Effect of B₂O₃ addition on optical and structural properties of TiO₂ as a new blocking layer for multiple dye sensitive solar cell application (DSSC)

Abstract

TiO₂–B₂O₃ sol–gel films were prepared using titanium(IV) isopropoxide as a Ti source and boric acid (H₃BO₃) as B precursor. B₂O₃ doping was found to improve the characteristics of TiO₂ films making them suitable to use as blocking TiO₂ layers in DSSC B₂O₃ acted as a flux and glass forming oxide leading to amorphous vitreous layers having an average thickness of 50 nm. Films were transparent, adherent and perfectly continuous without any leakage current. X-ray diffraction measurements prove that crystallinity decreases with boron content meaning that amorphous phase was favoured. Surface morphology was investigated by atomic force microscopy. It showed that film surface became more and more smooth. The E_g Raman-active phonon mode at 145 cm⁻¹ reveals the same arrangement of TiO₂ octahedra observed in the anatase phase. The Lorentzian multipeak fitting showed the emergence of a new mode at 152 cm⁻¹ whose mode intensity increased with boron content; we attributed it to the presence of boron. The influence of B₂O₃ dopant on the optical properties was examined by UV-visible spectroscopy and spectroscopic ellipsometry. Refractive index, extinction coefficient and optical band gap have been extracted by fitting ellipsometric spectra with the double new amorphous model. Difference between the optical gap values obtained from UV-visible spectra and those calculated by ellipsometry did not exceed 0.3 eV. The optical band gap increased from 3.4 to 3.9 eV by increasing boron content from 0 to 20%. The increase of E_g is expected to induce an enhanced output ddp into DSSC.

---

1. Introduction

Sol–gel technology is a convenient method to prepare thin glass coatings, especially for a variety of either ceramic or metallic surfaces. It remains the cheapest method compared to chemical vapor deposition, pulsed laser deposition and reactive sputtering techniques. The sol–gel method has been used to deposit various metal oxide materials such as SiO₂,¹ SnO₂,² ZrO₂,³ ZnO⁴ and TiO₂.⁵ Recently, this process has also been used to deposit perovskite compounds such as BaTiO₃.⁶ Sol–gel technology allows contol of the microstructure and film thickness. Among the different oxides, TiO₂ is currently the most interesting. Anatase films are used in many industrial applications, particularly as a photocatalyst for water,⁷ air purification⁸ and for self-cleaning surfaces.⁹ Additionally, it has antibacterial properties due to its high oxidation activity and superhydrophilicity.¹⁰ In the dye-sensitized solar cell (DSSC) field, the purpose of our study, TiO₂ was used since the first studies because of its high energy gap.¹¹ A standard DSSC is composed by a layer of transparent conducting oxide (TCO) made usually with a fluorine-doped tin-oxide (FTO) deposited on a glass substrate. Then a nanostructured layer of TiO₂ porous oxide on which the dye is adsorbed. The latter is full with liquid or solid electrolyte. At least there is the counter electrode. Various efforts have been made to improve DSSCs' efficiency such as changing TiO₂ particle morphologies¹² and adding an insulating layer situated between fluorine doped tin oxide (FTO) and the porous titanium oxide. This work is focused on the fabrication of such a layer named also “dense layer”, or “blocking layer” or “compact layer” whose role is to improve the adhesion of the nanoporous TiO₂ to FTO substrate, to avoid current leakage and to prevent electron recombination between TiO₂ and FTO.¹³ Many works were devoted to this topic. They used different metal oxides such as SiO₂,¹⁴ Al-doped ZnO¹³ and TiO₂,¹⁵ which remains the most used material for this purpose. TiO₂ compact layer was elaborated by pre-treatment of FTO with TiCl₄ on the transparent electrode,¹⁶ RF and pulsed DC magnetron sputtering.¹⁷

In view of the desired aims, to obtaining a transparent, continuous and perfectly insulating layer, all methods are possible although the benefits and disadvantages depend on the discretion of each. The use of TiCl₄ is very difficult because of its high reactivity with and even with air humidity, so the TiCl₄ bottles fume on opening in air laboratory and a bulk white precipitate occurs when it goes in contact with insufficient dried solid surfaces. RF and magnetron sputtering need high vacuum installations and are costly especially if large surfaces are wanted.

