Role of Nitrogen on the Porosity, Surface, and Electrochemical Characteristics of Activated Carbon.

Surface functionalities of activated carbon can be affected by the presence of heteroatoms such as oxygen, sulfur, and nitrogen. In this work, nitrogen-doped activated carbons (NACs) were prepared from shrimp shells, and the effects of the mixing ratio (raw material to an activating agent) on the porous texture and surface functionalities were investigated. It was found that, with increasing the mixing ratio (resulting in increasing N/C), the development of mesoporosity was significantly observed. This led to decreasing microporosity and specific surface areas (SSAs). The obtained NACs exhibited nitrogen functionalities in the forms of pyridinic and pyrrolic groups. It was found that although the pyridinic-N has a detrimental effect on the SSA, it does favor the pseudocapacitance, leading to an enhancement in the ion storage capability regardless of the low SSA.

Introduction

Activated carbons (ACs) are widely used in many applications, such as solid adsorbents in pollutant adsorption[1−4] and as electrodes in fuel cells, secondary batteries, and electrochemical capacitors.[5−9] ACs are commonly produced from carbonaceous materials such as wood,[10] coal,[11,12] and coconut shells.[13,14] In addition, they can be prepared from biomass wastes from agricultural and industrial processes.[2,15−18] Coal and coconut shells are commercially used as raw materials for AC preparation. However, since biomass wastes are generally inexpensive and abundant, their effective use is desirable.

The preparation of AC involves two steps: carbonization and activation. Activation of carbon materials increases their porosity and surface area, leading to an increase in the adsorption capacity. There are two types of activation that can be employed: physical and chemical.[19,20] For physical activation, the raw material is first carbonized at high temperature (600–900 °C) then activated by steam or carbon dioxide, usually in a temperature range of 600–1200 °C. In contrast, carbonization and activation can be performed simultaneously in chemical activation at a relatively low temperature of 450–900 °C. Generally, the activating agents (i.e., NaOH, KOH, K2CO3, and H3PO4) are impregnated into the precursors followed by pyrolysis. With chemical activation, the obtained ACs generally exhibit a much higher porosity than those prepared by physical activation.

It is also well known that chemical activation can introduce a significant amount of oxygen-containing functional groups into the carbon framework, which can increase the surface functionalities and enhance the adsorption ability of ACs.[21] In addition, the surface functionalities of ACs can be improved by the presence of heteroatoms, such as nitrogen and sulfur,[22,23] which can further enhance the adsorption ability of ACs. The common ways to introduce such heterofunctionalities into the carbon framework are (i) carbonization of a suitable heteroatom-rich precursor, (ii) post-treatment in an oxygen- or nitrogen-containing atmosphere, and (iii) molecular grafting of suitable functional groups.[24−27] However, processes (ii) and (iii) generally employ several environmentally unfriendly steps involving hazardous chemical reagents. Therefore, to avoid such an environmental burden, using suitable precursors that contain the heteroatoms would be preferable for cleaner production of ACs.

Shrimp shells can be considered as a potential precursor for AC preparation as they are a byproduct of seafood processing and contain nitrogen in the forms of chitin and nutrients. Therefore, the obtained AC would contain nitrogen without a further environmental burden process. Although there have been reports regarding AC preparation from nitrogen-containing precursors such as aniline,[28] pea skin,[29] tea leaves,[30] and shrimp/prawn shells,[31−33] the roles of nitrogen on the porous textures, surface, and electrochemical characteristics are still unclear. In this work, N-doped activated carbons (NACs) were prepared from shrimp shells by varying the mixing ratio (shell to an activating agent). The obtained NACs, with different surface characteristics and porous textures, were fully characterized by N2 adsorption and X-ray photoelectron spectroscopy to fully understand the roles of nitrogen on the porous textures and surface characteristics. In addition, their electrochemical capacitive behaviors with different nitrogen functionalities were also fully characterized.

Results and Discussion

Elemental Analysis

Elemental compositions of the raw materials and NACs, including carbon, hydrogen, and oxygen, are listed in Table . It can be seen that the amount of carbon in both precursors is relatively low. Therefore, the addition of lignin can ensure a sufficient amount of carbon to form the framework.[34] By increasing the mixing ratio, WS/L/K or RS/L/K, from 1:1:2 to 3:1:2, the amount of nitrogen in the ACs (in terms of N/C) is increased as expected.

