Stable Calcium Nitrides at Ambient and High ... - ACS Publications

Jul 18, 2016 - ... University of Saskatchewan, Saskatoon, Saskatchewan S7N 5E2, Canada. #. Canadian Light Source, Saskatoon, Saskatchewan S7N 2V3, ...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/IC

Stable Calcium Nitrides at Ambient and High Pressures Shuangshuang Zhu,† Feng Peng,*,‡,§ Hanyu Liu,⊥ Arnab Majumdar,∥ Tao Gao,*,† and Yansun Yao*,∥,# †

Institute of Atomic and Molecular Physics, Sichuan University, Chengdu 610065, China College of Physics and Electronic Information, Luoyang Normal University, Luoyang 471022, China § Beijing Computational Science Research Center, Beijing 10084, China ⊥ Geophysical Laboratory, Carnegie Institution of Washington, 5251 Broad Branch Road NW, Washington, D.C. 20015, United States ∥ Department of Physics and Engineering Physics, University of Saskatchewan, Saskatoon, Saskatchewan S7N 5E2, Canada # Canadian Light Source, Saskatoon, Saskatchewan S7N 2V3, Canada ‡

S Supporting Information *

ABSTRACT: The knowledge of stoichiometries of alkalineearth metal nitrides, where nitrogen can exist in polynitrogen forms, is of significant interest for understanding nitrogen bonding and its applications in energy storage. For calcium nitrides, there were three known crystalline forms, CaN2, Ca2N, and Ca3N2, at ambient conditions. In the present study, we demonstrated that there are more stable forms of calcium nitrides than what is already known to exist at ambient and high pressures. Using a global structure searching method, we theoretically explored the phase diagram of CaNx and discovered a series of new compounds in this family. In particular, we found a new CaN phase that is thermodynamically stable at ambient conditions, which may be synthesized using CaN2 and Ca2N. Four other stoichiometries, namely, Ca2N3, CaN3, CaN4, and CaN5, were shown to be stable under high pressure. The predicted CaNx compounds contain a rich variety of polynitrogen forms ranging from small molecules (N2, N4, N5, and N6) to extended chains (N∞). Because of the large energy difference between the single and triple nitrogen bonds, dissociation of the CaNx crystals with polynitrogens is expected to be highly exothermic, making them as potential high-energy-density materials.



INTRODUCTION Solid nitride materials have attracted great attention because of their intriguing properties, such as superconductivity,1 high energy density,2 and superior hardness,3 as well as extraordinary chemical and thermal stability.4 Alkaline-earth metal nitrides, in particular, have shown growing applications in recent years, for example, as nitriding agents for the formation of other nitrides,5−7 as catalysts for the synthesis of superhard materials,8 and as additives in the refinement process of steel.9 In the alkaline nitride family, calcium nitride is unique because it is the only member that can have both ionic (Ca3N2)10−12 and subnitride (Ca2N)13,14 forms. At ambient conditions, the ground-state compound is Ca3N2 consisting of Ca2+ and N3−. Ca3N2 has three crystalline forms, α (Ia3̅), β (R3̅c), and γ (Pbcn) phases, formed at different temperatures. In 1968, a metastable Ca2N form was synthesized by heating pure Ca and Ca3N2 to temperatures above 1300 K.13 Ca2N was later found to be a two-dimensional electride, i.e., an unusual ionic material with electrons serving as anions.14 This novel property of Ca2N could potentially lead to applications in electronics and photonics, which, as can be expected, generated enthusiasm in the study of calcium nitrides. In 2012, another metastable form, CaN2, was synthesized by controlled decomposition of calcium azide at high-pressure/high-temperature conditions.9 CaN2 contains double-bonded [N2]2− anions © XXXX American Chemical Society

that are similar to deprotonated N2H2 and was once considered as an energy material. Indeed, the family of calcium nitrides is fast expanding because of the development of synthesis methods. The versatility of calcium and nitrogen in forming chemical bonds surely allows for a large diversity of stoichiometries and crystalline structures in this group; many of them are perhaps unknown. The focus of previous investigations has been primarily on the calcium-rich side of this group. The nitrogenrich nitrides have been relatively less explored, both experimentally and theoretically. Compared with their calcium-rich counterparts, the nitrogen-rich nitrides are electron-deficient, which facilitates the formation of bonding between nitrogen atoms. By modifying the concentration of calcium (electron donors), one may manipulate the bonding pattern and enable single- and double-bonded polynitrogen molecules in these compounds. Such compounds may be utilized as high-energy-density materials because of the large amount of energy stored in the polynitrogens. The present study therefore aims to address the structures and properties of CaN compounds using state-of-the-art theoretical methods. Theoretical studies have often been used Received: April 19, 2016

