Letter pubs.acs.org/NanoLett
Study on the Catalytic Activity of Noble Metal Nanoparticles on Reduced Graphene Oxide for Oxygen Evolution Reactions in Lithium−Air Batteries Yo Sub Jeong,† Jin-Bum Park,† Hun-Gi Jung,†,‡,∇ Jooho Kim,† Xiangyi Luo,§ Jun Lu,§ Larry Curtiss,∥ Khalil Amine,§ Yang-Kook Sun,*,† Bruno Scrosati,*,⊥ and Yun Jung Lee*,† †
Department of Energy Engineering, Hanyang University, Seoul 133-791, Republic of Korea Department of Chemical Engineering, Hanyang University, Seoul 133-791, Republic of Korea § Chemical Science and Engineering Division, Argonne National Laboratory, Argonne, Illinois 60439, United States ∥ Material Sciences Division, Argonne National Laboratory, Argonne, Illinois 60439, United States ⊥ Italian Institute of Technology, 16163 Genoa, Italy ‡
S Supporting Information *
ABSTRACT: Among many challenges present in Li−air batteries, one of the main reasons of low efficiency is the high charge overpotential due to the slow oxygen evolution reaction (OER). Here, we present systematic evaluation of Pt, Pd, and Ru nanoparticles supported on rGO as OER electrocatalysts in Li−air cell cathodes with LiCF3SO3− tetra(ethylene glycol) dimethyl ether (TEGDME) salt-electrolyte system. All of the noble metals explored could lower the charge overpotentials, and among them, Ru-rGO hybrids exhibited the most stable cycling performance and the lowest charge overpotentials. Role of Ru nanoparticles in boosting oxidation kinetics of the discharge products were investigated. Apparent behavior of Ru nanoparticles was different from the conventional electrocatalysts that lower activation barrier through electron transfer, because the major contribution of Ru nanoparticles in lowering charge overpotential is to control the nature of the discharge products. Ru nanoparticles facilitated thin film-like or nanoparticulate Li2O2 formation during oxygen reduction reaction (ORR), which decomposes at lower potentials during charge, although the conventional role as electrocatalysts during OER cannot be ruled out. Pt-and Pd-rGO hybrids showed fluctuating potential profiles during the cycling. Although Pt- and PdrGO decomposed the electrolyte after electrochemical cycling, no electrolyte instability was observed with Ru-rGO hybrids. This study provides the possibility of screening selective electrocatalysts for Li−air cells while maintaining electrolyte stability. KEYWORDS: lithium−air batteries, catalysts, noble metals, electrolyte stability, catalytic mechanism
A
effectively lower the charge potential, but the decomposition of the electrolyte is the one that is catalyzed, not the Li2O2 oxidation, with carbonate and dimethyl ether (DME) electrolytes. Recently, evidence of catalyzed oxidation of Li2O2 by noble metals has been revealed with the Li2O2−carbon composite electrode system.9 Although there are a few reports on the use of noble metals as electrocatalysts in Li−air cells,8−18 systematic studies on the cell level under normal operation conditions, with the relatively stable and robust electrolyte-salt system such as LiCF3SO3−tetra(ethylene glycol) dimethyl ether (TEGDME), have been limited. In this study, we assessed Pt, Pd, and Ru noble metal nanomaterials supported on reduced graphene oxide (rGO) as OER electrocatalysts for Li−air batteries with a LiCF3SO3−
Li−air battery utilizes oxygen as an oxidizer in the air electrode, which differs from the conventional Li-ion configuration.1 Unlike the primary operation in fuel cells, the Li−air cell aims for use as secondary batteries, and the charging reaction on the positive electrode for this purpose is the oxygen evolution reaction (OER). Although the most attractive feature of Li−air batteries is that the active cathode material oxygen is not stored in the battery itself as in fuel cells,2 slow OER kinetics require proper electrocatalysts as in the water electrolysis cells.3,4 Noble metal nanoparticles are known to be the best materials in terms of catalytic activity in facilitating oxygen reduction reaction (ORR) in fuel cells5 and many other oxidation reactions.6 Noble metals have been also explored as OER electrocatalysts in Li−air cells. In Li−air cells, it is reported that ORR is hardly influenced by the presence of catalysts other than carbon or carbon itself catalyzes ORR sufficiently, whereas OER can be enhanced catalytically.7 In the initial study of low degree of discharge level (DOD),8 Pt could © XXXX American Chemical Society
Received: November 18, 2014 Revised: February 13, 2015
A
DOI: 10.1021/nl504425h Nano Lett. XXXX, XXX, XXX−XXX
Letter
Nano Letters
Figure 1. Discharge−charge voltage curves of lithium−air batteries using noble metal catalysts supported on rGO electrode. Results for 200 mA g−1, for 10 h. (a) 1st cycle of catalyst free and catalyst loaded rGO electrodes, (b) Pt-rGO hybrid, (c) Pd-rGO hybrid, and (d) Ru-rGO hybrid.
