Effect of Calcination Temperature on Lattice Contraction in Hydrothermally Grown Monoclinic BiVO₄ for Photocatalytic Dye Degradation

2.1 Materials

Bismuth nitrate pentahydrate (Bi(NO₃)₃·5H₂O, ≥98%, Sigma-Aldrich) and ammonium metavanadate (NH₄VO₃, ≥99%, Sigma-Aldrich) were used as the bismuth and vanadium sources. Methylene blue (MB, C.I. 52015, Sigma-Aldrich) and methyl orange (MO, C.I. 13025, Sigma-Aldrich) were employed as model pollutants of opposite ionic character. All reagents were used as received without further purification, and deionized water was used throughout for solution preparation and washing.

2.2 Hydrothermal Synthesis of BiVO₄ Nanoparticles

BiVO₄ nanoparticles were synthesized by a hydrothermal route following a previously reported protocol [8]. Bismuth nitrate pentahydrate (2.9 g) was dissolved in 50 mL of deionized water under continuous magnetic stirring until a clear solution was obtained. Ammonium metavanadate (0.70 g), corresponding to a Bi:V molar ratio of 1:1, was separately dissolved in 30 mL of deionized water under gentle heating at 60 °C. The vanadium solution was added dropwise to the bismuth solution under continuous stirring, which yielded a yellow suspension that was stirred for a further 30 min at room temperature. The suspension was transferred to a 100 mL Teflon-lined stainless-steel autoclave and held at 150 °C for 24 h. After natural cooling to room temperature, the solid was recovered by vacuum filtration, rinsed thoroughly with deionized water, and dried at 80 °C for 2 h. The as-dried powder was divided into six portions. One portion was retained without thermal treatment and is designated BVO. The remaining five portions were calcined in air for 2 h at 300, 400, 500, 600 and 700 °C at a heating rate of 5 °C min⁻¹, and are designated BVO-300, BVO-400, BVO-500, BVO-600 and BVO-700, respectively, where the numeral denotes the calcination temperature. Calcination temperature was therefore the only variable across the series.

2.3 Materials Characterizations

The crystalline phases were identified by X-ray diffraction (XRD) on a Rigaku SmartLab SE Basic diffractometer with Cu Kα radiation (λ = 1.5406 Å) operated at 40 kV and 50 mA. Patterns were recorded from 5 to 80° 2θ at a step size of 0.01° and a scan rate of 10° min⁻¹, using a 0.5° divergence slit, a 2.5° Soller slit and a Kβ filter. The six patterns were recorded consecutively in a single session with the same instrument alignment. Phase composition, lattice parameters and crystallite size were obtained by Rietveld refinement over 10 to 80° 2θ with Profex 5, which uses the BGMN refinement kernel [16]. The monoclinic clinobisvanite structure (space group I112/b) was taken from COD 9013437, and tetragonal zircon-type BiVO₄ (space group I4₁/amd) was included as a second phase, with its atomic positions fixed at reported values [17]. The instrumental profile was calculated from the diffractometer geometry, because no line-profile standard was measured on this diffractometer. The refined parameters were the lattice parameters and scale factors of both phases, a specimen-displacement term, the crystallite size of both phases, the microstrain and the preferred orientation of the monoclinic phase, the latter described by sixth-order spherical harmonics, and the background. No internal standard was measured with the diffraction patterns, so the lattice parameters are discussed as relative changes across the series, which share the same instrument, specimen preparation and refinement model. Surface morphology was examined by scanning electron microscopy (SEM) on a JEOL JSM-6510LA instrument operated at 10 kV.

Optical absorption was examined by diffuse reflectance spectroscopy (DRS) on a Shimadzu UV-2600 spectrophotometer equipped with an integrating sphere, using BaSO₄ as the reflectance standard. Spectra were recorded over 190 to 800 nm with a slit width of 5 nm and a sampling interval of 0.5 nm. The reflectance R was converted into the Kubelka-Munk function F(R) = (1 − R)²/2R. Band gaps were estimated by extrapolating the linear region of (F(R)hν)² against photon energy to the energy axis, which assumes a direct allowed transition. Monoclinic BiVO₄ has an indirect fundamental gap with a direct transition only about 0.2 eV higher [18], and the direct form was used because it is commonly applied to BiVO₄ photocatalysts. The indirect form, (F(R)hν)^1/2 extrapolated to the sub-gap baseline, was evaluated for comparison.

