Towards Integration of Two-Dimensional Hexagonal Boron Nitride (2D h-BN) in Energy Conversion and Storage Devices

: The prominence of two-dimensional hexagonal boron nitride (2D h-BN) nanomaterials in the energy industry has recently grown rapidly due to their broad applications in newly developed energy systems. This was necessitated as a response to the demand for mechanically and chemically stable platforms with superior thermal conductivity for incorporation in next-generation energy devices. Conventionally, the electrical insulation and surface inertness of 2D h-BN limited their large integration in the energy industry. However, progress on surface modiﬁcation, doping, tailoring the edge chemistry, and hybridization with other nanomaterials paved the way to go beyond those conventional characteristics. The current application range, from various energy conversion methods (e.g., thermoelectrics) to energy storage (e.g., batteries), demonstrates the versatility of 2D h-BN nanomaterials for the future energy industry. In this review, the most recent research breakthroughs on 2D h-BN nanomaterials used in energy-based applications are discussed, and future opportunities and challenges are assessed.


Introduction
The ever-growing importance of green energy production/storage units to reduce the carbon footprint of fossil fuels has motivated researchers to explore new materials to enhance the efficiency of conventional energy technologies [1,2]. Two-dimensional (2D) materials derived from layered structures have shown to be promising candidates for the future generation of energy devices. Their properties such as ultra-high specific surface area, low absorption energy, high thermochemical stability, and biocompatibility have opened a window for highly efficient energy conversion/storage units [3]. Pioneering 2D materials in this regard include graphene, graphene oxide and transition metal dichalcogenides (TMDs). Other recently emerged 2D morphologies such as transition metal carbides, nitrides or carbonitrides (MXene) [4], boron carbon nitride (B x C y N z ) [5] and borophene [6] have also been of great interest due to their tunable surface chemistry and stoichiometric-dependent physical/chemical characteristics.
2D hexagonal boron nitride (h-BN) has also recently become a highly demanded energy materials functioning in many applications, from electrodes and electrolytes in batteries to membranes or catalysts in fuel cell units [7,8]. A recent wave of research on 2D h-BN-based energy applications, shown in Figure 1a,b, originates from the synergistic impacts of high thermal and chemical stabilities, superior thermal conductivity, and the electrically insulative nature of 2D h-BN [8][9][10]. The term h-BN refers to a layered nanomorphology of boron (B) and nitrogen (N) atoms in a sp 2 hybridized honeycomb-shaped lattice [11,12]. The van der Waals interactions maintain the vertical stacking of the layers, resembling a graphitic structure [13], while the in-plane B-N bonds are covalent (σ) with partial ionicity due to electronegativity difference between the B and N (Figure 1c) [14]. The B-N bond strength offers excellent mechanical robustness, but on the other hand, the lack of pi (π) electrons on each plane minimizes the surface interactions caused by π-π forces [15]. Therefore, h-BN shows superior chemical stability when compared to other layered materials such as graphite or transition metal dichalcogenides (TMDs), where π-π conjugations dominate the surface interaction mechanisms [16]. Nevertheless, the partially charged sites on the surface (B positive, N negative) can be anchoring sites for Lewis acid-base interactions [17,18]. In fact, the existing empty orbital in B and a lone pair in N offers a promising pathway for selective functionalization of 2D h-BN nanomaterials [19]. The functionalization of 2D h-BN allows the modulation of surface chemistry to acquire the desired properties while the intrinsic characteristics are preserved [20,21]. For instance, hydroxylation of 2D h-BN enhances its water dispersibility, fluorination intensifies its photoluminescent emission, edge amination enhances its surface acid-base interactions, and point defects amplify its catalytic activities. These examples reveal that surface modification is the key to harnessing the full capability of 2D h-BN nanomaterials [22,23].  This review covers the recent advancements in the application of 2D h-BN nanomaterials in various energy sectors with a focus on interfacial interactions that improve the performance of 2D h-BN in different energy storage/conversion devices. To this end, we demonstrate the importance of surface modification for harnessing the full applicability of 2D h-BN. Electrochemical energy production/storage systems including fuel cells, batteries, and supercapacitors, together with thermoelectric and photovoltaic applications of 2D h-BN will be discussed, followed by examining the future trends, obstacles, solutions, and possibilities in the integration of 2D h-BN in energy applications.  [56][57][58][59]. It is known that the major drawbacks of conventional anodic electrodes, such as lack of safety at higher power densities or inadequate theoretical capacitance (i.e., 372 mAh.g −1 for graphite) can be addressed by the incorporation of materials with high swift charge/discharge rates (e.g., 2D h-BN) into the battery [60,61]. To harness the full capability of 2D h-BN, it is often hybridized with other materials. As an example, hybridization of only 2 wt.% 2D h-BN with reduced graphene oxide (r-GO) has demonstrated a high reversible capacity of 278 mAh g −1 , a current density of 100 mA g −1 with improved cycling stability up to 200 cycles [62]. In fact, the synergistic effect of the 2D h-BN/r-GO interface generated by van der Waals forces minimizes the interlayer volume expansion caused by Li-ion intercalation on the electrodes. Similarly, the enhanced polarization of the anode electrode of lithium titanium oxide/r-GO hybridized with 2D h-BN demonstrated almost two times lower charge transfer resistance compared to when 2D h-BN was not present. In 2D h-BN/silicene heterostructures, the formation of intermediate complexes at interface decreases the bandgap (to 1.85 eV), Li + binding energy, and Li + diffusion barrier [62]. With these advantages, the maximum capacity of the unit can reach 1015 mAh g −1 . Besides the bare 2D h-BN, functionalized structures of 2D h-BN with various oxygen-based functional groups have also shown reversible charge/discharge surface reactions with Li-ions, revealing specific capacity up to 78 mAh g −1 with good Coulombic efficiency for an electrode (Figure 2a) [43,63].

