Journal of Structural Chemistry. Vol. 58, No. 1, pp. 126-134, 2017. Original Russian Text © 2017 M. A. Kerzhentsev, E. V. Matus, I. Z. Ismagilov, V. A. Ushakov, O. A. Stonkus, T. V. Larina, G. S. Kozlova, P. Bharali, Z. R. Ismagilov.
STRUCTURAL AND MORPHOLOGICAL PROPERTIES OF Ce1–xMxOy (M = Gd, La, Mg) SUPPORTS FOR THE CATALYSTS OF AUTOTHERMAL ETHANOL CONVERSION M. A. Kerzhentsev1, E. V. Matus1*, I. Z. Ismagilov1, V. A. Ushakov1, O. A. Stonkus1,2, T. V. Larina1, G. S. Kozlova3, P. Bharali4, and Z. R. Ismagilov1,5
UDC 544.478.02:544.478.01
A complex of physicochemical methods (powder XRD analysis, transmission and scanning electron microscopy, electron spectroscopy of diffuse reflectance, low-temperature nitrogen adsorption) is used for the comparative study of structural and morphological properties of oxide supports Ce1–xMxOy (M = Gd, La, Mg; x = 0-0.5; 1.5 ≤ y ≤ 2.0) for catalysts for the autothermal reforming of bioethanol to a hydrogenbearing gas. It is shown that Ce1–xMxOy samples synthesized by the method of ester polymer precursors are mesoporous materials being the homogenous substitutional solid solutions with the fluorite-type cubic structure. The structural and textural properties of the Ce1–xMxOy materials are regulated by varying the type of the dopant cation (M = Gd, La, Mg), the molar ratio M/Ce (0, 0.1, 0.25, 1), and heat treatment conditions (temperature 300-800 °C; duration 4-24 h). The relationship between the synthesis parameters and the characteristics of the Ce1–xMxOy materials is found. DOI: 10.1134/S002247661701019X Keywords: cerium dioxide, dopants, powder XRD analysis, electron microscopy, electron spectroscopy of diffuse reflectance, nanomaterials.
INTRODUCTION Mixed cerium-containing oxides (Ce1–xMxOy, where M is the dopant cation, M = Zr, Gd, La, Mg, etc.) find a wide use as catalyst supports providing a high dispersion of the applied active component particles and catalyst resistance to coke formation. The structural, morphological, and textural properties of cerium-containing oxides depend on the type of the dopant cation M, the molar ratio M/Ce, and synthesis method [1-6]. A purposeful change in Ce1–xMxOy characteristics makes it possible to regulate the dispersion and redox properties of the active metal and, consequently, to control the catalytic reaction parameters [6-8].
1
Boreskov Institute of Catalysis, Siberian Branch, Russian Academy of Sciences, Novosibirsk, Russia; *
[email protected]. 2Novosibirsk National Research State University, Russia. 3Shared Research Center, Federal Research Center of Coal and Coal Chemistry, Siberian Branch, Russian Academy of Sciences, Kemerovo, Russia. 4Tezpur University, Napaam, Tezpur Assam, India. 5Institute of Coal Chemistry and Material Science, Federal Research Center of Coal and Coal Chemistry, Siberian Branch, Russian Academy of Sciences, Kemerovo, Russia. Translated from Zhurnal Strukturnoi Khimii, Vol. 58, No. 1, pp. 133-141, January-February, 2017. Original article submitted March 29, 2016. 126
0022-4766/17/5801-0126 © 2017 by Pleiades Publishing, Ltd.
