1. Introduction

Titanium dioxide (TiO₂) a white inorganic compound, is a naturally occurring oxide of titanium (Ti) a transition element with atomic number 22. Its outstanding non-toxic nature and stability [1] in terms of its opacity, electrical and physical properties, chemical structure and biocompatibilityhas made it toexcel in clusters such as biomedicine as anti-bacterial agents [2], photocatalysis [3,4,5], in dye sensitized solar cells (DSSC) as a photoanode and counter electrode [6], gas sensing, self- cleaning surfaces, water sputtering devices among other industrial applications [7]. The structure of TiO₂ comprises of Ti-O-Ti stretching vibrations emanating from the TiO₆ octahedral structure with different sharp orientations in edges and corners which gives it distinct properties including but not limited to specific surface area different phasestructureandcrystal size [8]. These features have caused the TiO₂ to be termed as polymorphic existing in three phases that is brookite, rutile and anatase [9]. The structure of rutile and anatase is a tetragonal crystalline structure while brookite takes orthorhombic crystalline structure [9]. Anatase has a refractive index of 2.49 and density of 3.9g/cm³, while rutile has a refractive index of 2.61 and density of 4.3 g/cm³ and brookite having a refractive index of 2.58 and density 4.1 g/cm³ [10,11]. The most stable thermodynamic phase is rutile while brookite and anatase are metastable where when subjected to different conditions including but not limited to annealing temperature, readily transform to rutile phase [12]. Other conditions that transform the anatase and brookite phases include the type of starting materials, its composition and deposition method [13]. Brookite-anatase, anatase-rutile or brookite-anatase-rutile mixed forms are other phases that can be obtained inTiO₂NPs [14].

When nanoparticles are to be applied key properties such as crystal phase, morphology and crystallite size are highly considered [15]. Therefore, in the application of TiO₂ in solar cells and as a photocatalyst the polymorph that suits the above threshold is the anatase having an optical band gap (E_g) of 3.2eV. The E_g of brookite is found to be close to that of anatase showing potentiality of it being used as photocatalyst however, few reports have been made as its thermodynamic instability emanating from its metastable state makes it to be looked down upon; rutile has a band gap of 3.0 eV [16].

Hydrothermal, solvo-thermal, hydrolytic precipitation and sol-gel have been applied to synthesize desired TiO₂ NPs [10]. NPs with uniform particle size, surface properties, control of reaction time and temperature, easiness in operation and cost of synthesis are the key features to be considered when choosing a synthesis method [17]. Sol-gel matches the entire threshold making it to be widely applied. Muaz et al [18] reported a slight transformation from anatase to rutile at 500°C when thin films of TiO₂ were annealed at 300°C, 500°C for 1 hour. Bakri et al [19] reported phase transformation of anatase to rutile at temperature above 900°C after annealing thin films for 1 hour. According to M.K. Singh et al. [20] a phase transformation occurred at 800°C. Muthee and Dejene [10] reported that anatase phase occurred at 450°C, mixed-phase (anatase/rutile) at 550°C - 650°C and rutile phase at 750°C. This prompts out that annealing temperature influences TiO₂ properties and that it needs to be investigated, analyzed and reported to help study its crystal size and structure, morphology, optical and photocatalytic activities. This study examines the effects of annealing temperature for 2 hours at (400, 500, 600, 700, 800, and 900°C) on the structural, optical and morphological properties for the synthesized TiO₂ NPs using sol-gel synthesis method.

2. Methodology

2.1. Chemicals

Tetra isopropyl orthotitanate (C₁₂H₂₈O₄Ti) of purity 99.9% used as a metal precursor supplied by Sigma-Aldrich, ethanol absolute (C₂H₅OH) of Purity 99.9% as a solvent precursor supplied by Sigma-Aldrich, di-ethanolamine ((CH₂CH₂OH)₂NH) of purity 99.3% supplied by Sigma-Aldrich, deionized water (DI) (H₂O), Ammonium hydroxide supplied by A.B.Chem.Co., ltd, Hydrochloric acid of supplied by A.B.Chem.Co., ltd. All reagents were of pure analytical grade and used as received from the manufacturer.

2.2. Synthesis Procedure

Sol-gel synthesis method was employed to prepare TiO₂ NPs by adding 6mL of isopropyl orthotitanate (TIP) to 24 mL of ethanol in a clean 50 mL glass beaker placed on top of a magnetic stirrer at room temperature and stirred continuously for 30 min. Thereafter, its pH was adjusted to pH 7 using stock solution of HCl or NH₄OH with the help of a pH meter and the homogenous solution stirred for further 15 min. The contents were transferred to a hot magnetic stirrer where 6mL of diethanolamine was added drop wise to the solution and stirred for further 45 min where it will have formed the gel. The gel was dried in the oven at 100°C for 2h to evaporate the residual solvents. The formed gel was left to age for at least 12hours thereafter combusted and annealed at different temperature of 400, 500, 600, 700, 800, 900°C for 2 h. The annealed NPs were ground into a fine NPs using a mortar and stored in a sample holder for further analysis.

