Suppressing cation segregation on lanthanum-based perovskite oxides to enhance the stability of solid oxide fuel cell cathodes

Abstract

The slow rate of oxygen reduction reaction (ORR) at the cathode side has been considered as a scientific challenge to improve the overall performance of solid oxide fuel cell (SOFC). Unfortunately, dopant cation enrichment on the surface of perovskite-type cathode materials often leads directly to the reduction in activity and stability of ORR. For this reason, we quantitatively assessed the main driving force of dopant segregation on LaBO₃(001) surfaces (B = Cr_0.50Mn_0.50, Mn, Fe, Co_0.25Fe_0.75, Co, and Ni) using density functional theory (DFT) calculations. Based on our findings, the minimization of elastic energy, which is closely related to the size of both A-site and B-site cations, plays an important role on the A-site dopant segregation. The degree of dopant segregation can be controlled by the proper choice of cations contained in LaBO₃-type perovskite oxides. We therefore suggest a valuable principle that can be applied to design high performance cathode materials by suppressing the dopant cation enrichment at the surface.

---

1. Introduction

Solid oxide fuel cells (SOFCs) have attracted wide interest for efficient and clean electric power generation. In spite of many advantages (i.e. high efficiency, low emission, fuel flexibility, and so on), there are still several obstacles that SOFCs should overcome.^1,2 In particular, the development of cathode materials with fast and stable oxygen reduction reaction (ORR) kinetics has been a significant challenge to improve the overall SOFC performance.^3,4 In order to design highly active cathodes, it is necessary to understand the surface chemistry closely related to the rate of ORR.^5–8 However, cation segregation phenomena have significantly influenced the surface chemistry via the changes in both cation concentrations and chemical phases. For example, the degree of Sr enrichment on La_0.7Sr_0.3MnO₃ (LSM) surface can be controlled by temperature and oxygen pressure.⁹ The segregated Sr often induces the formation of secondary phases (e.g., SrO, SrCO₃) on La_1−xSr_x-based perovskite surfaces,¹⁰ which can deteriorate the reactivity and stability of the cathode surfaces. Cai et al. reported that the electrochemical performance of La_0.6Sr_0.4CoO_3−δ (LSC) films is reduced due to the formation of a chemically heterogeneous phase on the surface, but recovered after chemical etching.¹¹ Chen et al. also demonstrated that a Sr excess layer, known to have a large band gap,^12,13 suppresses the oxygen exchange reaction by blocking the tunneling of electrons from the bulk phase on the SrTi_1−xFe_xO_3−δ (STF) surface.¹⁴ Based on these examples, it turns out that a comprehensive understanding of the segregation behavior plays an important role in designing new cathode materials.

Many researchers previously have analyzed the segregation behaviors of A-site cation dopants in SOFC cathode candidate materials such as LaSrMnO₃,^5–9 LaCaMnO₃,¹⁵ LaSrCoO₃,^8,11,16 LaSrCoFeO₃,¹⁷ and LaSrCrMnO₃ (ref. 18) perovskites. Nevertheless, each of them has focused on investigating the phenomena at the particular materials or conditions, and thus has not identified a universal driving force that can be generally applied to the various perovskite oxides. Therefore, the development of robust cathode materials with optimal surface cation chemistry has been delayed without the fundamental understating of the segregation phenomena.

In our earlier study,¹⁹ both elastic and electrostatic contributions were quantitatively assessed as the main driving forces of cation rearrangements on the surfaces of manganite perovskites using density functional theory (DFT) calculations, and were also proved by sequential surface-sensitive experiments. We successfully demonstrated that the elastic energy of dopants depends on the size difference between the host and dopant cations in the A-site, and therefore (R_dopant − R_host) can be a descriptor implying a main thermodynamic driving force under a specific electrostatic condition. However, it is not still a perfect descriptor for the elastic energy generated upon doping the A-site. The elastic energy may also be contributed by the properties of B-site cations because when a cation is doped in the A-site, it occupies the 12 coordinate cavities that are created by a BO₃ network.²⁰ Jung and Tuller reported that the surface segregation of lattice Sr cation is dependent upon the Fe concentration on the surface of SrTi_1−xFe_xO_3−δ (STF) because Fe atoms in the B-site may exert to the high compressive strain onto the Sr atoms in the A-site.²¹ Therefore, a new type of descriptor which can more reasonably capture the elastic energy originated from both A- and B-site cations must be devised to control the dopant segregation of (La,A)BO₃ perovskite surface.