2.4 Single Pollutant Degradation Tests

The photocatalytic activity of the six samples was evaluated against MB and MO. In each run, 50 mg of catalyst was suspended in 100 mL of an aqueous dye solution at 5 ppm and irradiated directly, without a dark adsorption step, by four 10 W Philips TUV germicidal lamps (λ ≈ 254 nm) arranged symmetrically around the reaction vessel. Aliquots of approximately 3 mL were collected every 30 min over 240 min and filtered through a syringe membrane filter to remove the catalyst. A no-catalyst control for MB was run under the same irradiation, dye concentration, volume and sampling protocol, on a separate day from the catalyst runs. Absorption spectra were recorded on a Shimadzu UV-1280 spectrophotometer. Because the MB band shifts to shorter wavelength during irradiation, MB was quantified from the integrated absorbance between 550 and 740 nm after subtracting the mean absorbance between 760 and 800 nm, whereas MO was quantified from the absorbance at 464 nm after subtracting the mean absorbance between 700 and 800 nm. The initial dye solution, measured once for each dye on each measurement day, served as the reference, and C/C₀ is the ratio of the integrated band (MB) or of the absorbance at 464 nm (MO) to that of the reference. Apparent rate constants k were obtained from linear fits of ln(C₀/C) against irradiation time, over 0 to 240 min for MB and over 30 to 240 min for MO, whose absorbance exceeds the initial value at 30 min for several samples, and were divided by the catalyst mass to give the mass-normalised rate constant k_m.

3.1 Material and Structural Characteristics of BiVO₄ Nanoparticles

Figure 1 shows the surface morphology of the six samples. BVO consists of irregular blocky and short rod-like particles with poorly defined edges, accompanied by finer particles below about 200 nm (Figure 1a). The particles range from a few hundred nanometres to about 1 μm across in the fields examined. This blocky and rod-like habit persists in BVO-300 and BVO-400, in which individual particles remain clearly separated (Figure 1b and 1c). From BVO-500 onward the particle outlines become rounded and adjacent particles are joined by necks (Figure 1d and 1e). BVO-700 presents a largely fused surface in which individual particles are no longer resolved and little interparticle porosity is visible (Figure 1f). The progressive rounding and neck formation are consistent with sintering, which becomes appreciable between 500 and 600 °C.

Figure 2a presents the X-ray diffraction patterns of the six samples together with the reference pattern of monoclinic BiVO₄ (PDF#14-0688). Every strong reflection is indexed to the monoclinic clinobisvanite structure, and the splitting of the (200) and (020) reflections near 34.5 and 35.2° is resolved in all six samples, including the as-dried BVO. However, BVO, BVO-300 and BVO-400 also show two weak reflections at 24.4 and 32.7°, which correspond to the (200) and (112) reflections of tetragonal zircon-type BiVO₄ (Figure 2b to 2d). These reflections are absent from BVO-500 onward (Figure 2e to 2g). The monoclinic phase therefore forms mainly during the hydrothermal treatment at 150 °C, and the minor zircon-type fraction disappears between 400 and 500 °C. This is consistent with the transformation of zircon-type BiVO₄ into the scheelite structure between 420 and 500 °C, which becomes monoclinic on cooling [17]. No other reflection is observed up to 600 °C. Moreover, weak additional reflections near 26.6, 27.2 and 31.7° appear in BVO-700, which suggests the onset of a minor second phase that the present data do not identify. Because that phase is not included in the refinement, the refined values for BVO-700 carry more uncertainty than those for the other five samples.

Rietveld refinements are shown in Figure 2b to 2g, and in each case the calculated profile reproduces the observed intensities across the full angular range, with R_wp between 10.6 and 12.7% (Table 1). In BVO, BVO-300 and BVO-400 the zircon-type phase refines to less than 2 wt%. The refined unit-cell volume of the monoclinic phase decreases monotonically from 310.179 ų for BVO to 309.750 ų for BVO-700, as plotted in Figure 2i. The total change of 0.43 ų is more than 30 times the mean standard deviation returned by the refinement, although such standard deviations are known to underestimate the true parameter uncertainty. Moreover, the decrease is already monotonic over the five samples up to 600 °C, so it does not rest on BVO-700 and its unindexed second phase. The contraction amounts to 0.138% and is anisotropic, with b shrinking by 0.081% while a and c shrink by 0.035 and 0.023%, respectively. The ratio b/a is insensitive to zero-point and specimen-displacement errors, because the (200) and (020) reflections that fix it lie only 0.7° apart. It changes by at most 0.03% up to BVO-500 and then falls by about 0.06% at BVO-600, so the b-dominated contraction above 500 °C is not an artefact of an angular error. Meanwhile, the monoclinic angle γ shows no systematic trend across the series, which indicates that the degree of monoclinic distortion is unchanged by calcination. The lattice therefore densifies without a change in symmetry, which is consistent with the progressive removal of defects retained from the hydrothermal synthesis.

