Prolonged Cycle Life for Ultrafast Li4Ti5O12//Li3V2(PO4)3 Superredox Capacitor

Abstract

1. Introduction One promising strategy of increasing the energy density of electric double layer capacitors (EDLCs) is the design of hybrid supercapacitors, which combine an activated carbon (AC) electrode with an electrode made from a high-capacity faradic (or pseudocapacitive) material such as lithium titanate (Li4Ti5O12, LTO). We have previously reported the LTO//AC hybrid capacitor system—called Nanohybrid Capacitor—shows the 3-fold energy density (30 Wh L-1) of EDLC, while attaining the high power (6,000 W L-1) comparative to EDLC.1) To achieve further increase of energy density from Nanohybrid Capacitor, replacement of the AC positive electrode is required by alternatives with higher capacity as well as ultrafast electrochemical characteristics and excellent cycle capability. Lithium vanadium phosphate [Li3V2(PO4)3, LVP] is promising candidate for positive electrodes due to its relatively high reaction potentials (3.9 V vs. Li/Li+), high reversible capacity (131.5 mAh g-1), and a large value of the Li+ diffusion coefficient (10-9 -10-11 cm2 s-1). In particular, we have successfully synthesized LVP/multiwalled carbon nanotubes (MWCNTs) composites which enable high C-rate operation of 96 mAh g-1 at 300C—more than twice the rate possible with AC electrodes—via our unique technique of ultracentrifugation (UC) treatment.2) Beyond the challenge of improving power performance, a more serious drawback is that the cycle performance of previous LVP-based cells has been limited to 500-4,000 cycles.3,4) In this study, we designed LVP-based full cells consisting of positive electrodes made from the uc-LVP/MWCNT composite paired with negative electrodes made from standard commercially available LTO—which we call SuperRedox Capacitors—that simultaneously achieve high energy, high power density, and long cycle life. We also successfully identify the mechanism of capacity degradation during full cell cycling and devise a strategy for minimizing its effect on the cycling performance.2. Experimental LTO/LVP full-cells were assembled using negative LTO and positive uc-LVP/MWCNT composite electrodes in laminate-type cells. The electrolyte was 1.0 M LiPF6/EC:DEC (1:1, volume ratio). Charge-discharge tests for LTO//LVP full cells were performed in constant-current charge and discharge modes between 1.5-2.8 V. Current densities for the full cells were 10C-rate for cycling tests and ranged from 1 to 480C-rate in rate tests, assuming that 1C-rate equals 131.5 mA g-1.3. Results and Discussion Ragone plots of LTO//LVP full cells calculated based on the total volume of two electrodes are shown in Fig. 1. LTO//LVP full-cells exhibit high volumetric energy density of 63.5 Wh L-1 within the region of low power requirement (100 W L-1) which corresponds to the 5-folds of EDLC. Even at a higher power (10,000 W L-1), 63% of the energy density (40 Wh L-1) can be maintained. Accordingly, the obtained Ragone characteristics for our LTO//LVP full cells demonstrated that this system can be operated at high power comparative to the EDLC, while showing the merit of this system in terms of volumetric capacity compared to EDLC and even Nanohybrid Capacitor. Our LTO//LVP full cells also demonstrated outstanding cyclability: capacity retention of 77% over 10,000 cycles (Fig. 2). Such stable cycle performance was achieved thanks to the electrochemical preconditioning [=Li preconditioning of LTO, state of charge (SOC)=25%] conducted prior to the full cell assembling. In the process of elucidation of the Li-preconditioning effects on cycle performances, it was found that minimization of the vanadium elution from the LVP and the subsequent deposition on the LTO surface play an important role for the stable cycling. Combined results of XPS, ICP-MS and SEM observation suggest that the deposited vanadium species on the LTO surface induces the decomposition of electrolytes and production of HF. The irreversible electrolyte decomposition leads to a decrease in the coulombic efficiency of Li+ intercalation/deintercalation into LTO crystals, resulted in the gradual shift between two electrodes: higher SOC for uc-LVP/MWCNT positive and lower SOC for LTO negative electrodes. Additionally, the produced HF is considered to induce further elution of vanadium from the uc-LVP/MWCNT, accelerating the SOC shifts and degradation in the full cell capacity. Li preconditioning of LTO was found to be effective as a countermeasure, because of the given capacity margin to minimize the effect of SOC shifts, and the formation of the protective coverture on the LTO surface—composed of such as LiF and Li2CO3—from the undesirable vanadium deposition. References 1) K. Naoi et al., Energy Environ. Sci., 5, 9363 (2012).2) K. Naoi et al., J. Electrochem. Soc., 162, A827 (2015).3) M. Secchiaroli et al., J. Mater. Chem. A, 3, 11807 (2015).4) C. Liu et al., Energy Storage Mater., 5, 93 (2016). Figure 1