In this paper, we aim to extend aforementioned results¹⁹ by quantitatively assessing how B-site cation as well as the A-site cation contributes to the surface segregation of A-site dopant cation based on DFT calculations. It is noteworthy that we only consider the case that A-site dopant cation segregates toward the surface as experimentally reported. To examine the effect of elastic interactions, a wide range of A-site cation (Ca, Sr, and Ba) doped LaBO₃ perovskites (B = Cr_0.50Mn_0.50, Mn, Fe, Co_0.25Fe_0.75, Co, and Ni) that have been commonly used as cathode candidates^3,18,22–25 are chosen. This paper is organized as follows. The segregation energy is firstly calculated as a quantitative indicator of dopant segregation. Then, we introduce a descriptor, which is change of free volume, to describe the effects of the A- or B-site cations on the elastic interactions of dopant cation. The descriptor can be used to elucidate the underlying principles of dopant segregation and predict the degree of surface segregation. Lastly, based on these results, we suggest possible ways to suppress the dopant segregation.

2. Computational methods

The Vienna Ab initio Simulation Package (VASP)^26–28 was used to perform density functional theory (DFT) calculations. All calculations employed the Perdew–Burke–Ernzerhof (PBE) functional based on the generalized gradient approximation (GGA).²⁹ The total energy calculations used a plane wave expansion with a cutoff of 400 eV and included spin-polarization to describe the magnetic properties. In our model system, Ca, Sr, and Ba (alkaline earth metals) with 2+ charge states were chosen as A-site dopant cations and Cr_0.50Mn_0.50, Mn, Fe, Co_0.25Fe_0.75, Co, and Ni (first row 3d-transition metals) with 3+ charge states were chosen as the B-site cations. This selection was aimed to systemically introduce the elastic energy caused by depending on the size of A- and B-site cations. Table 1 shows the Shannon's ionic radii³⁰ of cations where the charge state, coordination number, and spin state (only B-site cations) were considered.

Table 1 Shannon's ionic radii (Å)³⁰ of A- and B-site cations determined by charge state, coordination number and spin state (only B-site cations). The B-site cations were listed according to the atomic number. The weighted average ionic radii were used for Cr_0.50Mn_0.50 and Co₂₅Fe_0.75 (ref. 38 and 39)

A-Site	La	Ca	Sr	Ba
	1.360	1.340	1.440	1.610

---

B-Site	Cr0.50Mn0.50	Mn	Fe	Co0.25Fe0.75	Co	Ni
	0.630	0.645	0.645	0.620	0.545	0.560

---

To choose the proper ionic radii from the database of Shannon's ionic radii, it is required to identify the spin state of B-site cation. For that, we compared our calculated magnetic moments per respective B-site cations with the reported values.^31–37 For the mixed B-sites such as Cr_0.50Mn_0.50 and Co₂₅Fe_0.75, the weighted average of ionic radii were used.^38,39 To avoid the self-interaction errors in the standard DFT-GGA for strongly correlated electronic materials such as mid-to-late first low transition metal oxides,⁴⁰ the DFT+U approach⁴¹ was used with U_eff = 3.5 eV (Cr), 4.0 eV (Mn), 4.0 eV (Fe), 3.3 eV (Co), 6.4 eV (Ni).⁴²

Periodic asymmetric (stoichiometric) slab models were employed to assess the energetics of dopant segregation with a vacuum of 15 Å in the direction of the surface normal. Dipole corrections⁴³ were considered in z-direction to avoid the error that is derived in non-zero dipole models such as the asymmetric slabs. Since the (001) surface in perovskite-type oxides is the most stable at high operating temperatures of SOFC,^44–46 the slab model was constructed by truncating the bulk perovskite structure along the (001) plane. Under the SOFC operating conditions, AO-terminated surfaces have been mainly observed in various perovskite oxides.^47–49 We thus focused only on the AO-terminated surfaces to describe a segregated dopant cation on the top surface layer among the two types of surface terminations in the (001) surfaces of perovskites (AO and BO₂). A (2 × 2) surface unit cell with six atomic layers was selected while one bottom layer is fixed as in their bulk positions. Compared to the previous reports,^{6,7,19,50–53} it is thought that our choices of AO-terminated surfaces and slab layer thickness are appropriate to analyze the trends in dopant segregation thermodynamics. All calculations for slab model used a 4 × 4 × 1 Monkhorst–Pack⁵⁴ grid k-point mesh.

To calculate the change of free volume (ΔV^f) upon the doping, a 2 × 2 × 2 supercell containing 40 total atoms were used with a Monkhorst–Pack grid⁵⁴ of 4 × 4 × 4 k-points. In these calculations, we used (pseudo)cubic structures that are exhibited at high SOFC operating temperature,^55,56 which have also been applied in previous DFT studies.^57–59 Two La atoms were substituted with two dopant atoms to calculate the lattice constant (or free volume) of the doped perovskite that was clearly altered compared to that of the corresponding perfect perovskite (Fig. S1 of ESI†).

