Introduction

In the development of portable electronics such as digital telecommunication systems and systems using pulse laser techniques, the lack of lightweight, inexpensive, sustainable energy storage devices has been a limiting factor (Deng et al. 2016; Zhao et al. 2014b). Among the currently employed energy storage devices, supercapacitors, which are promising candidates because of their excellent cycling stabilities, higher power density than batteries, and higher energy density than conventional physical capacitors, have attracted significant research interest (Simon et al. 2014; Winter and Brodd 2004; Zong et al. 2016; Li et al. 2014, 2017a, b; You et al. 2017). Supercapacitors can be divided into two categories according to their charge storage mechanisms: pseudocapacitors and electrical double layer capacitors (EDLCs) (Wang et al. 2012; Winter and Brodd 2004). Pseudocapacitors store electrical energy by fast, reversible faradaic reactions (redox reactions, electrosorption or intercalation) occurring on the electrode materials, typically of metal oxides (Jeong et al. 2016; Wang et al. 2013a) or conducting polymers (Wang et al. 2014, 2015; Zhao et al. 2017). EDLCs are associated with an electrode-potential-dependent accumulation of electrostatic charge at the interface between the electrolyte and the electrode, typically of carbon (Ghosh and Lee 2012). The use of carbon here is merited by the many desirable properties of this material, such as low processing costs, chemical inertness, thermo-stability, non-toxicity, wide availability, high electrical conductivity, and a long and stable cycle life (Jiang et al. 2013; Xu et al. 2016; Zhai et al. 2011). Of the various carbon materials with different morphologies and distinct physical and chemical properties (Fan and Shen 2015; Zhai et al. 2011), carbon microspheres have several advantages, such as their regular spheroidal shape, and their adjustable porosity and particle size (which can decrease the resistance of ion diffusion and thus improve the electrochemical performance) (Duffy et al. 2012; Liu et al. 2013). Moreover, the macroporosity created as the carbon microspheres aggregate can promote the generation of ion-buffer reservoirs, which could decrease the diffusion distance of electrolyte ions to the carbon surface (Jiang et al. 2013; Zhao et al. 2014a). Hence, carbon microspheres have become one of the most promising electrode materials in supercapacitor research and development (Liu et al. 2013; Ma et al. 2014; Zhu et al. 2015). As a useful, simple approach, heteroatom doping (e.g. with nitrogen, sulfur, boron and/or phosphorus) is often used to modify the intrinsic electronic structure (Lee et al. 2012), enhance the wettability of the electrode surface (Zhu et al. 2015) or introduce new pseudocapacitive interactions (Biel et al. 2009) in order to improve the electrochemical performance of carbon electrodes (Chen et al. 2014b; Li et al. 2007). There are two main methods employed in heteroatom doping; one is by direct pyrolysis of heteroatom-containing precursors (Xu et al. 2015; Zhu et al. 2015) and the other is by post-treating carbons with dopants (Chen et al. 2014b; You et al. 2013).

Cellulose, the most abundant renewable biopolymer in nature (Klemm et al. 2005), is a promising pyrolysis precursor for providing carbonous species. Chitosan is a pseudonatural polysaccharide composed of glucosamine and N-acetyl glucosamine (Rinaudo 2006). As chitosan is derived from chitin, the second most abundant and renewable biopolymer after cellulose, it is also almost inexhaustible (Rinaudo 2006). Chitosan is also a commonly used source of carbon species self-doped with nitrogen. Cysteine, a natural amino acid containing both nitrogen and sulfur, is a good candidate for providing nitrogen and sulfur doping; it has been used to functionalize carbohydrates in the production of dual-doped carbon microspheres (Wohlgemuth et al. 2012). Recently, work on the reductive amination and preparation of 2,3-dialdehyde cellulose (DAC) beads (Lindh et al. 2014), l-cysteine-functionalized DAC (LC-DAC) beads (Ruan et al. 2016) and chitosan-crosslinked DAC (CS-DAC) beads (Ruan et al. 2018) has been reported by our group for use in applications such as palladium adsorption and Congo red dye adsorption. All three of these micrometer-sized biopolymer beads are interesting candidates for a carbon source, with or without heteroatom doping.

The aim of this study was to broaden the application of these cellulose beads to include energy storage by using DAC, CS-DAC and LC-DAC beads as a carbon source to provide microscale porous carbon beads doped with heteroatoms (N and/or S), or un-doped beads, for use as an electrode material for manufacturing supercapacitors.