In this work, we choose to use the sol–gel technique which is simple and allows the control of thickness as well as the introduction of doping oxides. We propose also introduce B₂O₃ as doping oxide. B₂O₃ doped TiO₂ layers were studied in two references, the first deals with the use B₂O₃–TiO₂ layers to degrade organic pollutants under visible light radiations¹⁸ and the second concerns the fabrication of the porous TiO₂ layer of DSSC cells¹⁹ In fact, the presence of B₂O₃ into the compact layer is more justified than its presence into the porous layer. This oxide is commonly used as a fluxing agent into the fabrication of ceramic enamels and glasses.^20,21 Its introduction into the blocking layer will lower the melting temperature and facilitate the napping of substrate area. The layer thickness will be decreased and the transparency and the adherence will be increased.

Many works reported the change of optical and physical properties induced by doping the TiO₂ layer. In general we note an improve of the overall energy conversion efficiency of the DSSC. Sangwook Lee studied Nb-doped TiO₂ as a new compact layer. As a result, the efficiency of the DSSC with Nb-doped TiO₂ and 80 nm thick enhanced by 4.1% compared to DSSC with undoped TiO₂.²² Zn-doped TiO₂ blocking layer were also investigated by Thanh-Tung Duong, the increase of efficiency with 20 nm thick Zn doped-TiO₂ blocking layer showed was about 1.32%, compared to undoped TiO₂ (20 nm).²³ Doping induce also a change into the gap energy of TiO₂. Suk In Noh reported that F-doped TiO₂ compact layer show a band gap equal to 3.8 eV instead of 3.56 eV for undoped TiO₂ film.²⁴

This work is focused on the synthesis of a compact TiO₂ layer which is adherent to glass substrate, transparent and homogeneous. We choose to use TiO₂ doped with up to 20% mol B₂O₃. This layer was deposited directly on glass substrate in order to separate the observed phenomena. Obviously, the presence of the transparent conducting oxide is necessary for further applications. However, there are now several possibilities to choice the conducting layer and the success of the concordance of the TiO₂ layer and the glass substrate will remain whatever the choice of the interlayer between them mainly because the fact this layer is extremely thin and chemically inert towards the adjacent compounds. In addition the conducting layer does not affect the optical properties of the doped TiO₂ layer but it may complicate the treatment of the results.

2. Experimental method

2.1. Preparation of B₂O₃-doped TiO₂ thin films

Boron-doped TiO₂ thin films were prepared by sol–gel process spin coating method in three steps. The aim of the process is to produce a homogeneous solution containing TiO₂–B₂O₃ precursors. The first step is to prepare an alcoholic B₂O₃ solution (A solution) by dissolving 20 g of boric acid (H₃BO₃) in 100 g of methanol.

The second step is to obtain a homogeneous TiO₂ solution (B solution). This was achieved by mixing ethanol as solvent, acetic acid as a stabilizing agent and titanium isopropoxide as a precursor. The final dipping solution was got by mixing two volumes V₁ and V₂ from A and B solutions respectively. Study of the boron effect was achieved using the TiO₂ solution doped with x% B. x is the atomic percent of the B ion and is defined as x = [B/(Ti + B)]/100. x took the values 0, 5, 10, 15, and 20%. Subsequently, different B₂O₃ solutions were added into TiO₂ under vigorous stirring. The resultant dip solution remains transparent after use.

The different films were deposited on glass substrates which are carefully cleaned by ultrasonic treatment with acetone, ethanol and dichloromethane each for 10 min.

The boron-doped TiO₂ coatings were prepared by spin-coating method with fixed spin speed of 3000 rpm for 1 min. All films undergo a heat treatment begins by drying at 100 °C for 10 minutes and ends by annealing in furnace at 500 °C.

2.2. Characterization methods

The structure of the boron-doped TiO₂ films was examined by X-ray diffraction using a Cu Kα radiation (1.5406 A) Bruker D8 Advance diffractometer. Raman spectroscopy was performed with a LabRam HR and T64 000 spectrometer. The Lorentzian fitting was performed using the Fityk software.

The surface morphology of the TiO₂–B₂O₃ thin films was carried out using an Auto Probe CP-Research AFM (Thermo-microscopes) with a lateral resolution of 5 nm. UV transmittance analyses were monitored by a near-infrared to ultraviolet Spectrophotometry (NIR-UV-VIS) PerkinElmer type in the wavelength range of 200–2250 nm. Ellipsometric spectra were recorded at room temperature using a phase modulated spectroscopic ellipsometer (PMSE) from JOBIN-YNON HORIBA model UVISELTM in the wavelength range of 190–2100 nm. The chosen configuration for the modulator and analyzer positions were M = 0° and A = 45° respectively.

