Mesophase Pitch-Derived Carbons with High Electronic and Ionic Conductivity Levels for Electric Double-Layer Capacitors.

We develop a temperature-programmed pretreatment strategy for converting aliphatic-rich petroleum pitch into a mesophase framework, which can then be activated using KOH to produce high-performance carbons for electric double-layer capacitors (EDLCs). In the pretreatment of pitch at an optimal temperature, both the temperature ramp and holding time influence the mesophase structure, which governs the pore structure and crystallinity of the resulting activated carbon. High carbon microporosity is beneficial to capacitance maximization but detrimental to ion transport. To resolve this problem, we develop a multistep ramp incorporating aliphatic species into the aromatic framework during mesophase formation. This incorporation process produces a mesophase framework that can be activated to form carbons with high crystallinity, thereby enhancing electronic conductivity and hierarchical porosity, which improves ionic conductivity. The resulting carbon electrode is used to assemble a symmetric EDLC, which exhibits a capacitance of 160 F g-1 and excellent high-rate retention in a propylene carbonate solution of N,N-diethyl-N-methylethanaminium tetrafluoroborate. The EDLC delivers a superior specific energy of 40 Wh kg-1 (based on the total carbon mass) within a voltage range of 0-2.7 V and sustained a high energy of 24 Wh kg-1 at a high power of 50 kW kg-1. The findings of this study demonstrate that incorporating aliphatic species into aromatic mesophase frameworks plays a crucial role in regulating the crystallinity and pore structure of pitch-derived carbons for charge storage.

Introduction

Electric double-layer capacitors (EDLCs) have attracted considerable attention as auxiliary high-power sources for rechargeable batteries and fuel cells.[1−5] In an EDLC, a double layer can be rapidly formed at the electrode–electrolyte interface; thus, an EDLC exhibits a higher charge–discharge rate relative to that of a battery that requires ions to travel over long distances in solid-state electrodes.[1−3] According to the mechanism underlying double-layer formation, a larger electrode surface area for electrolyte adsorption enables the achievement of greater energy storage capacity in an EDLC. Therefore, activated carbons (ACs) with large surface areas and high porosity levels are strong candidate materials for EDLC polarizable electrodes.[1,2] Porous carbons can be fabricated by etching carbon precursors during thermal treatment for carbonization. Graphitizable hydrocarbon materials, such as pitch, constitute an ideal precursor for the production of porous carbons that possess high electronic conductivity.[6] Studies have reported the efficacy of pitch activation for producing high-porosity carbons that can serve as electrodes for EDLCs.[7−11] In addition to porosity, pore geometry is essential for rapid energy delivery because it facilitates ion transport in the interior of carbons.[6] Developing a strategy to manipulate the pore structure of carbons derived from graphitizable materials (such as pitch) is critical for promoting the applicability of EDLCs.

Pitch refers to high-molecular-weight residues derived from coal and petroleum refineries. Pitch derived from coal tar has higher aromatic/aliphatic and elemental C/H ratios than does that derived from petroleum.[12] Both types of pitch primarily comprise polycyclic aromatic units such as anthracene, phenanthrene, and pyrene, but petroleum pitch has high concentrations of naphthene and alkyl side chains, which are prone to volatilization during pitch carbonization.[12] Although carbons are formed through the coalescence of aromatic rings, the structure of carbons may be affected by the transformation patterns of naphthene and alkyl side chains during pitch carbonization. Capitalizing on its high concentrations of naphthene and alkyl side chains, we used petroleum pitch as the precursor for producing ACs; thus, we manipulated the carbon structure for effective ion transport and charge storage.

ACs are typically derived from their precursors by using heat treatment for carbonization along with etching for porosity development.[13−17] Etching can be achieved through either gasification in oxidizing gases or interaction with an alkali metal hydroxide impregnated with a precursor.[18−21] In this study, we applied KOH etching to produce ACs by using petroleum pitch precursors; we preferred alkaline etching to gasification because the gasification mechanism could not effectively create pores in a graphitizable carbon framework.[8,18−21] In KOH etching, KOH facilitates the cross-linking of constituent macromolecules to form a rigid matrix that has a low susceptibility to volatile loss and volume contraction upon heating to high temperatures.[22−24] The key mechanisms for pore formation are associated with carbon gasification by KOH-derived compounds such as K2O and K2CO3, which can be reduced by carbon to form metallic K to achieve carbon gasification and thus pore formation.[21,25,26] The residual K metal can be removed through acid washing, leading to the formation of small micropores in the carbon products.

