Nitrogen and Sulfur Self-Doped Activated Carbon Directly Derived from Elm Flower for High-Performance Supercapacitors.

N,S-Doped activated carbon was directly prepared via a facile and cost-efficient hydrothermal reaction, followed by alkali activation of elm flower (EL)-derived biomass. The EL-derived activated carbon (ELAC) had N and S contents of 2.21 and 6.06 atom %, respectively, in addition to a high Brunauer-Emmett-Teller (BET) surface area of 2048.6 m2 g-1 and moderate pore volume of 0.88 cm3 g-1. Owing to its high BET surface area and N/S functional groups, ELAC achieved a specific capacitance of 275 F g-1 at a current density of 1 A g-1 and retained a capacitance of 216 F g-1 at 20 A g-1. In addition, a symmetric supercapacitor based on N,S-self-doped ELAC electrode provided a capacitance of 62 F g-1 at a current density of 10 A g-1, with maximum energy and power densities of 16.8 Wh kg-1 and 600 W kg-1, respectively. The capacitance retention was also high, at 87.2%, at 4 A g-1 after 5000 cycles.

Introduction

With the increasing global fossil-fuel consumption and aggravating environmental pollution, alternative eco-friendly energy sources are of great importance. n class="Chemical">Supercapacitors, also known as electrochemical capacitors, are promising energy-storage devices.[1−3] Owing to their long lifetime and fast charge–discharge properties, supercapacitors have been extensively studied and used in experimental settings.[4,5] On the basis of their energy-storage mechanism, they can be classified as either electrical double-layer capacitors (EDLCs) or pseudocapacitors. Pseudocapacitors are based on Faradic processes, whereas EDLCs are based on electrostatic charging at the electrode–electrolyte interface. Therefore, the specific surface area of its electrode significantly influences its capacitance.[6]

Among various carbon materials that can be used as electrodes in n class="Chemical">supercapacitors,[7−9] activated carbon is advantageous because of its high specific surface area and relatively low cost. Heteroatom doping contributes to the improvement in electrochemical performance of activated carbon by altering its electron–donor properties. Doping with nitrogen or sulfur improves capacitive performance by reducing charge transfer resistance and increasing the wettability between electrolyte and electrode, effectively enhancing electrochemical performance.[10−12] Nitrogen doping has been performed with both organic and inorganic nitrogen sources,[13−16] urea,[17,18] and ammonia.[19,20] In 2011, corncob-derived, nitrogen-doped activated carbon was obtained by KOH activation in NH3 atmosphere.[19] The resulting electrode exhibited a maximum specific capacitance of 185 F g–1 in the organic electrolyte. Similarly, sulfur doping is also important in energy-storage processes.[21−23] Sulfur-doped activated carbon was obtained by directly mixing sulfur flakes with activated carbon and subsequently pyrolyzing the mixture.[24] The sulfur-doped activated carbon contained many thiophene functional groups on its surface, which improve the conductivity of electrode materials. Alternatively, activated carbon can be simultaneously doped with sulfur and nitrogen.[25,26] The presence of sulfur improves the pseudocapacitive performance,[28,29] whereas that of nitrogen improves the electron-transfer properties of the carbon material.[19] For example, willow-catkin-derived nitrogen and sulfur co-doped porous carbon nanosheets were prepared by KOH activation with thiourea as a precursor to both N and S atoms.[27] Overall, these studies indicate that activated carbon with a high specific surface area and moderate heteroatom doping exhibits good capacitive performance.

Herein, we successfully prepared activated carbon samples derived from elm flower (EL) biomass, which were found to contain significant amounts of self-doped n class="Chemical">nitrogen and sulfur. Elm tree is a virescence tree, ubiquitous in northern China. Elm flowers mature at the end of spring and represent a cheap and accessible biomass source for activated carbon samples with high specific surface area.[30] Moreover, these are usually burned or piled up casually, presenting an environmental problem. Thus, their conversion into activated carbon offers a practical removal method. We showed that the EL-derived activated carbon (ELAC) exhibits a high specific surface area (2048.57 m2 g–1) and moderate N and S contents (2.12 and 6.06 atom %, respectively). As a supercapacitor electrode material, ELAC displays good capacitive performance and decent rate capability both in three- and two-electrode systems.