3. Results and discussion

3.1. Surface segregation energy of A-site dopant cation on LaBO₃(001)

To quantify the segregation tendency of A-site dopants, we calculated surface segregation energy, that is the energy required for a dopant cation to segregate toward the surface as below;

E_segr. = E_surf − E_bulk, (1)

where E_segr. is the surface segregation energy per a dopant cation, E_bulk and E_surf are the total energies of a slab with a dopant cation replaced by a host cation in the bottom and top AO-layer, respectively. Fig. 1 shows our employed slab models to calculate the total energies of E_bulk and E_surf. By our definition, a larger negative segregation energy implies that the dopant cation more strongly segregates toward the surface. As mentioned in Section 1, in order to identify the effect of B metals on the segregation of an A-site dopant, we considered the segregation energies of Ca, Sr, and Ba on the surface of LaBO₃ perovskite oxides (B = Cr_0.50Mn_0.50, Mn, Fe, Co_0.25Fe_0.75, Co, and Ni).

Fig. 1 Schematic illustration of the doped LaBO₃(001) slab models to calculate the segregation energy of a A-site cation dopant. (a) and (b) show slab models before and after the segregation of a A-site cation dopant, respectively. A-site host cations (La) are light blue spheres, A-site dopant cations (Ca, Sr, and Ba) are light green spheres, B-site transition metal (TM) cations are (Cr_0.50Mn_0.50, Mn, Fe, Co_0.25Fe_0.75, Co, and Ni) purple spheres, and oxygen anions (O) are red spheres, respectively.

Our DFT results for E_segr. on LaBO₃(001) are summarized in Fig. 2. Our DFT values seem to be somewhat overestimated compared to the reported experimental data of Sr segregation enthalpy in the La_1−xSr_xMnO₃ thin film.⁶⁰ This inconsistency, which has also been observed in preceding DFT papers,^{6,7,19,50,52,53} may be attributed to the deviations in the reaction conditions (temperature, gas pressure), surface terminations, and dopant cation concentrations from the theoretical models. Nevertheless, the segregation energies of three dopant cations (Ca, Sr, and Ba) on LaMnO₃(001) still show a similar tendency as shown in previous DFT studies^19,51–53 even though there are several differences in the computational details such as the employed slab models and the calculating methods of the segregation energy. Here, we therefore focus on the tendency rather than the absolute values of the segregation energies.

Fig. 2 A trend in E_segr. of Ca, Sr, and Ba dopant cations on LaBO₃(001) surfaces (B = Cr_0.50Mn_0.50, Mn, Fe, Co_0.25Fe_0.75, Co, and Ni). The x-axis is configured in increasing the order of atomic number of the first row 3d-transition metals. The solid lines serve as a guide for the eye.

Fig. 2 is plotted with regard to the increasing the atomic number of 3d-transition metals in the same period, where we found two interesting observations. First, the larger dopant cation segregates more strongly toward the surface. Second, a dopant cation tends to segregate more as larger transition metals occupy the B-site (Table 1). These trends indicate that the surface segregation of dopant cations significantly depends on the sizes of the A-site dopant and B-site transition metal cations.

3.2. Elastic energy contribution to the dopant segregation: change of free volume (ΔV^f)

As mentioned above, we previously demonstrated that two key factors governing the dopant segregation on the surface of manganite perovskites are the elastic and electrostatic interactions.¹⁹ Elastic energy was induced by a size mismatch between the dopant and host cation, whereas electrostatic energy was generated by the charge interactions between a dopant and the other ions. Since A-site host cation (La), A-site dopant cation (Ca, Sr, and Ba), B-site cation (Cr_0.50Mn_0.50, Mn, Fe, Co_0.25Fe_0.75, Co, and Ni), and atomic oxygen anion (O) have identical charge state, respectively, in LaBO₃ perovskite system, the electrostatic contributions to dopant segregation can be considered constant. Indeed, our previous DFT results¹⁹ showed that the electrostatic energies in Ca, Sr, and Ba-doped LaMnO₃ and SmMnO₃ can be analytically calculated from same formal charges at each lattice site. Considering that electrostatic energy has nearly constant value in our (La,A)BO₃ system, it is reasonable to focus only on the effect of elastic interactions to elucidate the underlying mechanism of the dopant segregation in this study.

