Biphenylene and Phagraphene as Lithium Ion Battery Anode Materials

May 31, 2017 - Materials. David Ferguson,. †. Debra J. Searles,. †,‡ and Marlies Hankel*,†. †. Australian Institute for Bioengineering and N...
0 downloads 0 Views 6MB Size
Research Article www.acsami.org

Biphenylene and Phagraphene as Lithium Ion Battery Anode Materials David Ferguson,† Debra J. Searles,†,‡ and Marlies Hankel*,† †

Australian Institute for Bioengineering and Nanotechnology and ‡School of Chemistry and Molecular Biosciences, The University of Queensland, Brisbane, QLD 4072, Australia ABSTRACT: We present results of density functional theory calculations on the lithium (Li) ion storage capacity of biphenylene (BP) membrane and phagraphene (PhG) which are two-dimensional defected-graphene-like membranes. Both membranes show a larger capacity than graphene, Li2C6 and Li1.5C6 compared to LiC6. We find that Li is very mobile on these materials and does not interact as strongly with the membranes. In the case of BP we also investigated the possible volume expansion on Li insertion. We find a 11% expansion, which is very similar to the one found in graphite. Our findings show that both membranes are suitable materials for lithium ion battery anodes. KEYWORDS: DFT, lithium ion battery, anode material, 2D materials, graphene



INTRODUCTION Carbon (C)-based materials have been widely pursued over the past decades as energy storage materials. Graphite is currently used commercially as an anode material in rechargeable lithium (Li) ion batteries. It is low in cost and has a low and flat voltage range, high Coulombic efficiency, good cyclability, and a small volume change on Li insertion.1−3 Graphite can store lithium at a capacity of one Li atom per six C atoms, LiC6, which corresponds to a specific capacity of 372 mAh/g.4 2D layered defected or porous graphene-like membranes have become quite popular in recent years as possible anode materials. Their layered structure should permit the Li to intercalate during the charging process between the layers. The defects or pores in the membrane can also offer sites for the Li to reside within the membrane plane, which should have a positive effect on the change in volume during intercalation and result in an acceptable volume change which is comparable to that of 10% in graphite.5 The porous materials often have uniform pores which can act as active sites for Li adsorption to increase capacity. If the defects or pores are large enough, that is, they are the size of the ion or larger, they could also facilitate high Li mobility through the layers of the material. One new exciting porous carbon material is graphdiyne6 which is made of 2D layers of sp- and sp2-bonded carbons and has been investigated as possible anode materials for lithium ion batteries.7,8 Sun et al.8 predicted via density functional theory calculations that graphdiyne could provide a capacity of 744 mAh/g (LiC3) and excellent mobility for lithium diffusion parallel to the membrane as well as between layers perpendicular to the membrane. The theoretical prediction of the enhanced properties of graphdiyne over graphene has recently been confirmed by experiment. Huang et al.9 obtained reversible capacities of up to 520 mAh/g after 400 cycles at a © 2017 American Chemical Society

current density of 500 mA/g and 420 mAh/g after 1000 cycles at an even higher current density of 2 A/g. In two very recent theoretical studies graphdiyne was also suggested as an anode material for sodium ion batteries.10,11 These theoretical findings have also been confirmed by an experimental study.12 Graphenylene, also denoted as BPC, is a possible isomerization product of graphyne13,14 and was predicted to be stable over 40 years ago by Balban.15,16 A few years ago Brunetto et al.17 proposed that BPC could be synthesized via dehydrogenation of porous graphene.18 Since Brunetto’s et al.’s study, BPC in the form of nanotubes and fullerenes has also been investigated.19,20 Song et al.21 studied BPC as a membrane for gas separation. Other theoretical studies showed that BPC has a small band gap of around 0.8−1.08 eV17,22 which can be tuned by hydrogenation or halogenation.23 In two theoretical studies relevant to the current study, Yu24 and Hankel et al.25 investigated BPC as an anode material for lithium ion batteries employing DFT. Both studies showed a larger capacity than graphite of Li3C6 and that both in-plane and out-of-plane barriers for lithium diffusion are small and can be easily overcome. In addition to the above ordered porous 2D materials, graphene with different defects has been investigated intensively by theory.26,27,29−32 These studies found that defects in graphene such as vacancies and Stone−Wales defects enhance the adsorption of Li on the membranes as well as the charge transfer from the Li to the membranes. They also found that larger defects such as divacancies show a more significant enhancement in Li adsorption and capacity. Also, the defect Received: March 23, 2017 Accepted: May 31, 2017 Published: May 31, 2017 20577