Results and Discussion

The preparation of nitrogen and n class="Chemical">sulfur self-doped ELAC is illustrated in Figure a. The biomass raw materials, EL, were collected from the ground and subjected to hydrothermal treatment and KOH activation. The harsh conditions of hydrothermal treatment led to the partial carbonization, hydrolysis, and partial dissolution of noncrystalline regions of the two-dimensional biomass materials.[33] Furthermore, the process decreased the crystallinity and connection between microfibers of the samples. However, their elemental composition was conserved after the hydrothermal process[34] and their porosity increased with the amount of KOH during the activation process. KOH activation also diminished the amount of spherical structures and increased that of porelike structures. The products were denoted by N,S-ELAC-x, where x is the mass ratio between KOH and HEL. Both macro- and micropores were observed within the pore structure, indicating successful KOH activation of the HEL (Figure b).

Figure 1

(a) Schematic illustration of synthesis process employed for N,S-ELAC-x. (b) Typical scanning electron microscopy (SEM) image of N,S-ELAC-2. (c) Nitrogen adsorption/desorption isotherms and inserted pore size distribution (PSD); (d) X-ray diffraction (XRD) patterns and (e) Raman spectra of the as-prepared N,S-ELAC-x.

(a) Schematic illustration of synthesis process employed for N,S-ELAC-x. (b) Typical scanning electron microscopy (SEM) image of n class="Chemical">N,S-ELAC-2. (c) Nitrogen adsorption/desorption isotherms and inserted pore size distribution (PSD); (d) X-ray diffraction (XRD) patterns and (e) Raman spectra of the as-prepared N,S-ELAC-x.

The specific surface area and pore structures of N,S-ELAC-x were further studied by n class="Chemical">nitrogen adsorption–desorption isotherm measurements. As shown in Figure c, N,S-ELAC-1 (KOH/HEL mass ratio 1:1) provided a typical type-IV adsorption–desorption isotherm,[35] which exists as an hysteresis loop at high relative pressure. The typical type-I adsorption–desorption isotherms obtained from increasing KOH mass ratios in the N,S-ELAC-2 and N,S-ELAC-3 samples suggest the presence of micropores.[36] Prominent peaks in the PSD (inset in Figure c) are observed at 0.58, 0.86, 1.17, 1.58, and 2.1 nm. Table summarizes the Brunauer–Emmett–Teller (BET) surface area and pore structure characterization parameters of the as-prepared materials. As shown in Figure S1, the HEL mainly exhibited spherical amorphous carbon structures with small specific surface area.

Table 1

Specific Surface Area, Pore Volume, Pore Diameter, and Element Content for As-Prepared HEL and N,S-ELAC-x Samples

				composition (atom %)d
samples	SBET (m2 g–1)a	Vtotal (cm3 g–1)b	Dpore diameter (nm)c	C	O	N	S
HEL	6.13		36	78.47	4.15	2.61	15.59
N,S-ELAC-1	2638.94	1.23	3.4	80.14	7.11	2.8	9.94
N,S-ELAC-2	2048.57	0.88	2.6	85.07	6.66	2.21	6.06
N,S-ELAC-3	1928.82	0.83	2.8	92.62	3.97	2.61	0.8

Specific surface area calculated by Brunauer–Emmett–Teller (BET) method.

Total pore volume of as-prepared materials.

Average pore diameter of as-prepared materials.

Element contents analyzed by X-ray photoelectron spectroscopy (XPS).

KOH activation led to a remarkable increase in the specific surface area, from 6.13 (n class="Chemical">HEL) to 2638.9 m2 g–1 (N,S-ELAC-1). Although pore volume was not detected in the preactivated material (Table ), indicating blocked pore texture, KOH activation resulted in a porous structure.[37] The N,S-ELAC-1, N,S-ELAC-2, and N,S-ELAC-3 KOH-treated samples had total pore volumes of 1.23, 0.88, and 0.83 cm3 g–1, respectively. The porous structure and significantly higher specific surface area of these samples improve the accessibility of ions at high charge/discharge rates and reduce the ion transport time from the electrolyte to electrode, which is one of the main factors necessary for satisfactory power performance.[17,38] At a KOH/HEL ratio above unity, increasing concentrations of KOH decreased the specific surface area, from 2048.6 m2 g–1 (N,S-ELAC-2) to 1928.8 m2 g–1 (N,S-ELAC-3), as excessive quantity of KOH leads to pore structure collapse.[39,40]

