Exploiting Anti-T-shaped Graphene Architecture to Form Low Tortuosity, Sieve-like Interfaces for High-Performance Anodes for Li-Based Cells.

Graphitic carbon anodes have long been used in Li ion batteries due to their combination of attractive properties, such as low cost, high gravimetric energy density, and good rate capability. However, one significant challenge is controlling, and optimizing, the nature and formation of the solid electrolyte interphase (SEI). Here it is demonstrated that carbon coating via chemical vapor deposition (CVD) facilitates high electrochemical performance of carbon anodes. We examine and characterize the substrate/vertical graphene interface (multilayer graphene nanowalls coated onto carbon paper via plasma enhanced CVD), revealing that these low-tortuosity and high-selection graphene nanowalls act as fast Li ion transport channels. Moreover, we determine that the hitherto neglected parallel layer acts as a protective surface at the interface, enhancing the anode performance. In summary, these findings not only clarify the synergistic role of the parallel functional interface when combined with vertical graphene nanowalls but also have facilitated the development of design principles for future high rate, high performance batteries.

Introduction

Lithium ion (Li ion) batteries are a popular choice for energy storage, mainly due to their good specific energy and energy density values, which make these small, lightweight devices suitable for applications such as portable electronics and electric/hybrid electric vehicles.[1−3] However, as demand for improved energy storage has increased, so too have the requirements for improved Li ion batteries. While most work in this area has focused on improving the most significant constituent parts of the battery (i.e., the cathode, anode, and electrolyte), significant advances may be achieved by enhancing battery performance through improved electrode design without alteration of the materials’ chemistry.[4]

Currently, one of the most significant limitations of current Li ion batteries, particularly at high rates, lies in the diffusion of Li ions in electrodes. This arises predominantly not from the insertion/extraction of Li ions within the particle’s bulk but instead from the diffusion of Li ions across the electrode surface. By using low tortuosity electrodes, it may be possible to minimize the electron and ion pathways and thus improve this property.[4] Consequently, the inherently fast diffusion of charge carriers through low tortuosity, morphologically anisotropic, vertical carbon materials—such as such as nanowalls,[5] carbon nanotube arrays,[6] and graphene nanowalls[7]—has attracted considerable interest.[8,9] However, in order to unlock the potential of these materials it will be necessary to fabricate materials with the required carefully designed architectures.

Another important consideration regarding the performance of Li ion battery electrodes lies in the solid electrolyte interphase (SEI). Conventionally, the formation of an electrode’s SEI typically occurs due to the interaction between the electrode and the electrolyte, and thus is strongly influenced by their respective natures. For example, in some cases the SEI may be relatively thick (and thus insulating against Li ions) or even not form at all (as is the case with some electrolytes commonly used in lithium–sulfur batteries). Thus, in order to exploit a wide range of new and promising electrolytes it will first be necessary to ensure that there is no decrease in performance relating to the SEI, especially in the protection of the electrode.

In order to address these two significant issues, we consider the use of a carefully designed electrode interface, based upon highly structured nanocarbon, which combines low-tortuosity channels with lateral layers offering surface protection to enhance interfacial stability. In this way, it is possible to exploit the synergy of this material’s properties to create electrodes suitable for use with a wide array of applications, and in particular for lithium–sulfur batteries, with improved electrochemical performance. Finally, we propose that this simple approach has considerable potential when considering the future design of electrodes.

Design Principles for the Electrode/Electrolyte Interface

The principal behind the design of these nanocarbon electrode/electrolyte interfaces is to exploit the low-tortuosity nature of vertical graphene, as well as the protective nature of horizontally aligned graphene layers, in order to facilitate the fabrication of high-performance electrodes (which may then be used with a broader range of electrochemical systems). This combination of vertical and horizontal graphene porous coating has led to the design of an “anti-T-shaped” graphene interface, which enables formation of sieve-like interfaces with high stability and high rate of ion intercalation/de-intercalation, with the vertical and parallel units acting as ion selective channels and protective layers, respectively.