DOI: 10.1021/acsami.7b04170 ACS Appl. Mater. Interfaces 2017, 9, 20577−20584

Research Article

ACS Applied Materials & Interfaces

3 eV, to enable desorption.7 The so-called cohesive energy has been calculated by theory similar to the one employed here24,27 and is in excellent agreement with the experimental value of 1.63 eV.28 The specific and volumetric capacities should be larger than graphite’s if it is to be of practical utility, and the volume change on lithium insertion should be small. In addition, the membrane should retain its structural integrity. In our calculations we employed monolayer membranes as well as a bilayer membrane for BP. The monolayer calculations for BP employed a 2 × 2 supercell (containing 24 carbon atoms) as shown in Figure 1 while the calculations for monolayer PhG used a unit cell

density plays a role, and a larger density of defects enhances the Li capacity of the membrane. In 2014, Müllen and co-workers published the synthesis of isomeric graphene built from biphenylene units.33 These new biphenylene (BP) membranes show six-membered carbon rings as well as four- and eight-membered carbon rings. The eightmembered rings are similar to a divacancy which has been found to increase Li capacity.26,27,29−32 BP membranes have also been the subject of several theoretical studies.34−39 These have mainly focused on the electronic properties of the BP membranes and its ribbons and nanotubes where it was found that the pristine sheet is metallic with no band gap. Denis et al.38 investigated hydrogen storage on lithium and calcium decorated BP membrane. They found that Li and Ca adsorb strongly in the eight membered rings, making it a possible candidate for an anode material for Li ion batteries with increased capacity. In 2015, Wang et al.40 proposed a new carbon allotrope, phagraphene, composed of five-, six-, and seven-membered carbon rings which features distorted Dirac cones and which was confirmed to be stable via phonon analysis. Sun et al.41 then investigated the mechanical properties and found it less stable than graphene under strain. Phagraphene represents a 2D carbon membrane with a large density of defects and might therefore prove to be a suitable candidate for a Li ion battery anode. To the best of our knowledge, neither BP membrane nor phagraphene has been investigated as a Li ion battery anode material.



Figure 1. (a) 2 × 2 supercell of biphenylene (BP); (b) unit cell of phagraphene (PhG). Gray balls represent carbon. The letters label sites that are referred to in the main text.

(containing 20 carbon atoms). For BP there are three different types of ring, indicated as A, D, and G, in Figure 1a, while for PhG there are four different types of ring, indicated as A, B, D, and E in Figure 1b. In the first instance we placed a single Li at different positions on the membranes. These included over the rings, bonds, and carbon atoms. We then performed a geometry optimization to find the preferred lithium adsorption sites for each membrane. To calculate the Eads for different Li loadings, first one Li atom was placed onto the membrane over 12 different positions for BP, and nine for PhG, guided by the results from the single Li adsorption calculations, and the geometry was optimized. Then the configuration with the lowest energy was selected and another atom was placed at this configuration, again in the remaining possible positions. This procedure was continued until all possible positions were filled or the adsorption energy per Li was lower than the cohesive energy of Li, which we have taken as 1.63 eV.24,27 In the majority of our calculations we only consider a single sheet of the membrane with only one-sided adsorption. The rationale of such an approach is that in a layered material the Li would insert between the layers. The membranes we consider in our study have an interlayer distance of around 3 Å in a bulk system. This means that the Li will form a “single” layer between the membrane sheets. If we now consider, for example, a five-layer system, then Li can adsorb between the sheets as well as on top of the top layer and below the bottom layer. If we now peel off the layers a single sheets, we would find that the top layer has Li on top and below, so Li on both sides. However, the second layer only has Li below as the Li above have already been associated with the top layer. The same is true for all the other layers. Considering this, the top layer has Li adsorbed on both sides while the other four layers have Li adsorbed only on one side. Therefore, we can consider the capacity of a one-sided loading as an approximation of the capacity in a bulk system where the interlayer spacing is around 3 Å. This is a good approximation if packing is the dominant factor determining the loading but is less reliable when bonding occurs. To investigate the possible volume expansion in a multilayer system, we use a bilayer membrane. Here we insert the maximum loading that was found for the monolayer membrane in between two layers of the membrane to see how far these two layers move apart due to the inserted Li atoms. This expansion can then be taken as an upper bound to the possible volume expansion of the material.

METHODS

Calculation Details. All calculations have been carried out with the Vienna ab initio Simulation Package (VASP).42,43 3D periodic boundary conditions were applied to simulate the infinitely large systems. A 20 Å vacuum space between sheets was set to prevent the interaction between two membrane layers. The cell parameters for a unit cell of BP were found to be a = 4.525 Å and b = 3.775 Å. Lattice parameters for a unit cell of PhG are a = 8.09 Å and b = 6.65 Å. The Brillouin zone for a 2 × 2 supercell of BP and unit cell of PhG was sampled by 4 × 4 × 1 k-points. The electronic structure of the system was treated using the generalized gradient approximation (GGA) with the PBE functional.44 The van der Waals interactions were added to the standard DFT description by Grimme’s D2 scheme.45 All calculations include spin polarization. In all calculations the convergence parameters were 10−6 eV for the energy, 0.01 eV/Å for the forces, with an energy cutoff of 500 eV. A Gaussian smearing of 0.05 eV was applied. Charge analysis was performed via Bader analysis,46−49 which included the core charges, and charge density difference analysis within VASP. To study the mobility of the Li on the two membranes, we employed the climbing image nudged elastic band (CI-NEB)50,51 method within VASP. The parameters used were the same as in the geometry calculations, and one image was employed for each system studied. For the two end points we selected final structures of the geometry calculations with a single Li atom. The strength of the Li−Mem interaction, where Mem is the pristine membrane, is given by the average adsorption energy, Eads