2.3. Characterization of the Samples

X-Ray diffractometer (XRD) measurements were made using ARL EQUINOX 100 at 40V, 0.9mA, at a scanning range of 20°-100° to obtain the crystal structure and the phase content of the synthesised TiO₂ NPs. The data obtained from the sample series was plotted, analysed and compared with the data base in Cambridge crystallographic data and the crystal sizes computed using Debye-Scherer’s equation. Surface morphology and composition of the TiO₂ NPs were analyzed using Tescan Vega 3 scanning electron microscope (SEM). A double beam UV-Vis 1800 Shimadzu model was used to analyze optical properties at a wavelength ranging 200 nm to 1100 nm at a fast scanning speed; Kubelka- Munk and Tauc equations were employed in the analysis. Fourier Transform Infrared spectrophotometer IR Spirit Shimadzu model was used to analyze the functional groups present in the samples.

3. Results and Discussion

3.1. FTIR Analysis

Figure 1 shows Infrared spectroscopy analysis for stretching and bending vibrations of the present functional groups in TiO₂ NPs obtained in the range of 400-4000 cm^-1. The infrared spectroscopy shows main bands both at the finger print region and functional group region (435, 416, 668, 1033 and 2360 cm^-1). The pronounced stretching vibration at 668 and 1033 cm^-1 is assigned to Ti-O-Ti stretching vibrations of TiO₆ where it increased as the annealing temperature increased [21,22] while 416 cm^-1 and 435 cm^-1 are related to Ti-O vibration [23]. The stretched vibration of –OH including Ti-OH absorption band were assigned to 2360 cm^-1 stretching present in the functional group band [24]. FTIR spectra confirm presence of vital functional groups of TiO₂ NPs that is Ti-O and Ti-O-Ti associated with bending from the tetrahedral structure in both anatase and rutile. The pronounced –OH at 2360cm^-1 is associated with the solvent precursor ethanol, diethanolamine, and titaniaorthotitanate forming Ti-OH stretching. Angular vibration are also witnessed between 400-500 cm^-1, 1300 to 2100 cm^-1, 3450-4000 cm^-1 making it hard for a number of peaks less easily spotted such as stretching and bending vibrations of -CH₂and -NH which are present but their bands are not so pronounced both of which are from diethanolamine (DEA).The stretching vibration -CH₂ and –NH occur in the range 2800- 3000 cm^-1 and 300-3500cm^-1 respectively [42,43]. However, the corresponding stretching vibration in the spectra is diminished which could be attributed to combustion of the gel during annealing process.

Figure 1. FTIR spectra showing the effects of annealing temperature on the functional groups of TiO2 NPs

3.2. XRD Analysis

Figure 2 shows the X-ray diffraction (XRD) patterns in the range of 20°-100° of TiO₂ NPs annealed at different temperature: 400, 500, 600, 700, 800 and 900°C. Total change of phase from anatase to rutile phase occurred at 600°C.

Figure 2. XRD patterns for TiO2 NPs (a) merged overlaid and (b) stacked graphs for samples annealed at various temperatures

The phase content of anatase and rutile was obtained using Eqs (1) and (2) [25]

(1)

(2)

Where I_A and I_R represent intensity of the anatase and rutile phases.

The most intense peak of anatase and rutile were used to calculate mole ratio between the two phases i.e. anatase and rutile using Spurr and Myers Eqs (3) and (4) [26].

(3)

(4)

Where I_A and W_A represent the intensity and mole fraction of the Anatase phase while I_R and W_R are the intensity and mole fraction of the Rutile phase respectively.

In comparison of the two methods as shown in figure 3 (a) and (b), the effects of annealing temperature shows a reciprocal of each phase with similar trends of increase in both phase content and mole fraction in rutile phase and a similar trend of decrease in phase content and mole fraction in the anatase phase [27]. This confirms transformation of TiO₂ from the decreasing phase of anatase to an increasing phase of rutile [28]. In figure 3 (a) both phase content and mole fraction of the anatase phase appear to decrease with increasing annealing temperature for which at temperature above 700°C their values are close to zero suggesting that at high temperature above 700°C anatase phase is completely transformed to rutile. In figure 3 (b) Phase content and mole fraction appears to be constant below 500°C annealing temperature and during transformation at temperature above 500°C it increases and after rutilation has occurred becomes constant suggesting that rutilation starts to happen at temperature above 500°C and at around 700°C transformation to rutile has occurred confirming visible rutile peaks of (101) at 600°C in the analyzed samples.

