Effects of Pore Widening vs Oxygenation on Capacitance of Activated Carbon in Aqueous Sodium Sulfate Electrolyte

The commercial activated carbon has a relatively low speci ﬁ c capacitance in the Na 2 SO 4 electrolyte, which hinder the development of asymmetrical supercapacitors with high voltage. Re-activation and oxidative etching methods were applied to change the pore structure of activated carbon, respectively, to study the capacitive behavior of carbon in the Na 2 SO 4 electrolyte. The pore distributions combining with capacitive properties deduce that 0.85 nm is the threshold diameter of the ion-accessible micropores for hydrated Na + and SO 42 − . The speci ﬁ c capacitances of both the carbon materials by re-activation and oxidative etching methods are increased by 40%, in comparison with the commercial activated carbon. The enhanced capacitive performances of the carbon materials were mainly attributed to the increased ion-accessible speci ﬁ c surface area and pseudocapacitance, respectively. The oxidative etching is a more facile and economical method for practice application. Combining with MnO 2 as the positive electrode, the asymmetrical supercapacitor with a high voltage of 1.8 V exhibits a maximum speci ﬁ c cell capacitance of 50 F g – 1 and speci ﬁ c energy of 22.5 Wh kg – 1 .

Supercapacitors, one of energy storage devices, possess higher energy density than conventional electrostatic capacitors and higher power density than secondary batteries. They can be used to complement or replace secondary batteries in the energy storage field, such as smart grid, electric vehicles, load cranes and emergency doors. [1][2][3] Activated carbons which are based on electrical double layer (EDL) mechanism without faradic reaction have the unique advantages of long lifetime, high power density and low cost. [4][5][6] Nowadays, they have been widely employed as commercial electrode materials for supercapacitors. [7][8][9] The main efforts in current research on supercapacitors focus on increasing their specific energy, which is proportional to the specific capacitance and the square of the working voltage. The working voltage is mainly associated with the property of electrolyte. 10,11 Although organic electrolytes have higher achievable voltages, aqueous electrolytes have higher electrical conductivity, better safety and environmental benignity and lower cost. The conventional aqueous electrolytes for AC are H 2 SO 4 and KOH solution, which have only 1 V cell voltage. Recently, much less corrosive neutral Na 2 SO 4 began to be used as the aqueous electrolyte for symmetrical or asymmetrical supercapacitors, the potential window of which reaches 1.6-2 V. [12][13][14] The great extension of the potential window in Na 2 SO 4 electrolyte makes it possible to increase specific energy of aqueous supercapacitors. 15 However, the same commercial AC generally has a much lower specific capacitance in the neutral Na 2 SO 4 electrolyte than in those acidic or alkaline ones. 16,17 So far, the relationship between the pores of carbon and solvated ions of organic electrolytes has been studied, 18 which asserted that the ions with highly distorted solvated shell could be squeezed into the pores.
However, similar studies in the aqueous Na 2 SO 4 electrolyte was rarely reported. 19 Therefore, it is essential to fully understand the capacitive behavior of carbon electrodes in order to improve their specific capacitances in Na 2 SO 4 based electrolytes.
In this work, we employed two methods to change the pore structure of a commercial activated carbon (AC, YP50f, Kuraray Group) by re-activation and oxidative etching with Zn(NO 3 ) 2 and KMnO 4 , which were denoted as Zn-AC and Mn-AC, respectively. Both modified samples possessed specific capacitances of 132 F g -1 , which were higher than 94 F g -1 for the commercial AC at 1 A g -1 in Na 2 SO 4 . However, the pore distributions of Zn-AC and Mn-AC were different. The relationship between pore and ions of Na 2 SO 4 was discussed through the specific capacitance. As we know, reactivation with Zn(NO 3 ) 2 needs annealing at high temperature above 500°C. Therefore, oxidative etching with KMnO 4 at room temperature is much more facile and economic approach to improve the specific capacitance of carbon for practical application. 20 Based on the Mn-AC and MnO 2 , the asymmetrical supercapacitors in the Na 2 SO 4 electrolyte exhibited 22.5 Wh kg -1 of specific energy at 0.5 A g -1 .