Application of 2D h-BN
Electrode separation is another major application of 2D h-BN in LIBs. The presence of a separator, a physical barrier between the electrodes, is vital to prevent short circuits and control ion migration during charge-discharge cycles [65,67]. The short circuit of a LIB often derives from uncontrolled battery overheating. Then, a thermally stable separator can improve the operability and decrease safety concerns [68]. The common separators such as polypropylene (PP) and polyethylene (PE) suffer from low wettability against liquid electrolytes and uneven Li + diffusion flux [69,70]. On the contrary, 2D h-BN with local surface polarities exhibits a uniform flux distribution of Li + as a separator, and the ionic conductivity of the electrolyte was shown to improve (from 0.160 to 0.255 mS m −1 ) [69].
More importantly, 2D h-BN suppresses the formation of Li-dendrites. The Li-dendrites formed by uncontrolled growth of Li on the anodes can penetrate through the separator and cause short circuits in LIB. As seen in Figure 2b, the SEM images of the PP separator's morphology demonstrates large pore sizes on the surface, leading to higher impedance resistance. The surface morphology of the PP separator after 100 cycles at the current density of 5 mA cm −2 also demonstrates the extreme distribution by Li-dendrites and gaps ( Figure 2c). The SEM images of 2D h-BN modified PP separator before and after the same charge-discharge cycles (shown in Figure 2d,e, respectively) show that the addition of 2D h-BN suppressed the growth of dendritic morphology and presented a smoother surface [64]. Further research on improvement of commonly used separators with h-BN nanomaterials such as boron nitride nano tube (BNNTs)/PP [71], PE/PP/PE modified 2D h-BN [67], PE-BN/poly (vinylidene fluoride-hexafluoropropylene (PVDF-HFP) [67] established 2D h-BN as a vital component for separators' proper operation. The polymer-free blade coated 2D h-BN was also employed as the separator in Li-manganese oxide/graphene/2D h-BN/Li cells, exhibiting a capacity retention of 82% after 100 cycles (Figure 2f) [65]. In fact, the proper deposition and growth of the 2D h-BN layer resulted in an excellent separation performance as a sole matrix.
As discussed, 2D h-BN improves the electrolyte performance in two ways: (1) enhancing the mechanical stability and fabrication robustness, and (2) increasing the ionic conductivity and Li transference number (t Li + ). Generally, to increase the mechanical stability of electrolytes, the gel polymer electrolytes (GPE) or gel ionic electrolyte (GIE) are used instead of liquid electrolytes and reinforced by inorganic fillers such as TiO 2 or Al 2 O 3 [72][73][74][75]. However, these fillers act as a barrier against ion transport and affect the overall cell capacity. In contrast, 2D h-BN provides both mechanical strength and ionic conductivity (upon functionalization) at the same time and can potentially address this issue [1]. It was demonstrated that the addition of only 0.5 wt% perfluoropolyether (PFPE) modified 2D h-BN into P(VdF-co-HFP) increases the t Li + from 0.41 to 0.63 while the electrolyte shear modulus (G) is enhanced five times [73]. Similarly, the addition of bulky h-BN into LiTFSI/1,3-dioxolane (DOL)-1,2-dimethoxymethane (DME) [76][77][78], amino-functionalized 2D h-BN into trimethylolpropane triacrylate (TMPTA) with PVDF-HFP [79], and 2D h-BN/polyacrylonitrile (PAN) into LiPF in EC/ethyl methyl carbonate (EMC) [80] demonstrated exceptional increase in both t Li + and G. Moreover, the thermal stability of the GPE and GIE is enhanced by incorporation of 2D h-BN. Accordingly, the addition of 2D h-BN and bentonite into LiTFSI/1-methyl-1-propyl piperidinium bis (trifluoro methyl sulfonyl) imide (PP13) is shown to maintain a capacity of 90 mAhg −1 at 120 • C for almost 20 days (Figure 2g) (more than 600 cycles) while 2D h-BN-free electrolyte malfunctions above 60 • C [66].

Mg Metal-Based Batteries (MMBs)
MMBs batteries with an electrode potential of −2.37 V propose a high theoretical capacity of 2250 mAhg −1 [81]. Along with the electrochemical properties, the abundance of Mg in nature makes it a competitive candidate for use in the secondary batteries market. One serious obstacle in way of commercializing these batteries is the insertion/extraction of Mg into the cathode due to the highly polarized nature of Mg 2+ ions [82]. Fluorine functionalized 2D h-BN nanomaterials were reported to be well-functioning cathodes in MMBs [83]. The cyclic voltammetry (CV) graphs shown in Figure 3a indicate that Mg 2+ ions are successfully intercalated on the edge sites of the 2D h-BN, resulting in an anodic peak at 1.75 V. The de-intercalation also occurs with a cathodic peak at 0.8 V while pristine 2D h-BN does not show any redox reaction. The calculated current density from the CV is 48 mAg −1 , and the discharge capacity after 80 cycles is 50 mAhg −1 . In contrast, the undoped 2D h-BN does not show a good electrochemical behavior ( Figure 3b). As seen in electrochemical impedance studies (EIS) (Figure 3c), the semi-circles in high and medium frequency regions are attributed to the solid electrolyte interphase (SEI) and charge transfer resistance, respectively. The linear trend at the low-frequency region displays the Warburg impedance correlated with ion diffusion resistance. The calculated R SEI (SEI resistance) of undoped and doped 2D BN is 128 Ω and 94 Ω, respectively. Similarly, R ct (charge transfer resistance) illustrates a considerable drop of 34% from 209 Ω for undoped BN cathode to 137 Ω for doped 2D h-BN. Therefore, it can be concluded that the doping of 2D h-BN with fluorine enhances the conductivity of the electrode. The footprint of this improved behavior in the doped 2D h-BN cathode is also observed in Figure 3d [83]. As seen, by increasing the charge/discharge cycles, the discharge capacity of doped 2D h-BN is raised continuously and reaches a maximum value of 50 mAhg −1 at 80 cycles. In another study, a graphene/2D h-BN composite was used as the cathode material and the nominal composition of graphene /20 wt.% 2D h-BN revealed the best performance with a current density of 25 mAg −1 and discharge capacity of 105 mAhg −1 (75 mAg −1 and 94 mAhg −1 for bare graphene cathode, respectively) up to 80 cycles [81]. The excessive increment of 2D h-BN content deteriorates the performance of the battery due to its insulative characteristics. The implementation of a few-layer MoS 2 cathode, compared to 2D h-BN, demonstrated better capacity (170 mAhg −1 ); however, the electrode stability was limited to only 50 cycles [84]. Although the electrochemical performance of the Mg-ion batteries is improved by addition of 2D h-BN, the mechanisms by which these improvements occur are not yet well understood.

