Aqueous Activated Graphene Dispersions for Deposition of High-Surface Area Supercapacitor Electrodes

High-surface area activated graphene has a three-dimensional porous structure that makes it difficult to prepare dispersions. Here we report a general approach that allows the preparatioon of stable water-based dispersions/inks at concentrations of ≲20 mg/mL based on activated graphene using environmentally friendly formulations. Simple drying of the dispersion on the substrate allows the preparation of electrodes that maintain the high specific surface area of the precursor material (∼1700 m2/g). The electrodes are flexible because of the structure that consists of micrometer-sized activated graphene grains interconnected by carbon nanotubes (CNTs). The electrodes prepared using activated graphene demonstrate performance superior to that of reduced graphene oxide in supercapacitors with KOH and TEA BF4/acetonitrile electrolytes providing specific capacitance values of 180 and 137 F/g, respectively, at a specific current of 1 A/g. The high surface area of activated graphene in combination with the good conductivity of CNTs allows an energy density of 35.6 Wh/kg and a power density of 42.2 kW/kg to be achieved. The activated graphene dispersions were prepared in liter amounts and are compatible with most industrial deposition methods.


Characterization of arGO.
Detailed characterization of GO precursor material and arGO materials prepared at different conditions was performed in our earlier study. 3 Commercial Abalonyx graphite oxide, prepared by the modified Hummers method with C/O = 1.99 and 0.94 at.% of sulphur impurity (determined by XPS) was used as the precursor. Excluding oxygen added to the material as a sulphate the C/O ratio is corrected to 2.3. The graphite oxide was subjected to rapid explosive thermal exfoliation by inserting it into hot furnace (235°C) for 2 min. The powder rGO was then used to prepare high surface area a-rGO. 2,4 The a-rGO material showed no diffraction peaks in XRD and low oxygen content (C/O=55 determined by comparing areas of C1s and O1s) according to XPS, Figure S1. Specific Surface Area (SSA) determined using analysis of nitrogen sorption isotherms shows some batch-to-batch variation. The arGO batch used for preparation of electrodes in this study showed Specific Surface Area (SSA) of ~2580 m 2 /g by BET and ~2150 m 2 /g using QSDFT model ( Figure S2). The analysis of the nitrogen sorption isotherm using QSDFT slit pore model shows that almost 100% of pore volume is in pores with diameter below 4 nm.
S5 Figure S2. Analysis of nitrogen adsorption isotherms using the QSDFT slit pore model: cumulative pore distribution (red curve, right axis) and pore size distribution (black curve, showing traces of layered structure inherited from rGO and observed even after activation. Nanometer sized pores were also imaged from similar a-rGO samples using STEM in our earlier publication. . 3

S7
Hydrophobic nature of a-rGO was confirmed also using Dynamic Vapor Sorption (DVS) method. The sample of a-rGO similar to the used here for preparation of electrodes showed almost negligible BET SSA value of ~8 m 2 /g by water sorption. However, relatively large pore volume of 0.87 cm 3 g -1 was measured using full isotherm which exhibit the shape which points out to condensation of water inside the pores at P/P 0 above ~0.7. 3

Preparation of dispersions and electrodes.
The aqueous dispersions were prepared using sequential mixing of the components After drying, part of the electrode was scraped off the steel foil, dried in a vacuum oven and analyzed by nitrogen adsorption in order to estimate the actual surface area of the coated electrode. The thickness of electrodes tested in supercapacitors was 100-500 µm.
Typical samples were prepared using 2.13-2.21 mg/cm 2 loading. The weight of electrodes was measured after vacuum degassing at 130 0 C.  Nearly amorphous structure of a-rGO is expected for the material composed by defect graphene sheets assembled in disordered 3D network. Precursor graphite oxide shows relatively strong but broad (001) reflection. Dispersing graphite oxide provides graphene S13 oxide in solution. Re-stacking of graphene oxide after electrode coating is not significant as evidenced by absence of (001) peak, as it is expected considering relatively low amount of GO relative to other components.. Annealing at 200 0 C do not result in significant change of XRD patterns for both a-rGO and rGO based electrodes confirming absence of crystalline GO in the not heated samples.