Experimental
Preparation of samples.-Re-activation.-1 g AC and 3.6 g Zn(NO 3 ) 2 were dissolved in 10 ml water and sonicated for 15 min and aged for 18 h. The extra Zn(NO 3 ) 2 was removed by vacuum filtration, then the left solid was dried at 90°C for 3 h and annealed at 500°C for 1 h in Ar. The product was dispersed into 1 ml hydrochloric acid in 100 ml water and subsequently stirred for 30 min. Finally, the Zn-AC sample was obtained by centrifugation, washing and desiccation.
Oxidative etching. 0.08 g AC and 0.164 g KMnO 4 were dissolved in 27 ml water and then aged for 5 days. Characterization of samples.-The CO 2 and N 2 adsorption/ desorption isotherms were obtained at 273 and 77 K using automatic volumetric adsorption instruments (ASAP 2020 V3.04 H and ASAP 2460 2.02). The morphology and elemental contents were investigated by scanning electron microscopy (SEM, FEI Nova 400 Nano SEM) with X-ray energy disperse spectrum (EDS). The Mn content was determined via inductively coupled plasma-optical emission spectroscopy (ICP-OES) using IRIS Advantage ER/S (Thermo Elemental). The X-ray diffraction (XRD) patterns were collected by an X' pert Pro MPD diffractometer with Cu Ka radiation. Static contact angle measurements by placing a droplet of water or 1 mol l −1 Na 2 SO 4 solution on the surface of the electrodes consisting of one of the carbon samples and polyvinylidene difluoride (PVDF) with a weight ratio of 9:1.
Electrochemical measurements.-Carbon electrodes were fabricated by pasting mixtures of one of the carbon samples, acetylene black and PVDF with a weight ratio of 8:1:1 on substrates. MnO 2 electrodes were prepared by the same procedure with a ratio of 7:2:1. The accurate weight of the electrodes was read by a high-precision balance (MS105DU, Mettler Toledo, d = 0.00001 g). For a threeelectrode cell, the carbon or MnO 2 electrode was used as a working electrode, while a platinum foil and a saturated calomel electrode (SCE) were used as a counter and reference electrode, respectively, in 1 mol l −1 Na 2 SO 4 electrolyte. Further, the Mn-AC and MnO 2 electrodes were assembled to construct asymmetrical supercapacitors. All cyclic voltammetry (CV) and galvanostatic charge discharge (GCD) measurements were performed by electrochemical workstation CHI660E and the cyclic performances were measured by a LAND CT2001A system.

Results and Discussion
The commercial AC had a Brunauer-Emmett-Teller (BET) specific surface area (SSA) of 1944 m 2 g -1 , giving rise to a specific capacitance of 150 F g -1 in 6 mol l −1 KOH electrolyte ( Fig. S1 is available online at stacks.iop.org/JES/167/040524/mmedia). However, only 94 F g -1 of specific capacitance was obtained when 1 mol l −1 Na 2 SO 4 was used. In order to desire the high-energy asymmetrical supercapacitors, it was essential to study the capacitive behaviors of the different carbon samples based on the AC in the Na 2 SO 4 electrolyte. The Zn-AC and Mn-AC could be obtained by re-activation and oxidative etching of the AC with yields of 89.9% and 96.9%, respectively. The EDS analyses exhibited that the O contents of the AC and Zn-AC were 4.3 and 5.8 at% while that of the O content of the Mn-AC 8.0 at%. Moreover, the EDS of the Mn-AC showed no Mn element. The ICP-OES was further applied to check the content MnO 2 produced by reaction between C and KMnO 4 21 after washing with HCl/H 2 O 2 , revealing that the Mn weight content of Mn-AC is below 0.001%. The EDS and ICP-OES results illuminated that the O content obviously increased by oxidative etching method with KMnO 4 and finally MnO 2 was completely removed by HCl/H 2 O 2 .
