Letter Cite This: Nano Lett. XXXX, XXX, XXX−XXX
pubs.acs.org/NanoLett
Nanocrevasse-Rich Carbon Fibers for Stable Lithium and Sodium Metal Anodes Wooseok Go,† Min-Ho Kim,† Jehee Park,† Chek Hai Lim,† Sang Hoon Joo,† Youngsik Kim,*,†,‡ and Hyun-Wook Lee*,† †
School of Energy and Chemical Engineering, Ulsan National Institute of Science and Technology (UNIST), Ulsan 44919, South Korea ‡ Energy Materials and Devices Lab, 4TOONE Corporation, UNIST-gil 50, Ulsan 44919, South Korea Nano Lett. Downloaded from pubs.acs.org by UNIV OF EDINBURGH on 01/23/19. For personal use only.
S Supporting Information *
ABSTRACT: Metallic lithium (Li) and sodium (Na) anodes have received great attention as ideal anodes to meet the needs for high energy density batteries due to their highest theoretical capacities. Although many approaches have successfully improved the performances of Li or Na metal anodes, many of these methods are difficult to scale up and thus cannot be applied in the production of batteries in practice. In this work, we introduce nanocrevasses in a carbon fiber scaffold which can facilitate the penetration of molten alkali metal into a carbon scaffold by enhancing its wettability for Li/Na metal. The resulting alkali metal/carbon composites exhibit stable long-term cycling over hundreds of cycles. The facile synthetic method is enabled for scalable production using recycled metal waste. Thus, the addition of nanocrevasses to carbon fiber as a scaffold for alkali metals can generate environmentally friendly and cost-effective composites for practical electrode applications. KEYWORDS: Alkali metal carbon composite, Li metal anode, Na metal anode, scalable production, carbon scaffold
A
chemical interactions, various carbon species such as hollow carbon spheres,21 carbon nanofiber films,22 layered reduced graphene oxide (rGO),12 and carbonized wood15 have shown promising electrochemical performance. Lin et al.12 fabricated layered rGO with nanoscale interlayer gaps to host lithium metal and demonstrated that this material achieved more stable galvanostatic cycling than conventional lithium metal foil. This promising result was attributed to the synergetic effect of the lithiophilic interface and the capillary force induced by interlayer nanogaps. Unfortunately, the reported Li-rGO composite requires a spark reaction to fabricate the lithiophilic sample that can be employed in the limited area. Various carbon nanostructures have been designed to trap Li/Na within the carbon framework through plating or deposition. Three previously reported strategies11,15,23 to infuse molten Li or Na metal into carbon nanostructures are
nodes based on lithium and sodium metal have emerged as promising alternatives to conventional graphite anodes in lithium- and sodium-ion batteries due to their high theoretical specific capacities and low redox potentials.1,2 However, the use of alkali metal-based anodes is limited by the formation of metal dendrites during electrochemical cycling, which promotes short-circuiting in batteries. Various strategies have been introduced to address this issue, including the addition of artificial solid-electrolyte interphase (SEI) layers,3−5 the introduction of additives to the electrolyte,6−8 the use of superconcentrated electrolytes,9,10 the use of threedimensional current collectors,11−18 and nanoscale interfacial engineering.19,20 Although these approaches have successfully improved the performances of alkali metal anodes, many of these methods are difficult to scale up and thus cannot be applied in the production of batteries in practice. Carbon materials have unique advantages in battery applications because of their low density, natural abundance, different functionalities, and numerous fabrication methods. When designed to have appropriate physical shapes and © XXXX American Chemical Society
Received: October 12, 2018 Revised: November 19, 2018 Published: November 28, 2018 A
DOI: 10.1021/acs.nanolett.8b04106 Nano Lett. XXXX, XXX, XXX−XXX
Letter
Nano Letters
Figure 1. Schematic showing the fabrication process of the Li and Na metal/carbon composites. (a) Li or Na metal is infused into commercial carbon cloth to form alkali metal/carbon composites. (b) A single carbon fiber can form many nanocrevasses after the heat treatment at 500 °C before molten Li or Na metal infusion. (c,d) Scheme for the comparison of Li plating/stripping behavior on (c) a pristine Cu electrode and (d) a nanocrevasse-rich carbon structure-modified electrode. (e) The simplicity of this method makes it suitable for scalable production. Large-scale Li or Na metal carbon composites can be produced with a simple machine.