The values obtained for TiO₂–B₂O₃ film thicknesses were estimated by a profilometer VeecoTM Dektak 6M and confirmed by ellipsometry. The photoluminescence (PL) spectra was measured with an excitation (360 nm) source using a 1000 W Xe arc lamp coupled to a monochromator as excitation source at room temperature.

3. Results and discussion

3.1. Structural analysis

Fig. 1 presents the XRD spectra of samples with different boron doping concentrations. We observe for pure titanium oxide thin films a broad peak at 2θ = 25° (Fig. 1a). Owing to literature, this peak is attributed to anatase titanium oxide phase. The large width of this peak indicates a poor crystallization state. As the film thicknesses of all the samples are almost similar, the intensity of this peak decreases with increasing boron contents which indicate that boron dopant promotes the formation of an amorphous structure; this is caused by the better melting of TiO₂ films.

Fig. 1 XRD patterns of TiO₂ films with different boron contents.

The Raman spectra of TiO₂ films with different B₂O₃ contents are revealed in Fig. 2.

Fig. 2 Raman spectra of TiO₂ films with different B contents.

For pure TiO₂, only the peak at 145 cm⁻¹ appeared with a very low intensity. This peak is attributed to E_g mode of anatase phase.^25,26 After doping, the intensity of this mode increased and other Raman peaks at 125, 197 and 639 cm⁻¹ appeared.

Owing to literature, they are attributed to Raman-active modes of anatase phase with the symmetries of 3E_g.²⁶

In order to study the lattice vibrations, Lorentzian fittings of the Raman spectra for TiO₂ doped with 15% (Fig. 3a) and 20% boron (Fig. 3b) were performed into the wavenumber range from 120 to 180 cm⁻¹. Besides 2E_g modes of TiO₂ anatase (125 and 145 cm⁻¹), we show the emergence of a new peak at 152 cm⁻¹. It showed that the intensity of 2E_g modes respectively at 125 cm⁻¹ and at 145 cm⁻¹ decreased, while that of mode at 152 cm⁻¹ increased with increasing of the boron amounts. The vibration mode at 152 cm⁻¹ had not been described in literature; it is due to the presence of boron dopant.

Fig. 3 Lorentzian fittings of the spectra in the wavenumber range from 100 to 200 cm⁻¹ for the TiO₂ films doped with (a) 15% boron and (b) 20% boron.

3.2. Morphologic analysis by atomic force microscopy (AFM)

Boron doped TiO₂ films deposited on glass substrates were fully transparent to the naked eye. As shows in Fig. 4, boron–TiO₂ thin films are characterized by a decrease in surface roughness with the increase of boron dopant. The variation of the roughness and average grains size for the TiO₂ films as a function of boron contents were plotted in Fig. 5. The RMS values and the average grain size are 1.05 nm, 8.98 nm for pure TiO₂ films and 0.44 nm, 4.27 nm for the 20% B₂O₃–TiO₂ films respectively. The roughness fall indicates that the presence of boron improves the homogeneity and the smoothness of the surface with small grain size.

Fig. 4 2D and 3D micrographs of TiO₂ single layer doped with (a) 0%, (b) 5%, (c) 150%, (d) 15% and (e) 20% B₂O₃.

Fig. 5 Roughness average and size particles of TiO₂ films as a function of boron concentrations.

3.3. UV transmittance analysis of TiO₂ thin films with various boron dopant concentrations

Transmittance spectra of TiO₂ films doped with different boron contents in the wavelength region from 200 to 800 nm are given in Fig. 6. Transmittance spectra revealed that the deposited thin films are transparent in visible region. They start absorbing in the range between 270 and 300 nm. After doping with boron oxide (5%, 15%, 20%), there is a shift of the absorption threshold towards lower wavelengths.

Fig. 6 Transmittance spectra for TiO₂ thin films doped with different boron contents.

The only exception is noted for TiO₂ film doped with 10% B₂O₃, there is a red shift of the absorption threshold. This result shows that doping affects the optical gap. Starting from transmittance spectra, we can determinate the optical band gap which is a prominent feature of TiO₂. The band gap values of all samples were obtained from the model proposed by Tauc,²⁷ where E_g related to the absorption coefficient by the following equation:

αhvⁿ = A(hv − E_g)

Here α is the absorption coefficient, E_g is the absorption band gap, A is a constant depending on the transition probability, and n depends on the nature of transition. It is well known that TiO₂ possesses a direct and indirect band gap.²⁸ Most authors found that the anatase TiO₂ has only an indirect band gap, whereas rutile phase has a direct and an indirect band gap.²⁹ For an indirect band gap, the value of n is 1/2 while for the direct band gap, the value of n is 2.^30,31 In our case, XRD and Raman spectroscopy showed that anatase phase is the most probable.

