Simultaneous Oxygen and Boron Trifluoride Functionalization of Hexagonal Boron Nitride: A Designer Cathode Material for Energy Storage

Simultaneous Oxygen and Boron Trifluoride Functionalization of Hexagonal Boron Nitride: A Designer Cathode Material for Energy Storage

Károly Németh Physics Department, Illinois Institute of Technology, Chicago, Illinois 60616, USA

Covalent functionalization is a way to tune the electrochemical properties of hexagonal boron nitride (h-BN) monolayers. The wide band gap insulator h-BN may become metallic conductor upon functionalization with strong oxidants, such as fluorosulfonyl radicals (OSOF), as known since 1978 [N. Bartlett et al., J. Chem. Soc. Chem. Comm. 5, 200 (1978)], with electrical conductivity of 1.5 S/cm [C. Shen et al., J. Solid State Chem. 147, 74 (1999)] that greatly surpasses commercial cathode material LiCoO while retaining excellent ionic conductivity. Functionalized boron nitrides (FBN-s) have great potential for cathode applications in energy storage devices, for example in solid state batteries. While fluorosulfonyl functionalization is unlikely to result in rechargeable cathodes, similarly to graphene fluoride (CF), some other FBN-s discussed here may do. In the present work, fluorene, oxygen and combined oxygen and boron trifluoride functionalizations are studied, on the basis of band structure calculations. Due to the open surfaces of FBN-s, fast ionic diffusion with Li, Na and Mg ions is possible, enabling batteries with voltages of 2.1-5.6 V, theoretical energy densities of 800-1200 Wh/kg and fast charge and discharge.

hexagonal boron nitride and functionalization and band gap engineering and solid state battery and supercapacitor and pseudocapacitor

I Introduction

Ideal electrochemical energy storage devices would simultaneously fulfill several desirable properties, such as high gravimetric and volumetric energy and power densities, safety of operation, economic and environmentally friendly composition, rechargeability, fast charging and discharging, and a large number of charging/discharging cycles (long cycle life) Armand and Tarascon (2008); Simon and Gogotsi (2008).

While most traditional electroactive materials consist of three-dimensional (3D) structures, such as crystals or amorphous materials, the recent emergence of 2D materials offers new and advantageous platforms for novel ways of energy storage Geim and Novoselov (2007); Wang et al. (2016); Pakdel et al. (2014a); Weng et al. (2016).

The recent realization of hybrid supercapacitor-batteries through the application of graphene oxide (GO) cathodes and Li or Na anodes demonstrated simultanous achievement of high energy and exceptionally high power densities Liu et al. (2014); Kim et al. (2014). Power densities of 4-45 kW/kg (on average) and specific energies of 100-500 Wh/kg have been achieved and stable capacity (300-450 mAh/g) for over 1000 cycles of charge and discharge demonstrated, with Li anode and GO cathode, where the GO cathode had as much as 21-32 w% oxygen content Liu et al. (2014). With Na anode and GO cathode, energy density was at 100-500 Wh/kg, power density at 55 kWh/kg and stable capacity demonstrated through 300 cycles Kim et al. (2014). The respective cell reactions during discharge involve the settling of one negative charge on the oxygen atoms of GO during the ring opening of surface-bound epoxy bonds and the conversion of edge-bound C=O (oxo) groups to phenolate-type ones Kim et al. (2014).

In functionalized monolayer 2D materials (F2D-s), cations can hop directly from the electrolyte to the reactive/intercalating surface sites of F2D-s, allowing for very high power densities (charge/discharge rates), similar to those in supercapacitors. This is not possible in 3D materials in which diffusion of ions is normally much slower. The advantage of 2D materials in designing high power density energy storage devices appears to be unparalleled among 3D materials for this simple geometric consideration.

There are many layered materials, e.g., MoS, WS, MoSe, WSe, and TiS, that can be exfoliated to monolayers. However, in contrast to graphene (G) and h-BN, these transition metal containing monolayers cannot offer high specific energy and power densities because of the presence of heavy transition metal elements (per weight properties become disadvantageous). Furthermore, the transition metals involved are often rare, expensive and environmentally unfriendly. Therefore, only C, N and B based 2D monolayers appear to be practical for industrial-scale energy storage technologies among the 2D materials.

Unfortunately, graphene oxide is thermally unstable, it easily and often explosively decomposes to CO, CO and carbon when heated to 200-300 C Kim et al. (2010); Krishnan et al. (2012). This is one of the reasons why the scalable production of GO is still in its infancy. Other than oxygen functionalization of graphene, such as -COOH, -OH, -NH, -OR, and –COOR ones, have also been considered for cathode applications, however not tested yet Liu et al. (2014). Another problem with graphene is that it turns from an electrically conductive material to an insulator when densely functionalized (such as in highly oxidized graphene). As opposed to graphene, h-BN may become a small or zero bandgap system when densely functionalized Bhattacharya et al. (2012) also illustrated by the BN(F)-BN(OBF) example and its discharged (lithiated, sodiated, etc) versions below. The most drastic band gap reduction has been observed in fluorinated or fluorosulfonated h-BN experimentally, resulting in good electrical conductivity Xue et al. (2013); Du et al. (2014); Bartlett et al. (1978); Shen et al. (1999). Thus, certain functionalized h-BN-s may simultaneously be good electric and ionic conductors in both charged (lithiated, cathionated) and discharged (delithiated, de-cathionated) states.

