Theoretical design of diamondlike superhard structures at high pressure
Li Quan†, , Zheng Wei-Tao‡,
Department of Materials Science, Key Laboratory of Automobile Materials of MOE and State Key Laboratory of Superhard Materials, Jilin University, Changchun 130012, China

 

† Corresponding author. E-mail: liquan777@jlu.edu.cn

‡ Corresponding author. E-mail: wtzheng@jlu.edu.cn

Project supported by the National Natural Science Foundation of China (Grant Nos. 51202084, 11474125, and 51372095).

Abstract
Abstract

Diamond, as the hardest known material, has been widely used in industrial applications as abrasives, coatings, and cutting and polishing tools, but it is restricted by several shortcomings, e.g., its low thermal and chemical stability. Considerable efforts have been devoted to designing or synthesizing the diamond-like B–C–N–O compounds, which exhibit excellent mechanical property. In this paper, we review the recent theoretical design of diamond-like superhard structures at high pressure. In particular, the recently designed high symmetric phase of low-energy cubic BC3 meets the experimental observation, and clarifies the actual existence of cubic symmetric phase for the compounds formed by B–C–N–O system, besides the classical example of cubic boron nitride.

1. Introduction

The ages of stone, bronze and iron in a chronological sequence describe the succession of periods based on the use of stone (and wood), bronze and iron respectively, which are named for their respective tool-making technologies. From the view of mechanical researcher, humans have been searching for new harder materials, and new harder materials have always fascinated them. Natural diamond has long been found and used for making tools; however, it is expensive because it is extremely rare. Since the artificial synthesizing of diamond in the 1950s, its industrial application has been widely developed. Then, the diamond age, in which diamond or other superhard materials with hardness above 40 GPa are used as ideal materials for machine tools and cutting tools, has set in.

Experimentally, diamond is still the hardest known substance with the hardness reaching 100–160 GPa, though over the past decades extensive experimental efforts have been devoted to exploring new materials that could be harder than diamond.[14] However, diamond is exceptionally weak for ferrous metal’s cutting and is burned into carbon dioxide at 800–900 °C in air, which restricts its applications. Another well-known superhard material is cubic boron nitride (c-BN) with cubic symmetry, which is artificially synthesized and has been considered as the second hardest material for a long time. It possesses fascinating properties, such as high thermal stability (approximately 1650 K) and low chemical reactivity. However, the hardness of c-BN is only in a range of 46–66 GPa, much lower than that of diamond.[5,6] Therefore, the search for new superhard materials with the properties comparable or even superior to those of diamond in hardness and higher resistance to oxidation and ferrous metals, is an intriguing and long-standing job.

Superhard materials are mainly formed by light elements (B, C, N, and O) compounds with short and strong three-dimensional covalent bonds.[625] Recently, a new family of materials formed by heavy transition metals and light elements is proposed to be potential superhard since heavy transition metals can basically introduce high valence electron density into the compounds to resist both elastic and plastic deformation.[2648] Experimentally, these compounds formed by heavy transition metals and light elements have been demonstrated to have high bulk modulus, indicating that those compounds possess excellent ultra-incompressibility. However, the hardness obtained from the asymptotic load-independent region of the hardness as a function of load is not so high as excepted. Drawn from theoretical calculations for these heavy transition metals’ compounds is a conclusion that the highly directional bonds between light-element atoms with large electron densities are short and strong; however, most bonds formed between transition-metal and light-element atoms with lower electron densities are long and weak. As is well known, the strong anisotropic bonding behavior is a severe problem for the application. Up to now, it still remains a challenge to obtain superhard materials within the compounds formed by heavy transition metals and light elements.

Diamond and c-BN are transparent crystals with tetrahedrally bonded atoms in a covalent network lattice (sp3) that crystallizes into the face centered cubic structure (Fig. 1). The search for new superhard materials has been greatly motivated following the successful synthesis of man-made diamonds and c-BN. Considerable efforts have been devoted to synthesizing various B–C–N compounds with covalent network lattice by sintering various starting materials at high pressure and temperature. The experimental challenge to the synthesis of diamondlike superhard materials still exist, since they can be only synthesized in a narrow temperature and pressure region, which blocks the industrial synthesis in batch. Encouragingly, several sp3 bonded B–C–N–O compounds have recently been synthesized by high temperature and high pressure method in laboratory and identified as “diamond-like” or “cubic (c-)” compounds. These densely packed, three-dimensional and strongly covalent bonded compounds have short bond length, low ionicity, high bulk modulus and high hardness. The material formed by B, C, N, and O system with 4 valence electrons per atom is classified as isoelectronic compound with diamond, which may possess similar chemical bonds and structural properties to those of isoelectronic compounds of c-BN,[5,6] B2O,[49] and c-BCxN (x = 1–6),[6,811] and asymmetrical and non-isoelectronic compounds of BC3,[21] BC5,[13] and B2CN.[12] As is well known, the crystal structures of materials are the basis for understanding relevant properties. In this paper, the theoretical superhard diamond-like structures for these isoelectronic and non-isoelectronic compounds are reviewed.

