Molecular Origin of Electric Double-Layer Capacitance at Multilayer

Dec 14, 2016 - Multilayer graphenes have been widely used as active materials for electric double-layer capacitors (EDLCs), where their numerous edges...
1 downloads 9 Views 3MB Size
Letter pubs.acs.org/JPCL

Molecular Origin of Electric Double-Layer Capacitance at Multilayer Graphene Edges Huachao Yang,†,‡ Xiaoliang Zhang,‡ Jinyuan Yang,† Zheng Bo,*,† Ming Hu,*,‡,§ Jianhua Yan,† and Kefa Cen† †

State Key Laboratory of Clean Energy Utilization, Institute for Thermal Power Engineering, College of Energy Engineering, Zhejiang University, Hangzhou, Zhejiang Province 310027, China ‡ Institute of Mineral Engineering, Division of Materials Science and Engineering, Faculty of Georesources and Materials Engineering, Rheinisch-Westfaelische Technische Hochschule (RWTH) Aachen University, 52064 Aachen, Germany § Aachen Institute for Advanced Study in Computational Engineering Science (AICES), RWTH Aachen University, 52062 Aachen, Germany S Supporting Information *

ABSTRACT: Multilayer graphenes have been widely used as active materials for electric double-layer capacitors (EDLCs), where their numerous edges are demonstrated to play a crucial role in charge storage. In this work, the interfacial structure and capacitive behaviors of multilayer graphene edges with representative interlayer spacing are studied via molecular dynamics (MD) simulations. Compared with planar graphite surfaces, edges can achieve a 2-fold increase in the specific capacitance at a wider interlayer spacing of ∼5.0 Å. Unusually, the molecular origins for achieved charge storage are predominantly attributed to the structural evolutions of solvents occurring in the double layer, going beyond the traditional views of regulating the capacitance by ion adsorption/separation. Specifically, diverse ionic distributions exhibit similar screening ability and EDLC thickness, while water molecules can counterbalance the interfacial electric fields more effectively at edge site. The as-obtained findings will be instructive in designing graphenebased EDLCs for advanced capacitive performances. ranges from ∼3.4,8 to ∼3.6,9 ∼4.3,10 and ∼4.9 Å.11 Meanwhile, vertically oriented graphenes, a class of networks of “graphitic” platelets that are arranged perpendicularly to the substrate surface, commonly consist of few-layer graphene with a layer number of 1−10 and an interlayer (002) spacing of ∼3.4 to ∼3.9 Å.12,13 Subnanometer channels are usually inaccessible to the typical electrolytes applied in EDLCs.14−16 Due to strong excludedvolume effects, there is an energy penalty to enter subnanometer channels as ions have to shed part of their hydration shells. For example, the free energy penalty for a Na+ ion to enter the carbon nanotubes with a diameter of ∼8.2 Å could be ∼120 kJ mol−1 (or equivalently ∼2.48 V).15,16 As a consequence, edges have been considered to play a crucial role in the enhancement of charge storage capability in graphenebased EDLCs.17−23 Randin et al. reported that the exposed edges of an annealed pyrolytic graphite presented a much higher specific capacitance (∼50−70 μF cm−2) than that on the basal region (∼3 μF cm−2).17 Recently, Shi et al. observed that the area-specific capacitance of monolayer graphene edges was about 4 orders of magnitude greater than that on the basal

E

ver-increasing global energy demands have boosted the continuous development of electric double-layer capacitors (EDLCs), an advanced electrochemical device providing high power density, an ultrafast charging/discharging rate, and superior lifespan.1,2 The active materials of EDLCs have stimulated tremendous research efforts, experiencing development from conventional porous structures (e.g., activated carbons) to low-dimensional (zero-, one-, and two-dimensional) nanomaterials.3 Typically, graphene, one-atom-thick carbon atoms densely packed into a honeycomb lattice, has garnered substantial attention as a promising active material for EDLCs.4−6 The widespread interest initially stems from its extraordinary monolayer properties, such as a huge surface area of ∼2630 m2 g−1 and a superior electrical conductivity up to ∼106 S cm−1. In fact, stacks of graphene layers, instead of monolayer graphene, are commonly preferred for practical energy storage applications.4,7 Layered graphenes usually have a high tendency to reagglomerate themselves into graphene stacks during all phases of preparation and subsequent electrode production procedures. Consequently, numerous subnanometer channels, composed of parallel graphene layers with the representative interlayer spacing of less than 10 Å, widely exist in stacked graphene networks. For conventional reduced graphene oxide powders or films fabricated through wet-chemistry routes (e.g., modified Hummer’s method), the typical interlayer distance © XXXX American Chemical Society

Received: November 14, 2016 Accepted: December 14, 2016 Published: December 14, 2016 153