Experimental section

Materials

Nanocellulose from Cladophora green algae was provided by FMC Biopolymer (Batch G-3095-10; USA). Sodium metaperiodate (NaIO4), chitosan (low molecular weight, 75–85% deacetylated), l-cysteine, sodium cyanoborohydride (NaBH3CN) and other chemicals used were of analytical or reagent grade and were purchased from Sigma-Aldrich. Deionized water was used for all the experimental procedures.

Preparation of 2,3-dialdehyde cellulose (DAC) beads

Porous DAC beads were prepared according to a previously described method (Lindh et al. 2014) with minor modifications. Briefly, Cladophora nanocellulose (20 g) was dispersed in 1.8 L of water and mixed with sodium metaperiodate (134 g, 5 mol per mol of anhydroglucose units) dissolved in 0.2 L water. The proposed periodate oxidation needs to be performed in the absence of light, which was catered for by covering the reaction vessel in aluminum foil. The reaction mixture was gently stirred at room temperature in the dark for 10 days. The reaction was then quenched via the addition of ethylene glycol and the mixture was washed four times by centrifugation/washing/re-dispersion in water to provide pure DAC. Never-dried DAC samples were kept in water at 4 °C for further use.

Preparation of chitosan-functionalized DAC (CS-DAC) beads

The CS-DAC beads were prepared according to a previously described procedure (Ruan et al. 2018). In short, never-dried DAC beads (about 500 mg dried sample) were dispersed in distilled water (250 mL) to obtain a DAC bead suspension with good dispersibility at a concentration of about 2 mg mL−1. Chitosan (960 mg) was dissolved in HCl solution (50 mL) to a concentration of 19.2 mg mL−1 and the pH of the solution was adjusted to be in the range of 2.5–3.5. The chitosan solution was added drop-wise (about 1 mL min−1) into the DAC bead suspension in a round-bottomed flask followed by the addition of the reducing agent NaBH3CN (1.25 mol per mol of chitosan). The Schiff base reaction and reduction reaction were carried out at room temperature. The reaction was stopped by centrifugation after 24 h and the products were washed twice by centrifugation/washing/redispersion with HCl solution (10 mM), twice by centrifugation/washing/redispersion with distilled water and twice by centrifugation/washing/redispersion with ethanol. The purified sample was dried in a vacuum oven at 45 °C overnight.

Preparation of l-cysteine-functionalized DAC (LC-DAC) beads

The LC-DAC beads were prepared as previously published (Ruan et al. 2016). Briefly, never-dried DAC beads (amount estimated to be equivalent to about 500 mg dried sample) were dispersed in acetate buffer solution (250 mL, 0.1 M, pH about 4.5) to obtain a DAC bead suspension with good dispersibility at a concentration of about 2 mg mL−1. l-cysteine (757 mg) was also dissolved in acetate buffer solution (50 mL, 0.1 M, pH about 4.5) to a concentration of about 15.1 mg mL−1. This solution was then added drop-wise into the DAC bead suspension at a controlled speed of about 1 mL min−1 followed by the addition of the NaBH3CN reducing agent (1.25 mol per mol of chitosan). The Schiff base and reduction reactions were performed at room temperature. The reaction was stopped by centrifugation after 24 h and the products were washed four times by centrifugation/washing/redispersion with distilled water and twice by centrifugation/washing/redispersion with ethanol. The purified samples were dried in a vacuum oven at 45 °C overnight.

Preparation of carbon microspheres from DAC, CS-DAC and LC-DAC

All as-prepared beads were pyrolyzed under a nitrogen flow at 400 °C (at a heating rate of 5 °C min−1) for 1 h and then at 900 °C for 1 h (at a heating rate of 5 °C min−1) and were then cooled to room temperature in a nitrogen atmosphere.

Scanning electron microscopy (SEM) and energy dispersive X-ray (EDX) analysis

The morphology of the produced materials was visualized in a LEO 1550 field-emission SEM instrument (Zeiss, Germany) at an accelerating voltage of 1 kV, with an in-lens secondary electron detector. EDX analysis was performed on the same LEO 1550 field-emission SEM instrument (Zeiss, Germany) but at an accelerating voltage of 20 kV, with a map model to obtain the relative content of the elements in the samples.

