Thickness of the films

To investigate the thickness of the different deposited TiO2 layers, two methods were used.

The first technique is a theoretical thickness measurement which consists in calculating the depth of the electrons penetration using the Castaing formula [52] given by:

$$ R\left( {\upmu{\text{m}}} \right) = 0.033E_{0}^{1.7} \;(A/\rho Z), $$

where ρ represents the volume density (g cm−3), E 0 is the energy of the electrons, A corresponds to the atomic mass, and Z denotes the atomic number of the film.

Experimentally, the primary energy (E 0) of the electrons was gradually increased by detecting the characteristic peaks of the coating by EDAX. At the appearance of the characteristic peaks of the substrate, we noted the value of E 0 used to determine the range R corresponding to the coating thickness. Figure S2 shows the evolution of the range as a function of the primary energy in the case of titanium dioxide (A = 25.99, Z = 12.40, ρ = 2.39).

The second method is cross-observations by SEM allowing the measurement of film thickness (Fig. 6). Table 2 presents the thickness of all TiO2 films obtained by applying the two previously mentioned techniques. The theoretical thickness measurements are in good agreement with the SEM measurements. Correspondingly, the feeble difference between theoretical and experimental values is in the range of error bar.

Figure 6
figure 6

SEM cross-sectional images of a non-annealed TiO2/SS and b annealed TiO2/CL

Full size image
Table 2 Thickness values of TiO2 films
Full size table

With annealing temperature, there is an observed increase in the films thickness for both TiO2/CL and TiO2/SS, which demonstrates that calcination considerably affects the thickness of the deposited film [53,54,55]. The film thickness growth was attributable to the increase in the grain sizes with the annealing temperature. At high temperatures, atoms had sufficient diffusion activation energy to inhabit the site in crystal lattice and the grains with inferior surface energy become larger [56]. This result is in agreement with XRD data shown in Table S1.

Studies in [57,58,59] revealed that TiO2 film thickness affects its optical performance as it is demonstrated in Fig. 10.

Mechanical properties

The mechanical strength of the coatings is a vital condition for industrial applications. Figure 7 illustrates the evolution of the normal force as a function of penetration depth obtained by nano-indentation for the TiO2 coating. First, it is worth noting that all curves exhibit a similar plastic deformation. The observed imprint was permanent and did not show an elastic relaxation. Table 3 depicts the hardness and the elasticity modulus of TiO2 thin films deposited over bare SS and CL, annealed (Fig. 7a–d) and non-annealed (Fig. 7b–c). The annealing temperature and the presence of a CL remarkably influenced the mechanical properties of the deposited film. The smooth and crack-free morphology of the TiO2/CL coatings led to a stable and adhesive coating to the substrate, which was attributed to the intermediate layer (CL) linking the TiO2 nanoparticles and the SS. Nevertheless, in the case of the coating prepared without CL and due to the important cracks and augmented surface roughness, the TiO2 film could be detached easily.

Figure 7
figure 7

Nano-indentation curves carried out on the TiO2 coating over: (a) SS with annealing, (b) SS without annealing, (c) CL without annealing, and (d) CL with annealing

Full size image
Table 3 Young’s modulus and hardness of the different TiO2 films
Full size table

There was an increase in the values of both the elasticity modulus and the hardness of the TiO2 thin films deposited over a CL compared to those on SS [16]. These results suggested an improvement of the coating adhesion. Moreover, annealing the films reduced the Young’s modulus, which badly affected the adherence of the film [30].

In fact, no considerable difference in E and H between all deposited films was observed. Besides, the annealed films had a slightly greater Young’s modulus and hardness. Generally, the microstructure of the thin film plays a serious role in defining its mechanical responses [60]. Increasing temperature creates structural modifications behind the variations in hardness levels. The rise of the film hardness, in our case, could be explained by the growth of the structure into a more compact one compared to the structure obtained before heat treatment [61]. It should be highlighted that porosity is of ultimate significance on material properties. The primary synthesis methods often lead to porosity and incomplete bonding among the grains. Processing imperfections, as porosity, may damage the properties of nano-crystalline materials as it was proved in other works [62]. Weertman et al. [63] showed the Young’s modulus decrease as a function of porosity for nano-crystalline Pd and Cu. The decrease in Young’s modulus with porosity has been recognized and definitely spoken in many mechanics simulations.

