Cellulose-based activated carbon aerogels as electrode materials for high capacitance performance supercapacitors

Abstract

Carbon aerogels (CAs), as a low-cost and high-performance carbon material, have attracted wide interest as electrode materials for supercapacitors (SCs). Here, a CA was built as a highly interconnected lamellar network by activation. The activation process was achieved using different ratios of activators (ZnCl₂ and KOH) mixed with them and then carbonized. CA-KOH-3 has greatly improved the specific surface area from 616.97 m² g⁻¹ to 1709.62 m² g⁻¹. It also showed the highest capacitance performance, which could reach 213.53 F g⁻¹, and an SC assembled with it showed good energy density performance. Meanwhile, the specific capacitance exceeded its starting value even after 5000 times of charging and discharging. The research shows that a high proportion of KOH in a certain range is conducive to etching micropores and increasing the specific surface area, thus improving the capacitance characteristics. This provides an option for effectively adjusting the porous structure of CA, thereby increasing the specific surface area and improving the capacitance performance.

---

1. Introduction

As a new type of energy storage device, supercapacitors (SC) not only offer the characteristics of fast charge and discharge of capacitors but also the energy-storage characteristics of batteries.¹ Electrochemical SCs can be divided into electric double layer capacitors (EDLCs) and pseudo-capacitors.^2,3 Compared with pseudo-capacitors, EDLCs have better cycle stability, higher reversibility and a longer cycle life.⁴ As a result, they have attracted increasing attention. Electrode material is a key component affecting the performance of SCs.⁵ Porous carbon materials are ideal electrode materials for EDLCs due to their high specific surface area, high electrical conductivity, low cost and optimized pore structure.^6,7 Therefore, various carbon materials, such as activated carbon, carbon nanotubes, carbon aerogels (CAs), and graphene, have attracted wide attention in recent years.⁸ Preparations of biomass-based carbon materials have the advantage of efficient treatment of biomass waste, and at the same time, industrial cotton waste,⁹ bamboo,¹⁰ peanut shell,¹¹ sisal leaf,¹² and lignin¹³ have been developed and tested for high-performance EDLC applications. Cellulose, the main component of plant biomass (more than 50%), is the most abundant and widely used biopolymer in nature,¹⁴ and it can provide sustainable carbon precursors for energy storage. CAs derived from cellulose precursors have demonstrated exceptional versatility across various advanced applications owing to their unique porous structures and functional surfaces. For instance, bio-carbon aerogels driven by dual-functions of solar and electric energy exhibit an ultrahigh light-absorption efficiency and significantly enhanced evaporation rates.^15,16 Nitrogen-doped cellulose-based carbon aerogel/graphene composites have shown outstanding electrochemical performance in supercapacitors, while cellulose-based aerogels designed for CO₂ capture and energy storage offer high adsorption capacities and efficiency.^17,18 High-barrier functional materials with multi-level reinforced structures,¹⁹ along with breathable silk-thread sensors that combine sensitivity, scrub resistance, and durability, further illustrate the potential of cellulose-based carbon aerogels in diverse fields ranging from energy storage and environmental protection to wearable electronics.²⁰ However, the porous carbon produced from biomass is greatly affected by impurities and the uncontrollability of pore structure limits its application.²¹ The ideal carbon material for SC electrodes should have a multi-stage porous structure. Micropores can provide high capacitance, while mesopores and macropores can provide ion-transport channels and reduce diffusion resistance.²²

CA is a kind of carbon material with a three-dimensional layered porous structure, with macropores, mesopores, and micropores, a high specific surface area, and good electrical conductivity and cyclic stability, so it is an ideal electrode material for SCs. Its layered structure enables ions to diffuse effectively through the carbon network.

In order to further improve the pore development of CA, increase the specific surface area of the material, and improve the capacitance performance of the electrode material, chemical activation can be applied by etching with an activator in high-temperature inert gas. Compared with physical activation, the graded porous carbon materials in chemically activated plant biomass can obtain a more microporous structures through etching and developed a higher specific surface area, thereby enhancing the electrochemical performance of the electrode materials. The yield of carbon under chemical activation is higher and the reaction temperature is lower than that under physical activation, and the preparation process is also simpler. The effect of chemical activation is influenced by the activator type, activator proportion, activation temperature, and activation time. By regulating these factors, graded porous carbons with various pore structures can be prepared, which makes chemical activation more flexible. Fischer et al.²³ prepared a SC using cellulose acetate as the precursor, and the specific surface area of the activated material could reach 1200 m² g⁻¹. Furthermore, in 4 M KOH, the active carbon electrode provided a specific capacitance of 186 F g⁻¹ at a scanning speed of 10 mV s⁻¹ and a high energy density of 40 W h kg⁻¹ at a power density of 15 kW kg⁻¹. Lima et al. explored biochar activated by KOH and ZnCl₂ as a template for preparing electrodes for supercapacitors.²⁴ The results showed that the devices made of ZnCl₂ have a lower internal resistance and higher phase angle, which was related to the higher mesoporous degree and the distribution of Zn residues in the matrix. The area capacitance of the biochar electrode activated by ZnCl₂ was 342 mF cm⁻², which was higher than that of the biochar electrode activated by KOH (138.49 mF cm⁻²).