While the benefits of the low-tortuosity, vertically aligned graphene have been explored, those which may be attained from a horizontally aligned interface have predominantly been neglected. Given the potential value of a protective layer to enhancing interface stability, developing routes to surface layers with both surface protection and channels for fast ion diffusion is an important area of research. Indeed, one area in particular where such work is of interest is when SEIs are not formed or lead to poor performance. For example, graphite anodes form a stable, sieve-like SEI with ethylene carbonate (EC) based electrolytes, which can protect the electrode and improve electrochemical performance.[10−12]

However, EC is incompatible with the polysulfides and/or sulfur radicals typically found in working Li–S batteries,[13,14] necessitating other approaches in order to exploit the potentially high energy densities of these systems (Li metal anode, ∼2600 Wh/kg; graphite anode: ∼600 Wh/kg;[15,16] see details in Supporting Information, under Supporting Information for Li-ion sulfur Battery). While carbon coating has been attempted, as it is fairly common way to modify surface properties to stabilize the interfaces between electrodes and electrolytes,[17,18] it has been found that the resulting disordered carbon structures and defects cannot act as the desired sieve-like interface (Figures a and 1b).

Figure 1

Schematic diagram of anti-T-shaped graphene interface leading a sieve-like interface based on different Li ion transport channel for graphene, amorphous carbon, and anti-T-shaped graphene. Schematic diagram of ion channels in (a) 2D graphene, (b) 3D amorphous carbon, and (c) anti-T-shaped graphene. Schematic diagram of (d) 2D graphene nanowalls under an electric field confine the matching ion transport, and (e) the effect for the carbon paper coated multilayer-graphene nanowalls and SEI films to sieving ion intercalation (R– means other negative ions, R1 represents molecules in the electrolyte).

By contrast, an anti-T-shaped graphene interface, consisting as it does of both protective layers and fast ion diffusion vertical channels (as shown in Figures c and 1d), is ideally suited as an interface for these types of systems.

However, when considering the use of anti-T-shaped graphene as an interface, it is important to note the following factors. The interlayer spacings between the 2D materials, such as in multiple-layered graphene, enable the 2D confinement of ions. Moreover, cations can pass quickly through if the 2D materials are negatively charged.[19] Thus, when selectively charged, graphene-based membranes can act as an efficient ion-sieving layer.[20] In short, the integration of these two mechanisms can lead to an effective roadmap for designing interfaces comprising a suitable size and charge selectivity.

Fabrication of the Nanocarbon Electrode/Electrolyte Interface

The use of chemical vapor deposition (CVD) to fabricate targeted graphene materials with tunable properties makes it ideal for this work, particularly given the potential for forming vertical graphene nanowalls with plentiful and unique interfaces (e.g., the possible creation of parallel graphenes), which in turn enables a high degree of tuning of the subsequent electrode/electrolyte interface and the formation of low tortuosity Li ion diffusion routes.

In essence, CVD occurs via bottom-up heterogeneous nucleation (predominantly as described by Laplace’s equation). According to the film nucleation theory, it is possible to enable interfacial growth of parallel graphene, amorphous carbon, or even carbides (depending on the nucleation conditions occurring on the substrate). Indeed, the morphology and thickness of graphene coatings on substrates can be defined by the carbon solubility in metal and the growth conditions.

For example, Cu substrates can induce graphene coating via a surface adsorption self-limited mechanism, while Ni substrates induce graphene coating by a precipitation mechanism;[21] the nucleation of surface carbide Al3C4 can result from Al substrate as a consequence of improved matching between substrate and graphene;[22] and low temperatures can lead to growth of amorphous carbon by consuming low active gas species.[18]

It is notable that highly active, dense hydrocarbon fragments may be synthesized via CVD in which a vertical architecture of graphene nanowalls may be obtained as a result of a thermodynamic nucleation mechanism.[22,23] This growth mechanism will be particularly favorable when assisted by an extra electric field, i.e., by using advanced plasma enhanced CVD (PECVD) technology.[24−26]

Subsequently, a new form of nanocarbon needs to be defined, where vertical graphene nanowalls are coupled with an immediate parallel graphene coating. Such two mirror-reflected-L graphenes can be described as an anti-T-shaped graphene integration, as shown in Figure . As shown in Figure c, the pore diameter is only ∼1.4 Å within each carbon hexagonal ring of graphene, smaller than that of Li+ ions (∼1.8 Å) and any molecules in the electrolyte.