The decrease in E, in our work, could be attributed to the modification of the microstructure properties from a compact and dense morphology without cracks into a structure having a lot of cracks and voids. As shown in the SEM images, with the annealing of the deposited film, the number of cracks increased.

Adhesion properties

The scratch test was carried out for the characterization of adhesion. It involves three distinct stages: the measurement of the initial profile of the surface to scratch, called “pre-scratch,” the scratch step, and finally the measurement of the final profile in scratch background “post-scratch.” The presence of brute spallation, a large area isolated even using the lower load in the machine, is a sign of low adhesion strength.

The different adherence failures can be identified by: L C1, the lower critical load, defined as the load at which the first cracks occurred (cohesive failure); L C2, the upper critical load at which the first delaminating at the edge of the scratch track happened (adhesion failure) [64]; and L C3, the load obtained when the damage of the film exceeded 50% [65]. Figure 8 presents the result of scratch test performed on a non-annealed TiO2 coating over CL.

Figure 8
figure 8

Scratch test of as-deposited TiO2/CL coating

Full size image

We got good adhesion since the spallation of the coating was slow as the applied force increased. Thin films with CL showed enhanced adhesion compared to those over a SS. Table 4 represents the theoretical load values of the different films obtained from the friction–load curves.

Table 4 Critical loads of TiO2 thin films
Full size table

In this work, we propose, for the first time, the use of a CL to improve the adhesion of the TiO2 film over the SS substrate. Comparing our results to those obtained in [16, 17] proves that the presence of a CL over a SS substrate increases the adhesion of TiO2 film. Indeed, LC1 in case of TiO2/CL (L C1 = 1.8–8.2 N) film is higher than L C1 in case of TiO2/SS as shown in references [16] (L C1 = 969.473–1177.968 mN) and [17] (L C1 = 53.78–345.74 mN).

In addition, the increase in the critical load associated with adhesive limit demonstrates that the annealing treatment improves the coating adhesion to the substrate. This improvement was due to the enhancement of the mechanical anchorage inside the substrate roughness. The chemical affinity between the TiO2 coating and the substrate as well as the existence of cracks and other flaws affected the mechanical properties of TiO2 coating and the adhesion of the deposited films.

Thin film wettability

Table 5 shows the result of contact angle measurements for water on the TiO2 coating films deposited over bare SS and a CL. For both cases, the contact angles for water on the TiO2 coating films decreased by annealing the thin film.

Table 5 Contact angle measurements of the different deposited films
Full size table

In case of non-annealed films, TiO2/CL presented lower contact angle than TiO2/SS. Besides, the functionalized surface with a CL did not contain hydroxyl and water could not come into the interior region of the coating films. The annealed TiO2 films had large particle sizes as shown by XRD analysis (12.65 nm for TiO2/SS and 9.74 nm for TiO2/CL) and in SEM pictures (Fig. 4a, c), which induces a progress of the roughness and the water contact surface. These important modifications were caused by conversions in the surface morphology and/or chemical condition of the film surface. Therefore, the contact angles for water after annealing were low. The increase in the adsorbed hydroxyl groups on the surface would upswing van der Waals forces and hydrogen bond interactions between water molecules and hydroxyl group [66]. Then, water could completely blow out across the surface and the hydrophilicity could be retained for long time.

It should be noted that, for both annealed films, photocatalytic activities could decrease due to their larger particles size and superhydrophilicity [67,68,69]. For the conversion layer, a large contact angle corresponding to hydrophobic surface (θ ≥ 90) was obtained. The water tended to form droplets over the surface of the coating. Then, it spread completely. Even slight variations in surface roughness can lead to improved wetting properties.

Optical properties

Photoluminescence (PL) spectroscopy is a nondestructive method employed to examine the electronic structure of materials. PL spectra demonstrate the existence of lattice defects, recombination of the photogenerated electron–hole pairs, self-trapped excitons, immigration, and oxygen vacancies [70, 71]. The photoluminescence spectra of the different deposited thin films are presented in Fig. 9A. From the deconvolution and peak fitting, we notice an intensive photoluminescence at a wavelength of around 415 nm and 418 nm corresponding to bare SS and the CL, respectively. This resulted from the oxygen vacancies in Cr or Fe oxides at the surface [72]. Both peaks coexist after the deposition of TiO2 film, which was due to the strong background obtained from SS and the conversion coating [73, 74]. Moreover, the intensities were remarkably reduced after TiO2 deposition, which was attributed to the lower PL property of TiO2. The comparison between the films over SS (Fig. 9B-c-d) and over the CL (Fig. 9B-e-f) indicates that the PL intensity declines by introducing the CL to adjust the optical properties of the substrates.