Eads = (nE(Li) + E(Mem) − E(Li nMem))/n

(1)

where n is the total number of adsorbed Li atoms and E(Li), E(Mem), and E(Lin@Mem) are the total energies of a single Li, the pure monolayer or bilayer membrane, and the interacting Li@Mem system, respectively. By this definition, a positive Eads indicates that the interacting system is stable and that the adsorption of Li is favorable. For a material to be suitable as an anode material it has to meet certain requirements. The lithium adsorption energy should be above that of bulk lithium, 1.63 eV,24,27 to avoid clustering, and below about 20578

DOI: 10.1021/acsami.7b04170 ACS Appl. Mater. Interfaces 2017, 9, 20577−20584

Research Article

ACS Applied Materials & Interfaces



RESULTS AND DISCUSSION

Monolayer Membranes. In the first instance we placed a single Li atom over positions A, G, and D for BP and A, B, D, E, and F for PhG, to find the preferred binding site. For BP the preferred site indicated as A in Figure 1a, see Figure 2. Li adsorbs with an energy of Eads = 2.44 eV on site A on BP. Figure 2b shows that there is no visible distortion to the membrane. Figure 4. PhG monolayer with five Li adsorbed on the membrane: (a) top view; (b) side view. Gray balls represent carbon and pink balls lithium.

adsorption energy that is too close to the cohesive energy of Li, 1.63 eV. For five Li we also checked other possible combinations of positions but found that the presented one is the most stable configuration. For a loading of five Li on PhG we can now see some distortion of the membrane. We find a higher loading for BP and no distortion to the membrane (see Figure 5). This could mean that BP would be

Figure 2. BP with one Li adsorbed in the preferred position, A: (a) top view; (b) side view. Gray balls represent carbon and pink balls lithium.

For PhG the preferred site is site A as shown in Figure 1b. Li adsorbs with an energy of Eads = 2.07 eV on PhG, and as with BP there is no visible distortion to the membrane (see Figure 3). This is consistent with previous findings for Li adsorption in

Figure 5. BP monolayer with eight Li adsorbed on the membrane: (a) top view; (b) side view. Gray balls represent carbon and pink balls lithium.

more stable during charging and discharging than PhG. For BP we can add up to eight Li with an adsorption energy of Eads = 1.83 eV. Here all eight- and six-membered rings are filled by a Li while the four carbon rings (squares) are still unoccupied. The bond length in Li2 is around 2.6 Å while the Li−Li distance in a Li crystal is around 3 Å. The Li−Li distance in the configuration of eight Li on BP is 2.95 Å. Adsorption energies well above the cohesive energy should ensure that the Li do not cluster. However, adding more Li to this configuration will give much shorter Li−Li distances coupled with adsorption energies much closer to the cohesive energy. We therefore did not proceed to add more Li to the membrane. These Li loading leads to capacities of Li1.5C6 for PhG and Li2C6 for BP which correspond to 487.47 and 623.72 mhA/g specific capacities, respectively. We should note here that we include the lithium atoms in overall mass of the material. Both of these capacities are larger than the capacity for graphite which is LiC6. The monolayer capacities therefore provide promise that both membranes could be excellent candidates for lithium ion battery anode materials. These capacities can be compared to those found by Datta et al.29 for graphene with divacancy and Stone−Wales defects. They investigated the Li storage capacities for different defect densities. They report capacities of Li4.5C6 (Li18C24) for high density divacancy defects (25%) and Li17C32 which is equivalent to Li3.1875C6, for high-density Stone−Wales defect coverage (100%). These should be considered in the context that Datta and co-workers allowed Li coverage on both sides of

Figure 3. PhG with one Li adsorbed in the preferred position, A: (a) top view; (b) side view. Gray balls represent carbon and pink balls lithium.

a divacancy and high coverage Stone−Wales defects and Li adsorption on a BP membrane.26,27,29,30,32,38 Apart from finding similar preferred positions we also find that the divacancy-like defect (A in Figure 1a) shows stronger adsorption than the Stone−Wales-like defect (A in Figure 1b). For BP the adsorption energies of the other possible positions are Eads = 2.15 eV and Eads = 1.96 eV for positions D and G, respectively. There are more distinct positions for PhG, and the adsorption energies are Eads = 2.00 eV, Eads = 2.02 eV, and Eads = 1.96 eV for positions B, E, and D, respectively. This shows that the adsorption energies between the different positions of the respective membranes are very similar. This may point to good Li mobility on the monolayer membranes. We now continue to add one more Li in all the still free possible positions to determine the preferred adsorption for the next Li. We again take the configuration with the lowest energy and continue adding Li, one by one, until either all possible positions are filled or the adsorption energy is lower than the cohesive energy. For PhG, we find that we can add up to five Li to the monolayer membrane with an adsorption energy of Eads = 1.7 eV per Li (see Figure 4). Adding one more Li will give an 20579