To explain the contribution of the elastic energy introduced by the A- and B-site cations in various lanthanum-based perovskites, we introduced the change of free volume (ΔV^f). The free volume implies the unit cell volume that excludes the total volume of the constituent ions.^61,62 Upon doping the A-site dopant cations (Ca, Sr, and Ba), the free volumes are changed depending on the sizes of A- or B-site cations. Therefore, ΔV^f can be calculated by

where V^f_{perfect (or  doped)} is the free volume of the perfect (or doped) bulk perovskite, a_{perfect (or  doped)} is the lattice constant of the perfect (or doped) perovskite, and r_{host (or  dopant)} is the Shannon's radius of the host (or dopant) cation. The lattice constants we used are summarized in Table S1 of ESI.† A more negative change in the free volume indicates that a dopant cation in the bulk phase is under the less stable condition due to the surrounding atoms, thereby generating a higher elastic energy of a dopant cation. Fig. 3 shows a linear relation between E_segr. and ΔV^f in LaBO₃ perovskites (B = Cr_0.50Mn_0.50, Mn, Fe, Co_0.25Fe_0.75, Co, and Ni). For all cases, the dopant cations segregate toward the surface to minimize the elastic energy. E_segr. decreases with decreasing ΔV^f due to the larger elastic energy driven by the substitution of dopant cations. It is noted that the segregation phenomena are dependent not only on the dopant cations doped on the A-site, but also B-site cations. Consequently, it is important to select the appropriate dopant and B-site cations in perovskite oxides for suppressing the dopant segregation. The full explanation of how to choose the cations is described in Section 3.5.

Fig. 3 Surface segregation energy calculated by eqn (1) as a function of the change of free volume calculated by eqn (2).

3.3. Contribution of additional factors to the dopant segregation

As shown in Fig. 3, although the changes of free volume are approximately zero (almost no elastic energies) in several materials, segregation energies are still negative. This implies that there exist additional contributions to the segregation energies except the elastic contribution. One extra contribution comes from the electrostatic interaction as mentioned above. Based on our previous report,¹⁹ a dopant cation has a charge of −1 upon replacing a host cation of 3+, thereby interacting with the other ions in the perovskite oxides even if they have no additional charged defects. In addition, the decrease in bond strength upon doping the cation can contribute to the surface segregation. From the previous reports,^63–67 the difference in bond strength has been considered as one of driving forces for surface segregation in metal alloy system, demonstrating that an atom that weakly binds with the nearby atoms strongly segregates toward the surface. Battaile et al.⁶⁸ also reported that the difference in bond strengths between cation and anion is an origin for surface segregation of solute cations in oxide solutions. In LaSrBO₃ perovskite oxides, Deml et al.⁵⁹ performed DFT calculations to estimate the bond strengths of La–O and Sr–O from the enthalpy of oxide formation which implies the average metal–oxygen bond strength. They demonstrated that bond strength of Sr²⁺–O is weaker than that of La³⁺–O from the reduction in enthalpy of oxide formation upon doping Sr in the A-site. It is therefore thought that in our system, the dopant²⁺–O bonds are weaker than La³⁺–O bonds, which facilitates the dopant cation to segregate toward the surface. Fig. 4 shows the segregation energy as a function of dopant–O bond length relative to La–O bond length in the bulk perovskites. The measurement details of the bond lengths are described in S2 section of ESI.† We previously showed that Sr–O bond length is closely related to the Sr segregation on (La,Sr)MnO₃ surface.⁷ Likewise, in a particular (La,A)BO₃ system, a larger A-site dopant cation that has larger bond distance with lattice oxygen tends to strongly segregate toward the surface due to the weaker bond strength (Fig. 4). This implies that the order of segregation tendency (Ba > Sr > Ca) caused by elastic contribution is fortified by the effects of bond strength.

Fig. 4 Calculated dopant segregation energy as a function of the dopant–O bond length relative to La–O.

Nevertheless, elastic energy is a dominant governing factor to the dopant segregation than the bond strength. Only by dopant–O bond strength, we cannot explain the dependence of the dopant segregation on B-site cations (Fig. 4). For instance, the relative dopant–O bond length in Sr-doped La(Co or Ni)O₃ is similar to that in Ba-doped La(Cr_0.50Mn_0.50, Mn, Fe, or Co_0.25Fe_0.75)O₃ even though the segregation energy is significantly different. As a result, we conclude that the dopant segregation behavior is mainly determined by elastic contribution, and enforced by the bond strength of dopant–O.

3.4. Identification of useful descriptors to predict the dopant segregation behaviors

As mentioned above in Section 3.2, it is important to obtain the change of free volume that enables the simple estimation of the segregation energy. In eqn (2), the change of free volume can be decoupled into two terms: (1) size (volume) mismatch between the dopant and host cations on the A-site and (2) change in the unit cell volume [a_doped³ − a_perfect³]. The size mismatch between the A-site dopant and host cations (1) can be simply obtained from a database of Shannon's ionic radii, whereas the change in unit cell volume (2) can be calculated from the change in the lattice constant of perovskite before and after doping the A-site cation, which requires additional experimental measurements or computational calculations. For simple estimation of the change of free volume, and thus the segregation energy, therefore, the term of change in the unit cell needs to be replaced with the corresponding term that is easily accessible.