As shown in Figure e, the multilayer-graphene nanowalls not only act as a protecting layer for exclusive Li ion intercalation but also provide more active sites (e.g., the open carbon edges of vertically oriented graphene nanowalls may act as ideal channels to facilitate ion transport). As opposed to active sites in defect-orientated materials, these may act as an effective charge-confined channel for Li+ diffusion. During intercalation the first protecting layer repels anions and molecules away from an electrode surface, while the second one suppresses large-sized ions to intercalate into the electrode. This synergistic effect is the main factor behind the observed improved electrochemical performance.

Results and Discussion

In order to investigate the anti-T-shaped graphene interface, carbon paper (CP) was used as the commercial current collector (as is typical for Li–S, Li ion, and Li–O2 batteries). Since the selected Toray carbon papers are abundant with the inherent graphite-containing components,[27] the woven fibers may be also treated as an active material for Li ion intercalation.

Figure a shows that the CPs consist of gray colored “knitted” fibrous carbon with an average diameter of ∼10 μm. In contrast, carbon paper coated with vertical graphene (CPVG) is darker (Figure S1a); notably, the porous structure of CP (Figure S1b) enables the penetration of the active plasma gas into the CP, guaranteeing sufficient coating (Figure S1c). As shown in Figure b, the CPs were put on the electrode plate (also referred to as the “bottom electrode” for the deposited area). Subsequently, it was found that it was possible to grow multilayer-graphene nanowalls on the CPs by applying radiofrequency (RF)-PECVD, leading to the formation of CPVGs. From the SEM image in Figure c it may be seen that the fabricated multilayer-graphene nanowalls were densely distributed on the CPs uniformly. Figure c also demonstrates that the thickness of carbon nanowall is ∼20 nm, which is consistent with the HRTEM measurement (5–20 nm). The interlayer spacings between each graphene are ∼0.343 nm (Figure S2a).

Figure 2

Structural and morphological characterization for the anti-T-shaped graphene interface coating on the CP. (a, c) SEM images of (a) CP and (c) CPVG. (b) Schematic diagram of coating multilayer-graphene nanowalls on CP via PECVD method. (d) SEM image of the FIB-processed lamella of CPVG. (e) HRTEM image of CP with the inset of electron diffraction. (f) STEM image of the interfaces between multilayer-graphene nanowalls and CP. (g) Intensity mapping of the fwhm for a Gaussian fitting on the π*-edge by using EELS measurement. (h, i) STEM image of (h) the substrate near the interface area and (i) another interface between the multilayer-graphene nanowalls and CP. (j) The enlarged image of the highlighted region in panel i.

Usually, for Raman spectra of vertical carbon nanowalls[28,29] a strong peak at ∼1580 cm–1 corresponds to the G band (visible here in Figure S3)[7,30] and indicates the formation of a graphitized structure; another peak at ∼1350 cm–1 correlates to the disorder-induced, D band phonon mode; finally, the 2D band at ∼2700 cm–1 is similar to the typical Raman spectra of carbon films.[26,31] It is worth noting that the large D peak observed in the CPVG spectra implies an abundance of defects and edges, which most likely resulted from the rapid nucleation and growth facilitated by PECVD.[7] The Raman spectrum also indicates that the carbon film prefers vertical growth because the D′ peak at ∼1620 cm–1 (Figure S3) represents the finite sp2 crystallite size[26] which was commonly found in vertical carbon materials due to the plentiful edges of vertical carbon materials.[7,26,28,32] It may also be seen that the C–C bonding of CP changed significantly after the anchoring of the multigraphene nanowalls. From the increase in the ID/IG ratio (increasing from 0.31 to 2.1) it may be inferred that the C–C bonding defects had increased. Moreover, the increase in D′ peak intensity also implies that there was an increase in carbon edges after the modification of the CP. From Figure S4, it may be seen that the innate properties of the CP remained even after the PECVD process, implying that the vertically aligned carbon coating only changed the surface of the CP.