Zinc Flow Batteries (ZFBs)
ZFBs are often considered to be reliable energy storage systems for stationary applications owing to their high energy density, cost-effectivity, and safe operation [85]. Despite these advantages, the scaled-up production and commercialization of these batteries are still lacking. One of the limiting factors is the zinc dendrite/accumulation, deteriorating the cycle-life of the batteries. This process is greatly influenced by the inhomogeneous temperature distribution on the electrode as well as the poor mechanical characteristics of the membrane [86]. Thus far, many efforts have been devoted to obviating these shortcomings, using 2D h-BN. The incorporation of 2D h-BN into ZFBs is made through three major pathways: (1) modification of the electrolyte chemistry, (2) reinforcing membranes, and (3) changing the battery configuration [87,88].
Recently, Hu and coworkers exploited a porous polyether sulfone membrane coated with 2D h-BN in an alkaline zinc-iron flow battery [89]. Based on the obtained results, 2D h-BN acts as a heat transporter when facing the anode, modifies the morphology of zinc, and reinforces the membrane. According to the voltage-time diagrams, the bare membrane was stable up to 40 h (Figure 4a), followed by a fluctuation due to the formation of zinc needle-like dendrite penetrating through the membrane and micro-short circuiting. This dendrite formation may also damage the membrane, leading to the crossover of the electrolyte. As a result, self-discharge and declining concentration of the active materials occur. The sharp rise of voltage in the final step pertains to this concentration polarization, since the potential of the Fe (CN) 6 3− /Fe (CN) 6 4− depends on the concentration of the electrolyte. Accordingly, the discharge capacity decreases considerably. In contrast, the 2D h-BN coated membrane demonstrates an outstanding performance with voltage stability after 500 cycles (800 h) (Figure 4a). The proposed mechanism is such that 2D h-BN helps the generated heat distribute uniformly in favor of the facile plating/stripping of zinc on the electrode. In addition, 2D h-BN strengthens the membrane against the damage induced by the non-uniform deposition of zinc when the battery is charging. Supreme Coulombic efficiency (CE) of 98.5% and energy efficiency (EE) of 87.6% after 500 cycles support the proposed mechanisms ( Figure 4b). The 2D h-BN-based battery has reached the highest ever reported efficiency of 80% after 200 cycles among all ZFBs ( Figure 4c) [90,91]. Inset: the corresponding discharge capacity and discharge energy for each cycle [89] (Reprinted with permission from WILEY-VCH, Copyright © 2020).

Other
In recent years, the abundance of Na and K resources and their lower costs as compared to Li have drawn attention to Na and K ion batteries. The storage mechanism in these batteries is to some extent similar to that of LIBs [43]. However, under the same weight, the volume of Na or K ions is considerably larger than that of Li and therefore, the energy density is much lower [43]. These drawbacks have considerably suppressed the extensive application of Na and K ion batteries. Similar to what was discussed in the Section 2.1.1, 2D h-BN was also incorporated in sodium ion batteries' (NIBs) design. For instance, 2D h-BN was demonstrated to be an effective capping agent for black-phosphorene (Pn) as an anode material of NIBs [92]. By h-BN modification, the anode performance was improved significantly due to enhanced binding energy of Na in the h-BN/black-Pn heterostructure (2.55 eV) compared to pristine Pn (1.59 eV). Moreover, the Na-ion barrier movement within the layers was reduced by the incorporation of 2D h-BN. Apart from that, the silicene/2D h-BN heterostructure was theoretically shown to be an ultrafast ion diffusion anode material due to the enhanced binding strength of Na on 2D h-BN [92]. Focusing on the wide bandgap of 2D h-BN and the possibility of band gap modification after the formation of vacancies and absorption of Na + and K + ions, monolayer h-BN can be a great candidate for an anodic material [93]. Accordingly, the storage capacities, open-circuit voltages, and energy densities of 571.698 mAh/g, 0.0099 V and 5.656 Wh/ kg can be obtained for h-BN reinforced NIBs [94,95]. A comparison of battery performance improvement by incorporating 2D h-BN is shown in Table 2.

Supercapacitors
Supercapacitors have recently been in high demand for energy storage due to their peculiar characteristics of high-power density, swift charge/discharge capability, and better safety records and a long-life cycle in comparison to batteries [106,109]. Over the last decade, 2D h-BN nanostructures (i.e., functionalized and/or doped ones) and their heterostructures have been introduced as promising inorganic materials to improve the performance of supercapacitors [35,[110][111][112]. The 2D h-BN modifications can generate available free electrons or enhance the mobility of the charge carriers in order to improve the electrochemical behavior [113][114][115][116]. Moreover, the partial polarity of the B-N bonds facilitates the charge separation within the 2D h-BN lattice. In fact, 2D h-BN demonstrates a pseudocapacitive behavior owing to this partial ionic character of the B-N bonds. In addition, the oxidation state of B and N atoms can be changed through intercalation with the electrolyte during the redox reaction, contributing to additional charge storage. Combining electrochemical double-layer capacitance (EDLC) with such pseudocapacitive materials amplifies the energy density and the specific capacitance. Furthermore, functionalization of 2D h-BN can introduce additional acceptor and donor electronic states depending on the functional group, resulting in more efficient energy storage.

