Catalytic Application of Oxygen Vacancies Induced by Bi3+ Incorporation in ThO2 Samples Obtained by Solution Combustion Synthesis.

Recognizing immense advantages of solution-based combustion synthesis, its applicability to determine the extent of dissolution of Bi3+ in fluorite-structured thoria has been examined to generate high-surface-area samples with massive defects. Up to 50 mol % of thorium could be substituted with bismuth retaining fluorite structure beyond which phase separation occurred. The lattice parameters from Le-Bail refinements of their powder X-ray diffraction patterns showed marginal increase with increase in bismuth content, suggesting the competing effect between the size of the cation and the oxygen vacancy concentration. Energy-dispersive X-ray spectrometry analysis and high-resolution transmission electron microscopy measurements have also confirmed the composition and structure of the limiting composition. With progressive bismuth content, the band due to the fluorite (at 460 cm-1) diffused and a defect band in the region 570-600 cm-1 emerged in the Raman spectra. From these changes, the oxygen vacancy concentrations in these samples have been determined, which increased with increase in bismuth content. Absorbance in the visible region was noticed for bismuth-containing samples, and band gap values determined from the Kubelka-Munk function were in the range 2.34-3.24 eV. In addition to the blue emission from oxygen vacancies, 3P1 → 1S0 transition of Bi3+ was noticed in the photoluminescence spectrum. From Brunauer-Emmett-Teller measurements, the surface area of Th0.50Bi0.50O2-δ obtained by solution combustion synthesis was measured to be 265.74 m2 g-1, higher than the value (39.00 m2 g-1) for the sample prepared by solid-state synthesis. All of these factors combined with oxygen vacancies as defect centers have been found to play critical control over their use as catalyst for the reductive transformation of nitroaromatics and oxidative decolorization of organic dye molecules (methyl orange and xylenol orange). A nice correlation between oxygen vacancy concentration and pseudo first-order rate constants of these catalytic conversions has been arrived. The catalyst was found to retain its efficiency up to four cycles without undergoing any structural change during these experiments.

Introduction

With ever increasing demand of energy, research work dealing with alternate energy generation and energy storage is on the rise. The classical oxygen-ion conducting materials are the fluorite-structured oxides consisting of a simple cubic oxygen lattice with occupation of eight coordinated cations in alternate body centers. Zirconia, thoria, ceria, and hafnia systems are some typical examples exhibiting fluorite structure. Doping these oxide lattices with cations of lower valence have been found to introduce oxygen vacancies, and these vacancies have been reasoned to be responsible for high ionic conductivity promoted by their migration.[1,2] The sustainability of fluorite-structured oxides to a high degree of substitution with other cations and thus resulting in nonstoichiometry has been an unique feature. Extensive investigations have been centered on doped zirconia owing to their use as solid electrolytes in batteries, oxygen sensors, and fuel cells.[3] For applications under extreme conditions (such as oxygen pressure and/or high temperature), thoria-based solid solutions could be ideal alternates, although conductivity of the thoria-based solid solution was measured to be lower than that of stabilized zirconia. The study of structure–property correlation has been facilitated by the structural stability of thoria over a wide range of temperatures. Other than these tetravalent containing oxides, perovskite-structured oxides, apatites, BiMEVOX, and K2NiF4 structure type oxides have been at the center stage of investigation to engineer the oxide-ion conductivity effectively.[1] High oxide-ion conductivity of one of the polymorphs of Bi2O3, i.e., δ-Bi2O3, as compared to that of stabilized zirconia has attracted the attention of researchers working in the area of solid oxide fuel cells.[4,5] The δ-Bi2O3 has a face-centered cubic defect fluorite structure containing statistically disordered two vacant oxide-ion sites per unit cell. Its stability at a higher temperature range of 1003–1097 K has been brought down to room temperature by the dissolution of trivalent lanthanides.[6−16] Enormous investigations have focused on correlating disordered oxygen vacancies influencing oxide-ion conductivity in rare earth stabilized δ-Bi2O3 systems.[17] Realizing similarity in the fluorite structure between thoria and δ-Bi2O3, a solid solution between the two could address various deficiencies found in the use of thoria-based systems under extreme conditions. Such a solid solution would also pave the way for the generation of oxygen vacancies because of the mismatch between valence of thorium and bismuth. Oxygen vacancies generated in an isostructural ceria (CeO2) have been found to enhance the oxygen mobility, reduction/oxidation capability, and thermal stability. Excellent reviews from time to time have described the advances made using these oxygen vacancies for various catalytic applications, including exhaust after retreatment, water–gas shift reaction, fuel cells, CO oxidation, production, and purification of hydrogen.[18−26] Although extensive utility of oxygen vacancies present in ceria-based catalysts has been exploited for oxidizing reactions, there exists plenty of opportunity to be explored for reactions involving reduction.[26,27]

