In-situ characterization of discharge products of lithium-oxygen battery using Flow Electrochemical Atomic Force Microscopy

Highlights

• A novel Flow Electrochemical Atomic Force Microscopy (FE-AFM) technique is described.

• The technique was tested for studying the discharge products of a Lithium-Oxygen Battery (LOB).

• The FE-AFM cell allows us to visualize the discharge of thin Li₂O₂ deposits on HOPG.

• The formation of larger nanoparticles, with elongated and spherical globular shape, is attributed to the solution mechanism of Li₂O₂ formation.

• The oxidation of the nanoparticles, follow a bottom-up oxidation mechanism, characterized by a significant reduction in the height of the particles.

Abstract

The increasing interest in lithium-oxygen batteries (LOB), having the highest theoretical energy densities among the advanced lithium batteries, has triggered the search for in-situ characterization techniques, including Electrochemical Atomic Force Microscopy (EC-AFM). In this work we addressed the characterization of the formation and decomposition of lithium peroxide (Li₂O₂) on a carbon cathode using a modified AFM technique, called Flow Electrochemical Atomic Force Microscopy (FE-AFM), where an oxygen-saturated solution of the non-aqueous lithium electrolyte is circulated through a liquid AFM cell. This novel technique does not require keeping the AFM equipment inside a glove-box, and it allows performing a number of experiments using the same substrate with different electrolytes without disassembling the cell. We study the morphology of Li₂O₂ on graphite carbon using lithium bis(trifluoromethane sulfonyl)imide (LiTFSI) in dimethyl sulphoxide (DMSO) as electrolyte under different operational conditions, in order to compare our results with those reported using other electrolytes and in-situ and ex-situ EC-AFM.

Introduction

Since it was described for the first time by Abraham and Jiang in 1996 [1], the Li-O₂ battery (LOB) has attracted enormous research attention, aiming to approach its outstanding theoretical energy density [2], close to 3500 Wh kg⁻¹, needed to improve the autonomy of electrical vehicles beyond the limit imposed by the Li-ion batteries. It is now well-known that there are important technical challenges for developing any practical LOB, namely, high charge potential, low discharge capacities, and poor electrolyte and cathode stability that strongly limit discharge−charge cycle efficiency and power density [3], [4], [5], [6].

The mechanisms of the oxygen reduction reaction (ORR) and the oxygen evolution reaction (OER) in a LOB operating with aprotic solvents have been clarified by several authors [7], [8], [9], [10], [11]. Thus, the ORR discharge reaction leading to the formation of Li₂O₂ on the cathode can occurs through surface or solution routes where, in both cases, the first step is the reduction of O₂ to the superoxide radical (O₂^−•). The solution mechanism takes place in solvents with high donor number, where the superoxide ion forms ion pairs and higher aggregates with Li⁺ ions in the solution, which diffuse before chemical disproportion into Li₂O₂, leading to the formation of a thick layer. Solvents with low donor number promote the surface mechanism, leading to thin films (< 10 nm in thickness) due to the insulating nature of the deposit that self-limits its growth.

The dissociation degree of the lithium salt also seems to affect the morphology of the Li₂O₂ deposits [12]. Thus, a “bottom-up” Li₂O₂ precipitation mechanism operates with highly dissociated Li salts, such as LiTFSI in low donor number (DN) glymes, forming a thin layer of the product that cover the surface till the electrons can not penetrate further and the discharge process stops. On the contrary, for highly associated salts, such as LiNO₃ in diglyme, a “top-down” precipitation mechanism that expands in all directions takes place, leading to large deposits. However, since the degree of the lithium salt association depends on a balance between the interaction of the Li⁺ ions with the solvent and the counteranion [13], it is not straightforward to predict the type of deposit that can be formed in a LOB under a given choice of salt and solvent.

For this reason, the direct observation of morphology of the cathodic product during the discharge, and its subsequent dissolution (oxygen evolution) during the charge cycle, is of crucial interest for the optimization of a LOB in order to improve the electrolyte performance. The use of Atomic Force Microscopy (AFM) is well known in the field of lithium batteries as a direct method to visualize the formation of the solid-electrolyte interphase (SEI) and the dendritric formation in lithium-ion batteries (LIB) [14,15]. A recent review by Zhao et al. [16] briefly mentioned a few AFM studies describing the use of AFM for the characterization of LOBs. Thus, it is worth to review here the studies reported during the last years, which analyze the morphology and conductivity of the discharge products of LOB using AFM and Electrochemical Atomic Force Microscopy (EC-AFM), performed in-situ [17], [18], [19], [20], [21], [22], and ex-situ [23], [24], [25].