Electrodes
Electrodes are the most crucial parts of a supercapacitor assembly. Therefore, most of the research on 2D h-BN-based compounds has been dedicated to the improvement of the electrodes' performance. As one of the earliest attempts, Gao et al. developed a 2D h-BN/r-GO nanocomposite electrode with the capacitance of 140 F g −1 , almost 75% higher than that of a bare r-GO electrode [102]. The capacitance of 2D h-BN/r-GO increased by 5.5% after 1000 cycles, indicating the high stability of 2D h-BN/r-GO, and attributed to the hindrance of r-GO sheets agglomeration. In a similar study, various 2D h-BN/r-GO architectures were prepared and their electrochemical behavior in different stoichiometric ratios was evaluated [104]. By increasing the GO proportion from 0.1 g (named h-BNG1) to 0.4 g (named h-BNG4) in the primary batch containing 0.3 g boric acid and 0.3 g urea, a superlattice architecture was obtained (Figure 5a,b). Additionally, the electrochemical behavior of the composite electrode changed from Faradic to EDLC. The formed superlattice structure led to an incremental increase in electrical conductivity through carrier confinement within the graphene pathways embedded in the h-BN domains and increasing π conjugations. The electronic work function of the composite was also reduced from 296 (in h-BNG1) to 22 meV (in h-BNG3). Cyclic voltammetry tests, shown in Figure 5c, revealed that h-BNG1 possesses a low specific capacitance of 218 F g −1 at a scan rate of 20 mV s −1 (could be attributed to 2D h-BN as the dominant phase). The potential window of h-BNG2 is located at both negative and positive areas (−0.5-0.5 V), and the curves show distorted features, indicating that the storage mechanism was neither EDLC nor pseudocapacitive ( Figure 5d). This behavior could be due to insufficient intercalation between r-GO and 2D h-BN nanosheets since the generation of the active sites depends on the intercalation process. The highest specific capacitance was recorded for h-BNG3 with 960 F g −1 at a scan rate of 10 mV s −1 , pertaining to the synergistic effect of the EDLC and pseudo capacitance as well as the presence of more active sites on the superlattice structure of the electrode (Figure 5e).   High stability (98%) even after 10,000 cycles. 2D h-BN has also been employed in a quaternary composite of ZnCo 2 S 4 (ZCS)/2D h-BN/CNT/Polypyrrole (PPY) as a high-performance electrode (Figure 5f,g) [103]. Considering the poor cycling stability of bare ZCS, the combination of ZCS with other pseudocapacitive materials including 2D h-BN and PPY will overcome this problem. This often occurs through stabilization of the system with the assistance of π-π interactions and Coulombic attractions, leading to a remarkable increment in the active sites for electrochemical interactions. This complex geometry exhibited high specific capacitance of 534 and 785 F g −1 in aqueous (1 M KCl) and organic (1 M TEABF 4 in acetonitrile) media, respectively.
Surface functionalization is an alternative approach for the incorporation of 2D h-BN into supercapacitor electrode designs. However, due to surface charges, 2D h-BN sheets tend to agglomerate; therefore, hybridization with other nanostructures or conducting polymers is required. Recently, hydroxyl (-OH) functionalized 2D h-BN/polyaniline (PANI)/CNT nanohybrid was employed as an electrode [107]. It was shown that 2D h-BN stabilizes PANI in the system via a Coulombic interaction. Then it reduces the agglomeration tendency of CNTs through charge surface blocking mechanism and minimization of surface interactions. Moreover, the surface area was enhanced from 50.49 m 2 /g to 89.77 m 2 /g due to the homogeneous deposition of PANI on CNTs and the diffusion of fiber-like PANI decorated CNTs into the 2D h-BN.
This nanohybrid electrode showed a specific capacitance of 515 F g −1 and cyclic stability of 98% after 10,000 charge/discharge cycles. The proposed mechanism for electron transfer is illustrated in Figure 5h. Under an applied potential, the PANI's LUMO interacts with the conduction band of 2D h-BN, and leads to a reduction in the LUMO energy of PANI and the conduction band of h-BN. In addition, an electron is transferred from the PANI's HOMO to LUMO and then the conduction band of h-BN. The presence of oxygen functional groups on the 2D h-BN surface is shown to facilitate this interaction owing to the additional acceptor and donor levels of the electronic states [22]. A higher specific capacitance of 787 F g −1 in a 1 M KCl solution is recorded for the Zn-CdS/2D h-BN/CNT electrode compound in which the crystallization of flower-like nanostructures by oxygen functionalized 2D h-BN and CNTs resulted in this outstanding performance (Figure 6a) [105]. Brunauer-Emmett-Teller (BET) measurements also showed that introducing 2D h-BN leads to more than 60% enhancement in specific surface area (from 151 m 2 /g to 246 m 2 /g).

Electrolyte
Improvement of electrolyte performance in both mechanical and thermal aspects can be achieved by the incorporation of 2D h-BN [108]. It has been reported that by adding only 0.025 mg ml −1 h-BN the ionic conductivity of PVA-H 2 SO 4 GPE increases from 9 to 29 mS cm −1 [117]. Since the charge carriers in PVA-H 2 SO 4 (H + ) transport by hopping mechanism between hydrogen bonds, an optimal amount of 2D h-BN launched swift ion transportation through its functional groups. An excessive amount of 2D h-BN caused restacking of the nanosheets and obstructed the ion movements (Figure 6b-d) [117]. As seen in the EIS graphs in Figure 6e, the presence of a small semi-circle at the high-frequency region and the slope at the low-frequency end represent an ideal EDLC behavior [117]. It should be noted that the first intersection of the curve with Z demonstrates the bulk resistance of the supercapacitor, the diameter of the semi-circle represents the charge transfer resistance of the electrode, and the slope corresponds to the diffusion resistance in the supercapacitor. Accordingly, it can be inferred that supercapacitors containing 2D h-BN have lower charge transfer resistance due to the emergence of the ionic highways. This high ionic conductivity and mechanical stability of 2D h-BN drew the researchers' attention towards the applicability of 2D h-BN as a separator of a solid-state planar asymmetric supercapacitor [118].

