Synthesis of Physically Activated Carbons from Vitellaria paradoxa Shells for Supercapacitor Electrode Applications

Abstract

This study investigates the processing of shea nut shells (SNSs), an abundant agricultural waste, into porous activated carbon for supercapacitor electrodes through a two-stage thermal treatment involving pyrolysis and physical activation with CO₂ and steam. The aim was to develop sustainable, high-performance electrode materials while addressing waste management. Carbonization followed by activation yielded 16.5% (CO₂) and 11.3% (steam) activation yields, with total yields of 4.3% and 2.9%, respectively. CO₂ activation produced carbon (AC_CO₂) with a specific surface area (S_BET) of 1528 m² g⁻¹ and a total pore volume of 0.72 cm³ g⁻¹, a graphitization degree (I_D/I_G = 1.0), and low charge transfer resistance (9.05 Ω), delivering a specific capacitance of 47.5 F g⁻¹ at 0.5 A g⁻¹, an energy density of 9.5 Wh kg⁻¹ at 299 W kg⁻¹, and a fast discharge time of 2.10 s, ideal for power-intensive applications. Steam activation yielded carbon (AC_H₂O) with a higher specific surface area (1842 m² g⁻¹) and pore volume (1.57 cm³ g⁻¹), achieving a superior specific capacitance of 102.2 F g⁻¹ at 0.5 A g⁻¹ and a power density of 204 W kg⁻¹ at 9.2 Wh kg⁻¹, suited for energy storage. AC_CO₂ also exhibited exceptional cyclic stability (90% retention after 10,000 cycles). These findings demonstrate SNS-derived activated carbon as a versatile, eco-friendly material, with CO₂ activation optimizing power delivery and steam activation enhancing energy capacity, offering tailored solutions for supercapacitor applications and sustainable waste utilization.

1. Introduction

Renewable energy sources, such as wind, hydroelectric power, and solar energy, have gained considerable traction as viable alternatives to fossil fuels. However, incorporating these renewable energy sources into the global energy grid presents a significant technical challenge [1]. One of the most critical issues is the unreliable nature of renewable energy production, which is also heavily affected by weather conditions [2]. Consequently, this leads to an imbalance between energy supply and demand, hence the need to look for reliable and efficient means to store excess energy. Supercapacitors (SCs) have become a promising technology in energy storage due to their unique characteristics, including rapid charge and discharge capabilities, high power density, and long cycle life. Unlike conventional batteries, SCs store energy electrostatically, enabling them to deliver rapid energy bursts and withstand numerous charge–discharge cycles with minimal degradation [3,4]. These unique properties make SCs well suited for applications that demand high power output over short durations, including electric vehicles, portable electronics, and renewable energy systems [3].

Despite their advantages, SCs are constrained by their comparatively low energy density compared to conventional batteries. This limitation restricts their application in situations that demand high energy and high-power densities. To address this challenge, extensive research has been focused on optimizing materials as electrodes in SCs [5,6], which play a vital role in defining the overall performance of these devices. Activated carbon from biomass (AC) is among the frequently utilized electrode materials because of their high specific surface area, outstanding electrical conductivity, chemical stability, and cost-effective production [7,8]. Agro-wastes such as walnut shells [9], coconut shells [10], rice husks [11], corn cobs [12], peanut shells [13], sugarcane bagasse [14], olive stones [15], banana and plantain peels [16], palm kernel shells [17], pineapple crown and peels [18], cassava peels [19], mustard seeds [20], orange peels [21], etc., have been employed in making activated carbon for electrodes in SCs.

However, several biomass sources need to be investigated when it comes to the synthesis of activated carbon. One such source is shea nut shells (SNSs), a biowaste from the shea butter production industry in Africa. With about 1,760,000 metric tons of shea nuts being processed into butter annually, around 1,000,000 metric tons of waste are produced [22]. Currently, the SNSs produced by most communities are used as waste in landfills, which serve as breeding sites for disease vectors like mosquitoes. At the same time, some industries utilize it as fuel to help sustain their production plant, contributing to carbon dioxide emissions [23]. The study of SNSs has not received much attention, hence the need to employ such agro-waste.

