Suppression of Leakage Current in Proton-Conducting BaZr_0.8Y_0.2O_3−δ Electrolyte by Forming Hole-Blocking Layer

Y-doped BaZrO₃ (Ba(Zr_0.8Y_0.2)O_3−δ; BZY82) was used as a base proton-conducting electrolyte layer. Zr-, Y-, and Yb- doped BaCeO₃ (BaZr_0.1Ce_0.7Y_0.1Yb_0.1O_3−δ; BZCYYb) and La_28−xW_4+xO_54+δ with a La/W ratio of 6.7 (LWO67) were investigated as hole-blocking layers formed on the air side surface of the base electrolyte layer, respectively. Table I indicates the physicochemical parameters, σ_ion, σ^O_h, σ^O_e, of BZY82, BZCYYb, and LWO67, respectively, where σ_ion is conductivity of ions. σ^O_h and σ^O_e are conductivities of holes, ${{\rm{h}}}^{\cdot },$ and electrons, e⁻, respectively, at oxygen partial pressure of 1. These partial conductivities of ions, holes and electrons were taken from the literatures^28–30 including our previous report.

Defect chemistry-based reaction of holes formation was assumed to be as Eq. 1, where ${{\rm{V}}}_{{\rm{O}}}^{\cdot \cdot }$ and ${{\rm{O}}}_{{\rm{O}}}^{{\rm{X}}}$ are an oxygen vacancy and an oxygen atom existing at the oxygen site, respectively. The equilibrium constant of the reaction, K_h, can be expressed as Eq. 2. Under this equilibrium, the conductivity of holes, σ_h, is proportional to the 1/4 power of P(O₂), because the concentrations of the oxygen vacancy and the oxygen atom are known to be independent to P(O₂) within the operation condition of SOFCs.^28–30 P(O₂) dependences of the ionic transport numbers, t_i,, of BZY82, BZCYYb, and LWO67 were calculated by using the parameters listed in Table I and Eqs. 3, 4, and were plotted in Fig. 2.

Zoom In Zoom Out Reset image size

The ionic transport numbers of BZY82 and BZCYYb showed almost the same P(O₂) dependence except for P(O₂) range of lower than around 10^–20 atm, and decreased largely with increasing P(O₂) in the high P(O₂) range as a result of hole conduction. On the other hand, the ionic transport number of LWO67 showed unique P(O₂) dependence. LWO67 can keep high ionic transport number even at high P(O₂), but decreased with decreasing P(O₂) in the P(O₂) range of lower than around 10^–15 atm.

$$\begin{eqnarray}&&\displaystyle \frac{1}{2}{{\rm{O}}}_{2}+{{\rm{V}}}_{{\rm{O}}}^{\cdot \cdot }={{\rm{O}}}_{{\rm{O}}}^{{\rm{X}}}+2{{\rm{h}}}^{\cdot }\end{eqnarray} \tag{ 1 }$$

$$\begin{eqnarray}&&{{\rm{K}}}_{{\rm{h}}}=\displaystyle \frac{{\left[{{\rm{h}}}^{\cdot }\right]}^{2}\left[{{\rm{O}}}_{{\rm{O}}}^{{\rm{X}}}\right]}{P{\left({O}_{2}\right)}^{\tfrac{1}{2}}\left[{{\rm{V}}}_{{\rm{O}}}^{\cdot \cdot }\right]}\end{eqnarray} \tag{ 2 }$$

$$\begin{eqnarray}&&{t}_{{\rm{i}}}=\displaystyle \frac{{\sigma }_{{\rm{ion}}}}{{\sigma }_{{\rm{T}}}}\end{eqnarray} \tag{ 3 }$$

$$\begin{eqnarray}&&{\sigma }_{{\rm{T}}}={\sigma }_{{\rm{ion}}}+{{\sigma }^{{\rm{o}}}}_{{\rm{e}}}\cdot {{P}_{{{\rm{O}}}_{2}}}^{\tfrac{1}{4}}+{{\sigma }^{{\rm{o}}}}_{{\rm{h}}}\cdot {{P}_{{{\rm{O}}}_{2}}}^{\tfrac{1}{4}}\end{eqnarray} \tag{ 4 }$$