In this study, we used two activators, KOH and ZnCl₂, to activate CA in three proportions: 1 : 2, 1 : 1, and 2 : 1, respectively. The effects of the activator type and activator mixing ratio on the pore structure of carbon aerogels under the same conditions were investigated, and the electrochemical properties of the materials were studied. The relationship between the microstructure and electrochemical performance was established by studying the influence of the hierarchical porous carbon structure on the electrochemical performance.

2. Experimental section

2.1. Materials

Cellulose powder was purchased by Shanghai Macklin Biochemical Technology Co., Ltd. All the reagents were analytically pure and were used as received.

2.2. Experimental methods

As shown in Scheme 1, in a typical process, 12 g NaOH, 7 g urea, and 81 g H₂O were thoroughly mixed and cooled to −12 °C. Then, 3 g cellulose was added into the NaOH/urea/H₂O solution. After violent shaking, mechanically stirring was performed for 1 h to promote cellulose dissolution. Afterwards the solution was cooled to −12 °C, refrozen, and then stirred again to obtain a translucent cellulose solution. The cellulose solution was placed in an oven at 50 °C until the cellulose hydrogel was obtained. The hydrogel was washed with diluted acetic acid (AA) solution, and then left to stand in AA solution to replace the water in the hydrogel. Next, this was rapidly frozen using liquid nitrogen, and then freeze-dried in a freeze dryer (−60 °C, 0 mbar) for 48 h to obtain cellulose aerogel. Afterwards, the cellulose aerogel was put in to a tube furnace for two-step heat treatment. In the first stage, the temperature was raised from room temperature to 200 °C in a nitrogen atmosphere at a rate of 5 °C min⁻¹, and the temperature was kept there at 200 °C for 2 h. In the second stage, the sample was heated at 200 °C to 800 °C at a heating rate of 5 °C min⁻¹, and then kept there at 800 °C for 2 h. The obtained product was cleaned with deionized water and dried to obtain CA-acetic acid (CA-AA). The CA-AA and activators (KOH and ZnCl₂) were fully mixed and dispersed in aqueous solution according to the mass ratios of 2 : 1, 1 : 1, and 1 : 2, respectively, and stirred continuously for 8 h, and the solid mixture was collected. In a nitrogen atmosphere, the temperature was raised to 800 °C at a rate of 5 °C min⁻¹, and the mixture was kept at this temperature for 2 h. The obtained product was repeatedly washed with 1 M HCl solution and deionized water until neutral, and then dried, ground, and sieved with 200 mesh. The obtained activated CA samples were named as CA-KOH-1, CA-KOH-2, CA-KOH-3, CA-ZnCl₂-1, CA-ZnCl₂-2, and CA-ZnCl₂-3, respectively, according to the activator and the mixing ratio.

Scheme 1 Preparation process of the activated CAs.

2.3. Structural characterizations

The morphologies and microstructures of the samples were characterized using by X-ray diffraction spectrometry (XRD, AXS D8 Advance, Bruker, Germany), using Cuα radiation (λ = 0.15418 nm) as an X-ray source. The specific surface area (BET) was evaluated using a specific surface area and aperture analyzer (BSD-PM, Beishide, China). Prior to each adsorption experiment, the sample was degassed at 250 °C for 6 h to ensure that the residual pressure was below 10 Pa. The morphology was observed by transmission electron microscopy (TEM, Tecnai G2F 20, FEI, USA). Raman spectra were recorded on a Raman spectrometer (LabRAM HR Evolution, HORIBA JobinYvon, France) operated with 532 nm laser irradiation. X-Ray photoelectron spectroscopy (XPS, Nexsa, Thermo Fisher Scientific, USA) was performed with Al Kα radiation (1486.6 eV) as the excitation source.

2.4. Electrochemical measurements

Titanium mesh was selected as the current collector, and cut into 18 × 18 mm square sheets, which were placed in absolute ethanol for ultrasonic cleaning for 30 min, and then the required current collector was obtained after drying.

According to the ratio of electrode material: conductive agent (carbon black): binder (PTFE) of 8 : 1 : 1, the materials were weighed and placed in an agate mortar, then a proper amount of anhydrous alcohol was added to dissolve them, and they were then ground until fully mixed to finally obtain a plasticine-like electrode mixture. This rubber-like mixture was then placed on a glass plate, and the mixture was rolled repeatedly for about 15 min with a glass rod. The rolled mixture was then kept at 15 MPa for 1 min to obtain a uniform carbon film. After drying, a 15 × 15 mm portion of the carbon film was cut, which was then flatly covered on the current collector and kept at 20 MPa for 30 s to obtain the final electrode sheet with a loading of 7–9 mg cm⁻².

CV, GCD, and EIS analyses were all performed using a three-electrode system, in which Ag/AgCl electrode was used as the reference electrode, platinum wire as the counter electrode, and the prepared electrode sheet as the working electrode, while the electrolyte was 1 M H₂SO₄ aqueous solution. Electrochemical measurements were performed using an electrochemical workstation (Autolad PGSTAT302, Metrohm, Switzerland) in the potential window of 0–0.8 V, with CV curves obtained with a potential scanning rate of 5–200 mV s⁻¹ and GCD curves with a current intensity of 5–10 A g⁻¹. Also, further utilizing the electrochemical workstation (Chi760E, Chenhua, China), EIS was performed in the frequency range of 0.01–100 000 Hz by applying an AC voltage with an amplitude of 10 mV.

