Cite this Article
Ma Ming, Chen Yan, Cui Yi-Min. Effect of oxygen content on dielectric characteristics of Cr-doped LaTiOx *

Project supported by the National Natural Science Foundation of China (Grant No. 51571006).

 . Chinese Physics B, 2018, 27(5): 057702  
Permissions
Effect of oxygen content on dielectric characteristics of Cr-doped LaTiOx *

Project supported by the National Natural Science Foundation of China (Grant No. 51571006).
Ma Ming1, Chen Yan2, †, Cui Yi-Min2
School of Aeronautic Science and Engineering, Beihang University, Beijing 100191, China
School of Physics, Beihang University, Beijing 100191, China

 

† Corresponding author. E-mail: chenyan@buaa.edu.cn

Project supported by the National Natural Science Foundation of China (Grant No. 51571006).

Abstract

The ceramics La0.85Cr0.15TiOx and La0.7Cr0.3TiOx are prepared by conventional solid-state reaction method. The dielectric properties of Cr-doped LaTiOx as a function of frequency (0.1 kHz ≤ f ≤ 1 MHz) and temperature (77 K ≤ T ≤ 360 K) are studied. The blocks are then annealed in a flowing O2 or Ar/H2 to convert their oxygen content and the tests mentioned above are performed. The highly oxygenated samples exhibit extremely high low-frequency dielectric constants at room temperature (∼106). The results show that the oxygen stoichiometry could significantly influence the dielectric properties of Cr-doped LaTiOx.

PACS: 77.84.Bw;77.22.Gm;61.72.-y
Keyword:oxygen content;dielectric characteristics;LaTiOx
1. Introduction

Perovskite oxides have been extensively studied in recent years due to their various exotic properties. This is mainly because of their superior properties in dielectric, ferroelectric, piezoelectric, superconductivity, etc.[1–4] Lanthanum titanium oxide (LTO), as an important titanate material, has attracted a great deal of researchers’ attention recently.[5] Many researchers have studied LTO materials from different perspectives. For example, Schmehl et al. reported the transport properties of LaTiOx films and heterostructures.[6] Schmitz et al. discussed the effect of the experimentally observed Jahn-Teller distortion of the oxygen octahedronin LaTiO3 on the magnetic exchange.[7] Lunkenheimer studied the dielectric properties and dynamical conductivity of LaTiOx from dc to optical frequencies.[8]

Perovskite has the flexibility of changing the effective negative charges by performing doping substitution on the cationic sites. This class of compound can go through a wide range of structural phase transition, which can strongly affect its physical and chemical properties. Among various types of titanate perovskites, LTO is considered to be a very promising material.[9,10] Cu doping in LTO can not only improve its dielectric properties, but also enhance the Fenton degradation of rhodamine B.[11,12] The Sr doping lowers the conduction band and reduces the band gap of La1−xSrxTiO3.[13] As can be seen, the enhancement in total conductivity of the LTO perovskite can be achieved only at lower doping concentration. The opposite effect is observed for higher doping concentration.[14]

The Cr is an important doping element commonly used to improve the properties of the material.[15–17] It has been verified by first-principles calculations that Cr doping improves the conductivity of LiFePO4.[18] The Cr-doped SrTiO3 single crystals exhibit current-driven insulator–conductor transition characteristics which can be used for fabricating the next-generation non-volatile memory devices.[19] In Li3V2−xCrx(PO4)3/C, the initial specific capacity decreases with the decrease in Cr content at a lower current rate. Under the condition of the moderate Cr doping, both the cycle performance and rate performance have been improved.[20]

