Elaeocarpus tectorius derived phosphorus-doped carbon as an electrode material for an asymmetric supercapacitor

Abstract

Phosphorus-doped porous carbon is prepared from a new biomass (Elaeocarpus tectorius) at three different temperatures using a facile H₃PO₄ activation approach. The physicochemical characterisation of the as-prepared carbons by X-ray diffraction, Raman spectroscopy, thermal analysis, scanning electron microscopy, N₂ adsorption–desorption isotherms and X-ray photoelectron spectroscopy indicates that the carbon obtained at 900 °C possesses a high phosphorus content, 2.5% (by mass), and a large interlayer distance of the porous carbon with more expanded channels facilitating the penetration of ions into the interlayers and a rapid adsorption of ions suitable for ultra-high volumetric capacitance. The optimized carbon (900 °C) delivers high gravimetric capacitance (385 F g⁻¹ at 0.2 A g⁻¹) and volumetric capacitance (543 F cm⁻³ at 0.2 A g⁻¹) in 1 M H₂SO₄. In 1 M Na₂SO₄ electrolyte, it still exhibits a gravimetric capacitance of 203 F g⁻¹ at 0.3 A g⁻¹ and a volumetric capacitance of 286 F cm⁻³ at 0.3 A g⁻¹. Additionally, a coin cell asymmetric device fabricated using this carbon works in a wide potential window from 0 to 1.5 V with 96% capacitance retention in 1 M H₂SO₄ aqueous electrolyte for 1000 cycles and yields a high energy density of 27 W h kg⁻¹, showing the utility for the development of wearable electronic devices.

---

1. Introduction

The production of green, renewable and sustainable energy systems for solving global issues (greenhouse gases and global warming) resulting from non-renewable fossil fuels is in demand. Carbon is extensively employed as an advanced functional material in materials science owing to its high specific surface area, high physicochemical stability, superior surface reactivity, and excellent electronic, catalytic and electrochemical properties, specifically in energy conversion and storage.^1–4 Supercapacitors are one of the advanced energy storage systems that depend on rapid ion sorption or reversible redox reactions on the surface to store energy.^5–7 Carbon-based porous materials including activated carbon,⁸ graphene,⁹ carbon nanotubes¹⁰ and carbide-derived carbon¹¹ have been comprehensively investigated for electrical double layer capacitors (EDLCs). Graphene and carbon nanotubes possess useful nanostructures for energy applications but are chiefly dependent on fossil fuel-based precursors (CH₄, pitch and phenol) and severe synthetic conditions (chemical vapor deposition, laser ablation, and electric arc-discharge methods).¹² Some synthetic routes employ toxic chemicals such as KMnO₄ and NaNO₃ which are environmentally harmful and expensive.¹³ Therefore, activated carbon materials, derived from renewable biomass, have been the focus of a more meaningful conversion of waste to value-added carbons due to its low cost, tunable pore structure/surface properties and environmentally benign and simple synthesis process. Conventional graphite has an increased interlayer distance (∼0.335 nm) and it can highly promote ion diffusion and storage in the interior of the carbon material. On the other hand, alterations of carbon materials with phosphorus, nitrogen and oxygen functional groups could possibly enrich the electron conductivity and better capacitive performance due to their electron donor properties, causing a shift of electrons from the Fermi level to the valence band in the carbon matrix. Recently a few studies have been reported on the development of functionalized carbon by doping heteroatoms (O, N, S and P) for enhancement of pseudocapacitance. High energy flexible solid state supercapacitors based on O, N, and S-tridoped carbon electrodes were reported by Ziyang Song et al.¹⁴ This work demonstrated an improvement in energy production of 90.9 W h kg⁻¹ by using a surface O, N, and S-functionalized carbon (14.85 wt%) and a polymer gel electrolyte with a high operating voltage of 3.5 V. Further, nitrogen and oxygen co-doped porous carbons have also been fabricated based on simple polymerization followed by a carbonization/activation process by Yan et al.¹⁵ Carbon obtained using this method consists of interconnected carbon spheres, which facilitate a conductive platform for high performance supercapacitors operating in a wide potential of 2.2 V in a water-in-salt electrolyte with superior cycling stability after 10 000 cycles. Activated carbon derived from various biomass precursors, such as corncob residue,¹⁶ pistachio nutshell,¹⁷ human hair,¹⁸ waste coffee beans¹⁹ and sugarcane bagasse²⁰ for supercapacitors, have been reported. There are many reports on nitrogen rich carbon from biomass for supercapacitors. Only a few reports are available for phosphorus-doped carbon electrodes for supercapacitors. Wang Yang et al. prepared a 3D-phosphorus-doped porous carbon with 367 F g⁻¹ at 0.2 A g⁻¹ in 6 M KOH.²¹ A hierarchically porous N/P co-doped carbon with high level heteroatom doping and a moderate surface area of 414 m² g⁻¹ exhibiting a specific capacitance of 212 F g⁻¹ at 0.5 A g⁻¹ was reported by Ying Zhang et al.²² All the above works indicate the necessity of a high surface area of carbon and an additional pseudocapacitive effect by heteroatoms for high performance supercapacitors. Feili Lai et al. proved dual doping of heteroatoms into carbon with an expanded interlayer distance for efficient energy storage.²³ In this work, with the purpose of selecting a new carbon source through which the above required properties could be easily achieved, a waste biosource possessing a large number of compounds having heteroatoms was chosen, namely, Elaeocarpus tectorius. Elaeocarpus tectorius (common name: bikki fruit) grown in the tropical and subtropical regions of Tamil Nadu (The Nilgiris) are generally thrown away as waste, and are accumulating continuously, creating environmental problems. Moreover, no reports on the use of these shells have been so far made in any of the research areas such as adsorption, catalysis, energy related fields, etc.

In order to achieve a high capacitance by doping phosphorus atoms into a carbon matrix, the shells were carbonized with H₃PO₄ at different temperatures of 700, 800 and 900 °C, the carbons were electrochemically characterized and the results are discussed in this work.

2. Experimental

2.1. Preparation of samples

The shells were washed well with deionized water to remove impurities, dried in an air oven for 24 h at 110 °C and then crushed into a fine powder. The powder was subjected to carbonization for 400 °C for 2 h in a muffle furnace and then cooled to room temperature.