The KOH-derived compounds K2O and K2CO3 can effectively interact with a carbonaceous framework at temperatures higher than 500 °C.[6] Pretreating a pitch precursor at temperatures less than 500 °C may be beneficial for conjugating the macromolecules in the precursor to suppress volatile evolution during heating. In addition, a controlled pretreatment process may result in a mesophase with an ordered arrangement of constituent macromolecules. Such a mesophase is beneficial for creating graphitic crystallites during carbonization, thereby achieving high electronic conductivity. The naphthene and alkyl side chains in petroleum pitch can be useful in the construction of a mesophase pitch framework to produce desired porous carbons; alternatively, they can be decomposed into volatiles during pretreatment.

The present study devised a temperature-programmed pretreatment strategy for manipulating the structure of mesophase pitch. We developed a staged heating strategy that entailed first applying low temperatures to generate local aggregates surrounded by naphthene and alkyl side chains and then applying increased temperatures to combine the aggregates for the formation of a hierarchically structured mesophase. When the derived mesophase pitch was subjected to KOH etching, it was transformed into an AC framework that had a high graphitic content for electron conduction as well as suitable proportions of mesopores and micropores for ion transport and storage. Compared to template-synthesized[27−29] or graphene-based[30,31] porous carbons that also contained nanostructured mesopores and micropores, our pitch-derived AC presents promising readiness in industry-scale production. The developed AC electrodes were used to assemble a symmetric EDLC in which 1 M N,N-diethyl-N-methylethanaminium tetrafluoroborate (TEMABF4) in propylene carbonate (PC) solvent was the electrolyte solution. The carbon electrodes exhibited a specific capacitance of 160 F g–1, which was superior to the reported value of 80–130 F g–1 for carbon electrodes in organic electrolytes.[32−35] The EDLC delivered an ultimate specific energy of 40 Wh kg–1 (based on the total carbon mass), within a voltage range of 0–2.7 V, and sustained a high energy of 24 Wh kg–1 at a high power of 50 kW kg–1. The findings of this study demonstrate the feasibility of our simple temperature-programmed pretreatment strategy for converting petroleum pitch into a hierarchically structured mesophase, which can be etched using KOH etching to produce a high-performance AC for EDLCs.

Results and Discussion

Characteristics of Pitch Precursors and Resulting ACs

We derived porous carbons from petroleum pitch samples by using KOH as the activation reagent. The pitch samples were subjected to thermal pretreatment to incorporate the constituent molecules, thereby suppressing volatilization and promoting the yield of porous carbons in the subsequent activation process. The thermal pretreatment process resulted in the formation of an ordered mesophase, which was determined to be suitable to graphitic domain growth with additional increases in temperature. Our auxiliary experiments indicated that 430 °C was the optimal temperature for obtaining a high carbon yield through pitch activation. We varied the heating rate and isothermal (430 °C) holding time to devise a temperature program for producing an ideal mesophase structure for activation into porous carbons exhibiting high capacitive performance.

Figure presents examples of the thermogravimetric analysis (TGA) profiles observed for the as-received pitch samples subjected to thermal pretreatment through the use of different ramp processes. The resulting pitch samples were designated as P-x-y, where P represents the pitch sample, x represents the heating rate (°C min–1), and y represents the holding time (h) at 430 °C. For example, P-2-1 represents the pitch sample obtained using a heating rate of 2 °C min–1 and a holding time of 1 h at 430 °C. For the multistage heating process, we heated the samples at a rate of 2 °C min–1 to 100 and 200 °C, maintained each of these temperatures for 0.5 h, continued heating to 430 °C, and finally maintained this temperature for 1 h; the resulting pitch sample was designated as P-m. Figure a presents the weight loss profiles of P-2-1. Weight loss was observed at 100–200 °C and attributed to the evaporation of adsorbed water. Weight loss was also observed at 200–430 °C and attributed to the decomposition of volatile hydrocarbon species, which may have primarily comprised naphthene and alkyl side chains.[12] The weight loss profiles observed for the samples subjected to the other single-step ramp procedures are presented in Figure S1; the total weight loss from volatile hydrocarbon decomposition of these samples was approximately 40 wt %. Polarization microscopic analysis (Figure S2a) indicated that the thermal pretreatment resulted in the formation of a mesophase framework from the pitch.

Figure 1

TGA profiles for the thermal pretreatment on the as-received pitch to produce different mesophase pitch samples: (a) P-2-1 and (b) P-m.

Figure b presents the weight loss profiles of the samples subjected to the multistage heating process, which substantially reduced the total weight loss from volatile hydrocarbon decomposition of the sample to approximately 16 wt %. The holding processes performed at 100 and 200 °C may have led to the interaction between the aliphatic and aromatic species, which suppressed the evolution of volatiles at 200–430 °C. Figure S2b depicts the polarization microscopic image of P-m, which reveals the formation of a streamlined mesophase framework. The incorporation of aliphatic species may have hindered the coalescence of the ordered mesophase domains, resulting in the formation of a streamlined framework comprising mesophase domains intersected by aliphatic borders.