Several researchers have worked with SNSs as a biowaste for synthesizing activated carbon using a chemical activating method. Abdullahi et al. [24] synthesized activated carbon from SNSs and desert dates using 1.0 M of H₂SO₄ and investigated how contact time, pH levels, and temperature influence the efficiency of removing lead and cadmium from water. The researchers concluded that the synthesized activated carbon effectively removed lead and cadmium from contaminated water. Musah et al. [25] synthesized activated carbon with good pores of different dimensions from SNSs using H₃PO₄ and KOH as activating agents. They reported that synthesized carbons are effective for treating wastewater. Ampong et al. [8] also synthesized oxygen-rich interconnected hierarchical activated carbon using SNSs via a two-stage activation process using KOH, which exhibited impressive electrochemical performance, achieving a high specific capacitance of 286.6 F g⁻¹ and enhanced rate capability. These remarkable properties stemmed from the material’s high specific surface area and O-rich content. The research works reported so far employed chemical activation when utilizing SNSs for synthesizing activated carbons, with most of the application focused on water treatment and very little attention on supercapacitor electrodes for energy storage.

Therefore, based on the above survey and using an experimental setup established in our previous work as described by Breitenbach et al. [26], activated carbons with high specific surface area are synthesized via a two-stage synthesis process by pyrolysis in a nitrogen atmosphere followed by physical activation using steam or CO₂ as activating agents. The prepared carbons are comprehensively characterized and utilized as SC electrodes. With the collected data, relationships between processing, properties, and electrochemical performance are evaluated and compared to existing data from the literature.

2. Results and Discussion

2.1. Material Characterization

Physicochemical characterizations were performed on the precursor sample, as well as moisture content and thermogravimetric analysis (TGA). In Figure 1a, the TGA and DTA results of the untreated precursor sample from room temperature up to 900 °C are presented. The TGA curve shows three distinct zones: A—drying stage, B—pyrolytic stage, and C—carbonization stage. TGA of the raw SNS sample resulted in a one-step thermogram, suggesting a single dominant decomposition process. In contrast, differential thermal analysis (DTA) revealed a broad peak accompanied by shoulder peaks, indicating the presence of multiple components within the raw SNS sample [23]. The thermal analysis began with a loss of moisture at the drying stage, about 6% of moisture of the total mass of the sample at 102 °C. This agrees with the initial moisture analysis conducted on the sample using the moisture analyzer. The degradation temperature for the pyrolytic stage began around 261 °C. This temperature marks the onset of the breakdown of hemicellulose, cellulose, and other volatile compounds in the material. Maximum degradation in the pyrolytic stage was recorded at 320 °C for the onset of degradation for lignin. The carbonization in zone C marks char production with about 26% carbon yield.

Figure 1b presents the overall carbon yield produced from the two modes of physical activation. An initial carbon yield of about 26% was produced after carbonization, which agreed with the thermal analysis results. Physical activation with steam produced an activation yield of 11.3% and a total yield of 2.9%, as compared to 16.5% and 4.3% for CO₂ activation. This result shows that the latter produced more carbon yield than the former, with CO₂ activation producing a double carbon yield as water activation. This phenomenon is attributed to the erratic nature of the steam activation mechanism, which attacks the active sites of the carbon structure to form pores while simultaneously attacking the walls of the pores [27]. CO₂ activation occurs more slowly than steam activation due to reduced diffusion rates at the active sites within the center of the pores [28]. Hence, CO₂ produced a higher carbon yield than steam activation. Detailed reaction mechanisms can be found in references [27,29].

The surface functional groups in the synthesized activated carbon samples were characterized using FTIR spectroscopy, as illustrated in Figure 1c. The spectra of both activated carbons look similar, with the only difference observed in the spectrum of AC_CO₂ at the peak at 1577.5 cm⁻¹ assigned to C-N. This is in accordance with previous findings, as it has been reported that heteroatoms are removed during steam activation [30]. However, this functional group might impact the electrochemical behavior of the prepared carbons as it has been reported to improve wettability and conductivity with the carbon structure [31]. Besides, both activated carbons show peaks for O-H (3658 cm⁻¹) and C-H groups like methylene or methyl (2982, 2890 and 1381 cm⁻¹) as well as carboxylic groups (1253 cm⁻¹), which have been reported for activated carbons in previous studies [32,33]. The double peak around 2349 cm⁻¹, visible for both carbon samples, can be attributed to anti-symmetric stretching of atmospheric CO₂ [34].

