Nanostructured porous carbons with high rate cycling and floating performance for supercapacitor application

Nanostructured porous carbons with high rate cycling and floating performance for supercapacitor application F. O. Ochai-Ejeh,1 D. Y. Momodu,1 M. J. Madito,1 A. A. Khaleed,1 K. O. Oyedotun,1 S. C. Ray,2 and N. Manyala1,a 1Department of Physics, Institute of Applied Materials, SARChI Chair in Carbon Technology and Materials, University of Pretoria, 0028, South Africa 2Department of Physics, University of South Africa, Florida Science Campus, Florida, Johannesburg 1709, South Africa


II. EXPERIMENTAL DETAILS
A. Material synthesis 10 g of raw cork (Quercus Suber) was washed using acetone and deionized water and dried in an oven at 60 • C for several hours. Then the raw cork was pre-treated with 0.8 ml of 0.5 M sulfuric acid and 80 ml deionized water in a hydrothermal system at 160 • C for 12 hours. The hydrochar obtained was washed with deionized water several times and dried in an oven. The hydrochar was then ground and re-weighed and activated with potassium hydrogen carbonate (KHCO 3 ) in a ratio of 1:1. The activated hydrochar was carbonized under argon flow in a chemical vapour deposition system at 850 • C for 2 hours. The product obtained was then washed several times using 3 M HCl and deionized water and dried in an oven at 60 • C overnight.

B. Material characterization
Scanning electron microscopy (SEM) measurements were performed using Zeiss Ultra plus 55 field emission scanning electron microscope (FE-SEM) operating at an accelerating voltage of 2.0 kV. Energy-dispersive X-ray (EDX) analysis, high-resolution transmission electron microscopy (HRTEM) and selected area diffraction (SAED) analysis of the sample was measured using JEOL 2100 (Tokyo Japan) operated at 200 kV. X-ray photoelectron spectroscopy (XPS) measurements were performed using a Physical Electronics Versa Probe 5000 spectrometer. X-ray Diffraction (XRD) analysis was performed using XPERT-PRO diffractometer (PANalytical BV the Netherlands). Fourier transform infrared (FT-IR) spectroscopy was carried out using Perkin Elmer Spectrum RX I FT-IR system in the 4000 -500 cm -1 range with a resolution of 2 cm -1 to determine the sample surface functional groups. The FT-IR measurements were performed by properly blending the samples with KBr in a ratio of 1:100 and making them into transparent pellets before measurements. The Brunauer-Emmett-Teller (BET) technique was used to obtain the specific surface area and the N 2 adsorption/desorption isotherm measurements was performed at -196 • C with a Micrometrics TriStar II 3020. Raman analysis was performed using a Jobin-Yvon Horiba TX 64000 micro-spectrometer.

C. Electrochemical characterization
Three-and two-electrode measurements of the ACQS sample was performed in a multichannel VMP300 potentiostat/galvanostat (Biologic, France) workstation at room temperature. The electrodes for the three-electrode measurements were made by adding 80 wt.% ACQS material, 15 wt.% carbon black (to boost the conductivity of the material) and 5 wt.% polyvinylidene difluoride (PVdF) as a binder in an agate mortar and then mixed properly by adding 1-methyl-2-pyrrolidinone (NMP) drop-wise to make a paste. This was then pasted on nickel foam (NF)/carbon cloth current collectors and dried in an oven at 60 • C for several hours. The paste for the two-electrode tests was prepared using the same method but was coated on 16 mm diameter Nickel foam (NF) and dried at 60 • C, pressed and then assembled in a coin cell with a microfiber glass filter paper as a separator. The three electrode tests were performed in 3 M KNO 3 , 6 M KOH and 1 M H 2 SO 4 aqueous electrolytes using Ag/AgCl and glassy carbon as reference and counter electrodes respectively. The two electrode tests were then performed in 3 M KNO 3 neutral electrolyte because of the good reversibility of the material in both positive and negative voltage windows in this electrolyte. The specific capacitance, C s for the half-cell electrode in a three electrode configuration is given by equation (1) The specific capacitance, C sp per single electrode of the full device was calculated from the chargedischarge profile using equation (2): 3 Where I is the applied current (A), ∆t is the discharge time (s), ∆U is the maximum cell voltage (V), and m is the total mass of the electrodes (g). Four (4) is a normalization factor of the symmetric electrodes to the mass of a single electrode. The energy density, E d and power density, P d were calculated from equation (3) and (4) 3,28 (using parameters from equation (2)).

