Hierarchical Porous Activated Carbon Derived from Pleurotus Eryngii and the Influence of Pore Structural Parameters on Capacitance Performance

Abstract

Hierarchical porous activated carbon derived from pleurotus eryngii was prepared by a one-step activation method. It was found that the specific surface area of the obtained sample increased with the increase in activation temperature (700–900 °C). The sample activated at 900 °C has a specific surface area of 2002.2 m² g⁻¹ and the highest specific capacitance (319 F g⁻¹), which is mainly attributed to the high utilization rate of specific surface area brought by the hierarchical porous structure. The assembled PEK-900//PEK-900 capacitor measured a specific capacity of 258 F g⁻¹ at a current density of 0.5 A g⁻¹. After 10,000 cycles of charging and discharging, the specific capacitance increased by 10%. Based on the correlation analysis of experimental data between the specific capacitance and pore structural parameters, Lasso dimensionality reduction and binary linear regression were used to reveal the relationship between the two. The residual sum of squares obtained by this method decreased by 38.4% compared to the univariate linear regression, providing a simple and reliable theoretical method for predicting the capacitance performance of biomass carbon materials.

1. Introduction

In recent years, biomass materials have received great attention and recognition from researchers due to their unique characteristics of universality, richness, renewability, and being pollution-free [1,2,3,4]. More and more biomass materials have already been demonstrated as appropriate carbon precursors to prepare electrode materials for supercapacitors applications [5,6,7]. For instance, Tian et al. [8] successfully synthesized ordered flute-shaped microporous carbon materials using cotton straw as carbon source. Tan et al. [9] prepared porous carbon materials via soaking bagasse in different concentrations of NaOH to change its composition. Su et al. [10] used willow catkins as carbon precursors and obtained hierarchical porous carbon materials through Ni(NO₃)₂·6H₂O etching and KOH activation, which retained the original tubular structure of willow catkins.

After years of research, various methods have been developed to prepare biomass carbon materials, such as activation method [8,10,11], template method [12,13,14], chemical vapor deposition [15,16], microwave radiation [17,18], spray pyrolysis [19,20], etc. Among them, activation method is the simplest, although it often requires a large number of activators. However, other methods are more complex and expensive, and some may even cause serious environmental pollution.

Compared with KOH and H₃PO₄, KHCO₃ is not only a mild activator, but also has excellent activation effects [21,22]. In our previous research [23], this has also been confirmed through comparative experiments. Compared to thermal reaction without activators, the specific surface area of paddy-derived carbon materials activated by KHCO₃ has significantly increased by 1425 m² g⁻¹, and its specific capacitance has also dramatically improved by over 200 F g⁻¹. This is mainly attributed to the activation mechanism of KHCO₃, which could be divided into two processes. Firstly, the CO₂ and water vapor released by the thermal decomposition of KHCO₃ could form three-dimensional interconnected honeycomb-shaped large pores in carbon materials. Moreover, other decomposition products further create mesopores and micropores at higher activation temperatures, ultimately forming hierarchical porous carbon materials.

As is known to all, the specific capacitance of carbon materials depends on various factors, such as specific surface area (SSA), pore size distribution (PSD), and conductivity, with the SSA being the most important. However, the relationship between specific capacitance and SSA is complex. Zhang et al. [24] and Hu et al. [25] obtained a linear relationship between the two through the analysis of experimental data. Niu et al. [26] found that the change in specific capacitance is consistent with the trend of the proportion of micropores in the SSA.

In this work, a one-step activation of pleurotus eryngii has been used to prepare hierarchical porous carbon materials by adjusting the activation temperature (700–900 °C). To study the relationship between the specific capacitance and pore structural parameters, Lasso dimensionality reduction and binary linear regression were used to reveal the relationship between the two, providing a simple and reliable theoretical method for predicting the capacitance performance of biomass carbon materials.

2. Materials and Methods

2.1. Material Synthesis

The cleaned pleurotus eryngii was cut into 1 cm³ small cubes and mixed with KHCO₃ powder in a mass ratio of 1:3. Subsequently, an appropriate amount of deionized water was added into the mixture and stirred evenly. After being dried at 80 °C in a vacuum drying oven, the mixture was placed in a high-speed ball mill for 15–30 min to form a powder, and then activated for 3 h in a tube furnace with nitrogen protection. When naturally cooled to room temperature, the activated sample was washed with sufficient dilute hydrochloric acid until acidic and filtered with excess deionized water until neutral. Finally, the product was dried in a vacuum drying oven at 80 °C. In order to study the effect of activation temperature, the activation temperatures were set to 700, 750, 800, 850 and 900 °C, and the corresponding products were named PEK-t, where t represents temperature. It should be noted that in our previous research [27], PEK-900 was used to study the impact of KHCO₃ activation on various representative biomass materials with typical morphology and composition.