In the dual-electrode system, the electrode materials were assembled into a button cell, with two electrodes with a similar mass used as the positive and negative electrodes of the symmetrical SC. A battery test temperature control system (MJS-SP250, Mojes, China) was used to test the capacitance retention rate over 5000 charge–discharge cycles at a current intensity of 1 A g⁻¹.

The mass specific capacitance obtained from the GCD curve was calculated by the following formula:

where I is the constant discharge current (A), Δt is the discharge time (s), ΔV is the discharge voltage (V), and m is the mass (g) of the sample used for the electrochemical test.

The energy density and power density were calculated by the following formulas:

where E is the energy density (W h kg⁻¹), P is the power density (W kg⁻¹), C_cell is the specific capacitance (F g⁻¹) of the fully symmetrical system, ΔV is the applied potential window (V), and Δt is the discharge time (s).

3. Results and discussion

3.1. Physicochemical structures of the hierarchical porous carbon aerogels

As shown in Fig. 1, the microstructure of CA-KOH-3 retained a highly interconnected lamellar network, which was accompanied by a developed pore structure and more abundant nanopores on the basis of the CA, indicating that it had a characteristic hierarchical porous structure, which was caused by the etching of carbon by KOH during the activation process and is beneficial as it can minimize the diffusion resistance on the inner surface and promote the rapid transfer of electrolyte in the whole material. The microstructure of CA-ZnCl₂-3 kept the lamellar stacking structure of CA as well, but its structure was not as dense as that of the CA, and many small layered carbons were broken. In Fig. 1d, it could be seen that there were a large number of short rod-shaped carbon structures, and at the same time, a large number of nanopores were distributed in the lamellar structure. This was mainly because ZnCl₂ mainly acts as a dehydrating agent during the activation process, and it also destroys the dense structure of CA while forming pores, resulting in a micro-morphology.

Fig. 1 SEM images of CA-AA (a) and (b), CA-KOH-3 (c) and (d), and CA-ZnCl₂-3 (e) and (f).

In order to further study the activation effects of the two activators on the microstructure of the CAs, the activated CAs were imaged by TEM, as shown in Fig. 2. The TEM images revealed that both CA-KOH-3 and CA-ZnCl₂-3 were lamellar structures. However, under electron beam irradiation, it was observed that the lamellar structure of CA-ZnCl₂-3 was relatively small. Both of them retained the hierarchical porous structure of the CA interconnections, which was consistent with the SEM results. As can be seen from Fig. 2e and h, the number of mesopores in the CAs activated by ZnCl₂ was far less than that of CA-KOH-3, which may be because ZnCl₂ mainly plays a dehydration role in the activation process, destroying some mesopores and macropores, and making the carbon materials directly become small-area fragments, while KOH mainly performs an oxidation role in the activation process, and etches more micropores to keep the mesopores, on the premise of retaining the original carbon skeleton. The existence of these micropores, as well as mesopores and macropores, provides numerous channels for the diffusion and absorption of electrolytes.²⁵ The carbon materials were prepared into carbon films and tested by TEM. The results showed that the microstructure of the materials before and after the preparation of carbon films was basically the same, which proves that the process of preparing carbon films was not easily affected by the nanoscale pore structure, and the electrode materials of supercapacitors could retain the excellent pore structure of the prepared carbon materials. This is also the reason for the performance gaps among the subsequently assembled supercapacitors.

Fig. 2 TEM images of activated (a) and (b), CA-KOH-3 (d) and (e), and CA-ZnCl₂-3 (g) and (h); and TEM images of carbon films prepared by them, respectively ((c), (f) and (i)).

In order to describe the porosity of the activated CAs more specifically, N₂ adsorption and desorption tests were carried out, as shown in Fig. 3a and b. Under relatively high pressure, all the samples showed the type IV isotherm mode, and hysteresis loops, indicating that there were a large number of mesopores and macropores in the activated CAs. Especially for CAs activated by KOH, the hysteresis loop of activation was more obvious in three areas, which shows that they had a large number of mesopores.²⁶ In addition, the N₂ adsorption–desorption curve of the CAs activated by KOH was obviously higher than that for the CAs activated by ZnCl₂, whereby it increased rapidly at a relatively low pressure, which means that the CAs activated by KOH had more abundant micropores.²⁷ In addition, it could be observed that both CA-KOH and CA-ZnCl₂ displayed the characteristics where the curves increased faster under the ground relative pressure with the increase in the activator ratio, especially the increase of CA-KOH, which means that the micropore content of the activated CAs increased with the increase in activator ratio, and the effect of using KOH as the activator became more obvious. In Fig. 3b, all the samples showed a multi-stage pore structure of macropore-mesopore-micropore. In the micropore structure, all the materials obtained more micropores after activation, and the activation effect of KOH was better than that of ZnCl₂. Also, with the increase in the proportion of the activator, the micropore content also increased, while the content of ZnCl₂ had no obvious influence on the micropore content, which shows that ZnCl₂ had a limited pore-forming effect on the micropores. In the mesoporous structure, the mesoporous content of CA-KOH was better than that of CA-AA, while the mesoporous content of CA-ZnCl₂ was worse than that of CA-AA. Both activators show that the mesoporous content became more developed with the increase in the activator content. In the macroporous structure, activated CAs had more macropores, but were little affected by the content of the activator, indicating that KOH and ZnCl₂ had limited activation effects on the macropores. The activation effect for large-medium-micropores was consistent with the micro-morphology and the law displayed in Fig. 3a.