The as-received and pretreated pitch samples were impregnated with KOH and carbonized at 800 °C to produce ACs. Similar to the derived pitch samples, the ACs were designated as AC-x-y, where AC represents the carbon derived from the corresponding pitch precursor, x represents the heating rate for the corresponding pitch precursor, and y represents the holding time. For example, AC-2-1 represents the AC produced from the activation of P-2-1. The carbon produced from the as-received pitch was designated as as-AC. Figure shows the high-resolution transmission electron microscopy (TEM) images of as-AC and AC-m. Both porous carbons exhibited an amorphous framework with a distribution of interconnected nanochannels that were interstices between turbostratic graphene sheets and clusters.[36−38] The high concentration of nanochannels in the carbons signifies that the KOH activation process engendered high porosity in the resulting carbons. The structure of as-AC (Figure a) was determined to be highly isotropic, whereas the framework of AC-m (Figure b) exhibited local molecular ordering, thus forming graphitic crystallites. These TEM results are consistent with the results obtained from comparing selected-area electron diffraction (SAED) patterns (insets of Figure ). Specifically, the SAED pattern of AC-m demonstrated clear diffraction rings on the graphitic planes of this carbon, whereas that of as-AC did not reveal any diffraction ring within the carbon structure.

Figure 2

TEM images of (a) as-AC and (b) AC-m. The insets of the TEM images show the selected-area electron diffraction patterns.

Figure S3 shows the TEM images and SAED patterns of the carbons derived from the other pitch samples. All pitch pretreatment processes induced crystallization during the subsequent carbonization process; this is because TEM revealed the presence of local crystallites and diffraction rings in the carbons. The presence of crystallites in the carbons can be attributed to the formation of the mesophase framework from the pitch precursors. A comparison of all SAED patterns revealed that AC-m exhibited additional diffraction rings, which were assigned to the (103) and (201) planes, indicating that the multistep pretreatment process engendered a pitch framework that was susceptible to graphitic crystallite formation during the activation process.

The X-ray diffraction (XRD) patterns of the ACs (Figure S4) were determined to support the TEM results. Specifically, the XRD pattern of as-AC did not exhibit a diffraction peak, indicating a highly amorphous structure. By contrast, the pattern of the ACs derived from the pretreated pitch precursors exhibited a broad diffraction peak at a 2θ value of approximately 25°, which was attributed to the graphitic (002) plane with an interlayer spacing larger than that of graphite (2θ = 26.6°); this result demonstrates the effectiveness of the thermal pretreatment of the pitch in promoting the stacking of the graphene layers of the resulting ACs. Compared with the 2θ values of the other ACs, the 2θ value of AC-m (25.3°) was closer to that of graphite, indicating that AC-m had a higher degree of graphitization relative to the other ACs. In addition, the crystallite size of AC-m, estimated from the full width at half-maximum of the (002) peak, was larger than those of the other ACs (Table S1), confirming that the multistep pretreatment process was superior to the other pretreatment processes in promoting crystallinity during activation.

Figure S5 illustrates the N2 adsorption–desorption isotherms of the ACs. The isotherms of the ACs exhibited fairly horizontal plateaus and were thus determined to be typical of microporous carbons.[39−42] Compared with the isotherm of as-AC, the isotherms of the ACs derived from the pretreated pitch precursors exhibited wide knees, indicating that the pitch pretreatment process resulted in a large micropore size distribution. Increasing the heating rate and maintaining the treatment temperature (at 430 °C) for 1 h (Figure S5b– d for AC-x-1, where x = 1.5, 2, or 5) resulted in an increase in adsorption at a P/P0 of >0.2; this result signifies that a high heating rate increased the formation of large micropores and mesopores. When the heating rate was 5 °C min–1, extending the holding time (at 430 °C) from 1 to 2 h (Figure S5d,e for AC-5-y, where y = 1 or 2) engendered an increase in N2 adsorption at a P/P0 of <0.25, but the total adsorption capacity remained unchanged; this result indicates that extending the holding time at 430 °C promoted micropore formation but suppressed mesopore formation. When the heating rate was 2 °C min–1, inserting two low-temperature treatment steps at 100 and 200 °C resulted in an increase in adsorption capacity throughout the P/P0 range (Figure S5f for AC-m), signifying that the multistep treatment process promoted both micropore and mesopore formation.