2.2. Material Characterization

The morphologies of the samples were captured using a Hitachi SU8010 scanning electron microscopy (SEM) system (Hitachi, Ltd., Tokyo, Japan). To analyze the X-ray diffraction (XRD) patterns, a Bruker D8 ADVANCE diffractometer (Bruker Corporation, Karlsruhe, Germany) was utilized with Cu Kα radiation. The X-ray photoelectron spectroscopy (XPS) analysis was conducted on a ESCALAB 250Xi system (Thermo Fisher Scientific, Waltham, MA, USA). Additionally, the SSA and PSD were evaluated with a Micromeritics ASAP 2460 instrument (Micromeritics Instruments Corporation, Norcross, GA, USA).

2.3. Electrochemical Testing

The production process of the working electrode was as follows: Firstly, mix the carbon material, carbon black and polyvinylidene fluoride in a ratio of 8:1:1 by mass. Then use a dropper to drop the appropriate N-methyl-2-pyrrolidone solvent into the mixture and stir thoroughly for 6 h. Subsequently, use a brush to evenly coat the mixed slurry onto the nickel foam and thoroughly dry it under vacuum at 80 °C for 24 h. After compression with the tablet press, the electrode was ready for testing. The approximate mass loading of the active material per single electrode was 3 mg cm⁻².

All the electrochemical tests were performed using a Metrohm Autolab PGSTAT302N electrochemical workstation (Metrohm China Ltd., Hong Kong, China). In the three-electrode system, 6 M KOH aqueous was used as the electrolyte, while a platinum plate and a Hg/HgO electrode were served as counter and reference electrode. In the two-electrode system, the supercapacitor was assembled from two PEK-900 electrodes with 6 M KOH aqueous as the electrolyte. For cyclic voltammetry (CV) testing, the scanning rate was set to 5–100 mV s⁻¹, which comprehensively reflected the electrochemical behavior of electrode materials under different conditions. For galvanostatic charge/discharge (GCD) testing, the current density range was 1–20 A g⁻¹, which helped to evaluate the charge discharge performance and rate performance of electrode materials. In electrochemical impedance spectroscopy (EIS) testing, a sine wave signal with an amplitude of 5 mV was applied to the tested electrode, covering a wide frequency range from 0.01 Hz to 100 kHz to obtain the impedance information of the electrode at different frequencies. During the cycle life test (CLT), the supercapacitor underwent 10,000 charge and discharge cycles at a current density of 4 A g⁻¹.

The calculation formula for the specific capacitance is as follows:

$${\mathrm{C}}_{\mathrm{g}}=\mathrm{k}\times \mathrm{I}\times ∆t/(\mathrm{m}\times ∆V)$$

(1)

where k takes 1 in the three-electrode system and 2 in the two-electrode system, m is the mass of the active material on each electrode, I, $∆t$ and $∆V$, respectively, represent the discharge current, time, and voltage drop.

3. Results and Discussion

Figure 1 shows the SEM images of samples at different magnifications. It can be observed that these five samples have similar morphology with macropores above 500 nm formed on the surface, presenting a three-dimensional honeycomb porous structure. These macropores were mainly attributed to the volatile substances in Pleurotus eryngii and CO₂, water vapor released by the thermal decomposition of KHCO₃. Moreover, the surface of samples became rougher as the activation temperature increased (Figure 1b,d,f,h,j), due to the further activation of KHCO₃ at higher temperatures.

The XRD diagram of all samples are displayed in Figure 2a. The two broad diffraction peaks appearing near 25° and 43° correspond to the (002) and (100) crystal planes of graphite, indicating that all samples have amorphous structures [28,29]. Figure 2b shows the Raman spectra of all samples. The two distinct absorption peaks correspond to the D-band and G-band. After peak fitting processing, the relative intensity ratio of D-band to G-band (I_D/I_G) of PEK-700, PEK-750, PEK-800, PEK-850, and PEK-900 are obtained, which are used to measure the degree of graphitization and defect in carbon materials [22]. Specifically, the I_D/I_G values of these samples are 2.16, 2.45, 2.56, 2.64, and 3.04, respectively. From the Figure 2, it can be seen that as the activation temperature increases from 700 °C to 900 °C, the intensity of the XRD diffraction peaks gradually decreases, and the value of I_D/I_G ratio increases, which indicate that the degree of graphitization gradually decreases. This is mainly because the high temperature enhances the activation effect of KHCO₃, leading to more pores and defects.