briefly described as follows. In the first strategy, the wettability of the carbon nanostructure is increased by coating with silicon. Liang et al.11 coated carbon with silicon via chemical vapor deposition to increase the wettability of carbon. The silicon coating then alloys with lithium to form LixSi, which has a high affinity for Li and thus facilitates the infusion of molten Li into the scaffold. In the second strategy, the carbon surface is imparted with lithiophilicity by coating with ZnO. Zhang et al.15 demonstrated that carbonized wood with channel walls for Li ion flux can be wet by coating with ZnO, generating a high affinity for molten Li. Finally, in the third strategy nitrogencontaining functional groups serve as lithiophilic sites in a graphene anode. Zhang et al.23 reported that doped graphene containing uniformly distributed N-containing functional groups (e.g., pyridinic and pyrrolic nitrogen) exhibited lithiophilic properties, leading to uniform Li nucleation in contact with molten lithium. Although these approaches facilitate the deposition of Li metal at the desired locations in the electrode, they require multiple labor-intensive synthetic and/or preparation steps. In a previous study,24 we prepared interwoven carbon nanofibers to suppress polysulfide dissolution. The viscous polysulfides were trapped between the narrow gaps in the matrix via capillary force. On the basis of this approach to suppress polysulfide dissolution, we hypothesize that the strong binding ability of carbon nanofibers could also enhance surface lithiophilicity to facilitate the infusion of alkali metal into a carbon fiber host. In this work, we present a facile synthetic method to fabricate alkali metal/carbon composites by immersing carbon cloth into alkali metal melts (Figure 1a).
This synthetic method is straightforward, fast, and enables scalable composite production. Commercial carbon cloth is lightweight, has excellent mechanical properties along with high electrical conductivity, and exhibits good thermal stability in the various area.25 Furthermore, carbon cloth has a rough, checkered three-dimensional shape with a relatively large surface area; this shape is ideal for alkali metal deposition and helps to lower the local current density.11 However, pristine cloth does not make a good scaffold because of the poor affinity between molten Li/Na metal and carbon, which leads to poor wettability. The affinity of the scaffold for molten metals is a critical factor in metal infusion because it affects the adhesion force between alkali metals and the cloth surface. Thus, to ensure that the molten metals are absorbed into the carbon cloth, the carbon should be modified to enhance its affinity for molten Li/Na. In this study, we enhanced the wettability of carbon by introducing nanocrevasses in the carbon cloth (Figure 1b), leading to the successful infusion of alkali metal. The heattreated carbon cloth (Figure S1) has excellent mechanical and chemical properties along with a larger active surface area than Li or Na foil. As a result, it effectively acts as a host scaffold for alkali metal. The composites prepared using this carbon cloth achieve long-term cycling and are not inhibited by the formation of alkali-metal dendrites because the local current density is low (Figure 1c,d). Furthermore, the fast and straightforward infusion of alkali metals into the carbon cloth allows the large-scale production of composite electrodes for use in seawater batteries and other real electronic applications. B
DOI: 10.1021/acs.nanolett.8b04106 Nano Lett. XXXX, XXX, XXX−XXX
Letter
Nano Letters
Figure 2. Changes in the physical and chemical properties of carbon cloth before and after heat treatment. (a−c) SEM images of the carbon scaffolds used for alkali metal infusion: (a) pristine carbon cloth; (b) enlarged image of the pristine cloth, showing the absence of pores and crevasses; and (c) the carbon fiber cloth after heat treatment, showing numerous pores and nanocrevasses on the fiber surfaces. (d,e) N2 adsorption data with BET surfaces areas of the carbon fiber cloth before (d) and after (e) heat treatment. The surface area was significantly increased after heat treatment. (f,g) Deconvoluted XPS C 1s spectra of carbon cloth before (f) and after (g) heat treatment. (h,i) Images showing the wetting test results of molten alkali metal on the carbon scaffolds: (h) pristine carbon cloth, which shows poor wettability; and (i) heat-treated carbon cloth, which shows excellent wettability. (j) Schematic showing how alkali metal infuses into the heat-treated carbon scaffold via nanocrevasses.