Nitrogen sorption isotherm analysis

The specific surface area (SSA) of all samples was recorded with an ASAP 2020 instrument (Micromeritics, USA) using a multipoint Brunauer–Emmett–Teller approach (Brunauer et al. 1938) involving N2 gas adsorption isotherm analysis. The pore size distribution was determined using density function theory (DFT) with DFT plus software from Micromeritics.

Fourier transform infrared (FTIR) spectroscopy

FTIR spectra were recorded on a Tensor 27 FTIR spectrometer (Bruker, Germany) with an attenuated total reflectance sample holder. The resolution was set to 4 cm−1 with 128 scans over a range of 2000–800 cm−1.

Thermal gravimetric analysis (TGA)

TGA was carried out on a Mettler Toledo instrument, model TGA/SDTA851, under nitrogen flow, in an inert aluminum oxide cup. The samples were heated from 30 to 900 °C at a heating rate of 5 °C min−1.

X-ray photoelectron spectroscopy (XPS)

High-resolution XPS spectra were prepared on an ESCA instrument (Physical Electronics, USA) equipped with an Al Kα source (1486.6 eV, 45 W); the analyzer was operated at 23.5 eV pass energy.

Raman spectroscopy analysis

Raman spectroscopy analyses were carried out with a Renishaw Ramascope equipped with a Leica LM optical microscope and an argon ion laser (l 1/4 514.5 nm) source.

Electrochemical characterization and analysis

Carbon electrodes composed of carbonized cellulose beads with a polyvinylidene fluoride binder in a ratio of 9:1 by weight were prepared by casting the obtained slurry onto graphite foil. The electrodes were then dried at 80 °C overnight in a vacuum oven. The carbon electrodes had a typical active mass loading of 1 mg cm−2.

All electrochemical experiments were performed at room temperature using an Autolab/GPES instrument (ECO Chemie, The Netherlands). For cyclic voltammetry (CV) measurements, the as-obtained carbon electrodes were used as the working electrode in a 3-electrode electrochemical cell; the other two electrodes were a Pt wire counter electrode and an Ag/AgCl reference electrode. An aqueous 1 M H2SO4 solution, which was purged with N2 for 10 min prior to the measurements, was used as the electrolyte.

Two-electrode symmetric supercapacitor devices hermetically heat-sealed in laminated aluminum foil were made as previously described (Wang et al. 2014, 2015). A piece of filter paper, pre-soaked with 1.0 M H2SO4 solution and used as the separator, was sandwiched between two carbon electrodes.

Galvanostatic charge–discharge measurements were carried out at different current densities between 0.5 and 20 A g−1 with a charging voltage of 0.8 V. The specific gravimetric electrode capacitance (Cg) was derived from the galvanostatic discharge curves using the formula: C g  = 4IΔt/(mΔV), where I denotes the discharge current, ΔV represents the potential change during the discharge time Δt, and m represents the total mass of carbon in the two electrodes.

Results and discussion

Preparation and characterization of CS-DAC and LC-DAC beads

DAC, CS-DAC and LC-DAC beads were prepared according to previously published work (Lindh et al. 2014; Ruan et al. 2016, 2018), with minor modifications. As shown in Scheme 1, the alcohol groups at C2 and C3 of Cladophora cellulose were oxidized by NaIO4, resulting in the formation of dialdehyde groups. Porous spherical DAC beads with a diameter of a few micrometers were obtained after washing with ethanol (Fig. 1a). A Schiff base reaction took place between the aldehyde groups in the DAC and primary amine groups in chitosan and l-cysteine to provide imine (–C=N–) groups (not shown in Scheme 1), which were reduced by NaBH3CN in situ to generate an amine. The porous morphology of CS-DAC and LC-DAC beads remained, as shown in the SEM images (Fig. 1b, c), after the reductive amination; this is similar to the surface morphology previously reported (Ruan et al. 2016, 2018). No obvious change was found after the functionalization with chitosan, but the beads modified with cysteine looked slightly more compact on the surface.