Annealing in different atmospheres may change the oxygen content of the material, thereby changing the properties of the material. After deposition annealing, the permittivity and the dielectric Q (1/tanδ) of single-phase thin film of Ba0.5Sr0.5TiO3 showed significant changes.[21] Titanium oxide films have better electrical properties after being annealed in oxygen.[22] Polycrystalline La0.67Ba0.33MnO3 bulk samples, which were obtained under the annealing conditions at 700 °C in the flowing 95%Ar:5%H2(Ar/H2) mixed gas or a different time, have significantly different insulator–metal transition temperature and AC magnetic susceptibility.[23] Oxygen content can significantly affect the dielectric properties of TiO2 crystals.[24]

In this study, the Cr-doped LaTiOx is synthesized using the solid-state method and the samples are annealed in O2 or Ar/H2 to change the oxygen content. The effects of Cr-doped ratio and oxygen content on the dielectric properties of Cr-doped LaTiOx are systematically investigated.

2. Experimental details

The ceramics La0.85Cr0.15TiOx and La0.7Cr0.3TiOx were prepared by solid-state reaction. The starting powders La2O3, Cr2O5, and TiO2 (with purity > 99.99%) were mixed as required and fired at 1450 °C for 600 min in the air with intermediate grinding in an agate mortar. The grinding and sintering process was repeated three times. Then the mixture was reground, pressed into pellets of 8 mm in diameter and sintered at 1520 °C in the air for another 600 min followed by surface cooling. In order to explore the influence of oxygen content, the samples were annealed at 700 °C in flowing O2 or Ar/H2 mixed gas for 60 min and then cooled down.

The structural analysis of the pellet was checked by x-ray diffraction (XRD) using Cu-Kα radiation with 2θ ranging from 20° to 80°. The chemical compositions and chemical states of the as-prepared pellets were characterized by x-ray photoelectron spectroscope (XPS, ESCALAB 250Xi, Thermo Scientific). The silver paste was coated on both sides of the disk as an electrode to test the dielectric properties. The dielectric properties of bulks were measured by a precise impedance analyzer 6500B (Wayne Kerr corp.). The temperature was automatically monitored and controlled by LakeShore temperature controller (Model: 331) with temperature ranging from 77 K to 360 K. Data measurement frequency was in a range from 0.1 kHz to 1 MHz.[25–32]

3. Results and discussion
3.1. Sample characterization

The XRD spectra of Cr-doped LaTiOx composites at room temperature are shown in Fig. 1. The intensity of the peak varies with the amount of chromium doped and the oxygen content. The analyses of the XRD spectra demonstrate that the lattice structure is tetragonal. For the original sample as shown in Figs. 1(a) and 1(b), except for the 200 peak, the peak intensities increase as the chromium concentration increases. The radius of Lanthanum ion is different from that of Chromium ion,[33] and thus Cr ionis substituted for La ion, the lattice structure of LaTiO3 can be distorted due to the different ionic radii. As shown in Figs. 1(c) and 1(d), comparing with La0.85Cr0.15TiOx+δ, the intensities of the peak (200) and (211) of La0.7Cr0.3TiOx+δ become weak and some impurity peaks appear. Comparing with the original samples, the intensities of the peaks (211) and (022) become weak after O2 annealing. As for the Ar/H2-annealed samples shown in Figs. 1(e) and 1(f), the intensity of peak (200) strengthens but that of (211) peak weakens. The (200) peak and (022) peak intensities become weak after Ar/H2 annealing. The reason for these results may be that the annealing conditions affect the crystallinity of the sample.[34]

Fig. 1. (color online) XRD spectra for (a) La0.85Cr0.15TiOx, (b) La0.7Cr0.3TiOx, (c) O2-annealed La0.85Cr0.15TiOx+δ, (d) O2-annealed La0.7Cr0.3TiOx+δ, (e) Ar/H2-annealed La0.85Cr0.15TiOx−δ, and (f) Ar/H2-annealed La0.7Cr0.3TiOx−δ at room temperature.