To synthesize carbon, 50 g of the shell precursor powder was mixed with 200 g H₃PO₄ (specific gravity: 1.68 kg L⁻¹) and stirred at 80 °C for 3 h using a magnetic stirrer. The slurry obtained was kept at 110 °C in an air oven for 24 h. A solid mass was obtained, which was pyrolysed at 700, 800 and 900 °C for 3 h in a muffle furnace, cooled to room temperature, washed several times with deionized water until neutral, and then dried in an oven at 60 °C. The activated carbon thus obtained was specified as Elaeocarpus tectorius carbon (ETC-700, 800 and 900), respectively.

The steps for ETC preparation are shown in Fig. 1. The bikki shell consists of aromatic indolizidine alkaloids and polyphenols such as rudrakine, elaeocarpine, quercetin and ellagic acid. The molecules enriched with hydroxyl and carbonyl groups in conjugation with the aromatic system are the efficient sources, which could modify the functionality of carbon and enhance the capacitive behaviour as an electrode for supercapacitors.

Fig. 1 Steps for the preparation of ETC.

2.2. Characterization techniques

The crystallinity and phase-purity of the synthesized carbons were assessed by X-ray diffraction studies using a diffractometer (X’-Pert Pro PAnalytic) equipped with CuK_α (λ = 1.5184 Å) radiation. The diffraction data were collected in the 2θ range of 10–75° at a scan rate of 10° min⁻¹. Fourier transform infrared (FT-IR) spectra were recorded using a PerkinElmer system using the KBr pellet technique. Raman spectra were measured using a Witec, confocal Raman instrument with Ar ion laser 750 nm CRM200. The surface morphology and structures of the carbon were analysed using field emission scanning electron microscopy (FE-SEM, Carl Zeiss SUPRA55VP) and high-resolution transmission electron Microscopy (HRTEM, JEOL JEM2100). The elemental composition analysis was carried out using X-ray photo-electron spectroscopy (PES-AIPES beamline of Indus). The surface area and pore diameter of the carbon was characterized using N₂ sorption isotherms (BET; Belsorb Mini-II, Belmaster Software-Japan).

2.3. Electrochemical characterization

Cyclic voltammetry (CV), galvanostatic charge–discharge (CD) and electrochemical impedance spectroscopy (EIS) were carried out using a Biologic SP-200 system. The electrochemical evaluation of the electrode was executed using a three-electrode cell in 1 M H₂SO₄ aqueous solution. The working electrode was prepared by mixing 80% of ETC, 15% of super-P-carbon black and 5% of PVDF binder, by weight in N-methyl pyrrolidone to form a paste and then coated on a carbon cloth (CC) of 1 cm² area. The ETC coated electrode was dried at 80 °C for 12 h. Prior to the coating, the CC was treated with conc. HNO₃ to make it a hydrophilic surface. The areal mass loading of the ETC in 1 M H₂SO₄ is 4 mg cm⁻². A three-electrode setup, comprising ETC, a standard calomel electrode (SCE) and a platinum rod as the working, reference and counter electrodes, respectively, was used to examine the electrochemical performance of the materials. The CV and CD experiments were carried out in the potential range of 0–1 V (vs. SCE) at various scan rates ranging from 10 to 100 mV s⁻¹ and diverse current densities ranging from 0.2 to 10 A g⁻¹.

The specific and volumetric capacitance for the three-electrode system were calculated by using eqn (1)–(3),

C_s = IΔt/mΔV (1)

C_vol = C_s × ρ (2)

ρ = m × V⁻¹ (3)

where C_s is the specific capacitance (F g⁻¹), I is the current (A), Δt is the discharge time (s), m is the active mass of the electrode material (g), ΔV is the working potential window (V), C_vol is the volumetric capacitance (F cm⁻³), ρ is the packing density of the material (g cm⁻³), m is the mass of the active material and V is the volume of the cylinder. The density of the material was measured using a pellet made by mixing 95% of active material and 5% of PVDF, by weight in NMP, and stirring for 2 h before drying in an oven.

In a two-electrode setup, a CR2032 type coin cell was fabricated using ETC-900 and reduced graphene oxide (RGO) as the positive and negative electrode materials with Whatman filter paper as a separator and a carbon cloth as the current collector in 1 M H₂SO₄ electrolyte. The mass ratio of the positive and negative electrodes was calculated from the formula,

where, m₊ and m₋ represent the mass loadings of the active material on the positive (ETC-900) and the negative electrode (RGO). C₊ and C₋ represent the specific capacitance (F g⁻¹). ΔV₊ and ΔV₋ represent the potential windows of the positive and negative electrodes. The electrochemical impedance spectra (EIS) were recorded in the frequency range from 100 kHz to 0.01 Hz with a perturbation amplitude of 5 mV.

In order to compare the electrochemical performance of ETC-900 the electrochemical analysis was carried out in 1 M Na₂SO₄ aqueous medium in the three electrode configuration. The areal mass loading of the working electrode is 1.6 mg cm⁻².

3. Results and discussion

3.1. Structural characterization of ETC

IR spectra of the ETCs are displayed in Fig. 2a. Though there are many common peaks in their spectra, a slight shifting and vanishing of some peaks could be observed. All ETCs show multiplet peaks centred at 3600 cm⁻¹ corresponding to ν_OH attached to an aromatic carbon in the surface. However, a broad peak near 3500 cm⁻¹ seen in the case of ETC-700 is found to be absent in the case of the other two carbons. This means that, during activation, OH groups linked to other groups by hydrogen bonding are eliminated. Furthermore, all the carbons have peaks at 1687 cm⁻¹ indicating that the C=O group is present.^24–26 Furthermore, the presence of sharp peaks at 1524 cm⁻¹ corresponding to C–C stretching in all carbons indicate that they are all of more graphitic nature. A medium broad band observed at 1100–1200 cm⁻¹ in ETC-700 is found to be changed to a single sharp band at 1084 cm⁻¹ in the other two carbons, (in particular ETC-900 shows a clear peak) leading to the understanding that P–C–P is formed strongly in the carbon matrix. Another interesting observation is that a small band observed at 675 cm⁻¹ for ETC-800 and ETC-700 becomes very clear in ETC-900 and corresponds to νP–C.²⁷ All the above observations substantiate the presence of stable functional groups such as, –C=C–, –P–C–, –P–O–C– and –P=O– groups on the surface of ETC-900.