In this work, in order to determine the regularities of the formation of nanostructured oxide materials and to obtain samples with the desired physicochemical properties, we synthesized Ce1–xMxOy materials by varying the type of the dopant cation (M = Gd, La, Mg), the molar ratio M/Ce (x = 0-0.5), and heat treatment conditions (temperature 300-800 °C; duration 4-24 h). Using a complex of methods (X-ray fluorescence spectroscopy, low-temperature nitrogen adsorption, powder XRD analysis, electron spectroscopy of diffuse reflectance, transmission and scanning electron microscopy) the physicochemical properties of Ce1–xMxOy were systemically studied. The obtained Ce1–xMxOy materials are intended to be used as supports in the composition of mono- (Ni, Co) and bimetallic (Ni(Co)+Pt, Pd, Rh, or Re) catalysts for the autothermal reforming of bioethanol to a hydrogen-bearing gas.
EXPERIMENTAL The Ce1–xMxOy materials were synthesized by the method of ester polymer precursors (the Pechini method) described in [6, 9]. The metal content in the materials under study was determined by X-ray fluorescence spectroscopy on an ARL ADVANT′X analyzer with the Rh anode of the X-ray tube (ThermoTechno Scientific, Switzerland). The textural characteristics of the materials (specific surface area SBET, pore volume Vpore, and average pore diameter Dpore) were examined on an automated ASAP 2400 volumetric system (Micromeritics, USA) by measuring and processing the isotherms of low-temperature nitrogen adsorption at 77 K. The powder X-ray diffraction (XRD) analysis of the samples was carried out on a HZG-4C diffractometer (Freiberger Prazisionmechanik, Germany) with monochromatic CoKα radiation (λ = 1.79021 Å). The phase composition was determined from the diffraction patterns measured by scanning in an angle range 2θ = 10-80° with a 0.1° step and an acquisition time of 6-15 s. The coherent scattering region (CSR) size was calculated from the broadening of the 1.1.1. diffraction peak of the fixed phases with a fluorite-type cubic structure. The electron spectra of diffuse reflectance (ESDR) were recorded on a Shimadzu UV-2501 PC spectrophotometer with an ISR-240 A diffuse reflectance attachment. The powdered samples were placed in a quartz cuvette with an optical path length of 2 mm. The spectra were measured relative to the BaSO4 reflectance standard in a range 190-900 nm (11 00053 000 cm–1). The obtained reflection factors R were converted into the absorption coefficients by the Kubelka–Munk function, F(R) = (1 – R)2/2R. All ESDR data are presented in the following coordinates: the Kubelka–Munk function F(R) vs. the wavenumber. The high-resolution transmission electron microscopy (HRTEM) images were measured on JEM-2010 (JEOL Ltd., Japan) and JEM-2200FS (JEOL Ltd., Japan) microscopes operating at the accelerating voltage of 200 kV. The spatial lattice resolution of the devices is 1.4 Å and 1 Å respectively. The electron microscopes are equipped with analytic attachments for the local elemental microanalysis (EDX spectroscopy). The JEM-2200FS microscope operated in TEM and STEM (scanning) modes. The EDX method on this device is compatible with both TEM and STEM modes allowing the EDX mapping with a resolution higher than 1 nm. The scanning electron microscopy (SEM) images were obtained on a JSM-6390LA (JEOL, Japan) electron microscope.