As opposed to GO, oxidized h-BN is stable even abovove 800 C Li et al. (2014); Cui et al. (2014). Both edge and surface functionalizations of h-BN have been carried out with various functional groups using radical species, such as OSOF (fluorosulfonyl, as early as in 1978 Bartlett et al. (1978); Shen et al. (1999)), OH Fedorov et al. (2013); Nazarov et al. (2012); Sainsbury et al. (2012); Lin and Connell (2012); Pakdel et al. (2014b); Nayak et al. (2014), NH Fedorov et al. (2013); Nazarov et al. (2012); Ikuno et al. (2007); Lin and Connell (2012) , CBr Sainsbury et al. (2014), F Xue et al. (2013); Du et al. (2014), and substituted phenyl Rajendran (2012) functional groups. Radical species preferencially bind to the electron-deficient B atoms, except double radicals (such as O, CBr) which bind to both B and N simultaneously Sainsbury et al. (2014); Anota et al. (2011). Covalent functionalization of h-BN has also been carried out using ionic species, such as OH Lee et al. (2015); Fu et al. (2016); Zhang et al. (2016), NO or SO which bind to positively charged B and negatively charged N atoms of h-BN Fedorov et al. (2013) and may be delivered in solution or melt phases. Furthermore, it has been known for decades from the study of refractory material h-BN, that the melt of certain salts, such as LiOH, KOH, NaOH-NaCO, NaNO and LiN etches h-BN Kumashiro (2000), whereby functionalized h-BN species form Yamane et al. (1987). The surface coverage of h-BN by -OH and -NH functional groups may be as high as about 30 atom% as shown by XPS analysis Nazarov et al. (2012). The reactions between functionalizing radicals or ions and the h-BN surface are based on Lewis acid/base reactions, utilizing the Lewis acid character of the B and Lewis base character of N atoms in h-BN that can react with strong Lewis bases/acids, respectively. A recent review by Bando, Golberg and colleagues discusses the advances in functionalization of h-BN in further details Weng et al. (2016).

The high thermal stability of oxygen functionalized h-BN and the great variety of h-BN functionalization options makes functionalized h-BN-s an attractive platform to design electroactive species for efficient batteries, as proposed recently by the author of the present work Nemeth (2015).

i.1 Charge storage mechanisms in functionalized h-BN

Charge can be stored in several ways in FBN-s:

1.) Charge storage in the bond attaching a radical to the surface of h-BN. The B atom of the h-BN monolayer is electron deficient, its valence shell can store up to two additional electrons. This becomes clear when one considers the specific one resonance structure of h-BN in which there are only single () bonds between the N and B atoms and the lone pair of N is fully localized on the N p orbital. In this resonance structure, the B valence shell has only six electrons while the N has eight. When a radical functionalizes the B atom, such as OH or NH, it donates one electron to the empty B p orbital which then transforms to an approximate sp hybrid orbital participating in a bond binding to the radical. This bond, however, still needs an additional electron to complete itself. The additional electron may come from the anode during the discharge of the device. Note that in reality the lack of this latter electron is distributed over all the local bonding environment of the B atom, creating holes (of electrons) in them and increasing the conductivity of the FBN as compared to h-BN. This hole creation effect is greatest if the functionalizing atom has a big electronegativity, such as fluorene, and indeed this is also experimentally observed by the good electrical conductivity of fluorinated h-BN Xue et al. (2013); Du et al. (2014). Such charge storage mechanism is justified by the existence of both charge neutral OH radical functionalized h-BN and OH anion functionalized h-BN where the large negative charge of the functionalized h-BN has also been pointed out, see Ref. Fu et al., 2016.

2.) Charge storage in the functionalizing group. It is also possible to store charge in other parts of the functional group, not only in the bond that links the functional group to h-BN. This charge storage may be based on the (reversible) reduction/oxidation of chemical bonds in the functional group, such as for example epoxy ring opening or C=O to C-O reduction during discharge. This charge storage mechanism may involve bond breaking and bond formation. A mixture of the two kinds of charge storage may also simultaneously occur as illustrated by the example of the BN(F)-BN(OBF) system discussed in the following.

Ii Results and Discussion

Perhaps the most fundamental FBN material is oxygen functionalized h-BN, the analogue of GO. Early studies focused on the high temperature oxidation processes of h-BN and its composites to BO and boric acide as well as boron glasses Jacobson et al. (1999a, b). Only recent studies attempted the production of oxygen functionalized nanoplatelets and mono/few layers of h-BN, using high temperature (800-1000 C heating on air in a sealed quartz tube Li et al. (2014); Cui et al. (2014)). When moisture is present during the high temperature heating, quick transformation to BO and boric acide happens Jacobson et al. (1999a, b). Exposing h-BN to air-plasma results in hydroxyl-functionalized h-BN, instead of oxygen functionalized one, due to the presence of water in the air Pakdel et al. (2014b). Theoretical studies of ozone absorption on h-BN also indicate the possibility of epoxy group formation on its surface Anota et al. (2011). Long exposure of h-BN to heating in air at a moderately high temperature of 600 C results in surface coverage by oxygen species with additional boric oxide formation Jin et al. (2016).

Another recent study investigated the fluoroboration reaction of GO with the ether adduct of BF Samanta et al. (2013). In this reaction the epoxy rings on the surface of GO open up forming a C-F and a C-OBF unit on the C atoms of the original epoxy unit. Such a functionalization of the surface is advantageous for two reasons: 1. Fluorene functionalization may lead to high voltage cells and 2. the -OBF unit may capture the F released during discharge forming -OBF: a mechanism to make carbon-fluoride batteries rechargeable as suggested in Ref. Jones and Hossain, 2011. Unfortunately, the C-OBF unit is not stable, and -OBF is generally a good leaving group in the presence of Lewis bases, such as thiols and amines Samanta et al. (2013). As opposed to the C-OBF bond on the surface of functionalized graphene, the B-OBF on the surface of functionalized h-BN may be much more stable owing to the electron-deficiency of boron in h-BN. Also, the B-O single-bond energy is much greater (536 kJ/mol) than that of C-O (358 kJ/mol) or N-O (201) Cottrell (1958); Darwent (1970). Furthermore, the existence of singly negatively charged tetravalent borate ions, such as the tetrahydroxyborate anion or the related polyborates, points toward the stability of negatively charged sp hybridized boron in FBN-s. The quantum-chemical calculations discussed below support this expectation. The respective functionalization process of h-BN to BN(F)-BN(OBF) is depicted in Fig. 1. The following sections will describe the band structure evolution of h-BN upon step(layer)-wise functionalization, first by oxygen, then by BF, followed by cell reactions of the product with Li, Na and Mg anodes, predicting the voltages and energy densities of these cells. Furthermore, the applicability of these materials as high voltage pseudocapacitors will also be discussed.