Fig. 2 (a) FT-IR spectra, (b) XRD patterns and (c) Raman spectra of ETC at an excitation wavelength of 750 nm.

XRD patterns of ETC-700, 800 & 900 were plotted and are shown in Fig. 2b. Two broad peaks at 23° and 43° can be indexed to the (002) and (100) crystal planes exhibiting the amorphous graphitic structure of the sample. Interestingly, all the samples show shifting of the (002) plane from 26.4° of graphite to lower values of 23.3–24°, which suggests the expansion of the interlayer spacing. Furthermore, from Bragg's law the calculated interlayer distance of 3.87 Å was observed for ETC-900 which is greater than graphite (∼3.36 Å) and is almost equal to the nitrogen-sulphur dual doped carbon reported by Feili Lai et al.²³ It has been reported that P is likely to be assimilated on the edges of the graphene, which may augment the interlayer spacing. This may be ascribed to the distortion of the graphitic planes caused by phosphorus doping.

Fig. 2c shows the Raman spectra of ETC-700, ETC-800 and ETC-900 samples. Two discrete broad peaks assigned as D and G bands are observed in all the samples. The occurrence of defects in the carbon lattice was denoted by the D band around 1320 cm⁻¹ corresponding to A_1g symmetry, whereas, the G band appearing at 1520 cm⁻¹ is attributed to the E_2g, due to the stretching vibration of the C–C bonds.²⁸ The level of disorder in the carbon samples is measured using the intensity ratio of the D and G bands. The incorporation of heteroatoms into the carbon scaffold prompts the edge defects on the carbon surface which is responsible for increased D band intensity. Hence, the ETC-900 sample shows more defects with increased defect intensity representing the disorderliness produced by P-functional groups.

To find the thermal stability of ETC-900 simultaneous TG-DTA was carried out and the thermogram is given in Fig. S1 (ESI†). The curves indicate that carbon undergoes dehydration at about 90 °C. This is due to the occluded water in the carbon. This dehydration is accompanied by a weight loss of about 25%. The second weight loss is observed with an exothermic decomposition at about 650 °C leaving no residue. From the results it may be inferred that the carbon is stable up to 550 °C.

X-ray photoelectron spectroscopy (XPS) is a surface probe technique which further attests the elemental composition and bonding state of carbon with surface functional groups. Fig. 3a shows the survey spectrum of ETC-900 with a peak at the binding energy of 285 (70.74%), 532.6 (21.16%), 133.3 (8.1%) and 192.2 eV representing C1s, O1s, P_2p and P_2s, respectively. The de-convoluted spectrum of C1s shown in Fig. 3b with four types of peaks at 284.8 (C=C), and 285 (C–C) eV in aromatic rings and at 288.1 eV (C=O). The satellite peak at 292 eV (π–π*) in the higher binding energy region, is attributed to the presence of a graphene-like microstructure carbon, which enhances the conductivity.²⁹ The O1s spectrum (Fig. 3c) has three peaks at the binding energy of 531.5 eV (C=O) carbonyl and C–O–P groups; 532.6 eV (C–O) and (C–O–C) ether functional group; 536 eV (O=C–OH) carboxyl group, all of which favor the electrochemical redox activity and wettability. Phosphorus is present in the form of P–O (pyrophosphate [P₂O₇]⁴⁻ and metaphosphate [PO₃]⁻) and it is covalently bonded to the carbon in the form of C–PO₃ at the binding energy of 133.3 eV ³⁰ as indicated in the survey spectrum (Fig. 3a). This implies that P functional groups may be positioned on the edges of the graphitic planes by bonding to more oxygens with the result of increased interlayer spacing, which is consistent with the XRD results. The distinct peak at 192.2 eV for P_2s (Fig. 3d) implies the formation of phosphorous pentoxide as an intermediate.³¹ The effective doping of P atoms into the carbon lattice is corroborated with the presence of P–C bonds, which are in agreement with FT-IR spectra. This result substantiates that ETC-900 has more oxygen and phosphorus functional groups which may produce pseudocapacitance. The higher oxygen content was mostly due to a dehydration reaction which converts the cellulose to carboxyl/ketone groups at a high temperature.

Fig. 3 (a) XPS Survey spectrum of ETC-900, (b)–(d) deconvoluted high-resolution spectra of C1s, O1s and P2s, (e) N₂-sorption/desorption isotherms, and (f) pore size distribution (PSD) curves from the NLDFT model of ETC.

N₂-Sorption analysis was employed to determine the surface area, pore size distribution (PSD) and textural properties of the ETC (Fig. 3e and f). N₂ adsorption–desorption isotherms display a steep rise in the adsorbed volume at P/P_o < 0.05, which indicates the type I isotherms of microporous materials.³² The specific surface area (SSA) of P-doped carbons escalated as the temperature is increased. ETC-900 has a high BET-SSA of 858 m² g⁻¹ with an average pore diameter of 1.8 nm and pore volume of 0.43 cm³ g⁻¹, respectively. Pore characteristics play a crucial effect on the electrochemical performance of carbon in supercapacitors. At a high temperature phosphoric acid forms its corresponding anions that can act as catalysts for the reduction of tar formation by producing polyphosphate bridges via crosslinking reactions.³⁰ Therefore, it may develop the micropores during pyrolysis. It is observed that the phosphorus-rich carbons have narrow micropores with little mesopores as tabulated in Table 1. In comparison with ETC-900, both ETC-800 and 700 have low BET-SSA and pore volumes of 0.22 and 0.24 cm³ g⁻¹ which are not very effective for ion transportation. Fig. 3f exhibits the pore distributions calculated from the non-local density functional theory (NLDFT) model. The pores for ETC samples are mainly composed of micropores (<2 nm), particularly sub-nanopores (<1 nm). The predominant sub-nanopores ranging from 0.5–1 nm provide accessible sites for penetration of aqueous electrolyte ions leading to capacitance enhancement.¹⁵ All the ETCs have the pore width centred at 0.4 to 1.8 nm with a high adsorption volume for ETC-900. The above result denotes that the increase in surface area at a higher temperature is owing to the gases resulting from decomposition and volatilization, which can generate new pores in the carbon surface.