DOI: 10.1021/acsami.7b04170 ACS Appl. Mater. Interfaces 2017, 9, 20577−20584

Research Article

ACS Applied Materials & Interfaces the membranes. To compare the maximum capacity for a divacancy defected membrane by Datta et al.29 to our current results and to test the achievable capacity if we consider coverage on both sides of the membrane, we have loaded BP with eight Li on each side of the membrane. We find an adsorption energy of Eads = 1.85 eV per Li and only a small distortion to the membrane. This double-sided coverage would corresponds to a capacity of Li4C6 which is very close to the one reported by Datta et al.29 Our results and those of Datta et al.29 indicate that in general divacancy-like defects lead to higher Li storage capacities than Stone−Wales defects, and both provide higher loadings than those obtained with graphene sheets. We can now look at the electron transfer from Li to the membranes. Figure 6 shows the charge density difference for

Figure 7. Charge density difference plots for (a) Li in position A on monolayer PhG. The isosurface level is set to 0.001 e/a03. (b) 5 Li on monolayer PhG. The isosurface level is set to 0.0025 e/a03. Blue indicates electron excess and red electron deficiency. Gray balls represent carbon and pink ones lithium.

A similar picture is true for the 8Li on BP and 5Li on PhG cases, but here the peaks from the Li(s) orbital are now broader. The significant charge transfer and the PDOS indicate that the interaction and nature of the bonding between the Li and the two membranes are ionic. Bilayer BP. BP showed the larger possible capacity, and we therefore investigated the bilayer membrane. The single-layer model provides an estimate of the maximum total capacity of the material while the bilayer model provides an estimate of the upper limit of the possible volume expansion when the Li are intercalated between the layers of the material. We tested a range of different stacking configurations which included AA and several different AB type configurations. The different AB configurations converged to six stable configurations which are shown in Figure 9. However, we found that the AA stacking is preferred. The interlayer distance for the pristine bilayer system is 3.17 Å. From the different AB configurations, AB3 is the most stable whereas AB1 is the least stable configuration. As for the single-layer model, we find that Li binds most strongly when it is above the C8 ring. Adsorption energies for the A, D, and G positions are 3.03, 2.94, and 2.49 eV, respectively, although the G position is not a minimum-energy site as will be seen later. This shows that the interaction between the Li and the membrane is stronger in a multilayer configuration. For a single intercalated Li at position A the distance between layers is 3.40 Å, which corresponds to a 7% expansion. Figure 10 shows the bilayer model where eight Li have been intercalated between the two layers. The eight Li form the same configuration as found for the monolayer with an average adsorption energy of 2.92 eV per Li, which is much higher than the average of 1.83 eV per Li for the monolayer. As for the monolayer we find no visible distortion to the membrane even at high loadings which should ensure good stability of the membrane. When the eight Li are intercalated we find that the interlayer distance expands to 3.52 Å, which corresponds to a 11% expansion which is similar to that of graphite at 10%.5 It should be noted here again that as this is a bilayer system this expansion can be taken as an upper limit. This small expansion and the doubling in capacity compared to graphite shows that BP could be an excellent lithium ion battery anode material. Figure 6c shows the charge density difference for the case of eight Li intercalated between the layers. It is clear that also for the bilayer membrane that the Li transfer significant charge to the membrane. In the case of eight Li, each Li transfers around 0.84 e− to the membranes. PDOS analysis, where we can see a large peak in the conduction band from the Li(s) orbitals,

Figure 6. Charge density difference plots for (a) Li in position A on monolayer BP. The isosurface level is set to 0.001 e/a03. (b) 8 Li on monolayer BP. The isosurface level is set to 0.0025 e/a03. (c) 8 Li between layers of bilayer BP. The isosurface level is set to 0.0025 e/a03. Blue indicates electron excess and red electron deficiency. Gray balls represent carbon and pink balls lithium.