During the course of this investigation, a paper by Kanrar et al.[28] has appeared, which elaborated oxygen vacancies, created as a result of bismuth substitution, influencing the magnetic ordering in these systems. The samples in this investigation have been prepared by the solid-state reaction oxides of thorium and bismuth. The wide window of thermal stability of solid thoria (3300 °C) and Bi2O3 (817 °C) coupled with mismatch in their reactivity could be the reason behind longer duration of reaction between the metal oxides at a moderately high temperature of 900 °C. Such a heating schedule would lead to nonstoichiometry in the metal-ion composition. Taking cognizance of all of these gaps and to exploit the disordered oxygen vacancies for heterogeneous catalysis, the extent of solid solution formation between thoria and δ-Bi2O3 has been investigated by solution-based combustion synthesis in the present study. It is worthy to note that solution combustion method has been applied to synthesize thoria and Bi2O3 independently in nanostructures.[29,30] In addition to determining the extent of dissolution of bismuth in thoria lattice, change in the optical properties arising from bismuth substitution has been examined. The oxygen vacancies as a consequence of bismuth substitution have been favorably utilized for imparting catalytic function to this system for the reductive transformation of nitroaromatics and for the oxidative decolorization of complex organic dye molecules, methyl orange (MO), and xylenol orange (XO).