Figure 1d shows the Raman spectra of the activated carbons. Raman spectroscopy was used to analyze the structure, defects, and extent of graphitization in the porous activated carbon. Raman shifts were recorded for two prominent peaks around 1358 cm⁻¹ for the D band and 1590 cm⁻¹ for the G band. The D band is associated with the structural defects and disorder of the synthesized activated carbon, while the 2D mode corresponds to the second order of the D band [35]. The G band is also associated with the crystalline or graphitic nature of the as-prepared samples arising from the E_2g vibrational mode and the in-plane stretching of the sp²-hybridized carbon atoms [36]. The D band is further deconvoluted and fitted into three peaks at 1354, 1522, and 1204 cm⁻¹ (Figure S1a), attributed to D₁, D₃, and D₄, respectively [37]. The intensity ratio of the D₁ and G bands (I_D1/I_G) is calculated and used to assess the degree of structural disorderliness of the carbon samples [38]. The I_D1/I_G values for AC_CO₂ and AC_H₂O, calculated from the ratio of the peak heights, are 1.14 and 1.36, respectively. This is in accordance with the overall lower degree of activation and lower porosity of the AC_CO₂ sample. A comparative analysis of the peak heights of the deconvoluted peaks according to the Lorentz function [39] is also given in

Table 1. Structural properties obtained from Raman analysis of physically activated carbons.

Sample ID	ID1 (a.u)	ID3 (a.u)	ID4 (a.u)	IG	ID1/IG
AC_CO2	1595	506	154	1395	1.14
AC_H2O	2275	740	438	1669	1.36

and Figure S1b.

The morphologies of carbonized and activated samples are shown in Figure 2. The surface of the precursor sample, SNS, as shown in Figure 2a, is relatively smooth and almost free of pores. As the precursor was carbonized in a nitrogen atmosphere at 900 °C, the surface of the sample was etched, which introduced voids to make the surface rough and irregular, as shown in Figure 2b. During the activation with CO₂ and steam, as shown in Figure 2c,d, respectively, porous structures were developed for both samples. However, the AC_H₂O-activated sample shows a more distorted surface structure, which can be attributed to the high activation energy of steam, which makes it easier to etch and penetrate the carbon structure whilst expanding the pores effectively.

Additionally, the structural characterization of the activated carbon samples was analyzed using nitrogen adsorption isotherms and recorded at 77 K (Figure 3). The calculated specific surface area (S_BET), the average pore size, and total volume are reported in

Table 2. Textural properties of activated carbon samples.

Sample	SBET (m2 g−1)	VT (cm3 g−1)	Pore Diameter (nm)
AC_CO2	1528	0.72	1.95
AC_H2O	1842	1.57	3.41

. Based on the IUPAC adsorption classifications, AC_CO₂ exhibits a Type I isothermal characteristic of a microporous material with strong adsorbent–adsorptive interactions, whereas AC_H₂O exhibits a Type II isotherm which indicates significant multilayer adsorption [40], as shown in Figure 3a. The sharp rise in adsorption within the low-pressure range of the adsorbed volume indicates the high-density micropores within the activated carbon structure. In contrast, AC_H₂O exhibited higher adsorption than AC_CO₂ in the low-pressure range. This is evident in the honeycomb structure, which has numerous mesopores and micropores developed by H₂O activation. In Table 2, AC_H₂O had a SSA of 1842 m² g⁻¹ and a total pore volume of 1.57 cm³ g⁻¹ higher than 1528 m² g⁻¹ and 0.72 cm³ g⁻¹ obtained by the AC_CO₂. Again, this is attributed to steam’s diffusion effect in etching the pores’ internal layers to form numerous mesopores [28]. The pore size distribution of the activated carbons calculated using the non-localized density functional theory (NLDFT) and assuming slit-pores, presented in Figure 3b, shows that the sample AC_H₂O has a significant share of pores in the mesopore range, while the porosity of AC_CO₂ is almost completely limited to the micropore range below 2 nm. However, it has to be noted that due to the slower activation process and the higher yield in the CO₂ process, there is potential for improving the porosity of the AC_CO₂ sample by increasing the activation time or temperature.