Table 1 Textural properties and pore parameters of ETCs

Sample	S BET (m^2 g^−1)	Micro surface area (m^2 g^−1)	Meso surface area (m^2 g^−1)	Total pore volume (cm^3 g^−1)	Micropore volume (cm^3 g^−1)	Average pore size (nm)
ETC-700	267	235	33	0.24	0.12	1.72
ETC-800	339	320	19	0.15	0.22	1.76
ETC-900	858	781	77	0.43	0.38	1.80

---

The microstructure of ETC was elucidated by field-emission scanning electron microscopy (FE-SEM) and high-resolution transmission electron microscopy (HRTEM) observations. The FE-SEM image of ETC-900 is depicted in Fig. 4a. The images display an ordered porous network structure for ETC-900 at higher magnification. The decomposition of cellulose and hemicellulose at a temperature higher than 700 °C results in uniform pores for the facilitation of ion transport and offers more accessible active sites in the carbon matrix of ETC-900. Moreover, the doping of phosphorus leads to the existence of numerous micropores in the carbon matrix. The amount of phosphorus, oxygen and carbon content were found to be 2.5%, 13.8% and 84% (by mass), respectively, from SEM-EDAX analysis given in Fig. S2 (ESI†). The HRTEM image (Fig. 4b) of ETC-900 clearly reveals the porous wrinkled structure with numerous pores on the carbon matrix. This implies that phosphorus has triggered a greater structural effect on the morphologies of the carbon matrix with the elevation in the temperature. As reported earlier, by Jens Peter et al., the greater percentage of phosphorus content transfers the morphology to a tube-like structure or crumbled carbon lumps that are also wrinkled.³³ The creation of enormous mixed micro and mesopores in the carbon matrix by H₃PO₄, owing to the escaped gas from partial decomposition of H₃PO₄, (4H₃PO₄ + 10C → P₄ + 10CO + 6H₂O) above 750 °C, with the formation of P₂O₅ as an intermediate. The transformation of bulk particles to uniform pores with wrinkles takes place as the temperature increased from 700° to 900 °C which is well attained at 900 °C. Hence, this kind of carbon surface provides an ion buffering pool to enrich the electrolyte transport kinetics during the electrochemical process. The SAED pattern exhibits a polycrystalline nature with a hexagonal lattice representing the reflections of the graphene plane given in Fig. 4c. Fig. 4(d–f) shows the elemental composition of ETC-900, substantiating the presence of phosphorus in the carbon matrix. This investigation concludes that the formation of uniform pores and a graphitic nature was successful with H₃PO₄ activation at a higher temperature.

Fig. 4 (a and b) FE-SEM images of ETC-900, (c–f) HRTEM images (inset of (f): SAED pattern) and elemental mappings of ETC-900 (g) carbon, (h) oxygen and (i) phosphorus.

3.2. Electrochemical assessment of ETCs

The electrochemical performances of ETC were investigated in a three-electrode configuration in both 1 M H₂SO₄ and 1 M Na₂SO₄. Fig. 5a illustrates the cyclic voltammetry (CV) profile of ETC at a lower scan rate of 10 mV s⁻¹. It is observed that all the CV profiles are characterized by a quasi-rectangular shape with a pair of small humps indicating the predominant electrical double layer formation (EDLC). The reversible redox peaks approximately at 0.2 V imply the pseudocapacitance contribution, which becomes an added mechanism of energy storage besides the double layer mechanism. The CV curves recorded at the scan rates 10 to 100 mV s⁻¹ show a trend of enlargement of area for the carbons when heated from 700 to 900 °C. However, ETC-900 maintains the shape of the curve comparatively in a better manner (Fig. 5b). The redox peaks observed at 0.2 V is due to the reaction given below presumably PO₄³⁻ + 2H⁺+ 2e⁻ →PO₃³⁻ + H₂O. According to the Pourbaix diagram of the phosphorus–water system, the conversion of PO₄³⁻ ↔ PO₃³⁻ occurs at pH 1, which is accompanied by the electrode potential of 0.2 V.³⁴ The same is observed in the present case also, where small amounts of PO₄³⁻ in the carbon surface undergo reduction at a potential of 0.2 V in the CV profile. It is inferred from the largest area that ETC-900 has superior charge storage abilities, which may be attributed to the combination of more microporous and mesoporous volumes. In addition, the increased phosphorus content acts as an electron donor for carbon causing a shift of the conduction band from the Fermi level, which enables the adsorption of electrolyte ions.³⁵

Fig. 5 CV profiles of (a) ETCs at 10 mV s⁻¹ and (b) ETC-900 at different scan rates, (c) CD profile of ETC at 0.2 A g⁻¹, (d) CD profiles of ETC-900 at different current densities, (e) specific capacitance as a function of current density for ETCs, and (f) specific capacitance and coulombic efficiency as function of cycle for ETC-900 at 2 A g⁻¹ (inset: first and last three cycles of ETC-900).

Galvanostatic charge–discharge (CD) profiles of ETCs in three electrode cells and within the optimized potential window of 0–1 V at 0.2 A g⁻¹ are displayed in Fig. 5c. The CD profiles (Fig. 5d) at diverse current densities from 0.7–10 A g⁻¹ for ETC-900 shows isosceles triangular shape, signifying the EDLC with better electrochemical stability and charge–discharge reversibility. The discharge curves of ETC-900 without an apparent voltage drop reveal the minor internal resistance and excellent electrical conductivity. The capacitive characteristics of ETC-700 and 800 (CV and CD profiles) are listed in Fig. S3 (ESI†).