X-ray diffraction (XRD) analysis of the KOH-activated materials (Figure d) revealed two broad and weak peaks at 22.3 and 43.8° that can be ascribed to (002) and (100) reflections of the n class="Disease">disordered carbon layer, respectively.[41] The presence of a weak and broad (002) peak suggests the formation of a microporous structure after direct KOH activation. The spectrum of N,S-ELAC-2 sample exhibited a relatively sharper (100) peak, which implies a relatively higher degree of graphitization among the carbon atoms.[41,42] The chemical structure and degree of disorder were also characterized by Raman spectroscopy (Figure e). The D band at 1337 cm–1 is attributed to structural defects and impurities, whereas the G band at 1594 cm–1 corresponds to in-plane stretching vibrations of sp2 hybridized carbon in graphite crystallites.[36,43] High-integrated intensity ratios (ID/IG) for N,S-ELAC-1 (1.16), N,S-ELAC-2 (1.17), and N,S-ELAC-3 (1.18) were attributed to increased disorder and defect structure in the KOH-activated carbon materials. In addition, these also indicate an increase in the content of oxygen-containing functional groups in as-prepared N,S-ELAC-x samples.[44]

Transmission electron microscopy (TEM) images were observed to identify the porosity of the samples. As shown in Figure a–c, N,S-ELAC-2 clearly contains abundant micropore structures, consistent with the BET results. In addition, a high-resolution TEM image proved the existence of a n class="Disease">disordered carbon structure at the edge of the material (Figure d). The elemental composition of the as-prepared materials was determined by X-ray photoelectron spectroscopy (XPS, Figure ), which revealed S, C, N, and O contents, as evidenced by the presence of peaks at 165, 284, 400, and 532 eV, respectively. HEL had a high S content (15.59 atom %) and a moderate N content (2.61 atom %) (Table ). After KOH activation, the HEL exhibited an average S content of 6.06 atom % with no significant change in the N content. On the basis of these findings, we expect that N,S-ELAC-2 will have optimum capacitive properties.

Figure 2

(a–c) TEM images and (d) high-resolution TEM image of as-prepared N,S-ELAC-2.

Figure 3

High-resolution (a) C 1s, (b) N 1s, (c) O 1s, and (d) S 2p spectra of the N,S-ELAC-2.

The high-resolution C 1s, N 1s, O 1s, and S 2p spectra of n class="Chemical">N,S-ELAC-2 are shown in Figure . The C 1s spectrum (Figure a) can be deconvoluted into four peaks centered at 284.8 (52.49%), 285.3 (12.82%), 286.3 (14.88%), and 289.3 eV (19.78%), assigned to sp2-C hybridized C=C bonds, C–O/C–N bonds, C–O/C–S bonds, and O–C=C/O–C–N bonds,[45,46] respectively. The deconvoluted N 1s spectrum (Figure b) presented peaks at 398.5 (3.5%), 400.0 (43.25%), 401.4 (44.7%), and 403 (37.8%) eV, corresponding to pyridinic, pyrrolic, quaternary N atoms, and N oxides, respectively.[47] The presence of pyridinic, pyrrolic, quaternary N atoms, and N oxides promotes the ion transport from the electrolyte to electrode material, effectively enhancing the capacitive properties. Quaternary N atoms and N oxides also increase the capacitance of electrode materials.[31] The O 1s spectrum can be deconvoluted (Figure c) into two peaks at 532 (32.17%) and 533.4 (67.83%) eV, corresponding to O–C and O–N bonds, respectively.[38] For the S 2p spectrum (Figure d), three peaks at 164.4 (52.16%), 165.6 (26.08%), and 169.3 (21.75%) eV were observed upon deconvolution, corresponding to S 2p3/2, S 2p1/2, and oxidized sulfur, respectively. The dominant S 2p3/2 and S 2p1/2 peaks result from the spin–orbit coupling between thiophene sulfur atoms. The presence of sulfur atoms can also enhance energy storage by increasing the pseudocapacitance of the electrode.[24]