XPS was further used to analyze the surface elemental components of the as-prepared carbon samples. The survey plot of AC_CO₂ and AC_H₂O is illustrated in Figure 4a, where two characteristic peaks for C1s and O1s were observed to be present in both samples. However, characteristic peaks for N1s and Ca2p were observed in the high-resolution spectra of AC_CO₂ (Figure 4d)_. From Figure 4b, C is mostly bonded via single bonds to C or H (284.97 eV) and also bonded to O to form O=C-O (289.4 eV) [41]. O is also mainly bonded to carbon to form carboxyl bonds in the structure of the AC_CO₂ sample at 531.93 eV, as seen in Figure 4c [42]. The remaining O atoms are bonded to other elements (Ca). Two unique peaks for N1s were also detected at 398.57 eV and 400.74 eV in the structure of AC_CO₂, which correspond to pyridinic-N and pyrrolic-N, respectively (Figure 4d) [43]. These characteristic peaks enhance electron transfer at higher current densities [44]. The high-resolution spectra of C1s and O1s for AC_H₂O are also presented in Figure 4e and Figure 4f, respectively. The intensity of the O1s peak shows that there is less O in the carbon structure of AC_H₂O than AC_CO₂, as evident from

Table 3. Elemental composition from high-resolution XPS.

Sample ID	Elemental Composition (at. %)
	C	O	N	Ca
AC_CO2	81.1	15.1	0.8	3.0
AC_H2O	96.5	3.5	-	-

. This is because all the oxygen atoms bond to the carbon. The absence of N-functional groups in the AC_H₂O sample can be attributed to the high energetic nature of the activation agent which removed most of the functional groups during activation, which was also reported by Breitenbach et al. [26]. The high-resolution spectra of Ca2p present in AC_CO₂ is shown in Figure S2. The calculated elemental distributions are further provided in Table 3. The high amount of O and N in the AC_CO₂ suggests that the sample is expected to show good retention capacity and conductivity in contrast to the AC_H₂O sample despite having a larger surface area [43].

2.2. Electrochemical Characterization

The as-synthesized electrodes were tested in a symmetric two-electrode set-up using a Swagelok cell and 1M TEMA-BF4 in propylene carbonate as an electrolyte and measured by CV, GCD, and EIS. Figure 5a,b show the CV profiles from 1 mV s⁻¹ to 300 mV s⁻¹ within the potential range of −2.7 V to 2.7 V. The curves of the electrodes have a quasi-rectangular shape, which is typical behavior for double-layer supercapacitors. This shape depicts the fast-transport nature of the electrolyte ions and the energy storage capacity of the activated carbons. The curves obtained by the AC_H₂O electrodes have a higher integral area than AC_CO₂. As the scan rates increased, AC_CO₂ maintained its quasi-rectangular shape through to 200 mV s⁻¹ compared to AC_H₂O, which only maintained its shape to the 100 mV s⁻¹ scan rate. This is due to good reduction/oxidation reaction and fast ion responsiveness at higher scan rates by AC_CO₂ electrodes, highlighting excellent charge storage and transport capabilities [13]. Figure 5c depicts the CV plot at a 10 mV s⁻¹ scan rate as a comparative analysis for the as-prepared electrodes. It is evident from the plot that AC_H₂O has a higher area than AC_CO₂. Redox peaks are also observed at low scan rates and are more severe in AC_CO₂. This might result from the residual N-containing compounds originating from the precursor (see Figure S3), which are present in the AC_CO₂ sample but are washed out during water activation of the AC_H₂O sample [30]. Due to the low N content (0.8 at. %), the effect of pseudo-capacitance is negligible, and the total capacitance values are dominated by the porosity, which is significantly higher for the AC_CO₂ sample. Despite the absence of the N heteroatoms within the carbon structure of AC_H₂O, it also displays humps in the CV profile. These can be attributed to the creation of a solid–electrolyte interphase (SEI) due to oxidation at high potentials that decreases the overall surface area of the activated carbon electrode and limits ion transport within the porous network, which ultimately reduces the specific capacitance with increased scan rates, as seen in Figure 5d [26].

The specific capacitance for the various scan rates is also reported in Figure 5d. The graph shows that the specific capacitance decreases with an increasing scan rate, which can be attributed to the insufficient time for ion diffusion at higher scan rates. AC_H₂O achieves higher capacitance values from lower scan rates and drops sharply after 100 mV s⁻¹. This is because, at higher potentials, electrolyte degradation tends to produce an SEI that reduces the overall surface area and consequently reduces the specific capacitance. In contrast, AC_CO₂ shows good capacitance retentivity as it maintains its quasi-rectangular shape (Figure 5a) and capacitance at higher scan rates (Figure 5d). This phenomenon can be attributed to the presence of O and N heteroatoms in the AC_CO₂ carbon structure, which tends to improve the conductivity and wettability of the electrode. The capacitance contributions for the devices were analyzed using Dunn’s method, considering the peak currents from various scan rates [45], and are reported in Figure 5e,f. Surface-controlled processes are more stable with increasing scan rates. In the AC_CO₂ device, the capacitance is highly controlled by this mechanism and increases significantly as the scan rate increases. In contrast, diffusion-controlled processes are severe within the AC_H₂O device. The pore structure of AC_H₂O facilitates ion diffusion, and as such, at low scan rates, diffusion-controlled processes are more pronounced compared to the exact mechanism in the AC_CO₂ electrode.