Table 1

Carbon, Hydrogen, and Nitrogen Content (Excluding Ash) and Surface Characteristics of the Raw Materials and the As-Prepared NACs.

		elemental composition
sample	mixing ratioa	C (% w/w)	H (% w/w)	N (% w/w)	N/C	SBETb (m2 g–1)	Vtotalc (cm3 g–1)	Vmicrod (cm3 g–1)	Vmesoe (cm3 g–1)
WSf		32.16	5.73	6.79	0.211
RSg		31.13	5.05	6.65	0.214
LhKi	1:2				0.000	1970	0.98	0.73	0.25
WS1	1:1:2	74.72	3.92	1.47	0.020	2245	1.44	0.83	0.61
WS2	2:1:2	79.26	4.00	3.10	0.039	2020	1.56	0.74	0.83
WS3	3:1:2	73.84	4.39	3.98	0.054	1940	1.50	0.71	0.79
RS1	1:1:2	77.53	3.76	1.68	0.022	2260	1.46	0.84	0.63
RS2	2:1:2	75.34	4.21	2.34	0.031	2140	1.50	0.79	0.71
RS3	3:1:2	75.76	4.27	3.49	0.046	1546	1.07	0.57	0.50

mixing ratio - WS/L/K or RS/L/K

SBET - specific surface area, determined by the BET method

Vtotal - total pore volume, determined at a PP0–1 of 0.95

Vmicro - micropore volume, determined by the DR method

Vmeso - mesopore volume, determined by subtracting the micropore from the total pore volume

WS - white shrimp

RS - red shrimp

L - lignin

K - potassium carbonate (K2CO3)

Porosity Analysis

The obtained NACs exhibit a combination of types I and IV of N2 sorption isotherms (Figure a), indicating the development of micro- and mesoporosity. The specific surface area of the NACs, calculated by using the BET equation (SBET), is also listed in Table together with the total pore volume (Vtotal), micropore volume (Vmicro), and mesopore volume (Vmeso). Note that, in this work, SBET was calculated from the data for 0.01 < PP0–1 < 0.05 as that obtained in a relative-pressure range of 0.1–0.3 is always overestimated for high-surface-area porous materials.[35] The obtained NACs have a high specific surface area ranging from 1546 to 2260 m2 g–1, which is comparable to those reported in the literature.[28,29,33,36]

Figure 1

(a) Nitrogen sorption isotherms at −196 °C and (b) DFT pore size distributions of the obtained NACs.

As previously mentioned, by increasing the mixing ratio, the amount of nitrogen in the ACs (in terms of N/C) is increased. However, SBET and Vmicro of the NACs are decreased with the increasing mixing ratio. In contrast, Vtotal and Vmeso tend to increase to a certain point and then decrease. These results are in good agreement with the density functional theory (DFT) pore size distribution, shown in Figure b, and TEM images of the NACs (Figure ). As the mixing ratio (or N/C) increases, the micropore tends to decrease since the intensity of the peak at about 1 nm is slightly decreased. Therefore, excessive nitrogen can widen the small pores and decrease SBET.

Figure 2

TEM images of the obtained NACs (scale bar is 100 nm).

To fully understand the effects of nitrogen on the porosity of NACs, their porous characteristics, including SBET, Vtotal, Vmicro, and Vmeso, were plotted versus N/C (Figure ). It can be clearly seen in Figure a that an increase in N/C decreases SBET. However, an unclear tendency is observed between Vtotal and N/C (Figure b). Although Vtotal tends to increase with N/C, an excessive amount (above 0.04) significantly lowers Vtotal. Since Vtotal consists of the micropores (Vmicro), mesopores (Vmeso), and even macropores, it is difficult to draw a clear relation between Vtotal and N/C. A significant change in Vmicro can be observed in Figure c. As N/C increases, Vmicro is decreased as a result of the excess amount of nitrogen, which can widen the pores. The relation between Vmeso and N/C (Figure d) is similar to that of Vtotal and N/C (Figure b) where the highest Vmeso is approximately at an N/C of 0.04. Several impacts from pore structures, Vtotal, Vmeso, and Vmicro, on SBET are also shown in Figure S1 (Supporting Information) where it is well known that Vmicro has a direct impact on SBET. As expected, SBET linearly increases with Vmicro.

