AMnO3 (A = Sr, La, Ca, Y) Perovskite Oxides as Oxygen Reduction Electrocatalysts
- 862 Downloads
A series of perovskite-type manganites AMnO3 (A = Sr, La, Ca and Y) particles were investigated as electrocatalysts for the oxygen reduction reaction. AMnO3 materials were synthesized by means of an ionic-liquid method, yielding phase pure particles at different temperatures. Depending on the calcination temperature, particles with mean diameter between 20 and 150 nm were obtained. Bulk versus surface composition and structure are probed by X-ray photoelectron spectroscopy and extended X-ray absorption fine structure. Electrochemical studies were performed on composite carbon-oxide electrodes in alkaline environment. The electrocatalytic activity is discussed in terms of the effective Mn oxidation state, A:Mn particle surface ratio and the Mn–O distances.
KeywordsPerovskite Manganite oxide A-site Oxygen reduction reaction Catalytic activity
Platinum-based catalysts remain the benchmark materials for the ORR, in acidic and alkaline environments, but significant efforts have more recently been focused on earth-abundant and lower-cost materials; such as perovskite oxides of the general structure ABO3 [7, 8, 9]. Significant efforts have been devoted to establishing electronic descriptors for rationalising and predicting the activity of this class of materials [10, 11, 12, 13]. Our recent studies have established links between changes in the redox state of the B-site and their activity towards ORR which, in the case of Mn based oxides, can occur at potentials close to the reversible oxygen potential [14, 15, 16]. We proposed that this is the key rational behind the high activity of LaMnO3+δ observed by many groups [17, 18].
Several studies have also focused on the effect of the A-site in AMnO3 structures, including Pr1−xAx (A = Ca, Sr, Ba) , La1−xCax , La1−xSrx , La1−xTex  and A = Lanthanide element (La, Pr, Nd, Sm, Gb, Y, Dy and Yb) . However, rationalising trends for different synthetic approaches and experimental conditions is rather complex due to the contribution of other phenomena such as A-site surface segregation  and oxygen vacancies [23, 24]. In this work, we perform a systematic study of the ORR activity of AMnO3 particles as a function of the nature of the A-site (Ca, Sr, La and Y). A common synthetic strategy is implemented to generate phase pure particles which are characterised by X-ray diffraction (XRD), X-ray photoemission spectroscopy (XPS) and X-ray absorption spectroscopy (XAS). Electrochemical studies of the oxide particles mixed with mesoporous carbon and deposited at a rotating ring-disc electrode suggest that overall activity and selectivity towards the four-electron ORR reaction show complex dependence on structural parameters such as the ionic radius of the A cation and the average Mn–O distance.
2 Materials and Methods
2.1 Synthesis and Characterization of the Materials
The perovskite oxide particles were synthesized employing a route described in previous works [14, 15, 16, 25, 26]. The synthesis involved the dispersion of Mn(NO3)3 (0.5 mL, 0.1 M in H2O) and 0.5 mL of the corresponding A-site metal nitrate (0.5 mL, 0.1 M in H2O) in 1-ethyl-3-methylimidazolium acetate (1 mL) under stirring conditions. After heating at 95 °C for 4 h to ensure all water was removed, cellulose (100 mg) is added and the mixture was stirred for 20 min. The gel-like mixture was calcined for 4 h, except LaMnO3 and CaMnO3 which were calcined for 2 h. The calcination temperatures for each material were as follows: 850 °C (CaMnO3), 950 °C (LaMnO3, BaMnO3 and SrMnO3) and 1100 °C for YMnO3. Calcination conditions were set to ensure single-phase formation.
Powder X-ray diffraction (XRD) was carried out on a Bruker D8 Advance using Cu Kα radiation. Experiments were run between 10° and 90°, using a step size of 0.02°. Scanning electron microscopy (SEM) was carried out on a JEOL SEM 5600LV scanning electron microscope. Transmission electron microscopy (TEM) was carried out on a JEOL JEM-1400Plus instrument.
X-ray photoemission spectroscopy (XPS) measurements were performed in a custom-designed UHV system equipped with an EA 125 Omicron electron analyser, working at a base pressure of 10−10 mbar. Core level photoemission spectra were collected in normal emission at room temperature with a non-monochromatized Al Kα X-ray source (1486.7 eV), using 0.1 eV steps, 0.5 s collection time and 20 eV pass energy. The binding energies (BE) were referenced to the C 1s peak at 284.6 eV.