In order to analyze the effect of activation temperature on the surface chemical properties and element content, XPS analysis was conducted on five samples [30,31]. The elemental contents of samples are shown in

Table 1. Surface element composition and relative contents of functional group in O 1s from XPS spectra of PEK samples.

Samples	C (Atom%)	O (Atom%)	N (Atom%)	C=O (%)	C-O-C (%)	O-C=O (%)
PEK-700	89.27	8.69	2.04	13.81	52.72	33.47
PEK-750	92.89	5.71	1.4	15.74	49.83	34.43
PEK-800	91.87	6.87	1.26	14.14	52.49	33.37
PEK-850	95.13	3.94	0.93	13.52	49.28	37.20
PEK-900	93.01	5.9	1.09	14.03	54.95	31.02

. It can be observed that all samples mainly contain carbon and oxygen, as well as a small amount of nitrogen element. In addition, the nitrogen content shows a decreasing trend with the increase of activation temperature. This is because nitrogen-containing functional groups are more easily removed as the temperature increases during the activation process. The full spectrum is shown in Figure 3a, where the two distinct peaks located near 284 eV and 533 eV correspond to C 1s and O 1s, respectively [32]. What is more, a weak peak of N 1s also appeared near 399 eV in the spectrum of PEK-700 [33] (the nitrogen content in other samples is too few to be detected). In order to deeply analyze the binding state of atoms in the samples, the high-resolution O 1s spectra are presented in Figure 3b–f. The O 1s spectra could be well-fitted with three peaks, corresponding to C=O, C-O-C, and O-C=O groups [34,35]. The relative contents of different types of oxygen functional groups are also listed in Table 1. From the table, it can be seen that as the activation temperature increased, the proportions of three groups only fluctuated slightly. It is worth mentioning that in aqueous electrolytes, the C=O group is the main contributor to pseudocapacitance [36]. Moreover, these functional groups not only provide pseudocapacitance, but also improve the hydrophilicity of carbon materials, reduce the contact resistance between carbon materials and electrolytes, thereby improving the conductivity of the materials [37].

To investigate the effect of activation temperature on the pore structure of carbon materials, nitrogen adsorption and desorption tests were applied on all samples. According to Figure 4a, the isotherms of PEK-700, PEK-750, PEK-800, PEK-850 increase sharply in the low-pressure region, while they have almost horizontal plateaus at high-pressure region. Moreover, the higher the activation temperature, the greater the adsorption, implying that the number of micropores increases with elevated temperature [38]. In contrast, the isotherm of PEK-900 shows a typical type IV isotherm with a clear hysteresis loop, implying the existence of abundant micropores and mesopores [39]. Thus, it can be seen that the activation temperature is a key factor affecting the specific surface area and pore size distribution of carbon materials. As shown in

Table 2. The pore structural parameters of all samples prepared at different activation temperatures.

Samples	SBET (m2 g−1)	Smicro (m2 g−1)	Vtotal (cm3 g−1)	Vmicro (cm3 g−1)
PEK-700	1049.1	1008.1	0.4853	0.4001
PEK-750	1270.7	1229.0	0.6006	0.5024
PEK-800	1502.4	1440.9	0.6920	0.5751
PEK-850	1852.0	1759.5	0.8981	0.7426
PEK-900	2002.2	1227.7	1.2963	0.6036

, when the activation temperature increases from 700 °C to 850 °C, the specific surface area of carbon materials increases from 1049.1 m² g⁻¹ to 1852.0 m² g⁻¹, while the specific surface area of micropores ranges from 1008.1 to 1759.5 m² g⁻¹. Combined with pore size distribution (Figure 4b), it can be inferred that more micropores were generated within this temperature range. When the activation temperature further increases to 900 °C, the specific surface area reaches a maximum value of 2002.2 m² g⁻¹, and the mesopores at 2–6 nm were significantly increased (Figure 4b); however, the specific surface area of micropores reduces from 1759.5 m² g⁻¹ to 1227.7 m² g⁻¹. The results indicated that the activation effect of KHCO₃ is the strongest at 900 °C, which can effectively promote the increase in pore size.