The variation of (αhν)^1/2 with photon energy is shown in Fig. 7 for B₂O₃-doped TiO₂ films.

Fig. 7 (αhν)^1/2 versus the photon energy for TiO₂ thin films doped with different boron contents.

The optical gap values were evaluated from the intercept of the linear portion of each curve hν in X-axis. The obtained values of TiO₂ thin films optical band gap are with the range reported in the literature of anatase form.^24,32 From Fig. 7, it is clearly seen that the optical band gap is affected by boron contents. Doping with 5%, 15% and 20% of B₂O₃ caused the TiO₂ optical band gap shift to higher energies. This result was earlier reported in another work on mesoporous TiO₂ modified by boron doping.³³ Adriana Zaleska has also found that the value of E_g increased from 3.29 eV (pure TiO₂) to 3.4 eV with 10% of boric acid.³⁴ This was explained by the increase of the carrier concentration blocking the lowest states in the conduction band. This phenomenon is well known as the Burstein-Moss effect.³⁵ Gap values are similar to those obtained for F-doped TiO₂ blocking thin films in DSSC.²⁴ He deduced that a degenerated n-type TiO₂ layer between the substrate and the nanoporous TiO₂ layer was obtained to furnish an ohmic contact, leading to the enhanced performance of DSSCs by reducing interfacial resistance.²⁴

The gap value of pure TiO₂ was 3.6 eV and it decreases to 3.56 eV after doping with 10% B₂O₃. This redshift for TiO₂ doped with 10% B₂O₃ was observed by some authors. They assumed that the red shift of the absorption edge is derived from the mixing of B 2p states and O 2p states.³⁶ It is noted that the 10% B₂O₃–TiO₂ film revealed the best photocatalytic activity both in visible and UV light.³⁷

The values of the gap will be confirmed by ellipsometry spectroscopic.

3.4. Spectroscopic ellipsometry modeling and analysis

As it is known, spectroscopic ellipsometry (SE) determines the complex reflectance ratio ρ defined in terms of the standard ellipsometric parameters Ψ and Δ:³⁸

where R_p and R_s are the reflection coefficients for light polarized parallel (p) and perpendicular (s) to the sample's plane of incidence respectively. It is an indirect technique, which requires an appropriate optical model to fit the experimental spectra with maximum accuracy. Ellipsometry allows to determinate the influence of boron doping on the optical constants (n, k and E_g), and the thin films thickness of all samples. The best optical model adopted for B-doped TiO₂ semiconductor thin films consists of three layers: void/glass/TiO₂.

Fig. 8 reveals a good agreement between experimental and theoretical spectrum achieved for B₂O₃-doped TiO₂ films. The good fit designates that the “Double new amorphous” dispersion formula describe perfectly the optical properties of B₂O₃-doped TiO₂ films.^39,40 Table 1 recapitulates the double new amorphous parameters used to analyse ellipsometry data obtained for TiO₂ films doped with different boron contents. The refractive index (n) and the extinction coefficient (k) are definite as follows:

where

Fig. 8 Typical ellipsometry spectra and best fit of TiO₂ doped with boron of parameters (a) Δ and (b) Ψ as functions of λ with “double new amorphous” dispersion formula.

Table 1 Parameter values of the double new amorphous model used for fitting the optical functions and the thickness of the TiO₂ thin films doped with different boron contents

Parameters	0% B	5% B	10% B	15% B	20% B
X^2	2.7	1.5	1.5	1.4	3.2
ε∞	4.49	3.8	5.54	3.55	3.52
Eg (eV)	3.43	3.64	3.31	3.68	3.95
A1	0.29	0.26	0.0002	0.001	1.38
B1	7.70	6.47	8.49	5.76	5.70
C1	17.8	17.76	18.69	17.53	17.49
A2	0.32	0.01	0.33	0.18	0.14
B2	8.62	8.25	8.76	7.68	6.86
C2	19.27	17.52	21.4	14.78	13.49
Thickness by SE (nm)	52.09	51.16	55.34	50.42	45.80
Thickness by profilometer (nm)	55.3	53.2	58.7	50.6	47.3

---

The term ε_∞: is an additional parameter corresponding to the high-frequency dielectric constant. It is at least superior to one and equal to the value of the dielectric function when ω → ∞.^41–44

A, B, C: are positive non zero parameters referring to the electronic structure of the material. For parameters A_n, B_n, C_n, n refers to the number of oscillators in our case n = 2 for double new amorphous and the parameters number is 8.

