Binary carbon-based additives in LiFePO4 cathode with favorable lithium storage

A pairwise coupling of 0D Super-P (SP), 1D carbon nanotubes (CNTs), and 2D graphene nanosheets (GNs) into binary carbon-based conductive additives was used here for the LiFePO4 cathode in lithium-ion batteries. For comparison, the LiFePO4 cathode with SP, CNT, or GN unitary conductive agent was also examined. Electrochemical test results suggest that the cathodes with binary conducting additives present greatly improved electrochemical performance than the traditional cathode system (only SP used). Especially, the LiFePO4 cathode containing 3% CNT component exhibits the highest specific capacity and the best cycling stability among all the cathodes with binary conducting additives, indicating that an appropriate amount of CNTs is critical in enhancing the conductivity and practical capacity output. However, an excess of CNTs leads to entangling with each other, hampering the uniform distribution of active materials and resulting in poor electrode performance. Furthermore, the combination of CNT and GN can effectively improve the capacity and cycling stability of the LiFePO4 cathodes due to the synergistic effect of 3D conductive networks constructed by the two.


Introduction
Recently, intensive research has been devoted to devising high-performance lithium-ion batteries (LIBs) by developing novel battery materials and designing innovative internal battery structure [1][2][3][4][5][6][7][8][9][10]. Practically, the electrochemical performance of LIBs mainly depends on the electrodes (cathode and anode), which are the key components of battery. In general, the electrodes are mainly comprised of active materials, conductive agents, and polymeric binders.
Nowadays, to effectively utilize the active materials, the proportions of the conductive carbon in the LiFePO 4 cathode system often reach up to about 3-5 wt% and even as high as 6-10 wt% especially in the case of highrate LIBs. This substantially reduces the energy and power densities of the resulting batteries. Therefore, it is imperative to develop high-efficiency conductive agents that can afford adequate conducting network with appropriate addition amount for future high-performance LIBs.
The normally used conductive additives often display relatively low conducting efficiency due to their small particle surfaces (only the outer carbon plane can contact with active materials), which can be improved by introducing carbon materials with large specific surface area. Recently, graphene (GN) and carbon nanotubes (CNTs), as novel but powerful conductive additives, have attracted increasing attention because of their unique nanostructure, flexible conducting network, and excellent electronic conductivity. Due to its one-atom thicklayered 2D structure, graphene has large specific surface area, which can effectively establish contact with the active materials and favor fast Li + transport [40][41][42][43][44][45][46]. On the other hand, CNTs with wire-like shape and high aspect ratio are conductive to constructing continuous conductive network for rapid electron transfer [47][48][49][50][51]. Thus, the combination of CNTs with GN as composite conductive additives is supposed to construct a 3D structure that may form well-distributed conductive network to improve the electrochemical performance of the LiFePO 4 cathode. Up to now, there are few reports on the synergistic effect of GN and CNTs on the LiFePO 4 cathode.
Herein, we fabricated a series of binary carbonbased conductive additives consisting of SP, CNTs, and/ or GN and investigated their impacts on the electrochemical performance of the LiFePO 4 cathodes. The electrochemical performances of the LiFePO 4 cathodes with single conductive agent (SP, GN, or CNTs) were also tested for comparison. This work will provide theoretical and practical bases for the application of high-performance conductive agents in LIBs.

Materials
LiFePO 4 was purchased from BTR New Material Group Co. Ltd (China). SP was purchased from Timcal (Switzerland). CNTs and graphene were obtained from Chengdu Institute of Organic Chemistry, Chinese Academy of Sciences, and Guangdong Xigu Tanyuan New Materials Co. Ltd (China). Poly(vinylidene fluoride) (PVDF) and N-methyl-2-pyrrolidone (NMP) were purchased from Arkema (Changshu, China) and Sinopharm Chemical Reagent Co. Ltd (China), respectively. All chemicals were of analytical grade and used as received without further purification.