Results and Discussion

Determination of Solubility Limit of Bi3+ in Thoria

The pedagogical way of steps involved in the synthesis has been outlined as a flowchart in Figure . The exothermicity of combustion reaction has been majorly controlled by the ratio of oxidant to fuel (O/R). Citric acid (C6H8O7) being a polyhydroxy acid could act both as a good complexing agent as well as a fuel in combustion reaction. In terms of propellant chemistry, the ratio of the oxidizing valency of the metal nitrates (O) to the reducing valency of the fuel (R), i.e., O/R, should be unity to get maximum exothermicity.[31] The oxidizing and reducing valences for the present set of reactions were calculated by considering 3+ and 4+ oxidation states for Bi and Th, respectively, corresponding to the starting materials employed and from the oxidation state determination by Kanrar et al.[28] The oxidizing valences were −15 and −20 for Bi(NO3)3 and Th(NO3)4, respectively, and the reducing valency of citric acid was 18+. According to Pederson’s reaction model, the following combustion reactions could be conceivedThe calculated C/N ratio was around 0.277 for the reactions corresponding to eqs and 2. In a similar fashion, the C/N ratio was calculated when the amounts of these two reactants were varied. The products from these reactions after calcination at 800 °C were examined by powder X-ray diffraction (PXRD). The results from these measurements have been reproduced in Figures and S1 (Supporting Information). From these patterns, it was evident that a monophasic fluorite structure could be possible up to the inclusion of 50 mol % of bismuth for thorium, beyond which phase separation occurred. A similar conclusion has been arrived at by Hund[32] in which solid-state reactions between thorium nitrate and bismuth oxide were employed. The Le-Bail refinement of PXRD patterns was carried out in Fm3̅m space group considering a defect fluorite structure (Figure S2, Supporting Information). The lattice parameter of these compositions revealed a marginal increase with progressive bismuth content in the samples (Figure i). This was matching with the trend reported by Hund[32] and differed from the report of Kanrar et al.[28] in which the cubic lattice constant showed a marginal decrease with increasing bismuth concentration. These variations could be the result of differing oxygen vacancy concentrations in the samples prepared by either different starting materials or a different route. On the basis of ionic radius considerations, Bi3+ (1.17 Å) replacing Th4+ (1.05 Å) in 8-fold coordination would be expected to result in a linear variation of a unit cell parameter. However, the oxygen vacancies arising out of mismatch in their oxidation states could provide coordination environment less than eight (probably six) around bismuth. These factors could result in non-observance of linearity in the unit cell constant variation with increasing bismuth content in the samples. The broad band at 540 cm–1 for thoria (T1u active mode) observed in the Fourier transform infrared spectra (FT-IR) became even more broader with the inclusion of bismuth in it (Figure S3 Supporting Information).[23,24] Additionally, this band shifted to lower wave numbers with increased bismuth content, suggesting a change in force constant by bismuth substitution. Raman spectroscopy has been found to be a very useful tool to study the finer local coordination environments arising from oxygen vacancy arrangements in fluorite-structured oxides.[33,34] In addition to the vibration mode characteristic for the fluorite-structured thoria at 460 cm–1, an additional band in the vicinity of 580–600 cm–1 was observed for bismuth-substituted samples that was attributed to the defects present in the lattice (Figure ii). Also, Raman active mode at 460 cm–1 (T2g) reduced in its intensity, indicating the randomness created at that crystallographic site. On comparing the area under the peak at 460 cm–1 and the band due to oxygen vacancies at around 580 cm–1, oxygen vacancy concentration in these samples could be estimated.[35] The vacancy concentration increased steadily from 0.37 (for 10 mol % bismuth substitution) to 0.88, 2.21, 2.31, and 2.69 for 20, 30, 40, and 50 mol % bismuth-substituted samples, respectively. This variation clearly suggested that bismuth existed in 3+ oxidation state, conforming well to the bond valence sum (BVS) results (3.2) obtained from the refinement of PXRD pattern of Th0.50Bi0.50O2−δ sample with a Bi–O bond distance of 2.466(8) Å. A similar conclusion has also been arrived at by Kanrar et al.[28] in which the samples have been made following solid-state reaction between constituent oxides. Two emission bands at around 418 and 437 nm were noticed in photoluminescence spectra of these samples when excited with λex = 380 nm. The band at 418 nm could be ascribed to the emission from oxygen vacancies, whereas the other emission at 437 nm was correlated to 3P1 → 1S0 transition of Bi3+ (Figure ).[36,37] The limiting member of this solid solution, viz., Th0.50Bi0.50O2−δ was further characterized by microscopy techniques. In scanning electron microscopy (SEM) image, flaky morphology of the crystallites was observed, indicating a huge amount of gas evolution during its formation (Figure a). Energy-dispersive X-ray spectrometry (EDX) spectra at various locations of this sample confirmed the uniform presence of thorium and bismuth in the sample, and from quantification, nearly equal concentrations of thorium and bismuth were evident (inset of Figure a). The unindexed and indexed selected area electron diffraction (SAED) pattern of this sample have been presented in Figure b,c, respectively, wherein diffused spots in the form of rings could be located. Indexation of this pattern yielded d-values corresponding to (111), (200), (220), (311), and (222) hkl planes of cubic fluorite structure. Similarly, lattice fringes with distances of 0.1766, 0.2076, and 0.2691 nm corresponding to (311), (220), and (200) hkl planes of fluorite structure, respectively, were observed in its high-resolution transmission electron microscopy (HR-TEM) image (Figure d).

Figure 1

Flowchart showing the methodology adopted for the synthesis of Th1–BiO2−δ (x = 0.0, 0.10, 0.20, 0.30, 0.40, 0.50, and 0.60) samples.

Figure 2

PXRD patterns of samples in the series, Th1–BiO2−δ (a) x = 0.00, (b) x = 0.10, (c) x = 0.20, (d) x = 0.30, (e) x = 0.40, and (f) x = 0.50. Digital images of the samples are shown in the inset.