A

DOI: 10.1021/acs.inorgchem.6b00948 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

ΔHf(CaNx) = [H(CaNx) − H(Ca solid) − xH(N solid)]/(1 + x). The energetically most favorable structures for solid calcium (face-centered-cubic, body-centered-cubic, and β-tin phases) and solid nitrogen (α, γ, ε, and cubic gauche phases) were used as reference structures in their corresponding stable pressure ranges. The thermodynamically stable phases at each pressure are represented by solid squares connected by the convex hull (solid lines), whereas the unstable or metastable phases are shown as open squares above the convex hull. The complete pressure−composition phase diagram of CaN compounds is shown in Figure 2, with the stable structures

to identify new materials prior to their realization. Recent advances of the structure search methods, in particular, made it possible to systematically investigate the phase diagram of unknown compounds. Through a survey of the free energy landscape, we predict the existence of several hitherto unknown but stable CaN phases at ambient and high pressures, ranging from the calcium-rich Ca2N to the nitrogen-rich CaN6, and provide suggestions on the synthesis conditions. Six novel stoichiometric CaN compounds with interesting structures were predicted that might be synthesized at high pressure and recovered to ambient conditions. A set of metastable polynitrogen forms were discovered, in these crystalline structures, including variants of diazenes ([N2]4−, [N2]2−, and [N2]1−), tetrazadiene ([N4]4−), pentazole ([N5]2−), hexazine ([N6]4−) and extended chains (N∞). Some of the polynitrogen molecules contain a high content of single N−N bonds, which may lead to applications as high energy carriers. The bonding nature revealed in these novel nitrides is of great importance to nitrogen chemistry and to the understanding of metal−nitrogen interactions.



METHODS

The search for thermodynamically stable crystalline structures was made on various stoichiometries of CaNx (x = 1/2, 2/3, 1, 3/2, 2, 3, 4, 5, and 6) using simulation cells containing up to four formula units. Structure searches for all stoichiometries were carried out at multiple pressures, e.g., 0, 20, 50, and 100 GPa at 0 K, using the particle swarm optimization methodology, as implemented in the CALYPSO code.15,16 Total energy calculations, geometrical optimizations, and electronic structure calculations were performed within the framework of density functional theory using the Vienna Ab Initio Simulation Package (VASP) program.17 Exchange and correlation of the electrons were treated by the generalized gradient approximation with the Perdew−Burke−Ernzerhof functional.18 The projector-augmentedwave19 method was employed with the calcium and nitrogen potentials adopted from the VASP potential library. The calcium and nitrogen potentials have 3s23p64s2 and 2s22p3 as valence states, respectively, and use an energy cutoff of 600 eV for the plane-wave basis set. A dense kpoint grid20 with a spacing of 2π × 0.03 Å−1 was used to sample the Brillouin zone, which was shown to yield excellent convergence for total energies (within 1 meV/atom).

Figure 2. Predicted pressure−composition phase diagram of CaN crystalline phases.