Leakage current should be caused by the existence of partial conductions of the holes and/or electrons in the electrolyte. Fuel will be consumed by both external- and leakage -currents. However, unlike the external current, the leakage current never generate any electrical power, so that the electrical efficiency will decrease with increasing a ratio of the leakage current to external current.²³ The leakage current density, i_leak, in a mono-layer electrolyte was estimated by using following equations as functions of electrical conductivity (σ_el) and ionic conductivity (σ_ion).^{7,24,25,28–30}

$$\begin{eqnarray}&&{i}_{{\rm{lon}}}=\displaystyle \frac{{\rm{R}}T}{4{\rm{F}}L}\displaystyle {\int }_{{{\mathrm{ln}}_{{P}_{{O}_{2}}}}_{,\mathrm{fuel}}}^{{{\mathrm{ln}}_{{P}_{{O}_{2}}}}_{,{\rm{air}}}}\displaystyle \frac{r{\sigma }_{{\rm{el}}}{\sigma }_{{\rm{ion}}}}{r{\sigma }_{{\rm{el}}}-{\sigma }_{{\rm{ion}}}}{\rm{d}}\,\mathrm{ln}\,P\left({{\rm{O}}}_{{\rm{2}}}\right)\end{eqnarray} \tag{ 3a }$$

$$\begin{eqnarray}&&{i}_{{\rm{ext}}}=\displaystyle \frac{RT}{4FL}\displaystyle {\int }_{{{\mathrm{ln}}_{{P}_{{O}_{2}}}}_{,\mathrm{fuel}}}^{{{\mathrm{ln}}_{{P}_{{O}_{2}}}}_{,{\rm{air}}}}\displaystyle \frac{\left(1+r\right){\sigma }_{{\rm{el}}}{\sigma }_{{\rm{ion}}}}{r{\sigma }_{{\rm{el}}}-{\sigma }_{{\rm{ion}}}}{\rm{d}}P\left({{\rm{O}}}_{{\rm{2}}}\right)\end{eqnarray} \tag{ 4a }$$

$$\begin{eqnarray}&&{i}_{{\rm{leak}}}={i}_{{\rm{ion}}}-{i}_{{\rm{ext}}}\end{eqnarray} \tag{ 5 }$$

$$\begin{eqnarray}&&r=\displaystyle \frac{{i}_{{\rm{ion}}}}{{i}_{el}}\end{eqnarray} \tag{ 6 }$$

where σ_el consists of conductivities of electrons and holes and has PO₂ dependence shown in Eq. 2, i_ext is the external current density, and i_el is a current density induced by conductions of holes and/or electrons. r is a ratio of ionic current density to electronic current density and equals to −1 in a condition of open-circuit.

The oxygen potential profile in the electrolyte at open-circuit was estimated by using following equation with σ_el and σ_ion:^7,24,25

$$\begin{eqnarray}&&\begin{array}{r}\displaystyle \frac{x}{L}=\displaystyle {\int }_{\mathrm{ln}\,{{\rm{P}}}_{{{\rm{O}}}_{{\rm{2}}}}({\rm{fuel}})}^{\mathrm{ln}\,{{\rm{P}}}_{{{\rm{O}}}_{{\rm{2}}}}(x)}\displaystyle \frac{{\sigma }_{{\rm{ion}}}{\sigma }_{{\rm{el}}}}{{\sigma }_{{\rm{ion}}}+{\sigma }_{{\rm{el}}}}d\,\mathrm{ln}\,P\left({{\rm{O}}}_{{\rm{2}}}\right)/\\ \,\displaystyle {\int }_{\mathrm{ln}\,{{\rm{P}}}_{{{\rm{O}}}_{{\rm{2}}}}({\rm{fuel}})}^{\mathrm{ln}\,{{\rm{P}}}_{{{\rm{O}}}_{{\rm{2}}}}({\rm{air}})}\displaystyle \frac{{\sigma }_{{\rm{ion}}}{\sigma }_{{\rm{el}}}}{{\sigma }_{{\rm{ion}}}+{\sigma }_{{\rm{el}}}}d\,\mathrm{ln}\,P\left({{\rm{O}}}_{{\rm{2}}}\right)\end{array}\end{eqnarray} \tag{ 7 }$$

where L is a thickness of the electrolyte, x is a distance from the surface of electrolyte at fuel side toward air side.