It is clear that the performances of the carbons vary with different temperatures, which is also interrelated with BET-SSA and porosity. Micropores (<2 nm) are necessary for double layer formation of charge storage at lower current density (<100 mA g⁻¹) and mesopores aid as an ion buffering source at higher scan rates (>100 mA g⁻¹). The specific capacitance of the prepared carbon decreases in the following order: ETC-900 > ETC-800 > ETC-700 as shown in Fig. 5e. The poor capacitance behaviour of ETC-800 and ETC-700 may be ascribed to reduced total pore volumes and BET-SSA. ETC-900 provides the highest gravimetric capacitance of 385, 250, 201, 189, 173 and 132 F g⁻¹ at the current densities of 0.2, 0.7, 1, 2, 5 and 10 A g⁻¹, respectively. The ETC-900 also displays a high volumetric capacitance of 543 F cm⁻³ at a current density of 0.2 A g⁻¹ and 186 F cm⁻³ at 10 A g⁻¹ as shown in Table S1 (ESI†). The high volumetric capacitance was due to the high packing density (1.41 g cm⁻³) of the ETC-900. The high packing density of the material is also due to the low mesopore/macropore content in the ETC-900 sample. Jiangying et al., reported the specific capacitance of 260 F g⁻¹ at 0.05 A g⁻¹ and 169 F g⁻¹ at 1 A g⁻¹ for nitrogen/oxygen/phosphorus decorated carbons from shrimp cells.³⁶ The highest capacitance witnessed for ETC-900 as compared to ETC-700 and ETC-800 was ascribed to the effect of the contribution of the p-electron of the P-functional groups which improve the wettability of the electrode and also induce direct pseudocapacitance through a redox reaction. The reaction is proposed below.³⁷

Another accepted reason is that pores considerably larger than electrolyte ion size and its solvation shells are necessary for higher capacitance values. The micropore with a pore size ranging between 0.77 and 1.8 nm could be completely accessible to the hydronium ions (0.36–0.42 nm)³⁸ and hydrated bisulphate ions (0.53 nm),³⁹ which agrees well in our case. The unstable oxygen functional group declines the capacitance which is suppressed by P₂O₅ formation as seen in the XPS results. However, ETC-900 could hold a specific capacitance of 132 F g⁻¹ and volumetric capacitance of 186 F cm⁻³ even at the higher current density of 10 A g⁻¹ showing outstanding rate capability. Compared to nano-sized carbons, micro-sized carbons are suitable for practical energy storage applications, as they empower a higher tap density.⁴⁰ The electrochemical performances of activated carbons derived from various biowaste sources are compared in Table 2. The long term cycle stability and coulombic efficiency of ETC-900 were reviewed at a high current density of 2 A g⁻¹, implying 71% of capacitance retention and 100% efficiency was perceived after 2000 charge–discharge cycles (Fig. 5f).

Table 2 Comparison of carbons from various biomass precursors

Biomass precursor	Activation method	S BET (m^2 g^−1)	C sp (F g^−1)	Electrolyte	Ref.
Beer lees	KOH	3560	188	0.1 M H2SO4	41
Natural wood	KOH	2925	200	6 M KOH	42
Wood saw dust	KOH	2960	236	1 M TEABF4	43
Willow catkin	KOH	645	279	6 M KOH	44
Corn grains	KOH	3199	257	6 M KOH	45
Tea leaves	KOH	2841	330	2 M KOH	40
Coconut shell	ZnCl2	1874	268	6 M KOH	46
Borassus flabellifer flower	H3PO4	633	234	1 M KOH	47
Pine cone	KOH	1515	137	1 M Na2SO4	48
Elaeocarpus tectorius shell	H3PO4	860	385	1 M H2SO4	Present work

---

Electrochemical impedance spectroscopic analysis was conducted to evaluate the electrode kinetics of the activated carbons. The total impedance can be classified into three components in Nyquist plots: (1) bulk electrolyte resistance (intercept of real axis at a high frequency region), (2) interfacial impedance has been assigned between the electrode and electrolyte solution (the diameter of the semicircle at the middle frequency region) and (3) diffuse layer resistance (i.e. impedance that is related to the intra-particle pores which occur at intermediate frequency regions).⁴⁹ The Nyquist plots for all the carbons (Fig. 6a) display a real axis with a short intercept at a high frequency region which is the indication of bulk electrolyte resistance (R_s) and it is estimated to be 1.8 Ω (ETC-900), 2.1 Ω (ETC-800) and 4 Ω (ETC-700), respectively. The equivalent circuit fitting was performed by a Randles circuit fitting and is represented in Fig. 6a. The circuit is composed of intrinsic resistance R_s, charge transfer resistance R_ct and Warburg diffusion element W. The impedance of the circuit is R_esr − j/ωC_util, where j is the imaginary unit and ω is the angular frequency (=2πf); the R_esr (equivalent series resistance) and C_util (utilizable capacitance) were calculated from the perpendicular region of real and imaginary parts in the Nyquist impedance. The rate capability of the porous carbons can be deduced from a smaller relaxation time constant using a simple relation τ = R × C.⁵⁰ The calculated relaxation time signifies the required time needed to discharge the efficient energy stored from the device, which are 2.7, 3.8 and 5.5 s for ETC-900, 800 and 700, respectively. For more detailed information regarding rate capability and capacitive characteristics, complex capacitance analysis was performed. The transformation of measured impedance to complex capacitance was carried out using the relationship .⁵¹ The real (C′(f)) and imaginary (C′′(f)) capacitance as a function of frequency are shown in Fig. 6b.

Fig. 6 (a) Nyquist plots of ETCs (inset: equivalent circuit fitting), (b) frequency dependent real and imaginary capacitance plots, (c and d) the imaginary capacitance vs. log f plots of ETC-900 and 800, respectively, and (e) Bode plots for the phase angle vs. frequency.

The capacitance as a function of frequency can be calculated using the following equations.