Figure 9
figure 9figure 9

A Photoluminescence spectra of the different thin films excited at 270 nm. B-a, B-b, B-c, B-d, B-e, and B-f deconvolution and peak fitting of PL emission peaks for all studied samples

Full size image

The PL intensity reduction was caused by charge separation property of TiO2 films. In general, the higher decrease in the intensity of photoluminescence is due to the lower charge carrier recombination in TiO2 film [72]. As shown in Fig. 9A, lower intensity was obtained by non-annealed TiO2/CL followed by the annealed one. From the deconvolution and the peak fitting of the PL intensity (Fig. 9B-c-d-e-f), we notice the presence of the same first band around 397 nm in all curves, which was accredited to the band-to-band transition of TiO2 [75].

The peak at 442 nm corresponds to the interaction of trapped electrons on Ti 3d orbital with holes existing in O 2p orbital of TiO2, which generated self-trapped excitons (STE) [76]. However, the emission band at 463 nm was related to the charge transfer transition from Ti3+ to oxygen anion in a \( {\text{TiO}}_{6}^{{8^{ - } }} \) complex associated with surface oxygen vacancies [70, 77, 78]. Finally, the low bands at 419 nm and 525 nm could be attributed to the oxygen vacancies [76]. The section of electrical measurements confirms the PL results.

Figure 10 illustrates the UV–Vis diffuse reflectance spectra (DRS) of the different TiO2 films over SS and CL. The band gap transition as distinguished above in the PL analysis was confirmed in all TiO2 thin films showing the classic absorption of anatase in the UV region λ < 400 nm.

Figure 10
figure 10

Reflectance spectra for (1) annealed TiO2/CL, (2) non-annealed TiO2/CL, (3) annealed TiO2/SS, and (4) non-annealed TiO2/SS; inset: first derivative of reflectance (dR) of the different films

Full size image

The differential diffused reflectance spectra were used to determine all TiO2 films optical transition energies. The E g values were 3.49 eV for annealed TiO2/CL, 3.48 eV for non-annealed TiO2/CL, 3.52 eV for annealed TiO2/SS, and 3.50 eV for non-annealed TiO2/SS (inset of Fig. 10). The non-annealed TiO2/CL showed lower E g value. The crystal structure of the TiO2-deposited film affected its band gap value. From the above-mentioned results, we may conclude that, the thinner the band gap is, the wider the spectral range response and the more efficient the photocatalytic activity are. Obviously, the semiconductor could increase poorer photon energy and produce photoelectrons and holes [79].

Photocatalytic decolourization of amido black (AB-10B)

The photocatalytic properties of TiO2 thin films under UV light irradiation were investigated using AB-10B as the model contaminant. The photocatalytic efficiency of the coated TiO2/SS before and after functionalization of the substrate was tested against the decolourization of AB in aqueous solution at natural pH. The decolourization during experimentation is presented in Fig. 11. The CL alone showed a limited photocatalytic efficiency of 13.63%.

Figure 11
figure 11

Amido black degradation as a function of time for the different photocatalysts under illumination

Full size image

According to the obtained curves in Fig. 11, the relationship of (Abs/Abs0) with respect to irradiation time suggested the first-order model. The effectiveness of dye decolourization for each TiO2 photocatalyst over SS or CL and after 300 min of illumination is presented in Table 6.

Table 6 Percentage of decolourization, correlation coefficients (R 2), and rate constant (k min−1) of the apparent first-order kinetic model for photocatalytic degradation of AB-10B over various photocatalysts
Full size table

It appeared that the most active one is non-annealed TiO2/CL. The degree of AB-10B decolourization, in this case, was 63.53% corresponding to an increase in photocatalytic efficiency in the presence of CL. The existence of the hematite from the CL affected significantly the percentage of the decolourization. Therefore, for the annealed film, the photocatalytic activity decreased. This behavior can be attributed mainly to the growth of the crystalline size upon annealing. In fact, decreasing the specific area and the transformation of anatase to rutile (Table S1) produced less photocatalytic activity. This finding was confirmed by several authors [36, 80].

