Black Liquor and Wood Char-Derived Nitrogen-Doped Carbon Materials for Supercapacitors

Herein, we present a synthesis route for high-efficiency nitrogen-doped carbon materials using kraft pulping residue, black liquor, and wood charcoal as carbon sources. The synthesized nitrogen-doped carbon materials, based on black liquor and its mixture with wood charcoal, exhibited high specific surface areas (SSAs) of 2481 and 2690 m2 g−1, respectively, as well as a high volume of mesopores with an average size of 2.9–4.6 nm. The nitrogen content was approximately 3–4 at% in the synthesized nitrogen-doped carbon materials. A specific capacitance of approximately 81–142 F g−1 was achieved in a 1 M Na2SO4 aqueous solution at a current density of 0.2 A g−1. In addition, the specific capacitance retention was 99% after 1000 cycles, indicating good electrochemical stability.


Introduction
Recently, highly efficient energy storage devices have become the focus of significant research efforts [1][2][3][4]. For clean and renewable energy, innovations in supercapacitor (SC) technology are necessary for advanced energy storage systems. A large body of research on SCs has been published [5][6][7][8][9][10][11][12] and this large interest in SCs is driven by their outstanding properties, including high power density, long cycle life, fast charge/discharge rate, ecofriendly nature, flexibility, and lightness. Their broad spectrum of possible applications ranges from portable electronic devices to hybrid electric vehicles [12][13][14][15].
SC performance and successful widespread commercialization is dependent mainly on the electrode material properties which rely heavily on the selection and fabrication of relevant, cheap, effective, and efficient raw materials [12,16,17]. Various novel advanced nanostructured materials, including hybrid derivatives of two or more components coupled with high surface area carbon-based materials or conductive polymers, have been intensively studied [18][19][20][21].
Electrode surface area is an important factor to boost the capacitive performance of SCs. Porous and activated carbon, carbon nanotubes, graphene, carbon fibers, and graphitic carbon are the most widely used materials because of their suitable physical and chemical properties [22][23][24].
SC performance depends on the surface structure, porosity, and nature of the raw materials as well as the doping of carbon materials with heteroatoms such as N, O, B, S, or P [6,11]. Doping is an alternative method to increase C s and creates low-resistance electron transfer channels by reducing ion transfer resistance and diffusion path lengths. Simultaneously, the heteroatom-doped structure provides a large number of active sites for electrochemical reactions, resulting in increased C s of the SC material.

Materials and Methods
Synthesized nitrogen-doped carbon materials using BL from Horizon Pulp & Paper Ltd. Supplier (Kehra, Estonia) and its mixture with WC from Fille Ltd. Supplier (Riga, Latvia) were named as NC-ABL and NC-ABL-WC, respectively. Activation (activator ratio to carbon 3 to 1, activation temperature 800 • C) and modification with nitrogen (dicyandiamide ratio to carbon 20:1, dimethyl formamide as solvent, doping temperature800 • C) are described in detail elsewhere [37]. The carbon-based materials fabrication scheme is shown in Figure 1. The porous structure of synthesized carbon materials was studied using a Nova 4200e instrument (Quantachrome Instruments, Boynton Beach, FL, USA), as described previously [38]. Specific surface area was determined using BET theory, micropores volume was calculated using Dubinin-Raguskevich theory, and pores size distribution was assessed using Density Functional Theory.
TEM analysis of the samples was performed using a Tecnai G2 F20 X-TWIN microscope (FEI, Eindhoven, The Netherlands) [39].
The chemical composition of the carbon samples was analyzed by XPS utilizing a Kratos AXIS Supra+ spectrometer (Kratos Analytical, Manchester, UK) as described in more detail elsewhere [40]. The porous structure of synthesized carbon materials was studied using a Nova 4200e instrument (Quantachrome Instruments, Boynton Beach, FL, USA), as described previously [38]. Specific surface area was determined using BET theory, micropores volume was calculated using Dubinin-Raguskevich theory, and pores size distribution was assessed using Density Functional Theory. TEM analysis of the samples was performed using a Tecnai G2 F20 X-TWIN microscope (FEI, Eindhoven, The Netherlands) [39].