Fig. 3 N₂ adsorption–desorption isotherms (a) and pore-size distributions (b) of the CA-AA, CA-KOH, and CA-ZnCl₂ samples; (c) XRD curves of the CA-AA, CA-KOH, and CA-ZnCl₂ samples; (d) Raman spectra of the CA-AA, CA-KOH, and CA-ZnCl₂ samples.

EDLC generally requires a relatively high specific surface area for carbon materials, which is generally contributed by small-sized micropores, which can provide effective space for carbon materials to adsorb electrolyte ions and give electrode materials a high specific capacitance value.²⁸Table 1 provides the specific surface area observed for each material, i.e., for CA-KOH-1 (937.97 m² g⁻¹), CA-KOH-2 (1140.46 m² g⁻¹), CA-KOH-3 (1709.62 m² g⁻¹), CA-ZnCl₂-1 (648.74 m² g⁻¹), CA-ZnCl₂-2 (703.18 m² g⁻¹), CA-ZnCl₂-3 (707.29 m² g⁻¹), and CA-AA (616.97 m² g⁻¹). The main reason why the specific surface area of CA-KOH was higher than that of CA-ZnCl₂ was due to the higher micropore content of KOH from the pore-forming process. With the increase in the proportion of KOH and ZnCl₂ activators, the specific surface areas of the CAs also improved. Increasing the proportion of the activator has thus become an effective means to improve carbon materials. However, an excessive activator ratio may lead to overactivation, and the specific capacitance will no longer increase linearly with the increase in specific surface area, because micropores smaller than electrolyte ions cannot provide capacitance value, and can even affect the migration rate of electrolyte ions in carbon materials.²⁹ These results show that both KOH and ZnCl₂ activation can produce more micropores, and the effect of the micropores produced by KOH activation was far better than that by ZnCl₂ activation, resulting in a higher specific surface area.

Table 1 Physical and electrochemical properties of various CAs

Sample	S BET (m^2 g^−1)	I D/IG	C (%)	Specific capacitance (F g^−1)	Inherent resistance (Ω)	Charge-transfer resistance (Ω)	Cyclic stability (%)
CA-AA	616.97	1.96	96.43	138.03	0.375	0.083	101.99
CA-KOH-1	937.97	2.48	94.88	178.72	0.755	0.145	104.18
CA-KOH-2	1140.46	2.40	95.55	186.57	0.758	0.143	105.40
CA-KOH-3	1709.62	2.05	93.46	213.53	0.771	0.178	101.61
CA-ZnCl2-1	648.74	2.48	97.22	141.34	0.872	0.257	103.58
CA-ZnCl2-2	703.18	2.74	96.97	139.52	0.795	0.172	101.66
CA-ZnCl2-3	707.29	2.38	96.90	141.43	0.899	0.181	104.02

---

Fig. 3c shows the XRD curves of CA-AA, CA-KOH, and CA-ZnCl₂. The peaks at 2θ = 23.5° and 43.3°, respectively, corresponded to the (002) and (100) reflections of graphite.³⁰ It could also be seen from the figure that the peaks at 23.5° are very wide, while the peaks at 43.3° are very weak, indicating that the sample have a low degree of graphitization.

The Raman curves in Fig. 3d show that all the samples exhibited a wide disorder-induced D-band (1310 cm⁻¹) and an in-plane vibration G-band (1590 cm⁻¹), which were due to the double resonance effect of the disordered carbon structure and in-plane vibration of the ordered graphite with the E₂g symmetry.³¹ The D-band intensity of all the samples was obviously higher than the G-band intensity, which indicates that the concentration of amorphous carbon in the two CAs was high.³² These results were in good agreement with the XRD and TEM results. The degree of graphitization can be determined by the ratio of the G-band intensity to the D-band intensity. The lower the I_D/I_G value, the higher the degree of graphitization and the less disordered carbon. According to the fitting calculation, CA-AA had the highest graphitization degree (I_D/I_G = 1.96), which was due to the pore structure formed by the use of acetic acid as an anti-solvent to change the hydrogen bonding strength and promote the self-aggregation of cellulose chains.³³ The graphitization degree of the activated materials decreased, and the I_D/I_G values of CA-KOH-1, CA-KOH-2, CA-KOH-3, CA-ZnCl₂-1, CA-ZnCl₂-2, and CA-ZnCl₂-3 were 2.48, 2.40, 2.05, 2.48, 2.74, and 2.38, respectively. Parts of the graphite and interconnected porous structures had well-developed layered pores, which could be expected to improve the electrochemical properties of the CAs.

