Xiubing
Huang
,
Tae Ho
Shin
,
Jun
Zhou
and
John T. S.
Irvine
*
School of Chemistry, University of St Andrews, St Andrews, Fife, UK. E-mail: jtsi@st-andrews.ac.uk; Fax: +44 (0)1334463808; Tel: +44 (0)133463680
First published on 28th May 2015
Hierarchically nanoporous materials based on layered perovskite oxides La1.7Ca0.3NixCu1−xO4−δ (x = 0, 0.25, 0.50 or 0.75) have been synthesized by a facile citrate-modified evaporation-induced self-assembly (EISA) method. These La1.7Ca0.3NixCu1−xO4−δ oxides have been evaluated as potential cathodes for intermediate-temperature solid oxide fuel cells (IT-SOFCs) with Ni–YSZ cermet supported type cells. It was found that La1.7Ca0.3CuO4−δ cathode exhibits the maximum power density at high temperature (e.g., 1.5 W cm−2 at 850 °C), while La1.7Ca0.3Ni0.75Cu0.25O4−δ cathode shows the highest power density at intermediate temperature (e.g. 0.71 W cm−2 at 750 °C) using humidified H2 and air as the fuel and oxidant, respectively. The electrochemical performance of single cells with La1.7Ca0.3Ni0.75Cu0.25O4−δ cathode materials with different morphologies demonstrated better performance in the intermediate temperature range when using the cathode prepared by the citrate-modified EISA method, which has a bigger grain size, but with higher surface area and pore volumes.
Some perovskite-type oxides have mixed ionic-electronic conducting (MIEC) and have been widely investigated as the cathode materials for IT-SOFCs due to their good performance for ORR.19,20 Recently, MIECs based on layered perovskites with K2NiF4-type structure have received extensive interest as cathode materials for IT-SOFCs, owing to their relatively high oxygen ion diffusivity, rapid surface exchange property, and compatible thermal expansion coefficients with solid electrolytes.20,21 Most of the studies have been focused on the oxygen over-stoichiometry layered perovskites, especially La2NiO4 based materials, which have been reported to have good ionic conductivity because of the interstitial oxygen.22–25 However, the low electronic conductivity of La2NiO4 limits its further practical use as SOFC cathodes.26
In addition, as another archetypal material with K2NiF4 structure, layered La2CuO4 has previously been investigated for its electrical conduction properties, such as superconductivity at low temperature, and metallic/semi-conducting properties at room temperature.27,28 According to many previous reports, introducing alkaline-earths (e.g., Sr, Ca, Ba) at the La-site provides effective routes to further improve their electrical conductivity through formation of electron–holes, oxygen vacancies or interstitial oxygen.28–31 Partially substituting Cu with other transition metals (e.g., Ni) could also improve their ionic conductivity which would enhance the performance of these substituted La2CuO4 oxides as cathode materials of SOFC,32,33 however, the Sr-doped La2CuO4 exhibits the issues of high-temperature synthesis, as well as low stability with electrolytes.34 Although some encouraging results have been reported, good evidence regarding SOFC cathode with high performance has not yet been available for this system.33 For the purpose of this study, we have chosen these methods: doping Ca on the La-site and Ni substituting on the Cu-site, to enhance both the oxygen diffusion and their electrochemical performance as a cathode of IT-SOFCs.
Besides the composition, the microstructure of cathode also has influence on the electrochemical properties and performance of SOFCs.35–38 Nanoscaled structure can accelerate oxygen reduction kinetics by providing enlarged number of active sites for surface oxygen exchange, which has great effect on the ORR of mixed ionic and electronic conductors.36,37,39,40 Thus, in this study, we examined the applicability of the hierarchically nanostructured La1.7Ca0.3NixCu1−xO4−δ (0 ≤ x ≤ 0.75) cathode materials which were prepared by a facile citrate-modified evaporation-induced self-assembly (EISA) method or conventional citric acid method with the purpose of enhancing electrochemical performance under IT-SOFCs operating conditions, as well as investigating the effect of nanoparticle size on the performance. We systemically investigated the ratio of nickel to copper on the B-site of La1.7Ca0.3NixCu1−xO4−δ (0 ≤ x ≤ 0.75) and nanostructure morphologies with an effort to optimize their electrochemical properties. We report dramatically high power densities for the cell using nanostructured La2CuO4 series cathodes which are optimized their electrochemical properties with YSZ electrolyte.
For electrochemical performance and impedance measurements of the unit cell, Ni–YSZ|YSZ|La1.7Ca0.3NixCu1−xO4−δ, were performed in humidified H2 and air or O2 used as the fuels and oxidant, respectively, using an IM6 Electrochemical Workstation (ZAHNER, Germany) with frequency ranged from 0.1 Hz to 100 kHz with amplitude of 20 mV. Impedance plots were fitted using the Zview software.