The chemical composition of the carbon samples was analyzed by XPS utilizing a Kratos AXIS Supra+ spectrometer (Kratos Analytical, Manchester, UK) as described in more detail elsewhere [40].
Raman spectra were recorded using an inVia Raman spectrometer (Renishaw, Wottonunder Edge, UK) equipped with a thermoelectrically cooled (−70 • C) CCD camera and microscope, as described elsewhere [41], except that the laser power at the sample was restricted to 0.2 mW to prevent sample damage. The excitation wavelength was 532 nm. Parameters of the Raman bands were determined by fitting the experimental contour with Gaussian and Lorentzian-Gaussian form components by using GRAMS/A1 8.0 (Thermo Scientific, Waltham, MA, USA) software.
All electrochemical measurements were performed using a three-electrode cell with cyclic voltammetry (CV). A Zennium electrochemical workstation (Zahner-Elektrik GmbH & Co.KG, Kronach-Gundelsdorf, Germany) was used. The prepared NC-ABL and NC-ABL-WC samples were coated on a titanium (Ti) sheet with a geometric surface area of 1 cm 2 and employed as the working electrode. A Pt sheet was used as the counter electrode and Ag/AgCl/KCl (3 M KCl) as the reference electrode. The carbon material inks were prepared as follows: 5 mg of each sample was dispersed ultrasonically in 0.250 mL of 2% polyvinylidenefluoride (PVDF) in an N-methyl-2-pyrrolydone (NMP) solution for 1 h. Then, the obtained slurry was sprayed onto the Ti electrode and dried in an oven at 80 • C for 2 h. The active material mass loading was 1.1 and 1.5 mg cat cm −2 for the NC-ABL and NC-ABL-WC samples, respectively.
Cyclic voltammograms (CVs) were recorded in a 1 M Na 2 SO 4 solution at scan rates between 5 and 100 mV s −1 . All solutions were deaerated with argon (Ar) for 15 min prior to the measurements. A galvanostatic charge/discharge (GCD) test was performed in a two-electrode configuration separated with a glass fiber filter (ROTILABO ® Type: CR259) for each of the carbon materials. The electrode material was prepared by mixing 80 wt% of synthesized NC-ABL or NC-ABL-WC carbons, 10 wt% polyvinylidene fluoride (PVDF), and 10 wt% carbon black. Thereafter, N-methylpyrrolidone (NMP) and ethanol were added to the mixture to make a slurry. After stirring for 12 h, the mixture was dropped onto titanium plates and the prepared electrodes were dried at 60 • C. The mass per area of the single electrode in the cell was approximately 1.25 mg cm −2 . The specific capacitance Cs (F g −1 ) values were calculated from the GCD test according to Equation (1) [42]: where C s is the specific capacitance (F g −1 ), I is the current in the charge-discharge process (A), m is the active material mass (g), ∆V is the voltage change in the supercapacitor between completely charged and discharged (V), and ∆t is the discharged time.
The electrochemical impedance spectroscopy (EIS) spectra were obtained at OCP in the frequency range from 100 kHz to 100 mMHz or 10 mHz, with a perturbation amplitude of 10 mV. The electrochemical impedance data were modeled with equivalent electric circuits (EEC) using Zview software version 3.4.

Results
This study investigated the synthesis and properties of N-doped porous carbon materials based on BL, WC, and BL composites. BL is produced in the kraft-pulping process by splitting the bonds of wood components, which generally consists of lignin and a small portion of hemicelluloses dissolved in an aqueous solution of sodium hydroxide (NaOH) and sodium sulfide (Na 2 S). In our case, BL refers to an aqueous solution of lignin residues, hemicellulose, and inorganic chemicals used in the kraft-pulping process (11.2% NaOH, 10.9% Na 2 CO 3 , and 21.9% NaHCO 3 ).