The total XPS spectra of CA-AA, CA-KOH, and CA-ZnCl₂ are shown in Fig. 4. All the samples were mainly composed of C and a small amount of O. The relative atomic contents of C in CA-KOH-1, CA-KOH-2, CA-KOH-3, CA-ZnCl₂-1, CA-ZnCl₂-2, and CA-ZnCl₂-3 were 94.88%, 95.55%, 93.46%, 97.22%, 96.97%, and 96.90%, respectively. The functional groups contained in the activated CA were consistent. The C 1s spectra could be roughly divided into three peaks at 284.8, 286.0, and 290.0 eV. The peak at 284.8 eV was related to the sp³ C–C band, which belonged to a disordered structure. The peak at 286.0 eV originated from C–O and the peak at 290.0 eV originated from C=O, which were attributed to the presence of oxygen functional groups due to the incomplete carbonization during the preparation of the charcoal. The results show that only physical changes in the pore structure and the degree of graphitization occurred in the activation process of the carbon materials, and no obvious chemical changes occurred.

Fig. 4 Total XPS spectra of the CA-AA, CA-KOH, and CA-ZnCl₂ samples (a); C 1s spectra of CA-KOH-1 (b), CA-KOH-2 (c), CA-KOH-3 (d), CA-ZnCl₂-1 (e), CA-ZnCl₂-2 (f), and CA-ZnCl₂-3 (g).

3.2. Electrochemical performances of the hierarchical porous carbon aerogels

Fig. 5 shows the cyclic voltammetry curves of CA-KOH and CA-ZnCl₂ in aqueous 1 M H₂SO₄ solution for the three-electrode system. As can be seen from the figure, the CV curves showed the typical rectangular shape of EDLC in the 0–0.8 V potential range. The CV curves of all the samples showed rectangular features at low scan rates and diamond-shaped features at high scan rates, indicating the formation of a double electric layer (reversible adsorption and desorption of ions) on the carbon network.³⁴ According to the literature,³⁵ CV curves become sharp with increasing the scan rate, which reflects the presence of larger Ohmic resistance in the pores of a material, as noted here, with CA-KOH-1 (Fig. 5a) and CA-ZnCl₂-3 (Fig. 5f) appearing to be sharper, reflecting the larger resistive cases. It is well known that the integration area corresponds to the capacitance value of SCs, and since CA-KOH-3 had the largest integration area, CA-KOH-3 had the highest specific capacitance among the six groups of materials, which was confirmed by the specific capacitance calculated by GCD later, which was directly related to the fact that CA-KOH-3 had the largest specific surface area. CA-ZnCl₂ exhibited approximately the same pore distribution structure, as can be seen in Fig. 5b, and the specific capacitance calculated from the CV integral area was also basically the same. The overall integral area of CA-KOH was larger than that of CA-ZnCl₂, which was mainly due to the fact that CA-KOH had more micropores, which provide more space for the storage of ions.³⁶ The low value of the specific capacitance of CA-ZnCl₂ was influenced by the activation process of the limited number of micropores created during the process and the loss of some mesoporous structures.

Fig. 5 CV curves of CA-KOH-1 (a), CA-KOH-2 (b), CA-KOH-3 (c), CA-ZnCl₂-1 (d), CA-ZnCl₂-2 (e), and CA-ZnCl₂-3 (f) at scan rates of 5–200 mV s⁻¹.

The capacitive behavior of the activated CAs was analyzed using the GCD test, and the results for CA-KOH and CA-ZnCl₂ at different current densities are shown in Fig. 6. The curves were linear and almost symmetrical, indicating excellent electrochemical reversibility and Coulombic efficiency for all the samples.³⁷ According to Table 2, the IR drop of the electrode was significantly smaller than that of the CA after activation. However, the IR drop still showed an increasing trend with the increase in current density.

Fig. 6 Charge/discharge curves of CA-KOH-1 (a), CA-KOH-2 (b), CA-KOH-3 (c), CA-ZnCl₂-1 (d), CA-ZnCl₂-2 (e), and CA-ZnCl₂-3 (f) at different current densities.

Table 2 IR drops of various CAs at different current densities

Sample	@0.5 A g^−1 (Ω)	@1 A g^−1 (Ω)	@2 A g^−1 (Ω)	@5 A g^−1 (Ω)	@10 A g^−1 (Ω)
CA-AA	0.076	0.124	0.131	0.192	0.362
CA-KOH-1	0.009	0.015	0.033	0.075	0.143
CA-KOH-2	0.019	0.033	0.064	0.159	0.324
CA-KOH-3	0.007	0.014	0.030	0.067	0.129
CA-ZnCl2-1	0.012	0.022	0.047	0.108	0.211
CA-ZnCl2-2	0.011	0.021	0.042	0.096	0.185
CA-ZnCl2-3	0.007	0.014	0.030	0.067	0.129