The N 2 and CO 2 adsorption/desorption isotherms of the AC, Zn-AC and Mn-AC are shown in Fig. 1a and its inset. Obviously, all the samples exhibit type I adsorption desorption isotherms, confirming that the microporous structures dominated. The AC had the maximum BET SSA of 1944 m 2 g -1 and total pore volume of 1.08 cm 3 g -1 . The SSAs of the Zn-AC and Mn-AC decreased to 1854 and 1329 m 2 g -1 , respectively. Strictly, the BET method is not applicable to micropores, and hence the obtained BET SSA should be considered as an apparent area. 22 However, total pore volumes of Zn-AC and Mn-AC also declined to 1.05 and 0.71 cm 3 g -1 , respectively. It was concluded that the oxidative etching decreased the SSA and pore volume more severely. Figure 1b displays the results of pore distribution from CO 2 and N 2 adsorption by applying a non-local density functional theory (NLDFT) kernel, in which CO 2 adsorption giving the ultramicropore distribution ranging from 0.45 to 0.93 nm while N 2 from 0.93 to 3.00 nm. The AC has multimodal pore distribution with the typical maxima at 0.84, 0.87 and 1.18 nm. After reactivation, pores at 0.84 nm for the AC disappear, instead, pores at 0.86 nm appear and the differential volume increases to the highest value of 2.05 cm 3 g -1 nm -1 of the three carbon samples, indicating that re-activation broaden the partial ultramicropores of the AC from 0.84 to 0.86 nm within which Zn 2+ was absorbed. The Mn-AC has a different porous structure from the Zn-AC and AC, which shows a maximum at 0.89 nm with a lower differential volume of 1.02 cm 3 g -1 nm -1 and absence of pores at 0.87 nm for the AC, suggesting that KMnO 4 consumed C within and around the dominant micropores.
In order to detect the electrolyte wettablity of the carbon materials, the Static contact angle was measured. It is found that the water contact angles of the AC, Zn-AC and Mn-AC are 121°, 125°, 101°and the Na 2 SO 4 contact angles are 133°, 137°, 124°, respectively, as shown in Figs. S2 and S3. Overall, the wettability of the three carbon materials were comparable. It was also concluded that the changed pore distribution by re-activation slightly diminished the wettability for the Zn-AC and the increased O content by oxidative etching slightly promoted the wettability for the Mn-AC, comparing with the original AC.
The electrochemical properties of the AC, Zn-AC and Mn-AC were estimated by CV and GCD measurements using three-electrode cells in the potential range from -1.0 to 0.6 V vs SCE in the Na 2 SO 4 Figure 1 (a) N 2 adsorption-desorption isotherms with CO 2 adsorption-desorption isotherms, and (b) pore distributions of AC, Zn-AC and Mn-AC. electrolyte. Figure 2a describes their cyclic voltammograms (CVs) at a scan rate of 10 mV s -1 , featuring quasi-rectangular shape without obvious redox peaks. Obviously, the AC exhibited the smallest area among all, suggesting that both the Zn-AC and Mn-AC had superior specific capacitances. Accordingly, Fig. 2b shows the GCD curves for the three samples. Considering that coulombic efficiency should not be beyond 100%, the specific capacitances were calculated from the curves with shorter time. 23 Both the Zn-AC and Mn-AC had the same specific capacitance value of 132 F g -1 at 1 A g -1 , which was 40% higher than 94 F g -1 for AC at the same specific current. On the other hand, the densities of the AC, Zn-AC and Mn-AC were calculated to be 0.66, 0.56 and 0.74 g cm −3 respectively according to the thickness (Fig. S4) and the weight of the electrode. 24 Therefore, the volumetric capacitances of the AC, Zn-AC and Mn-AC were determined to be 62, 74 and 98 F cm −3 respectively. The highest volumetric capacitance of the Mn-AC combining with its higher yield suggested that the oxidative etching method was more beneficial than the re-activation for practice application.