the BP membrane. From Figure 6a, we can see that the single Li has transferred a significant part of its charge to the membrane. This is confirmed by Bader charge analysis which show that the Li transfers about 0.89 e− to the BP monolayer. Figure 6b shows that also in the case of eight Li adsorbed on BP there is still significant charge transfer from the Li to the membrane. In the case of eight Li on BP each Li still transfers about 0.72 e− to the membrane which leads to a total charge transfer of 5.76 e−. Figure 7 shows the charge density difference for a single and five Li on PhG. Like for BP, Figures 7a,b show clearly that there is significant charge transfer from the Li to the membrane. In the case of a single Li adsorbed on the membrane the Li transfers about 0.88 e− to the membrane. This is very similar to what was seen for BP. In the case of five Li on PhG each Li transfers about 0.69 e− to the membrane which leads to a total charge transfer of around 3.48 e−. More details on the interaction of the Li with the membranes is revealed by the partial density of states (PDOS), shown in Figure 8. In the single Li case we can see a large peak in the conduction band from the Li(s) orbital, which shows no overlap with the C(p) orbitals of its neighboring carbon atoms. 20580

DOI: 10.1021/acsami.7b04170 ACS Appl. Mater. Interfaces 2017, 9, 20577−20584

Research Article

ACS Applied Materials & Interfaces

Figure 8. Partial density of states for Li adsorbed in position preferred on BP and PhG. (a) Single Li in position A on BP membrane. (b) Single Li in position A on PhG membrane. The PDOS, spin up and spin down, for the s orbital of Li and the p orbital of a representative C are shown. (c) 8Li on BP. (d) 5Li on PhG. The PDOS, spin up and spin down, for sum over all Li(s) orbitals and sum over all C(p) orbitals are shown.

Figure 9. Stable configurations of AB stackings of the BP membrane. Gray balls represent carbon.

(similar to Li on graphene which is 0.32−0.46 eV).5 This implies that even Li that are in the most stable position on the membrane, position A, are able to move away from it, indicating that the Li would be very mobile across the BP membrane. For PhG there are six possible Li movements (see Figure 12). As for BP the Li is located over a carbon−carbon bond. Table 1 shows that for PhG the barriers are slightly smaller than for BP again, indicating very good Li mobility. The barriers for the monolayer are similar to those found by other studies. Zhou et al.27 found barriers for a Li to move from

similar to the picture for the monolayer case (see Figure 8), shows that the bonding is mostly ionic. Transition State Barriers for Li Mobility. We will now look at the mobility of the Li on both membranes. For BP, possible paths for the movement of Li are between the C8 ring and the C6 ring (A ↔ D) in Figure 11a, the C8 ring and the C4 square (A ↔ G) in Figure 11b, between two C8 rings (A ↔ L) in Figure 11c, and the C6 ring and the C4 square (D ↔ G) in Figure 11d. In all the transition states the Li is located over a carbon−carbon bond. Table 1 shows that the barriers for the Li to move across the membrane are small, all below 0.45 eV 20581

DOI: 10.1021/acsami.7b04170 ACS Appl. Mater. Interfaces 2017, 9, 20577−20584

Research Article

ACS Applied Materials & Interfaces

Figure 10. (a) Top view of eight Li intercalated between layers of bilayer BP; (b) side view. Gray balls represent carbon and pink balls lithium.

Figure 12. Transition state configuration for Li movement in the PhG monolayer. The initial and final positions are indicated by the letter. The Li is shown in its position of the transition state. Gray balls represent carbon and pink balls lithium.

Figure 11. Transition state configuration for Li movement in the BP monolayer. The initial and final positions are indicated by the letter. The Li is shown in its position of the transition state. Gray balls represent carbon and pink balls lithium.

and thereby increase the Li storage capacity, the barriers for Li movement remain small. For BP we also investigated how the barriers would change in a bilayer configuration. If one considers that the adsorption energies for the bilayer membrane were higher than for the monolayer, we can expect that the barriers for Li movement might also be higher. This is indeed the case, but the barriers are still relatively small. The barriers in Table 1 show that now position G over the C4 square is not favored in a bilayer configuration and that the Li will most likely be in the C8 and C6 rings. Figure 13 shows the transition state configurations. Here we can see that for the A to G and A to D transition the transition state is now with the Li nearly above a carbon atom and not a carbon−carbon bond anymore. In a multilayer configuration for the hop from the divacancy into a smaller ring it is more favorable to go over a carbon atom than over a bond. These barriers can be compared to those found in graphite (0.218−0.4 eV).5 The barriers found here are slightly larger than those for graphite but still comparable. We can also compare these barriers to those found for lithium in graphdiyne which has been shown to be an excellent anode material experimentally. Barriers for movement from one pore to another are about ∼0.7 eV in graphdiyne.8 All barriers found for BP are smaller than this. Graphdyine with a barrier of around 0.7 eV (by theoretical studies similar to ours) has been proven to be an excellent anode material by an experimental study9 which is good evidence that the barriers found here should ensure a suitable mobility of lithium in BP membrane when used as an anode material. This indicates that based on Li

Table 1. Transition State Barriers for Li Hopping on BP, Bilayer BP (BL-BP), and PhG (Barriers Are Given in eV) transition