---

The variation curves of the specific capacitances of CA-AA, CA-KOH, and CA-ZnCl₂ at different current densities are shown in Fig. 7a. The discharge curves showed that at a current density of 0.5 A g⁻¹, the specific capacitances of the CA-KOH-1 (178.72 F g⁻¹), CA-KOH-2 (186.57 F g⁻¹), CA-KOH-3 (213.53 F g⁻¹), CA-ZnCl₂-1 (141.34 F g⁻¹), CA-ZnCl₂-2 (139.52 F g⁻¹), and CA-ZnCl₂-3 (141.43 F g⁻¹) samples had a higher specific capacitance than CA-AA (138.03 F g⁻¹), and when the current density was increased to 10 A g⁻¹, their specific capacitances decreased to 116.33, 165.88, 169.06, 116.24, 117.12, and 89.60 F g⁻¹, respectively. CA-KOH-3 demonstrated the optimum specific capacitance, which was related to the GCD discharge time and the integral area of the CV curve, which was in agreement with the phenomenon presented in the GCD test results, and was also superior to the reported results in the literature for nitrogen-doped cotton carbon fiber foam (107.5 F g⁻¹ at 2 mV s⁻¹)³⁸ and KOH-activated charcoal aerogel extracted from bagasse (268.4 F g⁻¹ at 2 mV s⁻¹).³⁹ These results showed that the specific capacitance of the ZnCl₂-activated charcoal aerogels increased only slightly and were little affected by the ratio of the activator. However, the specific capacitance of the KOH-activated charcoal aerogels increased significantly, and the specific capacitance showed an upward trend with the increase of the ratio of the activator, which was attributed to the fact that the micropores play an important role in the process of diffusive adsorption–desorption and that the mesopores in porous charcoal not only have a larger surface area, but also provide higher accessibility to the electrolyte by providing wider transportation channels for the micropores.

Fig. 7 (a) Specific capacitance curves at different current densities of the CA-AA, CA-KOH, and CA-ZnCl₂ samples; (b) Nyquist plots of the CA-AA, CA-KOH, and CA-ZnCl₂ samples; (c) cyclic stability curves and capacitance retentions of the CA-AA, CA-KOH, and CA-ZnCl₂ samples; (d) photos showing the prepared supercapacitor illuminating the connected LEDs.

At a current density of 10 A g⁻¹, the capacitance of all the samples remained in the range of 65–88%, while that of CA-AA remained at 87%, indicating a decrease in the multiplicity performance of the material. This may be due to the low percentage of macropores in the material since macropores provide larger openings and enough space for high current charging and discharging.⁴⁰ It is worth noting that the specific capacitance of CA-AA increased with increasing the current density, from 124.25 F g⁻¹ (2 A g⁻¹) to 128.23 F g⁻¹ (5 A g⁻¹). The specific capacitance of CA-AA demonstrated an upward trend with increasing the current density, rising from 124.25 F g⁻¹ at 2 A g⁻¹ to 128.23 F g⁻¹ at 5 A g⁻¹. This behavior indicates the excellent rate capability across all the materials, affirming their suitability for high-current-density applications. This advantageous rate performance was attributed to the porous sheet structure of CA-AA, which functions effectively as an electrolyte reservoir. This design feature minimizes diffusion resistance along the inner surface, thereby promoting swift electrolyte transfer throughout the material.⁴¹

KOH and ZnCl₂ activation mainly lies in the etching of micropores, and the enhancement of the pore volume of the macropores is much smaller than that of the micropores, which led to a decrease in the percentage of macropores. This indicates that the activated CAs are suitable for use in low to medium current density applications. As mentioned before, the voltage drop at the beginning of the discharge was very small even at a high current density of 10 A g⁻¹, indicating a very low equivalent series resistance.

The Nyquist plots of CA-AA, CA-KOH, and CA-ZnCl₂ at open-circuit potentials are illustrated in Fig. 7b. The internal resistance of the electrode materials was estimated from the onset of the Nyquist plot on the Z-axis.⁴² The charge-transfer resistance is related to the semicircle in the high-frequency region.⁴³ The almost vertical line parallel to the imaginary axis in the low-frequency region indicates the almost ideal capacitor behavior of the electrode sheet. The high slope in the low-frequency region clearly indicates a better pore accessibility of the electrolyte used to prepare the samples, which could be attributed to the high number of nanopores due to activation, as shown in Fig. 7b. The internal resistances of the samples were 0.755 Ω (CA-KOH-1), 0.758 Ω (CA-KOH-2), 0.771 Ω (CA-KOH-3), 0.872 Ω (CA-ZnCl₂-1), 0.795 Ω (CA-ZnCl₂-2), 0.899 Ω (CA-ZnCl₂-3), and 0.375 Ω (CA-AA), and this set of values corresponded exactly to the IR drop. Also, the charge-transfer resistances of the electrodes were 0.145 Ω (CA-KOH-1), 0.143 Ω (CA-KOH-2), 0.178 Ω (CA-KOH-3), 0.257 Ω (CA-ZnCl₂-1), 0.172 Ω (CA-ZnCl₂-2), 0.181 Ω (CA-ZnCl₂-3), and 0.083 Ω (CA-AA). Compared with CA-AA, the internal resistance and charge-transfer resistance of the activated CAs were slightly improved.

Each sample was assembled into a symmetrical SC, and the cycle stability curves of CA-AA, CA-KOH, and CA-ZnCl₂ at a current density of 1 A g⁻¹ are shown in Fig. 7c. The capacitance retention of all samples after 5000 charge/discharge cycles was slightly higher than the initial value, which was due to the excitation of the active sites in the electrode materials during the cycling process. The capacitance retention of the CAs activated by KOH and ZnCl₂ showed only small ups and downs with no obvious difference, while the overall values were higher than that of CA-AA. Therefore, activation of the CAs by KOH and ZnCl₂ could slightly enhance the cycling stability of the electrode materials. As shown in Fig. 7d, the prepared supercapacitor was connected to a simple circuit and could successfully illuminate connected LEDs.