identified using colors and space groups. Note that most structures undergo phase transitions below 100 GPa, the highest studied pressure, except for CaN5 and Ca2N3. It is very significant, and in fact surprising, that a new stoichiometry CaN is revealed to be stable at ambient pressure, along with the three known phases, Ca2N, CaN2, and Ca3N2. CaN undergoes three structural phase transitions within 100 GPa. The formation of CaN is very unusual because calcium has one less valence electron than gallium but forms the same stoichiometries as GaN. An obvious question to address here is how CaN fulfills the “octet rule”. Yet, this situation is even more common at high pressure, such that four new stoichiometries, Ca2N3, CaN3, CaN4, and CaN5, were found to become stable in varying pressure ranges (Figure 2). Specifically, Ca2N3 is predicted to be stable in the pressures range of 18−44 GPa; CaN3 is stable above 8 GPa; CaN5 is stable above 33 GPa; CaN4 is stable between 5 and 66 GPa with one phase transition at 20 GPa. The known stoichiometries, Ca2N, CaN2, and Ca3N2, were all predicted to have more than one phase transition below 100 GPa. The mechanical and dynamical stabilities of all newly predicted structures are confirmed by the calculated phonon dispersion relationships (see the Supporting Information (SI), Figures S1−16). The crystal structures of the predicted CaN phases are presented in Figure 3a−n, whereas the optimized parameters are provided in the SI. In what follows, these new structures will be categorized by their structure motifs and analyzed in line with potential energy storage capabilities. Crystalline CaN Phases with N3− Anions. In the calciumrich end, the Ca2N crystal adopts a layered rhombohedral R3̅m structure (Figure 3a, with structure parameters in Table S1) at ambient pressure. In this structure, nitrogen is fully reduced to the N3− anion with complete subshells (confirmed by Bader charge analysis). The N3− anion is often seen in alkali-metal nitrides, for example, in Li3N or Cs3N.21,22 Compared with alkali-metal nitrides, however, Ca2N has one more valence electron; thus, the formation of such a structure seems to violate the “octet rule”. The Pauli Exclusion Principle states that



RESULTS AND DISCUSSION The calculated enthalpies of formation, ΔHf, for the energetically most favorable CaNx structures obtained in the structure search at varying pressures are presented in Figure 1. ΔHf was calculated for each stoichiometry with respect to solid calcium and nitrogen, using a fractional representation CaN x ,

Figure 1. Relative enthalpies of formation of CaN phases with respect to solid calcium and nitrogen. The convex hulls for stable phases (solid squares) are connected by solid lines. Unstable/metastable phases are indicated by open squares. B

DOI: 10.1021/acs.inorgchem.6b00948 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

observations.23 This structure is formed by Ca2+ cations and N3− anions. Upon compression, the cubic Ia3̅ structure is predicted to transform to an orthorhombic Pmn21 structure (Figure 3d) above 11 GPa and then to a hexagonal P3̅m1 structure (Figure 3e) above 76 GPa. Our searched phase transformation sequence of Ca3N2 is consistent with the results of Hao et al.12 Ca3N2 is a typical ionic compound with insulator character (Figure S20). Under high pressure, the band gap of the P3̅m1 phase was found to reduce and eventually close at 100 GPa (Figure S21). Crystalline CaN Phases with Anionized N2. The newly predicted CaN phase adopts a monoclinic C2/m structure (Figure 3f) at ambient conditions. In this structure, nitrogen atoms are equally grouped into two forms, the fully reduced N3− anions and double-bonded [N2]2− anions. [N2]2− is analogous to deprotonated diazene N2H2. Previously, this anion has been synthesized in SrN2, BaN2, and CaN2 using a specialized autoclave system. In the C2/m structure, the [N2]2− anion has a bond length of 1.23 Å (at 0 GPa), which is very close to the value in N2H2 (1.25 Å). The calculated Mayer bond order (MBO)24 for the [N2]2− anion is 2.32, also is consistent with that of a double-bonding description. At high pressure, CaN is predicted to transform to an orthorhombic Cmc21 phase (Figure 3g) at 14 GPa, a monoclinic C2/m structure at 40 GPa, and another orthorhombic Pbam structure (Figure 3i) at 76 GPa. Notably, in these three high-pressure structures, the nitrogen atoms are grouped into 4-fold negatively charged and single-bonded [N2]4−. The [N2]4− anion, which is analogous to deprotonated hydrazine N2H4, has been previously synthesized in noble metal pernitrides MNM4+[N2]4− (MNM = Os, Ir, Pd, and Pt), which exhibit remarkable properties including superconductivity and high hardness. In the three high-pressure structures of CaN, the bond lengths of [N2]4− are 1.52, 1.48, and 1.49 Å (calculated at 20 GPa), very similar to the single bond length in N2H4 (1.45 Å). The calculated MBO values for the three structures are 1.67, 1.80, and 1.21, respectively, also consistent with the single-bond description. Compared with the [N2]2− anion, [N2]4− is stabilized by two additional electrons in the 2pπ* antibonding orbitals, similar to the electron configuration of a F2 molecule. The anionized N2 was also found in the hexagonal P6̅2m structure of Ca2N3 between 18 and 44 GPa (Figure 3j). In this structure, however, the N−N bond has mixed singleand double-bond character with a calculated bond length of 1.32 Å (at 20 GPa) and a calculated MBO value of 1.44. All CaN phases discovered here contain single and/or double nitrogen bonds. Because of the large energy difference between these bonds and the triple bond, these CaN compounds are potentially energy carriers because recently synthesized CaN2 revealing the possibility of forming energetic polynitrogen forms in this family of CaN compounds,9 which are expected to release a large amount of energy along with the triple-bonded N2 during the decomposition process. Crystalline CaN Phases with Polynitrogens. The CaN2 compound was recently synthesized by a controlled decomposition of Ca2N3 at high pressure and high temperature. The synthesized CaN2 has a tetragonal I4/mmm structure,9 which was later found to be a metastable phase. The theoretical ground state of CaN2 has a Pnma structure that is more than 0.08 eV/formula units (fu) lower in energy than the I4/mmm structure at ambient conditions.25 The Pnma and I4/mmm structures both consisted of [N2]2− and Ca2+ but differ in their crystal stacking. In the present study, the Pnma structure is