HRTEM, STEM, and EELS mapping were also utilized in order to acquire CPVG interfacial information. In Figure d, a lamella was prepared by using ion beam and e-beam etching. Figures e and 2h show that the substrate (CP) was mainly a mixture of amorphous carbon and nanocrystalline graphite (see the selected area diffraction (SAED) pattern). As shown in Figure f, a parallel multilayer-graphene coating initially grew on the surface of the CPs, contributing to the formation of protective layers. In this way, an intercalation channel linked to vertical multilayer-graphene was built up for Li ions. Figure g provides the mapping of the fwhm for a Gaussian fitting on the π*-edge measured by using electron energy loss spectroscopy (EELS). According to our results, the π* feature in substrate, observable at 3.5 eV, originated from the amorphous carbon and nanocrystalline graphite. However, the π* feature at 4.7 eV is attributable to the surface of CPVG. For the main composition of the multilayer-graphene, it might be the reason that the excitonic state in multilayer-graphene doubles with a splitting that appears in theoretical calculations. It was also observed in a highly oriented pyrolytic graphite via spectroscopic measurements.[33] As shown in Figures i and 2j and Figure S2b, the contacted L-shaped graphene can be observed in the interface regions, demonstrating that this interfacial layer consisted of two parts: the parallel and the vertical multilayer-graphene layers. This unique combination of protecting parallel and low tortuosity vertical graphene layers (which function as fast diffusion ion channels capable of ion-sieving) differentiates the CPVG significantly from amorphous carbon and nanocrystalline graphite.

To investigate the performance of CPVG, the EC-based electrolyte (EC/methyl 2-ethoxyethyl carbonate (EMC)/dimethyl carbonate (DMC) = 1:1:1, 1 M LiPF6, hereafter referred to as ECe frequently used in Li ion batteries) and ether-based electrolyte (1,3-dioxolane (DOL)/dimethyl ether (DME) = 1:1, 1 M lithium bis(trifluoromethane sulfonimide) (LiTFSI) and 3% LiNO3, frequently used in Li–S batteries and referred to as DOLe) were selected for the control experiment. Figure a shows that CP is more stable as the active material in ECe than in DOLe, in accordance with the impurity effect mentioned previously.[32] A stable SEI film can occur in CP as merits of the EC-based electrolyte. However, in the DOLe commonly used in Li–S batteries, a suitable coating layer cannot result from CP; the heavy co-intercalation leads to a poor cycling life (i.e., the capacity loss in the DOLe is ca. 80% after only eight cycles). Given that the total theoretical capacity resulting from Li ion storage in the graphene nanowalls is far below 2.6%, the main advantage provided by the anti-T-shape graphene layers is a fast Li ion transport channel (see details in Supporting Information, under Supporting Information for Li-ion sulfur Battery).

Figure 3

Electrochemical performance and the surface chemical group characterization for CP and CPVG. (a) Cycling performance of the CP in ECe and DOLe during a process of 60 mA/g. (b) The rate properties of the CP in ECe and DOLe, and the CPVG in DOLe. (c, d) The charging–discharging capacity and Coulombic efficiency of 50 cycles during a process of 100 mA/g in DOLe electrolyte, respectively. (e, f) EIS of (e) CP and (f) CPVG. (g–j) The C 1s fitting data of (g) CP, (i) CP after 50 cycles, (h) CPVG, and (j) CPVG after 50 cycles.

Figure b demonstrates that the CP, after a composite with multigraphene nanowalls is formed, exhibits enhanced rate performance, and that the CPVG in DOLe provides more capacity at high current densities. Initially, at a current density of 50 mA/g, the coin cells show little difference in terms of capacity. However, cycling of the CP-DOLe cell led to a significant decrease in capacity, which may be attributed to the formation of an SEI film free of selectivity.