The x-ray photoelectron spectrum (XPS) analysis is also performed to explore the surface chemical compositions of the Cr-doped LaTiOx samples. The O 1s core level spectra of the prepared samples are displayed in Figs. 2(a)–2(f). In spite of the different annealing atmospheres and doping ratios of the ceramics, it is important to note that the spectra are virtually identical. This assures that in all cases a similar oxygen valence can be estimated at the surface and the change of the dielectric properties with the same doping proportion is mainly attributed to the difference in oxygen concentration.[35–37]

Fig. 2. (color online) O 1s XPS spectra of (a) La0.85Cr0.15TiOx, (b) O2-annealed La0.85Cr0.15TiOx+δ, (c) Ar/H2-annealed La0.85Cr0.15TiOx−δ, and (d) La0.7Cr0.3TiOx, (e) O2-annealed La0.7Cr0.3TiOx+δ, and (f) Ar/H2-annealed La0.7Cr0.3TiOx−δ.

Figures 3(a)–3(f) show La 3d states of the perovskite under different oxygen conditions, respectively. The core and the satellite peaks of La 3d5/2 and La 3d3/2 each show a clear doublet structure specific for oxides. The primary peaks are located at ∼834.62 eV for La 3d5/2 and ∼850.87 eV for La 3d3/2, respectively. The satellite peaks are around ∼838.47 eV and ∼855.32 eV, respectively. The energy difference between core and satellite peaks is about 4 eV. The presence of satellite peak La 3d is due to the monopole excitation arising from a change in the screening of the valence electrons upon the removal of a core electron. It is clear that the valence states of Laions correspond to those of La3+.[35–37] It is important that the valence of Laions do not vary significantly with the difference in oxygen concentration.

Fig. 3. (color online) La 3d XPS spectra of (a) La0.85Cr0.15TiOx, (b) O2-annealed La0.85Cr0.15TiOx+δ, (c) Ar/H2-annealed La0.85Cr0.15TiOx−δ, (d) La0.7Cr0.3TiOx, (e) O2-annealed La0.7Cr0.3TiOx+δ, and (f) Ar/H2-annealed La0.7Cr0.3TiOx−δ.

The Cr 2p spectra of the samples are exhibited in Fig. 4. It is clear that two peaks at 576.37 eV and 586.17 eV are assigned to the Cr3+ 2p3/2 and Cr3+ 2p1/2 orbits, respectively.[38,39] As for the O2-annealed ceramic, the peak at about 579.17 is attributed to the 2p3/2 of Cr6+ species,[38] and the peak at 573.17 is assigned to the Cr metal.[38] The reason may be that the electrons around the chromium species combine with the oxygen, which causes the chromium ions to change their valence.

Fig. 4. (color online) Cr 2p XPS spectra of (a) La0.85Cr0.15TiOx, (b) O2-annealed La0.85Cr0.15TiOx+δ, (c) Ar/H2-annealed La0.85Cr0.15TiOx−δ, (d) La0.7Cr0.3TiOx, (e) O2-annealed La0.7Cr0.3TiOx+δ, and (f) Ar/H2-annealed La0.7Cr0.3TiOx−δ.

The high-resolution XPS spectra of Ti2p are shown in Fig. 5. Titanium is a more active metal. Its valence is usually bivalent, trivalent, and tetravalent. As shown in Fig. 5, two primary peaks can be explained as a combined effect of surface Ti4+ and near-surface Ti3+ interstitial defects. In order to further study the influence of oxygen atmosphere on the titanium ions, the measured data are split, which are displayed in Figs. 5(a)–5(f). The symmetric curve with binding energy located at 458.7 eV and 464.3 eV is in accordance with those of Ti4+ 2p3/2 and Ti4+ 2p1/2 in samples, respectively. The two peaks at 457.9 eV and 463.6 eV are assigned to the Ti3+ 2p3/2 and Ti3+ 2p1/2 orbits, respectively.[40–42] The peak height and area can be used to characterize the proportions of elements. By comparison, it can be seen that the proportion of tetravalent titanium ions in the ceramic annealed in an oxygen atmosphere was the largest. This is due to the binding of titanium ions with oxygen ions, which leads to the lose of electrons.