C′(f) = −Z′′(f)/ω|Z(f)|² (5)

C′′(f) = −Z′(f)/ω|Z(f)|² (6)

The real part of complex capacitance denotes the capacitance value as a function of frequency, whereas the imaginary part (C′′(f)) is correlated with the Kronig–Kramers (K–K) relationship. The appearance of peak shaped curves is noticed for ETC-900 and ETC-800 in the imaginary capacitance plots as a function of frequency in semi log scale as shown in Fig. 6c and d. As implied by the K–K relationship, the total capacitance (C_tot) can be calculated from the peak area (A_p) as C_tot = 1.466A_p.⁵² Furthermore, the peak frequency (f_p) at the maximum of (C′′(f)) corresponds to the representative frequency which is inversely proportional to the time constant (τ) of the system. The C_tot values and f_p are 71 F g⁻¹/45 mHz (ETC-900) and 65 F g⁻¹/20 mHz (ETC-800) with the FWHM values of 0.91 and 0.88 for ETC-900 and 800, respectively. The peak shaped curve was not observed for ETC-700 indicating the very low rate capability. The results indicate that capacitance and rate capability of the porous electrode can be evaluated well by C′′(f) vs. log f plot. The Bode plot in the low-frequency region (Fig. 6e) with a phase shift of −80° is close to −90° indicating the nature of the ideal capacitive charge storage behaviour. Among the three samples, ETC-900 has the rapid and better capacitive response as a superior electrode for a supercapacitor.

The electrochemical analysis for the ETC-900 sample in a 1 M Na₂SO₄ aqueous solution was also studied. The CV curves (Fig. 7a) are in a rectangular shape indicating the EDLC behaviour. From the GCD profiles shown in Fig. 7b and the Table S2 (ESI†), low gravimetric and volumetric capacitance was observed compared to that of an acidic electrolyte. This can be imputed to the smaller cationic radius, higher ionic mobility and larger molar ionic conductivity of H⁺ ions. The Na⁺ ions possess a large hydrated radius and the faradaic capacitance contribution is also insignificant which leads to low capacitance values.⁵³ The cycle stability of the electrode was executed at a constant current of 3 A g⁻¹ for 1000 cycles as shown in Fig. 7c. Electrochemical impedance spectroscopy (EIS) was used to reveal the ionic and electronic transport processes. As displayed in Fig. 7d, the Nyquist plot has a small ohmic resistance R_s = 2.6 Ω but is higher compared with that of an acidic electrolyte. The lower ohmic resistance is associated with the highest P_max values. As shown in the Bode plot, the phase angle is 80°, confirming that the capacitance was contributed by an EDLC behaviour.

Fig. 7 (a) CV profile of ETC-900 at 1 M Na₂SO₄, (b) GCD profile of ETC-900, (c) specific capacitance as a function of cycle number, (d) Nyquist plot of ETC-900, and (e) Bode plot of ETC-900.

3.3. Performance of asymmetric supercapacitor device

From the electrochemical and charge/discharge studies, it is ascertained that the ETC-900 electrode showed better capacitance compared to the other two samples. Therefore, ETC-900 was further subjected to its practical application in the two-electrode asymmetric cell. The asymmetric device was fabricated in the form of a coin cell bearing ETC-900 as the positive electrode and reduced graphene oxide (RGO) as the negative electrode in 1 M H₂SO₄ using a Whatman filter paper as the separator. The electrochemical property of the fabricated asymmetric device was investigated in a wide potential range of 0–1.5 V. The improvement of energy density is mutually related to the operating window. The P-doped carbon is able to achieve a high electrochemical window beyond the decomposition potential of water in an aqueous electrolyte. Hence, the assembled cell was tested at a higher operating voltage range 0–1.5 V by a gradual increase in potential above 1 V at 10 mV s⁻¹ (Fig. 8a). Fig. 8b depicts the CV profiles of the asymmetric device at different scan rates ranging from 10 to 100 mV s⁻¹. It is apparent that the CV profiles of the asymmetric device displayed a quasi-rectangular shaped curve with an operating cell voltage of 1.5 V. It should be noticed that with the increase of the scan rate, the areas under the CV curves showed a significant increase, which is usually observed for an ideal EDLC-based supercapacitor. Fig. 8c. illustrates the galvanostatic charge–discharge curves obtained for the fabricated asymmetric device at a current density of 0.4 A g⁻¹ by varying the voltage range. The charge–discharge profiles at various current densities are shown in Fig. 8d. Obviously, the presence of nearly symmetrical charge–discharge curves of the device within the investigated voltage range of 0–1.5 V was in good agreement with the three-electrode CV data.

Fig. 8 (a) CV curves of the asymmetric device at different voltage windows at 10 mV s⁻¹, (b) CV curves of the device at various scan rates from 10–100 mV s⁻¹, (c) GCD curves at 0.4 A g⁻¹ in various voltage ranges, (d) GCD curves at various current densities, (e) capacitance retention as a function of cycle number (inset: photograph of a powered red LED), and (f) a plot of specific energy vs. specific power.

The specific capacitance, energy density and power density for the two-electrode cell were calculated using the following equations:

C_sp (F g⁻¹) = 4it/(m₁ + m₂)V (7)

E = 1/2C_sp(ΔV)² (8)

P = E/Δt, (9)

where, C_sp (F g⁻¹) is the total specific capacitance of the electrode in the asymmetric device, (m₁ + m₂) is the mass of the material (mg), E (W h kg⁻¹) is the specific energy density and P (W kg⁻¹) is the specific power density of the supercapacitor device.

From the above calculations, the high specific capacitance is 100 F g⁻¹ at a current density of 0.2 A g⁻¹ and it still retains 79 F g⁻¹ at 1 A g⁻¹ with 79% capacitance retention signifying its excellent rate capability. The energy density has a direct relationship with the operating potential window. The operating potential window of phosphorus functionalized carbons is extended up to 1.3 V beyond the decomposition of water (1.23 V) as reported by Densia-Hulicova et al.⁵⁴ The result indicates that storage of energy depends mainly on the quantity of phosphorus content which could stabilize the carbon surface and ultramicropores (0.65–0.83 nm) which are responsible for double layer formation. As reported by Huang et al., the wide electrochemical window of 1.5 V in sulfuric acid for phosphorus rich carbons is due to the positive effect of phosphorus functional groups for the stabilization of carbon surface and some redox reactions for the overall improvement of energy storage.⁵⁵ The wide potential window of 0–1.5 V may be ascribed to the electrochemical reversible hydrogen storage in carbon materials. It is worth mentioning here that P-doped graphene can operate stably in aq. H₂SO₄ at 1.7 V with an excellent cycling performance and high energy densities.⁵⁶ In accordance with this result, it is observed that this system also works in a wide window up to 1.5 V. The sufficient polyphosphate functional groups in carbon materials could enrich the strength of hydrogen adsorption on the carbon surface for the electrochemically stable potential window (ESPW) in the aqueous supercapacitor. Hence, the fabricated asymmetric device delivered a maximum specific energy of 33.8 W h kg⁻¹ at a specific power of 648 W kg⁻¹. The specific power achieved the highest value of 3323 W kg⁻¹ at the specific energy of 27 W h kg⁻¹ as shown in Fig. 8f. The obtained results suggest that the ETC-900-based asymmetric device holds great promise for potential applications in energy storage devices. The long-term cycling stability of the fabricated asymmetric device was studied over 1000 cycles at 1 A g⁻¹ (shown in Fig. 8e). It is apparent that at the 1000th cycle, the asymmetric device exhibits almost 96% capacitance retention. The practical usability of the ETC-900-based asymmetric device electrode was examined by charging the device to a potential up to 1.5 V and then discharging it to light a commercial LED for more than 2 minutes.