Byon and coworkers carried out the first in-situ study of the formation[17] and decomposition[18] of Li₂O₂ on HOPG (highly oriented pyrolytic graphite) using the EC-AFM technique, with 0.5 M LiTFSI in TEGDME as electrolyte and metallic Li as counter-electrode. The AFM was located inside a glove-box and the contact mode AFM was realized with silicon nitride tips for the electrochemical scan and Si tips to observe the HOPG surface in tapping mode, once the cell was disassembled. The Li₂O₂ deposit on HOPG terrace edges exhibited nanoplate shape having several hundred of nm length and 5 nm height, and they disappear during the oxygen evolution reaction (OER). A similar study on a mesoporous gold electrode using the same electrolyte [19], allowed to visualize the formation of a film of Li₂O₂ nanoparticles with sizes around 10 nm, which grows up as the discharge proceeds, reaching a final size of 50 nm.

Liu and Ye [20] used in-situ EC-AFM to determine the morphology of the deposits on gold electrodes with 0.1 M LiClO₄ in DMSO as electrolyte. They found that the shape of the Li₂O₂ nanoparticles depend on the water content of the electrolyte, and concluded that there is no evidence of a disproportion of LiO₂ in solution for the formation of Li₂O₂ on the surface. The voltagram shows two cathodic peaks which were attributed to the formation of the superoxide ion followed by the formation of the peroxide. At a cathodic potential close to the first peak small particles were observed on the surface, which grow up when the second peak is reached. During the anodic scan, in an electrolyte containing 1000 ppm of H₂O, three peaks of low intensity were observed, indicating that the species formed during the ORR are not fully oxidized during the OER. The two first anodic peaks were assigned to partial oxidation of the peroxide and superoxide (in solution) on the surface, but the process that originates the third peaks was not identified. If the electrolyte contains only 33 ppm of H₂O, the cathodic voltagram only shows one peak, and the Li₂O₂ particles are smaller as the water content diminishes.

Shen et al. [21] studied by EC-AFM the ORR/OER processes on HOPG using 0.5 M LiTFSI in DMSO with the addition of DBBQ (2,5-di-tert-butyl-1,4-benzoquinone) as soluble catalyst. The voltagrams show two cathodic peaks, attributed to the superoxide and peroxide formation, and three anodic peaks, the first one assigned to the superoxide oxidation and the other two to the peroxide oxidation. The AFM images in absence of DBBQ show the initial formation of Li₂O₂ of size around 5 nm when the first peak is approached, while a toroidal morphology around 300–400 nm in diameter is formed when the second peak is reached. During the charge process a bottom-up mechanism of toroid decomposition is observed. In the presence of DBBQ the Li₂O₂ deposit adopts a flower-like shape that decomposes from the outside to the inside during the OER.

An AFM cell that allows the control of the atmosphere over the glassy carbon substrate was built by Virwani et al. [22]. They used 1 M LiNO₃1 M in TEGDME as electrolyte, with different water contents, and analyzed the topography of the deposits during the ORR and OER. A remarkable effect of the water content was observed on the deposit morphology and the discharge capacity of the cell. For instance, the cell capacity increases 11 times when the water content increase from 20 to 4600 ppm. The Li₂O₂ deposits formed with small water contents exhibit spherical or bar shape in the range between 200 and 400 nm, and they appear at the final of the discharge. With high water content, large toroidal particles (∼ 400 nm) are observed at the beginning of the discharge, while smaller ones are formed later.

Ex-situ AFM studies were performed by Calvo and coworkers using 0.1 M LiPF₆ in DMSO on HOPG [23,24]. In this case the cathodic voltagram has only one peak, while the anodic peaks were assigned to the oxidation of the superoxide, the solvent, and other products of reduction of oxygen. The particles deposited on the HOPG edges and terraces never exceed 12 nm. The few particles deposited at low polarization are supposed to form by disproportion of the LiO₂, while the massive deposit at higher cathodic potentials would be due to the superoxide reduction over the surface.

An ex-situ study on HOPG aimed to analyze the effect of the incorporation of chloride on the capacity of LOB [25], and it is the only AFM work where the electrical conductivity of the formed film was analyzed. The authors claim that an important increase in the film conductivity was detected by Conductive-AFM, when Cl⁻ ions are added to the electrolyte. However, some doubts on the conclusions arise due to the lack of care when the film deposited on the electrode is transferred to the AFM cell. It is probable that the Li₂O₂ was converted to LiOH before the conductivity was measured with the AFM and also, that the high conductivity reported be a consequence of the contact of the Pt AFM tip with the HOPG surface in some places. In fact, DFT studies on Li₂O₂ structures doped with Cl⁻ ions have probed that the conductivity of the systems is not altered by the presence of chloride or other halogen anions [26].

In this work we have developed a novel procedure for in-situ observation of the deposit in a LOB during discharge/charge cycles, which can be performed without resorting to the use of a glove-box, making the technique more simple that previously reported in the literature. The technique was used to analyze the morphology of the Li₂O₂ deposit on highly oriented pyrolytic graphite (HOPG), using lithium bis(trifluoromethane sulfonyl)imide (LiTFSI) in dimethyl sulphoxide (DMSO) as electrolyte, under different operational conditions.