Different carbon materials based on the commerical AC with changed pore structure provided good objects to study the relationship between pore and ions of the Na 2 SO 4 electrolyte. According to the literatures, 25,26 the radii of hydrated Na + and SO 4 2are 0.358 and 0.379 nm, respectively. In Fig. 1b, the orange dash line at 0.76 nm corresponds to the diameter of hydrated SO 4 2-. For reactivation, as mentioned above, the wettability of the Zn-AC was even slightly worse than the AC, which ruled out that the highly distorted solvated shell was merely squeezed into the ultramicropores in the Zn-AC rather than the AC. As the specific capacitance of the Zn-AC was evidently higher than that of the AC, meant that the differential volume (sum) of the ion-accessible micropores of the Zn-AC were higher than that of the AC. According to the pore distributions, it is only possible that the micropores with the diameter beyond 0.85 nm can provide efficient SSA for hydrated Na + and SO 4 2electrolyte ions, corresponding to the magenta dash line in Fig. 1b. Between 0.85 and 0.93 nm, the differential volume of the Zn-AC is 1.32 times as that of the AC, which is close to 1.40 times for specific capacitance. Therefore, the enhanced capacitive behavior of the Zn-AC was mainly attributed to the increased ionaccessible SSA arising from the more reconstructed efficient ultramicropores after re-activation. For oxidative etching, the contact angle of the Mn-AC was comparable to the orignal AC, especially for Na 2 SO 4 , while increased O content of the Mn-AC would certainly lead to lower conductivity, comparing with the AC. Hence, it could be ruled out that the highly distorted solvated shell was merely squeezed into the ultramicropores in the Mn-AC rather than the AC. Moreover, although the Mn-AC had the dominant ultramicropores with larger size of 0.89 nm, the corresponding differential volume is lower than that of AC. Thus, the enhanced capacitive behavior of the Mn-AC was mainly assigned to psudocapacitance arising from a larger oxygen content (8.0%) in the Mn-AC comparing with 4.3% for AC. The CV cathodic peak of the Mn-AC at −0.015 V in the Fig. 2a also indicates the existance of psudocapacitance. All in all, it was concluded that the threshold diameter of the ion-accessible micropores in the three carbon samples for hydrated Na + and SO 4 2− was 0.85 nm, meaning that that highly distorted solvated shell squeezed into the ultramicropores was ruled out.
Despite the same specific capacitances obtained by these two methods, oxidative etching had an obviously advantage of convenience and low cost, because the etching course occurred at room temperature, avoiding the high temperature re-activation. More importantly, due to its high yield and high volumetric capacitance, the Mn-AC was worth being further studied as the working electrode in three-electrode cells and negative electrodes in asymmetrical supercapacitors for pratice application. Figure 2c shows the CVs of the Mn-AC at different scan rates from 10 to 100 mV s -1 , indicative of a stable potential window of 1.6 V in 1 mol l −1 Na 2 SO 4 . It is observed that the distortion of the rectangle occurs at the fast scan, which indicates the electrode resistance is larger, causing the smaller actual potential than applied externally potantial. 27 According to the GCD curves in Fig. 2d, the specific capacitances of the Mn-AC were calculated to be 132 F g -1 at 1 A g -1 and 94 F g -1 at 5 A g -1 , respectively.
The positive electrode materials of the asymmetrical supercapacitors were prepared by a reaction between Mn 2+ and Mn 7+ . The XRD pattern of the positive electrode material is shown in Fig.  S5. The two broad peaks at 37°and 65.7°are indexed to ε-MnO 2 (JCPDS: 089-5171) with poor crystallinity. The EDS analysis revealed that the atomic ratio of Mn, O and K was 1:1.92:0.08, consisting with the stoichiometry of MnO 2 . Figure 3a shows the SEM images of MnO 2 , which exhibits ball-like morphology with an average diameter of 200 nm. The rectangular CVs are obtained for the MnO 2 at the scan rates of 10 and 100 mV s -1 with the potential range from 0 to 0.8 V vs SCE in the 1 mol l −1 Na 2 SO 4 electrolyte (Fig. 3b), demonstrating that MnO 2 also had a good capacitive behavior. According to the GCD curves in Fig. S6, the specific capacitance of MnO 2 reaches 230 F g -1 at 1 A g -1 .