The change in unit cell volume term may be expressed in terms of the ionic radii of A- and B-site cations that are closely related to the lattice constant of perovskite oxides.⁶⁹ Here, since the host cation (La) is constant in all perovskite oxides we examine, the change in unit cell volume may rely on the ionic radii of the dopant and B-site cations. The unit cell volume of LaBO₃ perovskite is determined depending on the size of B-site cation, and it generally decreases with decreasing the size of B-site cation. If an A-site dopant cation, which is always larger than B-site cation, is substituted in the bulk LaBO₃, thus, the unit cell with smaller B-site cations may increase more to expand the narrow internal space. For the same reason, expansion of the unit cell with a particular B-site cation may strongly induced upon a larger cation doping at the A-site. As a result, the ratio of size mismatch between the A-site dopant and B-site cation [(r_dopant − r_B)/r_B] is linearly proportional to the change in unit cell volume, as shown in Fig. 5. Therefore, the change of free volume can be rewritten as

Fig. 5 DFT-calculated change in the unit cell volume versus the ratio of size mismatch between the dopant and B-site cations from Shannon's ionic radii.³⁰

Furthermore, based on the linear correlation between E_segr. and ΔV^f in Fig. 3, the empirical equation of segregation energy is also obtained by linear fitting with two terms of eqn (3), and given as

From eqn (4), the dopant enrichment at the surface can be quickly and easily predicted only with the information for the ionic radii of cations.

3.5. Suggested strategies to suppress the dopant segregation

Here, we discuss each contribution from the two terms of eqn (4) to the segregation energy. We firstly examine the correlation between the segregation energy and size mismatch on the A-site. As shown in Fig. 6a, the larger dopant cation (in order of Ba, Sr, and Ca) tends to more strongly segregate to the surface within the same B-site transition metal. According to previous reports which investigated the origin of the dopant segregation on the surfaces of manganite perovskites,^19,51,52 this trend arises from the minimization of elastic energy induced by size mismatch between the dopant and host cations on the A-site. In this study, we more extensively support the previous arguments on a wide range of LaBO₃ perovskite surfaces; a relatively small difference between the dopant and host cations can suppress the dopant segregation (see Table S2 of ESI†). Indeed, this result has been applied to improve the long-term stability of SOFC cathode materials. Yoo et al.⁷⁰ expected the suppression of surface segregation on the surface of NdBaCo₂O_5+δ (NBCO) double-perovskite by only partially replacing Ba with Ca at the A-site. Then, they presented NdBa_1−xCa_xCo₂O_5+δ (NBCaCO) as a superior SOFC cathode material with high chemical stability. This example shows that our strategy to suppress the dopant segregation is successful.

Fig. 6 Contribution of the two terms in the predictive equation (eqn (4)) to the dopant segregation. Surface segregation energy as a function of (a) size mismatch between the A-site dopant and host cations and (b) size mismatch between the A-site dopant and B-site cations. The solid lines of (a) are used as a guide.

More importantly, it is notable that the size of B-site cation also affects the segregation energy (Fig. 6b). When the host A-site cation is substituted with a particular dopant cation, the larger B-site cation leads to the higher compressive strain of dopant cations (a more negative change of free volume). Therefore, the dopant cation more strongly segregate with increasing the size of the B-site cation, as can be seen in Fig. 6b. We can conclude that using the small B-site transition metals alleviates the surface cation segregation. Unlike our computational model, the oxidation state of B-site cation often becomes closer to 2+ with the formation of oxygen vacancies at SOFC operating conditions.⁸ Since B²⁺ is larger than B³⁺ as listed in Table S3 of ESI,† in this case, dopant cation would more strongly segregate toward the surface due to the fact that the larger B-site cation tends to generate the stronger elastic energy, as we demonstrated in this study. Nevertheless, we expect that the changes of elastic energy depending on B-site cations in LaB²⁺O₃ would not be significantly different from that in LaB³⁺O₃ because the changes in the size of B²⁺ are similar to that of B³⁺ (see Table S3 of ESI†). Since the formation of B²⁺ accompanies the creation of oxygen vacancy, both elastic and electrostatic interactions would affect the dopant segregation at the same time. Although it is difficult to perfectly decouple the two contributions, we believe that creating B-site cation of 2+ state, instead of 3+, allows the dopant cation to more strongly segregate toward the surface by two reasons; stronger elastic interaction generated by larger B-site cation and stronger electrostatic interaction induced by surface oxygen vacancy created from the change in the oxidation state of B-site cation.