RESULTS AND DISCUSSION Table 1 summarizes the chemical composition, textural and structural characteristics of the synthesized materials. The molar ratio M/Ce calculated from the chemical analysis data is in a satisfactory agreement with that of the initial formula of the mixed oxide. According to the N2 adsorption data, for unmodified cerium dioxide, type IV adsorption isotherm with a hysteresis loop of type H3 is observed (Fig. 1a). At the partial pressure p/po = 0.5-0.9 the hysteresis indicates the presence of both 127
TABLE 1. Chemical Composition, Textural and Structural Characteristics of the Ce1–xMxOy Materials: the Effect of the Dopant Cation Type (M = Gd, La, Mg) and Content
Sample*
Chemical composition, wt % Ce M
M/Ce
SBET, m2/g
V∑, cm3/g
Dpore, nm
Unit cell parameter a, nm 4 h** 24 h
Particle size (CSR), nm 4h 24 h 12.5 12.5 11.0 8.5
CeO2 Ce0.9Gd0.1O1.95 Ce0.8Gd0.2O1.9 Ce0.5Gd0.5O1.75
82.6 74.3 62.3 35.5
0 8.1 16.6 38.6
0 0.10 0.24 0.97
74 95 84 63
0.16 0.16 0.17 0.17
8.5 6.7 7.9 10.4
0.5414 0.5418 0.5426 0.5439
0.5414
Ce0.9La0.1O1.95 Ce0.8La0.2O1.9 Ce0.5La0.5O1.75
74.2 66.2 40.4
9.6 16.7 40.1
0.13 0.25 1.00
94 94 52
0.18 0.19 0.19
7.6 7.9 15.1
0.5446 0.5478 0.5562
0.5481
12.0 8.0 5.0
8.5
Ce0.9Mg0.1O1.9 Ce0.8Mg0.2O1.8 Ce0.5Mg0.5O1.5
80.9 77.4 66.3
1.5 3.2 9.1
0.11 0.24 0.80
69 54 38
0.19 0.19 0.20
11.2 14.0 25.0
0.5412 0.5410 0.5408
0.5409
8.0 7.0 6.5
9.0
0.5428
16.0 11.5
* The calcination temperature is 500 °C. ** The calcination duration.
Fig. 1. Nitrogen adsorption isotherms (a) and the dependence of the specific surface area SBET on the calcination temperature (b) for CeO2 and Ce0.8M0.2OY (M = Gd, La, Mg): the effect of the type of the dopant cation. primary (pores within the primary particles) and textural (pores between the primary particles) mesoporosity [10]. As evident from Fig. 1a, the initial part of the isotherm associated with the microporous region, is insignificant. With the introduction of the dopant cation, the shape of the adsorption isotherm remains unchanged, but the position of the hysteresis loop depends on the type and content of the dopant cation (Fig. 1a). A shift in the position of the hysteresis loop towards higher relative pressures on the Mg doping or with an increase in the molar fraction of the dopant cation (from 0.1 to 0.5) can mean an increase in the contribution of large interparticle pores to the porosity. Therewith, the average pore diameter increases almost twofold (Table 1). For unmodified cerium dioxide calcined at a temperature of 500 °C, the specific surface area is 74 m2/g, the pore volume is 0.16 cm3/g, and the average pore size is 8.5 nm. The obtained values of the specific surface area are in the range typical of CeO2 synthesized by the Pechini method and significantly exceed SBET (30-40 m2/g) of the samples synthesized by 128
the deposition or the citrate sol–gel method [11-13]. With the molar ratio M/Ce = 0.1-0.25, the introduction of Gd or La makes it possible to increase the specific surface area of the material by 15-30%. A further increase in the content of the dopant cation leads to a decrease in SBET and an increase in Dpore. Therewith, the porosity of the material remains at the same level. It should be noted that all Mg-containing samples feature lower SBET than CeO2 (Table 1). An increase in the calcination temperature from 300 °C to 500 °C leads to a 20-25% decrease in the specific surface area (Fig. 1b). With an increase in the calcination temperature to 800 °C, a further decrease in the specific surface area is observed due to the intensification of the sintering processes with a temperature rise [14]. From the ratio between the specific surface areas after the calcination at 300 °C and 800 °C it follows that the resistance to sintering increases in a series of modifying additives Gd < La < Mg. Therewith, the Gd introduction increases the sintering ability of the material, while La and Mg diminish it. Fig. 2 shows the typical diffraction patterns of CeO2 and Ce1–xMxOy (M = Gd, La, Mg). The analysis of the diffraction patterns shows that the obtained materials calcined at 300-800 °C are single phase crystal systems representing fluorite-like solid solutions based on cerium dioxide (JCPDS-34-394). An exception is the Ce0.5Mg0.5O1.5 sample, in which the traces of finely dispersed magnesium oxide (the diffraction peak at an angle of 50.45°) are observed (Fig. 2a). In order to study the uniformity