The isotherms revealed that the AC-m sample exhibited hysteresis at P/P0 values of >0.3 (Figure S5f, inset) but the other carbons did not. For AC-m, the adsorption and desorption branches of the hysteresis loop of the isotherm were horizontal and parallel to each other, which is typical of a type-IV hysteresis phenomenon.[38] We determined that AC-m was characterized by a hierarchical pore structure in which mesopores served as the main channels for molecular transport and micropores located in the mesopore walls accommodated molecules (see the AC derived from the pitch subjected to multistep pretreatment in Scheme ).[43−46] The other carbons that exhibited no hysteresis comprised mesopores at the pore entrance and micropores within the carbon particles (see the AC derived from the pitch subjected to single-step pretreatment in Scheme ).

Scheme 1

Schematic Illustration of the Possible Mechanisms for the Synthesis of the ACs Using Different Pitch Pretreatment Steps

(a) Single-step pretreatment; (b) multistep pretreatment.

The realistic total surface area (St), total pore volume (Vt), and pore size distribution of the ACs were derived from a simulation of N2 sorption data by using a two-dimensional nonlocal density functional theory model for carbons with heterogeneous surfaces (HS-2D-NLDFT).[39−42]Figure S6 and Table present the simulated results and corresponding values. Although the conventional Brunauer–Emmett–Teller (BET) equation is unsuitable for analyzing microporous carbons, for comparison, we applied the equation to derive the surface area (SBET) of the ACs; the results are presented in Table . All ACs had high St and SBET values (approximately 1900 and 2500 m2 g–1, respectively), reflecting the effectiveness of the KOH activation process. The St values were primarily contributed by the micropore area (Smi), which predominantly determines the energy storage capacity of a sample.[38,47−49] We observed that the pretreatment pattern affected the micropore/mesopore volume ratio. When the pretreatment holding time was 1 h (at 430 °C), reducing the heating rate from 5 to 1.5 °C min–1 resulted in decreased Vt, decreased mesopore fraction, and increased microporosity of the resulting ACs. Micropores typically have a higher contribution to the surface area per unit volume than do mesopores.[34] Therefore, the ACs derived at a low heating rate exhibited high St and Smi values even if they had a low Vt value. Regarding the effect of the holding time at 430 °C, when the heating rate was 5 °C min–1, increasing the holding time from 1 to 2 h increased the micropore proportion at the expense of the mesopore proportion, but the Vt value remained unchanged (Table ).

Table 1

Surface Characteristics of the ACs Derived from KOH Activation of Different Pitch Precursors

					pore size distributiona
sample	SBET	St	Smi	Vt	micro (%)	meso (%)
as-AC	2471	1969	1898	1.05	93	7
AC-1.5-1	2467	1966	1888	1.05	90	10
AC-2-1	2493	1879	1750	1.07	85	15
AC-5-1	2594	1878	1679	1.16	76	24
AC-5-2	2627	2053	1916	1.16	86	14
AC-m	2802	2028	1811	1.24	78	22

The distribution was in the volume ratio.

Based on the aforementioned pore characteristics, Scheme proposes the possible mechanism underlying the formation of the ACs; the purpose of the scheme is to elucidate the effect of the pretreatment procedure on the structure of the pitch precursors and subsequently on the structure of the ACs. When the as-received pitch samples were subjected to single-step heating to 430 °C (process (a) of Scheme ), the pitch molecules rearranged to form a framework consisting of merged mesophase domains partitioned by interstitial boundaries. Polarization microscopy (Figure S2a) revealed the formation of stacked mesophase domains after the single-step heating process. When the pitch samples were impregnated with KOH for activation, the mesophase domains were attacked by KOH on the basal plane, leading to the creation of wedge-like channels in which the near-entrance region corresponded to mesopores.[6,38,50] Some interstitial boundaries were transformed into mesopores after high-temperature activation. The K+ ions intercalated in the lamellar structure of the mesophase domains, resulting in the formation of micropores after the removal of K species through acid washing.[6] As revealed in Table , reducing the heating rate or increasing the holding time at 430 °C for the pitch led the resulting ACs to exhibit increased Smi and substantially decreased mesopore volume possibly because of the elimination of interstitial boundaries. An increase in Smi is beneficial to the process of energy storage in ACs, but the elimination of mesopores hinders ion transport and is detrimental to the power performance of ACs. We observed that compared with the ACs obtained from the pretreated pitch, the as-AC sample exhibited low mesoporosity, which can be attributed to the lack of mesophase domains and interstices for mesopore formation. Additionally, the yield of the as-AC from KOH activation was low because of the lack of mesophase formation that suppressed the evolution of volatiles. The as-AC was not subjected to further electrochemical analysis. The porosity data in Table demonstrate the critical role of mesophase domains in creating mesoporosity during pitch activation into porous carbon.