TEGDME electrolyte system. A two-dimensional graphene support was used because noble metals are known to interact much stronger with functionalized graphene surfaces such as rGO compared to the conventional carbon (Vulcan carbon, carbon black, etc.) due to the presence of surface functionalities and defects.19−21 Therefore, catalyst nanoparticles show better dispersion and less aggregation on rGO support. Large surface areas22 and inherent porous structures can also contribute to enhanced catalytic properties. Consistent with the previous reports, discharge potentials were relatively insensitive to the presence of noble metal nanoparticles, and charge overpotentials were substantially reduced with all of the noble metals employed in this study. Among the evaluated noble metals, Ru-rGO showed stable cycling performance and the lowest charge potentials. The mechanism of lowering charge overpotential has been investigated for Ru-rGO hybrids and the major role of Ru nanocatalysts was identified as changing the nature of discharge products to decompose at lower potentials. Pt- and Pd-rGO demonstrated fluctuating voltage profiles during cycling due to the catalyzed electrolyte decomposition. Noble metal nanoparticles loaded on rGO were synthesized in the one-pot method using modified polyol synthesis.23 Noble metal precursors of H2PtCl6, Na2PdCl4, and RuCl3 were first allowed to interact with GO templates and reduced simultaneously under the reflux or hydrothermal condition. Microstructural analysis of the synthesized hybrids was performed with X-ray diffraction (XRD) and transmission electron microscopy (TEM) in Figure S1 (Supporting Information). The peaks of nanocrystalline metallic Pt, Pd, and Ru along with the rGO are identified in XRD. A broad rGO peak is the characteristic of chemically reduced graphene. The noble metal contents of the hybrid catalysts were also estimated through thermogravimetric analysis (TGA) (Figure S2, Supporting Information) giving 49 wt % Pt, 45 wt % Pd, and
46 wt % Ru in the noble metal-rGO hybrids. Compared to the TGA curve of rGO only where the carbon backbone completely decomposes around 550−650 °C under air,24 noble-metal rGO hybrids displayed reduction in carbon decomposition temperature possibly because the intimate interaction with noble metals affects the decomposition behavior of graphene backbone. Chemical analysis by X-ray photoelectron spectroscopy (XPS) further revealed reduction of GO (Figure S3, Supporting Information). The C/O ratio of GO before reduction was 2.49 based on the peak area ratio of C 1s and O 1s, due to the presence of significant amount of oxygen-related surface defects. C/O ratio of the noble metal− rGO formed by the polyol method was estimated as 4.81. In the survey scan of the hybrids in Supporting Information Figure S2c−e, the elements carbon, oxygen, and the corresponding noble metal were verified. Surface oxidation state of the noble metals has been observed with XPS (Figure S4, Supporting Information). All the noble metals employed in this study were largely metallic with 20−30% of oxidized states consistent with previous reports.25 Particle size distribution was examined from TEM images (Figure S5, Supporting Information) and number-averaged particle size of the Pt, Pd, and Ru is shown in the corresponding graphs (Figure S1b, d, and f, Supporting Information). Pt, Pd, and Ru nanoparticles are formed exclusively on rGO supports in Figure S5 (Supporting Information). Well-distributed individual Pt, Pd, and Ru nanoparticles with size range of 1−4 nm were clearly visible. An average particle size of Pt, Pd, and Ru nanocatalysts is 2.29, 2.16, and 2.36 nm, respectively. Microstructures of the noble metals-rGO hybrid are also visualized with scanning electron microscopy (SEM) in Figure S6 (Supporting Information). All hybrids show the featured porous 3-dimensional (3D) networks of rGO composed of wrinkled 2D rGO sheets. There was no B
DOI: 10.1021/nl504425h Nano Lett. XXXX, XXX, XXX−XXX
Letter
Nano Letters
Figure 2. Morphology analysis of Ru-rGO hybrid cathode. (a) Typical discharge−charge profile of rGO and Ru-rGO hybrid electrode at a limited capacity of 10 000 mAh g−1. (b) and (c) Comparison of discharged electrodes. SEM images of the discharged electrodes at 10 000 mAh g−1. Red circles: discharge products. (b) rGO and (c) Ru-rGO hybrid cathodes. (d)−(f) SEM images of Ru-rGO hybrid electrodes at different charge capacities at points indicated in (a).