of the dopant cation distribution in the solid solution based on cerium dioxide, we performed the EDX mapping showing the elemental distribution in a chosen part of the sample. Fig. 3a depicts the image of a part of the Ce0.8Gd0.2Oy sample, from which the EDX maps of Ce (CeL, green) and Gd (GdM, red) distributions were recorded, obtained by the high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM). As it follows from the obtained data, Gd is uniformly distributed in the sample composition: there are no regions of an increased gadolinium concentration in the images (Fig. 3b). Similar results were obtained for lanthanum-doped cerium dioxide (Fig. 3d–f). A somewhat different picture is observed for cerium dioxide doped with Mg cations (Fig. 3g–i). From the EDX mapping, we identified the regions of an increased magnesium concentration (Fig. 3h), which is indicative of some deviation from the uniform Mg distribution in the sample and is consistent with the powder XRD results. The diffraction patterns of the samples calcined at 300-500 °C are characterized by broad symmetric peaks. The estimation of the average size of crystalline domains (CSRs) by the Scherrer equation shows that for unmodified cerium dioxide the CSR size is 12.5 nm (Table 1). The unit cell parameter of cerium dioxide calculated from the position of the 311 line is 0.5414 nm, which is somewhat higher than the value reported for bulk cerium dioxide (JCPDS-34-394). According to the literature data [15-17], the unit cell parameter a of cerium dioxide increases with a decrease in the particle size. This phenomenon is caused by a partial removal of oxygen atoms with the formation of oxygen vacancies and, consequently, a decrease in the effective oxidation state of cerium CeO2 → CeO2–х + x/2O2,
where 0 < x < 0.5;
Fig 2. Effect of the composition of the Ce1–xMxOy material (M = Gd, La, Mg) (a) and the calcination temperature of the Ce0.8M0.2Oy (M = Gd, La, Mg) material (b) on its phase composition. 129
Fig. 3. HAADF-STEM images of the parts of Ce0.8M0.2Oy samples (a, d, g); the EDX mapping showing the elemental distribution (b, e, h); the EDX spectra of the parts of the samples (c, f, i). M = Gd (a–c); La (d–f); Mg (g–i). The calcination temperature is 500 °C. 4Ce4+ + 2O2– → 4Ce4+ + Vo•• + 0.5O2 → 2Ce4+ + 2Ce3+ + Vo•• + 0.5O2. According to the data from [16], the molar ratio Ce3+/(Ce3++Ce4+) increases with a decrease in the particle size, approaching 1 at a particle size of ∼ 3 nm. The calculated average value of the unit cell parameter of CeO2 and CeO1.5 is 0.541 nm and 0.561 nm respectively. Thus, with a decrease in the particle size from 70 nm to 2.8 nm, the average a value increases from 0.541 nm to 0.554 nm. As can be seen from Fig. 2, the doped cerium dioxide samples are observed to have a shift of the diffraction maxima to the smaller (for Gd and La) or larger (for Mg) angle regions, which is associated with the introduction of the dopant cation into the cerium dioxide lattice and different ionic radii of the cations. The dopant cation introduction changes the lattice parameter and the average CSR size (Table 1), being in good agreement with the previous data [18, 19]. When the samples are doped by Gd or La cations, the lattice parameter increases due to the replacement of cerium cations (0.097 nm) by cations with a larger ionic radius (0.105 nm and 0.116 nm respectively), while in the case of Mg, it decreases due to the replacement of cerium cations by cations with a smaller ionic radius (0.072 nm) [20, 21]. The unit cell volume changes proportionally to an increase in the molar content of the dopant cation (Fig 4a), which evidences a graduate increase in its content in the cerium dioxide lattice and the formation of a homogenous solid solution. The introduction of a dopant cation whose valence is less than that of the Ce cation leads to the formation of oxygen vacancies. The concentration of oxygen vacancies can be regulated by varying the type and content of the dopant cation, thus determining the oxygen capacity of the material and affecting the redox properties of cerium ions [15, 22]. On doping cerium dioxide with trivalent cations (Gd or La), one oxygen vacancy forms per each two cations replacing a tetravalent cerium cation according to the lattice electroneutrality condition. On doping with a divalent cation (Mg), the formation of one oxygen vacancy corresponds to one substituent cation.