In order to further find out the effect of activation temperature on the specific surface area of carbon materials, Figure 5 shows the least squares fitting of the relationship between the specific surface area and activation temperature. From this, it can be inferred that the specific surface area of carbon materials is related to the activation energy of the activation reaction, and its relationship can be expressed by Equation (2):

$$S_{BET}=S_0\exp \left(\frac{-E_a}{kT}\right)$$

(2)

where S_BET is the specific surface area of carbon materials, S₀ is the thermodynamic factor, E_a is the activation energy of reaction, k is the Boltzmann’s constant, and T is the thermodynamic temperature.

By calculating the logarithms on both sides of Equation (2), Equation (3) could be obtained as follows:

$$\ln \left(S_{BET}\right)=\ln \left(S_0\right)-\frac{E_a}{kT}$$

(3)

Substituting the slope of the fitted curve (Figure 5) into Equation (3), it can be calculated that the activation energy is 31.77 kJ mol⁻¹ for the activation reaction. The results showed that the higher the temperature, the higher the energy obtained, and the more intense the activation reaction of KHCO₃, resulting in a larger specific surface area. Furthermore, this result has certain guiding significance for formulating experimental plans, and the specific surface area can be inferred from the fitted curve at different activation temperatures.

The CV curves obtained from the three-electrode test is shown in the Figure 6a. All samples exhibit the typical properties of double layer capacitance, as well as some minor distortions indicative of the pseudo capacitance provided by functional groups on the surface of carbon materials. Notably, PEK-900 shows the maximum current response, indicating its ideal double layer capacitance characteristics and maximum specific capacitance. Figure 6b shows the GCD curves of all samples at a current density of 1 A g⁻¹. These five curves all show the shape of an isosceles triangle slightly deviating from the linear symmetry, indicating that there is a small amount of pseudo capacitance in the electrode materials, which is consistent with the results of CV curves. The specific capacitance of PEK-900 is up to 319 F g⁻¹, higher than the 160 F g⁻¹ of PEK-700, 204 F g⁻¹ of PEK-750, 238 F g⁻¹ of PEK-800 and 279 F g⁻¹ of PEK-850. The variation trend of specific capacitance of each sample under different current densities are shown in Figure 6c. Compared to other samples, PEK-900 not only exhibits the maximum specific capacitance, but also maintains 225 F g⁻¹ even when the current density increases to 20 A g⁻¹, with a capacitance retention rate of 70.5%. From the above discussion, it comes down to the fact that the most excellent capacitive characteristics of PEK-900 should mainly ascribe to its 3D hierarchical porous structure which could quickly transfer electrolyte ions into the micropores inside the material, improving the utilization rate of the micropores and enhancing overall capacitance.

In order to further investigate the ion transport process of all samples in electrolyte, EIS testing was conducted, and the results are shown in Figure 6d. In the low-frequency region, the impedance curves are almost orthogonal to the real axis, exhibiting perfect capacitance characteristics. In the high-frequency region, the smallest arc radius of PEK-900 indicates the lowest charge transfer resistance, implying good charge–discharge ability and rate performance.

As is well known, there is a complex relationship between specific capacity and pore structure parameters, which has attracted many scholars to conduct in-depth research. Zhang et al. [24] and Hu et al. [25] obtained a linear relationship between capacitance and specific surface area. In addition, Niu et al. [26] found that the change in specific capacitance is consistent with the trend of the proportion of micropores in the specific surface area. In order to more accurately reveal their relationship, based on the correlation analysis of experimental data, Lasso regression was used to reduce the dimensionality of pore structure parameters, and a binary linear regression model was used to fit the relationship.

Figure 7 is a thermal diagram that employs a color scheme to represent the density or intensity of data, effectively highlighting the correlation between specific capacitance and pore structural parameters. In this figure, the strength of the correlation is indicated by the lightness of the blue color, where a lighter shade signifies a stronger correlation. It can be observed that specific capacitance is positive correlated with S_BET, S_micro, V_total, and V_micro. According to the degree of correlation, the order is S_BET > V_total > V_micro > S_micro. Meanwhile, the thermal diagram also shows obvious collinearity among the pore structural parameters. For example, the correlation coefficient between S_BET and V_total is 0.94, indicating the existence of redundancy. The same conclusion can also be drawn from the correlation matrix (

Table 3. Pearson correlation coefficient matrix.

	SBET	Smicro	Vtotal	Vmicro	C
SBET	1.00	0.60	0.94	0.85	0.99
Smicro	0.60	1.00	0.28	0.93	0.54
Vtotal	0.94	0.28	1.00	0.62	0.95
Vmicro	0.85	0.93	0.62	1.00	0.81
C	0.99	0.54	0.95	0.81	1.00

). Therefore, further screening of these parameters is required using Lasso regression.