Figure 3. Crystalline structures of the predicted stable CaN phases. (a) R3m ̅ structure of ̅ structure of Ca2N at ambient pressure. (b) Fm3m Ca2N at 50 GPa. (c) I4/mmm structure of Ca2N at 100 GPa. (d) Pmn21 structure of Ca3N2 at 50 GPa. (e) P3m ̅ 1 structure of Ca3N2 at 80 GPa. (f) C2/m structure of CaN at ambient pressure. (g) Cmc21 structure of CaN at 20 GPa. (h) C2/m and CaN at 50 GPa. (i) Pbam structure of CaN at 100 GPa. (j) P6̅2m structure of Ca2N3 at 20 GPa. (k) P21/c structure of CaN2 at 50 GPa. (l) Pbam structure of CaN2 at 100 GPa. (m) Pmma structure of CaN3 at 20 GPa. (n) C2/c structure of CaN3 at 100 GPa. (o) P4/mbm structure of CaN4 at 20 GPa. (p) P41212 structure of CaN4 at 100 GPa. (q) Pm structure of CaN5 at 50 GPa. The smaller spheres are nitrogen atoms, and the larger ones are calcium atoms.

this “additional” electron needs to occupy empty quantum states. In the R3̅m structure, the edge-sharing NCa6 octahedra are connected to form ionic (Ca2+)2N3− layers that stack along the c axis. These layers are separated by large van der Waals gaps that accommodate the additional electrons (e−). Electrons in the van der Waals gaps form quantized states and interact with the environment in the same way as anions, resulting in the formation of an electride ground state, (Ca2+)2N3−·e−. Thus, the R3̅m structure of Ca2N possesses two-dimensional electric conductibility parallel to the van der Waals gaps. Upon compression, the layered structure transforms to a threedimensional framework. We predict that the R3m ̅ structure transforms to a face-centered-cubic Fm3̅m structure (Figure 3b) above 10 GPa and then to a tetragonal I4/mmm structure (Figure 3c) at 84 GPa. Both the Fm3̅m and I4/mmm structures maintain the electride properties. The edge-sharing NCa8 hexahedra and NCa10 dodecahedra are formed in these two structures, respectively, leading to three-dimensional ionic [NCa2]+ networks with the additional electrons residing in the interstitial voids (Figure S17). Both phases show threedimensional electric conductibility due to the fact that the excess electron per formula unit can move freely. Similar to the calcium solid, the s → d transition has also been identified in Ca2N under high pressure by the electronic density of states (DOS; Figure S18). Ca3N2 was found to adopt a cubic Ia3̅ structure at ambient pressure, which is in good agreement with the experimental C