Scheme 1
scheme 1

Preparation process for doping carbonized chitosan-crosslinked 2,3-dialdehyde cellulose (CS-DAC-C) and l-cysteine-functionalized DAC (LC-DAC-C) beads

Full size image
Fig. 1
figure 1

Scanning electron microscopy images of a 2,3-dialdehyde cellulose (DAC), b chitosan-crosslinked DAC (CS-DAC), c l-cysteine-functionalized DAC (LC-DAC), d carbonized DAC (DAC-C), e carbonized CS-DAC (CS-DAC-C), and f carbonized LC-DAC (LC-DAC-C) beads

Full size image

FTIR spectra (Fig. 2a) revealed some differences between cellulose, DAC, CS-DAC and LC-DAC, such as the absorbance peak at 1730 cm−1 assigned to carbonyl groups, which appeared in DAC because aldehyde groups were produced after the oxidation of cellulose. However, the Schiff base and reduction reactions between DAC and chitosan consumed the aldehyde groups so that this absorbance peak disappeared in CS-DAC. Although similar reactions took place between DAC and l-cysteine, a strong absorbance peak at 1730 cm−1 was still observed in the LC-DAC beads; this was the result of the newly introduced carboxylic groups in l-cysteine.

Fig. 2
figure 2

a Fourier-transform infrared spectroscopy spectra and b thermogravimetric analysis curves of non-carbonized 2,3-dialdehyde cellulose (DAC), chitosan-crosslinked DAC (CS-DAC), and l-cysteine-functionalized DAC (LC-DAC) samples; c Raman spectra of carbonized DAC, CS-DAC, and LC-DAC samples

Full size image

The TGA curves (Fig. 2b) showed that DAC beads degraded at significantly lower temperatures than cellulose, demonstrating that the thermal stability of cellulose decreased after oxidation. The temperature at which the CS-DAC beads started to degrade was higher than that for the DAC beads. A similar increase in the temperature of initial degradation was not seen for the LC-DAC beads, most likely because of the difference in the amine group-containing compounds, since chitosan contains several amino groups in the molecule chain and can act as a crosslinking agent for DAC, while the single amino group in a molecule of l-cysteine cannot crosslink DAC molecules.

Carbonization and characterization of CS-DAC and LC-DAC beads

The purified, dried samples were pyrolyzed under nitrogen flow and then annealed to provide the non-doped, N-doped and N/S-co-doped carbon beads, named DAC-C (yield: 23.7%), CS-DAC-C (yield: 21.5%) and LC-DAC-C (yield: 18.5%), respectively, as illustrated in Scheme 1. The morphology of the DAC-C, CS-DAC-C and LC-DAC-C beads was analyzed by SEM, as shown in Fig. 1d–f. Obviously, the surface morphology of DAC-C changed extensively after carbonization, to form a more compact surface with some small pores. Unexpectedly, the shape of most of the DAC beads was not maintained after carbonization and only a fraction of the DAC-C retained a spherical bead shape. Interestingly, the CS-DAC-C and LC-DAC-C retained a spherical bead shape after pyrolyzation, which might have been the result of enhancement of the thermal stability as observed in the TGA curves (Fig. 2b), although this enhancement was small for LC-DAC. The porous surface morphology was maintained in the CS-DAC-C beads but disappeared from the LC-DAC-C beads (Fig. 1e, f), possibly because of the different thermal stabilities and the amine-containing reagent. Chitosan, a polyamine, served as a crosslinking reagent and can be considered as a spacer, as described by Lindh et al. (2016), to produce more porous beads whose structure was not totally broken during the carbonization. However, different processes can occur during the carbonization of LC-DAC, and these could have blocked the pores or caused the formation of smaller pores that could not be detected during the N2 sorption process, since the structure collapsed, as shown later.

Raman spectra of the carbonized cellulose beads are presented in Fig. 2c. All had very similar curves with two broad peaks located at 1335 cm−1 and 1590 cm−1. The D-band located at 1335 cm−1 corresponds to the defect-induced mode, while the G-band at 1590 cm−1 is associated with the vibration of sp2-bonded carbon atoms in a 2D hexagonal lattice. The intensity ratio of the D-band to G-band are close. The similar spectra and the presence of a broad, low-intensity D-band indicate that the carbon in all the carbonized beads is composed of thin graphene layers with a low degree of graphitization, which may contribute to a good capacitive response, which in turn would be of interest for energy storage applications (Chen et al. 2014a; Wang et al. 2013b).

The element mapping images of CS-DAC-C and LC-DAC-C recorded by an EDX detector verified the existence of nitrogen and sulfur in the doped carbonized samples (Fig. 3a, b). The blue and purple spots in the element mapping images represent the presence of the nitrogen and sulfur elements, respectively, demonstrating N-doping and N/S-co-doping in CS-DAC-C and LC-DAC-C beads. Both blue and purple spots were distributed evenly throughout the entire samples, providing evidence that the nitrogen and sulfur elements were well dispersed in the carbonized bead samples.