In the multistep heating process, heating at low temperatures (100 and 200 °C) resulted in long-range molecular ordering and cross-linking prior to molecular decomposition (process (b) in Scheme ). TGA (Figure ) demonstrated that the low-temperature treatment process suppressed the evolution of naphthene and alkyl side chains. When the heating temperature was increased to 430 °C, the pitch was transformed into a long-range mesophase framework sliced by defected splits that were composed of naphthene and alkyl side chains. Polarization microscopy (Figure S2b) demonstrated that multistep heating transformed the pitch into a basin-like coalesced mesophase framework. After activation with KOH at high temperatures, the resulting AC (i.e., AC-m) exhibited a high Smi value, which was attributed to the activation of the mesophase domains, and high mesoporosity, which was attributed to the defected splits in the pitch framework. The high mesoporosity of AC-m may facilitate ion transport for increasing the power rate of the resulting EDLC.[51,52]

In addition to pore structure regulation, multistep heating increased the crystallinity of AC-m, according to the TEM and XRD results. We applied a four-probe method to measure the electronic conductivity of each of the ACs used as electrodes. The electronic conductivity values of the AC-1.5-1, AC-2-1, AC-5-1, AC-5-2, and AC-m electrodes were 18.4, 17.5, 15.4, 18.4, and 23.2 S m–1, respectively. The conductivity decreased with an increase in the preheating rate because of the increase in mesopore content. When the heating rate was 5 °C min–1, increasing the holding time from 1 to 2 h increased the electronic conductivity; this can be attributed to the decreased mesopore content. However, the AC-m sample, which had a high concentration of mesopores, exhibited the highest electronic conductivity. This high conductivity can be attributed to the high crystallinity induced by the multistep heating of the pitch precursor for achieving long-range structural ordering.

Electrochemical Performance of ACs

Figure presents the cyclic voltammograms of symmetric two-electrode cells assembled using the ACs as the electrode and 1 M TEMABF4/PC as the electrolyte. The voltammograms were obtained by scanning the cells at voltage scan rates of 50 and 500 mV s–1. The voltammograms of the cells scanned at 50 mV s–1 were rectangular (Figure a), indicating that the cells exhibited an ideal capacitive feature for charge storage.[4,5,33] The AC-5-2 cell exhibited the highest capacitance value, which can be attributed to its high Smi. Increasing the voltage scan rate to 500 mV s–1 resulted in the distortion of the rectangular shape of the voltammograms (Figure b), indicating that charge transport resistance had a strong influence on the charge storage dynamics of the cells. The increased voltage scan rate had a relatively low effect on the shape of the voltammogram of the AC-m cell; this signifies that the high mesopore content and crystallinity of AC-m facilitated both electron and ion transport in the electrode.

Figure 3

Cyclic voltammograms of symmetric two-electrode cells assembled using the ACs as the electrodes and 1 M TEMABF4/PC as the electrolyte. The cells were scanned at different scan rates: (a) 50 mV s–1 and (b) 500 mV s–1.

Figure a presents the voltage–time curves of the symmetric two-electrode cells charged and discharged at a current density of 0.5 A g–1 and temperature of 25 °C. All galvanostatic charge–discharge curves revealed standard capacitive characteristics over a voltage range of 0–2.7 V.[33]Figure b shows the charge–discharge curves of the AC-m cell obtained at current densities of 0.5–18.75 A g–1. A linear voltage–time relationship was observed at each current density, demonstrating the high-rate capability of the AC-m electrode. The specific discharge capacitance (C) values of the electrodes were calculated using the charge–discharge data, as presented in the following equationwhere I is the discharge current, td is the discharge time, M is the total carbon mass of the two symmetric electrodes, and ΔV is the voltage difference in discharge, excluding IR drop. Figure c illustrates the variation of C with a discharge rate. At the lowest discharge current (0.5 A g–1), the ultimate C values obtained for the AC-1.5-1, AC-2-1, AC-5-1, AC-5-2, and AC-m electrodes were 160, 150, 126, 170, and 160 F g–1, respectively. We determined that the ultimate C values were primarily proportional to Smi, which predominantly contributed to the double-layer capacitance of the carbon electrodes.[4,53,54] Reducing the ramp rate and increasing the holding time at 430 °C for pitch pretreatment resulted in an increase in Smi, thus increasing the ultimate capacitance.

Figure 4

(a) Voltage–time curves of symmetric two-electrode cells assembled using different ACs at a current density of 0.5 A g–1. (b) Voltage–time curves of the AC-m cell at current densities of 0.5, 7.5, 12.5, and 18.75 A g–1. (c) Variation of the specific capacitance with the specific discharge current for the ACs. (d) Variation of IR drop with the cell discharge current for different AC cells.