A. Material characterization
The SEM images of the ACQS material are presented in Fig. 1. Figs. 1(a) and (b) display the SEM images of the ACQS sample at low and high magnifications. The micrographs display 3D granular porous network structures of the material. The surface morphology of the material shows that KHCO 3 activation leads to the production of porous granular carbons which could improve the fast diffusion of ions along the material surfaces, useful for high-performance supercapacitor application. 23 The carbonization reaction mostly proceeds by the decomposition of KHCO 3 between the temperature ranges of 100 -400 • C in the following reaction 23 At this temperature, the surface area and porosity of the material might be increased or modified because the CO 2 gas interacting with the carbon material to create more pores or modify existing ones. 23,30 The morphology of the as-synthesized ACQS material was further studied using HRTEM and is presented in Figs. 2(a) and (b). The high and low magnifications HRTEM images of the sample show no lattice fringes which are an indication that ACQS is mostly amorphous carbon. This is supported by the SAED pattern (inset to Fig. 2(b) which shows halo ring with no diffraction spots indicating the amorphicity of ACQS material.
EDX spectrum for the ACQS material is presented in Figure 2(c). The EDX plot confirmed the high carbon content of the as-synthesized ACQS material. The summary of the elemental composition is presented in Table I.
The oxygen content recorded is linked to the functional group present in the sample. The Cu recorded in trace amount was a contribution from the copper grid. The structure of the ACQS sample was studied by XRD technique. The XRD pattern of the ACQS sample ( Fig. 3(a)) shows a wide and broad peak at 26 • (002) and a weak peak at 51 • (012). The diffraction peaks of the ACQS sample can be indexed to a hexagonal crystal structure of graphite having a space group of P63mc utilizing the most appropriate matching Inorganic Crystal Structure Database (ICSD) card #31170. The peaks represent graphitic diffraction planes, but the broad and low intensity of the peaks is an indication that the ACQS material is mostly non-crystalline and can be referred to as amorphous carbon.  FT-IR measurement was carried out to determine the presence of surface functional groups in the as-synthesized ACQS sample in the wavenumber range of 500 to 4000 cm -1 ( Fig. 3(b)). The bands occurring at ∼3435 cm -1 is as a result of the O-H stretching vibrations of water molecule due to surface hydroxyl groups. The peaks at ∼2942 cm -1 indicate the C-H bond of the aliphatic group while the peak at ∼1608 cm -1 is attributed to the C=O stretching vibrations of the carboxyl groups. The peaks occurring at ∼1050 cm -1 and ∼582 cm -1 may be due to the -C-C stretching vibration. 3,32 The ACQS sample composition was further analyzed with XPS. The high-resolution survey spectrum of the as-received ACQS sample (without sputter cleaning) is presented in Fig. 4(a) and a summary of the sample composition showing high carbon content of the as-synthesized ACQS material is displayed in Table II. Fig. 4(b) shows core level spectrum of C 1s with intensity peaks associated with the graphitic carbon C=C or C-C and other peaks relating to C-O, C-O=C or C=O. Figure 4c shows the spectrum for the O 1s with peaks at binding energies 531.7 eV and 533.2 eV associated with C-O and O-C indicating the presence of functional groups as suggested by FT-IR measurements.