ii.1 Computational methodology

DFT calculations have been carried out using the Quantum Espresso program package Giannozzi et al. (2009, 2017), following the methodology in Refs. Németh, 2014a, b; Zhang et al., 2016. A plane wave basis set with 50 Ry wave-function cut-off has been used along the PBEsol exchange-correlation functional Perdew et al. (1996a, 2008) with ultrasoft pseudopotentials as provided by the software package. Functionalized h-BN monolayers were separated from each other by about 30 Å vacuum layers and 3D periodic boundary conditions were applied. Each simulation cell involved a double stoichiometric formula unit of the respective compound in order to allow for uniform functionalization on both the top and bottom surfaces of h-BN. Supercells of the simulation cells are shown in Fig. 2. Spinpolarized calculations have been carried out to allow for open shell systems characteristic for radicals-based functionlization of h-BN. The k-space grid was 10x10x4, this allows for a mRy convergence of the electronic energy with respect of k-space saturation and geometry convergence. This accuracy is satisfactory for the prediction of accurate cell voltages within the chosen exchange-correlation functional and surface model. The surface models have been relaxed until residual forces on the atoms became smaller than 0.0001 Ry/bohr and residual pressure on the simulation cell was less than 3 kbar. Energies of cell reactions were calculated from electronic energy differences of products and starting materials, all in the crystal phase. Cell voltages are computed as the negative of the cell energies divided by the number of electrons transferred in the reaction. The methodology has been validated on experimental data of lattice parameters and enthalpies of formation of -LiBN, LiN, and h-BN. Since the model systems have the same translational symmetry as h-BN, only with larger unit cells, the representation of the electronic band structure used the same high symmetry k-points. Enthalpies of formations were estimated as the change of electronic energy during the formation of the compounds from the corresponding elements at T = 0 K, i.e., from crystalline Li and B and N gas. Experimental lattice parameters have been reproduced within 2.5% error, while experimental enthalpies of formation were within 4% Németh (2014b). The coordinates of the high symmetry points in the k-space are (0,0,0), M(1/2,0,0) and K(2/3,1/3,0).

While the PBEsol functional typically underestimates band gaps, it usually provides good structural parameters and relative energies, such as in the above validation cases. Hybrid functionals, such as PBE0 Perdew et al. (1996b); Adamo and Barone (1999) and HSE Heyd et al. (2003) provide more accurate band gaps than PBEsol, as pointed out in band structure studies on h-BN as well Zolyomi and Kürti (2015), however, they are computationally much more expensive and not fully available in Quantum Espresso yet. Even if the PBEsol band gaps would underestimate the real ones by about 1 eV, it would likely not change the relative order of them, i.e. PBEsol is assumed to be accurate enough to separate large, medium and small band gap systems to estimate what can be expected for the electrical conductivity of the given systems. PBEsol band gap values can also be compared to known band gaps of similar FBN-s, such as oxidized, hydroxilated, fluorinated or fluorosulfonated ones, to check whether the magnitude of the band gaps have been qualitatively well described. For example, highly electron withdrawing fluorene or fluorosulfonyl radicals are experimentally known to create strong p doping that results in small bandgap FBN-s with strong magnetism. The fluorosulfonyl-analogous -OBF3 functionalization is expected to result in small band gap as well and PBEsol is capable to verify or at least support this expectation. Therefore, the accuracy of PBEsol is sufficient to provide reliable guidance for the selection of target cathode materials for experimental testing, which is the main goal of the current work.

ii.2 Band structure evolution due to functionalization

Fig. 1 depicts the process of the stepwise functionalization of h-BN. In the first step, oxygen radicals attach to neighboring B and N atoms, forming epoxy rings on the top or bottom surfaces of the h-BN monolayer. In the second step, these epoxy rings react with BF in an analogous fashion as formerly seen for GO Samanta et al. (2013). As a result, pairs of nearest B atoms become covalently functionalized with -F and -OBF units. Fig. 2 shows the evolution of a h-BN sheet during the functionalization in panels a-c, assuming all boron atoms of the sheet will be functionalized at the end leading to BN(F)-BN(OBF). Panels d-f indicate the discharge products of a BN(F)-BN(OBF) cathode with Li, Na and Mg anodes. As indicated by the bond lengths in Table 2, both the charged and discharged systems consist of strong covalent bonds withing the FBN and the cations intercalate the functional groups of the surface. Some of the B-N single bonds become as long as 1.59-1.62 Å due to the functionalization, however these values are well within observations in related B-N bond containing systems. For comparison, the B-N bond in h-BN is 1.45 Å long, in cubic-BN 1.57 Å (sphalerite type) and 1.55-1.58 Å (wurtzite type), in ammonia-borane 1.58 Å and in HN-BF or MeN-BF about 1.66 Å Jonas and Frenking (1994).

The band structures of the materials discussed in the present work are depicted in 3. The respective band gap and magnetization values are presented in Table 1.

Pristine (non-functionalized) h-BN has an indirect band gap of about 4.4 eV and a direct one of 4.7 eV (panel a of Fig. 3), about 1 eV smaller than the experimental band gap of 5.6 eV.

Oxidation of h-BN reduces the band gap to about 3.3 and 2.7 eV in (BN)O and BNO, respectively (panels b and g of Fig 3). Note that in (BN)O and BNO only non-spinpolarized solutions were found. The band-gap of hydroxilated h-BN has experimentally been found to be 3.9 eV Nayak et al. (2014). Doping of h-BN with O (substitution of N by O) may lead to an as low as 2.1 eV band gap, according to experimental observation Weng et al. (2017).