An important property of N-doped carbons is the pore size distribution, which determines the surface area and mass transfer of chemicals and ionic groups. The porous structures of the obtained carbon materials were studied using nitrogen sorption at 77 K, and the isotherms are shown in Figure 2a. The pore size distributions calculated according to quenched solid density functional theory (QSDFT) are shown in Figure 2b. The surface areas were calculated according to Brunauer-Emmett-Teller (BET) theory and are provided in Table 1. In Figure 2a, the NC-ABL adsorption-desorption isotherm forms a hysteresis, indicating significant mesopore presence in the porous structure. In the case of NC-ABL-WC, only a small hysteresis is observed, from which it can be concluded that both materials are micro-, mesoporous-in the case of NC-ABL-WC, the volume of microand mesopores is almost equal, and for NC-ABL-with the predominance of mesopores The structure and pore size distribution of the samples prepared herein were related to the chemical composition and structure of the raw material.  Resonance Raman spectroscopy was used to obtain insights into the structure of the newly prepared carbon materials. The two strong bands visible at 1350-1352 and 1601-1602 cm −1 belong to the prominent D and G modes of the graphite structure, respectively [43,44] (Figure 3). Both samples exhibited high background in the vicinity of 1100-1200 and 1500-1540 cm −1 . A more detailed analysis of such complex spectra of carbon material can be performed by fitting the experimental contour with five components [45]. The band denoted as D' has been ascribed to the disorder induced by crystal-defects, while the broad features denoted as D* and D″ bands were related with a disorder in graphitic lattice and the presence of amorphous carbon materials, respectively [45,46]. Claramunt et al., suggested that relative intensity I(D)/I(D) + I(G) of fitted components provides a measure of the crystallite size (Lα) on basal planes [45]. The dependence of this ration on the inverse value of crystallite size exhibited relatively good agreement with Cuesta model [47] for disordered carbon-based material [45]. We found that the intensity ratio I(D)/I(D) + I(G) increases from 56 to 62% comparing samples NC-ABL and NC-ABL-WC, respectively. In addition, the relative intensity of D″ band is higher in the case of sample NC-ABL-WC ( Figure 3). Thus, spectral analysis revealed that the structure of NC-ABL-WC sample is slightly more disordered compared with NC-ABL.  When liquid BL was used as a raw material, the specific surface area obtained during activation was smaller than when the WC and BL composite was used due to the formation of larger pores, with SSAs of 2481 (NC-ABL) and 2690 m 2 g −1 (NC-ABL-WC). Although the micropore volume, which directly affects the SSA, is similar for both samples, the increased total pore volume input of mesopores and reaches a maximum for the NC-ABL sample at 2.1 m 3 g −1 or 71.6% of the total pore volume. It should be noted that the mesopore volume was higher and average pore size increased (4.7 and 2.9 nm for NC-ABL and NC-ABL-WC, respectively).
Resonance Raman spectroscopy was used to obtain insights into the structure of the newly prepared carbon materials. The two strong bands visible at 1350-1352 and 1601-1602 cm −1 belong to the prominent D and G modes of the graphite structure, respectively [43,44] (Figure 3). Both samples exhibited high background in the vicinity of 1100-1200 and 1500-1540 cm −1 . A more detailed analysis of such complex spectra of carbon material can be performed by fitting the experimental contour with five components [45]. The band denoted as D' has been ascribed to the disorder induced by crystal-defects, while the broad features denoted as D* and D" bands were related with a disorder in graphitic lattice and the presence of amorphous carbon materials, respectively [45,46]. Claramunt et al., suggested that relative intensity I(D)/I(D) + I(G) of fitted components provides a measure of the crystallite size (L α ) on basal planes [45]. The dependence of this ration on the inverse value of crystallite size exhibited relatively good agreement with Cuesta model [47] for disordered carbon-based material [45]. We found that the intensity ratio I(D)/I(D) + I(G) increases from 56 to 62% comparing samples NC-ABL and NC-ABL-WC, respectively. In addition, the relative intensity of D" band is higher in the case of sample NC-ABL-WC ( Figure 3). Thus, spectral analysis revealed that the structure of NC-ABL-WC sample is slightly more disordered compared with NC-ABL. The surface elemental compositions and element-binding configurations of the NC-ABL and NC-ABL-WC samples were investigated by XPS. Figure 5 shows representative   The surface elemental compositions and element-binding configurations of the NC-ABL and NC-ABL-WC samples were investigated by XPS. Figure 5 shows representative The surface elemental compositions and element-binding configurations of the NC-ABL and NC-ABL-WC samples were investigated by XPS. Figure 5 shows representative XPS survey spectra for the NC-ABL (a) and NC-ABL-WC (b) samples, demonstrating the existence of N, C, and O.