Based on our results, we suggest that among the elements we examined, Ca is the most suitable for the A-site dopant cation due to the smallest size mismatch with La host cation while Co or Ni may be a good choice for the B-site transition metal due to the smaller size. Combining these, the level of dopant segregation would be the lowest at the surface of Ca-doped La(Co or Ni)O₃ perovskite among the materials we considered here. By just doping with the proper sizes of A- or B-site cations in perovskite oxides, SOFC cathode with better stability can be achieved. Since our proposals did not consider the kinetics or other factors of segregation that will be relevant in specific operating conditions, but only the thermodynamic aspect of segregation phenomena, we should be appropriately cautious. Nevertheless, our results still provide a useful basis for selecting the dopant elements where these other aspects of the dopant segregation could fruitfully be pursued. In fact, we previously demonstrated that at high temperatures, the tendency of dopant segregation was fairly predicted by just thermodynamics on the surface of manganite perovskites.¹⁹ However, the other contributions should be additionally considered to fully understand the dopant segregation phenomena, which will be a topic of our next study.

4. Conclusions

Cation segregation on perovskite oxide surfaces can often have an adverse effect on the SOFC cathode, especially the ORR. Thus, we quantitatively assessed the underlying mechanism of surface segregation to control the dopant enrichment at the Ca, Sr, and Ba doped-LaBO₃(001) surfaces using DFT calculations, where B = Cr_0.50Mn_0.50, Mn, Fe, Co_0.25Fe_0.75, Co, and Ni. To demonstrate the thermodynamic driving force, we introduced the change of free volume (ΔV^f). The minimization of elastic energy caused by the size of A- and B-site cations was determined to be a dominant factor governing the segregation energetics. This result, which expands our recent research,¹⁹ involves a significant finding in that the B-site cation also drives the surface segregation of dopants. The role of the change of free volume was strengthened by converting the lattice constant term into the ionic radius term, such as is done in eqn (3). By this alteration, the change of free volume (in other words, segregation energy) could easily be predicted from the ionic size of the cations. The useful principles were established for the development of new LaBO₃-type SOFC cathodes hindering dopant segregation as follows: select (1) a dopant A-site cation that shows smaller size mismatch with the host A-site cation and (2) a smaller size of B-site cation. These principles are helpful for reducing the elastic energy of the dopant cations. We believe that our findings will provide good insight in devising advanced SOFC cathodes.

Nevertheless, to develop new cathode materials with high stability, it is obvious that other factors should be also considered. Very recently, Tsvetkov et al. reported that reducing the concentration of surface oxygen vacancies in La_0.8Sr_0.2CoO₃ by surface additive cations can suppress Sr segregation by weakening the electrostatic attractions between Sr and surface oxygen vacancies.⁷¹ This result clearly implies that the electrostatic interaction also plays an important role to dopant segregation. However, it is still a big challenge to find the cathode materials to simultaneously reduce both elastic and electrostatic interactions. In the aspect of elastic interaction, in this study, selecting a dopant A-site cation that exhibits smaller size mismatch with the host A-site cation (Ca) and a smaller size of B-site cation (Co or Ni) was proposed to hinder the dopant segregation. However, (surface) oxygen vacancies which may accelerate the dopant segregation are easily formed in La(Co or Ni)O₃ among the candidate materials chosen in this study.⁷² It is therefore thought that there exists the opposite trend of contributions between elastic and electrostatic interactions to the dopant segregation, which brings us the fact that there will be an optimal point. Elucidating this more in detail will be our ultimate goal to design new cathode materials with high chemical and electrochemical stability.

Acknowledgements

The authors acknowledge support from the Global Frontier R&D Program on Center for Multiscale Energy System through the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT & Future Planning (NRF-2014M3A6A7074784 and NRF-2014M3A6A7074785) and the KIST Institutional Program (2E26081-16-053), and the supercomputing resources, including technical support from Supercomputing Center/Korea Institute of Science and Technology Information (KSC-2015-C3-045).