DOI: 10.1021/acs.jpclett.6b02659 J. Phys. Chem. Lett. 2017, 8, 153−160

Letter

The Journal of Physical Chemistry Letters plane.18 Seo et al. fabricated vertical graphene nanosheets from natural precursors, and it has been demonstrated that the reactivity and density of the sharp edges are critical factors in the energy storage performance.19,20 With an atomistic simulation, Pak et al. demonstrated that the edge of monolayer graphene could significantly improve the mass-specific capacitance by noticeably enhancing the ionic separations.21 Besides, our previous ab initio simulations pointed out that graphene edges tended to promote the kinetics of electrolyte movement, resulting in a higher ionic density at edge sites.22 Prior theoretical studies have greatly advanced our basic understandings of edge effects, especially on monolayer graphene. However, the corresponding issues of primary mechanisms/phenomena at multilayer graphene edges are still far from well understood. In this work, the interfacial structure and capacitive behaviors of multilayer graphene edges were investigated with molecular dynamics (MD) simulations. The electrochemical performances of planar (for comparison) and edge electrodes with typical interlayer spacings (d = 3.4, 4.5, and 5.0 Å) were examined. A comprehensive evaluation of the EDLC structures (e.g., accumulating density, packing location, and screening efficiency) at multilayer graphene edges was further carried out to unveil the underlying charge storage mechanisms. Especially, the contributions of ions and solvents to the achieved capacitance were quantified individually by decomposing the total electric potential distributions, where a novel mechanism beyond traditional views was proposed. Figure 1 shows the schematic diagrams for planar and edge electrodes. Typically, the planar electrodes are made up of two

results on a symmetry electrode surface (i.e., translation invariance along the in-plane direction) should not depend strongly on the algorithm used (i.e., constant charge density vs constant potential method).24−26 Moreover, prior density functional theory investigations (including ours) revealed that most charges tended to accumulate at the conductive graphene edges due to the quasi-localized pz states of edge sites.21,22 Hence, partial charges of ±15 μC cm−2 were placed on the electrode surface directly contacting the electrolytes to mimic the applied potentials, which was also used by other simulation studies.14,26,27 The applied charge density fell in the typical range of −20−30 μC cm−2 for EDLCs employing aqueous electrolytes,28 which was widely adopted by previous simulation studies.14,29 Additional details of simulation protocol and polarization effects, as well as convergence tests on our obtained key properties, are available in Figures S1 and S2, Supporting Information. Figure 2 represents the capacitance values of the cathode (Ccathode), anode (Canode), and whole capacitor cell (Ctotal) for

Figure 2. Specific capacitances for planar and edge electrodes of d = 3.4, 4.5, and 5.0 Å, respectively. Green color: negative electrode; red color: positive electrode; blue color: total capacitor cell.

both the planar and edge cases. Detailed methodologies for calculating the integral capacitance could be found in the Supporting Information. The as-calculated capacitances on the negative and positive planar electrodes were ∼5.96 and ∼4.40 μF cm−2, respectively, close to previously reported theoretical and experimental results.28,30 In an experimental study, a typical capacitance value of ∼10 μF cm−2 was obtained in the NaCl solution with carbon electrodes.28 With an atomistic simulation, Xu et al. suggested that the specific capacitances were ∼6.32 and ∼6.04 μF cm−2 near the negative and positive graphene electrodes.30 Especially, the capacitance in negatively charged planar electrodes was much larger than that in the positive counterpart, which was also observed in the edge case. This asymmetry was mainly due to the different electrolyte structures, in turn leading to a diverse screening ability.31 More importantly, the edge electrodes achieved a predominant advantage in charge storage over the planar case, which became more pronounced when enlarging the interlayer distance. For instance, on the positively electrified edge electrodes with d = 3.4 Å, the area-specific capacitance was about ∼25.7% higher than that of planar ones, which was drastically improved to ∼103.4% with a slight increase of interlayer spacing to ∼5.0 Å. Furthermore, our as-obtained

Figure 1. Representative snapshots of MD systems for (a) planar and (b) edge electrodes. Gray spheres: carbon atoms; purple spheres: Na+ ions; green spheres: Cl− ions; blue background: water molecules.

opposite graphene layers arranged in a parallel plate configuration (Figure 1a), while the edge ones consist of multilayer graphenes with a prismatic face exposed toward the electrolyte (2 M NaCl). Zigzag-type graphene edges stacked in an AA manner were employed as the EDLC electrodes (Figure 1b), where the interlayer spacings were set as ∼3.4, ∼ 4.5, and ∼5.0 Å. Recent studies have demonstrated that the simulation 154

DOI: 10.1021/acs.jpclett.6b02659 J. Phys. Chem. Lett. 2017, 8, 153−160

Letter

The Journal of Physical Chemistry Letters

Figure 3. Number density profiles for planar and edge electrodes at surface charge densities of (a) 0 and (b) −15 μC cm−2. Black square: planar electrodes; red circles: d = 3.4 Å; blue up triangle: d = 4.5 Å; green down triangle: d = 5.0 Å.