Generally, photocatalysis efficiency is a complex process depending on all the presented physical parameters (crystalline form, crystal size, electron–hole recombination, hydrophobic aspect, etc.). In photocatalytic applications, the anatase phase is more efficient than rutile because of its more open structure. The apparent kinetic constant presented a linear dependency with the band gap energy E g of studied electrodes (Fig. S3a). In addition, decoloration efficiency increased as the anatase percentage rose (Fig. S3b). With increasing WA, the integrated intensity fell down. The change in the structure of the films affects the PL intensity. The impurity levels can augment the electron–hole pair separation rate in TiO2 films. They are generated with oxygen vacancies and interstitial oxygen defects. As the redox reactions might take place on the surface of oxygen vacancies and interstitial oxygen defects, the oxygen defects can be considered to be the active sites of the TiO2 photocatalyst [81,82,83]. As shown in Fig. 9, the PL emission intensity increases with annealing temperature as well as CL absence, suggesting the reduction in the oxygen defects in TiO2 films. Therefore, the photocatalytic activity of the TiO2 films could not be improved. Obviously, the abundant surface oxygen vacancies or defects exist in TiO2 films, which may play a vital role in the photocatalytic activity.

The TiO2 over a CL had the maximum anatase mass fraction as indicated in Table S3. It also presented the less PL intensity, which can be contributed to its high degradation percentage.

As far as we know, the higher activity of the rutile–anatase mixture of Degussa P25 is due to a typical electron transfer from the conduction band of the anatase to that of rutile [70,71,72,73]. Consequently, the anatase–rutile mixed phase blocks the electron–hole recombination by locating electrons and holes in altered crystalline phase. This finding agrees with the results obtained by depositing the mixture of anatase, rutile, and brookite of our synthesized powder over a CL. In our study, the CL (Fe2O3) enhanced the photocatalytic activity of TiO2 coating.

The ternary mixtures of the TiO2 polymorphic phases were more efficient together with the presence of the hematite form. Hematite Fe2O3 was able to absorb photons from the green–blue part of the spectrum between 300 and 600 nm. Its small band gap of 2.1 eV makes it one of the most promising semiconductors for photocatalysis [74]. It was noticed that, among all samples, the non-annealed thin film with an interlayer of Fe2O3 showed the best result. Figure S4 reveals a schematic illustration of energy band and photogenerated charge carriers for the TiO2 and Fe2O3, which can explain the good photocatalytic activity of TiO2/CL and assign it to the creation of TiO2/Fe2O3 heterojunction.

These results were discussed in terms of energy band structure and microstructure [75]. The energy band and the photogenerated charge carriers for the TiO2 and Fe2O3 can explain the improved photocatalytic activity and assign it to the creation of TiO2/Fe2O3 heterojunction [84,85,86]. As shown in the energy diagram of the interface in Fig. S4, the photoinduced electrons generated in the conduction band (CB) of TiO2 favored the migration to the Fe2O3 phase having a lower CB potential. The electrons will be expended in a reduction process. The elementary field in TiO2/Fe2O3 bilayer and electrons in the valence bands (VB) of TiO2 were driven into the VB of Fe2O3, while photogenerated holes moved into the VB of TiO2 in an opposite direction [87]. Consequently, these holes would powerfully oxidize a mobile species and the charge transport between VBs of Fe2O3 and TiO2 can be considered as an effective process for stimulating the photocatalytic activity of the TiO2/Fe2O3 thin film electrodes, since this result indicates that the charge recombination was reduced in the process of electron transport. The TiO2 film over the CL showed crack as illustrated before by SEM images. Therefore, it can create sufficient defect states without blocking the hole transfer.

Furthermore, compared to the annealed hematite/TiO2 thin film, the small crystal size in the non-annealed one decreased the recombination occurring within the crystals. Besides, it contributed to an additional increase in the photogenerated holes reaching the interface of TiO2 and the pollutant.