Figure 3

(i) Variation of lattice constants of samples in the series Th1–BiO2−δ with bismuth content. (ii) Room-temperature Raman spectra of samples in the series, Th1–BiO2−δ (a) x = 0.00, (b) x = 0.10, (c) x = 0.20, (d) x = 0.30, (e) x = 0.40, and (f) x = 0.50.

Figure 4

Photoluminescence emission spectra of samples in the series Th1–BiO2−δ (a) x = 0.10, (b) x = 0.20, (c) x = 0.30, (d) x = 0.40, and (e) x = 0.50 with λex = 380 nm.

Figure 5

(a) SEM image along with its EDX spectrum and analysis, (b) unindexed, (c) indexed selected area electron diffraction (SAED) pattern, and (d) HR-TEM image of Th0.50Bi0.50O2−δ.

(i) Variation of lattice constants of samples in the series Th1–BiO2−δ with n class="Chemical">bismuth content. (ii) Room-temperature Raman spectra of samples in the series, Th1–BiO2−δ (a) x = 0.00, (b) x = 0.10, (c) x = 0.20, (d) x = 0.30, (e) x = 0.40, and (f) x = 0.50.

(a) SEM image along with its EDX spectrum and analysis, (b) unindexed, (c) indexed selected area electron diffraction (SAED) pattern, and (d) HR-TEM image of n class="Chemical">Th0.50Bi0.50O2−δ.

The optical absorption characteristics of the samples were evaluated by UV–visible diffuse reflectance spectra. The reflectance data has been converted into absorbance using the inbuilt software with the instrument and presented in Figure . The samples showed absorption from UV to visible region, with the absorption edge shifting progressively toward the visible range. The direct band gap energy values of bismuth-substituted samples were estimated using the Kubelka–Munk function. From analysis, it was abundantly clear that the band gap value reduced from 5.36 eV (for pure ThO2) to a range of 3.23 eV for bismuth-containing samples.[38] Within bismuth-containing samples, band gap was found to reduce with increase in bismuth content. The reduction in band gap reflected the creation of intermediate levels between the valence and conduction bands of thoria. It is worth noting that a band gap value of 2.39 eV is reported for δ-Bi2O3.[39]

Figure 6

UV–visible diffuse reflectance spectra of samples in the series Th1–BiO2−δ (a) x = 0.10, (b) x = 0.20, (c) x = 0.30, (d) x = 0.40, and (e) x = 0.50. Inset shows Tauc plots.

UV–visible diffuse reflectance spectra of samples in the series Th1–BiO2−δ (a) x = 0.10, (b) x = 0.20, (c) x = 0.30, (d) x = 0.40, and (e) x = 0.50. Inset shows Tn class="Chemical">auc plots.

Catalytic Evaluation of Th1–BiO2−δ [x = 0.50 and 0.40, Prepared by Solution Combustion Method and x = 0.50, by Solid-State Synthesis]

As combustion synthesis has been known always to yield products with high surface area due to the evolution of voluminous amounts of gases during the pyrolysis, Brunauer–Emmett–Teller (BET) surface area of Th0.50Bi0.50O2−δ was measured. The adsorption and desorption isotherms of nitrogen gas on this sample have been presented in Figure a. The sample was found to possess surface area and pore diameter of 265.74 m2 g–1 and 15.76 nm, respectively.[40] For comparison, Th0.50Bi0.50O2−δ was also synthesized by the solid-state reaction between nitrate salts of thorium and bismuth at 900 °C. The presence of fluorite structure for this sample was confirmed from its PXRD pattern, and it showed a surface area of 39.00 m2 g–1 (Figure b). As mentioned earlier, oxygen vacancies in fluorite-structured ceria (either autodoped or intentionally created by doping with aliovalent cations) have been greatly utilized for catalytic purpose,[26] higher oxygen vacancy concentration present in bismuth substituted thoria samples has been examined for its catalytic role. Nitroaromatics have been formed as by-products from pharmaceutical and dye industries, and therefore their catalytic reduction would be expected to yield value-added chemicals. p-nitrophenol (p-NP), p-nitroaniline (p-NA), and 2,4-dinitrophenol (DNP) have been chosen to be model substrates to be reduced in the presence of Th0.50Bi0.50O2−δ as catalyst.[41] To prove the correlation between oxygen vacancy concentration and reduction ability of the catalyst, Th0.60Bi0.40O2−δ has also been examined under similar experimental conditions for the reduction of these three substrates. Finally, to highlight the critical role of higher surface area of samples produced by solution combustion synthesis, Th0.50Bi0.50O2−δ sample obtained by the solid-state reaction has been used toward the reduction of these three substrates.