Electrode preparation
The LiFePO 4 cathode slurry with unitary and binary conductive agents was prepared via a facile mechanical stirring method. Typically, the preparation process of LiFePO 4 cathode slurry was performed in an agitated autoclave. First, 8 wt% dispersion liquid of PVDF in NMP was prepared and transferred into the agitated autoclave. Then an appropriate amount of conductive agents was added under continuous stirring until well blended. Last, the LiFePO 4 powders were added under constant stirring to obtain a homogeneous slurry with composition of 80 wt% active materials, 10 wt% conductive agents (unitary or binary), and 10 wt% PVDF. The solid content of the cathode slurry was controlled to be around 48 wt% to ensure the slurry viscosity of about 8,000 mPa s.
Then the cathodes were fabricated by compressing the as-prepared slurry onto an aluminum foil current collector followed by drying at 80°C in air for 24 h and 120°C in vacuum for 12 h. The electrode film was then compacted and punched into disks with a diameter of 10 mm for the half-cell assembly. Note that the LiFePO 4 cathode prepared with x% CNTs and y% GN was named as CNT-GN-x:y.

Cell assembly
Coin-type (CR 2032) half-cells were assembled in an argon-filled glove box, using the as-prepared LiFePO 4 electrode plate as a working electrode and lithium foil as a counter electrode. A Celgard-2400 microporous membrane was used to separate two electrodes. A 1-mol/L solution of LiPF 6 dissolved in EC/DMC (1:1, v/v) was used as the electrolyte.

Electrochemical testing
After assembly, the cells were stored and aged for 12 h before the electrochemical tests. The galvanostatic charge and discharge measurements were performed using a LAND CT2001A battery testing system with the voltage window between 2.5 and 3.7 V vs Li + /Li. Cyclic voltammetry (CV) studies were conducted by PARSTAT 4000A electrochemical workstation at a scan rate of 1 mV s −1 in the voltage range of 2.5-4.2 V. Conductivity tests were carried out on a four-point probe resistivity tester (ST2258B-F02, Beijing Bolun JingWei Tech Dev Co. Ltd) using a four-point probe method by coating LiFePO 4 cathode slurry onto polyethylene terephthalate. All electrochemical tests were performed at room temperature.

Materials characterization
X-ray diffraction (XRD) patterns were recorded using a D8 Advance diffractometer (Bruker, Germany) equipped with a Cu-Kα radiation source. Scanning electron microscope (SEM) images were taken on an SU-8010 field emission SEM (Hitachi, Japan). The slurry apparent viscosity was measured by an MSK-SFM-VT precision digital viscometer (HF-Kejing, China).

Raw materials characterization
The powder XRD patterns of LiFePO 4 , SP, CNT, and GN are given in Figure 1. As shown in Figure 1a, all the intense peaks could be readily indexed to a wellcrystallized LiFePO 4 (JCPDS card No. 40-1499), with an orthorhombic olivine structure. Two broad (002 and 001) peaks of SP revealed its poor crystallinity. In contrast,  Figure 3 shows the surface SEM images of the LiFePO 4 cathode plates with different unitary conductive agents. As shown in Figure 3a, the smaller SP particles could easily fall into the interspace between the larger LiFePO 4 grains and were very difficult to link up with each other [30]. As a result, the conductive paths in the cathodes are interrupted. By contrary, in cathodes with CNTs or GN, the conductive agents could easily form continuous conductive networks. In Figure 3b, the long wire-like CNTs connected with multiple LiFePO 4 particles and uniformly extended throughout the composite cathode without any obvious entanglement. In Figure 3c, the surface of individual fully exfoliated GN nanosheet was decorated by evenly dispersed LiFePO 4 particles, which could facilitate electron transfer across the surface of LiFePO 4 particles in short range, and ensured high utilization of GN, efficient long-range conductivity as well as sufficient contact between electrolyte and active materials [35]. Accordingly, it is promising to achieve high conductivity of the cathode with such continuous 2D conductive networks.