As can be seen from Fig. 8, the energy and power densities of the symmetric SCs were calculated, and at a current density of 0.5 A g⁻¹, the energy density of CA-KOH-3 was 18.577 W h kg⁻¹ and the power density was 197.863 W kg⁻¹, while the energy density of CA-ZnCl₂-3 was 11.991 W h kg⁻¹ and the power density was 195.331 W kg⁻¹. At a current density of 10 A g⁻¹, CA-KOH-3 had an energy density of 10.490 W h kg⁻¹ and a power density of 3342.056 W kg⁻¹, while CA-ZnCl₂-3 had an energy density of 2.734 W h kg⁻¹ and a power density of 2343.781 W kg⁻¹. CA-AA at a power density of 181.06 W kg⁻¹ had a high energy density of 10.06 W h kg⁻¹, and the energy density was maintained at 3.20 W h kg⁻¹ at a higher power density of 2188.52 W kg⁻¹. It could thus be seen that the energy density of the activated material was improved. The energy density of CA-KOH-3 was significantly greater than that of CA-AA and CA-ZnCl₂-3. Its energy density was close to that of N-doped cellulose-based carbon aerogels and biomass-based functionalized graphene materials, and much higher than that of other biomass-based electrode materials.^44–48

Fig. 8 Ragone curves of various materials in the present study and some reported in the literature.

4. Conclusions

The results showed that for cellulose-based carbon aerogels, both KOH and ZnCl₂ activation increased the microporosity, with KOH being more effective, resulting in a higher specific surface area. KOH activation also produced numerous mesopores and macropores that facilitate ion transport and reduce diffusion resistance, while ZnCl₂ activation disrupted the original mesopore structure. As the activator proportion increased, the specific surface area and pore structure were improved. CA-KOH-3, in particular, showed a high specific surface area, well-developed meso- and micro-pores, and a broad pore distribution, providing ample ion adsorption sites and short ion-transport paths.

Author contributions

Kai Yang: writing – original draft, methodology, data curation, investigation, methodology. Xinfeng Huang: investigation, visualization, software. Yuchun Zhang: writing – review & editing, supervision. Peng Fu: formal analysis, resources, validation, project administration.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Conflicts of interest

The authors declare no competing interests.

Appendix A. Supporting information

Supplementary data associated with this article can be found in the online version.

Acknowledgements

This work was financially supported by Shandong Provincial Natural Science Foundation [grant number ZR2022ME109], the National Natural Science Foundation of China [grant number 51976112], Special Project Fund of “Taishan Scholar” of Shandong Province [grant number tsqn202103066].

