Contrary interfacial effects for textured and non-textured multilayer solid oxide electrolytes

Abstract

The relationship between the microstructure and the conductivity for nanocrystallized oxygen ionic conducting thin films has been receiving great interest since it provides guidelines for designing electrolytes with high performances which might find applications in fuel cells and oxygen and fuel separation membranes. Here, we present a strategy for using the multilayered structure to tune the microstructures and ionic transport properties of solid electrolyte. Textured and non-textured Ce_0.8Sm_0.2O_2−δ/Al₂O₃ (SDC/AO) solid electrolyte multilayers were prepared, and the dependence of conductivity on layer number was studied. We found that non-textured and textured multilayers show a positive and a negative interfacial conduction contribution to the total ionic conductivity, respectively. The decrease of conductivity with the increase of layer number for textured SDC/AO was attributed to that the multilayered structure introduces random grain orientations to the interfacial region which results in more pronounced grain boundary blocking effects. In contrast, non-textured SDC/AO show rich structural defects in the interfacial regions which facilitate the oxygen ionic transport and lead to a higher ionic conductivity. These insights into the effect of the interfacial interaction on the structure and the conductivity allow a better control of the electrical properties of multilayered electrolytes, which might foster their applications in electrochemical devices operable at lower temperatures.

---

1. Introduction

There has been extensive research interest in solid oxide electrolytes recently due to their potential applications in various electrochemical devices including solid oxide electrolysis cells (SOECs), solid oxide fuel cells (SOFCs) and oxygen sensors.^1–4 Up to now, one major issue that hinders the practical applications of solid oxide electrolytes is the need for a high temperature (usually above 750 °C) to obtain sufficient ionic conductivity (0.01 S cm⁻¹).⁵ To bring down the temperature to intermediate temperature (<600 °C) can enhance the lifetime of devices, lower the demands for seal materials and reduce costs.^6,7 However, achieving a high conductivity for oxygen ionic conductors at temperatures below 600 °C is still a great challenge. Nanostructured solid electrolyte materials, particularly thin film electrolytes, have been investigated extensively recently due to the exceptional electrical properties.^8,9 When the size of functional materials decreases to the nanoscale, the micostructures and properties can differ remarkably from those in the bulk counterpart, owing to the nano effect.^10,11 For example, it has been reported that the mobility of oxygen ions in nanocrystallized solids is sensitive to the interface.^12,13 We have found that the migration of oxygen ions parallel to the GBs of scandia doped zirconia (10ScSZ) nanowires is significantly increased compared to that in the bulk, while the transport perpendicular to GBs is much slower.¹⁴ Thus, the ionic conductivity of nanostructured electrolyte strongly depends on the microstructures including grain sizes, grain orientations and morphologies which influence the interface resistances.^14–17

Highly textured and non-textured structures are two typical microstructures for polycrystalline thin film electrolyte which possess different electrical properties.¹⁷ It was found in our previous study that textured Sm and Nd co-doped ceria (SNDC) film exhibits a higher ionic conductivity than that of the non-textured one.¹⁸ Similar results have been also found for YSZ^19,20 and GDC²¹ thin films in literature. The extraordinary ionic conductivity for the textured electrolyte films has been mainly attributed to the reduction of grain boundary (GB) resistances.^21–23 So far, the relationship between the ionic conductivity and the crystallite orientation of single-layered electrolyte films has been extensively investigated. However, such relationship for the multilayered films which might possess higher ionic conductivities^24–29 can be more complex and need a better understanding. For example, for polycrystalline thin film heterostructures, the strong interfacial interaction between two materials can influence the growth mechanism and the microstructures of films.^15,30 In our previous work, we controlled the grain orientations of 10ScSZ electrolyte film using an amorphous alumina interlayer and observed enhanced ionic conductivity.³⁰ It is speculated that such interface effect on film structures can be achieved in polycrystalline oxide multilayers, which is promising to tune the conductivity of electrolyte by rational design of multilayered structure. Therefore, systematic investigations on the correlation between microstructures and layer number and its effect on the ionic conductivity of polycrystalline oxide multilayer, i.e. the structure–property relationships, are of great significance but are still lacking. Herein, highly-textured and non-textured SDC/AO multilayered electrolytes with different layer numbers were prepared and the electrical properties were studied. Opposite interfacial conduction contributions to the total ionic conductivity were observed for textured and non-textured multilayers. The apparent contradictory electrical behaviors were explained by the difference in microstructures including the grain sizes, grain orientations, GBs and structural defects of the interfacial regions.