Figure 3. Graphical representation of phase content/mole fraction of TiO2 NPs for (a) anatase phase and (b) rutile phase annealed at different annealing temperature

Debye-Scherrer [29] Eq (5) was used to estimate the average crystal size (D) of the TiO₂ NPs

(5)

Where is the shape constant (0.9), is the X-ray wavelength represents the full width at half maximum and θ is the Bragg’s angle in degrees.

As depicted in Figure 4(a) average crystallite sizes in (nm) are inversely proportional to their corresponding FWHM values which both appear to be constant at high temperature. Figure 4 (b) shows peak intensity vs crystallite size where the peak intensity in the anatase phase is found to decrease and their crystal size increase as the annealing temperature increases from 400 to 500°C, this is attributed to transformation and phase transition or coexistence with amorphous phase [30]. The peak intensity in the rutile phase increases as the annealing temperature increases due to formation of a more thermodynamic stable phase. This generally indicates improved crystal quality at high temperature; growth temperature rearranges the atoms which in turn increases their particle size [30]. A joint variation was deduced between FWHM, peak intensity and crystallite size where crystallite size is directly proportional to peak intensity and inversely proportional to its corresponding FWHM [10]: Crystal size .

Figure 4. Plot showing comparison on the (a) FWHM and (b) peak intensity with crystallite size of TiO2 NPs as the annealing temperature is varied

The difference in the rutile and anatase phase crystallite sizes proponents that nucleation and growth of the rutile phase was initiated between 500°C and 700°C. Apparently, high surface area energy contributes to low rate transformation from anatase to rutile [31] and transformation of phases occurs between 600°C to 900°C. In this study, complete elimination of anatase phase to rutile phase occurred at 700°C. It was also noted that the growth rate of particles differed when at low temperature and at high temperature. Slow growth rate was observed at lower temperature for which the rate increased at higher temperature. This is attributed by high activation energy at low temperature reducing its growth rate and low initiation energy at higher temperature increasing its growth rate [32].

As attributed, specific surface area (SSA) contributes to low anatase –rutile transformation. SSA was calculated by considering the total area covered by the crystals in unit mass [33] using Eq 6

(6)

Where is the density of TiO₂ NPs (4.23g/cm³) and D_p denoted the crystal size obtained by the Debye – scherrer equation [29]. The crystallographic defects or irregularities formed during crystal formation give rise to dislocations. Therefore, dislocation density is the length of the dislocation lines per unit volume of the materials crystal calculated using [34] Eq 7

(7)

Crystallite per unit surface area (N) was determined using [35]

(8)

Where D is the size of the NPs while d is the inter-planar spacing between the atoms.

Figure 5 shows that for both (a) anatase peak (105) and (b) rutile peak (101) SSA decreased as the annealing temperature was increased. SSA for anatase (105) shifted from 6.726×10¹⁴ M².g^-1 to 4.612× 10¹⁴ M².g^-1 and for rutile peak (101) from 2.6135× 10¹⁴ M².g^-1 to 2.2695× 10¹⁴ M².g^-1. This shows that crystallite size is inversely proportional to SSA for both the anatase and rutile peaks. TiO₂ generally has a low SSA [44].

Figure 5. graph showing effects of annealing temperature on SSA and crystal size for (a) rutile peak (101) and (b) anatase peak (105)

Higher surface area provides more active sites and surfaces for the reactants playing a vital role in increasing catalytic reactions, surface reactivity, photocatalytic efficiency, adsorption capacity, electrochemical performance and creation of surface defects and states [45]. Consequently, extremely high SSA leads to increased recombination rate of charge carriers prompting optimal SSA as akey factor in applications such as photocatalysis, photo-electrochemical cells among other applications. In line with TiO₂ applications and specifically as a photoanode, increase in SSA increases recombination rate. Minimizing surface related recombination rate helps improve the overall efficiency of the photoanode provided factors like band structure, charge transport and stability are optimized. Increase in annealing temperature is found to help reduce recombination rate and improving the optical properties of TiO₂.

The values of strain were estimated using Eq 9

(9)

Where θ is the diffraction angle and β is the FWHM.

Figure 6. A graph showing the effects of annealing temperature on crystal size and strain of TiO2 NPs

An inverse variation of crystallite size with strain is observed as a function of annealing temperature. As the annealing temperature increases the strain of the NPs decreases and appears to be constant as the crystal size becomes relatively constant.

3.3. Morphological Analysis

Figure 7(a) to (d) shows SEM images of TiO₂ NPs annealed at different temperature of 400°C, 500°C, 600°C and 800°C providing insights into their morphological changes. According to figure 7(a) TiO₂ NPs annealed at 400°C shows presence of primary particles of grains that are very small in size (nanometer) and as the annealing temperature increases agglomeration and aggregation is inevitable being more pronounced at 800°C where it shows an increased grain size.

