AVS2014 Session EN+EM+NS-TuA: Charge Storage Materials and Devices

We adopted a holistic approach in studying electrochemical processes and mechanism of the Li-O₂ battery using operando methods. Particularly, we introduced the microfocused synchrotron X-ray diffraction (μ-XRD) and tomographic (μ-CT) techniques for the spatiotemporal study on the phase and structural changes in Li-O₂ battery. These tools offered some unique capabilities to probe battery properties under the actual discharge-charging condition. For example, the μ-XRD has a spatial resolution at the micron scale of with a complete side penetration to the battery, rendering it feasible to study battery’s composition layer-by-layer without the interruption of battery operation. In this presentation, we will discuss our recent investigation of the Li-O₂ batteries under cycling condition in real time using the cells fabricated with the most representative design and materials. We were able not only to reveal individually the changes at anode, cathode and separator, but also to provide a comprehensive view between the regional chemical processes and their interdependence to the overall battery performance during the multiple discharge-charge cycles. More importantly, the finding of this study provides new insights on the catalytic process inside of Li-O₂ cell and calls for new design and materials which could lead to high capacity and longer battery life.

The work performed at Argonne is supported by DOE under Contract No. DE-AC02-06CH11357 by UChicago Argonne, LLC.

Nanostructured battery electrodes provide a design opportunity to achieve high power at high energy density, using thin active storage layers whose short ion diffusion pathways assure fast transport throughout the layers. However, this must be coupled with fast electron transport through current collectors to all regions of the ion storage layers, posing a design challenge in balancing and optimizing both charge transport components. Spatial inhomogeneity in the utilization of active material due to electronic or ionic transport limitations may lead to decreases in performance, but characterizing this effect with bulk electrochemical measurements is difficult. We address this challenge with a new state-of-charge (SOC) measurement scheme utilizing a patterned ultra-thin film battery electrode and spatially resolved XPS, and focus on the case of limited electronic transport by examining SOC as a function of distance from the current collector.

We fabricate electrode test structures by evaporating metallic strips as current collectors on an electrically insulating substrate. A patterned thin film of active material (V₂O₅) is then deposited using atomic layer deposition (ALD) and mechanical masking so that only a small fraction of the active material is in contact with the current collector. The use of ALD allows for an ultrathin (≤ 30nm) pinhole-free film. We discharge these electrodes in a liquid electrolyte under different rates and conditions and directly measure the state of charge as a function of distance from the current collector using small spot XPS, achieving a lateral resolution of better than 20µm. We find that a rate-dependent SOC gradient develops in the electrodes, with the SOC decreasing with distance from the current collector. Unlike microspot Raman or XRD, XPS provides a direct quantitative measurement of the SOC through the concentration of inserted ions and/or reduced vanadium ions. Additionally, in the ultrathin films relevant to nanostructured storage, XPS becomes a “quasi-bulk” measurement, because the escape depth of photoelectrons becomes a significant fraction of the film thickness. We also explore the depth dependence of the SOC using angle resolved XPS and ion beam depth profiling. We compare our observations with simulations using COMSOL Multiphysics, and attempt to resolve discrepancies between the two. We believe this approach can provide design guidance for heterogeneous nanostructures applied to electrical energy storage, and we anticipate it to be broadly applicable to other electrode materials and active ions.

Li-ion conversion batteries can store 2-3 times more charge than intercalation batteries by utilizing the full range of oxidation states of their constituent divalent or trivalent transition metal compounds during discharge. A prototypical conversion compound is CoO, which follows the reaction

2Li⁺ + 2e^- + Co⁽²⁺⁾O --> Li₂O + Co⁽⁰⁾.

Cobalt oxide and other transition metal oxides are attractive for use as Li-ion anodes in portable electronics due to their high charge storage capacity and moderate voltage versus Li⁺/Li⁰. However, the cycling stability of conversion electrodes is poor, and capacity losses have thus far prevented their implementation.