Figure 3

Relation between porous texture of the ACs prepared from the white shrimp (WS), red shrimp (RS), and lignin (LK), including (a) specific surface area (SBET), (b) total pore volume (Vtotal), (c) micropore volume (Vmicro), and (d) mesopore volume (Vmeso) versus N/C.

Surface Characteristics

The surface composition of the precursors and the obtained NACs was investigated by XPS analysis, and their N 1s spectra are shown in Figure . The peak at 398.0 eV reflects the presence of pyridinic-N, whereas the peak or shoulder at 400.2 eV can be ascribed to pyrrolic-N[37,38] or the amine group (R–NH2).[39] In addition, the peak at 395.4 eV can be assigned to the tetrahedral nitrogen-bonded to sp3 C or nitrate-bonded metal clusters.[40−42] The peak at 404.1 eV can be ascribed to the nitrogen of the acetyl amide group or pyridinic oxide.[39,42,43] It can be seen in Figure that the shrimp shell precursors (WS and RS) contain nitrogen (from the acetyl amide group), which originates from chitin and its derivatives.[39] Upon carbonization/activation, the obtained NACs exhibit two major nitrogen functionalities, pyridinic-N and pyrrolic-N (Figure c–f). It can be observed that after carbonization/activation, the acetyl groups from the precursors disappeared to form other nitrogen functionalities, especially pyridinic groups (Table S1, Supporting Information). In addition, by increasing the mixing ratio, the amount of pyridinic-N was increased even with the decreased SBET (see Table ). Therefore, the pyridinic-N clearly shows a detrimental effect on SBET.

Figure 4

N 1s spectra of (a) WS, (b) RS, (c) WS2, (d) RS2, (e) WS3, and (f) RS3 samples. Note that the solid and dotted lines represent the raw data and curve fitting using the Gaussian equation, respectively.

Electrochemical Behaviors

Figure shows the cyclic voltammograms of the obtained NACs. The voltammograms display a slightly distorted rectangular shape, which indicates the ion storage through the electric double-layer (EDL) mechanism. A broad peak at about 0.4 V can be observed, which can be ascribed to the presence of pseudocapacitance. The specific capacitance for each sample was calculated from the integrated voltammogram by the following equationwhere C is the specific capacitance (F g–1), E1 and E2 are the cut-off potentials for cyclic voltammetry (V), i(E) is the instantaneous current (A), ∫i(E)dE is the total charge obtained by integrating the positive and negative sweeps of the cyclic voltammograms (C), m is the mass of the active material in the sample (g), and v is the scan rate (V s–1). The calculated specific capacitances are listed in Table . All of the samples exhibit a specific capacitance in the range of 150–180 F g–1. In addition, charge/discharge measurements were performed on the obtained NACs (Figure ). The resulting specific capacitance was calculated by the following equationwhere C is the specific capacitance (F g–1), I is the current density (1 A g–1), and is the slope of the charge/discharge plot (V s–1). The resulting specific capacitances are also listed in Table and in the range of 246–277 F g–1, which is comparable with those reported for N-doped activated carbons in the literature.[29,33,44,45] Since the NACs have different SBET and functionalities, it is not straightforward to compare them in terms of the specific capacitance. Therefore, the capacitance was calculated in terms of normalized capacitance (F m–2) by dividing the average specific capacitance (from the cyclic voltammogram and charge/discharge measurement) with its SBET (Table ). By increasing the mixing ratio, the normalized capacitance is increased. This clearly shows that the nitrogen functionalities directly affect the charge storage since the effects from SBET have been already ruled out.

Figure 5

Cyclic voltammograms of the obtained NACs with different mixing ratios: (a) WS1, (b) WS2, (c) WS3, (d) RS1, (e) RS2, and (f) RS3. The scan rate was 5 mV s–1 with a three-electrode configuration in 0.5 M H2SO4.

Table 2

Specific Capacitance and Normalized Capacitance of the NACs.

sample	specific capacitance (F g–1)	normalized capacitance (F m–2)
WS1	170a, 259b	0.0955
WS2	173a, 253b	0.1054
WS3	167a, 258b	0.1095
RS1	174a, 273b	0.0989
RS2	159a, 246b	0.0946
RS3	175a, 277b	0.1462

Specific capacitance calculated from the cyclic voltammogram

Specific capacitance calculated from charge/discharge measurement

Figure 6

Charge/discharge plots (second cycle) of the obtained NACs with different mixing ratios. The current density was 1 A g–1 with a three-electrode configuration in 0.5 M H2SO4.