The deconvoluted XPS spectra of C 1s, N 1s, and O 1s for NC-ABL and NC-ABL-WC carbon samples are shown in Figure 6. Intensity / a.u.

C1s
N1s O1s   The overall nitrogen content of the NC-ABL and NC-ABL-WC carbon samples was 4.33 and 3.34 at%, respectively, but most was in the pyridinic-N form, which is widely recognized as an ORR-active species [51,52]. The O 1s spectra can be deconvoluted into three peaks at 530.4-531.0, 531.9-532.2, and 533.3-533.6 eV corresponding to C=O, C-OH phenolic hydroxyl/C-O-C ether, and COOH [53,54]. Oxygen functional groups greatly enhance the specific capacitance by introducing Faraday pseudo-capacitance [54]. Figure 7 presents the electrochemical performance of the NC-ABL and NC-ABL-WC samples using CV at scan rates from 5 to 100 mV s −1 . Both samples exhibit a rectangular shape, indicative that the specific capacitance is mainly induced by the electrical double layer capacitance (EDLC) and partly by the Faraday pseudo-capacitance (PS) (Figure 7a,b). The long-term stability of the NC-ABL sample was also evaluated by recording CVs at 100 mV s −1 for up to 1000 cycles (Figure 7c). The inset represents CVs after 1st and 1000 cycles. The specific capacitance retention was 99% after 1000 cycles, indicating good long-term electrochemical stability of the electrode. The deconvoluted XPS spectra of C 1s, N 1s, and O 1s for NC-ABL and NC-ABL-WC carbon samples are shown in Figure 6

O 1s
(c) (f) The overall nitrogen content of the NC-ABL and NC-ABL-WC carbon samples was 4.33 and 3.34 at%, respectively, but most was in the pyridinic-N form, which is widely recognized as an ORR-active species [51,52]. The O 1s spectra can be deconvoluted into three peaks  Figure 7 presents the electrochemical performance of the NC-ABL and NC-ABL-WC samples using CV at scan rates from 5 to 100 mV s −1 . Both samples exhibit a rectangular shape, indicative that the specific capacitance is mainly induced by the electrical double layer capacitance (EDLC) and partly by the Faraday pseudo-capacitance (PS) (Figure  7a,b). The long-term stability of the NC-ABL sample was also evaluated by recording CVs at 100 mV s −1 for up to 1000 cycles (Figure 7c). The inset represents CVs after 1st and 1000 cycles. The specific capacitance retention was 99% after 1000 cycles, indicating good longterm electrochemical stability of the electrode. The GCD performance of the NC-ABL and NC-ABL-WC samples was evaluated in a symmetrical two-electrode cell. Figure 7d shows the GCD curves for NC-ABL-WC at different current densities. At the current density of 0.2 A g −1 , the NC-ABL-WC device shows a higher specific capacitance of 142.23 F g −1 as compared with the device with NC-ABL sample (80.93 F g −1 ) (Figure 7e). The performance of electrode materials can be influenced  The GCD performance of the NC-ABL and NC-ABL-WC samples was evaluated in a symmetrical two-electrode cell. Figure 7d shows the GCD curves for NC-ABL-WC at different current densities. At the current density of 0.2 A g −1 , the NC-ABL-WC device shows a higher specific capacitance of 142.23 F g −1 as compared with the device with NC-ABL sample (80.93 F g −1 ) (Figure 7e). The performance of electrode materials can be influenced by the shape of pores and their size distribution as well as by particular types of pores (micro-and mesopores) in their total volume [55,56]. Micropores up to 2 nm in size, accessible to electrolyte ions, provide high capacitive characteristics of the material, whereas mesopores with sizes of 2-50 nm can play the role of transport channels during the processes of charge and discharge, and thus, their presence leads to a decrease in the resistance of the SC, which is mainly due to the transfer of ions in the porous structure of the electrodes [57][58][59]. Large mesopores, as well as macropores, can serve as a buffer reservoir, due to which the ion diffusion distance can be reduced. The ratio of micropores to mesopores is