σ_ion in the typical proton-conducting solid oxides consists of conductivities of protons and oxide-ions. Electrochemical potential gradient of proton, which is driving force of proton diffusion, should be different from that of oxide-ion when the chemical potential of H₂O has gradient across the electrolyte in terms of ionic current. Therefore, Eqs. 3–7 do not hold when potential gradient of H₂O exists. Vøllestad et al.³¹ and Zhu et al.^32,33 reported the analytical models that were more versatile in mixed conducting ceramics and should be able to address the effect of the H₂O gradient. Considering the aim of this study to evaluate the effect of the hole-blocking layer, we used the condition with flat chemical potential of H₂O across the electrolyte for simplicity and enabled Eqs. 3–7. Therefore, in this study, H₂O partial pressures in both the cathode side and the anode side were assumed to be the same value of 10% to eliminate the H₂O gradient resulting in same electrochemical potential gradient to protons and oxide-ions in terms of ionic current. In this condition, the leakage-current density and the oxygen potential profile in the electrolyte will be given by Eqs. 1–5 and 7, respectively as we previously reported.⁷

For the evaluation of integrals, P(O₂) was defined as e^y, so that lnP(O₂) = y, where e is Napier number. For example, Eq. 3 can be expressed as Eq. 8 when r = −1 (i.e. open-circuit condition):

$$\begin{eqnarray}&&{i}_{{\rm{lon}}}=\displaystyle \frac{{\rm{R}}T}{4{\rm{F}}L}\displaystyle {\int }_{{y}_{{\rm{fuel}}}}^{{y}_{,{\rm{air}}}}\displaystyle \frac{{\sigma }_{{\rm{el}}}{\sigma }_{{\rm{ion}}}}{{\sigma }_{{\rm{el}}}+{\sigma }_{{\rm{ion}}}}{\rm{d}}y\end{eqnarray} \tag{ 8 }$$

where σ_ion is independent to y, ${\sigma }_{{\rm{el}}}={\sigma }_{{\rm{e}}}^{{\rm{O}}}\cdot {\left({{\rm{e}}}^{{\rm{y}}}\right)}^{-\displaystyle \frac{1}{4}}+{\sigma }_{{\rm{h}}}^{{\rm{O}}}\cdot {\left({{\rm{e}}}^{{\rm{y}}}\right)}^{\displaystyle \frac{1}{4}},$ y_air = lnP(O₂)_air = −1.666 for 10% humidified air, and y_fuel = lnP(O₂)_fuel = −59.37 for equilibrium composition at 873.15 K in 10% humidified hydrogen. The calculation of the integral was performed by using a mathematical software, Maple 17.

Figure 3 shows the schematic illustration of the cross-section of bi-layered electrolyte with constituent layers-a and -b. The leakage current in the layer- a depends on P(O₂) at the interface between the constituent layers. The dependence of the leakage current in the layer-a on P(O₂) at the interface differs from that in the layer- b. Under steady-state condition, the constituent layers-a and -b must have same leakage current. Therefore, the P(O₂) at the interface should be determined to give the layers-a and -b same leakage current, respectively.

Zoom In Zoom Out Reset image size

Figure 4a shows the potential profile of oxygen calculated by Eq. 7 at open-circuit condition (r = −1) for the mono-layer BZY82 electrolyte having 10% humidified hydrogen and 10% humidified air as a fuel and an oxidant, respectively, at a temperature of 600 °C. The profile showed a sharper slope toward the fuel side due to lower electronic conductivity under lower oxygen partial pressures. Figure 4b indicates the dependence of the internal leakage current density in the BZY82 mono-layer electrolyte with a thickness of 10 μm on P(O₂) in the oxidant at a temperature of 600 °C with fixed fuel of 10% humidified hydrogen. The leakage current density was calculated by Eqs. 3–6 at open-circuit condition (r = −1). When 10% humidified air was used as an oxidant, the leakage current density was found to be around 200 mA cm⁻² as indicated by the right end of Fig. 4b.

Zoom In Zoom Out Reset image size

BZY82 should be a promising electrolyte for p-SOFCs thanks to its high chemical stability and high proton-conductivity. However, the leakage current should be largely reduced to realize highly efficient power generation. As clearly suggested in Fig. 4b, the leakage current density can be significantly suppressed by reducing P(O₂) at air side. Therefore, the reduction of P(O₂) at the interface between the constituent layers- a and -b in bi-layered electrolyte will be effective for the decrease in the leakage current when having BZY82 layer at fuel side.