Samples | a (Å) | c (Å) | Cell volume (Å3) | Grain size (nm) |
---|---|---|---|---|
La1.7Ca0.3CuO4−δ | 3.7875(3) | 13.1897(5) | 189.2124(5) | 42.5 |
La1.7Ca0.3Ni0.25Cu0.75O4−δ | 3.7916(6) | 13.0185(8) | 187.1647(2) | 74.3 |
La1.7Ca0.3Ni0.5Cu0.5O4−δ | 3.7958(9) | 12.8831(2) | 185.6306(9) | 78.7 |
La1.7Ca0.3Ni0.75Cu0.25O4−δ | 3.8109(3) | 12.7415(7) | 185.0488(5) | 73.2 |
The chemical compatibility between La1.7Ca0.3NixCu1−xO4−δ and YSZ after calcination at 900 °C for 5 h under static air was evaluated by XRD, as shown in Fig. 2. A lanthanum zirconate secondary phase (La2Zr2O7, JCPDS no. 17-0450) can be observed in the cases of low nickel content (x = 0, 0.25 and 0.5) (Fig. 2a–c) and the intensity of this secondary phase increases with the Ni content increment. However, in the case of La1.7Ca0.3Ni0.75Cu0.25O4−δ, only traces of lanthanum zirconate can be detected, showing that substituting La1.7Ca0.3CuO4−δ with 75% Ni on the B-site would prevent the formation of secondary phases. However, the detailed reasons for the different reactivity of these La1.7Ca0.3NixCu1−xO4−δ materials with YSZ are still unclear, which need further investigation. Whilst no degradation would be preferred, the degree of degradation is probably acceptable for potential La1.7Ca0.3NixCu1−xO4−δ electrodes at temperatures below 900 °C. Therefore, we have further analyzed the electrical and electrochemical properties as cathode materials of IT-SOFCs.
The SEM images of La1.7Ca0.3NixCu1−xO4−δ samples calcined at 750 °C and 900 °C for 2 h in air are shown in Fig. 3 and S2.† Although the primary particle size is about 50–400 nm, these samples possess hierarchically nanoporous structures after calcination at 750 °C for 2 h, as shown in Fig. 3a–d.41,42 This is maybe caused by the addition of P123 copolymer and citric acid during the synthesis process to favour the formation of hierarchically nanoporous layered perovskite oxides,43,44 which would be expected to positively increase the electrochemical performance. Even on further sintering the samples (Fig. 3e–h) at 900 °C for 2 h, their hierarchically nanoporous structures are still maintained although the particle sizes are a bit bigger than those of samples obtained at 750 °C. Their TEM images displayed in Fig. S3† also show that these particles are inter-connected with each other to form hierarchically nanoporous structures, in agreement with their SEM results.
The textural properties of La1.7Ca0.3NixCu1−xO4−δ samples after calcination at 750 °C and 900 °C for 2 h were evaluated by nitrogen adsorption/desorption. There are hysteresis loops between P/Po of 0.9 and 1.0, as presented in Fig. S4,† corresponding to the secondary mesoporous/macroporous structures produced by inter-aggregated particles.45,46 Their pore size distribution curves shown in Fig. S4† also demonstrate the co-existence of mesopores and macropores. Their BET specific surface areas (SSA), pore volume and average pore size calculated from N2 adsorption isothermal curves were summarized in Table S1.† The results indicate that the BET SSA, pore volume and average pore size after calcination at 900 °C for 2 h are smaller than those of samples obtained at 750 °C for 2 h. However the BET SSA and pore volumes are still high for samples obtained at 900 °C. Consequently, these samples obtained at 750 °C and 900 °C could possess hierarchically mesoporous/macroporous nanostructures with high BET SSA and pore volumes, which are attributed to the large amount of secondary pores among particles, in agreement with the SEM and TEM results.
In order to maintain the nanostructures, the hierarchically nanoporous cathode materials coated onto the YSZ electrolyte were fired at 900 °C for 2 h rather than being fired at higher temperature which would benefit the adherence between cathode and electrolyte if reaction could be avoided. The SEM images of the single cells using La1.7Ca0.3NixCu1−xO4−δ oxides shown in Fig. S5† indicate that the hierarchically nanoporous structures of La1.7Ca0.3NixCu1−xO4−δ cathode materials on the YSZ electrolyte were still remaining after sintering at 900 °C for 2 h. For all these four samples, the YSZ electrolyte is approximately 20 μm in thickness prepared by dip coating process and is adhered well to both cathode and anode layers without any delamination. The thickness of the La1.7Ca0.3NixCu1−xO4−δ cathode materials is around 15–20 μm with a highly hierarchically nanoporous structure that favours the gas diffusion. These cathode materials all show small particle size even after heat-treatment at 900 °C for 2 h. Furthermore, these particles connected with one another, forming the secondary mesopores/macropores, which would increase the tri-phase boundaries to enhance the ORR on the cathode surfaces.