By contrast, the capacity of CPVG in DOLe demonstrated the best performance at all current densities: at current densities ranging from 500 to 50 mA/g, the CPVG demonstrated a relatively stable capacity. Such high performance demonstrates that the artificial isolated layer was more stable than the spontaneous SEI film. The high performance on the part of the electrode demonstrates that CPVG may be treated as an excellent choice for a stable electrode with high energy densities.

To further explore the intercalation/de-intercalation of CPVG and CP in a DOLe electrolyte, the electrochemical performance of CP and CPVG was characterized by constant cyclic capacity and current density. Figures S5a and S5b show that the discharge capacities of CP and CPVG were almost the same in the first cycle, but the lower charge voltage implies that more driving force was required for the de-intercalation of CP. The discharge voltage profile in the first cycle also shows the enhanced rate performance, which is in agreement with the data in Figure b. However, the fifth cycle of CP shows considerable differences, with the voltage profile of CP suffering from significant fluctuations (i.e., irreversible (de)intercalation). By contrast, the smooth voltage profile of CPVG reveals that the multigraphene nanowalls enable good stability. The voltage profile from the 40th to 50th cycles in Figure S5c further verified that the CP only acts as the current collector (discernible from the symmetrical voltage profile). Thus, the Li ions follow the following chemical reactions: Li+ + e– = Li (discharge), Li – e– = Li+ (charge).[34,35] The stable voltage profile of CPVG shows that it also acts as an active material (i.e., it is capable of Li ion (de)intercalation). Based on the Coulombic efficiency (CE) in Figure S5d, the CE of CP exhibits irregular fluctuations after several cycles as the electrochemical processes relate to the stripping and plating of Li metal in CP. The CE of CPVG in Figure S5d reached above 99.95%, which shows that the multilayered graphene nanowalls can effectively stabilize Li ion intercalation/de-intercalation and prevent the intercalation of impurities. Thus, this result demonstrates that the rate performance is highly dependent on interface modification.

Figure S6 shows the elemental mapping of CP and CPVG. From these, the LiTFSI was identified as occurring, with various distributions, from the interface to the carbon fibers’ surface for both CP and CPVG. Since the CP surface cannot suppress the intercalation of impurities, a higher F content occurred in interface than surface, which was the opposite of CPVG. Figure S7a gives the EDS result of the CP. Sulfur (1 atom %) and fluorine (8.9 atom %) come from the LiTFSI. In the case of CPVG (Figure S7b), the rare element appeared in the interface and the carbon ratio up to 99.9%, suggesting that this layer can prevent co-intercalation. From the Raman spectra it can be seen that the multilayer-graphene nanowalls were well preserved after 50 cycles (Figure S8), as shown by the ID/IG of 0.38 and 1.46 for CP and CPVG, respectively.

From Figure S5, it may be observed that the SEI films originating from the various surfaces yielded distinct electrochemical performances. The cyclability plots for these materials, taken over 50 cycles at a current density of 100 mA/g, is shown in Figures c and 3d. In addition, Raman spectra were measured so as to provide additional information regarding the evolution of the carbon materials’ composition; XPS data were used to analyze the composition of SEI film; and electrochemical impedance spectroscopic (EIS) experiments were conducted so as to examine the mechanistic interfacial charge exchange.

The equivalent circuit model was utilized for the EIS data fitting (Figure S9). As shown in Figure e, the high frequency area of the EIS of CP demonstrates that the initial interfacial charge impedance is ∼50 ohm/cm2, while Figure f shows that this initial interfacial impedance is ∼60 ohm/cm2 for CPVG. This suggests that the vertical-graphene coating mainly influences the initial surface charge impedance. After the charging and discharging process, the SEI film was formed in the surface of CP and CPVG. The electrode was, therefore, activated. The EIS of the charged and discharged states shows that the resistance of SEI film for the diffusion of Li ions corresponds to the high frequency around 10 to 0.01 kHz, while the resistance of Li ion intercalation into active materials correlates to the medium frequency around 10 to 0.1 Hz.[36] As shown in Table S2, the resistance of Rs includes contributions from the solution, separator, and electrodes. The Rs of CPVG changed from 4.0 to 2.3 ohm/cm2 after the SEI was formed. Meanwhile, the R of CPVG mostly changed between the charged and discharged states (2.3 and 7.5 ohm/cm2), which means that the resistance of the electrode changed during the intercalation/de-intercalation of Li ions into/from the active materials.