The electrochemical performance of N,S-ELAC-x as a n class="Chemical">supercapacitor electrode material was first tested in a three-electrode system and evaluated by cyclic voltammetry (CV) and galvanostatic charge–discharge (GCD) measurements. The CV curves of activated N,S-ELAC-x were measured at the scan rates ranging from 2 to 50 mV s–1 in 6 M KOH electrolyte (Figure a–c). The rectangular voltammetry curve over a potential window of −1 to −0.1 V indicated that the electrochemical behavior corresponds to that of a typical electrical double-layer capacitor (EDLC) created at the electrode–electrolyte interface. The curves for the three KOH-activated materials are shown in Figure d. At 2 mV s–1, N,S-ELAC-2 exhibited higher current density responses and higher specific capacitances. Moreover, the redox peaks were related to several of heteroatom-containing functional groups in the as-prepared materials.[48]

Figure 4

CV curves of (a) N,S-ELAC-1, (b) N,S-ELAC-2, and (c) N,S-ELAC-3 at different scan rates. (d) Comparison of N,S-ELAC-x at the scan rate of 2 mV s–1.

The GCD curves of the KOH-activated n class="Chemical">N,S-ELAC-x samples tested at six different current densities ranging from 1 to 20 A g–1 are shown in Figure a–c. These exhibited relatively good linearity and had a nearly symmetrical triangular shape, suggesting charge storage via double-layer mechanism. The higher discharge time of N,S-ELAC-2 compared to that of either N,S-ELAC-1 or N,S-ELAC-3 indicates that N,S-ELAC-2 has a relatively higher specific capacitance. Columbic efficiency was also calculated on the basis of the ratio of discharge/charge time.[52−54] As shown in Figure S2, the as-obtained N,S-ELAC-x samples displayed higher efficiency. The specific capacitance of the three KOH-activated samples (calculated using eq ) at six current densities ranging from 1 to 20 A g–1 is illustrated in Figure d. At 1 A g–1, N,S-ELAC-2 provided a capacitance of 275 F g–1. Furthermore, at 20 A g–1, a maximum specific capacitance of 216 F g–1 was retained, suggesting excellent rate performance with 78.5% capacity retention (Figure d).

Figure 5

GCD curves of (a) N,S-ELAC-1, (b) N,S-ELAC-2, and (c) N,S-ELAC-3 at different current densities. (d) Specific capacitance of N,S-ELAC-x calculated by GCD curves.

The electrochemical properties of other heteroatom-doped carbon materials are shown in Table , which include those derived from willow catkin,[27] bamboo,[49] corncobs,[19] n class="Species">banana peels,[50] and cotton[51] biomass. However, the key difference between the heteroatom-doped activated carbon materials listed in Table and our ELAC is that the former adopted an extraneous source (or precursor) for their heteroatom, adding considerable cost and time to the procedure, whereas ours derived its N,S-doped heteroatom directly from the biomass material (i.e., they are self-doped). The specific surface area of ELAC is higher compared to that of previously reported paulownia flower-based carbon materials.[55] The abundant micropores improve electrolyte contact and enhance the electrochemical performance. The sulfur functional groups also increase wettability and electrical conductivity and, consequently, the electrochemical performance. Accordingly, we successfully prepared self-doped activated carbon material with a maximum specific capacitance of 216 F g–1 at 20 A g–1. Moreover, the ELAC also exhibited a relatively high specific surface area with capacitance comparable to the highest values reported in other studies. The main factors for a satisfactory specific capacitance in electrode include high specific surface area and the N/S surface functional groups.