Electrode characterization
The surface SEM images of the LiFePO 4 cathode plates with various binary conductive agents in different raw material ratios are shown in Figure 4. Figure 4a1-a3 consists of the SEM images of the cathodes prepared with SP/GN in ratios of 7:3, 5:5, and 3:7, respectively. In Figure 4a1, the conductive network in LiFePO 4 electrode was mostly composed of SP particles due to the low content of GN. While in Figure 4a2-a3, both SP and GN particles were dispersed well among active material particles in LiFePO 4 cathode, suggesting the formation of an effective electronic conducting network. Figure  4b1-b3 shows the SEM images of the cathodes prepared with SP/CNT in ratios of 7:3, 5:5, and 3:7, respectively. When the content of CNTs was 3% (Figure 4b1), active material particles were uniformly dispersed, and CNTs interpenetrated between SP and active material particles to form a conductive network, which was promising to achieve high conductivity (Figure 4b1). By gradually increasing the content of CNTs, obvious entanglement of CNTs could be observed, which resulted in an agglomeration and isolation of the active material particles (Figure 4b2 and b3) [52]. The cathodes prepared with both CNTs and GN presented a very distinctive morphology (Figure 4c1-c3). Obvious agglomeration of active material particles was observed (marked in Figure 4c2 and c3), and this phenomenon was getting worse with an increase in CNT addition. As the content of CNTs was increased to 7% (Figure 4c1), the CNTs became invisible due to their self-entanglement and agglomeration of the active material particles (the inserted SEM); while large GN sheets could be clearly observed. Figure 5 shows the electrochemical behavior of the LiFePO 4 cathodes with unitary conductive agent. Typical charges/discharge profiles of the LiFePO 4 cathodes at 1 C between 2.5 and 3.7 V are presented in Figure 5a. It could be seen that the initial charge and discharge profiles of the three LiFePO 4 cathodes were almost the same and no obvious polarization was observed, delivering discharge/ charge capacities of 131.8/130.7, 140.1/138.1, and 149.8/ 147.5 mA h g −1 , respectively. Obviously, the LiFePO 4 cathode with GN as conductive additive displayed the highest specific capacity, which could be attributed to its long-range electron conduction and large specific surface area. By contrary, SP conductive agent demonstrated low electrical conductivity and local electron conduction, thus its corresponding LiFePO 4 cathode displayed the lowest specific capacity among the three [53]. CV was performed in order to investigate the effect of the conductive agent on the electrochemical properties of the LiFePO 4 cathodes at 0.1 mV s −1 between 2.5 and 4.2 V. Figure 5b-d shows the initial three CV of the LiFePO 4 cathodes with SP, CNTs, and GN conductive agents, respectively. One pair of well-separated anodic/ cathodic peaks was discovered in all CV curves, which coincided well with the results of voltage profiles in Figure 5a. Besides, the anodic/cathodic peaks of the LiFePO 4 cathode with GN were more symmetrical and sharper, indicating its better electrochemical activity [40]. Furthermore, the redox pairs of the LiFePO 4 cathode with GN exhibited the minimum potential separation (0.3 V) among the three. Therefore, with the well-defined peaks and small peak potential separation, the LiFePO 4 cathode with GN conductive agent could achieve good electrochemical performance. Figure 5e compares the cycling properties of the resulted electrodes at 1 C. The capacity values of all electrodes exhibited a remarkable increase in the first 25 cycles, which then became stable, keeping 127.5, 137.9, and 141.8 mA h g −1 after 200 cycles. The initial capacity increases could be related to the electrode activation process. As expected, the LiFePO 4 cathode with GN showed both higher capacity and better cycling stability than that of the other. Moreover, when the current density increased stepwise from 0.2 to 5 C and returns to 0.2 C, a high reversible capacity (142.5 mA h g −1 ) and a high capacity retention (99.5%) could be achieved for the GN electrode (Figure 5f).