A_n (in eV): is related to the dipole matrix squared and gives the amplitude of the extinction coefficient peak. Generally, 0 < A_n < 2.

B_n/2 (in eV): is approximately the energy at which the extinction coefficient is maximum (peak of absorption). As the value of B increases, the absorption peak is shifted towards the UV region. Generally, 3 < B_n < 30.

C_n (in eV²): is related to the broadening term of the absorption peak. It depends on the energy difference between different states and on the lifetime of transition. Generally, 3 < C_n < 150.

3.4.1. Thickness, refractive index (n) and extinction coefficient (k). Firstly, it must be noted that ellipsometric measurements and optical parameters (n, k) of B₂O₃-doped TiO₂ thin films have not yet studied.

From Table 1, it is seen that the thicknesses determined by the fitting are in good agreement with the result obtained by profilometer DEKTAK. Fig. 9 shows a cross section realized by AFM for TiO₂ doped with 5% B₂O₃. The thickness is equal to 53.98 nm by comparing it with that found by ellipsometry (53.2 nm). It was found that the film thickness corresponded with the profilometer and ellipsometry technique. We observe that difference between the values obtained by the three methods does not exceed 2 nm. Table 1 show that thickness values decrease slightly from 52 to 45 nm when the boron contents increase from 0 to 20%. This slight decrease is due to the better melting of TiO₂ in presence of boron oxide, which was confirmed by the AFM. Fig. 10 shows the variation of refractive index (n) and extinction coefficient (k) respectively as a function of wavelength at different boron contents, calculated from the extracted best-fitted parameters. Spectra showed a similar shape than cited in literature.^45,46 We note that n and k are considerably affected by boron doping concentration. These optical factors increase with increasing of wavelength up to threshold with a maximum value that corresponds to the band gap. After doping with 10% B₂O₃, the maximum of n (n_max) undergoes a shift to the high wavelengths, but, for the remaining of samples (5%, 15% and 20% B₂O₃), this peak shifted to lower wavelengths. This result indicated the change of optical gap with doping, which is in good agreement with the results obtained by UV-visible spectroscopy. Fig. 10b demonstrates that all the samples have dielectric thin films.⁴⁷ For all samples; values of refractive index are greater than 2. These higher values make this material suitable for antireflection coatings.⁴⁸ After 370 nm, refractive index described by a gradual decrease when the wavelength increases for each TiO₂ thin films inclosing various boron doping. Moreover, the relation of Clausius Mossotti explained the correlation between the refractive index (n), the polarizability (α_m) and the density (ρ) by the following equation:

Fig. 9 Cross section AFM image of TiO₂ thin film doped with 5% B₂O₃.

Fig. 10 The calculated refractive index spectra (a), extinction coefficient spectra (b) of TiO₂ thin films with different B₂O₃ contents, based on the results of the double new amorphous model fitting.

where N_a is Avogadro's constant and M is the molecular weight. The density of different films was calculated from the values of the refractive index at 550 nm. As seen in Fig. 11b, refractive index (n₅₅₀) and density values are 2.31, 0.92 cm³ for pure TiO₂ films and 2.12, 0.82 cm³ for 20% B₂O₃–TiO₂ films respectively. It is clearly seen in Fig. 11a that refractive index decreases when boron contents increase from 0 to 20%. Obtained n₅₅₀ values are in good agreement with literature values of anatase TiO₂ films.²⁶ This result confirms that the decrease in refractive index is actually due to the decrease of density. Correspondingly, the value of the refractive index of our samples is smaller than the ones characteristic for the anatase phase; which is reported to be approximately 2.5. Several authors attributed the low values of the refractive index to the lack of cristallinity.^39,49 XRD results confirmed the low of cristallinity for TiO₂–B₂O₃ films (Fig. 12).

Fig. 11 The variation of refractive index of TiO₂ thin films (a) with various B-doped concentrations at 550 nm and (b) as function of density.

Fig. 12 Room-temperature PL emission spectra of TiO₂ film doped with different boron contents (a) and variation of PL intensity as a function of optical band gap of boron-doped TiO₂ films (b).