BP

BL-BP

transition

PhG

A→D A→G A→L G→D D→A G→A D→G

0.43 0.44 0.42 0.02 0.35 0.16 0.21

0.67 0.67 0.68 0.00 0.58 0.13 0.45

A→B A→E A→F B→D E→B E→F B→A E→A F→A D→B B→E F→E

0.37 0.42 0.30 0.22 0.35 0.15 0.31 0.38 0.20 0.17 0.33 0.08

a C8 ring to a C6 ring to be between 0.455 and 0.538 eV in a membrane with a single divacancy in the cell configuration. For a similar Li hop (BP membrane A−D) we find a barrier of 0.43 eV. They also report a barrier of 0.37 eV for the Li to move from a C7 ring to a C6 ring in a Stone−Wales defect configuration. We found the barrier for movement from configuration A to B or A to E for the PhG membrane to be 0.37 and 0.42 eV, respectively. These results are all very similar, and it shows that while defects enhance the adsorption of Li 20582

DOI: 10.1021/acsami.7b04170 ACS Appl. Mater. Interfaces 2017, 9, 20577−20584

Research Article

ACS Applied Materials & Interfaces

research was undertaken with the assistance of resources provided through The University of Queensland Research Computing Centre (RCC) and the Queensland Cyber Infrastructure Foundation (QCIF).



(1) Bhatt, M. D.; O’Dwyer, C. Recent Progress in Theoretical and Computational Investigations of Li-Ion Battery Materials and Electrolytes. Phys. Chem. Chem. Phys. 2015, 17, 4799−844. (2) Goriparti, S.; Miele, E.; De Angelis, F.; Di Fabrizio, E.; Proietti Zaccaria, R.; Capiglia, C. Review on Recent Progress of Nanostructured Anode Materials for Li-Ion Batteries. J. Power Sources 2014, 257, 421−443. (3) Zhu, G.; Lü, K.; Sun, Q.; Kawazoe, Y.; Jena, P. Lithium-Doped Triazine-Based Graphitic C3N4 Sheet for Hydrogen Storage at Ambient Temperature. Comput. Mater. Sci. 2014, 81, 275−279. (4) Zheng, F.; Yang, Y.; Chen, Q. High Lithium Anodic Performance of Highly Nitrogen-Doped Porous Carbon Prepared from a MetalOrganic Framework. Nat. Commun. 2014, 5, 5261. (5) Liu, J.; Wang, S.; Sun, Q. All-Carbon-Based Porous Topological Semimetal for Li-Battery Anode Material. Proc. Natl. Acad. Sci. U. S. A. 2017, 114, 651−656. (6) Li, G.; Li, Y.; Liu, H.; Guo, Y.; Li, Y.; Zhu, D. Architecture of Graphdiyne Nanoscale Films. Chem. Commun. 2010, 46, 3256−3258. (7) Hwang, H. J.; Koo, J.; Park, M.; Park, N.; Kwon, Y.; Lee, H. Multilayer Graphynes for Lithium Ion Battery Anode. J. Phys. Chem. C 2013, 117, 6919−6923. (8) Sun, C.; Searles, D. J. Lithium Storage on Graphdiyne Predicted by DFT Calculations. J. Phys. Chem. C 2012, 116, 26222−26226. (9) Huang, C.; Zhang, S.; Liu, H.; Li, Y.; Cui, G.; Li, Y. Graphdiyne for High Capacity and Long-Life Lithium Storage. Nano Energy 2015, 11, 481−489. (10) Xu, Z.; Lv, X.; Li, J.; Chen, J.; Liu, Q. A Promising Anode Material for Sodium-Ion Battery with High Capacity and High Diffusion Ability: Graphyne and Graphdiyne. RSC Adv. 2016, 6, 25594−25600. (11) Niaei, A. H. F.; Hussain, T.; Hankel, M.; Searles, D. J. SodiumIntercalated Bulk Graphdiyne as an Anode Material for Rechargeable Batteries. J. Power Sources 2017, 343, 354−363. (12) Zhang, S.; He, J.; Zheng, J.; Huang, C.; Lv, Q.; Wang, K.; Wang, N.; Lan, Z. Porous Graphdiyne Applied for Sodium Ion Storage. J. Mater. Chem. A 2017, 5, 2045−2051. (13) Baughman, R. H.; Eckhardt, H.; Kertesz, M. Structure-Property Predictions for New Planar Forms of Carbon: Layered Phases Containing sp2 and sp Atoms. J. Chem. Phys. 1987, 87, 6687−6699. (14) Ivanovskii, A. L. Graphynes and Graphdyines. Prog. Solid State Chem. 2013, 41, 1−19. (15) Balban, A. T.; Rentea, C. C.; Ciupitu, E. Chemical Graphs. VI. Estimation of the Relative Stability of Several Planar and Tridimensional Lattices for Elementary Carbon. Rev. Roum. Chim. 1968, 13, 231−247. (16) Randić, M.; Balaban, A. T.; Plavšić, D. Applying the Conjugate Circuits Method to Clar Structures of [n]Phenylenes for Determining Resonance Energies. Phys. Chem. Chem. Phys. 2011, 13, 20644−20648. (17) Brunetto, G.; Santos, B. I.; Autreto, P. A. S.; Machado, L. D.; dos Santos, R. P. B.; Galvao, D. S. Nonzero Gap Two-Dimensional Carbon Allotrope from Porous Graphene. J. Phys. Chem. C 2012, 116, 12810− 12813. (18) Bieri, M.; Treier, M.; Cai, J.; Aït-Mansour, K.; Ruffieux, P.; Gröning, O.; Gröning, P.; Kastler, M.; Rieger, R.; Feng, X.; Müllen, K.; Fasel, R. Porous Graphenes: Two-Dimensional Polymer Synthesis with Atomic Precision. Chem. Commun. 2009, 45, 6919−6921. (19) Koch, A. T.; Khoshaman, A. H.; Fan, H. D. E.; Sawatzky, G. A.; Nojeh, A. Graphenylene Nanotubes. J. Phys. Chem. Lett. 2015, 6, 3982−3987. (20) Paupitz, R.; Junkermeier, C. E.; van Duin, A. C. T.; Branicio, P. S. Fullerenes Generated from Porous Structures. Phys. Chem. Chem. Phys. 2014, 16, 25515−25522.