Fig. 5. (color online) Ti 2p XPS spectra of (a) La0.85Cr0.15TiOx, (b) O2-annealed La0.85Cr0.15TiOx+δ, (c) Ar/H2-annealed La0.85Cr0.15TiOx−δ, (d) La0.7Cr0.3TiOx, (e) O2-annealed La0.7Cr0.3TiOx+δ, and (f) Ar/H2-annealed La0.7Cr0.3TiOx−δ.

Finally, the different annealing atmospheres can affect oxygen concentration of the ceramic. However, the valence of oxygen ion is not affected. As shown in Fig. 3, the valence states of La ions remain unchanged. In addition to La0.7Cr0.3TiOx annealed in oxygen, the valence states of Cr ions remain constant too. Therefore, we believe that the structural stability of the material should be attributed to the difference in concentration of Ti4+ ions.

3.2. Dielectric properties

The dielectric constant is a characteristic of material insulation, which indicates the ability of a substance to hold the charge. Figure 6 shows the temperature-sensitive properties of the dielectric constant ε* = ε′−iε″ (ε* is the complex permittivity, ε′ is the real part and ε″ is the imaginary part of the complex permittivity) of Cr-doped blocks at different frequencies. Figures 6(a) and 6(b) show the temperature dependence of the dielectric constant ε′ of the Cr-doped LaTiOx. As shown in Fig. 6(b), the dielectric constant of sample La0.7Cr0.3TiOx is independent of temperature and frequency at the temperature below 150 K. However, at temperatures above 150 K, the dielectric constant of the sample increases with increasing temperature. For instance, the ε′ value increases rapidly as the temperature rises from 150 K to 220 K when the measured frequency is 0.1 kHz. The phenomenon may be caused by thermally exciting more electron–hole pairs when the temperature is over 150 K.[43] As the frequency increases, the ε′ value exhibits temperature dependence in higher temperature region. In the temperature range of 220–300 K, the dielectric constant increases slowly as the carrier reaches a saturation state. The ε′ value increases rapidly again when the temperature is above 300 K, which may be due to another type of carrier that is excited at a temperature above 300 K.[44] The trend of ε′ with temperature is also affected by the frequency. According to Figs. 6(a) and 6(b), it can be concluded that the values of dielectric constant ε′ increase as the concentration of doped chromium increases. The ε′ of the sample turns into an extremely high low-frequency dielectric constants at higher temperature. It increases rapidly with increasing temperature when the temperature is higher than a certain point. The temperature point is also influenced by the chromium concentration.

Fig. 6. (color online) Temperature-dependent ε′ and tanδ values of ((a), (c)) La0.85Cr0.15TiOx and ((b), (d)) La0.7Cr0.3TiOx in temperature range from 77 K to 360 K for different frequencies.

Figures 6(c) and 6(d) show the dielectric loss tangent tanδ (tanδ = ε″/ε′) curves and temperature curves of the samples at different frequencies. For the sample La0.7Cr0.3TiOx, it is shown that the dielectric constant ε′ is independent of temperature nor frequency at temperatures below 150 K, which corresponds to the independence of ε′ mentioned above. For both samples, if d(tanδ)/dT = 0 moves toward high temperature as the measured frequency increases, the temperature exhibits a thermal excitation relaxation process.[28–30] The Cr doping ratio has a significant effect on the dielectric loss. The doping of Cr may affect the conductivity and polarization of the sample, both of which determine the dielectric loss of the sample.[43]

It can be concluded that the dielectric constant ε′ increases as the concentration of doped chromium increases. The dielectric loss is also significantly affected by the Cr doping ratio. This may be because when chromium ions replace lanthanum ions, the conductivity and the polarizability change, affecting the dielectric constant of the sample.[43]