Fig. 1. Crystal structures of diamond (a) and c-BN (b).
2. Isoelectronic diamondlike compounds
2.1. B2O

An isoelectronic and asymmetrical compound of diamond, B2O, was synthesized by Endo et al. through the chemical reaction of BP with oxygen[49] under the condition of 2.0–6.0 GPa and 800–1350 °C. The measured Vickers hardness is in a range from 33.5 GPa to 40.5 GPa, which is close to the criterion of superhard materials. From the measured x-ray diffraction data, Endo et al. proposed a possible crystal structure consisting of diamond lattice sites with planes of O alternating with two planes of B along the [111] crystallographic direction. Subsequently, Grumbach et al.[50] theoretically proposed that the 111 structure be segregated into layers of BO and B and then proposed a 100 structure with stacking the B and O planes in the [100] direction which is energetically more stable than the 111 structure. However, the x-ray diffraction (XRD) peaks of the 111 and 100 structures could not match the experimental data.

Using ab initio evolutionary method[5153] for crystal structure prediction, we have investigated the candidate crystal structures of the synthesized superhard B2O.[54] We obtained an intriguing monoclinic C2/m structure which is most energetically favorable. The C2/m structure contains four B2O formula units in one unit cell, in which both B and O atoms occupy the Wyckoff 4i (z1, 0, z2) sites. Interestingly, B atoms possess coordination numbers ranging from 4 to 7. A novel tetragonal P42mc contains both sp3- and sp2-hybridized boron atoms, in which vertical flat 6-fold rings are linked together by sp3-hybridized B atoms along c axis. The inspection of the crystal structure of hexagonal P-62m along c axis reveals that it can be viewed as distorted graphite with one B atom shifting to the middle of the two flat layers and forming six-coordinated bonds. All O atoms in C2/m, P42mc, and P-62m structures are three-coordinated (sp2). The monoclinic C2/m of B2O is an excellent potential superhard material with a simulated hardness of 66.7 GPa. However, the XRD of C2/m cannot fit to the experimental XRD data[49] of the synthesized sample.

Fig. 2. Crystal structures of P-4m2 B2O,[54] I-42d B2CO,[55] R3m-2u BC2N,[56] and Imm2 B3NO.[57]

From the simulated XRD patterns, the most likely structure for the synthesized superhard B2O is the P-4m2 (Fig. 2) or even a mixture of P-4m2 and P42mc, which are not diamond-like. The simulated hardness values of P42mc and P-4m2 are 40.5 and 44.8 GPa, respectively, in accordance with the experimental values of 33.5–40.5 GPa.[54]

2.2. B2CO

Several B–C–O materials have successfully been synthesized at high pressure and high temperature, such as B6C1.1O0.33 to B6C1.28O0.31[58] and B(C, O)0.1555,[59] which demonstrate that superhard B–C–O compounds can form typical sp3 covalent hybridization, which contain B–C, B–O, C–O, and C–C bonds. Therefore, B–C–O materials can be expected to be good candidates for designing new superhard materials. B2CO has 16 valence electrons per formula unit (4 valence electrons per atom) and is thus isoelectronic with diamond.

Using the CALYPSO method[6062] for predicting the crystal structure, we design a potential superhard ternary material, B2CO, which is an isoelectronic example in the B–C–O system.[55] We predict two energetically competitive diamond-like structures with space groups of P-4m2 (tP4) and I-42d (tI16), see Fig. 2. The P-4m2-1u structure is isostructural to t-B2CN.[12,63,64] It is found that B atoms in both structures occupy four diagonal positions; while O atoms sit in the vertexes and two opposite surface positions in the tP4 structure, and sit on half of the vertexes and surface positions in the tI16 structure. Both structures adopt the same sp3 bonding environment: each B is tetragonally bonded with two C and two O atoms, and C and O atoms are tetragonally bonded with four B atoms, forming B–C and B–O covalent bonds in the diamondlike lattices. The results of total energy calculations at ambient pressure show that tI16 structure is 43 meV/f.u. lower than that of tP4. The dynamical and mechanical stabilities of tI16 and tP4 are examined by the calculations of phonon dispersions and elastic constants, respectively. The two structures exhibit similar mechanical properties, such as high bulk modulus (approximately 31 GPa), high shear modulus (approximately 260 GPa), high hardness (approximately 50 GPa), and low Poisson’s ratio (approximately 0.17).

2.3. c-BC2N

In 1994, Nakano et al.[11] experimentally observed the formation of c-BC2N from graphite BC2N at 7.7 GPa and approximately 2000–2400 K. Later, Knittle et al.[8] prepared a c-BCxN (x = 0.9–3.0) solid solution by using even higher static pressure of 30 GPa and laser heating, starting from either the microstalline BCN, or the mechanical mixture of graphitic carbon and boron nitride. The resulting BC2N has a bulk modulus of 355 GPa. Zhao et al.[6] have successfully synthesized millimeter-sized bulk samples of c-BC2N with the Hv about 62 GPa. Using shock-wave compression, Komatsu et al.[10] obtained c-BC2N with a bulk modulus of 401 GPa, which is larger than that of c-BN. Under high pressure above 18 GPa and high temperature over 2200 K, Solozhenko et al.[9] experimentally reported that the bulk modulus 282 GPa of synthesized c-BC2N is lower than that of c-BN but the measured hardness of 76 GPa is much higher than that of c-BN single crystal and slightly lower than that of diamond. By all the evidence, it is clear that c-BC2N is a superhard material. Diamond-like BC2N (c-BC2N) has gained extensive attention since it has been expected to be thermally and chemically more stable than diamond, and harder than c-BN.