Figure 13. Transition state configuration for Li movement in the BP bilayer. The initial and final positions are indicated by the letter. The Li is shown in its position of the transition state. Gray balls represent carbon and pink balls lithium.

mobility, BP and PhG could be useful candidates for anode materials.



CONCLUSIONS We presented results for Li storage on two different 2D membranes. The first is built from biphenylene building blocks and features regular defects similar to divacancy defects found in graphene and has already been synthesized. The second, called phagraphene and which is only theoretically predicted, features regular defects similar to Stone−Wales defects found in graphene. Our results show that these defects enhance the adsorption of Li on the membrane, leading to an increased capacity that is larger than found for graphite. Charge and electronic analysis showed that the Li atoms transfer a significant amount of charge to each of the membranes and that the bonding character is ionic. We also found that while the bonding of the Li to the membrane is strong on BP membrane and PhG, the barriers for Li movement remain low enough to be overcome under normal battery operation. The expected volume change for BP membrane on Li insertion is comparable to graphite. Our results show that 2D membranes designed with a regular defect structure present promising new materials for lithium ion batteries. With BP already been synthesized we hope that our study will spark new interest in the experimental realization of such materials on a large scale suitable for commercial production.



REFERENCES

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Debra J. Searles: 0000-0003-1346-8318 Marlies Hankel: 0000-0002-8297-7231 Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS We thank the Australian Research Council for support of this project through the LIEF and Discovery programs. This 20583