The Raman spectrum of the ACQS sample is presented in Fig. 5(a). The spectrum shows the D band at ∼1338 cm -1 and the G band at ∼1587 cm -1 which are characteristic of disordered carbons in the sp 2 carbon network and graphitic carbon tangential vibrations respectively. 33,34 The ratio of the intensity of the D and G peaks (I D /I G ) expresses the extent of graphitization of the material. This ratio  for the ACQS sample is 0.89 which is an indication of a low amount of graphitic crystallinity [34][35][36][37] further confirming the observed XRD result. The broadband in Fig. 5(a) was deconvoluted using Lorentzian distribution for the curve fitting, (see Fig. 5(b)). The peak 1 curve is ascribed to the lattice vibrations linked to sp 2 -sp 3 network and peak 2 curve characteristically stems from the dissemination of unstructured carbon in interstitial sites in the disturbed graphitic lattice. 38 In a nutshell, the Raman spectrum indicates that the ACQS material contains a substantial quantity of interstitial unstructured carbon in the lattice framework.  N 2 adsorption-desorption isotherms of the ACQS sample studied at -196 • C is presented in Fig. 6. Fig. 6 shows typical type IV with an H4 hysteresis loop signifying materials consisting of a microporous and mesoporous structure with relative pressure (P/P o ) > 0.2 and SSA of 1056.52 m 2 g -1 and pore volume of 0.64 cm 3 g -1 . The pore size distribution (PSD) (inset to Fig. 6 is in the range of 2-10 nm. The high SSA of the material and the PSD narrow range coupled with the granular microstructures can provide more active sites that ensure that the ACQS surface is sufficiently exposed for the electrolyte ions to access the micropores of the material. 39,40

B. Electrochemical analysis
The CV plots of the ACQS electrodes tested in the three-electrode configuration in 6 M KOH, 1 M H 2 SO 4 and 3 M KNO 3 aqueous electrolytes at 50 mV s -1 scan rate are shown in Fig. 7. Rectangular shaped CV profiles are observed for the ACQS sample in the positive and negative potential windows of 0.0 to 0.9 V and -0.9 to 0.0 V in 3 M KNO 3 aqueous neutral electrolyte, which are typical of electric double layer capacitive (EDLC) performance of the ACQS sample. It can be observed that the material is able to operate reversibly with typical rectangular EDLC curves and is stable within the positive and negative potential windows with di-hydrogen and oxygen evolutions only observed as the potential is increased to -1.   Fig. S2 of supplementary material). The disparity observed in electrochemical performances of the ACQS material in the various electrolytes could be due to the pH of the different electrolytes and/or pseudo-capacitance contribution from surface functional groups. 26,42,43 For example, in 3 M KNO 3 neutral electrolyte, the H + and OHions are fewer as compared to those in the acidic and alkaline electrolytes, and the pH is balanced on the H + /OHions such that the ions of the electrolyte are not biased to react with the electrode material in any preferred potential. 3,26,44 Whereas the acidic and alkaline electrolytes have an abundance of the OHand H + ions which can make the material have a biased reaction in a given potential. 44 Other parameters such as the sizes of the ionic radii, ionic conductivities, and radii of the ionic hydration spheres and the mobility of the ions of the electrolytes may play a role as they interact with the electrode material. 2,15,42,44 These electrolyte properties are itemized in Table III. From the Table III, it can be observed that the radius of hydration spheres for the electrolytes in any particular potential, decreases in this order:  for the higher specific capacitance of the material in the H 2 SO 4 and KOH electrolytes. However, the ACQS material is only able to operate reversibly in the positive and negative potential windows in the 3 M KNO 3 electrolyte, hence its utilization in the symmetric device fabrication. The Charge-discharge (CD) curves measured in the positive and negative potential windows of 0.0 to 0.9 V and -0.9 to 0.0 V, respectively in 3 M KNO 3 at 1.0 A g -1 to 5.0 A g -1 specific current are presented in Figs. 8(a) and (b). The CD profiles show triangular symmetrical shapes corresponding to the characteristic EDLC behavior observed for the CV curves.
The Specific capacitance (C S ) values calculated from the CD profiles in the positive potential window are from 133 to 94.0 F g -1 and in the negative potential window, from 167 to 139 F g -1 , at 1.0 to 5.0 A g -1 , respectively. The CD profiles for the test performed at specific currents of 1.  Fig. 8(e).