The presence of organic intermediates was in relation to the level of mineralization achieved during the photocatalytic decolourization of AB-10B. Their absence represented a proof of the complete destruction of organic compounds in water, which is of great importance in water treatment. Mineralization level can be evaluated by COD measurement. The COD change during photocatalysis was studied in the case of the electrode that showed the best photocatalytic activity, which is non-annealed TiO2/CL as demonstrated in Fig. 12b. The initial concentration was 10 mg/l at natural pH. The COD removal of the dye after 5-h irradiation with UV light is presented in Fig. 12a. Giving to the obtained curve, the relationship of ln(COD/COD0) with respect to irradiation time suggested the first-order kinetic, where the apparent rate constant is found to be 3.31 10−3 min−1 with a correlation coefficient of 0.987.

Figure 12
figure 12

a COD evolution during AB-10B photocatalytic decolourization compared to the percentage of decolourization in case of non-annealed TiO2/CL; b kinetics of AB decolourization over non-annealed TiO2/CL compared to its mineralization

Full size image

The percentage of COD reduction was 65.00% after 300 min equivalent to 63.53% of dye decolourization. With the extended irradiation time, we could possibly achieve a complete mineralization. After materials characterization, the application of the synthesized TiO2 films was explored in the photocatalysis degradation of AB-10B. The current samples are compared to other reports on TiO2 deposited with different and more common techniques. In the literature, a few modern studies [88,89,90,91,92,93] considered the mechanical properties of TiO2 layers using nano-indentation method. Nevertheless, in this work, mechanical properties were combined with photocatalytic characteristics. Table 7 presents an extensive literature overview of previous works on TiO2 properties.

Table 7 Techniques, conditions, and experimentally determined values published in the literature for mechanical and photocatalytic properties of TiO2 coatings
Full size table

These TiO2 thin films were deposited, for instance, by sol–gel dip-coating method [88, 89], pulsed laser deposition [90], filtered arc deposition [91], electrochemical deposition [92], sputtering [93], or EPD [16, 17] on a variety of substrates. The existing hardness, Young’s modulus data, and photocatalytic properties reported the average data of the different TiO2 films. Relying on the characteristics of the films and the type of substrates, the values of hardness and Young’s modulus values were between 22 MPa and 19 GPa for hardness, and from 7 to 209 GPa for Young’s modulus. In our case, the values of hardness and Young’s modulus values were more similar to those presented in references [16, 17]. This may indicate that these comparable properties are linked to the deposition method (pulsed EPD) due to its important influence on the films morphologies. The photocatalytic activities were affected also by the variation of the film thickness and the particles size of the TiO2 film as well as the presence of Fe2O3 as mentioned before (heterojunction effect). Consequently, more photocatalytic action can be achieved.

Conclusion

The microstructure of TiO2 films considerably affected the photocatalytic activity. In order to have an extremely active photocatalytic thin layer, new electrode materials were considered. This study examined the photocatalytic activity of nano-structured TiO2 thin films to be used in the degradation of organic pollutants. The effect of thermal annealing and the CL on the structural, morphological, and photocatalytic properties of nano-structured TiO2 thin films was also investigated.

A novel supported photocatalyst on SS was prepared by TiO2 deposition using pulsed EPD. The used TiO2 powder was prepared by hydrolysis of TiCl4 followed by dialysis. An interlayer CL composed mainly of Fe2O3 was formed on the SS surface with the intention to strongly improve the film adhesion. The electronic interaction between CL and the TiO2 coating was also marked to decrease electron–hole recombination increasing TiO2 layer efficiency.

The as-deposited nano-structured TiO2/CL thin film prepared by pulsed EPD exhibited highly enhanced photocatalytic activity toward UV light-driven photocatalytic degradation of AB dye in water, as compared to the annealed samples. This enhanced photocatalytic activity was attributed to the combined effects of three-phase structure, TiO2/Fe2O3 heterojunctions, and smaller size of TiO2 nanoparticles providing higher surface area for higher dye adsorption and efficient photocatalytic degradation upon UV irradiation.

The annealing treatment did not bring an improvement in the photocatalytic activity because of the growing size of TiO2 nanoparticles and the important change in the crystal structure of TiO2 thin film. The non-annealed TiO2/CL presented a COD removal of 65.00% and a dye decolourization of 63.53% after 5 h.

The current technique can be considered as a simple approach and low-cost heading for synthesizing an active photocatalyst. The deposited film had advanced mechanical properties and showed no adhesion difficulty. These characteristics are very promising in photocatalytic coatings for practical environmental applications.