3.3. Effect of O2 annealing on dielectric properties

The Cr-doped LaTiOx is oxidized when annealed in oxygen, which may affect the crystal structures of some types of perovskites, and influence the physical and chemical properties.[21] The bulks are annealed in oxygen and tested as mentioned above. Figures 7(a) and 7(b) show the temperature-dependent ε′ values of the oxygen-annealed Cr-doped samples. Figures 7(c) and 7(d) show the variations of dielectric loss tangent tanδ with temperature at different frequencies. According to Figs. 6 and 7, for all the samples, the values of dielectric constant ε′ and the dielectric loss tangent increase as the concentration of chromium doping rises. By comparing Fig. 6 with Fig. 7, the dielectric constant will increase after oxygen annealing.

Fig. 7. (color online) Temperature-dependent ε′ and tanδ values of the oxygen-annealed ((a), (c)) La0.85Cr0.15TiOx+δ and ((b), (d)) La0.7Cr0.3TiOx+δ in temperature range from 77 K to 360 K at different frequencies.

In conclusion, the dielectric properties of Cr-doped LaTiOx can be improved after being annealed in oxygen. It may be due to the fact that the number of activated carriers increases and the structure changes with rising oxygen content.[45–47] The high value of the dielectric constant may be attributed to the close-packed structure of oxygen ions. Oxygen ions with negative charges, because of two excess electrons, are more loosely bound to the nucleus than in a neutral atom or positive ion, so that the ion is readily distorted by an electric field. Also because of the close-packed structure, the number of polarized ions per unit volume is high, so when an electric field was applied, a high polarization increases, which results in a high dielectric constant.[24] In order to demonstrate the effect of oxygen content on the material property, the oxygen content of the material is reduced for further investigation.

The samples are reduced when annealed in Ar/H2 mixed gas. The oxygen content of the material became lower, which may affect the physical and chemical properties.[21] In order to study the effect of oxygen-deficient on the dielectric property, the samples are annealed in Ar/H2 and tested as done above. Figures 8(a) and 8(b) show the temperature dependent ε′ values and figures 8(c) and 8(d) show the temperature-dependent dielectric loss tangent tanδ values of Ar/H2-annealed samples at different frequencies. According to Figs. 6 and 8, for both samples, the dielectric constant ε′ increases as the concentration of doped chromium increases. However, the dielectric constant turns quite different after annealing. Compared Fig. 6(a) with Fig. 8(a), the dielectric constant ε′ values of the original samples will show obvious temperature dependence when the temperature is above 160 K, while those for theAr/H2-annealed samples are 240 K. The ε′ increases rapidly with increasing temperature when the temperature is higher than a certain point. The temperature point becomes higher after the sample has been annealed in Ar/H2. A comparison between Figs. 6(b) and 8(b) shows that the non-annealed sample exhibits a very high low-frequency dielectric constant at room temperature and the low-frequency dielectric constant of the annealed sample does not exceed 2 × 103. The dielectric constant of the La0.7Cr0.3TiOx after being annealed is also much smaller.

Fig. 8. (color online) Temperature-dependent ε′ and tanδ values of Ar/H2-annealed ((a), (c)) La0.85Cr0.15TiOx−δ and ((b), (d)) La0.7Cr0.3TiOx−δ in temperature range from 77 K to 360 K at different frequencies.

A comparison of Fig. 6 with Fig. 8 shows that the dielectric loss tangent values of the samples with two different doping ratios decrease. As shown in Fig. 8(d), only the 0.1 kHz frequency curve has a maximum value over the tested temperature range. Both curves of non-annealed La0.85Cr0.15TiOx and La0.7Cr0.3TiOx exhibit their maximum values, while the curves of Ar/H2-annealed samples each have only one extreme value. It is assumed that extreme point will appear in the high-temperature region. In conclusion, the dielectric loss tangent decreases, and most of the extreme value points disappear in the test temperature range. This may be due to the decrease in the number of activated carriers and the increase in the temperature at which the carriers are activated after Ar/H2-annealing.