3.4.2. Discussion of the optical band gap values obtained by UV-visible spectroscopy and that by ellipsometry. As shown in Table 2, optical band gap values obtained both by ellipsometry and by UV-visible spectroscopy are nearby to literature values.⁵⁰ Ellipsometry reveals that optical band-gap energy progressively increases from 3.4 eV to 3.95 eV with the increase of the boron concentration from 0% to 20%. We note an exception for 10% B₂O₃ where the gap decreases.

Table 2 Gap values for TiO₂ thin films doped with different boron contents

% B2O3	Gap by ellipsometry (eV)	Gap by UV-vis spectroscopy (eV)
0	3.34	3.6
5	3.68	3.64
10	3.31	3.56
15	3.69	3.66
20	3.95	3.70

---

This evolution is in accordance with the findings acquired by UV-visible spectroscopy. It was demonstrated that the increment of band gap for the TiO₂–B₂O₃ (5%, 15% and 20%) is due to the Burstein-Moss effect.⁴⁸ With widening boron doping, Fermi level shifts into the conduction band of the TiO₂–B₂O₃ films. Since, the donor electrons occupied the states underneath Fermi level, optical transitions can only occur between the valence band and the states above Fermi level in the conduction band, which results in the increment of the optical gap.⁵⁰ It is difficult to select which of two methods is more accurate. In reality each of two methods postures an accuracy problem.

In the UV-visible spectroscopy case, plotting the tangent to the curve (αhν)_n = f(hν) can easily induce errors. For ellipsometry, calculating E_g uses parameters that cannot be with accuracy.

In our case we found a difference of 0.3 eV which is not always in the same case.

3.5. Photoluminescence (PL)

Photoluminescence spectroscopy is employed to examine the electronic structure, the photo generated electron–hole pairs in semiconductors and the rate of recombination. The room temperature PL spectra obtained under excitation at 360 nm for TiO₂ thin films with different B₂O₃ concentrations (0–20%). Fig. 11a shows that all PL emission peaks are observed at about 544 nm and another smaller peak at 485 nm. The two peaks were attributed to the transitions from oxygen vacancies with one-trapped electrons and two-trapped electron to valence band of TiO₂ respectively.⁵¹

The corresponding emission peaks are respectively 2.28 and 2.56 eV. This indicates that the position of the energy levels associated with the two types of oxygen vacancies is respectively at 0.8 and 0.44 eV beneath the conduction band of TiO₂ respectively.⁵¹ Several factors can affect the PL intensity. Shuai Chen shows that the decrease of PL intensity is related to the better cristallinity.⁵² This finding is in good agreement with our results. As we have defined the values of the optical gap by ellipsometry, we note in Fig. 11a that quenching or enhancement intensity PL is related to the gap values. A high intensity was obtained for gap value of about 3.9 eV for TiO₂ doped with 20% B₂O₃, whereas a low intensity is observed for the low optical gap of about 3.31 eV for TiO₂ doped with 10% B₂O₃. Thus, it appears that doping of TiO₂ with boron has abundant influence on the band gap and the defect emissions of the top TiO₂ films.⁵³

4. Conclusions

B₂O₃-doped TiO₂ thin films have been prepared by sol–gel method and the coatings were annealed at 500 °C. Results indicate that the addition of boron dopant affected structural and optical properties of TiO₂ thin films. The cristallinity of TiO₂ thin films decreased when the boron oxide content increased from 0 to 20% explained by the flux effect of B₂O₃. The Raman spectra show that the E_g Raman-active phonon modes appeared at 145 cm⁻¹, 199, 638 cm⁻¹ justified the presence of anatase phase. The absorption edge of TiO₂ films shifted towards lower wavelengths from 370 to 309 nm with the increase of boron contents. Extinction coefficient and refractive index are affected by boron oxide content. As demonstrated by UV-visible spectroscopy and the spectroscopic ellipsometry, the band gap energy of TiO₂ thin films increases from 3.4 to 3.9 eV when the boron contents increase from 0 to 20%. The room temperature PL spectra suggest that the emission intensities increase with increasing boron contents. It is showed that the positions peaks are situated at about 2.28 and 2.56 eV. These are assigned to the transitions from oxygen vacancies with one-trapped electrons and two-trapped electron to valence band of TiO₂ respectively. The present results will be helpful for future applications of TiO₂-based semiconductor devices, especially, as a blocking layer in multiple dye solar cells application, to prevent the back electron transfer and resulted in the reduced interfacial resistance.