From the overall results and discussion above, we have demonstrated that the ETC-900 has all the desirable features given below, and could act as a promising electrode for an advanced supercapacitor:

(i) enhancement of electrode wettability: the huge quantity of phosphorus and oxygen functional groups on the carbon surface favor the impregnation of the electrolyte into the inner core of the electrode material,

(ii) involvement of pseudocapacitance: redox-active functional groups such as P=O and C=O can furnish electroactive sites for capacitance enhancement,

(iii) generation of structural defects: the least electronegative phosphorus forms a C–P bond with carbon which increases the charge delocalization on the carbon atom directly to influence the surface morphology forming more open edged sites and the electron donor property of phosphorus increased the electrical conductivity of the ETC-900.

4. Conclusions

In summary, naturally abundant and unexplored bikki shells are utilized as carbon precursors for the preparation of porous P-doped carbon through a simple H₃PO₄ activation process, for use as an electrode in high performance supercapacitors. The resultant carbon at 900 °C has a reasonable specific surface area (858 m² g⁻¹), suitable PSD (1.2 nm), with huge quantity (2.5% P and 13.81% O by weight) of surface functional groups and high electrical conductivity, with ensuing excellent electrochemical properties. Furthermore, it has been witnessed that phosphorus with a dual nature such as a doping and gas expanding agent plays a crucial role in the creation of uniform micropores and increase in interlayer distance of the carbon leading to the formation of EDLCs. The pseudocapacitance contribution was observed owing to the electron-donor properties of phosphorus on a carbon material, which also enhances the supercapacitive performance. More captivatingly, P-doped porous carbon-900 can withstand a wide operating potential window in 1 M H₂SO₄ aqueous electrolyte at 1.5 V with an excellent cycle stability and a high energy density. The formation of P–O and P–C bonds is cogitated as the main factor for supercapacitive improvement. This analysis paves a way for probing and adoption of renewable resources, such as bikki shells, in the field of energy storage.

Conflicts of interest

The authors declare no conflicts of interest.

Acknowledgements

This work was supported by the Centre of Excellence for Environmental Studies (CoE-ES), TEQIP-II, Government College of Technology, Coimbatore, India by providing instrumentation facilities. We thank Lenin Elumalai, Metrohm Private Ltd, Chennai for providing NLDFT analysis.