The capacitive nature of the synthesized electrodes was also investigated using the GCD technique at a current density from 0.1 to 4 A g⁻¹ with a potential window of 2.7 V to −2.7 V, as shown in Figure 6a,b.

The GCD profiles for both electrodes exhibit nearly triangular charge–discharge curves at lower current densities, indicating ideal capacitive behavior. Again, the linearity of the discharge curve distorts at higher current densities, which can be attributed to internal resistance and poor ion transport efficiency [46]. From Figure 6c, the curve for AC_H₂O shows a higher voltage drop than AC_CO₂ during discharge, which indicates lower power density capability due to high resistance to ion diffusion. The specific capacitance of the electrodes is also calculated from the GCD results, as shown in Figure 6d. The specific capacitance for both electrodes decreases with increasing current density, typically due to the limited ion diffusion at higher currents. The AC_H₂O electrode shows superior capacitance at lower current densities than the AC_CO₂ electrode due to a more developed porous structure coupled with a high SSA. In contrast, AC_CO₂ maintains its capacitance significantly. This characteristic behavior shows the excellent electrochemical property of AC_CO₂ due to the presence of oxygenated and nitrogen functional groups.

Electrochemical impedance spectroscopy (EIS) measurement from 1 MHz to 10 mHz using a symmetric cell was performed to better understand the capacitive behavior and the ion transport within the electrode’s pores. The Nyquist plot in Figure 7a clearly shows distinct electrochemical characteristics: a steep line in the low-frequency region is indicative of their capacitive nature and the diameter of semi-circular arcs in the high-frequency region represents charge transfer resistance (R_CT). In contrast, the intercept on the real axis represents the equivalent series resistance (Rs) [47]. The measured data was fitted to an equivalent circuit (inset of Figure 7a). The constant phase element, CPE1, is used to compensate for inhomogeneities in the system between electrode–electrolyte–current collector interphases. CPE2 is also employed to compensate for any pseudo-capacitance, while the Warburg impedance, W_o, also represents the total impedance due to the diffusion processes within the symmetric cell [48]. The R_s and R_ct of AC_CO₂ were calculated to be 4.65 Ω and 9.05 Ω, respectively, by fitting the circuit model shown in Figure 7a to a circuit. These values were lower than AC_H₂O (6.87 Ω and 47.09 Ω for R_s and R_ct, respectively). The Bode plot in Figure 7b also highlights the relationship between the impedance and frequency of porous activated carbon at high frequencies. Closer frequency response to a phase angle of −90° shows nearness to ideal capacitive behavior [49]. From Figure 7b, AC_CO₂ shows a phase angle of −72.15° due to its improved conductive nature and wettability. In contrast, a lower phase angle (−56.48°) was recorded for AC_H₂O due to the highly resistive nature of the electrode used for charge transfer. The capacitor frequency response at a phase angle of −45° shows the time constant (τ), which depicts the fast discharge nature of the supercapacitor device [50]. AC_CO₂ has a low discharge time of 2.10 s compared to 5.73 s for AC_H₂O.

To intrusively ascertain the practical application of the AC_CO₂ symmetric device (Figure 7c), its cyclic stability is analyzed at 4 A g⁻¹. The device maintained a good capacitance retention of 90% after 10,000 cycles, as shown in Figure 7d. A steady drop in retention was observed up to 7000 cycles; then, capacitance reached a plateau at approximately 90% of the initial value up to the end of the stability test at 10,000 cycles. The Ragone plot (Figure 7e) of AC_CO₂ and AC_H₂O symmetric devices displays a maximum energy density of 14.9 and 29.1 Wh kg⁻¹, respectively. At a current density of 0.5 A g⁻¹, energy densities of 9.5 Wh kg⁻¹ and 9.2 Wh kg⁻¹ at power densities of 299 W kg⁻¹ and 204 W kg⁻¹ were found for AC_CO₂ and AC_H₂O, respectively. These results are compared with other reported results at similar current densities of 0.5 A g⁻¹ and 1 A g⁻¹ (Figure 7f) [8,51,52,53,54,55,56].