A symmetric device was assembled based on the performance of the material in the threeelectrode configuration. The CV plots of the the ACQS//ACQS symmetric device fabricated from ACQS electrodes in 3 M KNO 3 at various voltage windows ranging from 1.4 to 1.8 V at 25 mV s -1 scan rate is presented in Fig. 9(a). The optimization of the working voltage windows is to obtain the most stable voltage window of the symmetric device. It has been shown in previous reports that neutral electrolytes are able to function with large operating potentials as a result of high over potential for di-hydrogen evolutions. 7,9,26,44 The CV curves displayed rectangular shapes within the operating voltages but a current leap begins to be observed as the voltage is increased to 1.8 V. This anodic current leap towards the more positive voltage might be as a result of oxygen or gas exudation as the voltage reaches 1.8 V. 45 However functional groups present in the sample can enhance the ion diffusion culminating in active electron mobility in the granular porous carbon sample. Hence, the working voltage of 1.8 V is stable for the symmetric device. Fig. 9(b) shows the CV plots of the symmetric device in the voltage window of 0 -1.8 V at scan rates of 10 to 100 mV s -1 . The CV curves maintain symmetric rectangular shapes as the scan rate increases up to 2.0 V s -1 as shown in Fig. 9(c). This is an indication of a highly capacitive electrochemical performance of the ACQS material. 14,45 The CD profiles recorded within the voltage window of 0.0 -1.8 V at varying specific currents of 0.5 A g -1 to 5.0 A g -1 are presented in Fig. 9(d). Linear CD profiles are observed which corresponds to typical EDLC behavior showing high reversibility of anodic and cathodic ionic transport.
The plot of the C sp values calculated from the CD profiles versus specific currents for the ACQS//ACQS symmetric device is shown in Fig. 10(a). A C sp value of 122 F g -1 is recorded for the  electrode at 0.5 A g -1 and 100 F g -1 at 5.0 A g -1 . A good rate capability is observed for the symmetric device as it displays more than 90% capacitance retention for each increase in specific current. This shows the ability of the material to retain reasonable capacitance even at high specific current. A plot of energy density versus power density (Ragone plot) for the symmetric device is shown in Fig. 10(b). The device displays an energy density of ∼14 W h kg -1 and power density of ∼450 W kg -1 at a specific current of 0.5 A g -1 (as determined from equation 3 and 4). This work displays remarkable electrochemical performance relative to other biomass-derived activated carbon. 6,18,46-50 A correlation of this results with other earlier reports from biomass-derived activated carbon (AC), which was activated with KOH, ZnCl 2 and MgO, is shown in Table IV. The high electrochemical performance of this device might be as result of the high SSA, high pore volume and large mesopore volume which resonate well with the granular porous network and good pore size distribution with good packing properties which enhance proper ion transfer. 1,18,51 The stability of the symmetric device studied at 5.0 A g -1 and 1.8 V maximum cell working voltage over 10,000 constant chargedischarge cycles is presented in Fig. 10(c). The device displayed ∼100% capacitance retention after the 10,000 th charge-discharge cycle with no degradation observed in the cell performance. The device was subsequently subjected to a voltage holding test to further investigate the long-term stability of the cell. Voltage holding as compared to the normal charging and discharging over many cycles helps to study the cell stability and shows the real resistance effects and physiochemical deterioration that might be taking place during the electrochemical process 41 after floating at 1.8 V maximum voltage. The voltage holding plot which is a plot of specific capacitance as a function of voltage holding time is presented in Fig. 10(d). The floating process involves periodic potentiostatic mode and subsequent galvanostatic charge-discharge cycle at 1.0 A g -1 . The floating and charge-discharge cycles are iterated for a period of 120 h. The voltage holding analysis showed an improvement in the performance of the device which increases after the first 10 h of floating which is an indication that more pore sites where accessed by the electrolyte as a result of prolonged floating, with only slight decay at the end of 120 h. The increase in capacitance might also be due to the expansion of the material granular porous network thereby allowing more adsorption and intercalation of ions into the material active sites thus increasing the cell capacitance. 