observations of a ∼24.8−103.4% increase in specific capacitance coincided well with recent theoretical studies on electrode surface topography. Vatamanu et al. reported that the specific capacitance of atomically corrugated electrodes (comprised of hexagonal graphite layers stacked in ABAB sequence) was around ∼25% larger than that on the basal plane using the same ionic liquid.32 Moreover, Pak et al. showed that a certain type of monolayer graphene edge immersed in an ionic liquid could almost double the mass-specific capacitance.21 In these studies, the superior charge storage capability was principally attributed to the remarkably improved ionic separations at low potentials. However, the solvent effects were not considered, which was demonstrated to have a significant influence on the charging mechanisms.31,33,34

The capacitive performances of EDLCs are closely correlated with the interfacial EDLC microstructures, for example, adsorbed density, layer location, as well as the corresponding screening efficiency and electric field distributions.29,35,36 In this regard, an in-depth investigation into these microscopic aspects was further carried out to reveal the molecular origins for achieved capacitance on edge electrodes. Figure 3a shows the electrolyte number density for neutral planar and edge electrodes. Due to the presence of solid electrodes, an alternating layer structure was observed at the interface. In particular, obvious discrepancies in density distribution between the planar and edge electrodes were observed. To be specific, water density near the planar electrode was much higher than that of the edge side, while 155

DOI: 10.1021/acs.jpclett.6b02659 J. Phys. Chem. Lett. 2017, 8, 153−160

Letter

The Journal of Physical Chemistry Letters

Figure 4. Charge screening coefficients of (a) ηion and (b) ηsolvent for planar and edge electrodes, respectively, at a surface charge density of −15 μC cm−2.

Similar results were also observed on atomically corrugated graphene electrodes immersed in ionic liquids.32,37 Although no electrolyte could intercalate inside of the edge electrode, its accumulating position was dragged obviously toward the electrified surface. In this regard, the thickness of the Helmholtz layer descended with widening interlayer spacing, from ∼4.6 Å in the planar case to ∼4.0 Å near the edge electrodes of d = 5.0 Å. Meanwhile, water molecules were drawn more markedly to the charged surface, especially for H atoms (by ∼1.6 Å for d = 5.0 Å; see Figure S4 in the Supporting Information). Such a finding was mainly due to the fact that the atomic interlayer distance (i.e., d from ∼3.4 to ∼5.0 Å) in the edge case was much larger than that of the planar electrode (∼2.46 Å). As a result, electrolytes could easily embed into the entrance of edge electrodes, accessing the electrified surface much closer. Meanwhile, this result also highlighted that the enhanced capacitance was attributed to the multilayer graphene edges, rather than ion penetrations, due to the strong entrance barriers. As the distance from electrode surface z further increased, the structure of the electrolytes gradually became bulk-like, in which the region was referred to as the diffusion area. Different with the Helmholtz layer, an unapparent but analogous electrolyte structure was presented for the planar and edge electrodes. Unusually, counterions manifested a higher intensity in the diffusion district of the edge site, which was ascribed to the broader ionic distributions induced by atomic roughness. A closer approach of counterions to the edge surface yielded superior capacitive performances, whereas the considerably lower densities were identified to have a detrimental effect.38,39 As a consequence, the contributions of ions to the achieved capacitance were, however, not clear and controversial. To solve such a puzzle, the effective EDLC thickness dcenter was further examined,39 which is calculated as

ions presented a similar distribution. This phenomenon was primarily due to the favorable van der Waals (vdW) interactions in the planar case. Considering the geometrical structure of electrodes, the surface atomic density of planar electrodes (∼0.38 #/Å2) was much larger than that of edge ones (∼0.12, ∼0.09, and ∼0.08 #/Å2 for d = 3.4, 4.5, and 5.0 Å, respectively), leading to stronger interfacial vdW interactions. For example, the vdW potentials between the water molecule and planar electrode (∼25.38 MeV) were almost two times greater than those of the edge case (∼12.71 MeV for d = 5.0 Å), which tended to accumulate more water molecules at the interface. However, such a vdW interaction was not sufficient to capture more ions with full hydration shells near the electrode surface. Figure 3b depicts the density profiles for negatively charged planar and edge electrodes of −15 μC cm−2. For convenience, the interfacial EDLC structure was decomposed into two portions according to the traditional Gouy−Chapman−Stern theory, that is, Helmholtz and diffusion layers. As the electrodes were electrified, a sharp density peak developed in the Helmholtz district, that is, the region between the first layer of counterions (Na+ ions) and the electrode surface. Especially, the edge electrode experienced a much lower and broader density distribution in comparison with the plane case, which became more prominent when enlarging the interlayer distance. For instance, the first Na+ density near planar electrodes was about ∼1.43 times higher than that at the edge site of d = 3.4 Å, which was further increased to ∼2.30 times for a wider interlayer distance of ∼5.0 Å. A similar but more noticeable trend could be observed in the water case, in which the O density near the planar electrode was about 4.67-fold greater than the edge counterpart. Besides, the water distribution of the edge side progressively developed into multilayer structures, whereas a mere single peak was recognized for the planar electrode. However, the as-obtained results of ionic packing properties in this work were in stark contradiction with earlier studies on a monolayer graphene edge, where a much higher ionic density was expected at edge sites.21,22 This contrast of behavior was principally attributed to the atomic rough nature of multilayer graphene edges. The roughness strongly perturbed the electrolyte layer structure at the interface, hence obviously lowering the maximum local density and broadening the peaks in distribution. The influences of surface roughness on the ion density are available in Figure S3, Supporting Information.