also an important factor in allowing enhanced capacitance or conductivity [60]. Although both NC-ABL and NC-ABL-WC materials are micro-, mesoporous and showed the same micropore volume, the total pore volume is significantly smaller in the case of NC-ABL (28.4%). The volume of micro-and mesopores is almost equal in the case of NC-ABL-WC, whereas mesopores predominance in the structure is determined for NC-ABL. The higher electrochemical performance of NC-ABL-WC may be attributed to the porous structure, larger accessible surface area, and the volume of micro-and mesopores and advisable proportion of micropore volume to total volume of 56.5%. Additionally, the high activity of nitrogen-doped activated carbon materials may be attributed to the presence of different types of O and N functional groups (such as C=O, C-OH phenolic hydroxyl/C-O-C ether, COOH, pyridinic-N, pyrrolic-N, graphitic-N and oxidized-N), which can directly generate pseudo-capacitance via redox reactions, improve the hydrophilicity and wettability, and conductivity of the activated carbon [53,60].
The impedance spectra shown in Figure 7f represent the capacitive behavior of the samples NC-ABL and NC-ABL-AWC. The high-frequency signal deviates from theoretical predictions, possibly due to the developed porous surface of the samples. The spectra were fitted with a typical Randles circuit, but also with an additional CPE element to represent low-frequency capacitance. This element was added for fitting quality but not analyzed. For NC-ABL the impedance magnitude is larger and equivalent circuit fitting data (shown in Table 2) reveal that the cause is mainly its larger R ct value (228.4 Ω cm 2 ). In comparison, the R ct of NC-ABL-AWC is 37.7 Ω cm 2, and consequently, the impedance is lower. The difference in capacitances is not significant, but for NC-ABL-AWC C dl is larger by~10 mF cm 2 in comparison to NC-ABL. In addition, a strong low-frequency capacitive response is seen for NC-ABL-AWC, which can be attributed to the accumulation SO 4 2on the material's surface. As was mentioned above, the capacitance of SC can be divided into the double layer capacitance caused by the adsorption-desorption process of the electrolyte ions and the Faraday capacitance induced by the redox reaction on the surface of the material [54,56]. The charge storage mechanism of SC can be evaluated using the following formula [56,61]: where the measured current I (A g −1 ) at a fixed potential obeys a power law relationship with the scan rate v, both a and b are adjustable parameters. Then b = 0.5, the charge storage mechanism is considered to follow the diffusion control process, and then b = 1, the surface control process is suggested [56]. Parameter b can be determined from the slope of the dependence of log(j) vs. log(v). Figure 8 presents the obtained b values under anodic scan and its corresponding fitting variance R 2 of the NC-ABL-WC. The determined b values varies in the range of 0.76 to 1 with the change in potential and the corresponded fitting variance R 2 is close to 1. This indicates that charge storage mechanism for NC-ABL-WC material is mainly dependent on surface control process [56]. Additionally, the CV analysis was performed on Duun's method [61], allowing e uating the contribution of the surface control capacitance and the diffusion control cap itance to the total capacitance (Equation (3)): where, k1v and k2v 0.5 represent the surface capacitive and diffusion-controlled current c tributions, respectively. Dividing the equation by v 0.5 on both sides gives the follow Equation (4): At a fixed potential, the jv −0.5 scaled linearly with the square root of scan rates (v and gave the slope of k1 and a y-intercept of k2. The k1v represents the surface capacit controlled current contribution. At other fixed potentials, a series of k1 and k2 were tained. Figure 9a shows the percentage of surface control capacitance contribution to t capacitance