Spin-polarized calculations result in 0.8 eV gap for BN(F)-BN(OBF) and zero gap for its alkalinated versions (panels c, d and e in Fig 3, respectively). The zero bandgaps of the latter systems appear close to the -point and the highest energy occupied band crosses the Fermi energy only to a small extent, therefore also the next smallest gaps are given in Table 1, and these are of 0.6 eV.

Note that the band gaps in BN(F)-BN(OBF) and its Li and Na reduced versions are quite small, much smaller than that of some semiconducting polymers, such as for example polyacetylene, in which the band gap is about 1.4 eV. In the most common Li-ion battery cathode material, LiCoO, the band gap is 2.7 eV Van Elp et al. (1991) and it reduces to zero as the cathode is charged and its Li content per formula unit decreases below 0.75 Molenda et al. (1989). Also note that the electrical conductivity of LiCoO is in the order of 0.0001 S/cm and can be increased to 0.5 S/cm by Mg doping Park et al. (2010); Tukamoto and West (1997). Fluorosulfonyl functionalized h-BN has a higher electrical conductivity of 1.5 S/cm Shen et al. (1999).

Calculations on the Mg reduced form of BN(F)-BN(OBF) resulted in only non-spinpolarized solution with a band gap of about 3.4 eV (panel f of Fig. 3).

The fluorene functionalized h-BN hasa direct zero band gap, characteristic of metals (panel h of Fig 3). With the above in mind, one possible way to introduce electrically conducting domains in h-BN may be based on fluorination. Also note that the good electrical conductivity of some fluorinated and fluorosulfonated h-BN has been experimentally observed Xue et al. (2013); Shen et al. (1999).

Another way to increase electrical conductivity in BN(F)-BN(OBF) is n-type doping that happens when Li, Na or Mg reduces BN(F)-BN(OBF) during the discharge of the battery. While the fully discharged systems may have small bandgaps, as a result of filling up the lowest energy conduction states, partial filling of these conduction states may result in better electric conductivity. Also note that the band structures in the respective systems substantially change during the discharge of the battery as one B-F bond breaks and another forms, therefore a simple band-filling approach is insufficient to understand how the band gap changes as a function of the extent of reduction (discharge). One formula unit of BN(F)-BN(OBF) can in principle take up a maximum of two electrons, as two electrons are missing to complete the valence shells of the two B-s that are being functionalized (either by F or by OBF radicals) per formula unit of BN(F)-BN(OBF). Thus, in BN(F)-BN(OBF)Li and BN(F)-BN(OBF)Na half of the conduction band , i.e. half of the holes in the nitrogen valence shells, remains unfilled, which results in zero bandgaps and leads to improved electrical conductivity of these systems as compared to the fully discharged BN(F)-BN(OBF)Mg. Note that full discharge is also possible with two Li or Na per formula unit, however in the present model the solvation shell for the second alkali cation could not be efficiently represented in the framework of a monolayer system, as it should probably be represented via the interlayer interactions that are not discussed here. Also note that small bandgap is in principle also possible with Mg anode when only a partial discharge is applied.

Even though BN(F)-BN(OBF) and its (partially) reduced versions may have small bandgaps, addition of small amounts of graphene may make these systems electrically conductive to the desired degree, as it is customary with most battery cathode materials, while such graphene composites would also be ionically conductive without the addition of liquid electrolytes, based on the ability of Li and Na ions to hop between the surface sites of BN(F)-BN(OBF). Thus these sytems are good candidates for cathodes in all-solid-state batteries that are highly desirable as especially safe and high energy density batteries.

Ionic conductivity is assumed to be good in all FBN-s due to the strong dipoles in the B-N bonds and in the functional groups on the surface, as well as sufficient space on the surface of the monolayers for virtually all kinds of ions to move. In fact FBN-s have been proposed and experimentally tested to some extent for use as ionic conductors Mofakhami and Fauvarque (2015); Kumar et al. (2010).

Magnetization values support that the combined oxygen and BF functionalization creates holes in all four nitrogen atoms of the simulation cell of BN(F)-BN(OBF) (two formula units used per simulation cell for top and bottom surface functionalizations), as significant magnetization values of 0.30-0.45 have been calculated only for the N-sites, while magnetization values on other atoms are at least an order of magnitude smaller. After the reaction with Li or Na, a reduction and rearrangement to BN-BN(OBF) ions happens. In these latter anions some N-s will have two B neighbors with OBF functionalization and one non-functionalized B, while some other N atoms will have two non-functionalized B neigbors and one OBF functionalized one. This is also reflected in the magnetization changes: the N-s with two functionalized B neighbors remain significantly magnetized,, while the other N-s essentially lose their magnetization. In the case of Mg anodes, two negative charge will be transferred to a formula unit of the cathode and doubly negative BN-BN(OBF) ions form. This results in a total filling of all holes in the N-s, a complete loss of magnetization and a large band gap of 3.4 eV.

A similar analysis holds to BNF, where all high magnetizations of about 0.45 happen on the N atoms, while only a tenth of this value occurs on the F-s. Fluorene or combined oxygen and BF functionalization represents a p-type doping of h-BN and is a powerful means to engineer conductivity and electrode reactivity / rechargeablity of functionalized h-BN. The holes created by these p-type dopings will be filled when the respective batteries discharge and the discharge of the battery can be viewed as n-type doping by alkaline or alkaline earth atoms. Thus h-BN provides a platform that is suitable for p-type doping and recombination of electrons and holes through a subsequent n-type doping that follows during the discharge of the battery.

Also note that total magnetization values (sum of up and down magnetizations) are near zero in all magnetic systems mentioned above. This implies an antiferromagnetic ordering of the individual atomic magnetizations.