z

dcenter =

∫z 1 z N (z − z 0)ρ(z) dz 0

z

∫z 1 z N ρ(z) dz 0

(1)

where z0 and z1 represent the location of the electrode surface and area of the EDLC dominant district (∼12 Å), ρ(z) is the space charge density of ions, and N is set to ∼1 for the planar and edge electrodes. Almost invariant dcenter values were observed, which were ∼6.55, ∼ 6.50, ∼6.48, and ∼6.51 Å for the planar and edge electrodes of d = 3.4, 4.5, and 5.0 Å, respectively. This unusual result was probably due to the fact 156

DOI: 10.1021/acs.jpclett.6b02659 J. Phys. Chem. Lett. 2017, 8, 153−160

Letter

The Journal of Physical Chemistry Letters

the dependence of ηwater on the distance z for planar and edge electrodes. Owing to the extended nature of the water charge distribution, a pronounced oscillatory pattern was observed in the Helmholtz region. Particularly, ηwater progressively converged to ∼0 within the diffusion area (i.e., ∼1 for ions), which was closely related to the electroneutral nature of water molecules. Besides, compared with the planar case, the amplitude of spatial fluctuations on edge electrodes declined conspicuously with increasing interlayer distance. This finding implied that the closer but multilayer structure of water molecules probably facilitated the balance of electrode charges, which was further verified by the dramatic descent of electric fields at the interface (Figure S6). With a closer examination of the inset of Figure 4b, the positions of achieving complete screening decreased with the enlarging interlayer spacing, qualitatively highlighting the better screening ability of water molecules in the edge case. As for the positively charged electrodes, a similar trend of rapid decline in the interfacial electric fields induced by water molecules was recognized at edge sites (Figures S7 and S8). The electric potential distributions across the EDLCdominant area were further explored. According to the Poisson equation, the potential of zero charge (PZC) for neutral planar and edge electrodes was computed (details of the electric potential calculation are given in the Supporting Information). The as-obtained PZC for the planar electrode was ∼0.39 V, consistent with previous reports on graphene-based EDLCs,30,45 while the edge counterpart had a slightly high value, which was further increased when widening the interlayer spacing (i.e., ∼0.43, ∼0.51, and ∼0.53 V for d = 3.4, 4.5, and 5.0 Å, respectively). This result was closely related to the reduced O density at edge sites, as mentioned above. Figure 5 shows the electric potentials as a function of z from the charged electrode surface. The total electric potential Utotal,

that a closer location of ions facilitated packing tightly at the interface, while the much sparser distributions seriously deteriorated the formation of the EDLC structure, ultimately inducing virtually identical dcenter values. According to a traditional parallel-plate capacitor, the capacitance varies as the inverse of the Debye length (i.e., effective EDLC thickness),40,41 suggesting that comparable capacitive performances were expected for the planar and edge electrodes. This finding seemed to be rather counterintuitive given that the capacitance behaved in the opposite way, indicating that the structural variations of ions probably failed to account for the enhanced charge storage. On the positively charged electrodes, a similar trend of electrolyte distribution was recognized, in which the density of edge sites was much lower but closer than the planar counterpart (Figure S5). Especially, the counterions (Cl−) were shifted more noticeably toward the electrified surface, owing to the relatively weak hydration shells.14,29 To ascertain the molecular origin for achieved capacitance unambiguously, the responses of ions and solvents to the interfacial electric fields were further evaluated individually. A charge screening coefficient η(z) was thus introduced and calculated as35,42 z

η(z ) = −

∫z ρ(z) dz 0

σ

(2)

where σ denotes the surface charge density of electrodes. Conceptually, η(z) = 1.0 indicates that the electrode charge became exactly balanced, while η(z) > 1.0 meant that charge overscreening occurred. As shown in Figure 4a, the screening coefficient ηion as a function of z relative to the electrode surface was displayed. Overall, the screening coefficient ηion increased prominently in the initial Helmholtz region and subsequently converged to ∼1 within the diffusion area. Unlike traditional organic electrolytes or ionic liquids,43,44 only a weak overscreening effect existed, arising from the relatively small ions in aqueous solutions. Despite the obviously different ionic distributions, ηion of planar and edge electrodes manifested an analogous fluctuation as the distance z increased and achieved complete screening at the almost same position of ∼12.0 Å. Such an unexpected finding could be interpreted by the balanced ionic distributions between Helmholtz and diffusion layers. The closer packing position of ions favored a higher screening coefficient, while the corresponding lower density severely suppressed the screening efficiency at the interface, where ηion was directly proportional to the adsorbed density. Consequently, decreased ηion values were observed in the Helmholtz area when enlarging the interlayer distance, which were ∼0.61, ∼0.56, ∼0.48, and ∼0.46 for the planar and edge electrodes of d = 3.4, 4.5, and 5.0 Å, respectively. On the other hand, the relatively broad counterion distributions at the edge site brought about a higher density in the diffusion district, which could compensate the lower ηion within the Helmholtz layer to some extent. Meanwhile, the ascalculated ηion was also in good accordance with the analogous electric field distributions induced by ions in the EDLCdominant area (Figure S6). Thus, our simulation evaluations pointed out that ions of planar and edge electrodes tended to exhibit a similar screening ability, regardless of their diverse distributions. In contrast, the screening coefficient of water molecules appeared to have more complex behavior. Figure 4b presents