On the theoretical side, much effort has also been made to uncover the crystal structure of the superhard BC2N.[6577] As is well known, the stable crystal structure has the lowest Gibbs free energy under the given conditions. Based on this principle, Sun et al.[70] proposed seven possible structures (struc-m, m = 1–7) with different atomic configurations in the eight-atom zinc-blende (ZB) cubic unit cell. They proposed that the struc-1 with Pmm2 space group having the lowest energy is the best candidate. While this structure has different appellations in the literature such as BC2N-1,[66] β -BC2N,[67] II-BC2N,[68] and HD1,[68] in fact, they are identical. Sun et al.[70] further predicted a chalcopyrite structure (cp-BC2N) with the tetragonal symmetry (space group, I-42d). The calculated hardness of cp-BC2N is 72.2 GPa very close to the experimental value. More recently, Luo et al.[78] suggested that the body-centered BC2N phase (bc6-BC2N) has lower density than ZB BC2N, wurtzite BC2N and cp-BC2N, and they have estimated that the hardness of bc6-BC2N is over 60 GPa. Among the above structures, struc-1 possesses the lowest free energy. Very recently, Zhou et al.[72] predicted a z-BC2N structure, which is is comprised of the sixteen-atom supercell of diamond. The simulated XRD spectrum agrees well with the experimental data, and importantly the total energy of z-BC2N is lower than that in struc-1. Luo et al.[79] further predicted three structures of wurzite-type BC2N, among which BC2N-w3 also has lower total energy than struc-1. Very recently, Chen et al.[74] proposed a series of short period (C2)n(BN)n (111) superlattices (denoted as BC2Nn×n, with n = 1,2,3,…). The authors suggested that these superlattices have the lowest energies in those of the previously proposed structures. However, the intense debate is still going on.[76,77] Also, it is worth noting that the crystal structure of superhard c-BC2N should be unique instead of a structural series.

To clarify the intensive debate on the superhard phase of c-BC2N, we have extensively explored the crystal structure of BC2N by using ab initio evolutionary method.[56] We have predicted three polytypic structural families: orthorhombic Pmm2-nu, hexagonal P3m1-nu, and rhombohedral R3m-nu, where n refers to that the structure contains n BC2N units per primitive cell (n = 1, 2, 4). Through total energy calculations, it is found that R3m structural series possesses the lowest enthalpy. The structural stability of R3m structural series has been confirmed by the phonon calculations. Analyses of the simulated x-ray diffraction pattern and energy-loss near-edge spectroscopy suggest that our predicted R3m-2u [Fig. 2] is the best candidate phase for the observed superhard BC2N. We have also demonstrated that the previously proposed high density and low density forms might be identical and their x-ray diffraction patterns could be reasonably understood by the single phase of R3m-2u. The calculated electronic structures show that R3m-2u BC2N is a wide gap semi-conductor with an indirect DFT band gap of 3.8 eV. The calculated bulk modulus of R3m-2u BC2N is 396 GPa, which is in good accordance with the experimental value of 401 GPa by Komatsu et al.,[10] but much higher than the measured value of 282 GPa by Solozhenko et al.[9] The estimated theoretical Vichers hardness of R3m-2u BC2N is 62 GPa, in agreement with the experimental value of 62–76 GPa.[6,9]

2.4. B3NO

Previous studies demonstrated that diamond-like superhard c-BN and B2O compounds can form typical sp3 covalent network bonds.[54] It is reasonable to explore whether those materials with mixed B–N and B–O bonds may possess high hardness values. We have performed superhard-driven search by using an unbiased structure search method based on CALYPSO method[57] in a ternary B–N–O system, B3NO, which is also an isoelectronic compound with diamond.[80] Crystal structure is critical to understand the hardness and relevant physical properties of B3NO. We have predicted various structures for B3NO under 0–100 GPa.[57] It is in contrast to the traditional ground-state prediction method where the total energy is solely used as the fittness function. We choose hardness as the fittness function in combination with the first-principles method to obtain the hardness versus energy map by seeking a balance between hardness and energy for a better mechanical description of given chemical systems. We predicted a variety of structures with high hardness values and relatively low energies.[80]