The crystallite size of the monoclinic phase is plotted in Figure 2h. It remains between 96 and 103 nm from BVO through BVO-500 and then rises to 158 nm for BVO-600 and 129 nm for BVO-700. This step between 500 and 600 °C indicates a threshold for thermally activated grain growth, and it coincides with the interval in which rounding and neck formation are first seen in Figure 1. Meanwhile, the refined microstrain decreases steadily above 300 °C (Table 1), which is consistent with the defect removal inferred from the lattice contraction. However, the instrumental profile was calculated rather than measured on a standard, and for BVO-600 and BVO-700 the sample broadening is comparable to the instrumental width. Their sizes therefore carry a larger uncertainty than the standard deviations suggest, and the difference between them is not interpreted. The crystallite sizes are smaller than most of the particles seen by SEM, which range from a few hundred nanometres to about 1 μm, so these particles are polycrystalline aggregates rather than single crystals.

Figure 3a shows the diffuse reflectance spectra of the six samples. Every sample reflects weakly below about 470 nm, including at the 254 nm line of the lamps used for the photocatalytic tests, and shows a steep absorption edge near 510 nm. The Tauc plots for a direct allowed transition in Figure 3b to 3g give band gaps of 2.54, 2.52, 2.53, 2.52, 2.50 and 2.49 eV for BVO through BVO-700 (Figure 3h). The band gap is therefore essentially unchanged up to 500 °C, although the cell volume contracts monotonically over the same interval, so the lattice contraction does not measurably shift the absorption onset. The indirect allowed form gives lower values with the same behaviour across the series, so this comparison does not depend on the assumed transition. The two highest temperatures give slightly lower values, and their reflectance above the edge is also lower, most clearly for BVO-700 (Figure 3a), whose Tauc plot extends to lower energy below the edge (Figure 3g). This additional sub-gap absorption is consistent with the minor second phase detected by XRD in BVO-700 and with its fused surface seen by SEM.

3.2 Photocatalytic Degradation of Single Pollutants

Figure 4a shows the removal of MB by the six samples. After 240 min, C/C₀ lies between 0.54 and 0.72, with the lowest value for BVO-500. The linear fits in Figure 4b give apparent rate constants between 1.1 and 2.1 × 10⁻³ min⁻¹, and the faster decrease during the first 30 min, which for most catalysts includes adsorption of the cationic dye, gives these fits a positive intercept. The mass-normalised rate constants in Figure 4c follow the same order, with the highest value for BVO-500. Part of this removal may arise from direct photolysis at 254 nm. In a no-catalyst control under the same conditions, which is not included in Figure 4, MB decreased by 17% after 180 min and by 22% after 240 min (k = 0.8 × 10⁻³ min⁻¹). This photolysis level lies within the range reported for MB under four 10 W lamps at 254 nm, from a slight decrease over 240 min [19] to about 33% after 180 min [20], depending on the arrangement of the lamps. Every catalyst ends below the photolysis level after 240 min, but BVO and BVO-300 remain close to it, and the clearest enhancement is obtained with BVO-500, whose rate constant is about 2.6 times that of photolysis. However, the runs were performed once, the control was recorded on a separate day, and BVO-600 and BVO-700 were measured on a different day from the other four samples, so the ranking between neighbouring samples should be read with caution.

In every catalysed run, the absorption maximum of MB shifted to shorter wavelength during irradiation, whereas it remained at its initial position in the control. A shift of this kind is a signature of successive N-demethylation, in which the chromophore is retained while methyl groups are removed [21], and its absence without catalyst suggests that this pathway is driven by the catalyst. Quantifying the residual dye at a fixed wavelength would therefore overstate its removal, and the integrated band area was used throughout.

Figure 4d shows the corresponding results for MO. With every sample, C/C₀ remains close to unity and lies between 0.95 and 1.01 after 240 min, and the mass-normalised rate constants stay below 0.007 min⁻¹ g⁻¹ (Figure 4e and 4f). BiVO₄ therefore does not degrade MO to a measurable extent under these conditions. The contrast with MB is consistent with the stronger interaction expected between the cationic dye and the catalyst surface, although the point of zero charge of these powders was not measured.

For MB, the mass-normalised rate constant is highest for BVO-500 and lower on either side of it (Figure 4c), while BVO and BVO-300 remain close to the photolysis level. This pattern follows neither the monotonic lattice contraction nor the essentially constant band gap described above, so in this series the MB activity is not controlled by the refined structural or optical parameters. The decrease at 600 and 700 °C coincides with the sintering seen in Figure 1 and with the growth of the crystallite size in Figure 2h, which is consistent with a loss of accessible surface. For MO, none of the samples gives measurable removal (Figure 4d). These conclusions rest on single runs, so moderate differences between catalysts should not be over-interpreted.