Thus, the proposed composite represents a significant step toward the commercialization of metal anodes. The alkali metal/carbon composite electrode was fabricated using a simple melt infusion method with a carbon scaffold. Most commercial carbon fibers contain thin polymer-based surface coatings (i.e., sizing agents; Figure S2) that help maintain the shape of the cloth. Because these sizing agents have no affinity for molten alkali metals and prevent direct contact between the carbon cloth and molten alkali metals during the infusion process, the carbon fibers in this study were heat-treated at 500 °C for 4 h under ambient atmosphere to remove the surface polymers. After heat treatment, the sizing agents were completely evaporated, and the carbon cloth had a strong affinity for molten metal. Thus, molten alkali metal could infiltrate the carbon cloth across the entire surface within a few seconds (Supplementary Movies S1 and S2).
Although a few previous studies have used a similar process for metal infusion into a scaffold, the scaffolds in these reports often suffered severe damage by a sudden reaction (e.g., a “spark” reaction).11 In contrast, the method described herein did not cause macroscopic damage to the carbon cloth after heat treatment, enabling the production of large-area composites. Our method is also applicable to various polyacrylonitrile (PAN)-based scaffold materials, such as carbon filters, carbon felt, and carbon paper (Figure S3). Thus, the composite shape and size can be easily controlled by selecting or modifying the carbon scaffold. To facilitate largescale production, we also developed an automated roller machine for alkali metal-infiltrated carbon cloth (Figure 1e). This machine allows mass composite production and may provide a foundation for future commercialization. The method demonstrated in this study drastically decreases the C
DOI: 10.1021/acs.nanolett.8b04106 Nano Lett. XXXX, XXX, XXX−XXX
Letter
Nano Letters
Figure 3. Electrochemical characterization of the Li composites. (a−c) Comparison of the cycling stabilities of the Li/carbon composite (green) and bare Li electrode (orange) symmetrical cells at current densities of 1 (a), 2 (b), and 3 (c) mA cm−2, respectively. The insets in (a) show enlarged voltage profiles with cycle numbers indicated. (d) Nyquist plots of Li-based symmetric cells after 10 galvanostatic cycles. (e,f) Top-view SEM images of the bare Li electrode (e) and the composites (f) after 100 cycles (scale bars = 50 μm).
resulting in the evolution of CO and CO2,26 which also account for the different wettability of the pristine and heattreated carbon fibers. X-ray photoelectron spectroscopy (XPS) was used to evaluate the surface chemical compositions of the pristine and heat-treated carbon cloth samples (Figure 2f,g and Figure S5). The XPS C 1s spectra can be deconvoluted into the following functional groups; aromatic CC (284.6 eV),24,25 hydroxyl COH (286 eV),27,28 C*OCO and epoxy groups (286.6 eV),27 carbonyl CO (287.6 eV),27 OC OH (289.1 eV),29 and the π−π* transition in aromatic groups (291 eV).28 The XPS C 1s spectra of the pristine carbon cloth (Figure 2f) show peaks corresponding to the oxygencontaining functional groups of sizing agents, including peaks of the π−π* transition, OCOH, and COH along with a distinct and strong peak of the COC epoxy group. In contrast, the COC epoxy peak disappeared in the spectra of the heat-treated carbon cloth,30 and CO peaks appeared (Figure 2g), as expected, along with slight shift due to the charging effect. Although the COH peak seems to be the remnants after the stabilization process, the peak can be newly created by the reaction with air during heat treatment (Figures 2g and S5). The XPS results indicate that heat treatment formed a new area consisting of oxygen-containing functional groups, suggesting the composite has strong interaction similar to alkali metal with rGO. In summary, the SEM, BET, and XPS data indicate that heat treatment enhanced the affinity of the
production cost by using recycled alkali-metal waste from exhausted batteries or leftover alkali metal foils. Thus, this method is expected to contribute to the environmentally friendly and cost-effective commercial production of alkali metal-based composite electrodes. Figure 2a shows a scanning electron microscopy (SEM) image of a pristine carbon cloth framework, which exhibits a checkered pattern comprising numerous overlapping carbon fibers. Figure 2b,c shows the SEM images of the carbon fiber cloth before and after heat treatment, respectively. The corresponding Brunauer−Emmett−Teller (BET) data of the pristine and heat-treated carbon cloth samples are shown in Figure 2d,e, respectively. The surface area of the pristine carbon is approximately 3.8 m2/g (Figure 2d), in agreement with the smooth surface observed in Figure 2b. In contrast, the surface area of the heat-treated carbon cloth is 454.5 m2/g (Figure 2e), indicating that heat treatment greatly increased the surface area by the numerous nanocravesses on carbon fibers, which corresponds to surface roughness in Figure 2c. The assertion is based on pore size distribution (Figure S4). After heat-treatment, numerous micropores (