2. Experimental

The target materials were purchased from Hefei Kejing Materials Technology Co. Ltd (Hefei, China). The diameter and thickness of Ce_0.8Sm_0.2O_2−δ and Al₂O₃ ceramic target were 50 mm and 3 mm, respectively; and the ceramic target was bonded with a copper plate with the same diameter and a thickness of 2 mm. The SDC/AO multilayer was deposited on the polycrystalline alumina substrate (with high purity of ∼99.99%) by radio frequency (RF) magnetron sputtering at 600 °C and at room temperature (RT), respectively. Prior to the film deposition, the main sputtering chamber was evacuated to background pressure below 10⁻⁶ mTorr. The working gas of Ar and O₂ was introduced into the sputtering chamber with the Ar/O₂ ratio of 4 : 1 and the work pressure of 5 mTorr. SDC and AO layers were sputtered sequentially with a control program that determined the sputtering power applied to each target, i.e. 200 W and 150 W for SDC and AO, respectively, and toggling of desired target shutters.

X-ray diffraction (XRD, D/max-2500, Rigaku, Tokyo, Japan) in both continuous and step-scan modes were applied to determine the phase structures, crystallite orientations and crystallite sizes of the multilayers. To confirm the crystallite orientation, pole figures were obtained in the tilt-angle (χ) range of 0–80° and the azimuth-angle (φ) range of 0–360° with χ and φ steps of 2.5°. Microstructural characterizations for the cross-section of the multilayers were performed by transmission electron microscope (TEM, FEI Tecnai F20) and high resolution TEM (HRTEM) to characterize the thickness of individual SDC layer, morphologies, crystallite sizes, GBs and structural defects. Scanning TEM (STEM) was performed under the high-angle annular dark field (HAADF) condition, and the contrast of STEM image depends on the atomic number. The composition separation across heterophase interfaces was revealed by line scanning energy dispersive X-ray (EDX) analysis performed in STEM mode. The impedances of multilayer in in-plane direction were measured by electrochemical workstation (IM6, Zahner, Kronach, Germany) from 450 °C to 700 °C in dry air over the frequency range from 8 MHz to 1 Hz. The electrical conductivity (σ) of multilayers was calculated by eqn (1):

where R is the resistance derived from the impedance spectrum, b is the width of two electrodes and L is the separation between the electrodes, d is the thickness of individual SDC layer and N is the total layer number.

3. Results and discussion

The crystallite orientation of ceria-based thin films is dependent on the deposition temperature of the film.¹⁸ Highly-textured SNDC films can be obtained at a high deposition temperature, while the non-textured counterpart was achieved when it is deposited at RT.¹⁸ Thus, the SDC/AO multilayers with different layer numbers were deposited at 600 °C and RT to obtain highly-textured and non-textured multilayers, respectively. The crystallite structures of these two series of multilayers and the SDC films were characterized using XRD (Fig. 1). As can be seen, except for the peaks of alumina substrate (marked by solid black squares), all diffraction peaks are indexed to the cubic phase of CeO₂ (JCPDS card no. 43-1002). The SDC films and the SDC layers for all multilayers deposited at 600 °C are (111)-textured (Fig. 1a), while the RT-deposited ones show a non-textured structure (Fig. 1b). The as-prepared samples were denoted as SDC/AO-T, SDC-T, SDC/AO-NT and SDC-NT where T stands for textured and NT refers to non-textured samples. No obvious shifts of the peak positions were found for both textured and non-textured samples when the layer number increases, indicating that the interface strain in these polycrystalline films is largely relaxed. It differs from previous findings^31,32 that an increase of the layer number for epitaxial multilayers usually leads to shifts in peak positions due to the interface strain.