Figure 5. Electric potential distributions as a function of distance z for planar and edge electrodes at a surface charge density of −15 μC cm−2. Black color: Utotal; red color: Uion; blue color: Usolvent; solid line: planar electrodes; dashed line: d = 3.4 Å; short dashed line: d = 4.5 Å nm; dotted line: d = 5.0 Å.

determined by both the ionic and water distributions, displayed remarkable oscillatory behavior in the Helmholtz district, which progressively grew smooth with increasing z. Particularly, the spatial fluctuation of Utotal, that is, the position of peaks, majorly depended on the water structures (Figure S9), arising from the larger amount of water molecules at the interface. Furthermore, in comparison with the planar case, the magnitude of Utotal on 157

DOI: 10.1021/acs.jpclett.6b02659 J. Phys. Chem. Lett. 2017, 8, 153−160

Letter

The Journal of Physical Chemistry Letters

current MD calculations (e.g., nonpolarizable electrode surface, no defects or functional groups, and rigid graphene sheets), the obtained results coincided well with the capacitance enhancement from experimental work. For instance, Hassan et al. fabricated edge-enriched graphene quantum dots with chemical activations, achieving a factor of 2 increase in the mass-specific capacitance.46 Also, our recent experimental research reported that preparing the vertically oriented graphenes with higher edge densities by employing different plasma sources could demonstrate a 2.3-fold increase in both the area- and massspecific capacitance.22 Moreover, Zhang et al. split the vertically aligned carbon nanotubes into graphene nanoribbons enriched with numerous edges, exhibiting a 4-fold increase in specific capacitance.23 In summary, MD simulations were employed to characterize the interfacial structure and capacitive performances of the EDLCs based on multilayer graphene edges. Results showed that the edge electrodes could remarkably enhance the specific capacitance in comparison with the planar case, which became more pronounced as the interlayer spacing increased (up to ∼103.4% for d = 5.0 Å). Different than the monolayer graphene edge, the dominant origins for achieved charge storage were strongly correlated with the solvent packing properties. Specifically, ions presented similar EDLC thickness and screening ability despite their diverse interfacial structures on the planar and edge electrodes, ultimately resulting in comparable electric potential distributions. Owing to the higher screening efficiency on edge electrodes induced by water molecules, the total electric potentials could be significantly reduced, corroborating the obvious capacitance enhancement at edge sites. The as-obtained novel perspective into multilayer graphene edges can provide instructive strategies for preparing active materials with rich edges to further advance the optimization of EDLCs and meanwhile highlight the crucial role of solvents in determining capacitive performances.

the edge electrode was decreased obviously with enlarging interlayer distance (from ∼2.13 to ∼0.85 V), corroborating the significant capacitance enhancement at edge sites. To quantitatively differentiate the contributions of ions and water molecules to the achieved capacitance, the total electric potential (Utotal) across the EDLC structure was further decomposed into two parts, that is, Uion (Na+/Cl− ions) and Usolvent (water molecules). As shown in Figure 5, Uion presented a positive potential value across the EDLC-dominant area, in which an almost linear relation was observed in the initial ion depletion region of the Helmholtz district. Especially, Uion of planar and edge electrodes exhibited a nearly identical distribution, demonstrating the unimportant contributions of ions to the enhanced charge storage. Such a result was intimately correlated with the as-obtained similar EDLC thickness and screening efficiency within the EDLC-dominant area. On the contrary, Usolvent exhibited a negative oscillatory distribution, where it behaved as an electric potential plateau in the water depletion region (labeled by a black dashed circle) and decreased sharply with the presence of water molecules. More especially, compared with the planar case, Usolvent of edge electrodes descended in a stepwise manner as the interlayer distance increased (from ∼−5.3 V to ∼−6.7 V), which could be interpreted as follows. The much lower water density at edge sites was adverse to the balance of electrode charges, leading to a gentle decrease of Usolvent in comparison with the rapid ones in the planar case. Nevertheless, the considerably closer packing position and multilayer structures could obviously diminish the potential plateaus, which shielded the charged surface more efficiently, ultimately yielding much lower Usolvent on edge electrodes. Hence, our observations highlighted that the structural variations of solvents were primarily responsible for the enhanced capacitance at multilayer graphene edges, which was strikingly different than the monolayer counterpart. Such an obvious discrepancy in charging mechanisms was probably correlated to the atomic rough nature of multilayer graphene edges. Similar electric potential distributions were also found on the positively charged electrodes, where a decreased Usolvent was recognized when enlarging the interlayer distance (Figure S10). The corresponding issues of primary mechanisms/ phenomena for a wider case will be explored in our further work. Besides, it is worthwhile to point out that solvent effects in determining the capacitive behaviors have been recognized by earlier simulation studies, where solvent packing properties could partly account for the origin of the achieved charge storage. Ho et al. demonstrated that the obtained 2-fold increase in area-specific capacitance of patterned electrodes was strongly correlated to both the ionic and water distributions.26 Jiang et al. reported that the specific capacitance exhibited a volcano-shaped trend as the dipole moment of solvents varied from 2.5 to 5.0 D.34 A very recent study further revealed that the variations of the water dielectric constant could cancel the influences of ion types, leading to a nearly constant capacitance value at diverse monovalent aqueous electrolytes.31 However, the predominant role of solvents in determining the achieved capacitance has never been reported before, indicating the novel charging mechanisms at multilayer graphene edges, which were probably related to the strong solvent polarity of water molecules. Finally, we make a brief connection with experimental studies. Despite the approximations and simplifications used in