A variety of newly structures of B3NO compound contain short, strong, and three-dimensional covalent bonds, which are responsible for the predicted superior mechanical properties.[57] Two energetically stable B3NO structures with space groups of Imm2 (oI20) and Pmn21(oP20) are uncovered. The oI20 structure (Fig. 2) is the most stable structure which possesses irregular 6-member rings, while oP20 is formed by 4-numbered, 6-numbered, and 8-numbered rings. To understand the electronic properties, we perform calculations on the electronic band structures and partial densities of states of oI20 and oP20. The calculated results reveal the semiconductor features of oI20 and oP20 structures with indirect band gaps of 0.87 eV and 0.12 eV, respectively. The electronic states near the Fermi level are mainly contributed by B-2p and a few N-2p, and O-2p states. The partial densities of states for B-2p, N-2p, and O-2p are very similar in an energy range from −12 eV to −4 eV, indicating the significant hybridizations between these orbitals and strong covalent interactions in both the B–N and B–O bonds. The calculated electron localization function (ELF) shows that the O atoms in oI20 and oP20 structures are threefold coordinated with three near-neighbor B atoms and a lone pair of electrons. The results show that oI20 and oP20 structure may have a broad prospect in industrial applications, such as solar battery, and luminous materials. The two structures exhibit high bulk moduli of 326–332 GPa, significantly higher than that of B2O (240 GPa),[54] indicating the structures are difficult to compress near their respective equilibrium volumes. The further hardness calculations show the two structures are likely to become superhard materials with Vickers hardness of 45.9 GPa (oI20) and 47.4 GPa (oP20), respectively, exceeding the criterion of superhard materials (40 GPa).

3. Non-isoelectronic diamond-like compounds
3.1. B2CN

Eproperties are of considerable demand for the creation of multifunctional device.[81] Since the discovery of superconductivity in boron-doped diamond (4 K),[82] the superconducting behaviors in “covalent metals” (dominated by strong covalent bonds, rather than the metallic bonding) have also been an exotic subject.[83] It is suggested that the strongly directional covalent characteristics in metallic compounds lead to large phonon frequencies and electron–phonon coupling potential, which contribute to the increase of Tc.[84] Thus, metallic B–C–N compound with exclusively covalent bond can thus be expected to simultaneously possess superhard and superconducting properties. Metallicity of ternary B–C–N compounds can be evaluated by the following simple rule: pZV(B) + mZV(C) + lZV(N) ≠ 4n.[85] The values p, m, l, and n are integers, and ZV(B), ZV(C), and ZV(N) are the atomic valence states (2s and 2p) for B, C, and N, respectively. Recently, diamond-like B2CN[12,63] and BCx,[13,21] crystals have been synthesized under the condition of high-pressure and high-temperature (1500–2500 K), which are suggested to be conductors.

We have extensively investigated the crystal structures of B2CN by using ab initio CALYPSO algorithm in a large pressure range of 0–100 GPa.[64] Five competitive structures with clear tetrahedrally sp3 hybridization are predicted: tetragonal P-4m2 (1 f.u. /cell), hexagonal P3m1 (1 f.u./cell), rhombohedral R3m (1 f.u./cell), and orthorhombic Pmm2 (2 f.u./cell), and Pmma (2 f.u./cell). P-4m2, Pmm2, and Pmma can be viewed as a periodic sandwich-like structure formed by stacking CBNB, BNBBCCBN, and CCBNBBNB along the ⟨100⟩ direction of diamond structure, respectively. Instead, the stacking sequences of P3m1 and R3m follows the ⟨111⟩ direction. With fixing the numbers of 1 and 2 formula units in the primitive cell, it is found that the energetically best structures are the diamond-[100] structures: t-B2CN and Pmma. The C–C and B–N chemical bonds are more favorable than B–B, B–C, and C–N for forming the diamond-like structure with sp3-hybridized, thus we define C–C and B–N bonds as stable bonds and the other relatively weaker bonds as less stable bonds. In the two most stable structures (Pmma [Fig. 3] and Pmm2), it is interesting to note that the ratio of the stable chemical bonds to less stable bonds reaches 5:3, which is the largest in the five structures. Further increasing the formula units (3 or 4 f.u./cell) in the simulation cell, we find that the produced structures tend to separate into bulk diamond + c-BN + boron layers. We find that all the B2CN structures have positive formation energies, indicating that these structures tend to phase decomposition. The calculated large bulk modulus (318–333 GPa) and high hardness (56–58 GPa) reveal that they are ultra-incompressible and superhard materials. Therefore, the synthesized B2CN can be expected to be applied to the high-pressure devices for investigating the electric properties of various materials under extreme conditions.

Fig. 3. Crystal structures of Pmma B2CN,[64] Pmma-a BC3,[86] P3m1 BC5,[87] and P-4m2 BC7.[88]

Due to the electron deficiency of sp3 hybridized B2CN in comparison with that of diamond, the five diamond-like B2CN structures show metallic with hole nature. Further electron-phonon coupling calculations predict that all the five phases are superconductors. The estimated Tc for diamond-[111] structures (R3m and P-3m1) reaches a very high value, 44–53 K, in contrast to the smaller value (2–9 K) in diamond-[100] structures. The large difference in superconductivity of B2CN is mainly related to the relatively low frequency (< 20 THz). In view of the excellent properties of diamond-like B2CN, it is necessary to pay much attention to the experimental synthesis and conductivity measurements.