Figure 7

(a, b) Show N2 adsorption–desorption isotherms of Th0.50Bi0.50O2−δ sample from solution combustion and solid-state reaction methods, respectively, using BET method.

The results from catalytic reduction of p-NP to p-aminophenol (p-AP) in the presence of NaBH4 using Th0.50Bi0.50O2−δ (prepared from solution combustion synthesis and solid-state reaction) and Th0.60Bi0.40O2−δ (prepared from solution combustion synthesis) have been presented in Figures a,b and S4b respectively. Without the addition of catalyst, the absorption band of p-NP underwent a redshift to 400 nm due to the formation of p-nitrophenoxide ions (alkaline conditions by the introduction of NaBH4). With the addition of these samples, dark yellow color of p-NP vanished and a new peak appeared at 300 nm, indicating the conversion of p-NP to p-AP. Although this conversion took place almost instantaneously (within 1–2 min) employing Th0.50Bi0.50O2−δ sample prepared by solution combustion synthesis, even after 30 min of reaction, this conversion was incomplete when Th0.50Bi0.50O2−δ sample from solid-state reaction was used as the catalyst. This illustrated the advantage of higher surface area of this composition for the catalytic application when prepared by solution combustion route. When Th0.60Bi0.40O2−δ (from solution combustion synthesis) was employed as catalyst, the conversion from p-NP to p-AP was completed in 10 min. This indicated in unequivocal terms the critical role played by oxygen vacancies in these samples in enhancing the rates. As the oxygen vacancy concentration of Th0.60Bi0.40O2−δ was less than that of Th0.50Bi0.50O2−δ sample, reduction rate was found to be slower. Although the reduction of p-NP to p-AP using aqueous NaBH4 has been found to be thermodynamically favorable (E0 for p-NP/p-AP = −0.76 V and H3BO3/BH4– = −1.33 V versus normal hydrogen electrode), the presence of kinetic barrier due to a large potential difference between donor and acceptor molecules was reasoned out to decrease the feasibility of this reaction. The samples in this study exhibiting a catalytic function might be facilitating electron relay from the donor BH4– to acceptor p-NP to overcome the kinetic barrier. The following steps could be involved in the mechanism of reduction of p-NP. Chemisorption of BH4– on the surface of catalyst followed by electron transfer from BH4– to the catalyst might be the first step. In the second step, Bi3+ might be accepting these electrons to form Bi(0). Adsorption of p-NP on the surface of the catalyst and further the donation of electrons from Bi(0), themselves getting reoxidized to Bi3+, might be providing p-NP molecules the required electrons to form p-AP, followed by their desorption in the final step.[42] The formation of p-AP was further confirmed from NMR and high-resolution mass spectrometry (HRMS) techniques. The peak value with proper chemical shifts for NMR (both 1H NMR and 13C NMR) and mass analysis has been listed below (Figure S5 Supporting Information): 1H NMR (400 MHz, DMSO) δ 7.14 (d, J = 8.72 Hz, 2H), 6.80 (d, J = 8.72 Hz, 2H), 3.50 (s, 2H); 13C NMR (100 MHz, CDCl3) δ: 155.18, 148.79, and 115.74. HRMS (ESI) [M]+ calculated and found for [C6H7NO] were 110.0528 and 110.0600, respectively. Along similar lines, reduction of p-NA was found to be efficiently catalyzed by Th0.50Bi0.50O2−δ sample prepared from solution combustion synthesis, wherein the reduction was found to be complete within 5 min, as indicated by decreased absorbance intensity due to p-NA at 381 nm along with the simultaneous appearance of peaks at 303 and 240 nm (Figure c). This reaction was found to be completed in 8 min using Th0.60Bi0.40O2−δ sample (prepared by the solution combustion synthesis) (Figure d). Using Th0.50Bi0.50O2−δ sample from solid-state reaction, this conversion was found to be complete in 20 min. Both Th0.50Bi0.50O2−δ and Th0.60Bi0.40O2−δ catalyzed selective reduction of 2,4-DNP to 2-amino-4-nitrophenol, in 5 min, as indicated by the appearance of a band at 295 nm in the UV–visible spectra. Even after 20 min of reaction, this conversion was not complete employing Th0.50Bi0.50O2−δ sample obtained from solid-state reaction.