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpclett.6b02659. Details of simulation protocol and capacitance calculations as well as additional data on the EDLC microscopic properties within the positively charged electrodes (PDF)



AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected] (Z.B.). *E-mail: [email protected] (M.H.). ORCID

Zheng Bo: 0000-0001-9308-7624 Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS Financial support for this work was provided by the National Natural Science Foundation of China (No. 51306159), the Zhejiang Provincial Natural Science Foundation of China (No. LR17E060002), and the China Scholarship Council (CSC). Simulations were performed with computational resources granted by the Shanghai Supercomputer Center and the Jülich 158

DOI: 10.1021/acs.jpclett.6b02659 J. Phys. Chem. Lett. 2017, 8, 153−160

Letter

The Journal of Physical Chemistry Letters

(21) Pak, A. J.; Paek, E.; Hwang, G. S. Impact of Graphene Edges on Enhancing the Performance of Electrochemical Double Layer Capacitors. J. Phys. Chem. C 2014, 118, 21770−21777. (22) Yang, H.; Yang, J.; Bo, Z.; Zhang, S.; Yan, J.; Cen, K. Edge Effects in Vertically-Oriented Graphene Based Electric Double-Layer Capacitors. J. Power Sources 2016, 324, 309−316. (23) Zhang, C.; Peng, Z.; Lin, J.; Zhu, Y.; Ruan, G.; Hwang, C.-C.; Lu, W.; Hauge, R. H.; Tour, J. M. Splitting of a Vertical Multiwalled Carbon Nanotube Carpet to a Graphene Nanoribbon Carpet and Its Use in Supercapacitors. ACS Nano 2013, 7, 5151−5159. (24) Merlet, C.; Péan, C.; Rotenberg, B.; Madden, P. A.; Simon, P.; Salanne, M. Simulating Supercapacitors: Can We Model Electrodes as Constant Charge Surfaces? J. Phys. Chem. Lett. 2013, 4, 264−268. (25) Wang, Z.; Yang, Y.; Olmsted, D. L.; Asta, M.; Laird, B. B. Evaluation of the Constant Potential Method in Simulating Electric Double-Layer Capacitors. J. Chem. Phys. 2014, 141, 184102. (26) Ho, T. A.; Striolo, A. Capacitance Enhancement Via Electrode Patterning. J. Chem. Phys. 2013, 139, 204708. (27) Feng, G.; Li, S.; Zhao, W.; Cummings, P. T. Microstructure of Room Temperature Ionic Liquids at Stepped Graphite Electrodes. AIChE J. 2015, 61, 3022−3028. (28) Conway, B. E. Electrochemical Supercapacitors: Scientific Fundamentals and Technological Applications; Kluwer Academic/ Plenum: New York, 1999. (29) Bo, Z.; Yang, H.; Zhang, S.; Yang, J.; Yan, J.; Cen, K. Molecular Insights into Aqueous Nacl Electrolytes Confined within VerticallyOriented Graphenes. Sci. Rep. 2015, 5, 14652. (30) Xu, K.; Ji, X.; Chen, C.; Wan, H.; Miao, L.; Jiang, J. Electrochemical Double Layer near Polar Reduced Graphene Oxide Electrode: Insights from Molecular Dynamic Study. Electrochim. Acta 2015, 166, 142−149. (31) Jiang, G.; Cheng, C.; Li, D.; Liu, J. Z. Molecular Dynamics Simulations of the Electric Double Layer Capacitance of Graphene Electrodes in Mono-Valent Aqueous Electrolytes. Nano Res. 2016, 9, 174−186. (32) Vatamanu, J.; Cao, L.; Borodin, O.; Bedrov, D.; Smith, G. D. On the Influence of Surface Topography on the Electric Double Layer Structure and Differential Capacitance of Graphite/Ionic Liquid Interfaces. J. Phys. Chem. Lett. 2011, 2, 2267−2272. (33) Jiang, D.-e.; Jin, Z.; Henderson, D.; Wu, J. Solvent Effect on the Pore-Size Dependence of an Organic Electrolyte Supercapacitor. J. Phys. Chem. Lett. 2012, 3, 1727−1731. (34) Jiang, D.-e.; Wu, J. Unusual Effects of Solvent Polarity on Capacitance for Organic Electrolytes in a Nanoporous Electrode. Nanoscale 2014, 6, 5545−5550. (35) Feng, G.; Li, S.; Atchison, J. S.; Presser, V.; Cummings, P. T. Molecular Insights into Carbon Nanotube Supercapacitors: Capacitance Independent of Voltage and Temperature. J. Phys. Chem. C 2013, 117, 9178−9186. (36) Park, S.-W.; DeYoung, A. D.; Dhumal, N. R.; Shim, Y.; Kim, H. J.; Jung, Y. Computer Simulation Study of Graphene Oxide Supercapacitors: Charge Screening Mechanism. J. Phys. Chem. Lett. 2016, 7, 1180−1186. (37) Hu, Z.; Vatamanu, J.; Borodin, O.; Bedrov, D. A Comparative Study of Alkylimidazolium Room Temperature Ionic Liquids with Fsi and Tfsi Anions near Charged Electrodes. Electrochim. Acta 2014, 145, 40−52. (38) Merlet, C.; Rotenberg, B.; Madden, P. A.; Taberna, P.-l.; Simon, P.; Gogotsi, Y.; Salanne, M. On the Molecular Origin of Supercapacitance in Nanoporous Carbon Electrodes. Nat. Mater. 2012, 11, 306−310. (39) Li, S.; Feng, G.; Fulvio, P. F.; Hillesheim, P. C.; Liao, C.; Dai, S.; Cummings, P. T. Molecular Dynamics Simulation Study of the Capacitive Performance of a Binary Mixture of Ionic Liquids near an Onion-Like Carbon Electrode. J. Phys. Chem. Lett. 2012, 3, 2465− 2469. (40) Huang, J. S.; Sumpter, B. G.; Meunier, V. A Universal Model for Nanoporous Carbon Supercapacitors Applicable to Diverse Pore