References

1. E. D. Wachsman and S. C. Singhal, Am. Ceram. Soc. Bull., 2010, 89, 22–32.
2. E. D. Wachsman and K. T. Lee, Science, 2011, 334, 935–939.
3. S. B. Adler, Chem. Rev., 2004, 104, 4791–4844.
4. Z. Shao and S. M. Haile, Nature, 2004, 431, 170–173.
5. K. Katsiev, B. Yildiz, K. R. Balasubramaniam and P. A. Salvador, Appl. Phys. Lett., 2009, 95, 092106.
6. H. Jalili, J. W. Han, Y. Kuru, Z. Cai and B. Yildiz, ECS Trans., 2011, 35, 2097–2104.
7. H. Jalili, J. W. Han, Y. Kuru, Z. Cai and B. Yildiz, J. Phys. Chem. Lett., 2011, 2, 801–807.
8. Z. Cai, Y. Kuru, J. W. Han, Y. Chen and B. Yildiz, J. Am. Chem. Soc., 2011, 133, 17696–17704.
9. T. T. Fister, D. D. Fong, J. A. Eastman, P. M. Baldo, M. J. Highland, P. H. Fuoss, K. R. Balasubramaniam, J. C. Meador and P. A. Salvador, Appl. Phys. Lett., 2008, 93, 151904.
10. P. A. W. van der Heide, Surf. Interface Anal., 2002, 33, 414–425.
11. Z. Cai, M. Kubicek, J. J. Fleig and B. Yildiz, Chem. Mater., 2012, 24, 1116–1127.
12. G. P. Summers, Phys. Rev. B: Condens. Matter Mater. Phys., 1979, 20, 5275–5279.
13. M. Abbate, H. Ascolani, F. Prado and A. Caneiro, Solid State Commun., 2004, 129, 113–116.
14. Y. Chen, W. Jung, Z. Cai, J. J. Kim, H. L. Tuller and B. Yildiz, Energy Environ. Sci., 2012, 5, 7979–7988.
15. S. Estradé, J. M. Rebled, J. Arbiol, F. Peiró, I. C. Infante, G. Herranz, F. Sánchez, J. Fontcuberta, R. Córdoba, B. G. Mendis and A. L. Bleloch, Appl. Phys. Lett., 2009, 95, 072507.
16. J. P. Kemp, D. J. Beal and P. A. Cox, J. Solid State Chem., 1990, 86, 50–58.
17. S. P. Simner, M. D. Anderson, M. H. Engelhard and J. W. Stevenson, Electrochem. Solid-State Lett., 2006, 9, A478–A481.
18. A.-K. Huber, M. Falk, M. Rohnke, B. Luerssen, L. Gregoratti, M. Amati and J. Janek, Phys. Chem. Chem. Phys., 2012, 14, 751–758.
19. W. Lee, J. W. Han, Y. Chen, Z. Cai and B. Yildiz, J. Am. Chem. Soc., 2013, 135, 7909–7925.
20. O. Müller and R. Roy, The Major Ternary Structural Families, Springer-Verlag, USA, New York, 1974.
21. W. Jung and H. L. Tuller, Energy Environ. Sci., 2012, 5, 5370.
22. C. Sun, R. Hui and J. Roller, J. Solid State Electrochem., 2009, 14, 1125–1144.
23. A. J. Jacobson, Chem. Mater., 2010, 22, 660–674.
24. Y. Choi, M. C. Lin and M. Liu, J. Power Sources, 2010, 195, 1441–1445.
25. A. M. Ritzmann, A. B. Muñoz-García, M. Pavone, J. A. Keith and E. A. Carter, Chem. Mater., 2013, 25, 3011–3019.
26. G. Kresse and J. Furthmüller, Phys. Rev. B: Condens. Matter Mater. Phys., 1996, 54, 11169–11186.
27. G. Kresse and J. Furthmüller, Comput. Mater. Sci., 1996, 6, 15–50.
28. D. S. Sholl and J. A. Steckel, Density Functional Theory: A Practical Introduction, John Wiley & Sons, Inc, USA, New Jersey, 2009.
29. J. P. Perdew and K. Burke, Phys. Rev. Lett., 1996, 77, 3865–3868.
30. R. D. Shannon, Acta Crystallogr., Sect. A: Cryst. Phys., Diffr., Theor. Gen. Crystallogr., 1976, 32, 751–767.
31. W. C. C. Koehler and E. O. O. Wollan, J. Phys. Chem. Solids, 1957, 2, 100–106.
32. T. Saitoh, T. Mizokawa, A. Fujimori, M. Abbate, Y. Takeda and M. Takano, Phys. Rev. B: Condens. Matter Mater. Phys., 1997, 55, 4257–4266.
33. P. M. Raccah and J. B. Goodenough, J. Appl. Phys., 1968, 39, 1209–1210.
34. M. A. Señarís-Rodríguez and J. B. Goodenough, J. Solid State Chem., 1995, 118, 323–336.
35. K. Sreedhar, J. M. Honig, M. Darwin, M. McElfresh, P. M. Shand, J. Xu, B. C. Crooker and J. Spalek, Phys. Rev. B: Condens. Matter Mater. Phys., 1992, 46, 6382–6386.
36. T. Mayeshiba and D. Morgan, Phys. Chem. Chem. Phys., 2015, 17, 2715–2721.
37. A. B. Muñoz-García, M. Pavone and E. A. Carter, Chem. Mater., 2011, 23, 4525–4536.
