Rapid synthesis of transition metal dichalcogenide–carbon aerogel composites for supercapacitor electrodes

Transition metal dichalcogenide (TMD) materials have recently demonstrated exceptional supercapacitor properties after conversion to a metallic phase, which increases the conductivity of the network. However, freestanding, exfoliated transition metal dichalcogenide films exhibit surface areas far below their theoretical maximum (1.2 %), can fail during electrochemical operation due to poor mechanical properties, and often require pyrophoric chemicals to process. On the other hand, pyrolyzed carbon aerogels exhibit extraordinary specific surface areas for double layer capacitance, high conductivity, and a strong mechanical network of covalent chemical bonds. In this paper, we demonstrate the scalable, rapid nanomanufacturing of TMD (MoS2 and WS2) and carbon aerogel composites, favoring liquid-phase exfoliation to avoid pyrophoric chemicals. The aerogel matrix support enhances conductivity of the composite and the synthesis can complete in 30 min. We find that the addition of transition metal dichalcogenides does not impact the structure of the aerogel, which maintains a high specific surface area up to 620 m2 g−1 with peak pore radii of 10 nm. While supercapacitor tests of the aerogels yield capacitances around 80 F g−1 at the lowest applied currents, the aerogels loaded with TMD’s exhibit volumetric capacitances up to 127% greater than the unloaded aerogels. In addition, the WS2 aerogels show excellent cycling stability with no capacitance loss over 2000 cycles, as well as markedly better rate capability and lower charge transfer resistance compared to their MoS2-loaded counterparts. We hypothesize that these differences in performance stem from differences in contact resistance and in the favorability of ion adsorption on the chalcogenides.


THEORETICAL SURFACE AREA CALCULATION
For each TMD, we calculated the theoretical surface area by extrapolating from the unit cell and neglecting edge effects. As an example calculation for WS 2 , the specific surface area is where a is the (100) lattice parameter (a = b = 3.1532 Å and c = 12.3230 Å), MW is the molecular weight of the TMD unit cell, and N a is Avogadro's number. For WS 2 , MoS 2 , and NbSe 2 , the theoretical specific surface areas are 482.77, 752.14, and 468.87 m 2 /g, respectively.

SCHERRER ANALYSIS OF TMD-LOADED AEROGELS
We quantified the size of the TMD crystals loaded into the aerogels using the Scherrer equation, where the crystal thickness L a in the (002) direction is a function of the x-ray wavelength λ Kα , the position of the (002) diffraction peak θ, and the full width at half maximum (FWHM) of the peak β, which we correct for instrument broadening with β ¼ where β m and β ref are the measured FWHM of the TMD (002) peak and the measured FWHM of a peak of a corundum standard that occurs at a similar Bragg angle to the TMD (002) peak 1 .

VOLUMETRIC CAPACITANCE CALCULATION FROM GALVANOSTATIC DISCHARGE
We calculated volumetric capacitance of a single electrode from galvanostatic discharge profiles using the equation adapted from gravimetric capacitance 2 : where I is the constant discharge current, m a is the total mass of active material in both electrodes, ρ is the bulk density of the active material, and (t 1 , V 1 ), (t 2 , V 2 ) are two chosen points on the voltage-time curve. Our reported values use the full potential window of our discharge tests (0.9-0.1 V). We note that, while gravimetric capacitance is frequently used as a figure of merit for supercapacitors, volumetric capacitance represents a more pragmatic metric for high density energy storage and a more reliable parameter in evaluating the chargestorage performance of low bulk density electrode materials, like aerogels 3,4 . Additionally, volumetric capacitance normalizes results for significant changes in density. Finally, many earlier studies on the supercapacitor applications of TMD-carbon composites employed thin film-like electrodes obtained through filtration 5-8 . Owing to the very small mass of active material, this can lead to artificially high values of gravimetric capacitance that do not translate to larger-scale devices.

VOLUMETRIC CAPACITANCE CALCULATION FROM EIS
Volumetric capacitance can also be calculated from electrochemical impedance spectroscopy (EIS) using the equation 9 : where f is the frequency and Z" is the imaginary component of impedance, m a is the total mass of active material in both electrodes, and ρ is the bulk density of the active material. Supplementary Figure S5 plots the EIS-derived volumetric capacitances of the aerogels as a function of frequency. Even at the lowest frequency of 10 mHz where volumetric capacitance is maximized, it is less than one-third of the maximum value obtained from galvanostatic testing. We note that the relatively low capacitance obtained from EIS versus other methods is welldocumented for supercapacitors based on porous carbons 9-12 . For example, Lufrano et al. observed that the capacitance of carbon composite electrodes with H 2 SO 4 electrolyte was 20% lower when measured with EIS than with galvanostatic charge-discharge tests 9 . Various explanations have been proposed for the discrepancy-such as the increased hindrance for alternating current penetration into the bulk electrode 10,11 ; or the presence of deeply trapped ions that are immobile under AC conditions but can be released at low potentials in DC methods 13 . However, nothing conclusive has emerged to date.

Figure S4
Electrochemical characterization of MA-17 and WA-17 supercapacitors. This includes Nyquist plots from EIS (a: full range, b: detail of high to mid frequency), cyclic voltammetry at a sweep rate of 20 mV s − 1 (c), and the specific volumetric capacitance (d) as a function of applied current density from galvanostatic tests. Note on Figure S4. The Nyquist plot of MA-17 exhibits unusual characteristics at lower frequencies that are difficult to interpret. Between the semicircular R|C region and the steeply sloped constant phase region, there is a small loop followed by a larger arc that curves back on the real impedance axis, instead of the expected 45°Warburg line. These features indicate low-frequency inductive behavior in the system, which is not typically observed in supercapacitors. We note, however, that these features also appeared for duplicate coin cells made with the same active material. One possible explanation is offered by Bisquert et al., who extensively studied inductive phenomena in the context of a porous, heterogeneous electroactive material composed of two different solids in contact with electrolyte 14 -a good model for our TMD/carbon aerogel composite electrodes. They found that inductive behavior resulted from the coupled dielectric relaxation in the two solid phases, with the relaxation being driven by changes in the electrochemical potential of the phases due to charge transfer between them.

Figure S5
Specific volumetric capacitance, derived from electrochemical impedance spectroscopy, as a function of frequency for the different aerogel supercapacitors. Figure S6 Specific volumetric capacitance of WA-17 during galvanostatic cycling at 0.25 A g − 1 , derived from discharge curves at selected cycles.

Figure S7
Nitrogen sorption isotherms with BET surface area (a) and BJH pore size distribution (b) of pyrolyzed WS 2 -loaded aerogel before processing, after grinding and sieving, and after forming into a supercapacitor electrode with PTFE binder and carbon black.

Figure S8
Raman (a) and XRD (b) characterization of the NbSe2-loaded aerogel after pyrolysis at 800°C.

Figure S9
Energy dispersive x-ray spectroscopy of MoS 2 powder and MoS 2 -loaded aerogel after pyrolysis.

Figure S10
Energy dispersive x-ray spectroscopy of WS 2 powder and WS 2 -loaded aerogel after pyrolysis.