For each electrode, the derived C value decreased with the discharge current (Figure c) because of ion transport limitations. At the highest discharge current (100 A g–1), the AC-1.5-1, AC-2-1, AC-5-1, AC-5-2, and AC-m electrodes exhibited capacitance retention levels of 61, 61, 65, 57, and 70%, respectively. In principle, pretreating a pitch precursor to increase the mesoporosity of the resulting ACs increases the high-rate capacitance retention of the ACs. We determined that the AC-m electrode exhibited the highest retention, although its mesopore content was lower than that of the AC-5-1 electrode. The networked mesopores in AC-m must have facilitated ion transport for the formation of double layers in the micropores.

Because IR drop was excluded from the capacitance calculation, some of the derived retention values were misleading. We evaluated IR drop in this study. Figure d presents the variation of IR drop with a discharge rate. The IR drop value increased linearly with the discharge rate for each cell; the slope of this linear relationship was determined to correspond to the direct current (DC) equivalent series resistance (RDC). Table presents a summary of the RDC values of the cells. For the cells that were assembled using the ACs obtained from the pitch samples subjected to single-step heating, the RDC values were primarily an increasing function of the mesopore content. This result indicates that the electronic conductivity of the carbon electrodes governed the RDC value of the corresponding cells and that lower electrode mesoporosity was associated with higher electronic conductivity. The AC-1.5-1 electrode had the lowest mesoporosity, and its corresponding cell exhibited the lowest RDC. By contrast, the RDC value of the AC-m cell, which was composed of the AC-m electrode with high mesoporosity, was as low as that of the AC-1.5-1 cell. The high crystallinity and thus high conductivity of the AC-m electrode must have contributed to the low RDC of the corresponding cell.

Table 2

Resistance Components of the Two-Electrode Cells Assembled Using Different ACs

			TMLa
electrode	RDC (Ω)	Rs (Ω)	Rp1 (Ω)	Cp1 (mF)	Rp2 (Ω)	Cp2 (mF)
AC-1.5-1	2.4	2.1	5.3	120	190	65
AC-2-1	3.0	2.1	5.4	110	150	53
AC-5-1	3.4	2.1	5.7	104	120	48
AC-5-2	3.1	2.0	6.2	136	240	67
AC-m	2.5	2.1	4.3	124	100	61

Transmission line model for resistance and capacitance in pores (Rp and Cp, respectively).

Figure a illustrates the Nyquist impedance spectra of the two-electrode cells; the spectra were measured at a voltage bias of 2 V. Applying a voltage bias ensured the penetration of the electrolyte ions into the relatively deep or small pores of the electrodes.[55] This figure presents experimental data (points) as well as simulation results (lines) obtained using an equivalent circuit. As revealed in the high-frequency region of the spectra (inset of Figure a), all cells exhibited a 45° line, corresponding to a transport-limited kinetics, followed by a vertical line featuring the capacitive process.

Figure 5

(a) Nyquist impedance spectra of the two-electrode cells obtained for carbons: AC-m, AC-5-1, AC-2-1, AC-1.5-1, and AC-5-2. (b) EDLC equivalent circuit consisting of the electrolyte resistance in bulk solution (Rs) and electrolyte resistance and capacitance in pores (Rp and Cp, where x = 1 or 2) simulated using a transmission line (TML) model. The impedance data (points) in (a) were obtained at a voltage bias of 2 V in a frequency range of 10 mHz to 100 kHz; the lines are the simulation results obtained using an equivalent circuit shown in (b).

Figure a presents that the vertical lines at low frequencies became increasingly tilted with the decrease in the mesoporosity of the ACs that were obtained from the pitch samples subjected to single-step pretreatment; this finding signifies that the mesopores reduced the charge storage resistance of the cells. The AC-m cell exhibited the lowest charge storage resistance, although the AC-m electrode did not have the highest mesoporosity. Figure b presents the equivalent circuit used for the simulation; in this circuit, the electrolyte resistance in the bulk solution (Rs) and the charge storage resistance and capacitance in the pores were simulated using the transmission line (TML) model.[56−59] Because of the wide pore size distribution, we used two parallel TML elements (TML1 and TML2, with the corresponding pore resistance and capacitance of Rp and Cp, respectively, where x = 1 or 2) to simulate the entire impedance spectra.[60−63] TML1 was used to characterize pores that were easily accessible to electrolytes, whereas TML2 was used to characterize small or deep pores that might require dissociation of counterions for penetration.