In order to understand phase progression during the conversion reaction of CoO, high-purity CoO thin films grown in UHV were sequentially exposed to atomic lithium. The electronic structure of the pristine films and of the products of lithiation was studied using x-ray photoemission spectroscopy (XPS), UV photoemission spectroscopy, and inverse photoemission spectroscopy. The crystal structure and film reorganization were probed in parallel with transmission electron microscopy (TEM) and scanning tunneling microscopy.

The amount of CoO reduction for a given Li dose was observed to be highly dependent upon the temperature at which lithiation was performed. At 150^oC, Li mobility in the active material was sufficient to allow full reduction of the CoO film as confirmed by XPS. Consistent with electrochemically lithiated CoO electrodes, precipitation of Co nanoparticles in a Li₂O matrix was observed in TEM images. However, at room temperature, the Li-rich overlayers that formed on the CoO film after initial lithiations inhibited further Li diffusion. This could be due to the intrinsically poor kinetic properties of Li₂O or to the formation of Li₂O₂ and/or LiOH passivating films.

The reactivity of CoO films was also found to depend on the orientation of the film. CoO(100) films exhibited a higher degree of conversion for a given Li exposure than polycrystalline films. STM and angle-resolved XPS of these films have been used to investigate the differences between these two film morphologies upon exposure to Li.

Electrochemical power sources based on metal anodes have specific energy density much higher than conventional Li ion batteries, due to the high energy density of the metal anode (3842mAh/g¹ for Li). Rechargeable Li-O₂ batteries consume oxygen from the surrounding environment during discharge to form Li oxides on the cathode scaffold, using reactions

(1) [anode] Li(s) ↔ Li⁺ + e⁻

(2) [cathode] Li⁺ + ½ O₂ (g)+ e− ↔ ½ Li₂O₂ (s), ~2.959 V vs Li/Li⁺

(3) [cathode] Li⁺ + e⁻ + ¼ O₂ (g) ↔ ½ Li₂O (s), ~ 2.913 V vs Li/Li⁺

The cathode reaction requires large over-potentials for charging due to the mass transfer resistance of reagents to the active sites on its surface, decreasing the round trip efficiency, making recharge of the Li-O₂ cell difficult. To overcome these problems, the cathode needs good electrical conductivity and a porous structure that enables facile diffusion of oxygen and can accommodate the reduced oxygen species in the pores.

Two significant challenges exist in the use of the traditional activated carbon material as the cathode of the Li-O₂ system. First, in the presence of Li₂O₂ the carbon electrode becomes relatively unstable even at low voltages (~3V). Second, cathode structures must be porous to accommodate a substantial amount of Li–peroxide (Li₂O₂) without blocking ion transport channels in the cathode. While a few studies have been reported on the effect of catalyst on the onset potentials for the oxygen evolution reaction (OER) and oxygen reduction reaction (ORR) in the Li-O₂ cell, the results were inconclusive due to the lack of systematic study in a single system and conditions.

We report here results from a model cathode system which enable determination of the effects of various catalysts on the OER/ORR reactions in the non-aqueous Li-O₂ cell. Mesoporous CNT sponge is used as the model cathode material, decorated with catalyst nanoparticles by nucleation-controlled atomic layer deposition (ALD) of Ru, RuO2, MnO2, and Pt catalyst components whose loading and composition are controlled by manipulating the ALD conditions. Using a custom Li-O₂ battery cell, we have studied the effect of different catalysts on the voltage of the OER and ORR, and on the cycling performances of the cell. We demonstrate a Li-O₂ cell that sustains >3000 mAhg_c^-1 over more than 15 cycles at current density of 200 mAg_c^-1. To our knowledge, this is the first comparison of a variety of catalysts with a well-defined morphology (controlled by ALD and monitored by TEM), and under the same electrochemical conditions.