DOI: 10.1021/acsami.7b04170 ACS Appl. Mater. Interfaces 2017, 9, 20577−20584

Research Article

ACS Applied Materials & Interfaces (21) Song, Q.; Wang, B.; Deng, K.; Feng, X.; Wagner, M.; Gale, J. D.; Müllen, K.; Zhi, L. Graphenylene, a Unique Two-Dimensional Carbon Network with Nondelocalized Cyclohexatriene Units. J. Mater. Chem. C 2013, 1, 38−41. (22) De La Pierre, M.; Karamanis, P.; Baima, J.; Orlando, R.; Pouchan, C.; Dovesi, R. Ab Initio Periodic Simulations of the Spectroscopic and Optical Properties of Novel Graphene Phases. J. Phys. Chem. C 2013, 117, 2222−2229. (23) Liu, W.; Miao, M.-S.; Liu, J.-Y. Band Gap Engineering of Graphenylene by Hydrogenation and Halogenation: A Density Functional Theory Study. RSC Adv. 2015, 5, 70766−70771. (24) Yu, Y.-X. Graphenylene: A Promising Anode Material for Lithium-Ion Batteries with High Mobility and Storage. J. Mater. Chem. A 2013, 1, 13559−13566. (25) Hankel, M.; Searles, D. J. Lithium Storage on Carbon Nitride, Graphenylene and Inorganic Graphenylene. Phys. Chem. Chem. Phys. 2016, 18, 14205−14215. (26) Fan, X.; Zheng, W. T.; Kuo, J.-L. Adsorption and Diffusion of Li on Pristine and Defective Graphene. ACS Appl. Mater. Interfaces 2012, 4, 2432−2438. (27) Zhou, L.-J.; Hou, Z. F.; Wu, L.-M. First-Principles Study Lithium Adsorption and Diffusion on Graphene with Point Defects. J. Phys. Chem. C 2012, 116, 21780−21787. (28) Kittel, C. Introduction to Solid State Physics, 8th ed.; Wiley: New York, 2005. (29) Datta, D.; Li, J.; Koratkar, N.; Shenoy, V. B. Enhanced Lithiation in Defective Graphene. Carbon 2014, 80, 305−310. (30) Yildirim, H.; Kinaci, A.; Zhao, Z.-J.; Chan, M. K. Y.; Greeley, J. P. First-Principle Analysis of Defect-Mediated Li Adsorption on Graphene. ACS Appl. Mater. Interfaces 2014, 6, 21141−21150. (31) Wan, W.; Wang, H. First-Principle Investigation of Adsorption and Diffusion of Ions on Pristine, Defective and B-Doped Graphene. Materials 2015, 8, 6163−6178. (32) Okamoto, Y. Density Functional Theory Calculations of Lithium Adsorption and Insertion to Defect-Free and Defective Graphene. J. Phys. Chem. C 2016, 120, 14009−14014. (33) Schlütter, F.; Nishiuchi, T.; Enkelmann, V.; Müllen, K. Octafunctionalized Biphenylenes: Molecular Precursors for Isomeric Graphene Nanostructures. Angew. Chem., Int. Ed. 2014, 53, 1538− 1542. (34) Hudspeth, M. A.; Whitman, B. W.; Barone, V.; Peralta, J. E. Electronic Properties of Biphenylene Sheet and its One-Dimensional Derivatives. ACS Nano 2010, 4, 4565−4570. (35) Wang, X.-Q.; Li, H.-D.; Wang, J.-T. Prediction of a New TwoDimensional Metallic Carbon Allotrope. Phys. Chem. Chem. Phys. 2013, 15, 2024−2030. (36) Denis, P. A. Stability and Electronic Properties of Biphenylene Based Functionalized Nanoribbons and Sheets. J. Phys. Chem. C 2014, 118, 24976−24982. (37) Karaush, N. N.; Baryshnikov, G. V.; Minaev, B. F. DFT Characterization of a New Possible Graphene Allotrope. Chem. Phys. Lett. 2014, 612, 229−233. (38) Denis, P. A.; Iribarne, F. Hydrogen Storage in Doped Biphenylene Based Sheets. Comput. Theor. Chem. 2015, 1062, 30−35. (39) Ge, H.; Wang, G.; Liao, Y. A Theoretical Investigation on the Transport Properties of Armchair Biphenylene Nanoribbons. Chem. Phys. Lett. 2016, 648, 97−101. (40) Wang, Z.; Zhou, X.-F.; Zhang, X.; Zhu, Q.; Dong, H.; Zhao, M.; Oganov, A. R. Phagraphene: A Low-Energy Graphene Allotrope Composed of 5−6-7 Carbon Rings with Distorted Dirac Cones. Nano Lett. 2015, 15, 6182−6186. (41) Sun, H.; Mukherjee, S.; Singh, C. V. Mechanical Properties of Monolayer Penta-Graphene and Phagraphene: A First-Principles Study. Phys. Chem. Chem. Phys. 2016, 18, 26736−26742. (42) Kresse, G.; Furthmüller, J. Efficient Iterative Schemes for Ab Initio Total-Energy Calculations Using a Plane-Wave Basis Set. Phys. Rev. B: Condens. Matter Mater. Phys. 1996, 54, 11169−11186. (43) Blöchl, P. E. Projector Augmented-Wave Method. Phys. Rev. B: Condens. Matter Mater. Phys. 1994, 50, 17953−17979.

(44) Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized Gradient Approximation Made Simple. Phys. Rev. Lett. 1996, 77, 3865−3868. (45) Grimme, S. Semiempirical GGA-Type Density Functional Constructed with a Long-Range Dispersion Correction. J. Comput. Chem. 2006, 27, 1787−1799. (46) Tang, W.; Sanville, E.; Henkelman, G. A Grid-Based Bader Analysis Algorithm without Lattice Bias. J. Phys.: Condens. Matter 2009, 21, 084204. (47) Sanville, E.; Kenny, S. D.; Smith, R.; Henkelman, G. An Improved Grid-Based Algorithm for Bader Charge Allocation. J. Comput. Chem. 2007, 28, 899−908. (48) Henkelman, G.; Arnaldsson, A.; Jónsson, H. A Fast and Robust Algorithm for Bader Decomposition of Charge Density. Comput. Mater. Sci. 2006, 36, 354−360. (49) Yu, M.; Trinkle, D. R. Accurate and Efficient Algorithm for Bader Charge Integration. J. Chem. Phys. 2011, 134, 064111. (50) Henkelman, G.; Uberuaga, B. P.; Jónsson, H. A Climbing Image Nudged Elastic Band Method for Finding Saddle Points and Minimum Energy Paths. J. Chem. Phys. 2000, 113, 9901−9904. (51) Henkelman, G.; Jónsson, H. Improved Tangent Estimate in the Nudged Elastic Band Method for Finding Minimum Energy Paths and Saddle Points. J. Chem. Phys. 2000, 113, 9978−9985.

20584

DOI: 10.1021/acsami.7b04170 ACS Appl. Mater. Interfaces 2017, 9, 20577−20584