Fig. 1 XRD patterns for (a) SDC/AO-NT and (b) SDC/AO-T multilayers with different layer numbers.

For an epitaxially grown oxide multilayer, the change of lattice constant at the hetero-interface can influence ionic conductivity significantly.³¹ In contrast, for our multilayered nanostructures, the effects of microstructures such as crystallite sizes, crystallite orientations and structural defects in the heterophase interfacial region are more pronounced due to the polycrystalline nature. The average crystallite sizes were determined by Scherrer's formula and the texture coefficient f₍₁₁₁₎ was calculated using Harris equation,³³ which are plotted as a function of layer number in Fig. 2a and b, respectively. It is found that the crystallite size of SDC/AO-T decreases with the increase of layer number while that of SDC/AO-NT is less sensitive to the layer number. Moreover, the f₍₁₁₁₎ of all SDC/AO-NT samples is smaller than 1 even though it changes slightly as the layer number increases, which confirmed that they are all non-textured.¹⁶ In contrast, SDC/AO-T are highly (111)-textured since they all show f₍₁₁₁₎ much higher than 1. The f₍₁₁₁₎ decreases by ∼0.75 as the layer number increases from 1 to 12, indicating a lower degree of (111) orientation. Pole figures of (111) planes for SDC/AO-T (Fig. 3) were obtained to further confirm the dependence of texture on layer number. It is observed that a larger layer number results in a broader distribution of (111) planes, which is consistent with the texture coefficient values. Therefore, the decrease of average crystallite size for SDC/AO-T is accompanied by a decrease in the degree of (111) texture, which is clearly correlated with the increased interfacial regions. That is, the multilayered structure changes the crystallite sizes and orientations of SDC in the interfacial regions of multilayers. However, such interfacial effect is more pronounced for the textured SDC/AO multilayers.

Fig. 2 (a) The average crystallite size and (b) the texture coefficient f₍₁₁₁₎ with respect to the layer number for both SDC/AO-T and SDC/AO-NT multilayers.

Fig. 3 Pole figures for (a) SDC film (N = 1) and SDC/AO-T multilayers with layer numbers of (b) N = 4, (c) N = 8 and (d) N = 12.

TEM was also employed to reveal the chemical composition and microstructures of multilayers. Fig. 4a shows a typical cross-sectional HAADF STEM image for multilayer, where well-defined and homogeneous bi-layers composed of SDC phase and AO phase are clearly seen. Typical EDX line scan spectra for Al element and Ce element obtained from the area marked by the red arrow in Fig. 4a are shown in Fig. 4b, which reveals a sharp chemical separation between the adjacent layers. Furthermore, compositions for the SDC phase in the multilayers with different layer numbers were determined by quantitative EDX. It is found that multilayered structure would not significantly alter the composition of SDC phase relative to that of the single-layered film, and the composition is also independent of the orientation as well. For example, the stoichiometric proportion of SDC phase is Ce_0.810Sm_0.190O_2−δ and Ce_0.808Sm_0.192O_2−δ for the SDC/AO-T (N = 12) and the SDC/AO-NT (N = 10), respectively, which are very similar.

Fig. 4 (a) Cross-sectional HAADF STEM image for SDC/AO-T. (b) EDX line scan spectra for Al and Ce element in the area marked by red arrow in (a). (c) High-magnification cross-sectional TEM image for SDC/AO-T. The left panel of (d) and (e) are HRTEM images for interfacial regions marked by 1 and 2 in (c), respectively. The right upper and lower panels of (d) and (e) are FFT and IFFT images for areas marked by the red square in left panels.