Fig. 3
figure 3

Elemental analysis [C (green), O (red), N (blue) and S (purple)] of the composition of a carbonized chitosan-crosslinked 2,3-dialdehyde cellulose (CS-DAC-C) and b carbonized l-cysteine-functionalized 2,3-dialdehyde cellulose (LC-DAC-C) by energy dispersive X-ray spectroscopy. (Color figure online)

Full size image

XPS detection was used to investigate the surface chemical compositions and the nature of the elements in the composites further. As shown in Fig. 4a, C 1s (285 eV) and O 1s (531 eV) were clearly visible for all samples while N 1s (398 eV) was observed in the spectra for CS-DAC, CS-DAC-C, LC-DAC and LC-DAC-C. However, S 2p (164 eV) was only seen for LC-DAC and LC-DAC-C. These results confirm the presence of nitrogen in the CS-DAC and CS-DAC-C samples, and of nitrogen and sulfur in the LC-DAC and LC-DAC-C samples. Figure 4a shows that the intensity of C 1s in the carbonized samples became stronger at the expense of the intensity of O 1s in the non-carbonized samples, indicating that the oxygen was consumed during pyrolysis. The high resolution spectra for N 1s in Fig. 4b, c can be deconvoluted into two individual peaks at binding energies of 398.8 and 400.7 eV, assigned to pyridinic N and pyrrolic N (Artyushkova et al. 2013; Kelemen et al. 2002), respectively, which were produced during the carbonization process. The high resolution spectrum of S 2p (Fig. 4d) shows two peaks at 164.2 and 165.3 eV, assigned to S 2p3/2 and S 2p1/2, respectively, for the –C–S–C– covalent bond, similar to thiophene-type sulfur (Qie et al. 2015).

Fig. 4
figure 4

a X-ray photoelectron spectroscopy survey spectra of the 2,3-dialdehyde cellulose (DAC), carbonized DAC (DAC-C), chitosan-crosslinked DAC (CS-DAC), carbonized CS-DAC (CS-DAC-C), l-cysteine-functionalized DAC (LC-DAC), and carbonized LC-DAC (LC-DAC-C) samples, and fitted high-resolution spectra for b N 1s of CS-DAC-C, and c N 1s and d S 2p of LC-DAC-C

Full size image

To further understand the porous structure of the as-prepared samples, N2 sorption isotherms were recorded and the pore volume distribution was evaluated for the carbonized beads, as shown in Fig. 5a, b. The SSA and pore-size distributions for all the samples are summarized in Table 1. The steep increase at low pressure in the sorption isotherm for DAC-C (Fig. 5a) indicates that most of the pores in this material were micropores, while Fig. 5b suggests the presence of many small micropores with diameters below 1 nm. In CS-DAC-C, the number of small micropores that were detectable by the instrument was significantly lower and some larger micropores (diameter around 1.5 nm) and macropores appeared, which significantly reduced the SSA of CS-DAC-C (Table 1). The N2 sorption isotherm of LC-DAC-C (Fig. 5a) indicated nearly no nitrogen sorption and the pore volume of the micropores decreased by three orders of magnitude for LC-DAC-C compared to DAC-C. These results confirm the changes after carbonization observed with SEM (Fig. 1).

Fig. 5
figure 5

a The nitrogen sorption isotherms (T = 77 K) and b pore-size distribution graphs for pyrolyzed 2,3-dialdehyde cellulose (DAC-C), chitosan-crosslinked DAC (CS-DAC-C), and l-cysteine-functionalized DAC (LC-DAC-C) beads. The inset in b presents an enlargement of the dashed part in the main panel

Full size image
Table 1 The specific surface areas (SSA) and pore volumes of 2,3-dialdehyde cellulose (DAC), carbonized DAC (DAC-C), chitosan-crosslinked DAC (CS-DAC), carbonized CS-DAC (CS-DAC-C), l-cysteine-functionalized DAC (LC-DAC), and carbonized LC-DAC (LC-DAC-C) bead samples
Full size table