The power generation property of the cells using the La1.7Ca0.3NixCu1−xO4−δ powder as cathode, humidified H2 as fuel and air as oxidant, respectively, with respect to the temperature, for the Ni–YSZ support and YSZ electrolyte systems is shown in Fig. 4. The maximum power densities (MPD) of the cells with the La1.7Ca0.3NixCu1−xO4−δ (x = 0, 0.25, 0.50, and 0.75) cathode materials were summarized in Fig. 5, in which their MPD at 850 °C were 1.5, 0.67, 0.74 and 0.89 W cm−2, respectively. Compared to previous data on electrochemical performance of cuprate cathodes fabricated at higher temperature,32 much improved performance is achieved by the hierarchically nanoporous La1.7Ca0.3CuO4−δ cathode. This suggests that low-temperature ceramic processing to fabricate nanostructured cathode layer offer a suitable potential on the cuprate cathode of SOFCs to effectively overcome its lower chemical compatibility in the long term. For La1.7Ca0.3NixCu1−xO4−δ (x = 0.25, 0.50, and 0.75) samples, their power density slightly increased with Ni content over the whole temperature range, which can be attributed to their increased oxygen ion mobility with Ni content. The highest MPD achieved by the La1.7Ca0.3CuO4−δ cathode at high temperature range can be attributed to its high electrical conductivity while doping La-site with Ca would result in enhanced oxygen ion mobility by the formation of oxygen vacancies and/or interstitial oxygen.28,31 As shown in Fig. 5, the MPD of La1.7Ca0.3Ni0.75Cu0.25O4−δ was slightly higher than that of La1.7Ca0.3CuO4−δ at lower temperatures (e.g., 700 and 750 °C). The MPD of La1.7Ca0.3Ni0.75Cu0.25O4−δ cathode achieved 0.51 and 0.71 W cm−2 at 700 and 750 °C, respectively.
![]() | ||
Fig. 4 Single cell performance of a cell with La1.7Ca0.3NixCu1−xO4−δ (x = 0, 0.25, 0.50, and 0.75) cathodes at various temperatures using humidified H2 and air as the fuel and oxidant, respectively. |
![]() | ||
Fig. 5 Maximum power density as a function of Ni content of La1.7Ca0.3NixCu1−xO4−δ (x = 0, 0.25, 0.50, and 0.75) cathodes using humidified H2 and air as the fuel and oxidant, respectively. |
In order to further study the cathodic polarization (Rp) during cell performance, impedance spectroscopy results were considered, taking the impedance intercept of high, mid and low frequency with the real axis of the Nyquist plot in order to determine the ohmic resistance (Rs) and non-ohmic resistance (Rp total = Rp1 + Rp2 + Rp3). Fig. 6 shows the impedance spectra of the cell, Ni–YSZ|YSZ|La1.7Ca0.3CuO4−δ, with simulated fitting curve under open circuit conditions at 700 °C and the whole impedance spectra for La1.7Ca0.3NixCu1−xO4−δ as thermal variation were represented in Fig. S6 and S7.† As shown in Fig. 6, two or three semicircles were roughly observed on impedance fitting plot and considering response frequency, semicircle at high (Rp1) and mid frequency (Rp2) could be assigned to the surface activation, oxygen and charge transfer while lower frequency (Rp3) should be associated with gas diffusion. Clearly the high frequency interception (Rs) is barely changed or slightly decreased by elevating temperature because ohmic resistance mainly arises from electrolyte.
![]() | ||
Fig. 6 Impedance spectra measured for a cell using Ni–YSZ|YSZ|La1.7Ca0.3CuO4−δ at 700 °C and simulated fitting by Zview software. |
Fig. 7a and b show Arrhenius plots of the thermal variation area-specific resistance (ASR) for ohmic and non-ohmic, respectively. As shown in Fig. 7c, activation energy for Rp1 and Rp2 were dramatically decreased with the increase of Ni content on the Cu-site which can be due to the increased oxygen over-stoichiometry, similar to that of La2Ni1−xCuxO4+δ.47 The slight improvement of electrochemical performance (e.g., MPD) with Ni–Cu couple might also reflect the changing of anisotropic oxygen transport mechanism (e.g., orthorhombic distortion, Jahn–Teller distortion of Cu2+, increased oxygen vacancies) in tetragonal layered perovskite structure with respect to Ni contents.47 For La1.7Ca0.3CuO4−δ, it also exhibits structural distortion and oxygen content change with the increasing of temperature, which may contribute to its excellent performance at high temperatures.28 On the other hand, the slight increase of activation energy for Rp3 with increasing Ni is because Rp3 would be related with microstructure factor and gas diffusion in electrode. Accordingly, nanoporous factor, as shown in Table S1,† might be in reasonable agreement with Rp3 polarization result. Considering the same anode and electrolyte for these cells, excepting La1.7Ca0.3CuO4−δ, it can be concluded that Ni–Cu couple of octahedral site in layered perovskite improves surface reactivity for oxygen transfer. In addition, the thermal expansion coefficient (TEC) values of La1.7Ca0.3NixCu1−xO4−δ were reduced with the increasing Ni contents in Cu site, suggesting La1.7Ca0.3Ni0.75Cu0.25O4−δ cathode shows relatively adaptable thermal compatibility with the typical SOFC electrolyte materials, as shown in Table 2. Therefore, considering a better stability, lower operating temperature, higher MPD at low temperature and more suitable TEC, hierarchically nanoporous La1.7Ca0.3Ni0.75Cu0.25O4−δ would be promising as a potential cathode for IT-SOFCs.