The microstructures for the cross-section of SDC/AO-T multilayers were carefully inspected. As shown in Fig. 4c, the SDC layer in the multilayer shows a columnar morphology that is typical for textured films.³⁴ Fig. 4d and e are typical HRTEM images for transitional regions between a SDC layer and its two adjacent AO layers, i.e. the interface 1 and the interface 2, respectively. We found that small granules with average sizes of 5 nm form near interface 1, while columns with an average width of 30 nm span across the rest of layer reaching interface 2. It is estimated that small SDC granules form at the initial stage of film deposition and the crystals with (111) orientation possess the highest growth rate and grow into large columns,³⁵ resulting in a textured structure. Moreover, fast Fourier transformation (FFT) and inverse FFT (IFFT) images were obtained to get a closer look at GBs between the granules and the columns in these two regions (right upper and lower panels). For example, it is found that granules near interface 1 exhibit random crystallite orientations which result in high angle GBs (Fig. 4d). For columns in Fig. 4e, the grain orientations were determined by the indices of two lattice planes and the dihedral angle, which are both 〈101̄〉 orientation. Therefore, the textured columns in SDC/AO-T show a much lower GB orientation compared with the non-textured granules grown near the heterointerface. Similar results have been also reported by Rupp et al., and they found that the YSZ lattice is continuous at the GB of columns in textured YSZ films which form low-angle GBs.^36,37

The morphologies and microstructures of SDC/AO-NT multilayer were also characterized by TEM. It can be seen from Fig. 5a that the SDC layers of SDC/AO-NT show the typical morphology for non-textured ceria-based oxide films,¹³ namely non-columnar granules with a homogeneous size stack in the film. Further revealed by HRTEM, it is found that the crystallite sizes and crystallite orientations of SDC in the interfacial regions for SDC/AO-NT are independent of the multilayered structure. Instead, the transitional regions are segregated with structural defects. A typical HRTEM image for the interfacial regions of SDC/AO-NT is shown Fig. 5b. Except for the ubiquitously observed vacancies and dislocations, amorphous phases (less crystalline phases) are also found. EDX analysis showed that these amorphous phases correspond to SDC instead of AO. The structures of these nanodomains can be well identified from images shown in Fig. 5c and d. It should be also noted that these structurally disordered nanodomains cannot be found in the interfacial regions of SDC/AO-T. Conclusively, the XRD and TEM results confirm the polycrystalline SDC/AO multilayers with both highly (111)-textured structures and non-textured structures. Fig. 6a and b are schematic diagrams for the microstructures and physical models i.e. ionic conduction model, of SDC/AO-T and SDC/AO-NT, respectively. For both multilayers, the structure of interfacial regions is different from that of the SDC volume phase. SDC/AO-NT show the interfacial regions with similar grain sizes and orientations but tailored structural defect models relative to the SDC volume phases. In contrast, the grain sizes and orientations of SDC in interfacial regions of SDC/AO-T are changed, and thus each SDC layer consists of a random interfacial region and a textured region with the thickness of δ and d − δ, respectively. Thus, the ionic conduction in the interfacial regions of both multilayers differs from that in the SDC volume phase, and the total ionic conductivity would be changed as the interfacial conduction contribution increases.

Fig. 5 (a) Cross-sectional TEM image for SDC/AO-NT. (b) Cross-sectional HRTEM image for SDC/AO-NT. (c) and (d) are FFT and IFFT images for the area marked by the red square in (b).

Fig. 6 Schematic diagram for the ionic conduction pathways in (a) SDC/AO-NT multilayers and (b) SDC/AO-T multilayers.