XAFS spectra were recorded in transmission mode at the Mn K-edge (6539 eV), on beamline B18 at Diamond Light Source (UK) operating with a ring energy of 3 GeV and at a current of 300 mA. The monochromator comprises Si(311) crystals operating in Quick EXAFS mode. Calibration of the monochromator was carried out using a Mn foil. The samples were prepared as pellets by mixing the ground sample with cellulose to form a homogeneous mixture. A total of three spectra were averaged for each sample. The spectra were aligned using the Mn foil response. The data was analysed using the Athena and Arthemis software , which implement the FEFF6 and IFEFFIT codes . Fits were carried out using a k range of 3.5–12 Å−1 and a R range of 1.0–4 Å with multiple k weightings of 1, 2 and 3. To perform the fittings, the coordination numbers were fixed to the crystallographic values; interatomic distances and Debye–Waller factors were fitted.
2.2 Electrochemical Measurements
A three-electrode cell was used to conduct the electrochemistry experiments, using a rotating ring-disk electrode (RRDE) fitted to an ALS rotation controller and connected to a CompactStat bipotentiostat (Ivium). The RRDE electrode consisted of a 4 mm glassy carbon disk surrounded by a Pt ring; a graphite rod and a Hg/HgO (in 0.1 M NaOH) were used as counter and reference electrodes, respectively. Measurements were recorded in 0.1 M KOH solution saturated with either purified Ar or O2. A two-step drop-casting process was used to prepare the working electrodes. First, an aqueous Vulcan/Nafion suspension was dropped onto the surface of the electrode, followed by an aqueous suspension of the metal oxide nanoparticles. The sample loading on the electrode surface was 250 µgOxide cm−2, 50 µgVulcan cm−2, and 50 µgNafion cm−2.
3 Results and Discussion
Figure 1a shows the XRD patterns of the as-grown AMnO3 (A = Y, Ca, La and Sr) oxide nanoparticles, revealing a high degree of phase purity based on their close match with the reported standards. YMnO3 was indexed to the hexagonal P63cm structure characterized by tilted layers of corner-linked MnO5 trigonal bipyramids, separated by layers of Y ; whereas SrMnO3 was indexed to the P63/mmc hexagonal space group. LaMnO3 was indexed to the rhombohedral (R-3c) phase, while CaMnO3 exhibited the orthorhombic (Pnma) phase.
Figure 1b displays a representative SEM image of CaMnO3 illustrating the microstructure of the material, while the TEM image in Fig. 1c shows the nanoscale dimensions of the particles (TEM images of the other oxides can be found in Fig. S1 of the Supporting Information). Particle size distributions (Fig. S2) revealed an increase in the mean particle size as the synthesis temperature increases, from 25.9 ± 4.3 nm for CaMnO3 to 166.1 ± 28.5 nm for YMnO3. Specific surface areas calculated from the mean diameters as measured by TEM and the theoretical densities are summarised in Table S1.
Figure 2 shows the XPS spectra for Mn 2p and the corresponding A site regions. The Mn 2p5/2 is observed at around 642.0 eV for all the samples. This broad peak contains contributions from Mn3+ (641.9 eV) and Mn4+ (642.2 eV) states , which are rather difficult to deconvolute. The Sr 3d line shows two components; the one at lower BE corresponds to the Sr in the perovskite lattice (132.0 eV), whereas the one at higher BE is attributed to the formation of SrO/Sr(OH)2 at the surface (133.8 eV) [31, 32]. Two components are also observed in the Y 3d region, one at 156.2 eV, associated with the formation of Y2O3, and the another at 158.5 eV related to Y in the lattice of YMnO3. La 3d line shows a double splitting due to the interaction between an electron from the oxygen valence band and the empty La 4f level. The La 3d5/2 BE is located at 834.4 eV, corresponding to La3+ compounds [30, 33]. In the Ca 2p line, the Ca2+ species of the perovskite lattice are represented by the component at 345.0 eV, while the Ca 2p3/2 component centered at 346.7 eV can be attributed to the formation of CaO in the surface due to the segregation of Ca [14, 34, 35].
The surface ratio of A and Mn sites, as estimated from the Mn 2p emission and the regions of the A cations, is summarized in Table S2. A slight A-site surface enrichment is observed in the case of La, Ca and Y, while near stoichiometric ratio is estimated in the case of Sr. In a previous study, we have shown that La has stronger tendency to segregate at the surface than Ca ions .
Figure 3a shows the normalized XANES spectra at the Mn K edge of YMnO3, CaMnO3, LaMnO3 and SrMnO3 samples; as well as MnO, Mn2O3 and MnO2 standards. A chemical shift of the main Mn K-edge (considered as the inflection point of the main rise) of 4 eV is observed for Mn2O3 with respect to MnO2, and for LaMnO3/YMnO3 with respect to CaMnO3/SrMnO3, as expected for Mn3+ and Mn4+ [36, 37]. The differences in absolute values between the binary and perovskite-like oxides may be related to geometrical effects, as it is widely acknowledged that the Mn K-edge position can be affected by local distortions of the MnO6 octahedra [36, 38, 39, 40].