3. Materials and Methods

3.1. Materials

SNSs, lignocellulosic biomass, were sourced from Bolgatanga in the Upper East region of Ghana. These shells served as the precursor material in the synthesis process. Graphite (99.5% pure, C-NERGY™, Imerys Graphite & Carbon Switzerland SA, Bodio, Switzerland), carbon-coated aluminum foil (z-flo 2651, Coveris Management GmbH, Vienna, Austria), polypropylene separator foil (Celgard® 3401, Celgard LLC, Charlotte, NC, USA), triethylmethylammonium tetrafluoroborate (98% pure, from TCI Deutschland GmbH, Eschborn, Germany), polytetrafluoroethylene (Fluorogistx CT LLC, Greenville, DE, USA), and propylene carbonate (99.5% pure, Acros Organics N.V., Geel, Belgium) were used for this experiment without modification. Distilled water (DW) was utilized throughout this work.

3.2. Activated Carbon Preparation

Figure 8 shows the schematic process for synthesizing activated carbon, similar to the work performed by Breitenbach et al. [26]. SNSs were washed with DW to remove any debris and dried in the sun for a day. The shells were ground using a universal mill (PULVERISETTE 19, FRITSCH GmbH, Idar-Oberstein, Germany) and sieved under 500 μm. Next, 50 g of the ground SNSs was dried in a vacuum overnight at 120 °C in an oven (Binder Inc., Bohemia, NY, USA), pyrolyzed in a nitrogen-rich chamber furnace (HTK8 Carbolite Gero GmbH, Neuhausen, Germany) using a heating rate of 5 K min⁻¹, and held isothermally at a temperature of 900 °C for 30 min and left to cool to room temperature. Then, 10 g of carbonized SNSs was activated in a rotary furnace (RSR-B 120/500/11, Nabertherm GmbH, Lilienthal, Germany) at a temperature of 880 °C under a nitrogen atmosphere and kept isothermally for 30 min before actual activation with CO₂ at a gas flow of 50 L h⁻¹ for 120 min. Steam activation was also carried out using the same setup and program with a peristaltic pump attached using water at a flow rate of 1 mL min⁻¹. The rotary tube was jacked up at one end of the rotary furnace with an installed cold trap to prevent condensed water from destroying the activated carbon. The samples activated with steam and CO₂ are hereafter labeled as AC_H₂O and AC_CO₂, respectively. The yield of each thermal process was determined according to Equation (1).

$$\mathrm{Y}=\frac{{\omega }_0}{{\omega }_{\mathrm{a}}}\times 100\mathrm{\%}$$

(1)

where w₀ = the mass of carbon before carbonization or activation, w_a = mass of carbon after carbonization or activation, and Y = obtained percentage yield of carbon.

3.3. Characterization of Lignocellulosic Biomass and Activated Carbon

The moisture content of the ground SNSs was measured using a moisture analyzer (MX-50, A&D Company, Tokyo, Japan). Thermal analysis of the precursor was performed with a STA449F3 TGA analyzer (NETZSCH, Selb, Germany). Carbonization, activation, and total yield were determined using Equation (1). The prepared samples’ specific surface area and total pore volume were investigated by gas adsorption using an Autosorb-iQ system (Anton Paar QuantaTec Inc., Graz, Austria). The morphology of the activated carbon and electrodes were characterized using a Phenom ProX scanning electron microscope (SEM) (Thermo Fisher Scientific, Waltham, MA, USA). Fourier transform infrared (FTIR) measurements were performed using a VERTEX70 spectrometer (PIKE GladiATR, Bruker Corporation, Billerica, MA, USA) in transmittance mode. Raman spectroscopy was also employed to examine the crystalline properties of the as-synthesized carbon samples using a HORIBA XploRA (HORIBA France SAS, Palaiseau, France) with a 532 mm excitation wavelength and a grating of 1800 grooves/mm. The surface functional groups were further investigated via X-ray photoelectron spectroscopy using a NEXSA G2 XPS system (Thermo Fisher Scientific, Waltham, MA, USA). The samples were probed by an Al-K_α monochromatic X-ray source (1486.6 eV) using a spot size of 400 μm in diameter. The hemispherical analyzer was operated in a constant analyzer energy mode with a pass energy of 200 eV and energy step of 1 eV for the survey spectra, whereas 20 eV and 0.05 eV steps were taken for recording the high-resolution spectra. A dual flood gun, which provides low-energy electrons and Ar ions as a simultaneous beam, was used to compensate the surface charge. A standard charge shift referencing the spectra via a C1s peak of adventitious carbon at 285.0 eV was applied.