To measure the ionic conductivity parallel to the interfaces of multilayers, a two-probe method was used and side-deposited electrodes were applied (Fig. 7a). As can be seen, the multilayer was grown at the center of alumina substrate (Fig. 7b) with the aid of a mask (Fig. 7c) covering the two opposite sides. Then, Pt pad was deposited adjacent to the two sides of multilayer at blank areas on the substrate with the help of another mask (Fig. 7d). Ag paste was applied over Pt pads to enhance the electrical contact and fix Ag wires. Fig. 8a shows the typical impedance spectra for SDC/AO multilayers measured at different temperatures. Only a single arc was seen in all spectra, agreeing with those of nanocrystalline SDC ceramics.³⁸ The conductivities for SDC/AO-T and SDC/AO-NT are plotted in Fig. 8b as a function of the layer number. It is found that SDC-T film shows a higher conductivity than SDC-NT film, which is consistent with our previous findings on the enhanced conductivity for textured SNDC films than the non-textured ones.¹⁸ Interestingly, the conductivity of SDC/AO-NT increases with layer number, whereas an opposite trend was observed for SDC/AO-T. SDC/AO-T and SDC/AO-NT show a negative and a positive interfacial conduction contribution to the total ionic conductivity, respectively. As a result, SDC/AO-NT exhibit even higher conductivities than SDC/AO-T when the layer number exceeds 6.

Fig. 7 Schematic illustration of the side-deposited Pt electrode configuration for electrical measurement of the film samples: (a) the side view and (b) the top-view. The distance between two electrodes is denoted as L. (c) and (d) are schematic illustration of the masks which are used to deposit the multilayer and Pt electrodes, respectively. The areas with italic dashes are solid and the white areas are void.

Fig. 8 (a) Typical impedance spectra for SDC/AO-T multilayers measured at different temperatures. (b) The ionic conductivity which is measured at 600 °C for SDC/AO-T and SDC/AO-NT multilayers as a function of layer number.

The dependence of the conductivity for SDC/AO multilayers on layer number was further investigated. We elaborate the electrical conductivities of SDC/AO-T as an example; SDC/AO-NT show contrary relationships that can be deduced in a similar way.²⁸ Three sets of SDC/AO-T with layer numbers of 4, 8 and 12 were used. The thickness of individual SDC layer and the layer number were tuned to ensure the total thickness of SDC layers remains almost unchanged (Table 1). Fig. 9a shows the Arrhenius plots of the conductivity for SDC film (N = 1) and for SDC/AO-T (N = 4, 8, 12). It can be seen that the conductivity of multilayers decreases with the increase of layer number. Moreover, it is accompanied by an increase of activation energy E_a; the E_a increases by 0.2 eV as N varies from 1 to 12, which indicates a higher energy barrier for the oxygen ion transport in interfacial regions. The relationship between the total ionic conductivity (σ_tot) and the conductivities of textured region (σ_t) and random region (σ_r) can be derived by the classic rule of mixture²⁹ as described in eqn (2):

Table 1 List of investigated SDC/AO-T multilayers. Nd/2 is total thickness of SDC layers

N	1	4	8	12
d [nm]	252	243	115	85
Nd/2 [nm]		486	460	510
f(111)	2.89	2.70	2.47	2.14

---

Fig. 9 (a) The Arrhenius plots of ionic conductivity as a function of measuring temperature for SDC/AO-T with different layer numbers. (b) Variation of the conductivity of SDC/AO-T at different temperatures with the reciprocal layer thickness. (c) Arrhenius plots of the derived σ_r and σ_t for SDC/AO-T along with the conductivity of SDC-T film. (d) The correlation between the conductivity of SDC/AO-T and the texture coefficient f₍₁₁₁₎.