By contrast, the CP mainly acts as a current collector. Here the ionic conductivity passing through the SEI film is 50 ohm/cm2 (R). Thus, Li ion intercalation into CP is hindered due to the high resistance of Rct (∼1000 ohm/cm2, shown in Table S2). Moreover, there is a little change between the charged and discharged states, demonstrating the slow electrochemical dynamics illustrated by Figures e and 3f.

The resistance, Rs, of CPVG changes from 2 ohm/cm2 (in the charged state) to 7 ohm/cm2 (in the discharged state), demonstrating the resistance of the electrode changed with delithiation/lithiation. The SEI film (due to the low tortuosity vertical multilayer-graphene channels) exhibits little charge transfer resistance, R, as shown in Table S2 (resistance is ∼6 ohm/cm2). However, notably R disappeared during the lithiation of CPVG. This may be attributable to the influence of surface multigraphene nanowalls. During delithiation, this region suffers from a paucity of Li ions, where it may be active for Li ion intercalation. In contrast, during the lithiation this region is abundant with Li ions, and lacks active sites for Li ion intercalation. In essence, the SEI film is a hierarchical mixture. The SEI film on CP exhibits slower electrochemical dynamics than that found for CPVG. In short, the rate capability is predominantly affected by the composition of the SEI film. The anchored multilayer-graphene nanowalls contribute to ionic and electronic conductivity, which results in enhanced rate performance, and the disordered inert carbon surface leads to the suppression of the electrode activation at high current density and of rapid Li ion intercalation.

In addition to improved electrode lifespan, the composite electrode also shows enhanced rate capability. This indicates that the interface does indeed facilitate Li ion transport. In contrast to heteroatom coating,[37] our results highlight the significant influence on rate performance that the composition and surface state of carbon structures can have. From Figure c it is also possible to see that the CPVG can provide more capacity than CP, at a current of 100 mA/g, even from the first discharge. This result suggests that the rate performance may be attributed to the modification of CP, rather than capacity loss due to an irreversible intercalation. The CPVG exhibited a high initial CE, and this composite structure is, therefore, suitable for the mass production of batteries, considering that the total amount of Li ions has been defined in the cathode materials in most commercial Li ion batteries.

Finally, X-ray photoelectron spectroscopy (XPS) was employed in order to investigate the differences between the surfaces of CPVG and CP after 50 cycles. As shown in Figure S9, the surface chemical groups mainly contain F, N, Li, S, N, O, and C from the LiTFSI and substrate (CP or CPVG). Examination of the C 1s fitted data, shown in Figures g and 3h, reveals that the surface chemical composition changed after the growth of the multigraphene. Due to the ambient atmosphere during a PECVD process (e.g., the plasma gas of CH3+, H+, etc.), the surface carbon is mainly formed by C–H or C–C dangling bonds. Consequently, there are almost no C–O bonds observable in the CPVG C 1s data. The two peaks located at 284.5 and 285.5 eV may be attributed to C sp2 and sp3 bonds, respectively.[38] By contrast, the C 1s of CP indicates that the CP is mainly composed of C sp2 bonds (peak at 284.5 eV) and C sp3 bonds (peak at 285.5 eV), with peaks at higher binding energies resulting from C–OH, C–OH, and COOH (as shown in Table S3).[39]

Examination of the C 1s data for CP and CPVG after 50 cycles, shown in Figures i and 3j, reveals peaks at 286.7 and 286.5 eV, which may be attributed to C–O or C–N functional groups; the peaks at 289 and 288 eV were assigned to the C=O based groups; and the peaks at 290.2 and 289.6 eV represent the −COO– or CO32– based groups. In addition, the peak at ∼293 eV corresponds to the chemical group of CF3 (which likely originated from the TFSI–). The higher VI peak ratio during the discharged state (CP, 30%; CPVG, 15%) confirmed the existence of TFSI– in the surface of CP, and at a higher concentration. The higher binding energy of C 1s of CP and high percent of peak V (7.9% inTable S3), demonstrated that the SEI film of CP mainly contains Li2CO3.[40] So the SEI film induced by the composite structure suppresses the formation of Li2CO3 (peak V in Figure j: 1.4%). Meanwhile, it is clear that the SEI film in CP exhibits more sp2 C–C (25.1%) and the SEI film in CPVG holds a little of sp2 C–C (6.9%) and more peak II content (54.6%). It might be that the ions intercalated/de-intercalated into/from the surface vertical graphene layer, which leads to a high binding energy shift.[41]