Therefore, the conclusion is that it is the presence of holes in the N atoms of BN(F)-BN(OBF) and its Li, Na and Mg doped (reduced) versions that is responsible for small band-gaps and magnetizations. The electrical conductivity of these systems is a function of hole-concentration: too high hole concentration in BN(F)-BN(OBF) is not optimal but partially n-type doped (reduced) BN(F)-BN(OBF) may lead to zero bandgap and good electrical conductivity.

ii.3 Electrochemical features

During the discharge process, the triangle of the epoxy bond of BNO opens up in the N-O bond while it stays intact in the B-O bond and converts to a singly negatively charged BNO ion that is approximately tetrahedrally shaped and is linked to the rest of the 2D sheet through its N-corners. During discharge, the BN(F)-BN(OBF) cathode material detaches a F ion from its BN(F) unit and captures it on a nearby OBF unit according to the intramolecular surface conversion process of


thereby forming a BF-like ion, OBF that is covalently bound to the h-BN surface through its O-atom. The resulting discharged 2D monolayer acts as a 2D array of linked anions, a 2D polyanion. In fact, such a charged 2D polyanion has been observed in OH functionalized h-BN and its large negative charge (zeta-potential) has been measured Fu et al. (2016). The discharge product negative ions are charge-balanced by nearby cations, such as depicted in Figs. 2 d-f. Note that this mechanism enables the making of a rechargeble battery from a fluorinated 2D FBN material (BN(F)-BN(OBF)), as opposed to the well known and commercially available fluorinated graphite cathode material, CF, that is not rechargeable due to irreversible formation of fluoride salts. Also note that a similar mechanism to make CF rechargeable was proposed in Ref. Jones and Hossain, 2011, though it was based on an inter-, instead of an intra-molecular F transfer process. Simple fluorination of h-BN that leads to BNF would be insufficient to mantain rechargeablity for the same reason CF is not rechargeable. It is only the more sophisticated BN(F)-BN(OBF) functionalization that allows for rechargeability.

As summarized in Table 3, the batteries based on BNO are of relatively small open circuit voltage ( 2 V), though this is approximately the voltage of intensely researched Li-sulfur batteries, while the ones based on BN(F)-BN(OBF) are of high voltage (3.6-5.6 V) as these latter ones represent reactions between bouound fluorene radicals and Li, Na or Mg. Both types of batteries have high theoretical gravimetric energy densities and capacities.

ii.4 Pseudocapacitive applications

Two fundamental features of pseudocapacitors involve 1. a “sloping” electrode potential as a function of charge withdrawn or added to the electrode (classically associated with capacitors, not with batteries where the potential should be approximately constant during discharge) and 2. chemical (redox) reactions happening during the charging/discharging process on the surface or in the bulk of the electrode (classically associated with batteries, not with capacitors) Conway (1991); Dubal et al. (2015); Conway (1999).

2D functionalized h-BN-s can, at least to some extent, contribute to pseudocapacitive effects. As it is seen from the above described examples of BNO and BN(F)-BN(OBF), chemical reactions may happen on the surface of these materials, such as epoxy ring opening/closing or fluorene reduction/oxidation accompanied with fluoride ion capture in the OBF side-chain. Since the negative charges will be localized in the monolayer surface, all positive ions should also fit in the surfaces, such as indicated in Fig. 2, in order to have a battery. Positive ions, however, may not always all fit in the surface, for several reasons. The primary such reason is the size of these ions. If their solvation shells, or the ions themselves are large enough, they cannot be packed as dense on the surface of the monolayers as the negative ions and therefore the positive ions must form multiple surface layers resulting in a sloping electrostatic potential. To maintain such a potential, the negative ion side-chains must be bound by strong covalent bonds in the surface. It is expected that pseudocapacitive behaviour of FBN-s increases with the increasing size of the cations used. For example, large organic cations, such as quaternary ammonium ions are expected to provide large pseudocapacitance with FBN-s.

The use of large organic cations may allow for a rechargeable use of BNF (h-BN fluorinated on all B atoms) as a cathode material in supercapacitors and pseudocapacitors. With Li, Na or Mg cations, BNF would decompose to the respective fluoride salts and h-BN, similarly to CF, as the close vicinity of these small cations would polarize the B-F bonds too much, the cations would rip the F ions off from the h-BN surface and form fluoride salts, such as in the BN(F) unit in BN(F)-BN(OBF) in reaction Eq. 1. While the OBF side-chains capture the detached F ions in BN(F)-BN(OBF), there is no such option in BNF. It is however reasonable to assume that the F ions would not be detached from the h-BN surface with sufficiently large organic cations, both for the smaller polarizing effect of these cations and for the thermodynamically less energetic formation of the corresponding organic fluorene salts (or ionic liquids) as compared to Li/Na/Mg-fluorides. While BNF would act as the electron withdrawing component of an organic charge transfer complex, no such behaviour is known for CF. The reason is that BNF has holes in electronic orbitals of high electronegativity N atoms, generated by electron withdrawing effect of nearby B-F bonds (more precisely B-bound F radicals), while CF has no such holes.

While most pseudocapacitors are known to operate at small maximum voltages (typically 1 V), some FBN-s may allow for pseudocapacitors with unusually large voltages, even as high as about 3 V, when using cathodes like BN(F)-BN(OBF), owing to the large energy of fluorene reduction/oxidation in the FBN electrode. As the amount of energy stored in a pseudocapacitor is linearly proportional to its capacitance and quadratically to its voltage, the applications of FBN pseudocapacitors may open the way of storing more energy per weight than FBN batteries.

Iii Summary and Conclusions

The present work proposes new functionalized 2D materials,derivatives of exfoliated h-BN as positive electrode electroactive materials for use in improved electrochemical energy storage devices. The choice of functionalization can tune the electrochemical properties of these materials. Three such materials have been analyzed in detail: BNO, BNF and BN(F)-BN(OBF).