It can be seen that σ_tot is linearly dependent on the reciprocal thickness 1/d on conditions when δ ≦ d. Fig. 9b shows the conductivity of SDC/AO-T measured at 600 °C, 650 °C and 700 °C as a function of 1/d. As can be seen, the total conductivity decreases linearly with the decrease of d at all temperatures. Moreover, the conductivity of (111)-textured phase σ_t can be determined from the intercept with σ_tot axis by a linear fit to datas in Fig. 9b, e.g. it is ∼0.0153 S cm⁻¹ at 600 °C. This is in agreement with the ionic conductivity of (111)-textured SDC film, i.e. 0.0191 S cm⁻¹ at 600 °C (see Fig. 9a). From the slope δ(σ_r − σ_t) (δ ≦ 80 nm) it can be also estimated that the conductivity of random region σ_r is at most 0.00468 S cm⁻¹ at 600 °C, which is ∼80% lower than that of textured regions. In addition, we plot the derived values of σ_r and σ_t and that for SDC film as a function of the measuring temperature in Fig. 9c and the E_a was determined from the slope of plots. It is observed that the derived values of σ_t agree well with that of the (111)-textured SDC film, which shows a comparable E_a of ∼0.5 eV. But the conductivity of the random region is significantly lower than that of the textured SDC phase, which shows a E_a of ∼0.7 eV, confirming that oxygen ion migration in interfacial random regions is remarkably impeded compared with that in textured regions. It is also confirmed by the correlations shown in Fig. 9d that the ionic conductivity of SDC/AO-T increases with the increase of texture coefficient. The total ionic conductivity decreases with the increase of the proportion for interfacial regions, which indicates a lower conductivity for non-textured regions than that for textured regions. We believe that this is the fundamental mechanism for the decrease of conductivity of SDC/AO-T with the increase of layer number.

It is concluded that the difference in microstructures of the textured and non-textured SDC/AO multilayers leads to different electrical properties. The negative interfacial effect for SDC/AO-T, i.e. a lower ionic conductivity for the interfacial random region than that for the textured region is attributed to the different density and species of GBs in these two regions. It is recognized that in the nanostructured solid electrolyte with a high density of GBs, GBs that are vertical to the direction of ionic transport decrease the ionic conductivity, which is several orders of magnitude lower than that in the bulk.^17,39–44 More specifically, low-angle GBs are preferred ionic transport pathways over high-angle GBs.³⁹ For instance, Yoon et al. have found that ∑5 (310) GB exhibits a higher E_a and a more pronounced blocking effect on the oxygen ionic diffusion than ∑13 (510) GB.⁴⁵ Thus, the ionic transport across interfacial random regions (containing high-angle GBs) undergoes higher barriers compared with textured regions (containing low-angle GBs), which leads to a decrease of ionic conductivity for SDC/AO-T as the layer number increases. In contrast, slight changes in grain orientations of SDC/AO-NT cannot contribute significantly to the GB conductivity due to their non-textured structures. SDC/AO-NT show positive interfacial effect since disordered interfacial regions with rich structural defects provide fast oxygen ionic migration pathways.²⁸ Therefore, the seemly contradictory interfacial effect for SDC/AO-NT and SDC/AO-T is inherently associated with their differences in grain orientations, densities of structural defect as well as the proportion and species of GBs. This is different from the opposite interfacial effects observed from YSZ/Y₂O₃ and YSZ/Sc₂O₃ multilayers, where the dilatative strain and the compressive strain that is tuned by the difference in lattice constants, are responsible for the change of ionic conductivity.²⁵ In the current study of polycrystalline multilayered solid electrolytes, altered ionic transport properties for the interfacial regions are also found, but it is attributed to a mechanism different that for the epitaxially grown multilayers.

4. Conclusions

Textured and non-textured SDC/AO multilayer electrolytes with different layer numbers were deposited by magnetron sputtering, and their dependences of conductivity on the layer number were investigated. It is found that SDC/AO-T show negative interface conduction contribution to the total conductivity, which is opposite to the positive interfacial effect observed from SDC/AO-NT. The multilayered structure introduces random orientations to the interfacial region of SDC/AO-T, which blocks the migration of oxygen ions across high-angle GBs. In contrast, it increases the conductivity of SDC/AO-NT where interfacial regions show disordered nanodomains. The reported new concept of using a multilayer design to tune structures and electrical properties of solid oxide electrolyte is a promising way to apply polycrystalline multilayered electrolytes as the building block for a range of electrochemical devices.