Figure 8

(a, c, e) Show UV–visible spectra of (a) p-NP, (c) p-NA, and (e) 2,4-DNP in the presence of NaBH4 and Th0.50Bi0.50O2−δ sample from solution combustion synthesis; (b, d, f) show UV–visible spectra of p-NP, (d) p-NA, and (f) 2,4-DNP in the presence of NaBH4 and Th0.60Bi0.40O2−δ sample from solution combustion synthesis.

(a, c, e) Show UV–visible spectra of (a) p-NP, (c) n class="Chemical">p-NA, and (e) 2,4-DNP in the presence of NaBH4 and Th0.50Bi0.50O2−δ sample from solution combustion synthesis; (b, d, f) show UV–visible spectra of p-NP, (d) p-NA, and (f) 2,4-DNP in the presence of NaBH4 and Th0.60Bi0.40O2−δ sample from solution combustion synthesis.

The applicability of these three samples for the oxidative decolorization of dye molecules, MO and XO has been examined. As the absorbance maxima of MO and XO occurred at 500 and 430 nm, respectively, progress of the reaction was monitored by measuring the absorbance maxima with time.[43−45] From the results shown in Figures and 10, it was found that nearly 89 and 88% of MO and XO were decolorized in 30 and 20 min, respectively, employing Th0.50Bi0.50O2−δ sample prepared from solution combustion synthesis. The pseudo first-order rate constants extracted for these heterogeneous reactions were 0.0701 min–1 (for MO) and 0.0642 min–1 (for XO). When Th0.60Bi0.40O2−δ sample prepared from solution combustion reaction was used as the catalyst, it took up to 70 min (for MO) and 60 min (for XO dye) for complete decolorization with estimated rate constants of 0.0266 and 0.0362 min–1, respectively. These variations indicated a direct link between oxygen vacancy concentration and the catalytic action of Th0.50Bi0.50O2−δ and Th0.60Bi0.40O2−δ samples. No effect toward decolorization of these dye molecules was observable when Th0.50Bi0.50O2−δ sample (prepared from solid state reaction) was used as catalyst. Hydroxide free radicals from hydrogen peroxide could not be produced in the presence of MO alone, and therefore addition of catalyst containing Bi3+ might be facilitating the decomposition of H2O2 to produce free radicals, which in turn might be rupturing the bonds between azo groups of MO, eventually leading to decolorization. Electroactive nitrogen, oxygen, sulfur atoms, aromatic rings, and electron-rich para quinanoid aromatic rings constituted XO molecule. Hydroxyl radical emerged from hydrogen peroxide due to its interaction with catalyst could be attacking the benzene ring containing SO3– group, further leading to shifting of π bonds of aromatic rings and decolorization of XO.[46] In Figure a,b, results from recyclability experiments for the reduction of p-NP and for the oxidative decolorization of MO employing Th0.50Bi0.50O2−δ (from solution combustion synthesis) have been presented. The catalyst was found to retain its efficiency up to four cycles of use in both cases. The PXRD pattern of recycled catalyst showed no change in its structure, during these experiments, hinting at its robustness (Figure c).