3.2. c-BCx

Boron-doped diamond has been widely explored as better oxidation and ferrous resistant materials to cover the shortage of diamond and expand the applications to electric devices.[89,90] Electrical measurements demonstrate that Boron-doped diamond becomes metallic and even turns into a superconductor with a transition temperature Tc of 4 K at a higher doping level (2%–3%).[82] Further experimental and theoretical studies indicate a Tc increases with increasing dopant concentrations.[9193] However, it remains a major challenge to introduce such a large amount of B into diamond in experiments, and the maximum B concentration has been limited up to approximately 5%.[93] An alternative route to achieving high B content is to directly synthesize the B–C compounds. Recently, Solozhenko et al.[13] successfully synthesized a diamond-like B–C phase (c-BC5) with a high B content ever achieved (∼ 16.7%). A subsequent hexagonal structure constructed from a six-atom supercell of diamond for c-BC5 was proposed[9496] and a large Tc (45 K) was predicted.[97] The obtained c-BC5 sample has been measured to possess a large bulk modulus (335 GPa), unusually high fracture toughness (9.5 MPa·m1/2), and high thermal stability (up to 1900 K). Remarkably, the indentation experiment has demonstrated an extremely high hardness (71 GPa), harder than cubic BN.[13]

We have uncovered four other low-energy structures: tetragonal I-4m2 (1 f.u./cell), orthorhombic Imm2 (1 f.u./cell), hexagonal P-3m1 (2 f.u./cell), and orthorhombic Pmma (2 f.u./cell).[98] It is found that except for Imm2, other structures are built up by the sandwich-like layers with different stacking sequences. Specifically, I-4m2 and Pmma structures having the atomic packing of ABCABC··· along the [100] crystallographic direction of diamond can be viewed as diamond-[100] structure. Instead, P3m1 and P-3m1 structures are diamond-[111] structures as the atomic packing follows the [111] direction. Subsequently, a new P3m1_2u [Fig. 3] with slightly lower formation energy was theoretically proposed for the candidate structures for the synthesized BC5.[87] All the predicted structures are tetrahedrally bonded with clear sp3 hybridization. Further increasing the formula units (3 or 4 f.u./cell) in the simulation cell, the produced structures tend to decompose into bulk diamond + boron-carbon layers. This fact is understandable in view of the metastable nature of the synthesized BC5 compound supported also by the calculated formation energy.

These diamond-like BC5 possess similar structural and mechanical properties. The calculated bulk modulus is in a range from 373 GPa to 398 GPa, in good accordance with the experimental value of 335 GPa.[98] On the basis of the microscopic hardness model,[99,100] the Vickers hardness of Pmma structure is estimated to be 61–74 GPa, in satisfactory agreement with the experimental data (71 GPa). It was previously found that a small metallic component has a strong negative effect on hardness and thus the correction of metallic bonding is necessary to account for the experimental hardness for electron conductors. The underlying mechanism is that those electrons delocalized to contribute to the conduction should be excluded from the hardness calculation. However, for the hole conductors, the major carriers are holes and the valence electrons are mainly localized to form covalent bonds. It is thus unnecessary to include the metallic correction in the hardness calculation. Indeed, if we force to include the correction of metallic bonding, the simulated hardness decreases by about 15 GPa, which is largely deviated from the experimental data.

Using CALYPSO methodology for crystal structure prediction of the potential superhard materials of BC7, we find four potential superhard structures of tetragonal P-4m2, hexagonal P3m1, orthorhombic Pmm2, and hexagonal R3m, which are energetically much superior to the previously proposed P-43m structure.[88] All the predicted structures are tetrahedrally bonded with sp3 hybridization. The energetically best structure is the tetragonal P-4m2 (Fig. 3). The calculated Vickers hardness and B0 of the lowest energy P-4m2 polymorph reaching very high values (75.2 GPa and 377 GPa) indicate that diamond-like BC7 is a superhard and ultra-incompressible material. Further phonon and elastic constant calculations imply that these structures are all mechanically stable.

Encouragingly, diamond-like c-BC3 with high B content is also successfully synthesized under the condition of high pressure and high temperature[21] by using the precursor materials. Numerous structural models have been theoretically proposed to understand the structural nature of diamondlike c-BC3.[86,87,101] However, none of these proposed sp3 structural models is consistent with the experimental observation of cubic diamond structure. A lack of accurate structural determination for these materials has severely impeded further study of their properties. We report a novel high symmetric phase of low-energy cubic BC3 (I-43m symmetry and 64 atoms per unit cell, d-BC3) consisting of unique eight B–B bond distributions identified by using an unbiased structure search method based on particle-swarm optimization algorithms in combination with density functional theory calculations.[102] At ambient pressure, the optimized lattice parameter of d-BC3 is 7.330 Å (Fig. 4), with boron occupying Wyckoff 8c (0.89711, 0.89711, 0.10289) and 8c (0.24273, 0.24273, 0.75727), and carbon occupying 12e (0.26755, 0.0, 0.0), 12d (0.0, 0.5, 0.75), and 24g (0.61792, 0.87580, 0.87580). A Bader charge analysis reveals that the charge density at the B–B bond critical point is 0.463 electron charges per Å3 with a Laplacian value of −1.14, indicating the covalent nature of the B–B bond in d-BC3. As shown in Fig. 5, the strong covalent B–B bonding can also be shown in the two-dimensional ELF on the (110) planes passing through the B–B bonds as shown in Fig. 5. The proposed d-BC3 becomes stable structure with exclusively sp3 hybridization above 41.3 GPa, in good agreement with the reported synthesis pressure of 39 GPa.[21] The simulated x-ray diffraction pattern and Raman modes of this d-BC3 phase are in excellent agreement with the experimental data.[21]