Table 2

Comparison of Different Nitrogen- and Sulfur-Doped Carbon for Supercapacitor Electrode Materials

samples	carbon sources	nitrogen/sulfur sources	SBET (m2 g–1)	capacitance (F g–1)	measurements condition (A g–1)	electrolyte	ref
nitrogen-doped porous carbon	potato	melamine	1052	192	10	2 M KOH	(14)
nitrogen-doped porous carbon foam	banana peel	NH3	1357.6	210.6	0.5	6 M KOH	(50)
nitrogen-doped activated carbons	corncobs	NH3	2859	185	0.4	Organic	(19)
N-doped porous carbon	cotton	melamine	617	360	0.5	6 M KOH	(51)
sulfur-doped hierarchically porous carbon	glucose	thiourea	735	252	4.0	1 M H2SO4	(23)
sulfur-doped nanoporous carbon sphere	glucose	sulfur	3357	405	0.5	6 M KOH	(21)
N,S-doped activated carbon	willow catkin	thiourea	1533	298	0.5	6 M KOH	(27)
N,S-doped activated carbon	elm flower		2048.6	216	20	6 M KOH	this work

A two-electrode symmetric capacitor was assembled with N,S-ELAC-2, the best-performing electrode in the study. This was tested in 6 M n class="Chemical">KOH electrolyte with a maximum mass loading of 10 mg. The symmetric capacitor performance at different voltages and a fixed scan rate of 50 mV s–1 is shown in Figure a. At 1.2 V, the rectangular shape of the CV curve persisted, demonstrating that the symmetric capacitor can be reversibly cycled within the voltage window of 0–1.2 V. The GCD curves of the symmetric capacitor at different current densities of 0.5–10 A g–1 (Figure b) show that their good linearity and symmetrical triangular shape were retained. The specific capacitance values, calculated using eq (Figure c), achieved up to 84 F g–1 at a current density of 0.5 A g–1 for the entire electrode. Moreover, at 10 A g–1, the maximum capacitance was 62 F g–1. The average power density of the symmetric capacitor increased from 600 to 12 000 W kg–1 with the current density (Figure d). The energy density of the symmetric capacitor remains at 12.4 Wh kg–1 at a power density of 12 000 W kg–1, higher than that of previously reported carbon symmetric supercapacitors in aqueous electrolyte, such as glucose-based porous carbon (7.01 Wh kg–1 at 7200 W kg–1),[56] mesoporous carbon nanospheres (9.1 Wh kg–1 at 3200 W kg–1),[57] and chemical-modified graphene (5.2 Wh kg–1 at 4000 W kg–1).[58] After 5000 cycles, the capacitance retention of the as-prepared symmetric capacitor was 87.2%, suggesting excellent long-term cycling stability (Figure e,f). The cycling durability of the symmetric supercapacitor was confirmed by the GCD curves after 5000 cycles in Figure d. Observably, the curves were nearly symmetric, resembling that from previous cycle. These findings demonstrate that nitrogen and sulfur co-doped carbon derived from elm flower is a good material for applications in energy conversion and storage devices. The electrochemical results revealed that ELAC material is viable as a high-performance supercapacitor electrode.

Figure 6

(a) CV curves of symmetrical supercapacitor in different operation potentials. (b) Charge–discharge curves at different current densities. (c) The specific capacitance values at different current densities. (d) Ragone plot of the symmetric cell. (e) Long-cycle stabilities at current density of 4 A g–1. (f) The charge–discharge curves of first two cycles and last two cycles in the 5000 cycle life test.

Conclusions

A novel N,S-doped activated n class="Chemical">carbon has been successfully fabricated from elm flowers via a facile prehydrothermal reaction and KOH activation for supercapacitors. Through adjusting the mass ratio of KOH to the prehydrothermal carbonized elm flowers, the resultant N,S-doped activated carbon possesses a high specific surface area, moderate pore volume, and abundant functional groups. And the as-synthesized N,S-doped activated carbon electrodes exhibited a high specific capacitance (275 F g–1 at 1 A g–1 and 216 F g–1 at 20 A g–1). In addition, the assembled symmetric supercapacitors based on this material demonstrated an energy density of 16.8 Wh kg–1 and power density of 600 W Kg–1, as well as a stable cycle life over 5000 at 4 A g–1. These high performances demonstrate that the N,S-doped activated carbon derived from elm flowers is a good potential material in energy conversion and storage devices.