The Li-O₂ battery system is one of the prime candidates for next generation energy storage. Like other metal-O₂ systems, this technology is known for its impressive theoretical specific energy due to use of metallic anodes and because the cathode active material (oxygen) is not stored in the battery, but is available in the cell environment. A typical cell consists of a pure lithium metal anode, an organic electrolyte (in this study), and a porous positive electrode (usually made of carbon or oxides) which acts as a reaction scaffold for oxygen reduction to Li₂O₂ or Li₂O during discharge. Despite remarkable scientific challenges within every component of the cell, the positive electrode is particularly complicated by its role in the oxygen evolution (OER) and reduction (ORR) reactions, leading to strict requirements for electrode architecture and physicochemical stability for optimal performance. We present herein one of the first experimental realizations of a controlled macroscale 3D carbon nanotube architecture with a practical carbon loading of 1mg/cm² in an attempt to satisfy these requirements.

The O₂ cathode highlighted in this work features a macroporous nickel foam current collector coated with dense forests of vertically aligned carbon nanotubes (VACNT). This freestanding, hierarchically porous system is the first to feature VACNT robustly and electrically connected to a 3D current collector without a binder, and without requiring delamination of the CNT from the growth substrate. Grown via LPCVD with an Fe catalyst on a thin ALD interlayer, the micron-length VACNT provide a very promising electrode material due to their high electrical conductivity, physicochemical stability, and a high surface area architecture that is conducive to ionic mobility and storage of the reduced oxygen discharge product. As a result, this structure exhibits significant capacity (>2Ah/g-carbon) at high ORR voltages (>2.76V) without requiring a catalyst.

Electrochemical performance results as a scaffold for oxygen reduction in various non-aqueous electrolytes will be presented with SEM/TEM/XPS of pristine/discharged electrodes.

The 3D micro-supercapacitor reported here utilizes a novel bottom-up assembly method that combines genetically modified Tobacco mosaic virus (TMV-1Cys) with deposition of RuO₂ on multi-metallic microelectrodes. The nanostructured RuO₂ coating is selectively deposited on the electrodes due their unique composition, which is a significant advantage for microfabrication process integration. Test results show electrode capacitance as high as 18 mF/cm² in 1.0M H₂SO₄ electrolyte and 7.2 mF/cm² in solid Nafion electrolyte.

The device fabrication involves the photolithographic patterning of titanium nitride (TiN) microelectrodes with Au cap on top of polyimide micropillars supported by a silicon wafer. A schematic cross section of the device is shown in Figure 1 and a photograph of the fabricated chip in Figure 2. The complexity of the self-assembly process in multiple chemically reactive solutions required the development of a special kind of micro-electrode. The TiN functions as a chemically resistant current collector, the Au cap as an adhesion layer for the TMV-1cys, and the Ni pad as a sacrificial material during the RuO₂ deposition process. After microfabrication, each chip is submerged in TMV-1Cys solution for 24 hours and then transferred to a 0.5% solution of RuO₄. A nanostructured coating of RuO₂ forms on all exposed electrode areas as the Ni is sacrificed in a galvanic displacement reaction. EDX spectral imaging of the constituent elements on the electrode demonstrates selective RuO₂ coating (Figure 3), and SEM images of the electrodes before and after TMV/RuO₂ coating shows the TMV-1Cys/RuO₂ nanostructures (Figure 4).

Cyclic voltammetry (CV) was performed from 0-800mV versus Ag/AgCl at 10 mV/s in 1.0M H₂SO₄ electrolyte. Figure 5 shows the CV curves, and Figure 6 shows the associated capacity fading, which was insignificant after 100 cycles for electrodes annealed at 150°C. Separately prepared chips were coated with Nafion dispersion and tested in a controlled humidity environment. The measured capacitance drops from 18 to 7.2 mF/cm² per electrode due to ionic conductivity limitations, but 80% capacity is retained after 12,000 cycles (Figure 7). Associated rate capability (Figures 8-9) shows 60% capacity is retained when comparing 3 uA/cm² to 3000 uA/cm², and the low leakage current of only 5 nA (Figure 10) enables use in a wide variety of energy storage applications.