35,52 The electrochemical impedance spectroscopy (EIS) measurements were carried out in an open circuit potential and the Nyquist plots for the ACQS//ACQS symmetric device before and after stability is presented in Fig. 11(a). The EIS study is important for understanding the electron and ion transport in the electrolyte and the electrode material. 21 The EIS study was performed in an open circuit potential from 0.01 Hz to 100 kHz frequencies. Nearly vertical lines and small semicircles (inset to Fig. 11(a)) can be observed at the high and low-frequency regions of the Nyquist plots indicating the presence of charge transfer resistance R ct of 0.2 Ω showing good electron and ion conductivity and typical electric double layer capacitive behavior. R ct of 0.5 Ω was obtained after cycling stability test, this large increase of the R ct value after stability could be attributed to the fact that electron transfer and ion diffusion within the electrode material/electrolyte has slowed down as a result of the prolonged cycling of the electrode. 40 The intersection of the semicircle with the Z' axis shows the solution resistance (R s ) before and after cycling stability, with values of 0.75 Ω and 0.74 Ω respectively. This R s element is made up of the electrolytic resistance, intrinsic electrode material resistance and the resistance between the electrode material and the current collector. 35,53 The equivalent series circuit fitted for the Nyquist plot before stability is as presented in Fig. 11(b). The equivalent series circuit shows a series connection between the solution resistance, Rs and the constant phase element, Q 1 which is connected in parallel with R ct, the charge transfer resistance. The Q 1 element shows the double layer capacitance associated with the Warburg diffusion element (W ) which is connected in series with R ct . Q 2 is the mass capacitance and is associated with the ideal polarizable capacitance connected in parallel to an impedance characteristic element R L , which accounts for the shifting of the vertical line from the typical EDLC characteristic. 54,55 The deviation from the typical vertical behavior connotes that there is a leakage resistance, R L which is a resistive element connected with Q 2 .
The impedance (Z) of Q 1 , is expressed as; 55,56 Where e and b are frequency-independent constants and ω is angular frequency. C (ω) and C (ω) which are the real and imaginary parts of the capacitance versus frequency is presented in Fig. 11(c). A value of 1.38 F was recorded for C at a frequency of 10 mHz is which is the real attainable capacitance of this symmetric device at this frequency. C describes the frequency transition between a typical capacitive and resistive characteristics of the symmetric cell 57 with a relaxation time (τ) of 1.32 s corresponding to a frequency of ∼1.0 Hz which was calculated for the symmetric cell using the equation: The relaxation time simply means that the stored energy contained in the ACQS//ACQS symmetric cell can be retrieved in a period of 1.32 s. The phase angle of the cell versus frequency is presented in Fig. 11(d). The phase angle of the cell is recorded at approximately -83 • (which is close to -90 • ) and a value of -81 • was recorded for the phase angle after cycling stability, which is indicative of the good capacitive performance of the symmetric cell similar to the ideal device.

IV. CONCLUSIONS
Activated carbon derived from cork (Quercus Suber) (ACQS) was successfully produced via a two-step environment-friendly synthesis route. This green synthesis route makes the synthesized material less toxic for utilization as an electrode material for energy storage application. The activated carbon has well-defined microporous and mesoporous structures which provide a good interface for fast rate ion transfer. The granular network structures also provide good packing properties for better ion diffusion and intercalation. The electrochemical study of the symmetric device assembled from the synthesized activated carbon as electrode materials displayed a specific capacitance of 122.2 F g -1 , energy density of ∼14.0 Wh Kg -1 and power density of 450.0 W kg -1 at a current density of 0.5 A g -1 with an excellent long-term cycling life after 10,000 CD cycles and subsequent floating at maximum voltage for 120 h. The excellent stability after floating test and the ∼100% capacitance retention after 10000 CD cycles makes the Cork derived activated carbon material a potential material for energy storage application.

SUPPLEMENTARY MATERIAL
See supplementary material for additional information on the electrochemical characterization of the ACQS material.