Fig. 4. Crystal structures of d-BC3.
Fig. 5. The two-dimensional ELF on the (110) planes passing through the B–B bonds.

To deal with the metastable material with a positive formation enthalpy, it is interesting to investigate whether the newly predicted ordered d-BC3 is more energetically favorable than diamondlike disordered structure. We have used the special quasirandom structure (SQS) approach[103] to randomly distribute B and C in the diamond lattice with different n-atoms (n = 16, 32, 64, and 128). The results show that the enthalpies of all the quasirandom diamond-like structures are higher than those of d-BC3, indicating that the current d-BC3 is clearly much more stable than the disordered phase at zero temperature. Since the d-BC3 was synthesized under the condition of high-pressure and high-temperature, the temperature effect on energy should be taken into account. We have calculated the Gibbs free energies with the consideration of the contribution of vibrational and configurational entropy for d-BC3 and the 128-atom quasirandom BC3 structure. In any case, our calculations show that with or without the vibrational and configuration entropy contributions, the Gibbs free energy of predicted d-BC3 are more energetically favorable than that of quasirandom structure below 1166 K (Fig. 6), indicating that our predicted d-BC3 structure is more favorable for the experimentally recovered sample at atmospheric temperature than the disordered phase.

Fig. 6. Gibbs free energy per atom of quasirandom 128-atom BC3 structure relative to d-BC3 at 40 GPa.

The simulated hardness of BC3 reaches 62 GPa. We have performed first-principles calculations of the stress–strain relations, which provide an insight into the local bond deformation and breaking mechanisms that determine the incipient plasticity in a crystal.[104106] Its deformation modes under tensile and x shear strains show intriguing bond elongation and sequential bond-breaking processes that lead to remarkable extended ductility and elastic response. The [111] tensile stress decreases very gradually past the peak and the peak-to-valley drop extends over a wide range of tensile strain from 0.08 to 0.18, and the system undergoes a second elastic response regime from tensile strain 0.18 to 0.31. This behavior is highly unusual for a superhard material, and it is in stark contrast to the results for diamond and c-BN where the stress drops precipitously past the peak. The ratio between shear strength and tensile strength (53.0/52.5), which is nearly unity, is the lowest in superhard covalent materials, indicating the superior ductility of d-BC3 to those of top superhard materials. The results provide a crucial insight into the structural assignment for BC3 and clarify the actual existence of cubic symmetric phase for other boron carbide.

4. Summary

The review describes the theoretical prediction of the crystal structures for diamond-like covalent compounds formed by B–C–N–O system, which are the ideal targets in search for new superhard phases. Theoretical methods of designing new superhard materials are essential to improve the feasibility and effectiveness of experimental synthesis and/or find some extricable approaches. Material design techniques are greatly desirable to assist experiment, and free energy minimization methods and the recently developed superhard-driven methodology are quite successful for the applications to theoretically design superhard materials.

