(Invited) Characterization of Mxene/Graphene Supercapacitor Yarns Using Time and Frequency Models

Abstract

Supercapacitors (SC) turned out to be efficient devices for electrochemical energy storage thanks to their high energy density vs. classical surface capacitors- and high power density vs. batteries. In the absence of electrode components prone to react with the electrolyte, charges are stored in the pore space in a purely capacitive way under the form of electrical double-layer EDL at the solid-liquid interface [1, 2]. In several cases, additional pseudo-capacitive energy storage appears through redox reactions, underpotential deposition of metals or -to a lower extent- to ions intercalation [3]. This may also lead eventually to charge losses. The optimization of supercapacitors can be facilitated using electrode and electrolyte models that capture the physics at the local scale to better understand the charge and mass transfer phenomena governing their behavior. Indeed, SC characterization is still based on relatively simple models based on macroscopic parameters that are generally determined according to voltage evolution during constant current charge and discharge. The electrical equivalent circuit (EEC) often includes only an equivalent series resistance and a capacitance tied up to the average pore population governing their dynamic behavior [1, 2]. In order to account for pore size distribution within the electrodes, EEC consisting of two or more branches of resistances and capacitances in series are sometimes used [1, 2], although some parameters may appear to be correlated. In addition to the charge/discharge behavior, Electrochemical Impedance Spectroscopy (EIS) is another extensively used tool for characterizing electrochemical systems [2, 4]. Since single or multiple branches RC EECs based on surface reaction assumption are not adapted to the frequency domain, SC impedance spectra are usually modelled using Transmission Line Models (TLM) [2, 4] accounting for the volumetric character of the electrodes. TLM are however rarely applied for SC characterization in the time domain, and in the few cases that exist, not under the form of simple analytical expressions [2]. In order to yield a basis for SC modelling and characterization, we present a parameter study comparing different model approaches. They can be applied to purely capacitive and pseudo-capacitive SC with the ultimate objective to derive a model able to capture part of the physics at the local scale and to describe the SC behavior in both the time and frequency domains. As an example we compare the parameters identified in the time and frequency domains on different types of supercapacitor yarns developed by Gao’s group at NCSU [5], including pure double-layer capacitive reduced Graphene Oxide (rGO) yarn SCs and titanium carbide (MXene) – rGO yarn SCs presenting an additional pseudo-capacitive behavior and potentially ion intercalation [3]. In these flexible 2D material based devices, the active components, 2D flakes, are stacked together in distinct morphologies (either smoothly packed with grooves on fiber surfaces, or highly wrinkled via cation crosslinking), wherein different porosity and pore distributions exist. Experimental results as reported by Dr. Gao’s group exhibit dramatically different specific capacitances and rate capabilities. Figure Caption Left: Scheme of the MXene – rGO yarn SC structure; right: Nyquist plot of the MXene-rGO Sc impedance data and interpolation curves obtained with a purely capacitive (dashed line) and a pseudo-capacitive (full line) TLM model.