Table presents the simulated resistance and capacitance components of the cells. All cells exhibited similar Rs values because the same electrolyte was used in the cell assembly. Cp1 primarily contributed to the capacitance of the cells because the pores characterized by TML1 exhibited considerably lower resistance levels compared with those characterized by TML2 (i.e., Rp1 ≪ Rp2). For the ACs obtained from the pitch that was subjected to the single-step pretreatment processes involving a constant holding time of 1 h, the Cp1 values of the resulting cells increased as the pretreatment ramp rate decreased; this can be attributed to the increase in Smi. However, the Rp1 values decreased as the ramp rate decreased; this decrease in the ramp rate caused a decrease in the mesoporosity of the ACs. This result indicates that the electronic conductivity of the ACs was the governing factor for the value of Rp1. Furthermore, the Rp2 and Cp2 values increased as the ramp rate decreased; this finding can be attributed to the increase in microporosity. The preceding results demonstrate that the electronic conductivity of the ACs governed charge transport in TML1 and that ion transport resistance governed the dynamics in TML2.

For the ACs obtained from the pitch subjected to single-step pretreatment at a ramp rate of 5 °C min–1, both the Cp and Rp values of the corresponding cells increased when the holding time at 430 °C was increased from 1 to 2 h; the increase can be attributed to the substantial increase in the microporosity of the electrodes. The Rp1 value of the AC-5-2 cell was higher than that of the AC-5-1 cell; an explanation for this finding is that increasing the holding time engendered a minor increase in electronic conductivity, as revealed by the IR drop measurements (Figure d and Table ). We observed that the AC-m cell had the lowest Rp1 and Rp2 values compared with the other cells; this observation can be attributed to the high electronic conductivity of the AC-m carbon framework and the hierarchical pore structure that facilitated ion transport in small pores. Figure S7 presents the impedance angle values of the cells at varying frequencies based on the impedance data of Figure a. In the low-frequency regime, the AC-m cell exhibits phase angles closer to −90° (a pure capacitive behavior)[5] relative to the angles of the other cells, justifying the facile ion transport in the AC-m for charge storage.

We used the galvanostatic discharge data to correlate the specific power and energy of the two-electrode cells according to eqs and 3where E is the specific energy, and P is the specific power. Figure a displays Ragone plots providing a summary of the E and P data for the cells. The AC-m cell was superior to the other cells with respect to both E and P; the cell achieved a maximum E value of 40 Wh kg–1 (based on the total carbon mass) or 20 Wh L–1 (with a carbon tap density of 0.5 g cm–3) within a voltage range of 0–2.7 V and sustained an E value of 24 Wh kg–1 (12 Wh L–1) at a high P value of 50 kW kg–1 (25 kW L–1). The excellent performance of this cell can be attributed to the multistep pretreatment of the petroleum pitch precursor to produce a porous carbon that exhibited high electronic conductivity and comprised hierarchically structured pores that facilitated ion transport. To confirm the stability of AC-m, the AC-m cell was charged and discharged for 20,000 cycles at voltages ranging between 1.35 and 2.7 V and a current density of 5 A g–1. As shown in Figure b, the cell exhibited a high capacitance retention rate of 91% after the charge and discharge processes, and the Coulombic efficiency values were nearly 100% throughout the whole cycling; this result demonstrates the high chemical durability of AC-m as an EDLC electrode. We developed a strategy for activating pitch with high concentrations of naphthene and alkyl side chains to produce a high carbon yield and a carbon framework suitable for ion and electron transport.

Figure 6

(a) Ragone plots of the two-electrode cells assembled using different ACs. The data were determined based on the galvanostatic discharge within a voltage range of 0–2.7 V. (b) Variation of the specific capacitance and Coulombic efficiency with the cycle number for the AC-m cell. The cell was galvanostatically charged and discharged at 5 A g–1 within 1.35–2.7 V.