In comparison to figure 7 (a), (b), (c) and (d) sintering and coarsening took place as a result of thermal treatment offering a glimpse into the dynamic rearrangement of NPs; increase in the annealing temperature made the adjacent particles coalesce and the large particles to grow at the expense of the smaller particles [31]. Figure 7(a) shows some nucleation on primary particles reducing its texture which are absent at figure 7(d) which is as a result of sintering and coalescing where the large crystalline structures tends to fill in the interstitial spaces, surface porosity decreased as the void spaces between the particles became more densely packed forming a more compact structure. This shows that, as the annealing temperature increases and precisely at around 600°C, the nucleation that was observed at 400°C sintered and coalesced leading to a smoother surface with no deposits on the surface of the crystals. Uniform grain distribution with a higher clarity and connections is observed as the annealing temperature increases as shown in figure 7(d).

Figure 7. Representative SEM images of NPs annealed at (a) 400°C, (b) 500°C, (c) 600°C and (d) 800°C

The agglomeration formed resulted from overlapping of the smaller and medium particles. This overlap were caused by lack of water on the surface structure, soft agglomerates from the weak surface and hard agglomerates from strong chemical bond (10).

3.4. Optical Analyses

The reflectance spectra for NPs annealed at different temperature are depicted in figure 8(a) between the ranges of 200 nm to 1100 nm. TiO₂ NPs annealed at 400°C had the highest reflectance of 85.26% while that annealed at 700°C had the least reflectance of 23.38%. Band gap energy (Eg) was estimated using the Kubelka- Munk formula Eq 10 [37] where (F(R)) is the Kubelka-Munk function and ∞ denotes sufficient thickness of the sample layer that is thick enough to completely hide the support for which the desired material is placed and deposited

(10)

Where (k) is the absorption coefficient (m^-1), (s) is the scattering coefficient (m^-1).

The plot of (F(R)hv)^1/2 with photon energy (hv) as shown in figure 8(b) was used to get the estimated band gap; From Tauc equation Eq 11 [38]

(11)

Where is the linear absorption coefficient, hv is the photon energy, Eg the separation gap between conduction band and valence band and is a constant which depends on the probability of transition; it takes values such as 2 and 3 for indirect allowed, indirect forbidden, direct allowed and direct forbidden transitions respectively [39].

Replacing with Kubelka function F(R∞) (Eq 12)

(12)

For purposes of optical analysis of this study of indirect transition was used as it gave good aquasi. From Eq 12 which corresponds to a linear equation y = mx+c where when extrapolating of the aquasi line to zero absorption coefficient the x intercept gives an estimate value of the band gap(Eg) value. Increased crystal size gives smaller band gap therefore as the annealing temperature increases the crystal size increased resulting in a decreased band gap [40]. The estimated band gap ranges from 3.26 to 3.50 eV; rutile phase has a lower optical band gap as compared to anatase in line with other works [41].

Figure 8. Graph showing (a) reflectance spectra and (b) band gap plot for TiO2 prepared at different annealing temperature

4. Conclusions

In summary, the structural, morphological and optical properties of TiO₂ NPs synthesized by sol- gel method were greatly influenced by the increasing annealing temperature. FTIR showed in the finger print region four stretching modes at 1033 and 668 cm^-1 associated with Ti-O-Ti and other stretching associated with Ti-O at 435 and 416 cm^-1. Other stretching modes were noted in the functional group associated with Ti-OH associated with the bonds formed with the solvent precursor. XRD analysis displayed presence of rutile phase and anatase phase. The anatase phase was found to end at 500°C with the depletion of the peak 105 for TiO₂ NPs annealed at 600°C. Phase content and mole fraction were found to decrease in the anatase phase and increase in the rutile phase. Similarly crystallite size was found to decrease in anatase phase and increase in rutile phase. Crystallite size increased with a consequential decrease in strain as the annealing temperature increased. SEM analysis showed that the surface morphology of the TiO₂ improved as the annealing temperature increased with no agglomeration and aggregation at 800°C. Optical analysis depicted that the E_g reduced from 3.50 eV to 3.26 eV. The high band gap of 3.5 eV confirms that the anatase phase has a higher optical value compared to the 3.26eV of the rutile phase. The reduced band gap energies at raised annealing temperature confirm the increase in grains as observed in SEM images and XRD crystals sizes. The study confirms that higher annealing temperature yields larger crystal sizes that influence the nanomaterials properties.

ACKNOWLEDGEMENTS

This research has been successful through the support received from African AI Research grant Award: DSA and Deep Learning Indaba. The authors wish to thank this team and Murang’a University of Technology for giving us access to various synthesis and characterization techniques for the research.

Declaration of Conflict of Interest

The authors have no conflict of interest to declare.