In the three-electrode cells, the Mn-AC and MnO 2 electrodes could be measured within potential ranges of -1 to 0.6 V and 0 to 0.8 V vs SCE at a scan rate of 10 mV s -1 (Fig. S7), respectively, suggesting that the asymmetrical (−) Mn-AC | 1 mol l −1 Na 2 SO 4 | MnO 2 (+) (Mn-AC | MnO 2 in short) supercapacitors had 1.8 V of potential window. For asymmetrical cells, the amount of charge Q stored in each of the positive and negative electrodes must be the same, and is governed by the following Eq. 1, where C p and C n are the capacitance, and U p and U p are the working potential range of  the positive and negative electrodes in the cell. [28][29][30] The negative and positive working potential ranges were respectively from −1 to 0 V and 0 to 0.8 V vs SCE, the voltage ratio U n /U p was determined to be 1.25. Thus, the optimal capacitance ratio C p /C n was also 1.25 in the asymmetrical supercapacitor. The asymmetrical two-electrode supercapacitors were constructed using the ε-MnO 2 , Mn-AC and 1 mol l −1 Na 2 SO 4 as positive and negative electrode and aqueous electrolytes. Figures 4a and S8 show CVs of the asymmetrical supercapacitor with a voltage of 1.8 V at different scan rates. It was worth noting that the asymmetrical supercapacitor exhibited ideal capacitive behavior with a nearly rectangular shape even at a high scan rate of 500 mV s -1 . Figure 4b shows the GCD curves of the asymmetrical Mn-AC | MnO 2 supercapacitor at different specific current densities. The specific capacitances of the asymmetrical supercapacitor were 50 F g -1 at 0.5 A g -1 and 29 F g -1 at 10 A g -1 according to the discharge curves, respectively. Therefore, the maximum specific energy for the Mn-AC | MnO 2 supercapacitor was calculated to be 22.5 Wh kg -1 based on the total mass of active materials on the two electrodes, which was 22% higher than 18.5 Wh kg -1 for the AC | MnO 2 supercapacitor. In addition, 9000 W kg -1 of the specific power was obtained when the Mn-AC | MnO 2 supercapacitor kept a reasonable specific energy of 13.1 Wh kg -1 while 8804 W kg -1 when the AC | MnO 2 supercapacitor 11.3 Wh kg -1 (Fig. 4c). The superior performance of the Mn-AC | MnO 2 supercapacitor was completely assigned to higher specific capacitance of Mn-AC. The cyclic performances of the asymmetrical cell are shown in Fig. 4d. 71.8 and 66.7% of the initial capacitance could be retained at 1 A g -1 after 3000 and 10000 cycles, respectively. The capacitance fading may be mainly due to the potential polarization on the MnO 2 electrode.

Conclusions
In summary, the porous structure of the commercial AC has been changed by re-activation with Zn(NO 3 ) 2 and oxidative etching with KMnO 4 , respectively. Re-activation broadened the partial ultramicropores of the AC by absorbing Zn 2+ within them. Oxidative etching severely decreased the SSA and pore volume but increased the O content. The specific capacitances of the Zn-AC and Mn-AC were increased to 132 F g -1 from 94 F g -1 for the original AC. Combining the pore distribution with specific capacitances of the three carbon samples, it was inferred that the threshold diameter of the ion-accessible micropores in the three carbon samples for hydrated Na + and SO 4 2− was 0.85 nm, meaning that that highly distorted solvated shell squeezed into the ultramicropores was ruled out. The enhanced capacitive behaviors of the Zn-AC and Mn-AC in 1 mol l −1 Na 2 SO 4 were mainly attributed to the increased ionaccessible SSA and pseudocapacitance, respectively. For pratice applicaiton, the oxidative etching with KMnO 4 at room temperature was much more facile and economic approach to improve the specific capacitance of carbon. The asymmetrical Mn-AC | MnO 2 supercapacitors assembled in 1 mol l −1 Na 2 SO 4 aqueous electrolyte had a high operating voltage of 1.8 V. A maximum specific capacitance of 50 F g -1 and specific energy of 22.5 Wh kg -1 was obtained at 0.5 A g -1 . Moreover, a high specific power of 9000 W kg -1 was achieved at 13.1 Wh kg -1 . Therefore, it is believed that either increasing the efficient micropores or the O content could improve the capacitive behavior of carbon in the Na 2 SO 4 electrolyte. Oxidative etching method is also a general and low-cost method comparing with re-activation for carbon based supercapacitors.