Acknowledgements

This work was supported by Toyota Motor Corporation and National Natural Science Foundation of China (Grant no. 51323001, 51272120, 51072088).

References

1. D. M. Bierschenk, J. R. Wilson and S. A. Barnett, Energy Environ. Sci., 2011, 4, 944–951.
2. C. Graves, S. D. Ebbesen, S. H. Jensen, S. B. Simonsen and M. B. Mogensen, Nat. Mater., 2015, 14, 239–244.
3. S. Sengodan, S. Choi, A. Jun, T. H. Shin, Y. W. Ju, H. Y. Jeong, J. Shin, J. T. S. Irvine and G. Kim, Nat. Mater., 2015, 14, 205–209.
4. P. P. Dholabhai, J. B. Adams, P. Crozier and R. Sharma, Phys. Chem. Chem. Phys., 2010, 12, 7904–7910.
5. B. Li, J. M. Zhang, T. Kaspar, V. Shutthanandan, R. C. Ewing and J. Lian, Phys. Chem. Chem. Phys., 2013, 15, 1296–1301.
6. E. D. Wachsman and K. T. Lee, Science, 2011, 334, 935–939.
7. Z. Gao, V. Y. Zenou, D. Kennouche, L. Marks and S. A. Barnett, J. Mater. Chem. A, 2015, 3, 9955–9964.
8. Y. Ma, X. D. Wang, S. H. Li, M. S. Toprak, B. Zhu and M. Muhammed, Adv. Mater., 2010, 22, 1640–1644.
9. S. Sanna, V. Esposito, D. Pergolesi, A. Orsini, A. Tebano, S. Licoccia, G. Balestrino and E. Traversa, Adv. Funct. Mater., 2009, 19, 1713–1719.
10. X. Guo, Scr. Mater., 2011, 65, 96–101.
11. M. G. Campbell, D. C. Powers, J. Raynaud, M. J. Graham, P. Xie, E. Lee and T. Ritter, Nat. Chem., 2011, 3, 949–953.
12. I. Kosacki and H. U. Anderson, Ionics, 2000, 6, 294–311.
13. T. Suzuki, I. Kosacki and H. U. Anderson, Solid State Ionics, 2002, 151, 111–121.
14. L. Yao, W. Liu, G. Ou, H. Nishijima and W. Pan, J. Mater. Chem. A, 2015, 3, 10795–10800.
15. L. Yao, G. Ou, H. Nishijima and W. Pan, Phys. Chem. Chem. Phys., 2015, 17, 23034–23040.
16. P. Arunkumar, R. Ramaseshan, S. Dash, J. Basu, T. R. Ravindran, S. Balakumar and K. Suresh Babu, RSC Adv., 2014, 4, 33338–33346.
17. M. V. F. Schlupp, B. Scherrer, H. Ma, J. G. Grolig, J. Martynczuk, M. Prestat and L. J. Gauckler, Phys. Status Solidi A, 2012, 209, 1414–1422.
18. W. Liu, B. Li, H. Liu and W. Pan, Electrochim. Acta, 2011, 56, 3334–3337.
19. J. Jiang, D. Clark, W. D. Shen and J. L. Hertz, Appl. Surf. Sci., 2014, 293, 191–195.
20. I. Kosacki, C. M. Rouleau, P. F. Becher, J. Bentley and D. H. Lowndes, Solid State Ionics, 2005, 176, 1319–1326.
21. G. Baure, R. M. Kasse, N. G. Rudawski and J. C. Nino, Phys. Chem. Chem. Phys., 2015, 17, 12259–12264.
22. J. S. Park, Y. B. Kim, J. An and F. B. Prinz, Solid State Commun., 2012, 152, 2169–2171.