A series of distinct phases with variable oxygen stoichiometry exists in the Cr-doped LaTiOx system. Starting with fully oxidized compounds, the oxygen content is then reduced and the perovskite-related structure changes. The structural and physical properties vary drastically as oxygen quantity becomes stoichiometric.[47] In the Ar/H2-annealed samples, due to the low oxygen content, vacancies are formed in the samples. Charges trapped by the vacancies cannot participate in the long-distance conduction, but form a space charge. Many oxygen vacancies exist in the sample with low oxygen content, which can capture carriers. At lower voltage, most of the charges injected from the electrode fill the vacancies and cannot participate in the electrical conductivity resulting in a decrease in dielectric constant. For samples with high oxygen content, the oxygen vacancies are less and the oxygen ions are tightly linked, most of the charges are involved in the conductance. Due to the close-packed structure of oxygen ions, the number of polarizable ions is high. When an electric field is applied, a high polarization is established, resulting in a high dielectric constant. Space charges limit the current, thus dielectric loss varies with oxygen content. The current density decreases as the oxygen content decreases, resulting in a decrease in dielectric loss. This also shows that under the experimental conditions, the dielectric loss is determined mainly by the macroscopic charge transfer.[24,48]

The oxygen ions (O2 −) act as a bridge between Ti3+ and Ti4+ and play an essential role in the Ti3+–O2–Ti4+ superexchange interaction. Naturally, the extent to which the number of oxygen ions changes makes the crystal lattice structure collapse down, so the dielectric property of the material changes significantly.

4. Conclusions

In summary, doped Cr ion concentration can significantly influence the dielectric property of the LTO. The doping ratio increases from 15% to 30%, the dielectric constant is increased by almost an order of magnitude. Annealing in flowing O2 or Ar/H2 can change the oxygen content and influence the dielectric properties of the samples. The sample with high oxygen content shows higher dielectric constant, which means that O2 annealing is an effective method to improve the dielectric characteristics of Cr-doped LTO.