In order to reduce the P(O₂) at the interface of bi-layered electrolyte having BZY82 as base proton-conducting electrolyte layer, two types of electrolytes, BZCYYb and LWO67, were applied as hole-blocking layer formed on the air side surface of the base electrolyte layer.

Figure 5a indicates the potential profile of oxygen calculated by Eq. 7 for the bi-layered electrolyte consisting of 8 μm BZY82 and 2 μm BZCYYb with 10% humidified hydrogen and 10% humidified air as a fuel and an oxidant, respectively, at a temperature of 600 °C. In the case of the BZY82-BZCYYb bi-layered electrolyte, the potential profile was found to be almost the same as that in mono-layer BZY82 electrolyte due to the similar ionic transport number between BZY82 and BZCYYb as shown in Fig. 2, so that the significant effect on the reduction of the leakage current will not be expected. Figure 5b shows the potential profile of oxygen calculated by Eq. 7 for the bi-layered electrolyte consisting of 8 μm BZY82 and 2 μm LWO67 with 10% humidified hydrogen and 10% humidified air as a fuel and an oxidant, respectively, at a temperature of 600 °C. In the case of the BZY82-LWO67 bi-layered electrolyte, the potential profile largely changed from mono-layer BZY82 electrolyte, and P(O₂) at interface was found to be reduced largely.

Zoom In Zoom Out Reset image size

Figure 6a indicates the leakage current densities calculated by Eqs. 3–6 as a function of P(O₂) at the interface for each layer of the bi-layered electrolyte, BZY82 (8 μm) and BZCYYb (2 μm), respectively, at open-circuit condition (r = −1). 10% humidified hydrogen and 10% humidified air were assumed as a fuel and an oxidant, respectively. The intersection indicates the leakage current and P(O₂) at the interface at a steady-state condition. As predicted from the profile of chemical potential of oxygen as shown in Fig. 5a, the leakage current density was found to almost equal to that in the mono-layer BZY82 electrolyte. On the other hand, Fig. 6b indicates leakage current densities in each layer of the bi-layered electrolyte consisting of BZY82 (8 μm) and LWO67 (2 μm). As predicted from the profile of chemical potential of oxygen as shown in Fig. 5b, the leakage current density was found to be largely suppressed as compared with that in the mono-layer BZY82 electrolyte. The estimated leakage current density of the bi-layered electrolyte was found to be almost 200 times lower than that of mono-layer BZY82 electrolyte. These results suggest that LWO67 is promising material for hole-blocking layer combined with BZY82.

Zoom In Zoom Out Reset image size

In the condition of power generation, the higher ratio of the leakage current to external current will more deteriorate the electrical efficiency. Fortunately, the ratio will decrease with increasing the external current density, which leads to higher Faradaic efficiencies. However, the high value of Faradaic efficiency should be kept even at turn-down operation for highly efficient power generation. LWO67 hole-blocking layer should enable high Faradaic efficiencies at all range of external current densities. In addition, the hole-locking effect will be more important in the condition of steam electrolysis operation.

Figure 7a shows the chemical potential profiles of oxygen in the bi-layered electrolyte having various thickness ratio of BZY82 to LWO67 at a temperature of 600 °C. 10% humidified hydrogen and 10% humidified air were assumed as a fuel and an oxidant, respectively. P(O₂) at the interface decreased with increasing the thickness ratio of LWO67. Figure 7b shows the leakage current density and area specific resistance (ASR) of 10 μm bi-layered electrolyte as a function of thickness of BZY82 layer. By increasing the thickness ratio of LWO67, leakage current density could be reduced but ASR largely increased. However, 8 um BZY82 with 2 um LWO67 showed a compatibility with enough low leakage current density and ASR.

Zoom In Zoom Out Reset image size

To enable our findings, electrochemical tests with the bi-layer electrolyte with LWO67 hole-blocking layer are very important. Therefore, we are now developing the bi-layer electrolyte. At this stage, there are some technical issues to be solved to obtain LWO67 thin layer having good compatibility with a base electrolyte in terms of chemical stabilities.³⁴