Charge storage properties

The capacitive performances of the carbon beads were firstly evaluated in a three-electrode configuration using 1 M H2SO4 as the electrolyte. Figure 6a presents the CV curves for the carbon beads at a scan rate of 10 mV s−1. All the carbon bead samples had similar quasi-rectangular shapes with small redox peaks; the pseudocapacitive behavior was in good agreement with the nearly linear galvanostatic charge and discharge (GCD) profiles (Fig. 6b). The CS-DAC-C electrode had the largest CV loop area and the longest discharge time (compared to DAC-C and LC-DAC-C). This electrode showed nearly linear GCD profiles even at high current densities up to 20 A g−1 (Fig. 6c), which indicates that the capacitive response was fast and the resistance was relatively low. As seen in Fig. 6d, the specific capacitance of CS-DAC-C (derived from the GCD curves) was up to 242 F g−1 at a current density of 1 A g−1, which was much higher than those obtained for DAC-C (181 F g−1) and LC-DAC-C (180 F g−1), and comparable to results previously reported for cellulose-based carbon electrodes (Chen et al. 2014a; Jiang et al. 2015, 2016; Li et al. 2016c). The higher capacitance of CS-DAC-C compared to LC-DAC-C was probably the result of the higher SSA of the former (Li et al. 2016a, b; Wang et al. 2017). The fact that the DAC-C electrode, which had the highest SSA of the samples under study, exhibited a lower capacity than CS-DAC-C indicates that the doped heteroatoms contributed more to the capacitive behavior than the SSA.

Fig. 6
figure 6

a Voltammograms recorded at a scan rate of 10 mV s−1; b galvanostatic charge/discharge (GCD) curves obtained at a current density of 1 A g−1 for the carbon beads in 1 M H2SO4 electrolyte; c GCD profiles for the CS-DAC-C electrode obtained at various current densities; d rate-dependent specific electrode capacitance for the carbon beads in a three-electrode setup; e cyclic voltammetry (CV) curves at different scan rates from 5 to 200 mV s−1 and f cycling performance recorded at a scan rate of 20 mV s−1 for the CS-DAC-C-based symmetric supercapacitor. DAC 2,3-dialdehyde cellulose, DAC-C carbonized DAC, CS-DAC chitosan-crosslinked DAC, CS-DAC-C carbonized CS-DAC, LC-DAC l-cysteine-functionalized DAC, LC-DAC-C carbonized LC-DAC

Full size image

To evaluate the actual device performance of the CS-DAC-C electrode, further measurements were conducted in a symmetric supercapacitor with a two-electrode configuration in 1 M H2SO4 electrolyte. The CV curve (Fig. 6e) maintained the relatively rectangular shape without apparent distortion even at a high scan rate of 200 mV s−1, indicating a good rate performance for the CS-DAC-C-based symmetric supercapacitor. The cell capacitance calculated from the CV curves of the symmetric supercapacitor was 42 F g−1 at a scan rate of 5 mV s−1. Moreover, the CS-DAC-C-based supercapacitor showed a remarkable cyclic durability, retaining over 91.5% of its initial capacitance even after 1000 cycles at a scan rate of 20 mV s−1 (Fig. 6b). The reasonable specific capacitance and good cycling performance of the CS-DAC-C beads indicates that cellulose beads offer a promising carbon source for supercapacitor applications.

Conclusions

Micrometer-sized, porous beads based on cellulose derivatives (CS-DAC and LC-DAC) were prepared by one-pot reductive amination of DAC beads followed by pyrolysis to obtain carbon beads doped with nitrogen (CS-DAC-C) and nitrogen and sulfur (LC-DAC-C) and non-doped beads (DAC-C). Characterization by SEM, FTIR and TGA showed the changes in morphology, molecular groups and thermostability of the beads during/after the processes of modification and carbonization. XPS and EDX verified the nitrogen and sulfur doping, and the nitrogen sorption analysis results showed the different hierarchical porosities of the differently functionalized carbon beads. A suitable porosity and degree of nitrogen doping was achieved for CS-DAC-C beads and, when these were employed as the active material in electrodes, they showed the best specific capacitance of 242 F g−1 at a current density of 1 A g−1 and retained more than 91.5% of their initial capacitance after 1000 charge/discharge cycles. The reasonable specific capacitance and good cycling performance of the CS-DAC-C beads indicate that the reductive amination of DAC beads is an effective approach to providing a heteroatom-doped carbon source whose corresponding carbonaceous active materials can be used as active electrode materials in energy storage.