at a scan rate of 10 mV s −1 . The contribution of surface control capacita reaches 79.6% at 5 mV s −1 and 92.0 % at 100 mV s −1 , indicating a diminution of the con bution of diffusion control capacitance to charge storage capability with the increas scan rate (Figure 9b). Additionally, the CV analysis was performed on Duun's method [61], allowing evaluating the contribution of the surface control capacitance and the diffusion control capacitance to the total capacitance (Equation (3)): where, k 1 v and k 2 v 0.5 represent the surface capacitive and diffusion-controlled current contributions, respectively. Dividing the equation by v 0.5 on both sides gives the following Equation (4): At a fixed potential, the jv −0.5 scaled linearly with the square root of scan rates (v 0.5 ), and gave the slope of k 1 and a y-intercept of k 2 . The k 1 v represents the surface capacitivecontrolled current contribution. At other fixed potentials, a series of k 1 and k 2 were obtained. Figure 9a shows the percentage of surface control capacitance contribution to total capacitance at a scan rate of 10 mV s −1 . The contribution of surface control capacitance reaches 79.6% at 5 mV s −1 and 92.0 % at 100 mV s −1 , indicating a diminution of the contribution of diffusion control capacitance to charge storage capability with the increasing scan rate (Figure 9b A comparison of the supercapacitive behavior of various carbon-based electrode materials reported in the literature and the present work is presented in Table 3, exhibiting the suitable specific capacitance of the newly prepared electrode materials. The capacitance values are close to the capacitance reported for carbon-based materials.

Conclusions
Herein, we present a simple synthesis route for high-efficiency nitrogen-doped carbon materials using kraft pulping residue, black liquor, and wood charcoal as carbon sources. The synthesized nitrogen-doped carbon materials from black liquor and wood char exhibited high SSAs of 2481 and 2690 m 2 g −1 , respectively, and contained a large volume of pores with an average size of 2.9−4.6 nm. The nitrogen content was approximately 3-4 at% in the synthesized carbon materials doped with nitrogen. A specific capacitance of approximately 81−142 F g −1 was achieved in a 1 M Na2SO4 aqueous solution at a current density of 0.2 A g -1 . In addition, the specific capacitance retention was 99% after 1000 cycles, indicating good electrochemical stability. The obtained results demonstrate that the N-doped activated carbon materials obtained from BL (NC-ABL) and its mixture with WC (NC-ABL-WC) are promising electrode materials for SC applications. A comparison of the supercapacitive behavior of various carbon-based electrode materials reported in the literature and the present work is presented in Table 3, exhibiting the suitable specific capacitance of the newly prepared electrode materials. The capacitance values are close to the capacitance reported for carbon-based materials.

Conclusions
Herein, we present a simple synthesis route for high-efficiency nitrogen-doped carbon materials using kraft pulping residue, black liquor, and wood charcoal as carbon sources. The synthesized nitrogen-doped carbon materials from black liquor and wood char exhibited high SSAs of 2481 and 2690 m 2 g −1 , respectively, and contained a large volume of pores with an average size of 2.9−4.6 nm. The nitrogen content was approximately 3-4 at% in the synthesized carbon materials doped with nitrogen. A specific capacitance of approximately 81−142 F g −1 was achieved in a 1 M Na 2 SO 4 aqueous solution at a current density of 0.2 A g -1 . In addition, the specific capacitance retention was 99% after 1000 cycles, indicating good electrochemical stability. The obtained results demonstrate that the N-doped activated carbon materials obtained from BL (NC-ABL) and its mixture with WC (NC-ABL-WC) are promising electrode materials for SC applications.