130
Fig. 4. Results of the powder XRD analysis: the average unit cell volume of Ce1–xMxOy (M = Gd, La, Mg) depending on the molar fraction of the dopant cation x (a). The calcination temperature is 500 °C; the effect of the calcination temperature of the material on the CSR size (b). For the materials calcined at 300-500 °C, the average CSR size is ∼5.0-12.5 nm. The average size of Ce1–xMxOy crystallites decreases with an increase in the molar fraction of the dopant cation in the composition of the material and in the following metal series Gd < La < Mg (x = 0.1-0.2). A decrease in the size of primary particles on doping has been a widely known phenomenon for the CeO2 system caused by the inhibition of the crystallite growth in the presence of the dopant cation [14, 23, 24]. When the duration of the calcination of CeO2 and Ce0.8M0.2Oy (M = Gd, La, Mg) materials at 500 °C extends from 4 h to 24 h, a weak tendency for increasing the CSR size is observed (Table 1). In the case of cerium dioxide doped by Gd or La cations, the CSR size remains almost unchanged. As we have already mentioned above, an increase in the calcination temperature to 800 °C does not lead to a change in the cubic structure. The diffraction peaks become narrower due to particle coarsening (their sintering). The crystallite size increases to 30 nm in the case of La-containing samples and to 50 nm for other materials. As seen from Fig. 4b, which illustrates the effect of the calcination temperature of the material on the CSR size in more detail, when the calcination temperature is elevated from 300 °C to 500 °C the CSR size remains unchanged. A further increase in the calcination temperature leads to a regular increase in CSR. The intensity of this process depends on the type of the dopant cation. For example, in the case of cerium dioxide, CSR increases ~2 times at a temperature of ∼600 °C. A similar dependence can also be observed for the Gd-doped sample. For the Mg-containing sample and, in a greater extent, for the La-containing sample, a ~2 times increase in CSR occurs at a higher temperature. Hence, with an increase in the calcination temperature, the sintering resistance of the materials increases in the series of dopant cations Gd < Mg < La. The obtained data are in good agreement with the data on the textural properties of the samples (Table 1, Fig. 1). Based on the results of the HRTEM study of the samples, it is found that cerium dioxide particles are represented by ∼10 nm-crystallites that form polycrystalline agglomerates mainly in the form of plates. On doping CeO2 by Gd, La, or Mg cations, the material retains the nanostructured organization: Ce0.8M0.2Oy particles are composed of crystallites with a characteristic size of up to 10 nm (Fig. 5). According to the HRTEM data, the average crystallite size increases in a series of dopant cations Mg ∼ La < Gd and is ∼5 nm and ∼7 nm respectively, which agrees with the powder XRD results (Table 1). According to the ESDR data, for the cerium dioxide samples calcined at different temperatures (300 °C, 500 °C, and 800 °C), there is strong absorption in a region above 25 000 cm–1 due to the appearance of ligand-to-metal charge transfer bands (CTBs) of Ce3+ and Ce4+ cations. Therewith, the fundamental absorption edge (FAE) energy of cerium dioxide calcined at a temperature of 300 °C is close to the FAE energy for the CeO2 single crystal (Eg = 3.15 eV [25]), being 3.09 eV. A rise in the calcination temperature of the samples from 300°C to 800°C leads to a blue shift of FAE of cerium dioxide due to a slight increase in the band gap: from 3.09 eV to 3.20 eV respectively. In addition, in the region above 25 000 cm–1 the
131
Fig. 5. Electron microscope images of the Ce0.8M0.2Oy samples: the effect of the type of the dopant cation on the nanostructure of the material M = Gd (a), La (b), Mg (c). The calcination temperature is 500 °C. absorption band intensity gradually decreases, which most likely indicates a decrease in the number of defects in the CeO2 structure with an increase in the calcination temperature of the samples. For cerium dioxide doped by La cations, similar changes in the ESDR spectra are observed on varying the calcination temperature of the material (Fig. 6a). The comparison of the ESDR and powder XRD data evidences that an increase in the crystallite size (CSR) and the degree of crystallinity of the materials is accompanied by a FAE