catalytic activity of Ru to other noble metals. At a capacity limit of 2000 mAh g−1, charge potentials of Ru-rGO electrode was reduced to ∼3.9 V from ∼4.3 V for rGO only electrode. After the first cycle, however, noble metal-rGO hybrids showed different cycling behaviors. The Ru-rGO hybrid electrodes exhibited stable voltage behavior with no significant change during the following cycles (Figure 1d). On the other hand, the voltage profiles of Pt- and Pd-rGO hybrids fluctuated and were very unstable (Figure 1b and c). Moreover, in the enlarged discharge voltage scale (Figure S7, Supporting Information), the discharge voltages of Ru-rGO hybrid were slightly increased from 2.71 V at the first cycle to 2.75 V at the sixth cycle, whereas those of Pt- and Pd-rGO hybrids were gradually decreased as the cells cycled suggesting that internal resistance increased due to undesirable reactions. When cycled longer than 6 cycles, cells with the Pt- and Pd-rGO cathode rapidly failed after 15th cycle (Figure S8a, and S8b, Supporting Information), whereas the Ru-rGO cathode could be stably cycled for more than 30 cycles (Figure S8c, Supporting Information) as in our previous report.26 The catalytic activity was also examined using cyclic voltammetry (CV). CV curves of rGO, Ru-rGO hybrid, and Pt-rGO hybrid measured in the voltage range of 2.4−4.3 V are presented in Figure S9 (Supporting Information). In Supporting Information Figure S9a and b, Ru-rGO exhibits a higher current density compared with rGO indicating enhanced electrochemical reactions during cycling. Pt-rGO electrode showed unstable CV curves consistent with the galvanostatic test results. It showed a current enhancement compared to the rGO only electrode at first, but a very large fluctuating current was observed at the fourth cycle in Supporting Information Figure S9c. The current level of Pt-rGO was lower than that of the Ru-rGO electrodes, implying a lower catalytic activity of Pt than that of Ru. When we focus on the performance of Ru-rGO hybrid, there are a few points to consider. Ru-rGO hybrid electrodes showed a higher current density in both cathodic
distinguished difference between the morphologies of Pt, Pd, and Ru-rGO hybrid. Energy dispersive X-ray spectroscopy (EDX) analysis also present the presence of carbon, oxygen, and the corresponding noble metal elements. Brunauer− Emmett−Teller (BET) specific surface area of pristine rGO only powder, Pt-rGO, Pd-rGO, and Ru-rGO hybrids is 51.7 m2 g−1, 207.4 m2 g−1, 99.7 m2 g−1, and 124.0 m2 g−1, respectively. The pore volume is 0.049 cm3 g−1, 0.2401 cm3 g−1, 0.2073 cm3 g−1, and 0.184 cm3 g−1 for the corresponding samples. It is clear that the catalyst nanoparticles loaded on rGO effectively suppress the restacking of rGO sheets during the drying process from the colloidal solutions. Although enhanced surface area and pore volume of rGO with supported catalysts could contribute to the enhanced reversible capacity, the main focus of the current investigation is the catalytic performance on the electrochemical reactions that could lower the overpotential issues in Li−air cells. In this regard, the small particle size and well dispersed uniform distribution of catalysts on the conducting support in hybrid system could be more important and dominating factors by providing higher surface area of active catalytic sites. The catalytic activity of supported noble metals-rGO hybrids was investigated in Li−air cells under capacity-controlled conditions with the fixed capacity regime of 2000 mAh g−1 and compared with that of rGO only electrodes. The specific capacity was calculated based on the total weight of noble metals plus rGO. In Figure 1a, all noble metal-rGO hybrid electrodes showed pronounced OER electrocatalytic effects (decreasing overall charge potential), whereas ORR discharge potentials did not change with the presence of noble metal catalysts consistent with the previously reported results.11 As observed in the charge voltage profiles at first cycle (Figure 1a), the overall charge voltage regions of all noble metal-rGO hybrid electrodes were remarkably lower than that of catalyst free rGO electrode. Especially, the average charge voltage of Ru-rGO hybrid electrode was 3.5 V indicating superior OER electroC
DOI: 10.1021/nl504425h Nano Lett. XXXX, XXX, XXX−XXX
Letter
Nano Letters