Acknowledgements

This work was realized in collaboration between the photonics laboratory (LPHIA) from Angers, France and the Laboratory of applied Mineral chemistry (UR11ES18), Tunisia. The authors acknowledge and thank Miss Emmna Kadri from applied physics laboratory faculty of science Sfax, Tunisia for her help in the AFM observations and for the photoluminescence measurements.

Notes and references

1. H. Tung Li, C. A. Kuo and H. W. Yao, J. Nanosci. Nanotechnol., 2013, 13, 279–287.
2. M. Marikkannan, V. Vishnukanthan, A. Vijayshankar, J. Mayandi and J. M. Pearce, AIP Adv., 2015, 5, 027122.
3. S. Seidel, A. Sabelfeld, R. Strohmeyer, G. Schreiber, V. Klemm, D. Rafaja, Y. Joseph and J. Heitmann, Thin Solid Films, 2016, 606, 13–18.
4. T. Demes, C. Ternon, D. Riassetto, H. Roussel, L. Rapenne, I. Gélard, C. Jimenez, V. Stambouli and M. Langlet, J. Phys. Chem. Solids, 2016, 95, 43–55.
5. G. Kenanakis, D. Vernardoua, A. Dalamagkas and N. Katsarakis, Catas. Today, 2015, 240, 146–152.
6. L. Huang, Y. Dain, H. Xiao, X. Pei and W. Chen, Ceram. Int., 2016, 42, 9046–9050.
7. E. Boonen and A. Beeldens, J. Coat., 2014, 4, 553–573.
8. D. Mardare, A. Yildiz, M. Girtan, A. Manole, M. Dobromir, M. Irimia, C. Adomnitei and N. Cornei, J. Appl. Phys., 2012, 112, 073502.
9. S. R. Meher and L. Balakrishnan, Mater. Sci. Semicond. Process., 2014, 26, 251–258.
10. S. M. Kim, I. In and S. Y. Young Park, Surf. Coat. Technol., 2016, 294, 75–82.
11. L. Kavan and M. Gratzel, Electrochim. Acta, 1995, 40, 64–652.
12. J. Akilavasan, M. Al-Jassim and J. Bandaraa, J. Nanophotonics, 2015, 9, 1–12.
13. D.-Y. Chen, J.-Y. Kao, C.-Y. Hsu and C.-H. Tsai, J. Electroanal. Chem., 2016, 766, 1–7.
14. L. Liang, Y. Liu and X.-Z. Zhao, Chem. Commun., 2013, 49, 3958–3960.
15. Y. Chen, H. Zhang, D. Fox, C. C. Faulkner, D. Jeng and M. Bari, Mater. Sci. Eng., B, 2013, 71–76.
16. N. G. Park, G. Schlichthörl, J. van de Lagemaat, H. M. Cheong, A. Mascarenhas and A. J. Frank, J. Phys. Chem. B, 1999, 103, 3308–3314.
17. Q. Q. Liu, D. W. Zhang, J. Shen a, Z. Q. Li, J. H. Shi, Y. W. Chen, Z. Sun, Z. Yang and S. M. Huang, Surf. Coat. Technol., 2013, 231, 126–130.
18. X. Lu, B. Tian, F. Chen and J. Zhang, Thin Solid Films, 2010, 519, 111–116.
19. H. Tian, L. Hu, C. Zhang, S. Chen, J. Sheng, L. Mo, W. Liua and S. Dai, J. Mater. Chem., 2011, 21, 863–868.
20. C. R. Kurkjian and W. R. Prindle, J. Am. Ceram. Soc., 1998, 81, 795–813.
21. X. Wang, G. Wu, B. Zhou and J. Shen, ACI Mater. J., 2013, 6, 2819–2830.
22. S. Lee, J. Hong Noh, H. Soo Han, D. Kyun Yim, D. Hoe Kim, J.-K. Lee, J. Y. Kim, H. Suk Jung and K. S. Hong, J. Phys. Chem. C, 2009, 113, 6878–6882.
23. T.-T. Duong, H.-J. Choi and S.-G. Yoon, J. Alloy Compd., 2014, 591, 1–5.
24. S. I. Noh, K.-N. Bae, H.-J. Ahn and T.-Y. Seong, Ceram. Int., 2013, 39, 8097–8101.
25. M. C. Kao, H. Z. Chen, S. L. Young, C. Y. Kung and C. C. Linc, Thin Solid Films, 2009, 517, 5096–5099.