Aachen Research Alliance-High Performance Computing (JARA-HPC) from RWTH Aachen University under Project No. jara0146. H.Y. gratefully acknowledges the insightful suggestions from Zhenxing Wang (University of Kansas) and Jenel Vatamanu (University of Utah) on the MD calculations.



REFERENCES

(1) Zhang, S.; Pan, N. Supercapacitors Performance Evaluation. Adv. Energy Mater. 2015, 5, 1401401. (2) Simon, P.; Gogotsi, Y.; Dunn, B. Where Do Batteries End and Supercapacitors Begin? Science 2014, 343, 1210−1211. (3) Yu, Z.; Tetard, L.; Zhai, L.; Thomas, J. Supercapacitor Electrode Materials: Nanostructures from 0 to 3 Dimensions. Energy Environ. Sci. 2015, 8, 702−730. (4) Yang, W.; Ni, M.; Ren, X.; Tian, Y.; Li, N.; Su, Y.; Zhang, X. Graphene in Supercapacitor Applications. Curr. Opin. Colloid Interface Sci. 2015, 20, 416−428. (5) Lv, W.; Li, Z.; Deng, Y.; Yang, Q.-H.; Kang, F. Graphene-Based Materials for Electrochemical Energy Storage Devices: Opportunities and Challenges. Energy Storage Mater. 2016, 2, 107−138. (6) Chen, J.; Li, C.; Shi, G. Graphene Materials for Electrochemical Capacitors. J. Phys. Chem. Lett. 2013, 4, 1244−1253. (7) Yang, Y.; Han, C.; Jiang, B.; Iocozzia, J.; He, C.; Shi, D.; Jiang, T.; Lin, Z. Graphene-Based Materials with Tailored Nanostructures for Energy Conversion and Storage. Mater. Sci. Eng., R 2016, 102, 1−72. (8) Stankovich, S.; Piner, R. D.; Nguyen, S. T.; Ruoff, R. S. Synthesis and Exfoliation of Isocyanate-Treated Graphene Oxide Nanoplatelets. Carbon 2006, 44, 3342−3347. (9) Liu, W.; Yan, X.; Lang, J.; Chen, J.; Xue, Q. Influences of the Thickness of Self-Assembled Graphene Multilayer Films on the Supercapacitive Performance. Electrochim. Acta 2012, 60, 41−49. (10) Lobato, B.; Vretenár, V.; Kotrusz, P.; Hulman, M.; Centeno, T. A. Reduced Graphite Oxide in Supercapacitor Electrodes. J. Colloid Interface Sci. 2015, 446, 203−207. (11) Chen, Y.; Zhang, X.; Zhang, D.; Yu, P.; Ma, Y. High Performance Supercapacitors Based on Reduced Graphene Oxide in Aqueous and Ionic Liquid Electrolytes. Carbon 2011, 49, 573−580. (12) Bo, Z.; Mao, S.; Jun Han, Z.; Cen, K.; Chen, J.; Ostrikov, K. Emerging Energy and Environmental Applications of VerticallyOriented Graphenes. Chem. Soc. Rev. 2015, 44, 2108−2121. (13) Bo, Z.; Wen, Z. H.; Kim, H.; Lu, G. H.; Yu, K. H.; Chen, J. H. One-Step Fabrication and Capacitive Behavior of Electrochemical Double Layer Capacitor Electrodes Using Vertically-Oriented Graphene Directly Grown on Metal. Carbon 2012, 50, 4379−4387. (14) Kalluri, R. K.; Ho, T. A.; Biener, J.; Biener, M. M.; Striolo, A. Partition and Structure of Aqueous Nacl and Cacl2 Electrolytes in