shift to a short-wave region and a decrease in the absorption band intensity in the UV region of the ESDR spectra (Table 1, Figs. 2, 6). As seen from the analysis of the ESDR spectra of the cerium dioxide samples doped by different-type cations, (Fig. 6b), in a range 22 000-25 000 cm–1 the absorption edge smears out in a series Gd < La < Mg and Eg takes the values of 3.05 eV, 3.02 eV, and 2.99 eV for the samples doped by Gd, La, and Mg respectively. Moreover, when 50% of cerium oxide are replaced by Gd, La, and Mg cations, the absorption band intensity in the UV band of the ESDR spectra slightly increases, indicating some disordering of these systems as compared to initial CeO2 and an increase in the number of their defects. We used SEM to examine the effect of the calcination temperature of cerium dioxide on the morphology of its particles (Fig. 7). It can be seen that after calcination at a temperature of 300 °C, the sample contains particle agglomerations with irregular shapes and different sizes. The particles have a loose structure like a sieve or sponge (Fig. 7a). The pore walls are thin plates (Fig. 7b). The formation of a developed polydisperse pore system is a result of the burn off of the organic matrix of a polymer precursor. With an increase in the calcination temperature from 300 °C to 800 °C the particle morphology of cerium dioxide does not change (Fig. 7c, d). The sample retains high texture porosity. The study of the effect of the type (M = Gd, La, Mg) and content (molar ratio M/Ce = 0.1, 1.0) of the modifying additive on the particle morphology
Fig. 6. ESDR spectra of the CeO2 and Ce1–xMxOy samples: the effect of the calcination temperature M = La, x = 0.2 (a); the effect of the type of the dopant cation M = Gd, La, Mg, x = 0.5 (b).
132
Fig. 7. SEM images of CeO2: the effect of the calcination temperature on the material morphology. The calcination temperature is 300 °C (a, b), 800 °C (c, d). in the solid solution based on cerium dioxide shows that the introducing of the dopant cation into the composition of cerium dioxide does not make a noticeable effect on the particle morphology. The use of Ce1–xMxOy supports with different textural, structural, and morphological characteristics opens the way to regulate the dispersion and redox properties of active metal and, correspondingly, the parameters of autothermal ethanol reforming.
CONCLUSIONS With the aim to determine the regularities in the formation of nanostructured oxide materials and to obtain supports with the desired textural, structural, and morphological properties, a representative series of Ce1–xMxOy materials was synthesized by the method of ester polymer precursors. It is shown that the synthesized Ce1–xMxOy samples are mesoporous materials, being the homogenous substitutional solid solutions with a fluoride-like cubic structure. The unit cell parameter of Ce1–xMxOy is changed as compared to that of CeO2 according to the radius and content of the dopant cation. The textural and structural properties of Ce1–xMxOy (x = 0-0.5, 1.5 ≤ y ≤ 2.0, M = Gd, La, Mg) materials were regulated by varying the type of the dopant cation (M = Gd, La, Mg), the molar ratio M/Ce (0, 0.1, 0.25, 1), and heat treatment conditions (temperature 300 °C, 500 °C, and 800 °C; duration 4-24 h). The possibility to change the average crystallite size of Ce1–xMxOy in a range from 5 nm to 50 nm, the specific surface area of the material from 15 m2/g to 120 m2/g, the pore volume from 0.07 cm3/g to 0.22 cm3/g, the average pore diameter from 6 nm to 30 nm was shown. We revealed the relationship between the synthesis parameters and the characteristics of Ce1–xMxOy materials. It is established that in a series of the materials CeO2 < Ce1–xGdxOy < Ce1–xLaxOy < Ce1–xMgxOy the structure defectiveness increases, while the band gap and the average crystallite size decrease. The average size of Ce1–xMxOy crystallites also decreases with an increase in the molar fraction of the dopant cation and with a decrease in the calcination temperature and duration. It is determined that the best way to enhance the textural characteristics of Ce1–xMxOy and to increase the sintering resistance is to use La as the dopant cation. The work was supported by RFBR (Project No. 15-53-45039 IND_a) and preformed with the equipment of the Shared Research Center of the Federal Research Center of Coal and Coal Chemistry of the Siberian Branch of the RAS.