electrodes have quasi-film-like nanoparticle discharge products, a dramatic morphologic change such as toroids to films is not likely in our case. However, the role of Ru nanocatalysts in reducing discharge product sizes and inducing thin film like morphology is obviously identified. The discharge products formed were completely decomposed during the following recharging step. The morphologies of Ru-rGO hybrid electrodes at discharged state and at different charged states were examined by SEM in Figure 2d−f. When discharged to 10 000 mAh g−1, electrode surface was covered with film-like discharge products composed of fine nanoparticles. During the subsequent charge process, surface coverage of discharge products reduced at the charge capacity of 5000 mAh g−1 (Figure 2e). Charging to 10 000 mAh g−1 completely removed discharge products from the electrode (Figure 2f), implying the high reversibility of the Ru-rGO hybrid electrode. In Figure S10 (Supporting Information), rGO electrode also showed similar reversibility with Ru-rGO hybrid electrode. We suggest the dual role of noble metal catalysts in our Li− O2 system (Figure 3): (1) promoting thin film-like or
and anodic scans indicating enhanced reaction kinetics in both formation and decomposition of the discharge products. Although discharge potentials did not change with the presence of noble metal catalysts in the galvanostatic test in Figure 1, the reduction peak at ∼2.7 V in the cathodic scan of the rGO only electrode was broadened in the cathodic scan of the Ru-rGO electrode. More importantly, the onset potentials in anodic scans of Ru-rGO and rGO electrodes are almost identical at ∼3.2 V and the peak of the rGO only electrode at ∼4 V was shifted to 3.7 V in the Ru-rGO electrode (Figure S9d, Supporting Information). This kind of behavior has been observed by a few groups with Co3O4/RGO27 and RuO2/ CNT28 catalyst systems. These catalysts were not regarded as conventional electrocatalysts that lower the activation energy through electron transfer.27,28 However, Ru nanoparticles have been reported to electrocatalytically decompose bulk Li2O2 particles when mixed with chemically generated Li 2 O 2 particles.9,16 Therefore, it is evident that Ru has catalytic activity toward decomposition of bulk Li2O2 through electron transfer shifting oxidation peak to have a lower overpotential. We surmised that the peak at ∼4 V might correspond to the decomposition of bulk Li2O2, which has been shifted to a lower potential after Ru addition. The origin of the peak at ∼3.2 V has not been clearly identified yet. In our own and many other reports,16,26−28 electrochemically generated Li2O2 behaved differently from chemically produced bulk Li2O2. It may suggest that electrochemically produced Li2O2 is not totally the same with bulk Li2O2, which is stoichiometric and crystalline. Enhanced reduction reaction with the presence of Ru nanoparticles might account for the generation of kinetically different Li2O2 in the electrochemical cell. Very recently, researchers have focused on the presence of Li2O2 electrochemically different from crystalline bulk Li2O2, generated under kinetically limited conditions in the Li−O2 cells. The kinetic parameters that can induce the formation of this phase include high reaction rates (high current),29,30 large surface area (porosity),31 and the presence of catalysts.17,28 These conditions could influence the oxygen adsorption, superoxide formation, and disproportionate reactions, the widely recognized reaction pathways leading to Li2O2. The formation of this phase is often accompanied by the morphological transition from toroids to defective amorphous films or small nanoparticles.17,28−30 Morphological change might accompany because this phase is suggested to be produced under the surface diffusion controlled regime,29 and limited surface diffusion suppresses the growth of individual particles into toroids and reduces the feature size.28 Quasiamorphous thin films of Li2O2 have been reported to have lower overpotentials than toroidal ones.28,29 We speculate the activity of noble metals supported on rGO could be the favorable formation of this thin-film like phase and have traced the morphological difference between the catalyzed and noncatalyzed electrodes in Figure 2 and Figure S10 (Supporting Information). For better visualization, the electrodes were discharged and charged to 10 000 mAh g−1. In Figure 2b and c, discharge products produced in both rGO and RurGO hybrid electrodes assumed small nanoparticulate morphologies because the current densities applied in our system are rather high (fast rates). However, it is clear that the feature size is much fine for Ru-rGO (