Electrochemical performance
To figure out the synergy between two different conductive agents, we examined the galvanostatic charge/discharge properties of a series of LiFePO 4 cathodes with varied compositions (Figure 6). Figure 6a compares the charge/discharge profiles of the LiFePO 4 cathodes with different binary conductive agents in the same ratio. The similar charge and discharge profiles of all electrodes indicated similar electrochemical behavior of the LiFePO 4 cathodes with different binary conductive agents. And SP-GN-5:5 electrode exhibited higher reversible capacities (151.2 mA h g −1 ) than that of SP-CNT-5:5 (138.1 mA h g −1 ) and CNT-GN-5:5 electrodes (115 mA h g −1 ). Figure 6b-d presents the effect of the ratio of binary conductive agents on the charge/discharge property of the LiFePO 4 cathodes. In SP-CNT binary systems, the capacity of the electrodes showed negligible improvement by increasing the content of CNTs (Figure 6b). However, a large number of curled CNTs could be entangled with each other when combined with GN, which adversely affected the dispersion of the active materials, and resulted in poor electrode performance. Therefore, the electrode capacity decreased with an increase in CNT content (Figure 6d). In addition, both SP-GN electrode (0.08 V) and CNT-GN electrode (0.07 V) exhibited smaller potential difference than that of SP-CNT electrode (0.08 V), which could be attributed to the excellent electrical conductivity and effective "plane-to- point" conducting mode of GN [34,35]. The results indicated that CNTs exert a great effect on the electrode capacity and GN can effectively improve the conductivity of LiFePO 4 cathodes. The cycling performance and rate capability of the LiFePO 4 cathodes with different binary conductive compositions are shown in Figure 7. It can be found that the electrodes with binary conductive agents displayed better electrochemical performance than those with SP only, while the electrodes with the CNT-GN combination achieved both high capacity and good cycling stability (Figure 7a-c), which may be related to their respective merits and the synergistic effect [38,54]. Based on Figure 7b, as expected, the electrode with 5% GN (SP-GN-5:5) showed the highest reversible capacity, but the electrode capacity dropped sharply with further increase in the GN content (SP-GN-3:7), probably due to the steric effect from the large planar structure of GN nanosheets [55]. On the other hand, excessive GN may cause re-stacking of GN nanosheets, which is unfavorable for smooth Li + diffusion. Furthermore, from Figure  7a and c, the capacity of the electrode containing 3% CNTs (SP-CNTs-7:3 and CNTs-GN-3:7) was obviously higher than that with 7% CNTs (SP-CNTs-3:7 and CNTs-GN-7:3). This means that adding excess CNTs does not benefit the electrochemical performance of the electrode, which was consistent with the aforementioned results. The rate performances of the electrodes with different binary conductive agents in the same ratio are compared in Figure 7d. All electrodes with binary conductive agents exhibited stable capacity and high capacity retention at various current densities. The low capacity of SP-GN-5:5 electrode at 0.2 C may be related to its slow electrode activation process.
The conductivity test results of LiFePO 4 cathodes with different binary conductive agents are shown in Figure 8. It can be found that the electrical conductivity improved with an increase in the CNT or GN content. SP-CNT cathodes with 3%, 5%, and 7% CNTs exhibited electrical conductivities of 73.8, 104.9, and 144.1 mS mm −1 (Figure 8a electrical conductivities of the electrode containing GN were higher than that of the electrode with CNTs, which could be attributed to the higher electrical conductivity of GN than that of CNTs. In addition, the electrodes containing both CNTs and GN exhibited high and stable electrical conductivities, resulting from their excellent conductivity and the synergistic effect originating from the 3D conductive network constructed by CNT nanowire and GN nanosheets (Figure 8c). It should be noted that the electrical conductivity fluctuated significantly when the content of the CNTs increased to 7% (Figure 8a and c), indicating a heterogeneous distribution of the conductive network, which was directly reflected in the decreased electrochemical performance of the electrodes. All these results corresponded well with the SEM and electrochemical test results.

Conclusions
In summary, binary conductive agents were made by adding suitable amount of commercial CNTs and/or GN to SP to improve the conductivity of the LiFePO 4 cathodes in LIBs. The impact of single conductive agent and the binary conductive agents on the electrochemical performance of the LiFePO 4 cathodes was inspected. Meanwhile, the effect of the ratio of binary conductive agents on the properties of the LiFePO 4 cathode was also investigated. Results indicated that the binary conductive agents can provide more effective conductive network and faster Li + transport compared to traditional cathode system (only SP used). Especially, the electrode prepared with CNTs and GN as binary conductive agents exhibited both high capacity and good cycling stability. The best electrochemical performance of the cathode was obtained with the combination of 3% CNTs and 7% GN. Therefore, the binary conductive agents consisting of CNTs and GN present a promising candidate for devising high-performance LiFePO 4 cathodes. Such facile and scalable preparation procedure could be extended to various prospective electrode materials.