Reference
1Brazhkin VDubrovinskaia NNicol MNovikov NRiedel RSolozhenko VZhao Y 2004 Nat. Mater. 3 576
2Tian YXu BZhao Z 2012 Int. J. Refract. Met. Hard Mater. 33 93
3Kurakevych O O 2009 J. Superhard Mater. 31 139
4Xu BTian Y 2015 Sci. China Math. 58 132
5Wentorf R HJr 1957 J. Chem. Phys. 26 956
6Zhao YHe D WDaemen L LShen T DSchwarz R BZhu YBish D LHuang JZhang JShen GQian JZerda T W 2002 J. Mater. Res. 17 3139
7Kaner R BGilman J JTolbert S H 2005 Science 308 1268
8Knittle EKaner R BJeanloz RCohen M L 1995 Phys. Rev. 51 12149
9Solozhenko V LAndrault DFiquet GMezouar MRubie D C 2001 Appl. Phys. Lett. 78 1385
10Komatsu TNomura MKakudate YFujiwara S 1996 J. Mater. Chem. 6 1799
11Nakano SAkaishi MSasaki TYamaoka S 1994 Chem. Mater. 6 2246
12He J LTian Y JYu D LWang T SLiu S MGuo L CLi D CJia X PChen L XZou G T 2001 Chem. Phys. Lett. 340 431
13Solozhenko V LKurakevych O OAndrault DLe Godec YMezouar M 2009 Phys. Rev. Lett. 102 015506
14Hubert HGarvie L A JDevouard BBuseck P RPetuskey W TMcMillan P F 1998 Chem. Mater 10 1530
15Oganov A RChen J HGatti CMa Y ZMa Y MGlass C WLiu Z XYu TKurakevych O OSolozhenko V L 2009 Nature 457 863
16Solozhenko V LKurakevych O OOganov A R 2008 J. Superhard Mater. 30 428
17Godec Y LKurakevych O OMunsch PGarbarino GSolozhenko V L 2009 Solid State Commun. 149 1356
18Bullett D W 1982 J. Phys. C: Solid State Phys. 15 415
19Kurakevych O OSolozhenko V L 2007 Acta Crystallogr., Sect. C: Cryst. Struct. Commun. 63 i80
20Li QWang HMa Y M 2010 J. Superhard Mater. 32 192
21Zinin P VMing L CIshii H AJia RAcosta THellebrand E 2012 J. Appl. Phys. 111 114905
22Li QLiu HZhou DZheng W TWu ZMa Y 2012 Phys. Chem. Chem. Phys. 14 13081
23Zhang MLiu HDu YZhang XWang YLi Q 2013 PCCP 15 14120
24Li QMa Y MOganov A RWang H BWang HXu YCui TMao H KZou G T 2009 Phys. Rev. Lett. 102 175506
25Xia YLi QMa Y 2010 Comput. Mater. Sci. 49 S76
26Crowhurst J CGoncharov A FSadigh BEvans C LMorrall P GFerreira J LNelson A J 2006 Science 311 1275
27Gregoryanz ESanloup CSomayazulu MBadro JFiquet GMao HHemley R J 2004 Nat. Mater. 3 294
28Crowhurst J CGoncharov ASadigh BZaug JAberg DMeng YPrakapenka V B 2008 J. Mater. Res. 23 1
29Young A FSanloup CGregoryanz EScandolo SHemley R JMao H 2006 Phys. Rev. Lett. 96 155501
30Jiang CLin ZZhao Y 2009 Phys. Rev. Lett. 103 185501
31Ono SKikegawa TOhishi Y 2005 Solid State Commun. 133 55
32Tolbert S HCumberland R WClark S MGilman J JWeinberger M BKaner R B 2005 J. Am. Chem. Soc. 127 7264
33Gu QKrauss FSteurer W 2008 Adv. Mater. 20 3620
34Chung H YWeinberger M BLevine J BKavner AYang J MTolbert S HKaner R B 2007 Science 316 436
35Qin J QHe D WWang J HFang L MLei LLi Y JHu JKou Z LBi Y 2008 Adv. Mater. 20 4780
36Dubrovinskaia NDubrovinsky LSolozhenko V L 2007 Science 318 1550c
37Chung H YYanga J MTolbert S HKanerb R B 2008 J. Mater. Res. 23 1797
38Wang MLi YCui TMa YZou G 2008 Appl. Phys. Lett. 93 101905
39Knappschneider ALitterscheid CGeorge N CBrgoch JWagner NBeck JKurzman J ASeshadri RAlbert B 2014 Angew. Chem. Int. Ed. 53 1684
40Zhang XQin JXue YZhang SJing QMa MLiu R 2013 Phys. Status Solidi-R 7 1022
41Gou HDubrovinskaia NBykova ETsirlin A AKasinathan DSchnelle WRichter AMerlini MHanfland MAbakumov A M 2013 Phys. Rev. Lett. 111 157002
42Xie MMohammadi RMao ZArmentrout M MKavner AKaner R BTolbert S H 2012 Phys. Rev. 85 064118
43Niu HWang JChen X QLi DLi YLazar PPodloucky RKolmogorov A N 2012 Phys. Rev. 85 144116
44Wu HSun HChen C 2014 Appl. Phys. Lett. 105 211901
45Chen X QFu C LKrčmar MPainter G S 2008 Phys. Rev. Lett. 100 196403
46Li QZhou DZheng WMa YChen C 2013 Phys. Rev. Lett. 110 136403
47Wang HLi QLi Y EXu YCui TOganov A RMa Y M 2009 Phys. Rev. 79 132109
48Wang HLi QWang HLiu HCui TMa Y 2010 The Journal of Physical Chemistry C 114 8609