S15
The TGA provide insight into the composition of a-rGO-based electrodes. The first weight loss (2.3%) step corresponds to thermal deoxygenation of GO. Complete oxidation and removal of carbon result sin 8.25% residual mostly due to SiO2.

Methods
The BET specific surface area was evaluated by analysis of nitrogen sorption isotherms acquired using Autosorb iQ XR analyzer by Quantachrome. A slit pore quenched solid density functional theory (QSDFT) equilibrium model was used to evaluate cumulative S16 SSA, pore volume and pore size distribution of the samples. SEM images were recorded using a Zeiss Merlin FEG-SEM microscope.
Electrochemical characterization of electrodes prepared using dispersions was specific current values of 0.5, 1, 2, 3, 5, 10, 20 and 40 A/g. Specific capacitance, energy S17 density and power density were estimated based on the galvanostatic discharge data using the standard procedure reported in the literature. [5][6][7] Specific capacitance, energy density and power density calculations.
Specific capacitance of a single electrode , the energy density of the whole electrode , stack and the power density for two electrode cell were calculated using procedure described below.

Linear profile:
In the case of a linear discharge profile, the capacitance of the electrodes stack can 2 be determined as where is the charge, Δ the measured voltage drop 2 = ∆ = ∆ S21 due to the discharge ( ), is the applied current in A and is the time elapsed during the discharge ( ). -On the example given Figure S10, the discharge takes place between 0.999 V and 3.96 in series in the device, the capacitance can be written as: Here therefore resulting in capacitance for one electrode: Specific capacitance is calculated by normalizing experimentally measured capacitance to the electrode weight. Here the total amount of electrode material (without considering S22 the steel supports) was 2.10 mg, meaning the mass of one electrode was 1.05 mg. In the given example, the resulting specific capacitance of the electrode is therefore 200 In practical applications, only the first part of the discharge (until ) is considered. In However, as highlighted in Figure 2, the capacitor discharge curve is not linear. Therefore, in order to increase the accuracy of the measurement, the integration of the discharge curve was used instead, leading to: The integration of the discharge curve from Figure 1 gives a value of 22.472 Vs, resulting in a capacitance for the stack of 0.095 F. Calculating the value for one electrode using equation 2 therefore gives a specific capacitance of 180 F/g.

S24
When considering a discharge until only , the integration of the partial discharge curve 2 gives a value of 15.507 Vs, resulting in a capacitance for the stack of 0.087 F, meaning the specific capacitance for the single electrode is 165 F/g.

Energy density and power density:
The energy density of the supercapacitor can be determined for each charge/discharge as: where is the equivalent distributed resistance and M is the mass of the electrode stack (here the total mass of the active material of both electrodes). When considering only the first half of the discharge, this equation then becomes = 1 1 It should be noted that in equations (4) and (5), the specific capacitance can be S25 When standard units are used, the energy density will be in J/g. Reported values are usually in Wh/kg, therefore the energy density needs to be divided by 3.6 in order to convert it to Wh/kg from J/g. When considering integration method, the obtained specific capacitance of 180 F/g results in an energy density of 22.5 J/g (6.24 Wh/kg). For the figures presented in this paper we used values obtained by integration method and full discharge curve.
The power density in W/kg is then derived from the energy density in Wh/kg using: where is the discharge time in seconds. For the example shown above, this results in charge-discharge curves at the current density of 1 A/g; note the much shorter discharge times and significantly higher voltage drop measured for the RGO-based electrodes; (e,f) specific capacitance vs current density, calculated using the galvanostatic discharge data.
Mechanical decomposition of the rGO-based electrodes was observed at higher current densities.

Additional information about rheological properties and stability of dispersions.
S32 Figure S16. The dispersions prepared without GO (arGO:SiO 2 :CNT= 10:1:1) are not stable and some precipitation is observed already after 5 minutes (a, b). In contrast, dispersions prepared using GO (arGO:GO:CNT=10:1:1) are more stable even without addition of SiO 2 (c,d,e). Some precipitation on the bottom of the vial was observed only after several hours of storage.