The O 1s fitted data for CP, CPVG, and CP/CPVG, after 50 cycles, are shown in Figures S11a and S11c and Table S4 and suggest that the surface of CP was terminated by O or C dangling bonds (inferred from peak 1 in Figure S11a). From Figure S11c, it may be seen that the surface of CPVG appears to be terminated by the H or C dangling bonds due to peaks 1 and 2 in Figure S11c, which are likely the result of surface oxygen adsorption. There is no observable C–O peak, which is in keeping with the C 1s data.[42]

From examination of Figures S11b and S11d, it may be seen that there are considerable differences between peaks 1 and 2 for the CP and CPVG materials, notably the ∼1 eV shift in peak 1 binding energy after 50 cycles, implying that there are significant differences between the chemical groups present in the two SEI films. Figure S11b exhibits that peak 1 of CP50 at 531.8 eV corresponds to the Li2O2 or LiOH, and peak 2 of CP50 at 532.6 eV may correspond to the Li2CO3.[40] Combined with the data of Figure S12, the emerged peak at ∼61 eV was assumed to originate from the peak of oxidized Li.[43] So peak 1 of CPVG50 at 530.4 eV may be probably R–O–Li between the surficial vertical graphene. All the XPS data have demonstrated that the surface of CPVG provides a reversible ion intercalation channel with a function of sieving for Li ions. The vertical graphene structures, as the SEI film replacing Li salts such as Li2CO3, may induce a higher electronic and ionic conductivity for enhancing the rate performance. Finally, the discharge for a full cell consisting of CP or CPVG as the anode, coupled with a Li2S cathode (shown in Figure S13), demonstrated that the anti-T-shaped graphene coating layer does indeed render the CPVG active in a full Li–S cell.

Conclusions

In summary, PECVD technology has been used to grow the multigraphene nanowalls on a CP electrode, forming carefully designed anti-T-shaped graphene interfaces.

The protective, parallel graphene layer improved cycle life and prevented the intercalation of impurities, making it particularly suitable for systems which are unfavorable for the formation of an SEI layer with good ion diffusion properties. In this way, this hierarchical structure can be useful for carbon-based anodes, particularly in Li–S batteries. Meanwhile, the traditional carbon coating field for Li ion battery as well as other fields may march into alternative graphene coating, as this delicate graphene structure holds better electronic and ionic conduction, also especially sieving-like performance.

The low tortuosity, vertically aligned channels proved regions of fast Li ion diffusion, and afford plentiful active sites to grow an SEI film for fast Li ion transport, leading to the observed enhanced rate performance. Thus, by exploiting the multilayer-graphene nanowall coating, it is possible to forge a stable electrode via the control of the surface intercalation channels and charges.[24,25]

It is also worth noting that the proof-of-concept formation of anti-T-shaped 2D materials may be applicable to other electrode materials (e.g., MoS2, MXene, Ni(OH)2 etc.), and these new interface designs may represent a new avenue for stabilizing electrodes and enhancing the rate performances of electrodes in a wide range of systems (e.g., ion intercalation for dual-ion battery cathodes, alkali ion intercalation for the cathodes in Prussian blue and its analogous systems, etc.).[44−46]

Thus, the results presented here not only represent a route to high-performance electrodes for Li–S systems but also prove the design principles which could be key for the development of future high-performance electrode interfaces in general. In this way, the design strategies and methodology employed here could potentially be applicable for a wide range of systems (e.g., different electrolytes, cathodes, and anodes), in order to improve the performance of future battery technologies.