Thermoelectrics
Considering the increasing importance of green energy production, thermoelectric materials converting heat to electricity have become an area of great interest [119][120][121]. To characterize the thermoelectric conversion efficiency via the material-dependent figure of merit, Equation (1) is used: where σ, S, T, K e , and K ph represent the electrical conductivity, Seebeck coefficient, absolute temperature, and electronic and phononic thermal conductivity, respectively [122,123]. Therefore, a large Seebeck coefficient and electrical conductivity, and a small thermal conductivity are in favor of high-performance thermoelectric devices. However, the interdependence of the Seebeck coefficient, and electrical and thermal conductivity are serious challenges in thermoelectric materials research, as changing one parameter could adversely affect the other two in any given compound. According to Equation (1), the electrical insulation and high phonon conductivity of h-BN suppress the ZT value [124][125][126]. Nevertheless, introducing defects to the h-BN structure and hybridization with other 2D materials could enhance its thermoelectric properties [127]. In a theoretical study, the figure of merit of zigzag and armchair h-BN nanoribbons (ZBNNRs and ABNNRs) were compared with that of zigzag and armchair graphene nanoribbons (ZGNRs and AGNRs) [128]. The diagrams of phonon thermal conductance with respect to the absolute temperature of the 2D h-BN system are shown in Figure 7a. As seen, when T is almost 0 K, the thermal conductance approaches 3K 0 , attributing to contributions from three acoustic phonon modes. By raising the temperature, the thermal conductance is enhanced due to the activation of higher modes. In wider ribbons (Figure 7b), the enhancement of the thermal conductance becomes more apparent owing to higher accessible channels for phonon transport [128]. Calculations in Figure 7e,f suggest that ABNNRs demonstrate the best thermoelectric performance of all, with a maximum figure of merit of about 0.2. Moreover, the value of ZT can be tailored by introducing lattice defects in the ribbons, whose positions highly affect the ZT, as depicted in Figure 8a,b.
Although the thermoelectric properties of h-BN nanomaterials are improved by introducing structural defects, their values are still not high enough to be used in real devices. One leading approach to address this issue is to assemble h-BN heterostructures with graphene and other 2D materials [130,131]. With very high electrical conductivity, graphene could be a good option for thermoelectric applications; however, its high thermal conductivity needs to be tailored to obtain an acceptable ZT value [132]. Although graphene and h-BN represent high in-plane thermal conductance, their cross-plane conductance is low, deemed to be an advantage for thermoelectric devices. To explore the thermoelectric transport properties at the interface of graphene and h-BN [133], Raman spectroscopy is often used to apply a temperature gradient between the top and bottom graphene layers. Considering the zero lateral temperature gradients, the diagram of thermoelectric voltage as a function of the cross-plane temperature gradient shows a linear trend, resulting in a Seebeck coefficient of −99.3 µV/K. The obtained result is nearly two times higher than that of single-layer graphene at room temperature (50 µV/K) [134]. The calculated power factor is~1.51 × 10 −15 W/K 2 , offering a quite low value due to the insulating character-istics of h-BN. Indeed, a thick enough h-BN layer would diminish quantum mechanical tunneling, whereas both electron tunneling and thermionic emission across the energy barrier contribute to the cross-plane Seebeck coefficient [135,136]. Finally, a low value of 1.05 × 10 −6 was calculated as the figure of merit of the device, which is due to the thick h-BN layer, and consequently, the large energy barrier at the interface of graphene/h-BN (0.5 eV). By tailoring the energy barrier height and the thickness of h-BN, higher ZT values can be achieved [137,138]. A remarkable enhancement of ZT up to 0.7 was also predicted theoretically in superlattice armchair ribbons consisting of graphene and h-BN sections [139]. However, the insulating behavior of h-BN causes a high electron scattering and shifts the maximum ZT values to higher chemical potentials. Further improvement of phonon scattering in these heterostructures requires complex simulations to understand the transport phenomena at graphene/2D h-BN interfaces [129]. For instance, Figure 8c,g demonstrates h-BN flakes periodically attached to armchair graphene nanoribbons. In this case, phonon thermal conductance is diminished while electronic conductance is almost conserved. Moreover, introducing 2D h-BN causes bandgap broadening, leading to an increment of the Seebeck coefficient. In this situation, the ZT value will increase from 0.6 for a short heterostructure system (n BN = 1) to 0.81 for longer systems (n BN > 5). Interestingly, these high ZT values are predicted for small and achievable chemical energy levels of around 0.4 eV. In addition, introducing vacancies in the central region of graphene layers may increase the ZT up to 1.48 at room temperature, mainly due to extra suppression of phonon conductance [129].  Doping the G/h-BN heterostructure with chemical species can also impact the ZT, as witnessed in the case of tetrathiafulvalene (TTF) and tetracyanoethylene (TCNE) doped G/2D h-BN heterostructures. TCNE, as an electron acceptor, causes the Fermi level to shift between the valence band of the ribbon and its LUMO. In contrast, TTF as an electron donor leads to observing the Fermi level between the conduction band of the ribbon and HOMO of TTF (Figure 9a) [140]. The Fermi level shift triggers a notable increment of the thermopower and electronic figure of merit. Moreover, the presence of dopants in the structure leads to extra phonon scattering. Accordingly, ZT of 0.1 and 0.9 were predicted for TCNE and TTF doped graphene/2D h-BN.
Recently, it was shown that heterostructures of 2D h-BN and 2D GeP 3 can be promising thermoelectric materials [141]. Pristine GeP 3 depicts supreme charge carrier mobility and good chemical stability; however, its small bandgap degrades the thermoelectric characteristics. It is noteworthy to mention that the higher bandgap values for GeP 3 can be obtained in mono and bilayer structures due to the dominance of quantum confinement; however, no reports on their experimental synthesis have been published so far. One approach to enlarge the band gap of GeP 3 is capping 2D h-BN layer as an insulating material, preserving the in-plane electrical conductance of GeP 3 . As illustrated in Figure 9b-e, GeP 3 is an indirect bandgap material with E g of 0.23 and 0.12 eV for monolayer (1L) and bilayer (2L) GeP 3 , respectively [141]. These values increase to 0.34 eV when a monolayer 2D h-BN is inserted. However, the number of 2D h-BN layers is not of paramount importance regarding the band structure. Evaluations unveiled that p-type GeP 3 /BN 1_1 (i.e., monolayer GeP 3 and monolayer 2D h-BN) with a doping level of 0.004 e/u.c, could have a maximum ZT of 5.13 at room temperature (Figure 9f). In absence of experiential research, thersitical studies suggest that controlled integration of 2D h-BN into thermoelectric materials and devices can open pathways for explorations in this field in the net future [136,137].