Experimental Section

Sample Preparation

Elm flowers were collected from Shihezi University (China, Xinjiang Province) in April. To remove impurities, elm flowers were repeatedly washed with deionized water. The samples were dried at 100 °C for 12 h, and the flowers were crushed into powder subsequently. Powdered elm flowers of 2 g were added to 2.5 mL of concentrated n class="Chemical">sulfuric acid and 50 mL of deionized water in a 100 mL stainless-steel autoclave. The hydrothermal reaction occurred at 180 °C for 48 h. Thereafter, the mixture was cooled at room temperature and atmospheric pressure, was filtered, washed with deionized water, and then dried at 80 °C for 12 h. The resulting hydrothermally treated sample (HEL) was thoroughly mixed with KOH in separate agate mortars at mass ratios of 1:1, 1:2, and 1:3. They were preactivated at 400 °C for 30 min and then heated to 700 °C for an additional 1 h under an Ar atmosphere in a tube furnace. The activated products were then washed with 10% v/v HCl and deionized water and dried at 80 °C for 12 h. Finally, the activated carbon samples, containing both nitrogen and sulfur, were accordingly labeled N,S-ELAC-x, where x represents the mass ratio between KOH and HEL in the corresponding sample.

Material Characterization

X-ray diffraction (XRD) analysis was performed using a Bruker D8 Advance X-ray diffractometer with Cu Kα radiation. The specific surface areas were calculated according to the Brunauer–Emmett–Teller (BET) method. The pore size distribution (PSD) of the samples was obtained from density functional theory method. The surface morphologies and microstructures of the samples were observed by scanning electron microscopy (SEM, SU8010) and transmission electron microscopy (TEM, Tecnai G2 F20). Trace elements were analyzed by energy-dispersive X-ray mapping using an ISIS-300 spectrometer. X-ray photoelectron spectroscopy (XPS) was performed on an ESCALAB 250Xi (Thermo Fisher Scientific) spectrometer using monochromatic Al Kα radiation (h = 1486.6 eV) with 210 W. Raman characterization was conducted on a Horiba Jobin Yvon LabRAM HR800 Raman spectrometer.

Electrochemical Measurements

N,S-ELAC-x (5.0 mg) and n class="Chemical">poly(tetrafluoroethylene) solution (1.0 μL, 60 wt %; Aladdin) were added to 1.0 mL of ethanol. The corresponding suspension was sonicated for at least 30 min until homogenization. This was then transferred into a rectangular Ni foam current collector (1 cm × 1 cm) and vacuum dried at 80 °C for 12 h. The electrochemical performance of the as-prepared electrode material was tested using a three-electrode system in 6 M KOH electrolyte on an electrochemical workstation (CHI 760E, Shanghai). A platinum sheet and saturated calomel electrode served as counter and reference electrodes, respectively. Cyclic voltammetry (CV) and galvanostatic charge–discharge (GCD) measurements were performed over a potential window of −1 to −0.1 V at different scan rates and current densities. Finally, the specific capacitances of the as-prepared materials were calculated from the GCD data by eq (31,32)where C (F g–1) is the specific capacitance, I (A) is the charge/discharge current, Δt (s) is the discharge time, m (mg) is the mass of the active material in the electrode, and ΔV is the potential window.

The electrochemical performance of the two-electrode system was evaluated by assembling a symmetrical supercapacitor based on n class="Chemical">N,S-ELAC-2, with electrodes prepared as previous. N,S-ELAC-2 (5 mg) and a poly(tetrafluoroethylene) solution (1.0 μL, 60 wt %, Aladdin) were added to 1.0 mL of ethanol and then sonicated until homogenization. The solution was then coated onto circular Ni foam current collectors (0.785 cm2) and vacuum dried at 80 °C for 12 h. The dried electrodes were symmetrically assembled with cellulose membrane as the separator and 6 M KOH as the electrolyte in a CR2032 stainless-steel coin cell. CV and GCD analyses were conducted to evaluate the electrode performance.where m (mg) is the total mass of the electrode, I (A) is the charge/discharge current, Δt (s) is the discharge time, m (mg) is the mass of the active material in the electrode, and ΔV is the potential window.where E (Wh kg–1) is the specific energy density, P (W kg–1) is the specific power density of the symmetrical supercapacitor system, C (F g–1) is the specific capacitance of the total symmetrical system, and ΔV is the potential window of discharge.