Figure 9

(a, c) Show plot of absorbance of MO dye versus wavelength in the presence of H2O2 and Th0.50Bi0.50O2−δ obtained by the solution combustion synthesis; (b, d) show the plot of absorbance of MO dye versus wavelength in the presence of H2O2 and Th0.60Bi0.40O2−δ obtained by the solution combustion synthesis.

Figure 10

(a, d) Show the plots of percentage of MO and XO decolorized with increasing durations in the presence of H2O2 and Th0.50Bi0.50O2−δ sample obtained by solution combustion synthesis. Plots of ln(C0/C) and C/C0 versus time for the decolorization experiments of MO and XO in the presence of H2O2 and Th0.50Bi0.50O2−δ sample are presented in (b, e) and (c, f), respectively.

Figure 11

Results from recyclability experiments employing Th0.50Bi0.50O2−δ prepared by solution combustion synthesis as catalyst for the reduction of 1 × 10–4 M solution of (a) p-NP and (b) oxidative decolorization of MO. (c) PXRD pattern of catalyst before and after experiments (reduction and oxidation).

(a, c) Show plot of absorbance of MO dye versus wavelength in the presence of H2O2 and n class="Chemical">Th0.50Bi0.50O2−δ obtained by the solution combustion synthesis; (b, d) show the plot of absorbance of MO dye versus wavelength in the presence of H2O2 and Th0.60Bi0.40O2−δ obtained by the solution combustion synthesis.

Conclusions

Following a solution-based combustion synthetic approach, limit of dissolution of bismuth in thoria retaining fluorite structure was determined. Up to 50 mol % of bismuth could be substituted for Th4+ in thoria, which brought about dramatic changes in optical and catalytic properties in this system. The band gap of 5.4 eV of thoria reduced to values in the semiconducting range of 2.34–3.42 eV after substituting with bismuth. This could be due to the introduction of intermediate levels within the valence and conduction band. Oxygen vacancies as defects were created by the introduction of bismuth, which confirmed its 3+ oxidation state. The existence of Bi3+ was supported additionally from BVS calculations and emission band in PL spectra. The oxygen vacancy concentration was found to increase with increase in bismuth content in the samples. The oxygen vacancies combined with high surface area have been demonstrated to append catalytic function to these samples for the reduction of nitroaromatics and for the oxidative decolorization of complex dye molecules. A nice correlation of reaction rates of catalysis and the oxygen vacancy concentration in these samples has been reached. The catalytic efficiency of Th0.50Bi0.50O2−δ was retained up to four cycles following pseudo first-order kinetics and without undergoing any structural change. These results are believed to enhance their further applicability in gas-phase detection and catalytic conversion of harmful gases.

Experimental Section

Synthesis

(BiO)2CO3 (assay of 82–85% based on metal, Central Drug House), Th(NO3)4·5H2O (99%) (Speck pure), citric acid (C6H8O7) (99% Bombay Drug House), and HNO3 (68% Merck) were used as the starting materials. Although Th(NO3)4·5H2O was soluble in double distilled water under constant stirring and heating, bismuth oxycarbonate was converted to bismuth nitrate in situ by reacting with minimum amount of 1:1 HNO3 solution. The amount of citric acid for varying amounts of thorium and bismuth nitrates was calculated using the ratio of propellant chemistry, (citrate (fuel) to nitrate (oxidizer) ratio equal to 0.3). After the addition of citric acid to the mixture of metal nitrates, the reaction mixture was stirred and heated at ∼200 °C. The mixture transformed into viscous foam, followed by auto ignition, leaving behind a black carbonaceous foamy product. This ash product was collected and calcined at 800 °C for 3 h in a muffle furnace, followed by switching off to cool naturally to room temperature. The details of oxidizer (starting materials) and fuel (citric acid) taken to generate other members of the series Th1–BiO2−δ series have been compiled in Table . Th0.50Bi0.50O2−δ sample was also synthesized by the solid-state reaction between 0.2425 g (0.50 mmol) of Bi(NO3)3·5H2O [Thomas Baker 99%] and 0.2850 g (0.50 mmol) of Th(NO3)4·5H2O. The homogenized mixture of these reactants, after grinding, was heated in a platinum crucible at 900 °C for 12 h, followed by switching off the furnace. This step was repeated one more time.