The primary challenge of nanomaterials is often integration into microfabrication processes. The RuO₂ electrode developed here is optimized for compatibility with standard microfabrication steps by using a novel bottom-up assembly approach for manufacturing micro-supercapacitors.

Characterization of nickel oxide supercapacitor electrodes utilizing Tobacco mosaic virus (TMV) nanotemplates is presented. NiO was formed on Ni coated TMV nanotemplates by annealing at high temperatures (Figure 1). The resulting electrode showed excellent electrochemical performance with remarkable cycle stability. The TMV/Ni/NiO nanostructured electrodes were also integrated with a solid electrolyte to demonstrate their potential application as solid flexible supercapacitors.

NiO supercapacitor electrodes have been prepared in literature using various methods, and it has been found that the crystallinity of the NiO is critical for its electrochemical charge capacity [1]. The NiO electrode presented in this work was thermally grown on Ni coated TMVs. Gold electrodes (0.5cm²) were immersed in TMV solution for virus self-assembly followed by electroless deposition of Ni uniformly coating the TMV nanostructure [2]. TMV/Ni electrodes were annealed in a furnace at three different temperatures (200°C, 300°C, and 400°C) and the NiO formation on TMV/Ni surface was characterized by XPS (Figure 2). The results indicate that thermal growth of NiO layer on TMV/Ni electrodes starts at temperatures higher than 300°C, in good agreement with previously reported results.

Electrochemical testing was performed in aqueous 2M KOH electrolyte in a three-electrode configuration. The electrodes annealed at 300°C showed the highest areal capacitance (148mF/cm²) measured by a galvanostatic (2mA/cm²) charge/discharge test shown in Figure 3a. The redox charge storage mechanism was confirmed by cyclic voltammetry (CV) with good rate capability up to 100mV/s (Figure 3b). Excellent cycle stability was measured with little degradation over 500 cycles as shown in Figure 4. This is attributed to the conformal layers of Ni/NiO over the TMV nanostructure, and the stabilizing effect of KOH on NiO. The continuous electrical contact between the Ni and NiO layers ensures an optimized current collector configuration.

A PVA-KOH-H₂O polymer was prepared to study the performance of the nanostructured electrodes with a solid electrolyte. Polymer electrolyte solution was poured onto the nanostructured NiO electrodes and the Pt foil was assembled on top as an anode. The polymer electrolyte film formed after 24hours was flexible and strong enough to support both electrodes. Figure 5 shows CV curves measured with the assembled cell, verifying proper operation of the nanostructures in both liquid and solid electrolytes. The successful integration of TMV/Ni/NiO electrodes with polymer electrolytes highlights the potential of this approach to develop flexible solid-state supercapacitor devices.

Among devices that have been used to store antimatter, Penning-Malmberg trap has become the device of choice because of its simplicity and versatility. However, the challenge involved in these traps is when the number of particles increases inside the trap to the densities of energy harvesting interest, the confining fields rise to unpractical values. One of the authors has proposed a design of microtube arrays with much lower end barrier potentials. The microtraps are designed for non-neutral plasma storage such as positrons. Here, we present fabrication, simulation studies, and trapping milestones so far. The fabrication involved advanced MEMS techniques including photolithography, deep reactive ion etching of silicon wafers, sidewalls smoothening, gold sputtering, wafers aligning, and thermo-compression gold bonding. Alignment of less than 2 microns was achieved using a micro-machined jig and precision ground sapphire rods. Simulation using a WARP Particle-In-Cell code showed that density of 1.6×10¹¹cm^-3 is achievable with the new trap design while the end barrier potentials are several order of magnitudes smaller compared to the conventional traps. However, positron losses occur in experimentation by both trap imperfections such as misalignment of wafers, asymmetries, and physical imperfections on the surface, and also field misalignment and perturbations. The loss rates were also compared to the results from simulation in order to study and distinguish each effect. This project will open the door to a wide range of new and exciting research areas. The size of these traps along with the low confining potentials is a big step to make them portable. It could be used as a source of energy or in propulsion system where alternate sources are not feasible.