49Endo TSato TShimada M 1987 J. Mater. Sci. Lett. 6 683
50Grumbach M PSankey O FMcMillan P F 1995 Phys. Rev. 52 15807
51Oganov A RGlass C W 2006 J. Chem. Phys. 124 244704
52Glass C WOganov A RHansen N 2006 Comput. Phys. Commun. 175 713
53Oganov A RGlass C WOno S 2006 Earth Planet. Sci. Lett. 241 95
54Li QChen WXia YLiu YWang HWang HMa Y 2011 Diamond Relat. Mater. 20 501
55Li YLi QMa Y 2011 EPL 95 66006
56Li QWang MOganov A RCui TMa YZou G 2009 J. Appl. Phys. 105 053514
57Zhang XWang YLv JZhu CLi QZhang MLi QMa Y 2013 J. Chem. Phys. 138 114101
58Garvie L A JHubert HPetuskey W TMcMillan P FBuseck P R 1997 J. Solid State Chem. 133 365
59Bolotina N BDyuzheva T IBendeliani N A 2001 Crystallography Rep. 46 734
60Wang YLv JZhu LMa Y 2012 Comput. Phys. Commun. 183 2063
61Wang YLv JZhu LMa Y 2010 Phys. Rev. 82 094116
62Lv JWang YZhu LMa Y 2011 Phys. Rev. Lett. 106 015503
63He J LGuo L CWu ELuo X GTian Y J 2004 J. Phys.: Conden. Matter 16 8131
64Li QZhou DWang HChen WWu BWu ZZheng W 2012 Solid State Commun. 152 71
65Sun HJhi S HRoundy DCohen M LLouie S G 2001 Phys. Rev. 64 094108
66Pan ZSun HChen C 2005 J. Phys.: Condens. Matter 17 3211
67Zhang R QChan K SCheung H FLee S T 1999 Appl. Phys. Lett. 75 2259
68Mattesini MMatar S F 2001 Int. J. Inorg. Mater. 3 943
69Kim EPang TUtsumi WSolozhenko V LZhao Y 2007 Phys. Rev. 75 184115
70Sun JZhou X FQian G RChen JFan Y XWang H TGuo X JHe J LLiu Z YTian Y J 2006 Appl. Phys. Lett. 89 151911
71Luo XGuo XXu BWu QHu QLiu ZHe JYu DTian YWang H T 2007 Phys. Rev. 76 094103
72Zhou X FSun JFan Y XChen JWang H TGuo XHe JTian Y 2007 Phys. Rev. 76 100101
73Luo XGuo XLiu ZHe JYu DXu BTian YWang H T 2007 Phys. Rev. 76 092107
74Chen S YGong X GWei S H 2007 Phys. Rev. Lett. 98 015502
75Chen SGong X GWei S H 2008 Phys. Rev. 77 014113
76Chen C FSun H 2007 Phys. Rev. Lett. 99 159601
77Chen SGong X GWei S H 2007 Phys. Rev. Lett. 99 159602
78Luo XGuo XXu BWu QHu QLiu ZHe JYu DTian YWang H T 2007 Phys. Rev. 76 94103
79Luo XGuo XLiu ZHe JYu DXu BTian YWang H T 2007 Phys. Rev. 76 92107
80Li QWang JZhang MLi QMa Y 2015 RSC Adv. 5 35882
81Dubitskiy G ABlank V DBuga S GSemenova E EKulbachinskii V AKrechetov A VKytin V G 2005 JETP Lett. 81 260
82Ekimov E ASidorov V ABauer E DMel’nik N NCurro N JThompson J DStishov S M 2004 Nature 428 542
83Crespi V H 2003 Nat. Mater. 2 650
84Blase XBustarret EChapelier CKlein TMarcenat C 2009 Nat. Mater. 8 375
85Matar S FMattesini M2001Acad. Sci. Paris4255
86Liu HLi QZhu LMa Y 2010 Phys. Lett. 375 771
87Mikhaylushkin A SZhang XZunger A 2013 Phys. Rev. 87 094103
88Liu HLi QZhu LMa Y 2011 Solid State Commun. 151 716
89Gildenblat G SGrot S ABadzian A 1991 Proc. IEEE 79 647
90Jones L EThrower P A 1991 Carbon 29 251
91Ma YTse J SCui TKlug D DZhang LXie YNiu YZou G 2005 Phys. Rev. 72 014306
92Lee K WPickett W E 2004 Phys. Rev. Lett. 93 237003
93Takano YTakenouchi TIshii SUeda SOkutsu TSakaguchi IUmezawa HKawarada HTachiki M 2007 Diamond Relat. Mater. 16 911
94Calandra MMauri F 2008 Phys. Rev. Lett. 101 016401
95Liang YZhang WZhao JChen L 2009 Phys. Rev. 80 113401
96Lazar PPodloucky R 2009 Appl. Phys. Lett. 94 251904
97Moussa J ECohen M L 2008 Phys. Rev. 77 064518
98Li QWang HTian YXia YCui THe JMa YZou G 2010 J. Appl. Phys. 108 023507
99Gao FHe JWu ELiu SYu DLi DZhang STian Y 2003 Phys. Rev. Lett. 91 015502
100He JWu EWang HLiu RTian Y 2005 Phys. Rev. Lett. 94 015504
101Liu ZHe JYang JGuo XSun HWang H TWu ETian Y 2006 Phys. Rev. 73 172101
102Zhang MLiu HLi QGao BWang YLi HChen CMa Y 2015 Phys. Rev. Lett. 114 015502
103Zunger AWei S HFerreira L GBernard J E 1990 Phys. Rev. Lett. 65 353
104Zhang YSun HChen C F 2004 Phys. Rev. Lett. 93 195504
105Krenn C RRoundy DCohen M LChrzan D CMorris J WJr 2002 Phys. Rev. 65 134111
106Li Jvan Vliet K JZhu TYip SSuresh S 2002 Nature 418 307