Niobium Tungsten Oxide in a Green Water-in-Salt Electrolyte Enables Ultra-Stable Aqueous Lithium-Ion Capacitors

Highlights A green water-in-salt electrolyte was developed using lithium acetate as solute with a wide electrochemical stability window of 2.8 V. Molecular dynamics simulation confirmed the nature of water-in-salt electrolyte, where hydrogen bonds of water–water were disrupted and ionic interactions became stronger than dilute solution. Nb18W16O93-based lithium-ion capacitors delivered unexceptionable stability over 50,000 cycles. Electronic supplementary material The online version of this article (10.1007/s40820-020-00508-z) contains supplementary material, which is available to authorized users.


S1.2 Materials Characterization
The crystal structures of the obtained samples were characterized by XRD (PANalytical Empyrean) with Cu Kα radiation (λ = 1.54056 Å). The microscopic morphologies were characterized using field-emission scanning electron microscopy (FE-SEM, Hitachi S-4800) and transmission electron microscopy (TEM, JEOL JEM-ARM300F). The X-ray photoelectron spectroscopy (XPS) data were collected on a PerkinElmer spectrometer (PHI 550) with the Al-Kα (1486.6 eV) as X-ray source. Thermogravimetric/differential scanning calorimetry (TG/DSC) data were collected on a NETZSCH (STA 409 PC) thermal analyzer under air atmosphere at a temperature ramp of 5 ℃ min -1 .

S1.3 Electrochemical Measurement
All the half-cell tests were performed in three-electrode Swagelok cell setup, where free-standing activated carbon (AC) films were employed as a counter electrode and Ag/AgCl electrode (Sat. KCl) served as the reference electrode. AC was purchased by Kuraray TM . AC counter electrodes were prepared by mixing AC, Ketjen Black and PTFE (mass ratio of 8:1:1), with a mass loading of about 20 mg cm -2 . The aqueous 13 m LiAc were employed as the electrolytes. The OECG working electrode was prepared by mixing OECG, Ketjen Black, and carboxymethylcellulose (CMC)/styrene-butadiene rubber (SBR) (CMC/SBR = 2:1) binder with a mass ratio of 8:1:1 in water, where the homogeneous slurry was then casted onto the carbon fiber paper (CFP), followed by drying at 60 ℃ for 12 h under vacuum. NbWO electrode was prepared by mixing NbWO, CNT, and CMC/SBR with a mass ratio of 8:1:1. It is worthwhile mentioning that NbWO and CNT (mass ratio = 8:1) were first ball-milled with moderate absolute ethanol at 400 rpm for 2 h. The obtained powder was then dried in air at 80 ℃ overnight. The active mass loading of NbWO electrode is about 1.5 mg cm -2 . We also employed polytetrafluoethylene (PTFE) as the binder to prepare the free-standing films of the working electrodes for the ex situ XPS, XRD, and thick electrode measurements. Aluminum and titanium rods were used as negative electrode and positive electrode current collectors, respectively. Filter paper was used as the separator. The electrolyte was purged by N2 for 30 minutes before use, and the cell was assembled in a simply equipped nitrogen glovebox. Cyclic voltammetry (CV) was carried out on an EC-Lab VMP300 multichannel workstation at various rates. The second cycle of CVs was used for kinetic analysis. Galvanostatic chargedischarge (GCD) measurements were conducted on a LAND CT2001A cell test instrument.
Electrolyte conductivity (σ, S cm -1 ) of electrolytes was calculated from the electrochemical impedance spectroscopy (EIS) on a Biologic EC-Lab VMP300 with a frequency range of 100 kHz to 100 Hz at 25 ℃. The test cell consisted of two parallel titanium rod electrodes on either side of a Φ0.5 × 0.1 cm (rradius × l) electrolyte reservoir. The uncompensated resistance (R, Ω) from the Nyquist plot was assumed to be dominated by electrolyte resistance and used to calculate conductivity.
The specific energy density (E, Wh kg −1 ) and power density (P, W kg −1 ) of the ALIC can be obtained as follows [S2, S3]: where I is the current density normalized by the total active mass in both electrodes (A g −1 ), and t1 and t2 are the start and end time during the discharge process (h), U is the voltage (V).

S1.4 Molecular Dynamics (MD) Simulations
Molecular dynamics (MD) simulations were performed for 1 m LiAc and 13 m LiAc electrolytes, in order to examine the composition of the hydration shell of cations. Relevant data for the systems can be found in Table S1. The OPLS-AA force field is chosen for its ability to accurately simulate liquid phase systems [S4-S7]. The MD computer simulation process is a combination of three steps. The first step is energy minimization. When packmol [S8] is used to fill a certain percentage of molecules into a cubic box with periodic boundary conditions in all three Cartesian directions, it will produce some unreasonable contact between molecules, and the system energy is in a higher state. Therefore, we pass the first step to minimizing the energy of the system. The conjugate gradient algorithm is used to minimize energy and the tolerance is set to 10 kJ mol −1 nm −1 . The second step is the equilibrium phase. In order to achieve the lowest energy isomer state of each molecule in the system, we perform a single annealing method in the NVT ensemble. The temperature is maintained by the Velocity-rescale thermostat with a time constant of 0.1 ps [S9]. First, the system is heated to 500 K in 30 ps, then keep the temperature for 10 ps, and finally cooled to 298.15 K in 2 ns. The final step is the production phase. The production simulations are performed in the NPT ensemble at constant pressure and constant temperature. The temperature is maintained by the Velocity-rescale thermostat at 298.15 K for all other systems with a time constant of 0.1 ps. A constant pressure of 1 bar is controlled by the Parrinello-Rahman barostat [S10, S11] with a coupling constant of 2 ps. Electrostatic interactions are treated using the Particle -Mesh-Ewald (PME) method [S12, S13] with the Fourier spacing of 0.118 nm and 0.9 nm real-space cut-off. The visualization of the molecular structures (see Fig. 1) was realized utilizing the noncommercial Visual Molecular Dynamics (VMD) software [S14].   Obviously, OECG electrode exhibits a reversible adsorption/desorption behavior with Ac − in 13 m LiAc electrolyte with the potential window of 0-0.9 V vs. Ag/AgCl. However, in 1 m LiAc electrolyte, it doesn't work well in the same potential window. The capacity still retains 24.0 mAh g −1 even at 15 A g −1 . After 20,000 cycles, the capacity retention is about 96% at 2 A g −1 .