Carbon-Slit Electrodes. J. Phys. Chem. C 2013, 117, 13609−13619. (15) Peter, C.; Hummer, G. Ion Transport through MembraneSpanning Nanopores Studied by Molecular Dynamics Simulations and Continuum Electrostatics Calculations. Biophys. J. 2005, 89, 2222− 2234. (16) Feng, G. A.; Qiao, R.; Huang, J. S.; Sumpter, B. G.; Meunier, V. Atomistic Insight on the Charging Energetics in Subnanometer Pore Supercapacitors. J. Phys. Chem. C 2010, 114, 18012−18016. (17) Randin, J. P.; Yeager, E. Differential Capacitance Study of StressAnnealed Pyrolytic Graphite Electrodes. J. Electrochem. Soc. 1971, 118, 711−714. (18) Yuan, W.; Zhou, Y.; Li, Y.; Li, C.; Peng, H.; Zhang, J.; Liu, Z.; Dai, L.; Shi, G. The Edge- and Basal-Plane-Specific Electrochemistry of a Single-Layer Graphene Sheet. Sci. Rep. 2013, 3, 2248. (19) Seo, D. H.; Yick, S.; Pineda, S.; Su, D.; Wang, G.; Han, Z. J.; Ostrikov, K. Single-Step, Plasma-Enabled Reforming of Natural Precursors into Vertical Graphene Electrodes with High Areal Capacitance. ACS Sustainable Chem. Eng. 2015, 3, 544−551. (20) Mao, S.; Wen, Z.; Ci, S.; Guo, X.; Ostrikov, K. K.; Chen, J. Perpendicularly Oriented Mose2/Graphene Nanosheets as Advanced Electrocatalysts for Hydrogen Evolution. Small 2015, 11, 414−419. 159

DOI: 10.1021/acs.jpclett.6b02659 J. Phys. Chem. Lett. 2017, 8, 153−160

Letter

The Journal of Physical Chemistry Letters Regimes, Carbon Materials, and Electrolytes. Chem. - Eur. J. 2008, 14, 6614−6626. (41) Huang, J.; Sumpter, B. G.; Meunier, V. Theoretical Model for Nanoporous Carbon Supercapacitors. Angew. Chem., Int. Ed. 2008, 47, 520−524. (42) Zhang, S.; Bo, Z.; Yang, H.; Yang, J.; Duan, L.; Yan, J.; Cen, K. Insights into the Effects of Solvent Properties in Graphene Based Electric Double-Layer Capacitors with Organic Electrolytes. J. Power Sources 2016, 334, 162−169. (43) Merlet, C.; Salanne, M.; Rotenberg, B.; Madden, P. A. Influence of Solvation on the Structural and Capacitive Properties of Electrical Double Layer Capacitors. Electrochim. Acta 2013, 101, 262−271. (44) Shim, Y.; Jung, Y.; Kim, H. J. Graphene-Based Supercapacitors: A Computer Simulation Study. J. Phys. Chem. C 2011, 115, 23574− 23583. (45) Zhan, C.; Zhang, Y.; Cummings, P. T.; Jiang, D.-e. Enhancing Graphene Capacitance by Nitrogen: Effects of Doping Configuration and Concentration. Phys. Chem. Chem. Phys. 2016, 18, 4668−4674. (46) Hassan, M.; Haque, E.; Reddy, K. R.; Minett, A. I.; Chen, J.; Gomes, V. G. Edge-Enriched Graphene Quantum Dots for Enhanced Photo-Luminescence and Supercapacitance. Nanoscale 2014, 6, 11988−11994.

160

DOI: 10.1021/acs.jpclett.6b02659 J. Phys. Chem. Lett. 2017, 8, 153−160