Fuel Cells 2.4.1. Membrane
Direct methanol fuel cell (DMFC), as an essential member of the fuel cells family, faces some technical problems, e.g., passing of methanol through the polymer electrolyte membrane (PEM), slow methanol oxidation and oxygen reduction reactions. The methanol crossover reduces the efficiency of DMFCs by the creation of a mixed potential at the cathode. The undesired oxidation, decreasing the active sites for reduction of oxygen at the cathode, and carbon monoxide poisoning of the cathode catalyst are other methanol crossover challenges [142][143][144][145]. Nafion, as the most common PEM used in DMFCs, suf-fers from the methanol crossover despite its supreme proton conductivity and outstanding chemical, thermal, and mechanical stability [146][147][148]. One attempt to decrease the methanol crossover in Nafion is to use 2D h-BN as reinforcement. Polarized covalent bonds of B and N lead to valence electron accumulation around the N atom and create an uneven distribution of electron cloud, resulting in high proton conductivity. This high proton conductivity along with its large surface area, low density, thermomechanical stability, barrier effect, and excellent chemical resistance, make 2D h-BN a great candidate as a PEM filler employed in DMFCs. For this purpose, 2D h-BN can be functionalized via -OH groups to escalate the side-chain reactions with -mercaptopropyl trimethoxysilane to form continuous ionic channels in the PEM. The proton conductivity of bare Nafion improved from 135 to 214 mS.cm −1 after adding 0.75wt.%BN. The mechanism is explained through the formation of-SO 3 H groups in the functionalized 2D h-BN. At the same time, the methanol crossover is reduced by the addition of 2D h-BN in the membrane [149]. Considering the ratio of the proton conductivity and methanol crossover current density, the electrochemical selectivity of Nafion/0.75wt.%BN can be as high as 6.63 × 10 −2 S/mA compared to 1.81 × 10 −2 S/mA in pristine Nafion. This composite membrane exhibits a high peak power density of 165 mW/cm 2 at 70 • C, which is much higher than that of pristine Nafion (65 mW/cm 2 ) [149].
Regarding the high cost of the Nafion, many researchers actively examine other alternatives, including poly(2, 6-dimethyl-1,4-phenylene oxide) (PPO) owing to its high T g , as well as mechanical and hydrolytic stability [150]. The transport behavior and remarkable hydrophobic/hydrophilic phase separation of Sulphonated PPO (sPPO) renders outstanding electrochemical performance [151]. However, since a higher degree of sulfonation might reduce its stability, incorporating different additives would be beneficial [152]. Covalent functionalization of h-BN with the hydrophilic acidic group makes them well-dispersed in the sPPO matrix [153]. Experimental results show that adding 5 wt.% sBN can result in 168 % increase in Young's modulus of sPPO. This large increase is attributed to π − π interactions between sBN and sPPO. Moreover, availability of more sulfonic acids in sPPO/sBN and the formation of hydrogen bonding with water molecules, the water sorption of the sPPO/sBN was 50.61%, almost 50% higher than that of bare sPPO. The proton conductivity was also enhanced by 67% in sPPO/sBN due to the presence of excess sulfonic acid groups and hydrogen bonding bridges, introducing long-range proton-conducting pathways. In this case, vehicular and Grotthuss mechanisms were suggested for the proton conduction. In the vehicular mechanism, protonated water molecules in the forms of H 3 O + aq , H 5 O 2 + aq , and H 9 O 4 + aq diffuse through the media and act as ion transport carriers, whereas the latter mechanism deals with the stationary water molecules in which ion hopping occurs from one water molecule to another through hydrogen-bonded ionic channels [43]. Figure 10a depicts how sBN simplifies proton conduction by the vehicular and Grotthuss mechanisms. A low methanol permeability of 2.04 × 10 −7 cm 2 /s was also obtained in sPPO/sBN, which is significantly lower than sPPO and Nafion (three and nine times, respectively). This is due to interfacial interactions of sPPO and sBN, as well as the barrier effect of sBN owing to its 2D structure, which hinders the methanol crossover [153]. Moreover, the overall performance of the cells indicated a high peak power density of 122 mW/cm 2 for sPPO/sBN, which is two times higher than that of sPPO. In another survey, a polyether ether ketone (PEEK) membrane reinforced with polyaniline coated activated boron nitride (PANI-BN) was used to improve the performance of DMFC [154]. Although the addition of PANI-BN improved the performance of the membrane, it was not as high as the results obtained from sPPO/sBN. For instance, the best results for SPEEK/0.1 wt.% PANI-BN membranes were a water uptake of 58 %, proton conductivity of 4.13 mS/cm, methanol permeability of 3.08 × 10 −7 cm 2 /s, and power density of 11.38 mW/cm 2 .