26. W. F. Zhang, M. S. Zhang, Z. Yin and Q. Chen, J. Appl. Phys. B, 2000, 70, 261–265.
27. J. Tauc, R. Grigorovici and A. Vancu, Phys. Status Solidi, 1966, 15, 627.
28. K. H. Leong, H. Y. Chu, S. Ibrahim and P. Saravanan, Beilstein J. Nanotechnol., 2015, 6, 428–437.
29. S. Valencia, J. M. Marín and G. Restrepo, Open Mater. Sci. J., 2010, 4, 9–14.
30. S. J. Darzi, A. R. Mahjoub and A. Nilchi, Phys. E, 2009, 42, 176–181.
31. Y. Zhao, C. Li, X. Liu, F. Gu, H. Jiang, W. Shao, L. Zhang and Y. He, Mater. Lett., 2007, 61, 79–83.
32. P. Karasin ski, E. Gondek, S. Drewniak, A. Kajzer, N. Waczyn ska-Niemiec, M. Basiaga, W. Izydorczyk and Y. E. L. Kouari, Opt. Mater., 2016, 56, 64–70.
33. M. A. Islam, M. J. Haither, I. Khan and M. Islam, J. Electr. Electron. Eng., 2012, 3, 18–24.
34. H. Tian, L. Hu, C. Zhang, S. Chen, J. Sheng, L. Mo, W. Liua and S. Dai, J. Mater. Chem., 2011, 21, 863–868.
35. A. Zaleska, J. W. Sobczak, E. Grabowska and J. Hupka, Appl. Catal., B, 2008, 78, 92–100.
36. W. Zhao, W. Ma, C. Chen, J. Zhao and Z. Shuai, J. Am. Chem. Soc., 2004, 126, 4782–4783.
37. X. Lu, B. Tian, F. Chen and J. Zhang, Thin Solid Films, 2010, 519, 111–116.
38. R. Oommen, P. U. Rajalakshmi and S. Sudha, Int. J. Appl. Phys. Math., 2012, 2, 439–441.
39. H. G. Tompkins and E. A. Irene, Handbook of Ellipsometry, Springer-Verlag, 2005.
40. A. V. Manole, M. Dobromir, M. Girtan, R. Mallet, G. Rusu and D. Luca, Ceram. Int., 2013, 39, 4771–4776.
41. E. D. Palik, Handbook of Optical Constants of Solids II, 1998, ch. 7.
42. A. R. Forouhi and I. Bloomer, Phys. Rev. B: Condens. Matter Mater. Phys., 1986, 34, 7018–7026.
43. A. R. Forouhi and I. Bloomer, Phys. Rev. B: Condens. Matter Mater. Phys., 1988, 38, 1865.
44. N. Bsiri, M. A. Zrir, A. Bardaoui and M. Bouaïcha, Ceram. Int., 2016, 42, 10599–10607.
45. D. Saha, R. S. Ajimsha, K. Rajiv, C. Mukherjee, M. Gupta, P. Misra and L. M. Kukreja, Appl. Surf. Sci., 2014, 315, 116–123.
46. C. M. Maghanga, G. A. Niklasson and C. G. Granqvist, Thin Solid Films, 2009, 518, 1254–1258.
47. A. Dakka, J. Lafait, C. Sella, S. Berthier, M. Abd-Lefdil, J. C. Martin and M. Maaza, J. Appl. Opt., 2000, 39, 2745–2753.
48. D. Saha, R. S. Ajimshaa, K. Rajiv, C. Mukherjee, M. Gupta, P. Misra and L. M. Kukreja, Appl. Surf. Sci., 2014, 315, 116–123.
49. W. Chiappim, G. E. Testoni, R. S. Moraes, R. S. Pessoa, J. C. Sagas, F. D. Origo, L. Vieira and H. S. Maciel, Vacuum, 2016, 123, 91–102.
50. Q. H. Lia, D. Zhua, W. Liua, Y. Liuc and X. C. Ma, Appl. Surf. Sci., 2008, 254, 2922–2926.
51. J. Yuan, E. Wang, Y. Chen, W. Yang, J. Yao and Y. Cao, Appl. Surf. Sci., 2011, 257, 7335–7342.
52. S. Chen, X. Zhao, H. Xie, J. Liu, L. Duan, X. Ba and J. Zhao, Appl. Surf. Sci., 2012, 258, 3255–3259.
53. X. D. Zhou, X. H. Xiao, J. X. Xu, G. X. Cai, F. Ren and C. Z. Jiang, EPL, 2011, 93, 57009.

---