After Lasso regression, the coefficients of pore structural parameters were obtained as shown in

Table 4. Coefficients of pore structural parameters.

Parameters	SBET	Smicro	Vtotal	Vmicro
Coefficients	0.164	−0.019	0	0

. The coefficients of V_total and V_micro decrease to 0, leaving coefficients of 0.164 and −0.019 for S_BET and S_micro, respectively. Therefore, S_BET and S_micro were used as variables for binary linear regression.

Based on the above analysis, the relationship between specific capacitance and pore structural parameters fitted by a binary linear regression model can be expressed as

$$C=13.304+0.164S_{BET}-0.019S_{micro}$$

(4)

For comparison, a univariate linear regression fitting was performed using S_BET as a variable. The fitting results showed that the residual sum of squares (RSS) obtained by the binary linear regression model (124.06) decreased by 38.4% compared to the RSS obtained by the univariate linear regression model (201.30). Figure 8 shows the relationship between the specific capacity and specific surface obtained from the univariate linear regression and binary linear regression (S_micro = 500, 1000 and 1500 m² g⁻¹). It can be seen that S_micro is an important factor affecting the specific capacitance in addition to S_BET. The use of binary linear regression can more accurately reflect the relationship between pore structural parameters and specific capacitance, providing a simple and reliable theoretical method for predicting the capacitance performance of biomass carbon materials.

In order to further evaluate the practical commercial application prospects of PEK-900, two electrode tests were conducted using PEK-900 as positive and negative electrodes, and 6 M KOH as electrolyte to assemble capacitors. Figure 9a shows the CV curves of the supercapacitor operated in different voltage ranges at a scanning rate of 50 mV s⁻¹. When the voltage increases from 1.2 to 1.3 V, the current significantly increases and the rectangle undergoes severe deformation. Therefore, 0–1.2 V is used as the operating voltage window for the PEK-900//PEK-900 supercapacitor.

Figure 9b shows the CV curves of the supercapacitor at a scanning rate of 5–100 mV s⁻¹. All curves exhibit a nearly rectangular shape, especially without significant deformation at 100 mV s⁻¹, indicating that the supercapacitor has fast I–V response characteristics and fast charging and discharging capabilities. The GCD curve of the capacitor at a current density of 0.5–20 A g⁻¹ is shown in Figure 9c, and all curves exhibit good symmetry. The specific capacity and its variation trend are shown in Figure 9d. At current densities of 0.5, 1, 2, 5, 10, and 20 A g⁻¹, the specific capacities are 258, 229, 218, 205, 192, and 174 F g⁻¹, respectively, exhibiting good rate performance (67.4%).

Cycle life test (CLT) was applied to study the long-term cycling stability of the supercapacitor. As shown in Figure 9e, after 10,000 cycles charging and discharging at a current density of 4 A g⁻¹, the specific capacitance of the supercapacitor increased by 10%. From the illustrations in Figure 9e, it can be seen that there was a significant improvement in charge and discharge time in the first 2500 cycles, while the charge and discharge curves basically overlapped in the 2500, 5000, and 10,000 cycles. It is not difficult to speculate that the first 2500 cycles of charge and discharge improve the wettability of the electrode material, making it easier for electrolyte ions to penetrate deeper into the electrode material, thus forming more double layer capacitors. In the subsequent cycles, it tends to stabilize, demonstrating excellent cycle stability.

4. Conclusions

In summary, hierarchical porous activated carbon derived from pleurotus eryngii was prepared by a one-step activation method. By adjusting the activation temperature (700–900 °C), it was found that the specific surface area of the obtained sample increased with the increase of activation temperature. Among them, the sample activated at 900 °C has a specific surface area of 2002.2 m² g⁻¹ and the highest specific capacitance (319 F g⁻¹), which is mainly attributed to the high utilization rate of specific surface area brought by the hierarchical porous structure. The assembled PEK-900//PEK-900 capacitor measured a specific capacity of 258 F g⁻¹ at a current density of 0.5 A g⁻¹. After 10,000 cycles of charging and discharging, the specific capacitance increased by 10%. Based on the correlation analysis of experimental data between the specific capacitance and pore structure parameters, Lasso dimensionality reduction and binary linear regression were used to reveal the relationship between the two. The residual sum of squares obtained by this method decreased by 38.4% compared to the univariate linear regression, providing a simple and reliable theoretical method for predicting the capacitance performance of biomass carbon materials.