The position of the pre-edge feature as a function of the oxidation state of the Mn is displayed in Fig. 3b. The mean Mn oxidation state was probed from the pre-edge energy position following the approach reported by Croft et al. , choosing the minimum point in the derivative spectrum. Figure 3c shows a linear relationship between the pre-edge position and the Mn oxidation state of the standards Mn3O4, Mn2O3 and MnO2 [41, 42]. The mean Mn oxidation state of the different AMnO3 oxides was estimated from this trend, providing values consistent with the composition of the A-site. The manganite oxides containing an alkali metal (2+) at the A-site presented a Mn oxidation state higher (Mn4+) than those containing La/Y (Mn3+). The slightly higher oxidation state of Mn in LaMnO3 could be a reflection of the oxygen vacancies in the structure or the surface segregated manganese sites forming single manganese oxide, and thus, the mean oxidation state would indicate a mixture of Mn2+/Mn3+/Mn4+ sites.
Figure 3d shows the phase-corrected Fourier transform (FT) of the EXAFS region for YMnO3, LaMnO3, CaMnO3 and SrMnO3. All of the FT spectra show two strong peaks below 4 Å. The first one at around 2 Å corresponds to the first coordination shell (Mn–O). It is interesting to notice how the amplitude of this peak is smaller in the case of YMnO3, for which Mn is only fivefold coordinated with an oxygen atom. The second peak, above 3 Å, is associated with the second shell with Mn–A, Mn–Mn and Mn–O contributions together with multiple scattering paths. The best-fit parameters of the analysis are summarised in Table S3. For a fixed Mn oxidation state, a slight increase of the Mn–O distance is observed as the radius of the A-site cation increases. The fit on k-space can be found in Fig. S3.
Figure 4 shows cyclic voltammograms (CVs) of the various AMnO3 nanostructures with a fix particle loading (250 μgOxide cm−2) in argon-saturated 0.1 M KOH at 0.010 V s−1. The CVs are characterised by complex responses associated with the redox properties of surface Mn sites. As the open circuit potential of all of the oxides is located at values more positives than the reduction responses, the initial oxidation state of the Mn site can be taken from the XANES analysis . LaMnO3 features two cathodic reduction peaks, located at 0.90 and 0.50 V, related to the reduction of Mn from an oxidation state + 3.2 (as determined in Fig. 3c) to 2+ [14, 15, 18]. The CV of CaMnO3 is characterised by a broad reduction peak, ascribed to the reduction of + 4 to + 2 [14, 43, 44]. A similar response is observed from SrMnO3, although current magnitude is significantly weaker. Surprisingly, YMnO3 shows no clear redox peaks in the voltammetric range investigated. The difference in current responses cannot be simply rationalised in terms of the change in Mn redox state and the mean particle size. For instance, the specific surface area (SSA) of CaMnO3 is about three times larger than SrMnO3 (see Table S1), while the Mn redox responses is an order of magnitude larger. The rationale for the contrasting behaviour of current responses associated with surface Mn sites remains under investigation.
Figure 5 displays the current measured at the disk (iDISK), containing a fix carbon-oxide composition, and the Pt ring (iRING) at 1600 rpm in O2-saturated 0.1 M KOH electrolyte. The responses measured for a solely Vulcan carbon electrode have been included as a comparison. It could be clearly seen that LaMnO3 and CaMnO3 exhibit an ORR onset potential significantly more positive than the Sr and Y manganite. Indeed, the onset potential of SrMnO3 and YMnO3 is close to the value observed in the absence of oxide, i.e. ORR at the Vulcan support (ca. 0.7 V) [15, 16]. It can also be observed that the diffusional limiting current is smaller for SrMnO3 and YMnO3, while a larger iRING is observed for the latter in the range of 0.2–0.6 V. which is linked to the trend observed in HO2− current value. Considering the collection coefficient of the RRDE electrode (N = 0.42), the effective number of transferred electrons (n) and the hydrogen peroxide yield (% OH2−) were calculated from the disk and ring currents and displayed in Fig. S5. The trend the OH2− yield is YMnO3 > SrMnO3 > CaMnO3 > LaMnO3. The effective number of electron ranges from effectively from around 2.6 (YMnO3) to 3.6 (LaMnO3) in the potential range between 0.2 and 0.6 V. The deviation from the four-electron limit is due to the ORR at the Vulcan support, which effectively goes through a two-electron reaction leading to the production of OH2−.