Table 1

Experimental Synthesis Details of the Starting Materials for the Synthesis of Th1–BiO2−δ

composition	Th(NO3)4·5H2O	(BiO)2CO3	citric acid
x = 0.00	0.5701 g (1.00 mmol)		0.2305 g (1.20 mmol)
x = 0.10	0.5131 g (0.90 mmol)	0.0255 g (0.05 mmol)	0.2247 g (1.17 mmol)
x = 0.20	0.4561 g (0.80 mmol)	0.0510 g (0.10 mmol)	0.2190 g (1.14 mmol)
x = 0.30	0.3990 g (0.70 mmol)	0.0765 g (0.15 mmol)	0.2132 g (1.11 mmol)
x = 0.40	0.3420 g (0.60 mmol)	0.1020 g (0.20 mmol)	0.2074 g (1.08 mmol)
x = 0.50	0.2850 g (0.50 mmol)	0.1275 g (0.25 mmol)	0.2017 g (1.05 mmol)
x = 0.60	0.2280 g (0.40 mmol)	0.1530 g (0.30 mmol)	0.1959 g (1.02 mmol)

Characterization

Powder X-ray diffraction (PXRD) patterns were recorded using a high-resolution PANanalytical X’pert diffractometer, equipped with a Xe detector employing Cu Kα radiation (λ = 1.5418 Å) with a scan rate of 4.5 s per step and a step size of 0.04° over the 2θ range of 20–70° at 25 °C. The structure refinement of the PXRD patterns was carried out by the Le-Bail method using TOPAS3 software.[47] Fourier transformed infrared red (FT-IR) spectra were recorded using a PerkinElmer 2000 spectrometer using KBr disks. Raman spectra were collected using a Renishaw spectrometer via a microscope system operating with an Ar+ laser (λ = 514 nm). The morphological studies (SEM) and qualitative elemental analysis of the samples were performed by scanning electron microscopy employing a JEOL 6610LV microscope. High-resolution transmission electron microscopic (HR-TEM) images and selected area electron diffraction (SAED) patterns were obtained using an FEI Technai G2 20 electron microscope operating at 200 kV. Surface area measurements were carried out using the adsorption–desorption method with an automated surface area and pore size analyzer (Autosorb IQ Quantachrome instrument). UV–visible diffuse reflectance spectra of the samples were recorded using a PerkinElmer Lambda 35 scanning double beam spectrometer equipped with a 50 mm integrating sphere. BaSO4 was employed as the reference. Photoluminescence spectral measurements were carried out on solid samples using the Horiba Jobin Yvon Fluorolog 3 spectrofluorometer at room temperature.

Catalytic Activity Evaluation

The oxidative decolorization of XO and MO dye molecules and reduction of p-NP, p-NA and 2,4-DNP were chosen to be model systems to examine the catalytic role of bismuth-containing samples. For the oxidative decolorization experiments, 30 mg of the sample was added to 20 mL of the dye solution (1 × 10–4 M) and 10 mL of H2O2 (30 v/v % Merck). For catalytic reduction experiments, 30 mg of the sample along with 5 mg of NaBH4 was added to 50 mL of aqueous solutions of p-NP, p-NA, and 2,4-DNP, with an initial concentration of 1 × 10–4 M. The progress of catalytic experiments was followed by recording the absorbance of the solutions at periodic time intervals using a UV–visible spectrophotometer (Shimadzu 1800). The products from reduction experiments were characterized by 1H NMR (400 MHz) and 13C NMR (100 MHz) spectra recorded on a Bruker Avance-400 spectrometer using tetramethylsilane as the internal standard (chemical shifts in ppm) and using DMSO-d6 as solvent. ESI-MS spectra were recorded on Agilent Technologies 6530 Accurate-Mass Q-TOF LC/MS.