Notes and references

1. X. Tan, S. Liu, Y. Liu, Y. Gu, G. Zeng, X. Hu, X. Wang, S. Liu and L. Jiang, Bioresour. Technol., 2017, 227, 359–372.
2. L. J. Brennan, M. T. Byrne, M. Bari and Y. K. Gun’ko, Adv. Energy Mater., 2011, 1, 472–485.
3. L. Yu, L. Hu, B. Anasori, Y. T. Liu, Q. Zhu, P. Zhang, Y. Gogotsi and B. Xu, ACS Energy Lett., 2018, 3, 1597–1603.
4. R. Zhong, Y. Liao, R. Shu, L. Ma and B. F. Sels, Green Chem., 2018, 20, 1345–1353.
5. P. Simon and Y. Gogotsi, Acc. Chem. Res., 2013, 46, 1094–1103.
6. M. Beidaghi and Y. Gogotsi, Energy Environ. Sci., 2014, 7, 867.
7. P. Simon and Y. Gogotsi, Nanoscience and Technology, Co-Published with Macmillan Publishers Ltd, UK, 2009, pp. 320–329.
8. M. Wahid, D. Puthusseri, D. Phase and S. Ogale, Energy Fuels, 2014, 28, 4233–4240.
9. L. L. Zhang, R. Zhou and X. S. Zhao, J. Mater. Chem., 2010, 20, 5983.
10. G. Lota, K. Fic and E. Frackowiak, Energy Environ. Sci., 2011, 4, 1592.
11. J. Chmiola, C. Largeot, P.-L. Taberna, P. Simon and Y. Gogotsi, Science, 2010, 328, 480–483.
12. H. Dai, Acc. Chem. Res., 2002, 35, 1035–1044.
13. W. S. Hummers and R. E. Offeman, J. Am. Chem. Soc., 1958, 80, 1339.
14. Z. Song, H. Duan, L. Li, D. Zhu, T. Cao and Y. Lv, Chem. Eng. J., 2019, 372, 1216–1225.
15. A. J. Yan, D. Zhu, Y. Lv and W. Xiong, Chin. Chem. Lett. DOI:10.1016/j.cclet.2019.05.035.
16. A.-H. Lu, W.-C. Li, Y.-Y. Xu, W.-H. Qu and X.-Q. Zhang, Bioresour. Technol., 2015, 189, 285–291.
17. J. Xu, Q. Gao, Y. Zhang, Y. Tan, W. Tian, L. Zhu and L. Jiang, Sci. Rep., 2015, 4, 5545.
18. W. Qian, F. Sun, Y. Xu, L. Qiu, C. Liu, S. Wang and F. Yan, Energy Environ. Sci., 2014, 7, 379–386.
19. C. Huang, T. Sun and D. Hulicova-Jurcakova, ChemSusChem, 2013, 6, 2330–2339.
20. K. Zou, Y. Deng, J. Chen, Y. Qian, Y. Yang, Y. Li and G. Chen, J. Power Sources, 2018, 378, 579–588.
21. G. Shao, X. Qin, W. Yang, A. Song, L. Kong and W. Yang, Carbon, 2017, 127, 557–567.
22. Y. Zhang, Q. Sun, K. Xia, B. Han, C. Zhou, Q. Gao, H. Wang, S. Pu and J. Wu, ACS Sustainable Chem. Eng., 2019, 7, 5717–5726.
23. F. Lai, G. Zhou, F. Li, Z. He, D. Yong and W. Bai, ACS Sustainable Chem. Eng., 2018, 6(3), 3143–3153.
24. Y. Guo and D. A. Rockstraw, Microporous Mesoporous Mater., 2007, 100, 12–19.
25. X. Li, E. Xin and J. Zhang, Electron. Mater. Lett., 2015, 11, 143–148.
26. Z. Lei, J. Zhang and X. S. Zhao, J. Mater. Chem., 2012, 22, 153–160.
27. L. W. Daasch and D. C. Smith, Anal. Chem., 1951, 23, 853–868.
28. T. Panja, D. Bhattacharjya and J. S. Yu, J. Mater. Chem. A, 2015, 3, 18001–18009.
29. C. Shi, K. Guo, T. Zhai, L. Hu and H. Li, Adv. Sustainable Syst., 2017, 1, 1600011.
30. X. Ma, G. Ning, C. Qi, C. Xu and J. Gao, ACS Appl. Mater. Interfaces, 2014, 6, 14415–14422.
31. Y. Wang and P. M. A. Sherwood, Surf. Sci. Spectra, 2002, 9, 159–165.
32. L. Miao, X. Qian, D. Zhu, T. Chen, G. Ping, Y. Lv, W. Xiong, Y. Liu, L. Gan and M. Liu, Chin. Chem. Lett., 2019, 30, 1445–1449.
33. J. P. Paraknowitsch and A. Thomas, Energy Environ. Sci., 2013, 6, 2839–2855.
34. H. H. Huang, Metals, 2016, 6, 23.
35. M. Guo, J. Huang, X. Kong, H. Peng, H. Shui, F. Qian, L. Zhu, W. Zhu and Q. Zhang, New Carbon Mater., 2016, 31, 352–362.
36. J. Qu, C. Geng, S. Lv, G. Shao, S. Ma and M. Wu, Electrochim. Acta, 2015, 176, 982–988.
37. Y. Wang, M. Zhang, Y. Dai, H.-Q. Wang, H. Zhang, Q. Wang, W. Hou, H. Yan, W. Li and J.-C. Zheng, J. Power Sources, 2019, 438, 227045.
38. L. Eliad, G. Salitra, A. Soffer and D. Aurbach, J. Phys. Chem. B, 2001, 105, 6880–6887.
39. M. Endo, T. Maeda, T. Takeda, Y. J. Kim, K. Koshiba, H. Hara and M. S. Dresselhaus, J. Electrochem. Soc., 2001, 148, A910.
40. Z. Li, L. Zhang, B. S. Amirkhiz, X. Tan, Z. Xu, H. Wang, B. C. Olsen, C. M. B. Holt and D. Mitlin, Adv. Energy Mater., 2012, 2, 431–437.
41. S. G. Lee, K. H. Park, W. G. Shim, M. S. Balathanigaimani and H. Moon, J. Ind. Eng. Chem., 2011, 17, 450.
42. L. Chen, T. Ji, L. Brisbin and J. Zhu, ACS Appl. Mater. Interfaces, 2015, 7, 12230–12237.
43. L. Wei, M. Sevilla, A. B. Fuertes, R. Mokaya and G. Yushin, Adv. Energy Mater., 2011, 1, 356–361.
44. K. Wang, N. Zhao, S. Lei, R. Yan, X. Tian, J. Wang, Y. Song, D. Xu, Q. Guo and L. Liu, Electrochim. Acta, 2015, 166, 1–11.
45. M. S. Balathanigaimani, W.-G. Shim, M.-J. Lee, C. Kim, J.-W. Lee and H. Moon, Electrochem. Commun., 2008, 10, 868–871.
46. L. Sun, C. Tian, M. Li, X. Meng, L. Wang, R. Wang, J. Yin and H. Fu, J. Mater. Chem. A, 2013, 1, 6462.
47. M. Sivachidambaram, J. J. Vijaya, L. J. Kennedy, R. Jothiramalingam, H. A. Al-Lohedan, M. A. Munusamy, E. Elanthamilan and J. P. Merlin, New J. Chem., 2017, 41, 3939–3949.
48. A. Bello, N. Manyala, F. Barzegar, A. A. Khaleed, D. Y. Momodu and J. K. Dangbegnon, RSC Adv., 2016, 6, 1800–1809.
49. B. A. Mei, O. Munteshari, J. Lau, B. Dunn and L. Pilon, J. Phys. Chem. C, 2018, 122, 194–206.
50. H. D. Yoo, J. H. Jang, J. H. Ryu, Y. Park and S. M. Oh, J. Power Sources, 2014, 267, 411–420.
51. J. Lee, M. A. Abbas and J. H. Bang, ACS Appl. Mater. Interfaces, 2019, 11, 14126–14135.
52. J. H. Jang and S. M. Oh, J. Electrochem. Soc., 2004, 151, 571–577.
53. B. Liu, M. Yang, H. Chen, Y. Liu, D. Yang and H. Li, J. Power Sources, 2018, 397, 1–10.
54. D. Hulicova-Jurcakova, A. M. Puziy, O. I. Poddubnaya, F. Suárez-García, J. M. D. Tascón and G. Q. Lu, J. Am. Chem. Soc., 2009, 131, 5026–5027.
55. C. Huang, T. Sun and D. Hulicova-Jurcakova, ChemSusChem, 2013, 6, 2330–2339.
56. B. Wang, D. Hulicova-Jurcakova, C. Huang, Y. Wen and L. Wang, Chem. – Eur. J., 2014, 21, 80–85.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c9nj04813h

---