These functionalizations represent p-type doping of h-BN that creates holes in the nitrogen valence shells of h-BN. These holes recombine with electrons during the discharge of the respective batteries whereby the discharge represents a subsequent, and in the case of BNO and BN(F)-BN(OBF), reversible n-type doping through the reaction with Li, Na or Mg atoms.

BNO and BN(F)-BN(OBF) go through reversible surface conversion reactions during the cycling of the electrochemical devices and have potential for simultaneously achieving high energy and power densities. BNO, BNF and BN(F)-BN(OBF) can potentially also be used in pseudocapacitors due to the dense packing of covalently bound negatively charged functional groups in the monolayer surfaces. The careful choice of doping levels may lead to good electrical conductivity of these systems as well. Since such functionalized h-BN-s are also good ionic conductors, they could form a basis for lightweight all-solid-state batteries with high energy densities.


The author thanks Prof. Leon L. Shaw (IIT) for helpful discussions and NERSC (U.S. DOE DE-AC02-05CH11231) for the use of computational resources.


  • Armand and Tarascon (2008) M. Armand and J.-M. Tarascon, Nature 451, 652 (2008).
  • Simon and Gogotsi (2008) P. Simon and Y. Gogotsi, Nature materials 7, 845 (2008).
  • Geim and Novoselov (2007) A. K. Geim and K. S. Novoselov, Nature materials 6, 183 (2007).
  • Wang et al. (2016) X. Wang, Q. Weng, Y. Yang, Y. Bando, and D. Golberg, Chemical Society Reviews (2016).
  • Pakdel et al. (2014a) A. Pakdel, Y. Bando, and D. Golberg, Chemical Society Reviews 43, 934 (2014a).
  • Weng et al. (2016) Q. Weng, X. Wang, X. Wang, Y. Bando, and D. Golberg, Chemical Society Reviews (2016).
  • Liu et al. (2014) C. Liu, A. Zhamu, D. Neff, and B. Jang, Lithium super-battery with a functionalized nano graphene cathode (2014), uS Patent 8,795,899.
  • Kim et al. (2014) H. Kim, Y.-U. Park, K.-Y. Park, H.-D. Lim, J. Hong, and K. Kang, Nano Energy 4, 97 (2014).
  • Kim et al. (2010) F. Kim, J. Luo, R. Cruz-Silva, L. J. Cote, K. Sohn, and J. Huang, Advanced Functional Materials 20, 2867 (2010).
  • Krishnan et al. (2012) D. Krishnan, F. Kim, J. Luo, R. Cruz-Silva, L. J. Cote, H. D. Jang, and J. Huang, Nano today 7, 137 (2012).
  • Bhattacharya et al. (2012) A. Bhattacharya, S. Bhattacharya, and G. Das, Physical Review B 85, 035415 (2012).
  • Xue et al. (2013) Y. Xue, Q. Liu, G. He, K. Xu, L. Jiang, X. Hu, and J. Hu, Nanoscale research letters 8, 49 (2013).
  • Du et al. (2014) M. Du, X. Li, A. Wang, Y. Wu, X. Hao, and M. Zhao, Angewandte Chemie 126, 3719 (2014).
  • Bartlett et al. (1978) N. Bartlett, R. Biagioni, B. McQuillan, A. Robertson, and A. Thompson, Journal of the Chemical Society, Chemical Communications pp. 200–201 (1978).
  • Shen et al. (1999) C. Shen, S. G. Mayorga, R. Biagioni, C. Piskoti, M. Ishigami, A. Zettl, and N. Bartlett, Journal of Solid State Chemistry 147, 74 (1999).
  • Li et al. (2014) L. H. Li, J. Cervenka, K. Watanabe, T. Taniguchi, and Y. Chen, ACS nano 8, 1457 (2014).
  • Cui et al. (2014) Z. Cui, A. J. Oyer, A. J. Glover, H. C. Schniepp, and D. H. Adamson, Small 10, 2352 (2014).
  • Fedorov et al. (2013) V. Fedorov, A. Nazarov, and V. Demin, Method of producing soluble hexagonal boron nitride (2013), russian Patent RU 2478077 C2.
  • Nazarov et al. (2012) A. S. Nazarov, V. N. Demin, E. D. Grayfer, A. I. Bulavchenko, A. T. Arymbaeva, H.-J. Shin, J.-Y. Choi, and V. E. Fedorov, Chemistry–An Asian Journal 7, 554 (2012).
  • Sainsbury et al. (2012) T. Sainsbury, A. Satti, P. May, Z. Wang, I. McGovern, Y. K. Gun’ko, and J. Coleman, Journal of the American Chemical Society 134, 18758 (2012).
  • Lin and Connell (2012) Y. Lin and J. Connell, Method for exfoliation of hexagonal boron nitride (2012), uS Patent 8,303,922.
  • Pakdel et al. (2014b) A. Pakdel, Y. Bando, and D. Golberg, ACS nano 8, 10631 (2014b).
  • Nayak et al. (2014) A. P. Nayak, A. Dolocan, J. Lee, H.-Y. Chang, T. Pandhi, M. Holt, L. Tao, and D. Akinwande, Nano 9, 1450002 (2014).
  • Ikuno et al. (2007) T. Ikuno, T. Sainsbury, D. Okawa, J. Fréchet, and A. Zettl, Solid State Communications 142, 643 (2007).
  • Sainsbury et al. (2014) T. Sainsbury, A. O’Neill, M. K. Passarelli, M. Seraffon, D. Gohil, S. Gnaniah, S. J. Spencer, A. Rae, and J. N. Coleman, Chemistry of Materials 26, 7039 (2014).
  • Rajendran (2012) G. Rajendran, Surface modified hexagonal boron nitride particles (2012), uS Patent 8,258,346.
  • Anota et al. (2011) E. C. Anota, H. H. Cocoletzi, and E. R. Rosas, The European Physical Journal D 63, 271 (2011).
  • Lee et al. (2015) D. Lee, B. Lee, K. H. Park, H. J. Ryu, S. Jeon, and S. H. Hong, Nano letters 15, 1238 (2015).
  • Fu et al. (2016) L. Fu, G. Lai, G. Chen, C.-T. Lin, and A. Yu , ChemistrySelect 1, 1799 (2016), ISSN 2365-6549.