References

1. Y. Bai, W. Wang, R. Wang, J. Sun and L. Gao, J. Mater. Chem. A, 2015, 3, 12530–12538.
2. M. Conte, Fuel Cells, 2010, 10, 806–818.
3. J. Zhao and A. F. Burke, J. Energy Chem., 2021, 59, 276–291.
4. P. Sharma and T. Bhatti, Energy Convers. Manage., 2010, 51, 2901–2912.
5. M. D. Stoller and R. S. Ruoff, Energy Environ. Sci., 2010, 3, 1294–1301.
6. L. L. Zhang and X. Zhao, Chem. Soc. Rev., 2009, 38, 2520–2531.
7. Y. Wang, L. Zhang, H. Hou, W. Xu, G. Duan, S. He, K. Liu and S. Jiang, J. Mater. Sci., 2021, 56, 173–200.
8. P. Fu, W. Yi, Z. Li and Y. Li, Bioresour. Technol., 2019, 271, 136–142.
9. M. Vijayakumar, A. Bharathi Sankar, D. Sri Rohita, T. N. Rao and M. Karthik, ACS Sustainable Chem. Eng., 2019, 7, 17175–17185.
10. Y. Gong, D. Li, C. Luo, Q. Fu and C. Pan, Green Chem., 2017, 19, 4132–4140.
11. T. Purkait, G. Singh, M. Singh, D. Kumar and R. S. Dey, Sci. Rep., 2017, 7, 15239.
12. Y. Li, Q. Zhang, J. Zhang, L. Jin, X. Zhao and T. Xu, Sci. Rep., 2015, 5, 14155.
13. S. Hu, S. Zhang, N. Pan and Y.-L. Hsieh, J. Power Sources, 2014, 270, 106–112.
14. Z. Gui, H. Zhu, E. Gillette, X. Han, G. W. Rubloff, L. Hu and S. B. Lee, ACS Nano, 2013, 7, 6037–6046.
15. Z. Li, M. Wang, L. Chen, H. Ji and H.-Y. Yu, Sustainable Mater. Technol., 2023, 36, e00649.
16. M. Aghabararpour, M. Naderi, S. Motahari and S. H. Jafari, J. Electroanal. Chem., 2023, 951, 117922.
17. J. Xiao, X. Yuan, W. Li, T. C. Zhang, G. He and S. Yuan, Int. J. Biol. Macromol., 2024, 269, 132078.
18. Z. Miao, Y. Song, Y. Dong, D. Ge, J. Shui, X. He and H.-Y. Yu, Nano Res., 2023, 16, 3156–3167.
19. L. Yan, S. Y. H. Abdalkarim, X. Chen, Z. Chen, W. Lu, J. Zhu, M. Jin and H.-Y. Yu, Compos. Sci. Technol., 2024, 245, 110364.
20. Y. Dong, D. Xu, H.-Y. Yu, Q. Mi, F. Zou and X. Yao, Nano Energy, 2023, 115, 108723.
21. W. Li, L. Zhu, Y. Xu, G. Wang, T. Xu and C. Si, Adv. Mater., 2024, 14, 2415761.
22. T. Xu, Q. Song, K. Liu, H. Liu, J. Pan, W. Liu, L. Dai, M. Zhang, Y. Wang, C. Si, H. Du and K. Zhang, Nano-Micro Lett., 2023, 15, 98.
23. J. Fischer, K. Thümmler, S. Fischer, I. G. Gonzalez Martinez, S. Ostwald and D. Mikhailova, Energy Fuels, 2021, 35, 12653–12665.
24. R. M. A. P. Lima, G. S. dos Reis, M. Thyrel, J. J. Alcaraz-Espinoza, S. H. Larsson and H. P. de Oliveira, Nanomaterials, 2022, 12, 866.
25. B. L. Vijayan, I. I. Misnon, C. Karuppaiah, G. M. Anil Kumar, S. Yang, C.-C. Yang and R. Jose, J. Power Sources, 2021, 489, 229522.
26. W. Li, Y. Li, C. Yang, Q. Ma, K. Tao and L. Han, Dalton Trans., 2020, 49, 14017–14029.
27. M. B. Arvas, H. Gürsu, M. Gencten and Y. Sahin, J. Energy Storage, 2021, 35, 102328.
28. T. Xu, H. Du, H. Liu, W. Liu, X. Zhang, C. Si, P. Liu and K. Zhang, Adv. Mater., 2021, 33, 2101368.
29. H. Zhou, L. Yan, D. Tang, T. Xu, L. Dai, C. Li, W. Chen and C. Si, Adv. Mater., 2024, 36, 2403876.
30. Y. Zhai, Y. Dou, D. Zhao, P. F. Fulvio, R. T. Mayes and S. Dai, Adv. Mater., 2011, 23, 4828–4850.
31. X. Tian, B. Xiao, X. Xu, L. Xu, Z. Liu, Z. Wang, M. Yan, Q. Wei and L. Mai, Nano Res., 2016, 9, 1012–1021.
32. M. K. Chini and S. Chatterjee, FlatChem, 2017, 1, 1–5.
33. Z. Tong, J. Meng, S. Liu, Y. Liu, S. Zeng, L. Wang, Q. Xia and H. Yu, Carbohydr. Polym., 2021, 272, 118473.
34. X. Zhang, Y. Wang, X. Yu, J. Tu, D. Ruan and Z. Qiao, J. Energy Storage, 2021, 44, 103378.
35. T. Cai, M. Zhou, D. Ren, G. Han and S. Guan, J. Power Sources, 2013, 231, 197–202.
36. K. Ait Kaci Azzou, A. Terbouche, C. Ait Ramdane-Terbouche, H. Belkhalfa, K. Bachari, D. Hauchard and D. Mezaoui, J. Energy Storage, 2022, 47, 103559.
37. Y. Zhang, Y.-P. Zhao, L.-L. Qiu, J. Xiao, F.-P. Wu, J.-P. Cao, Y.-H. Bai and F.-J. Liu, Diamond Relat. Mater., 2022, 129, 109331.
38. C. Wang, Y. Li, X. He, Y. Ding, Q. Peng, W. Zhao, E. Shi, S. Wu and A. Cao, Nanoscale, 2015, 7, 7550–7558.
39. P. Hao, Z. Zhao, J. Tian, H. Li, Y. Sang, G. Yu, H. Cai, H. Liu, C. P. Wong and A. Umar, Nanoscale, 2014, 6, 12120–12129.
40. P. Xu, J. Tong, L. Zhang, Y. Yang, X. Chen, J. Wang and S. Zhang, J. Alloys Compd., 2022, 892, 162129.
41. K. Yang, Q. Fan, Y. Zhang, G. Ren, X. Huang and P. Fu, RSC Adv., 2024, 14, 1123–1133.
42. Y. Hou, Y. Cheng, T. Hobson and J. Liu, Nano Lett., 2010, 10(7), 2727–2733.
43. R. Madhu, V. Veeramani, S. M. Chen, P. Veerakumar and S. B. Liu, Chemistry, 2015, 21, 8200–8206.
44. C.-S. Yang, Y. S. Jang and H. K. Jeong, Curr. Appl. Phys., 2014, 14, 1616–1620.
45. Z. Chen, X. Peng, X. Zhang, S. Jing, L. Zhong and R. Sun, Carbohydr. Polym., 2017, 170, 107–116.
46. D. Puthusseri, V. Aravindan, S. Madhavi and S. Ogale, Energy Environ. Sci., 2014, 7, 728–735.
47. W. Tian, Q. Gao, Y. Tan, Y. Zhang, J. Xu, Z. Li, K. Yang, L. Zhu and Z. Liu, Carbon, 2015, 85, 351–362.
48. W.-H. Qu, Y.-Y. Xu, A.-H. Lu, X.-Q. Zhang and W.-C. Li, Bioresour. Technol., 2015, 189, 285–291.

---