Summary and Conclusions

This study demonstrated that thermal pretreatment of petroleum pitch for forming mesophase pitch is essential for the production of ACs through KOH activation. In the pitch pretreatment processes, the temperature ramp rate and holding time (at 430 °C) significantly affected the structure of the resulting ACs. Reducing the ramp rate and increasing the holding time increased the microporosity Smi and ultimate double-layer capacitance of the ACs but disrupted ion transport into the interior of the ACs. Nevertheless, reducing the ramp rate afforded a longer time for molecular arrangement in the pitch samples, thus increasing the crystallinity and electronic conductivity of the resulting ACs. By applying a multistep pretreatment process, we obviated the necessity of a trade-off between the electronic conductivity and ionic mobility of ACs caused by varying the ramp rate. This multistep pretreatment process incorporated aliphatic species into the aromatic mainframe of the pitch samples, forming a pitch framework consisting of mesophase domains intersected by aliphatic borders. We subjected a pitch sample to this multistep pretreatment process, yielding a carbon electrode (i.e., AC-m) that exhibited high crystallinity for electron conduction, high Smi for double-layer charge storage, and a hierarchical pore structure for ion transport. The AC-m electrode was used to assemble a symmetric EDLC, which registered an unprecedented high energy of 40 Wh kg–1 (20 Wh L–1) within a voltage range of 0–2.7 V and sustained an E value of 24 Wh kg–1 (12 Wh L–1) at a high P value of 50 kW kg–1 (25 kW L–1). In summary, we developed a unique and advantageous strategy for pretreating petroleum pitch to produce high-performance EDLC electrodes. The defining characteristic of this pretreatment strategy is its incorporation of aliphatic species into the framework of mesophase pitch, which can increase the carbon yield of petroleum pitch and produce carbons with properties advantageous for EDLCs. The proposed pretreatment strategy may provide the foundation for extensive applications of petroleum pitch in energy storage devices.

Experimental Section

Sample Preparation

Petroleum pitch powder provided by Chinese Petroleum Co., Taiwan, served as the precursor for producing porous carbons that could be used in EDLCs; the pitch had 3–6% primary quinoline insolubles and a softening point of 185 °C. Prior to KOH activation, we thermally pretreated the as-received pitch at 430 °C to cross-link the constituent molecules. Before the pretreatment procedure was performed, the pitch powder was ground and sieved to an average grain size of approximately 0.3 mm. In the pretreatment procedure, the as-received pitch powder was placed in a horizontal cylindrical furnace under N2 flow, heated from room temperature to 430 °C, maintained at 430 °C for 1 or 2 h, and finally cooled to room temperature. To heat the powder from room temperature to 430 °C, we used a single-step ramp at a heating rate of 1.5, 2, or 5 °C min–1 or a multistep ramp that involved linearly increasing the temperature at 2 °C min–1 to 100 and 200 °C and then holding each of these temperatures for 0.5 h.

The pitch samples were impregnated with KOH at a KOH/pitch ratio of 4/1 and then thermally treated to produce ACs. Our preliminary studies suggested that the KOH/pitch ratio (4/1) resulted in optimal carbon yield and porosity from activation. Specifically, the impregnated samples were thermally treated at 800 °C for 1 h under N2 flow and then cooled to room temperature. The products were washed using 1 L of 0.5 M HCl solution at 85 °C for 30 min and then leached using 1 L of distilled water at 85 °C several times until the pH value of the water–carbon mixture was higher than 6. The leached products were then dried in a vacuum at 110 °C for 24 h to yield KOH-activated carbons. All samples were heated in an Ar atmosphere at 700 °C to remove surface oxides.[64]

Carbon Characterization

The microstructures of the carbon specimens were examined using a high-resolution TEM instrument (JEOL 2100F, Japan). The crystal structure was analyzed using an XRD system (Rigaku, Ultima IV, Japan) with Cu Kα radiation excited at 40 kV and 40 mA. A TGA instrument (PerkinElmer, TGA7, USA) was used in a N2 atmosphere to estimate the decomposition behavior. In addition, a polarizing microscope (Nikon, Eclipse LV100POL, Japan) was used to characterize the formation of the mesophase in the pitch specimens. The pore structure of the carbons was determined through gas adsorption measurements; the measurements were performed using an automated adsorption apparatus (Micromeritics, ASAP 2020 HD88, USA) that was operated at 77 K in a N2 atmosphere at relative pressures (P/P0) that ranged from 10–7 to 1.

EDLC Assembly

The EDLCs used for examination in this study exhibited a symmetrical two-electrode capacitor configuration. To prepare the carbon electrodes, 2 mg of carbon powder was dispersed in ethanol and then coated on a carbon-coated Al foil under stress without using a binder. Symmetrical cells consisted of two opposing carbon electrodes (with an active area of 1 cm2 and a thickness of approximately 40 μm) that were separated by an electrolyte-loaded cellulose filter paper. The EDLCs were assembled under stress to ensure close contact at the carbon–carbon and carbon–Al foil interfaces.

Electrochemical Measurements

All electrochemical measurements were performed at room temperature (approximately 25 °C). Cyclic voltammetry and galvanostatic charge–discharge measurements were performed under the two-electrode configuration by using an electrochemical analyzer (Solartron Analytical, 1470E, UK). An ac impedance spectrum analyzer (Zahner-Elektrik, IM6e, Germany) was used to analyze the impedance characteristics of the EDLCs. The measurements were conducted at 0 V by applying an ac potential amplitude of 5 mV and a frequency ranging from 10 mHz to 100 kHz.