  • Zhang et al. (2016) F. Zhang, K. Németh, J. Bareño, F. Dogan, I. D. Bloom, and L. L. Shaw, RSC Advances 6, 27901 (2016).
  • Kumashiro (2000) Y. Kumashiro, Electric refractory materials (CRC Press, 2000).
  • Yamane et al. (1987) H. Yamane, S. Kikkawa, and M. Koizumi, Journal of Solid State Chemistry 71, 1 (1987).
  • Nemeth (2015) K. Nemeth, Functionalized boron nitride materials as electroactive species in electrochemical energy storage devices (2015), wO Patent App. PCT/US2014/045,402.
  • Jacobson et al. (1999a) N. Jacobson, S. Farmer, A. Moore, and H. Sayir, Journal of the American Ceramic Society 82, 393 (1999a).
  • Jacobson et al. (1999b) N. S. Jacobson, G. N. Morscher, D. R. Bryant, and R. E. Tressler, Journal of the American Ceramic Society 82, 1473 (1999b).
  • Jin et al. (2016) H. Jin, Y. Li, X. Li, Z. Shi, H. Xia, Z. Xu, and G. Qiao, Materials Letters 175, 244 (2016).
  • Samanta et al. (2013) K. Samanta, S. Some, Y. Kim, Y. Yoon, M. Min, S. M. Lee, Y. Park, and H. Lee, Chemical Communications 49, 8991 (2013).
  • Jones and Hossain (2011) S. Jones and S. Hossain, Polymer materials as binder for a cfx cathode (2011), uS Patent App. 13/010,431.
  • Cottrell (1958) T. Cottrell, The Strengths of Chemical Bonds, Butterworths scientific publications (Butterworths Scientific Publications, 1958).
  • Darwent (1970) B. Darwent, Bond dissociation energies in simple molecules, NSRDS-NBS (U.S. National Bureau of Standards; for sale by the Supt. of Docs., U.S. Govt. Print. Off., 1970).
  • Giannozzi et al. (2009) P. Giannozzi, S. Baroni, N. Bonini, M. Calandra, R. Car, C. Cavazzoni, D. Ceresoli, G. L. Chiarotti, M. Cococcioni, I. Dabo, et al., Journal of Physics: Condensed Matter 21, 395502 (19pp) (2009), URL
  • Giannozzi et al. (2017) P. Giannozzi, O. Andreussi, T. Brumme, O. Bunau, M. B. Nardelli, M. Calandra, R. Car, C. Cavazzoni, D. Ceresoli, M. Cococcioni, et al., Journal of Physics: Condensed Matter 29, 465901 (2017), URL
  • Németh (2014a) K. Németh, International Journal of Quantum Chemistry 114, 1031 (2014a).
  • Németh (2014b) K. Németh, The Journal of chemical physics 141, 054711 (2014b).
  • Perdew et al. (1996a) J. P. Perdew, K. Burke, and M. Ernzerhof, Phys. Rev. Lett. 77, 3865 (1996a).
  • Perdew et al. (2008) J. P. Perdew, A. Ruzsinszky, G. I. Csonka, O. A. Vydrov, G. E. Scuseria, L. A. Constantin, X. Zhou, and K. Burke, Phys. Rev. Lett. 100, 136406 (2008).
  • Perdew et al. (1996b) J. P. Perdew, M. Ernzerhof, and K. Burke, The Journal of chemical physics 105, 9982 (1996b).
  • Adamo and Barone (1999) C. Adamo and V. Barone, The Journal of chemical physics 110, 6158 (1999).
  • Heyd et al. (2003) J. Heyd, G. E. Scuseria, and M. Ernzerhof, The Journal of chemical physics 118, 8207 (2003).
  • Zolyomi and Kürti (2015) V. Zolyomi and J. Kürti, Physical Review B 92, 035150 (2015).
  • Jonas and Frenking (1994) V. Jonas and G. Frenking, Journal of the Chemical Society, Chemical Communications pp. 1489–1490 (1994).
  • Weng et al. (2017) Q. Weng, D. G. Kvashnin, X. Wang, O. Cretu, Y. Yang, M. Zhou, C. Zhang, D.-M. Tang, P. B. Sorokin, Y. Bando, et al., Advanced materials 29 (2017).
  • Van Elp et al. (1991) J. Van Elp, J. Wieland, H. Eskes, P. Kuiper, G. Sawatzky, F. De Groot, and T. Turner, Physical Review B 44, 6090 (1991).
  • Molenda et al. (1989) J. Molenda, A. Stokłosa, and T. Ba̧k, Solid State Ionics 36, 53 (1989).
  • Park et al. (2010) M. Park, X. Zhang, M. Chung, G. B. Less, and A. M. Sastry, Journal of Power Sources 195, 7904 (2010).
  • Tukamoto and West (1997) H. Tukamoto and A. West, Journal of The Electrochemical Society 144, 3164 (1997).
  • Mofakhami and Fauvarque (2015) A. Mofakhami and J. Fauvarque, Material for an electrochemical device (2015), uS Patent 9,105,907.
  • Kumar et al. (2010) B. Kumar, J. Kumar, R. Leese, J. P. Fellner, S. J. Rodrigues, and K. Abraham, Journal of The Electrochemical Society 157, A50 (2010).
  • Conway (1991) B. E. Conway, Journal of the Electrochemical Society 138, 1539 (1991).
  • Dubal et al. (2015) D. P. Dubal, O. Ayyad, V. Ruiz, and P. Gomez-Romero, Chemical Society Reviews 44, 1777 (2015).
  • Conway (1999) B. Conway, Electrochemical Supercapacitors: Scientific Fundamentals and Technological Applications (Springer US, 1999), ISBN 9780306457364.
material E (eV) magnetization ()
total absolute
h-BN 4.4 (K) / 4.7 (KK) 0.0 0.0
(BN)O 3.3 0.0 0.0
BN(F)-BN(OBF) 0.8 0.13 3.28
BN-BN(OBF)Li 0.0 (0.6) 0.01 1.53
BN-BN(OBF)Na 0.0 (0.6) 0.02 1.50
BN-BN(OBF)Mg 3.4 0.0 0.0
BNO 2.7 0.0 0.0
BNF 0.0 0.0 1.74
Table 1: Calculated band gaps (E), total and absolute magnetizations for functionalized h-BN-s. For BN-BN(OBF)Li and BN-BN(OBF)Na the band gaps are zero close to the -point where the highest energy occupied band crosses the Fermi-energy to a small extent, for more information also the next band gap value is presented in parenthesis.