23. A. M. Saranya, D. Pla, A. Morata, A. Cavallaro, J. Canales-Vázquez, J. A. Kilner, M. Burriel and A. Tarancón, Adv. Energy Mater., 2015, 5, 1500377.
24. J. Maier, Nat. Mater., 2005, 4, 805–815.
25. N. Schichtel, C. Korte, D. Hesse and J. Janek, Phys. Chem. Chem. Phys., 2009, 11, 3043–3048.
26. C. Korte, A. Peters, J. Janek, D. Hesse and N. Zakharov, Phys. Chem. Chem. Phys., 2008, 10, 4623–4635.
27. J. Garcia-Barriocanal, A. Rivera-Calzada, M. Varela, Z. Sefrioui, E. Iborra, C. Leon, S. J. Pennycook and J. Sanramaria, Science, 2008, 321, 676–680.
28. L. Yao, W. Liu, G. Ou, H. Nishijima and W. Pan, Electrochim. Acta, 2015, 158, 196–201.
29. A. Peters, C. Korte, D. Hesse, N. Zakharov and J. Janek, Solid State Ionics, 2007, 178, 67–76.
30. M. Hasunuma, H. Toyoda and H. Kaneko, Microelectron. Reliab., 1999, 39, 1631–1645.
31. N. Schichtel, C. Korte, D. Hesse, N. Zakharov, B. Butz, D. Gerthsen and J. Janek, Phys. Chem. Chem. Phys., 2010, 12, 14596–14608.
32. S. Schweiger, M. Kubicek, F. Messerschmitt, C. Murer and J. L. M. Rupp, ACS Nano, 2014, 8, 5032–5048.
33. H. Jimenez, E. Restrepo and A. Devia, Surf. Coat. Technol., 2006, 201, 1594–1601.
34. R. Frison, S. Heiroth, J. L. M. Rupp, K. Conder, E. J. Barthazy, E. Müller, M. Horisberger, M. Döbelie and L. J. Gauckler, Solid State Ionics, 2013, 232, 29–36.
35. N. Sonnenberg, A. S. Longo, M. J. Cima, P. Chang, K. G. Ressler, P. C. McIntyre and Y. P. Liu, J. Appl. Phys., 1993, 74, 1027–1034.
36. J. L. M. Rupp, Solid State Ionics, 2012, 207, 1–13.
37. J. Jiang, D. Clarek, W. D. Shen and J. L. Hertz, Appl. Surf. Sci., 2014, 293, 191–195.
38. M. G. Bellino, D. G. Lamas and N. E. Walsöe de Reca, Adv. Funct. Mater., 2006, 16, 107–113.
39. L. Cheng, W. Chen, M. Kunz, K. Persson, N. Tamura, G. Y. Chen and M. Doeff, ACS Appl. Mater. Interfaces, 2015, 7, 2073–2081.
40. J. An, J. S. Park, A. L. Koh, H. B. Lee, H. J. Jung, J. Schoonman, R. Sinclair, T. M. Gur and F. B. Prinz, Sci. Rep., 2013, 3, 2680.
41. C. Rongeat, M. A. Reddy, R. Witter and M. Fichtner, ACS Appl. Mater. Interfaces, 2014, 6, 2103–2110.
42. C. Ma, K. Chen, C. D. Liang, C. W. Nan, R. Ishikawa, K. More and M. F. Chi, Energy Environ. Sci., 2014, 7, 1638–1642.
43. X. Guo and J. Maier, J. Electrochem. Soc., 2001, 148, E121–E126.
44. J. H. Shim, C. C. Chao, H. Huang and F. B. Prinz, Chem. Mater., 2007, 19, 3850–3854.
45. H. C. Huang, P. C. Su, S. K. Kwak, R. Pornprasertsuk and Y. J. Yoon, Fuel Cells, 2014, 14, 574–580.

---