Cathode Catalyst
Although 2D h-BN can participate in catalytic reactions [14], there are not many reports on 2D h-BN being employed in fuel cell electrodes. Recently, a novel nanostructure of rGO-2D BN-Fe 2 O 3 was introduced as the cathode catalyst of a microbial fuel cell (MFC) [155]. The concept is based on the insertion of 2D h-BN and Fe 2 O 3 onto the r-GO nanosheets to enhance the surface area for facile transport of reactants and improve electron transport to graphene. The CV graphs in Figure 10b indicate that the limiting current of the doublelayer (DL) in r-GO-2D h-BN-Fe 2 O 3 electrode is higher than that of the single (SL) and trilayer (TL) (5.5 mA, 4.9 mA, and 4.4 mA, respectively); however, it was lower than that of Pt/C (5.9 mA). The superior limiting current of DL is described by the larger surface area and electrical conductivity of r-GO and 2D h-BN nanosheets. The EIS evaluations also confirm that Pt/C cathode has better performance since its correlated semicircle in the high-frequency area has a smaller diameter than that of rGO-BN-Fe 2 O 3 (Figure 10c). Power density measurements of the cells also reveal that rGO-BN-Fe 2 O 3 is a promising cathode catalyst for MFCs with a power density of 1673 mW/m 2 , which is 81% of the obtained power density for MFC using Pt/C cathode (Figure 10d) [155]. All in all, the abundance and cost effectiveness of this new cathode material makes it a competitive catalyst compared to other conventional ones.

Hydrogen Environmental Barrier Coating
A new application of 2D h-BN in fuel cells is the hydrogen environmental barrier coatings (EBCs), which hinder the hydrogen diffusion and embrittlement of materials. Studies revealed that hydrogen embrittlement can occur in fuel cells working at tempera-tures up to 1273 K [156]. In addition to preventing embrittlement, it is necessary to retain hydrogen fuel. The deposition method and obtained microstructure significantly affect the effectiveness of 2D h-BN barriers. For instance, employing ion beam assisted deposition can form (002) planes oriented orthogonal to the substrate, not capable of preventing H 2 diffusion [157]. Radiofrequency magnetron sputtering, however, results in barriers with considerably lower permeability of H 2 [158]. Deposition of h-BN flakes parallel to the substrate demonstrated remarkable outcomes by trapping H 2 as stable bubbles up to 1073 K [159]. Recently, atomic layer deposition (ALD) was exploited to deposition 2D h-BN onto ZrO 2 powders to evaluate their high temperature H 2 diffusion behavior [160]. ALD can provide low defect density thin films that are strongly bonded to the substrate [161]. Figure 11a shows TEM images of a thin 2D h-BN layer of 2.4 nm deposited onto ZrO 2 . The stability of h-BN at high temperatures was evaluated by exposing bulk BN powder to 20% H 2 in Ar balance for three hours at 1773 K. Almost all the N contents of the 2D h-BN lattice remained intact, which can be explained by the high N vacancy formation energy of 5.6 eV at elevated temperatures [160].

Seal
Another novel application of 2D h-BN in fuel cells is as a compressive seal in a solid oxide fuel cell (SOFC). A 2D h-BN/glass nanocomposite formed by tape casting method was used as a practical sealing material, where the sealing phenomenon occurs gradually through running thermal cycles [162]. The reported leakage rates decreased to less than half after only 10 thermal cycles under different input gas pressure, as shown in Figure 11b. Figure 11c illustrates the structure of seals after one thermal cycle. This porous structure explains the high leakage rate. Conversely, Figure 11d displays a denser microstructure formed after 10 thermal cycles. Finally, more tests revealed that a liquid B 2 O 3 film was formed on the 2D h-BN surface over time at elevated temperatures shown schematically in Figure 11e-g.