In order to better illustrate the effect of the A-site on the ORR pathway, we estimated the phenomenological electron transfer rate constant for the four-electron step (kdirect) employing the Damjanovic model [15, 46, 47, 48, 49]. Figure 6a, b contrast the value of kdirect (0.65 V) as a function of the A-site ionic radius and the mean Mn–O bond length obtained from the EXAFS analysis (Table S3). kdirect effectively represents the rate of conversion to OH−, based on the O2 flux to the surface (iDISK) and the rate of conversion to HO2− given by iRING. As this analysis considers mass transport fluxes, the geometric surface area is taken into account as opposed to the effective surface area. It could be observed that as the ionic radii increases, kdirect goes through a maximum value at A = La (1.93 Å). On the other hand, no clear trend is observed of kdirect as a function of the mean Mn–O bond distance. The plots also show the values corresponding to the carbon layer (in the absence of catalyst) as dotted line. The values show that Sr and Y based manganite show a modest improvement towards ORR in comparison to Ca and La manganite. Although aspects concerning the effective surface area can play a role in this analysis, the experimental data suggest that A-site ionic radii in the range of 1.15–1.20 Å promote an optimal electronic configuration of the Mn sites for the four-electron ORR.
Figure 6 also show the kinetically limit current density (jk) at 0.65 V, which is normalised by the specific surface area (SSA) of the oxide. As shown in Table S1, SSA values range significantly from 7.2 ± 1.8 (YMnO3) to 51.9 ± 1.8 (CaMnO3). Although jk exhibits a step increase in the case of LaMnO3, this value appears little dependent of the A-site ionic radius as shown in Fig. 6a. On the other hand, a slightly clearer pattern emerges when plotting this data as function of the Mn–O distance (Fig. 6b). A very weak dependence of jk is observed with increasing mean Mn–O distance until reaching the value associated with LaMnO3. The jk baseline value associated with Vulcan is also shown as dotted line, considering the specific surface area of the mesoporous carbon (218 m2 g−1 ). Overall, the analysis show that the overall activity (turnover rate) of the oxide catalysts is larger than the carbon support and significantly increases above a mean Mn–O distance threshold of 1.915 Å.
Manganite perovskite particles with the general formula AMnO3 (A = Y, Ca, La, Sr) were synthesised and assessed as electrocatalysts for the oxygen reduction reaction in alkaline solutions. The synthetic approach led to highly phase pure materials as characterised by XRD and EXFAS, although the different thermochemistry of the oxide phases required various calcination temperatures which led to a range of particle sizes. XPS studies showed that all oxide particles exhibit A-site rich surfaces, although the extent of segregation was weakly dependent of the A-site nature. The kinetics of the ORR as a function of the perovskite composition was quantitatively investigated using rotating ring-disk electrode in O2 saturated alkaline solutions. The analysis show that all the perovskite particles catalyse the ORR reaction in comparison to the mesoporous Vulcan support, although the effect of the A-site is rather complex to assess. Our data suggest that A-site with ionic radii in the range of 1.15–1.20 Å strongly promote the four-electron ORR process, while the overall activity increases as the mean Mn–O bond length increases above 1.915 Å. It is also interesting to point out that the Mn sites in the most selective catalysts towards the four-electron process (CaMnO3 and LaMnO3) are characterised by an octahedral coordination with equal Mn–O distances. Mn is coordinated to 5 oxygen sites with two different Mn–O bond lengths in YMnO3, while SrMnO3 has Mn sites coordinated to 6 oxygens with two different bond lengths. The electronic configuration of Mn sites in these particles yields different redox behaviour (Fig. 4) and, in agreement with our previous studies, the most active ORR catalysts (LaMnO3) show the most positive onset potential for the reduction of surface Mn sites [14, 15].
Although our main conclusions call for further studies correlating structure–activity of these materials, we have uncovered some fascinating trends linking selectivity and ORR with structural parameters such as A-site ionic radius and Mn–O bond length. Theoretical studies linking dioxygen bonding to Mn sites under the various configurations highlighted in these studies can shed more light into the complex activity trends observed of these materials.
V.C., A.E.R. and D.J.F. kindly thank the UK Catalysis Hub for resources and support provided via the membership of the UK Catalysis Hub Consortium (EPSRC Grants EP/K014706/1 and EP/K014714/1). TEM studies were carried out in the Chemistry Imaging Facility at University of Bristol with equipment funded by UoB and EPSRC (EP/K035746/1 and EP/M028216/1). The authors wish to acknowledge the Diamond Light Source for provision of beamtime (SP10306). All the data presented in this paper can be freely accessed from the Bristol’s Research Data Repository ( https://doi.org/10.5523/bris.8hwxpcse3lpz237ne8jd9crp4).
- 42.García J, Sánchez MC, Subías G, Blasco J (2001) J Phys: Condens Matter 13(14):3229Google Scholar
- 45.Bard AJ (1980) Electrochemical methods: fundamentals and applications/Allen J. Bard, Larry R. Faulkner. Wiley, New York. Accessed from http://nla.gov.au/nla.cat-vn2698966
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.