Reference
[1] Ahn C H Triscone J M Mannhart J 2003 Nature 424 1015
[2] Cheng J G Meng X J Li B Tang J Guo S L Chu J H 1999 Appl. Phys. Lett. 75 2132
[3] Maurya D Pramanick A An K Priya S 2012 Appl. Phys. Lett. 100 172906
[4] Suzuki H Bando H Ootuka Y Inoue I Yamamoto T Takahashi K Nishihara Y 1999 J. Phys. Soc. Jpn. 65 1529
[5] Moetakef P Zhang J Y Kozhanov A Jalan B Seshadri R Allen S J Stemmer S 2011 Appl. Phys. Lett. 98 112110
[6] Schmehl A Lichtenberg F Bielefeldt H Mannhart J Schlom D G 2003 Appl. Phys. Lett. 82 3077
[7] Schmitz R Entin-wohlman O Aharony A Harris A B Muellerhartmann E 2005 Phys. Rev. 71 144412
[8] Lunkenheimer P Rudolf T Hemberger J Pimenov A Tachos S Lichtenberg F Loidl A 2003 Phys. Rev. 68 245108
[9] Madhavan B Ashok A 2015 J. Sol-Gel Sci. Technol. 73 1
[10] Fasquele D Carru J C Gendre L L Paven C L Pinel J Cheviré F Tessier F Marchand R 2005 J. Eur. Ceram. Soc. 25 2085
[11] Chen Y Xu J X Cui Y M Shang G Y Qian J Q Yao J E 2016 Prog. Nat. Sci.: Mater. Int. 26 158
[12] Zhang L L Nie Y L Hu C Qu J H 2012 Appl. Catal. 125 418
[13] Gao L H Ma Z Fan Q B 2011 J. Electroceram. 27 114
[14] Bradha M Hussain S Chakravarty S Ashok A Amarendra G 2014 Ionics 20 1343
[15] Kim J K Sang S S Kim W J 2006 Appl. Phys. Lett. 88 132901
[16] Janousch M Meijer G I Staub U Delley B Karg S F Andreasson B P 2007 Adv. Mater. 19 2232
[17] Borgarello E Kiwi J Graetzel M Pelizzetti E Visca M 1982 J. Am. Chem. Soc. 104 2996
[18] Shi S Liu L Ouyang C Wang D S Huang X J Chen L Q 2003 Phys. Rev. 68 195108
[19] Watanabe Y Bednorz J G Bietsch A Gerber C Widmer D Beck A 2001 Appl. Phys. Lett. 78 3738
[20] Chen Y H Zhao Y M An X N Liu J M Dong Y Z Chen L 2009 Electrochim. Acta 54 5844
[21] Chang W Horwitz J S Carter A C Pond J M Kirchoefer SW Gilmore C M Chrisey D B 1999 Appl. Phys. Lett. 74 1033
[22] Lee M K Huang J J Wu T S 2005 Semicond. Sci. Technol. 20 519
[23] Cui Y M Zhang L W Wang C C Shi K Cao B S 2006 J. Magn. Magn. Mater. 297 21
[24] Bamzai K K Koohpayeh S M Kaur B Fort D Abell J S 2008 Ferroelectrics 377 1
[25] Cui Y M Wang R M 2007 Appl. Phys. Lett. 91 233513
[26] Cui Y M Cai W Li Y Qian J Q Xu P Wang R M Yao J E 2006 J. Appl. Phys. 100 034101
[27] Chen C Xu K B Cui Y M Wang C C 2012 Mater. Lett. 89 153
[28] Lu C Cui Y M 2012 Physica 407 3856
[29] Xu J X Cui Y M 2013 Mater. Sci. Eng. 178 316
[30] Cui Y M Zhang L W Xie C L Wang R M 2006 Solid State Commun. 138 481
[31] Cui Y M Zhang L W Wang R M 2006 Physica 442 29
[32] Cui Y M Liu W Wang R M 2013 Phys. Chem. Chem. Phys. 15 6804
[33] E R N Jr 1959 Biochim. Biophys. Acta 63 1381
[34] Vidyasagar C C Muralidhara H B Naik Y A Gururaj H Ilango M S 2015 Energy Environ. Focus 4 54
[35] López G P Castner D G Ratner B D 1991 Surf. Interface Anal. 17 267
[36] Miao J P Lü Z Li L P Ning F L Liu Z G Huang X Q Sui Y Qian Z N Su W H 2000 Mater. Lett. 427 267
[37] Marshall M S J Newell D T Payne D K Egdell R G Castell M R 2011 Phys. Rev.B 83 035410
[38] Daulton T L Little B J 2006 Ultramicroscopy 106 561
[39] Zhong L Cai W Zhong Q 2014 Rsc Adv. 4 43529
[40] Drera G Salvinelli G Brinkman A Huijben M Koster G Hilgenkamp H Rijnders G Visentin D Sangaletti L 2012 Phys. Rev. 87 1081
[41] Zhou Y Chen C H Wang N N Li Y Y Ding H M 2016 J. Phys. Chem.C 120 6116
[42] Guillemot F Porte M C Labrugere C Baquey C H 2002 J. Colloid Interface Sci. 255 75
[43] Dholam R Patel N Adami M Miotello A 2009 Int. J. Hydrogen Energy 34 5337
[44] Pesci F M Wang G Klug DR Li Y Cowan A J 2013 J. Phys. Chem. 117 25837
[45] Wang G J Wang C C Huang S G Lei C M Mei J Y Sun X H Li T 2012 J. Electroceram. 28 172
[46] Zhang Y T Wang C C He M 2009 J. Phys. D: Appl. Phys. 42 055309
[47] Lichtenberg F Widmer D Bednorz J G Williams T Reller A 1991 Z. Phys. 82 211
[48] Chang F G Song G L Fang K Wang Z K 2007 Acta Phys. Sin. 56 6068