3.4. Electrode Fabrication and Characterization

The working electrode was prepared by mixing as-synthesized activated carbon, graphite, and PTFE in an 80:10:10 mass ratio into a kneadable dough of 100 µm thickness. To ensure optimal performance, 8 mm electrodes were punched out from the dough and dried in a vacuum overnight at 110 °C. The electrochemical performance of the fabricated electrodes was evaluated using a symmetric two-electrode system. For this setup, two electrodes of similar masses, a 9 mm separator, two 8 mm carbon-coated aluminum foils serving as current collectors, and 200 µL of 1 M TEMA-BF4 solution in propylene carbonate as the electrolyte were assembled into Swagelok^®-type test cells (Figure S4). The schematic for the cell assembly is provided in Figure S5. Electrochemical impedance spectroscopy (EIS), cyclic voltammetry (CV), and galvanostatic charge–discharge (GCD) analyses were performed using a multi-channel electrochemical potentiostat (Vertex. One, IviumTechnologies BV, Eindhoven, The Netherlands).

EIS measurements were carried out with an AC voltage amplitude of 10 mV, spanning a frequency range of 1 MHz to 0.01 Hz. Specific capacitance (C_s), energy density (E_s), and power density (P_s) were calculated using data from CV and GCD experiments. These calculations considered the total mass (twice the mass of active material) and the single mass of the active material (80% of the electrode’s mass) for the device and electrode analysis, respectively. The equations for the analysis are referenced below:

$$C_{s,Cv}=\frac{{\int }_{V_i}^{V_f}i=dV}{2m_E·r}$$

(2)

$$C_{s,GCD}=\frac{2·I·t}{m_E\mathsf{\Delta }V}$$

(3)

$C_{s,Cv}$ in Equation (2) is the specific capacitance of the electrode, V_f and V_i are the initial and final voltages of the potential window ΔV, r is the scan rate, and m_E is the active mass of the electrode as reported by Breitenbach and co-researchers in 2021 [26]. Equation (3) was used to calculate the specific capacitance from the GCD results, where the current = I, active mass of electrode = m_E, discharge time = t, and potential window = ΔV. Equations (4) and (5) were also used to calculate energy density, E_s, and power density, P_s, for the symmetric device.

$$E_s=\frac{1}{8}·C_{s,GDC}{\left(\mathsf{\Delta }V\right) }^2$$

(4)

$$Ps=\frac{E_s}{t}$$

(5)

4. Conclusions

This study successfully demonstrates the potential of SNSs as a sustainable source for producing activated carbon electrodes in a two-stage activation process for supercapacitor applications. CO₂ and H₂O steam were used as physical activation agents for synthesis. The analysis showed that steam activation created a highly porous structure with a larger SSA (1842 m² g⁻¹) and higher total pore volume (1.57 cm³ g⁻¹). These properties of AC_H₂O translated into superior specific capacitance of 102.2 F g⁻¹ at 0.5 A g⁻¹, coupled with an energy density of 9.2 Wh kg⁻¹ at a higher power density of 204 W kg⁻¹ in a two-electrode symmetric Swagelok cell. In contrast, CO₂ activation produced carbon with a higher yield but reasonable SSA of 1528 m² g⁻¹, thus showing potential for further improvements through optimized activation time and/or temperature. The sample already shows better graphitization degree, a lower charge transfer resistance of 9.7 Ω, and the presence of O and N heteroatom functional groups. The AC_CO₂ symmetric device delivers a specific capacitance of 47.5 F g⁻¹ at 0.5 A g⁻¹ and an energy density of 9.5 Wh kg⁻¹ at 199 W kg⁻¹ power density, and it also achieved 90% capacitance retention after 10,000 cycles. This research highlights an innovative way to repurpose agricultural waste into high-performance energy storage materials, contributing to sustainable waste management and renewable energy solutions.