To fully understand the effects of nitrogen on the charge storage behaviors, the specific capacitance and normalized capacitance were plotted versus N/C. As can be seen in Figure a, the amount of nitrogen (in terms of N/C) does not affect the specific capacitance. However, after taking SBET out of consideration, the normalized capacitance increases proportionately with N/C (Figure b). The presence of nitrogen heteroatoms can promote improvements in wettability, conductivity, and most importantly, pseudocapacitance.[33,44,46,47] To clarify the effects of nitrogen groups on the capacitance, the normalized capacitance is plotted versus the amount of each group (Figure c,d). It can be seen that the pyridinic group has a direct impact on the normalized capacitance (Figure c). This further supports the idea that pyridinic-N is responsible for the enhancement of pseudocapacitance, which is in good agreement to those reported elsewhere.[48,49] In contrast, a pyrrolic group has a detrimental effect on the normalized capacitance (Figure d). These results clearly show the effects of different types of nitrogen functionalities on the electrochemical characteristics where the pyridinic-N is found to be beneficial.

Figure 7

Relations between (a) specific capacitance and N/C, (b) normalized capacitance and N/C, (c) normalized capacitance and % pyridinic-N, and (d) normalized capacitance and % pyrrolic-N.

Conclusions

Nitrogen-doped activated carbons (NACs) were successfully prepared from shrimp shells without a chemical-doping treatment. The amount of nitrogen can be simply controlled by the mixing ratio. Excessive nitrogen led to the development of mesoporosity, which reduced the micropores and specific surface area. Although the obtained NACs contain nitrogen in the forms of pyridinic and pyrrolic groups, the pyridinic-N was found to play a role in the porous texture and the ion storage. The specific area of NACs decreases significantly with pyridinic-N. However, regardless of the effects from the surface area, pyridinic-N has a direct impact on the ion storage capability, thanks to the contribution from the pseudocapacitance.

Materials and Methods

AC Preparation

Two types of shrimp shells, red (RS) and white (WS), were used as raw materials. Boiled shrimp shells were ground into a powder form and were mixed with lignin and K2CO3 at different mixing ratios (shell/lignin/K2CO3, w/w), varying only the weight of shrimp shells (as listed in Table ) with a sufficient amount of water. They were heated at 200 °C until most of the water was evaporated and were further dried at 105 °C in an oven overnight. The impregnated shrimp shells were placed into an alumina boat and inserted into a tubular horizontal furnace. The samples were then activated at 900 °C (5 °C min–1) for 1 h under a nitrogen atmosphere. After that, the samples were soaked with 1 M HCl for an hour to eliminate the lime.[34] The samples were then filtered and dried at 105 °C overnight. For comparison, the activation of lignin using K2CO3 (lignin/K2CO3 of 1:2 w/w) was also done at 900 °C for 1 h under a nitrogen atmosphere.

Characterizations

The amounts of carbon, hydrogen, and oxygen of the shrimp shells and the obtained NACs were characterized by CHN analysis. Porous textures of the samples were determined by N2 sorption measurements at −196 °C. The specific surface area was calculated by the Brunauer–Emmett–Teller method (SBET). The total pore volume was estimated from the N2 adsorption amount at a relative pressure of 0.95 (Vtotal). The micropore volume was calculated from the Dubinin–Radushkevich (DR) equation (Vmicro).[50] Some of the NACs were characterized by X-ray photoelectron spectroscopy (XPS) using Mg Kα radiation at 12 kV and 25 mA. Some of the AC samples were subjected to transmission electron microscopy (TEM; JEOL: JEM 3100F), X-ray powder diffraction (XRD; Rigaku Miniflex 600 with a Cu Kα source, 40 kV and 15 mA), and Raman (HORIBA T64000, 532 nm) analyses.

Electrochemical characterizations were performed by using the three-electrode configuration. For the electrode preparation, the obtained NACs were mixed with carbon black and polytetrafluoroethylene at a weight ratio of 9:0.5:0.5 to form a uniform solid sheet. After that, the sheet was cut into a square shape and was pressed into the stainless-steel mesh. It was then impregnated with the electrolyte (overnight) before the measurements. Cyclic voltammetry was performed to observe the electrochemical capacitive behavior of the NAC electrodes. The electrolyte, reference, and counter electrodes are 0.5 M H2SO4, Ag/AgCl in 3 M KCl, and platinum, respectively.