Cathode material B-N B-O B-F
charged Li,Na,(Mg) charged Li,Na,(Mg) charged Li,Na,(Mg)
BNO 1.57-1.60 1.55-1.62 1.42 1.44-1.45 - -
BN(F)-BN(OBF) 1.56-1.59 1.48-1.59 1.45 1.43-1.46 1.32, 1.42 1.32-1.48
BNF 1.55-1.57 - - - 1.38 -
Table 2: Calculated characteristic bond lengths (Å) at the optimum geometries in the cathode materials BNO and BN(F)-BN(OBF), as well as those in BNF, for comparison. Note that BNF is not stable as a cathode material with Li, Na and Mg anodes, as the discharge would result in the respective fluorene salts. The anode materials were Li, Na and Mg, the charged systems’s anode is referred to as ”charged”. Mg anode has been used only with BN(F)-BN(OBF). For comparison, the B-N bond in h-BN it is 1.45 Å long, in ammonia-borane 1.58 Å and in MeN-BF3 about 1.66 Å Jonas and Frenking (1994). Also note that the B-F bonds longer than 1.38 Å are due to the vicinity of a cation.
Cathode material OCV (eV) GED (Wh/kg) GC (mAh/g)
Li Na Mg Li Na Mg Li Na Mg
BNO 2.15 2.10 - 1209 868 - 561 413 -
BN(F)-BN(OBF) 5.60 5.10 3.6 1067 874 1223 191 171 340
Table 3: Calculated open circuit voltages (OCV), gravimetric energy (GED) and capacity (GC) densities with two functionalized h-BN cathode materials: 1. fully epoxy-functionalized h-BN, BNO, and 2. fluoroborated half-epoxy-functionalized h-BN, BN(F)-BN(OBF). The anode materials are bulk metallic Li, Na and Mg. The corresponding cell reactions involve one electrons for Li and Na and two for Mg per formula unit of the cathode material.
Figure 1: Proposed synthetic steps toward BN(F)-BN(OBF) on a two-BN-unit fragment of h-BN: the epoxy-functionalization of half of the B=N bonds is followed by the reaction with strong Lewis acid BF. Green indicates the charge of the B and N atoms of the h-BN monolayer in the given resonance structures, “” (“dot” and “plus”) indicates half of the lone pair in N, i.e. a hole localized in the N lone pair on the N p orbital. The B=N double bonds in the given resonance structure are due to the donation of the lone pair of N toward B in a -bond. The charge distribution reflects this resonance structure, in reality the N charge is (much) more negative, the B charge is more positive due to the much greater electronegativity of N as compared to B. Red indicates the atoms and bonds introduced by functionalization. The oxygen radical in the first step may come from various sources, such as ozone, for example. The second step of the process is analogous to that recently seen in the GO BF reaction Samanta et al. (2013).
Figure 2: Fragments (supercells) of examples of functionalized h-BN. Color code: B - magenta, N - blue, O - red, F - green (small), Li - violet (small), Na - violet (large), Mg - green (large). Systems shown: h-BN, (BN)O, BN(F)-BN(OBF), BN-BN(OBF)Li, BN-BN(OBF)Na, BN-BN(OBF)Mg, BNO and BNF, in panels a through h, respectively. The order indicates step(layer)-wise funtionalization of a h-BN monolayer, first by epoxide groups on every 2 B-N bonds, then the epoxyde groups are reacted with BF forming BN(F)-BN(OBF) units, then discharge products of the latter with Li, Na and Mg cations and BN-BN(OBF) anionic units (n=1 or 2), finally the fully epoxy-functionalized (BNO) and the fully (on all B atoms) fluorene functionalized h-BN (BNF) is shown for comparison.
Figure 3: Spinpolarized calculated band structures of h-BN and its functionalized derivatives. Different colors indicate up and down spin bands. Systems are identical to those shown in Fig. 2 and in the same respective order: h-BN, (BN)O, BN(F)-BN(OBF), BN-BN(OBF)Li, BN-BN(OBF)Na, BN-BN(OBF)Mg, BNO and BNF in panels a through h. The origin of the energy scale is the Fermi energy of the respective system.
Comments 0
Request Comment
You are adding the first comment!
How to quickly get a good reply:
  • Give credit where it’s due by listing out the positive aspects of a paper before getting into which changes should be made.
  • Be specific in your critique, and provide supporting evidence with appropriate references to substantiate general statements.
  • Your comment should inspire ideas to flow and help the author improves the paper.

The better we are at sharing our knowledge with each other, the faster we move forward.
The feedback must be of minimum 40 characters and the title a minimum of 5 characters
Add comment
Loading ...
This is a comment super asjknd jkasnjk adsnkj
The feedback must be of minumum 40 characters
The feedback must be of minumum 40 characters

You are asking your first question!
How to quickly get a good answer:
  • Keep your question short and to the point
  • Check for grammar or spelling errors.
  • Phrase it like a question
Test description