Solar Cell
Solar photovoltaic technology is considered to be mainstream renewable technology with a great potential to generate clean, reliable, scalable, and affordable electricity [163]. However, from the materials engineering point of view, as regards circuit design, many issues still need to be resolved [164]. In this regard, 2D layered materials have been in the forefront of solar cell industry due to their unique electrical and optical properties [165]. Nevertheless, 2D materials-based solar cells performance might be disturbed by environmental conditions such as certain gases and humidity, signifying the need for protective layers. In the last decade, scientists have used 2D h-BN in various solar cells to passivate the surface of active materials or design the interface of heterojunction solar cells. The peculiar characteristics such as lack of dangling bond, large bandgap, excellent chemical and thermal stability, excellent crystallinity, medium dielectric constant, and easy transfer to substrates without affecting its bulk properties make 2D h-BN a top candidate for this application [166][167][168].
In a research study, 2D h-BN was used as a passivation layer on top of a MoS 2 /WSe 2 van der Waals heterojunction solar cell to decrease the cell's electrical loss [169]. As a result, the average recombination lifetime of the electron-hole increased from 112 ps to 131 ps, indicating the reduced recombination rate at the active area as well as leakage path passivation. Thus, the charge diffusion length and charge extraction rate were enhanced. Recently, 2D h-BN has been introduced as a promising passivation layer for InP as a highperformance III-V semiconductor. Raj et al. [170] exploited five and seven-layer thickness of 2D h-BN to passivate InP solar cells in order to reduce the recombination phenomenon caused by the surface defect states (Figure 12a-c). The calculated number for the interface defect density was 2 × 10 12 eV −1 cm −2 , which is better than the results of other passivation layers such as Al 2 O 3 , HfO 2 , and SiO 2 with interface defect density of 2.6 × 10 12 eV −1 cm −2 , 5.4 × 10 12 eV −1 cm −2 , and 5 × 10 12 eV −1 cm −2 , respectively [171,172]. In addition, by introducing 2D h-BN on InP, the efficiency, V OC , J SC , and fill factor (FF) of the solar cell increased from 11.5% to 17.2%, 0.72 V to 0.78 V, 27.4 mAcm −2 to 29.4 mAcm −2 , and 58.6% to 75.2%, respectively. They postulated that the surface passivation occurred through charge transferring from the surface defect states of InP to 2D h-BN, which made them inactive. Notably, they demonstrated that the passivation is dependent on the thickness of 2D h-BN. This thickness dependence is due to a quantum mechanical tunneling current which is directly proportional to the charge carrier density at the dielectric/semiconductor interface (h-BN/i-InP), and inversely proportional to the thickness of the dielectric layer and band offsets. Accordingly, a thinner 2D h-BN leads to higher J sc , while a thicker one decreases the tunneling current, attributing to lower contact recombination, lower dark current, and ultimately higher V oc (Figure 12d).  Exploiting 2D h-BN as a passive layer can be considered a milestone for InP since the lack of lattice-matched and wide bandgap passivation layer was a limiting factor for its development. Except for passivation of heterostructure solar cells, 2D h-BN has been employed in dye-sensitized solar cells (DSSCs) to passivate the anode (TiO 2 nanoparticles). The anode material in DSSC should provide a high surface area in order to host a large number of photoactive dye molecules [174]. However, by increasing the surface area of TiO 2, the number of surface states increases, leading to higher rates of recombination [175]. Moreover, the heat treatment of TiO 2 at temperatures around 450 • C to form the anatase phase which is favorable for DSSC, leads to the formation of oxygen vacancies [176]. Consequently, there will be more surface deficiencies deteriorating the cell's performance. Therefore, a passivation layer is necessary to reduce the surface energy and hinder the recombination of electron-holes by introducing an energy barrier. Oxide ceramics with high-k dielectrics are conventional candidates for the passivation of TiO 2 . However, there are several critical issues associated with them, i.e., poor conformal coverage over a large area, blocking photo-generated electrons from the dye due to relatively high thickness, and reducing the porosity of TiO 2 [177]. In contrast, an ultrathin and high crystalline layer of h-BN can be deposited on TiO 2 to modify its semiconducting behavior. It was reported that by applying a 2D h-BN layer on TiO 2 the J SC , V OC and efficiency of the DSSC were enhanced by 23%, 1.6%, and 57%, respectively, meaning well suppression of the surface states at the TiO 2 /electrolyte interface [178].
Another application of 2D h-BN in solar cells is interface engineering for heterojunction structures. Schottky junction solar cells, such as graphene/Si, have become popular due to their high potential for inexpensive photovoltaic applications [179]. However, their low Schottky barrier leads to the interface recombination of charge carriers that remarkably suppresses their efficiency. One effective way to address this issue is to insert an insulating 2D material between graphene and Si. Accordingly, 2D h-BN would be an ideal candidate for this application. In a study on graphene/h-BN/Si solar cell platforms [173] it was shown that 2D h-BN can hinder the diffusion of electrons (majority carrier) from n-Si to graphene, and the transport of holes (minority carrier) from n-Si to graphene. Based on the band alignment demonstrated in Figure 12e, without a 2D h-BN interlayer, the built-in electric field at the interface causes the separation of photogenerated electron/hole pairs and pulls the electrons to n-Si, whereas holes drift towards graphene. Simultaneously, the built-in electric field hinders the diffusion of electrons from n-Si to graphene (dashed line in Figure 12e,f). However, the small surface potential of n-Si (Φ= 0.6-0.7 eV) will become smaller under light irradiation. The inevitable consequence of this would be a significant current leakage [173]. By introducing 2D h-BN to the system, the unwanted recombination will be reduced since the conduction band offset (∆EC) of h-BN/Si is larger than 4.05 eV, meaning there is a large obstacle for electron transport from n-Si to graphene. On the other hand, the small valence band offset (∆EV) of 0.63 eV facilitates the drift of holes from n-Si to graphene under the built-in electric field, which is an essential factor to prevent an increase in the series resistance of cells. Hence, the reverse saturation current density was significantly reduced from 4.09 × 10 −3 mA/cm 2 to 2.56 × 10 −4 mA/cm 2 for graphene/Si and graphene/h-BN/Si, respectively. Using 2D h-BN in other heterostructure solar cells such as MoS 2 /h-BN/GaAs [180] and graphene/h-BN/GaAs [181] has also improved the interface characteristics in a similar approach. Table 3 provides a summary of energy conversion devices based on 2D h-BN.

Future Challenges and Remarks
To sum up, increasing 2D h-BN integration into energy production/storage units demonstrates the great intrinsic potential of this two-dimensional material for the energy industry. The electrical insulation of 2D h-BN allows the prevention of close contact between electrodes in a battery or where the separation of conductive parts is needed. The superior heat dissipation due to thermally conductive pathways of 2D h-BN provides much better control of heat management and safety concerns by minimizing heat accumulation. The partially ionic nature of bonds between boron and nitrogen in 2D h-BN lattice provides surface polarization and offers a prominent pathway for proton exchanges. Similarly, 2D h-BN can be incorporated into cathodic materials. The enhancement of electrolyte ionic conductivity, mechanical and thermal stability, and engineerability are other benefits of using 2D h-BN in energy research. The surface functionalization of 2D h-BN and its hybridization with other 2D materials have been effective for bandgap modulation.
Nevertheless, there are still challenges and gaps in our understanding of thicknessdependent characteristics of layered h-BN, charge transfer mechanisms on 2D h-BN surface via chemical reactions, electron tunneling through thin layers of h-BN, surface catalytic sites, and charge extraction mechanisms. Addressing these challenges could be a viable solution to enhance the applicability of 2D h-BN for energy applications. Generally, these challenges can be tackled by two different approaches: (1) integration of 2D h-BN into existing technologies and process optimization, (2) development of new products with h-BN as the core cell. In either case, the enhanced performance of the device is entangled with the surface chemical composition of 2D h-BN. From the formation of percolation networks in h-BN-based nanocomposites to enhancement of the lithium transference number of LIB, to ionic conductivity and transparency for solar cell and proton exchange rates, they are all dependent on 2D h-BN surfaces characteristics. Simple and cost-effective synthesis methods, such as liquid exfoliation, often produce 2D h-BN with various functional groups/dopants. Therefore, regulating the synthesis condition leads to the enhancement of device efficiency. This becomes particularly important in thermoelectric applications where the negligible thermoelectric properties of 2D h-BN are considerably improved by functionalization of the basal plane, introducing point defects, modifying the edge chemistry, and hybridization with other 2D materials. In contrast, the need for bare 2D h-BN for voltaic cell applications may necessitate using more complex and costly approaches such as CVD.
The application of 2D h-BN in energy research is still at preliminary stages and besides batteries, supercapacitors, thermoelectrics, and fuel cells, other sustainable energy fields such as photovoltaics and solar cells are envisaged to benefit from the great separation characteristics of 2D h-BN.