REFERENCES 1. Z. R. Ismagilov, V. V. Kuznetsov, L. B. Okhlopkova, L. T. Tsikoza, and S. A. Yashnik, Oxides of Titanium, Cerium, Zirconium, Yttrium, Aluminium: Properties, Application, and Synthesis Methods [in Russian], Izd. SO RAN, Novosibirsk (2010). 2. V. K. Ivanov, O. S. Polezhaeva, and Yu. D. Tret′yakov, Russ. J. Gen. Chem., 80, No. 3, 604 (2010). 3. 4. 5. 6.
T. G. Kuznetsova and V. A. Sadykov, Kinet. Catal., 49, No. 6, 886 (2008). G. Nahar and V. Dupont, Renewable Sustainable Energy. Rev., 32, 777 (2014). R. Perez-Hernandez, A. Gutierrez-Martınez, and J. Palacios, IJHE, 36, 6601-6608 (2011). I. Z. Ismagilov, E. V. Matus, V. V. Kuznetsov, et al., Catal. Today, 210, 10 (2013). 133
7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25.
134
P. Maki-Arvela and D. Yu. Murzin, Appl. Cat. A, 451, 251 (2013). I. Z. Ismagilov, E. V. Matus, V. V. Kuznetsov, et al., IJHE, 39, 20969 (2014). H. Kaneko, S. Taku, and Y. Tamaura, Sol. Energy, 85, 2321 (2011). T.-D. Nguyen-Phan, M. B. Song, E. J. Kim, et al., Microporous Mesoporous Mater., 119, 290 (2009). B. Zhang, D. Li, and X. Wang, Catal. Today, 158, 348 (2010). F. Liu, L. Zhao, H. Wang, et al., IJHE, 39, 10454 (2014). R. O. Fuentes and R. T. Baker, J. Power Sources, 186, 268 (2009). T. Zhang, D. Tang, Y. Shao, et al., J. Mater. Eng. Perform., 19, 1220 (2010). F. Zhang, Q. Jin, and S.-W. Chan, J. Appl. Phys., 95, 4319 (2001). L. Wu, H. J. Wiesmann, A. R. Moodenbaugh, et al., Phys. Rev. B, 69, 125415 (2014). O. O. Stoianov, V. K. Ivanov, A. B. Shcherbako, et al., Russ. J. Org. Chem., 59, 15 (2014). L. Katta, P. Sudarsanam, G. Thrimurthulu, et al., Appl. Catal. B, 101, 101 (2010). Q. Yu, X. Wu, C. Tang, et al., J. Colloid Interface Sci., 354, 341 (2011). M. Chen, H. Zheng, C. Shi, et al., J. Mol. Catal. A, 237, 132 (2005). W. Y. Hernandez, O. H. Laguna, M. A. Centeno, et al., J. Solid State Chem., 184, 3014 (2011). Z.-P. Li, T. Mori, J. Zou, et al., Mater. Res. Bull., 48, 807 (2013). D. H. Prasad, S. Y. Park, H.-I. Ji, et al., J. Phys. Chem. C, 116, 3467 (2012). X. Yao, C. Tang, Z. Ji, et al., Catal. Sci. Technol., 3, 688 (2013). A. Bensalem, J. C. Muller, and F. Bozon-Verduraz, J. Chem. Soc. Faraday Trans., 88, 153 (1992).