Samples | TEC50−850 |
---|---|
La1.7Ca0.3CuO4−δ | 15.4 |
La1.7Ca0.3Ni0.25Cu0.75O4−δ | 15.2 |
La1.7Ca0.3Ni0.5Cu0.5O4−δ | 14.9 |
La1.7Ca0.3Ni0.75Cu0.25O4−δ | 14.4 |
8YSZ (ZrO2 with 8 mol% Y2O3) | 10.7 |
LSGM (La0.9Sr0.1Gd0.8Mg0.2O3−δ) | 11.8 |
CGD (Ce0.9Gd0.1O2−δ) | 13.2 |
The single cells were prepared by using Ni–YSZ|YSZ anode supported electrolyte, in which the thickness of electrolyte is ca. 30 μm. The cells were tested using pure H2 and O2 as the fuel and oxidant, respectively. The electrochemical performance, OCV, and MPD of the cells with La1.7Ca0.3Ni0.75Cu0.25O4−δ cathodes are shown in Fig. 8. The results indicate that La1.7Ca0.3Ni0.75Cu0.25O4−δ cathode prepared by citrate-modified EISA method shows a bit higher MPD than that of La1.7Ca0.3Ni0.75Cu0.25O4−δ by conventional citric acid method during the whole temperature range, especially at low temperature (e.g., 750 °C), suggesting its higher ORR activity at low temperature based on the consideration of the same Ni–YSZ anode and YSZ electrolyte for these two cells. However, with the increasing of testing temperature, the difference between the two cathodes decreased and their MPD are almost the same at 850 °C, indicating high ORR activity in both cathode samples at 850 °C. However, during the whole testing temperature range, the OCV values are relatively higher when using La1.7Ca0.3Ni0.75Cu0.25O4−δ obtained by citrate-modified EISA method as cathode, as shown in Fig. 8c, which might be obviously caused by its bigger grain size, higher surface area, pore volume and surface activity for ORR.
![]() | ||
Fig. 8 Single cell performance (a and b), OCV and MPD (c) of a cell with La1.7Ca0.3Ni0.75Cu0.25O4−δ cathodes at various temperatures using pure H2 and O2 as the fuel and oxidant, respectively. |
The impedance spectra, IR loss and overpotential η are displayed in Fig. 9. The results in Fig. 9a and b show that the Rp is a bit smaller when using La1.7Ca0.3Ni0.75Cu0.25O4−δ from citrate-modified EISA method as cathode materials while there were no obvious differences in non-ohmic resistance. The smaller Rp can be attributed to the improved oxygen ion mobility due to its bigger grain size. However, the overpotential η results shown in Fig. 9c indicate that the total overpotential η in the single cell using La1.7Ca0.3Ni0.75Cu0.25O4−δ from citrate-modified EISA method as cathode was smaller than that from conventional citric acid method. It might be attributed to the more three-phase boundary active site, and better oxygen ion mobility and adhesion between electrolyte and the La1.7Ca0.3Ni0.75Cu0.25O4−δ cathode from citrate-modified EISA method, since it has a bit bigger grain size, higher surface area and pore volume. However, which is the predominating effect, grain size, surface area or pore volume, is still under investigation.
![]() | ||
Fig. 9 Impedance spectra (a and b), and IR loss and & overpotential η of La1.7Ca0.3Ni0.75Cu0.25O4−δ cathodes. |
Footnote |
† Electronic supplementary information (ESI) available: Rietveld refinements from XRD patterns, SEM and TEM images, N2 adsorption/desorption isothermal curves and the textural properties obtained from N2 adsorption/desorption curves, impedance spectra of single cell with different cathode materials and different operating temperatures, XRD patterns, SEM images and N2 adsorption/desorption curves for La1.7Ca0.3Ni0.75Cu0.25O4−δ from conventional citric acid method. See DOI: 10.1039/c5ta00983a |
This journal is © The Royal Society of Chemistry 2015 |