38. N. Trofimenko and H. Ullmann, Solid State Ionics, 1999, 118, 215–227.
39. A. B. Muñoz-García and M. Pavone, Chem. Mater., 2016, 28, 490–500.
40. E. A. Carter, Science, 2008, 321, 800–803.
41. S. L. Dudarev, G. A. Botton, S. Y. Savrasov, C. J. Humphreys and A. P. Sutton, Phys. Rev. B: Condens. Matter Mater. Phys., 1998, 57, 1505–1509.
42. L. Wang, T. Maxisch and G. Ceder, Phys. Rev. B: Condens. Matter Mater. Phys., 2006, 73, 195107.
43. G. Makov and M. Payne, Phys. Rev. B: Condens. Matter Mater. Phys., 1995, 51, 4014–4022.
44. Y. Choi, D. S. Mebane, M. C. Lin and M. Liu, Chem. Mater., 2007, 19, 1690–1699.
45. J. W. Han and B. Yildiz, J. Mater. Chem., 2011, 21, 18983–18990.
46. Y. A. Mastrikov, E. Heifets, E. A. Kotomin and J. Maier, Surf. Sci., 2009, 603, 326–335.
47. J. Druce, H. Téllez, M. Burriel, M. D. Sharp, L. J. Fawcett, S. N. Cook, D. S. McPhail, T. Ishihara, H. H. Brongersma and J. A. Kilner, Energy Environ. Sci., 2014, 7, 3593–3599.
48. J. Druce, T. Ishihara and J. Kilner, Solid State Ionics, 2014, 262, 893–896.
49. A. Staykov, H. Téllez, T. Akbay, J. Druce, T. Ishihara and J. Kilner, Chem. Mater., 2015, 27, 8273–8281.
50. S. Piskunov, E. Heifets, T. Jacob, E. A. Kotomin, D. E. Ellis and E. Spohr, Phys. Rev. B: Condens. Matter Mater. Phys., 2008, 78, 121406.
51. V. Sharma, M. K. Mahapatra, P. Singh and R. Ramprasad, J. Mater. Sci., 2015, 50, 3051–3056.
52. Y.-L. Lee and D. Morgan, Phys. Rev. B: Condens. Matter Mater. Phys., 2015, 91, 195430.
53. V. Sharma, M. K. Mahapatra, S. Krishnan, Z. Thatcher, B. D. Huey, P. Singh and R. Ramprasad, J. Mater. Chem. A, 2016, 4, 5605–5615.
54. H. J. Monkhorst and J. D. Pack, Phys. Rev. B: Condens. Matter Mater. Phys., 1976, 13, 5188–5192.
55. A. C. Momin, E. B. Mirza and M. D. Mathews, J. Mater. Sci. Lett., 1991, 10, 1246–1248.
56. S. P. Jiang, J. Mater. Sci., 2008, 43, 6799–6833.
57. M. Pavone, A. B. Muñoz-García, A. M. Ritzmann and E. A. Carter, J. Phys. Chem. C, 2014, 118, 13346–13356.
58. M. Pavone, A. M. Ritzmann and E. A. Carter, Energy Environ. Sci., 2011, 4, 4933–4937.
59. A. M. Deml, V. Stevanović, C. L. Muhich, C. B. Musgrave and R. O'Hayre, Energy Environ. Sci., 2014, 7, 1996–2004.
60. R. Herger, P. Willmott, C. Schlepütz, M. Björck, S. Pauli, D. Martoccia, B. Patterson, D. Kumah, R. Clarke, Y. Yacoby and M. Döbeli, Phys. Rev. B: Condens. Matter Mater. Phys., 2008, 77, 085401.
61. R. L. Cook and A. F. Sammells, Solid State Ionics, 1991, 45, 311–321.
62. A. F. Sammells, R. L. Cook, J. H. White, J. J. Osborne and R. C. MacDuff, Solid State Ionics, 1992, 52, 111–123.
63. L. E. Murr, Interfacial Phenomena in Metals and Alloys, Addison-Wesley Publishing Company, USA, MA, 1975.
64. F. F. Abraham, T. Nan-Hsiung and G. M. Pound, Surf. Sci., 1979, 83, 406–422.
65. P. Wynblatt and R. C. Ku, Surf. Sci., 1977, 65, 511–531.
66. B. Lezzar, O. Khalfallah, A. Larere, V. Paidar and O. Hardouin Duparc, Acta Mater., 2004, 52, 2809–2818.
67. J. W. Han, J. R. Kitchin and D. S. Sholl, J. Chem. Phys., 2009, 130, 124710.
68. C. C. Battaile, R. Najafabadi and D. J. Srolovitz, J. Am. Ceram. Soc., 1995, 78, 3195–3200.
69. O. Fukunaga and T. Fujita, J. Solid State Chem., 1973, 8, 331–338.
70. S. Yoo, A. Jun, Y.-W. Ju, D. Odkhuu, J. Hyodo, H. Y. Jeong, N. Park, J. Shin, T. Ishihara and G. Kim, Angew. Chem., 2014, 126, 13280–13283.
71. N. Tsvetkov, Q. Lu, L. Sun, E. J. Crumlin and B. Yildiz, Nat. Mater., 2016 DOI:10.1038/nmat4659.
72. Y.-L. Lee, J. Kleis, J. Rossmeisl and D. Morgan, Phys. Rev. B: Condens. Matter Mater. Phys., 2009, 80, 224101.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra15253h

---