DOI: 10.1021/acs.inorgchem.6b00948 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry predicted to transform to the I4/mmm structure at 3 GPa, a monoclinic P21/c structure (Figure 3k) at 18 GPa, and then to an orthorhombic Pbam structure (Figure 3l) at 92 GPa. Along with these transitions, the [N2]2− anions are brought closer to each other by high pressure and rearranged into [N4]4− anions in the P21/c and Pbam structures. Previously, we have predicted similar [N4]4− anions in the high-pressure structures of CsN. In the [N4]4− anion, the terminal atoms are bonded to the internal atoms by a single bond (σ), leaving other electrons to lone pairs. The two internal atoms are double-bonded by a σ bond plus a π bond. The Lewis structure of the [N4]4− anion therefore has one more single bond than the neutral N4, also noting here that these molecules may stabilize several resonance structures. The calculated N−N distances for the single and double bonds in CaN2 are 1.29 and 1.37 Å (calculated at 20 GPa), respectively, very close to the standard bond lengths for the single and double bonds (1.25 and 1.45 Å). The calculated MBO values (1.67 and 1.01) confirmed this interpretation. The newly predicted CaN3 phase was calculated to become thermodynamically stable above 8 GPa (Figure 2). At 8 GPa, CaN3 has an orthorhombic Pmma structure (Figure 3m) consisting of Ca2+ and [N2]2−. The calculated bond length and MBO value in the Pmma structure are 1.19 Å and 2.10, respectively, consistent with a double-bonding description. At high pressure, the monoclinic C2/c structure (Figure 3n) is predicted to become more stable than the Pmma structure at 36 GPa. Along with this phase transition, the N2 units are rearranged into more compact [N6]4− anions. The [N6]4− anion has a chair conformation (Figure 4), with nearly identical

substantially to a nonplanar conformation without aromaticity. In the crystalline environment, however, the [N6]4− anion is substantially stabilized by the cation−π interaction. In the C2/c structure (Figure 3n), for example, the Ca2+ cations reside along the 6-fold axis of the [N6]4− anion and interact favorably with the delocalized π electrons above and below the anion planes. In a previous study,27 the binding energy between the Ca2+ cation and [N6]4− moiety in a Ca2N6 complex was calculated to be −1928.8 kcal/mol, indicating such an arrangement as energetically much favored. As a potential high-energy-density material, dissociation of CaN3 is highly exothermic because of the existence of weak nitrogen bonds in the [N6]4− anion. The dissociation process of the [N6]4− anion in cation−π systems has been previously studied in a Ca2N6 complex in which two Ca2+ ions are located on both sides of the [N6]4− anion along the 6-fold axis.27 Dissociation of this system involves two steps, (1) the separation of calcium from the N6 and (2) the simultaneous breaking of three N−N bonds in N6, and produces three N2. The activation barriers for these two steps were calculated (at the B3LYP/6-311+G* level) as 24.6 and 7.9 kcal/mol, respectively, which indicates a possible kinetic stable region for crystalline phases.27 The CaN5 phase is predicted to form at above 33 GPa in a Pm structure (Figure 3q), which is stable to at least 100 GPa. In this structure, the nitrogen atoms are grouped into slightly puckered [N5]2− anions (Figure 5a), whereas the calcium atoms

Figure 5. Calculated electronic properties of the Pm phase of CaN5 at 50 GPa. (a) Isolated planar [N5]2− anion in the Pm structure. (b) ELF map shown on the (0, 1, 0) cross section of the Pm structure. (c) Electronic band structure and projected DOS of the Pm structure.

Figure 4. Calculated electronic properties of the C2/c phase of CaN3 at 50 GPa. (a) Isolated chaired N64− anion in the C2/c structure. (b) ELF map shown on the (16.5, 1, 9.38) cross section of the C2/c structure. (c) Electronic band structure and projected DOS of the C2/ c structure. Numbers in part a next to the N−N bond represent the bond length (black, in angstroms) and MBO value (red).

serve as electron donors. Within the anion plane, the nitrogen atoms are bonded by five σ bonds with five lone pairs outside the atoms (Figure 5b). Above and below the plane, seven π electrons form a set of π bonds that are equally distributed among the five nitrogen atoms. Calculated N−N distances in the [N5]2− anion are different, e.g., 1.33, 1.38, and 1.36 Å at 50 GPa, which are between the experimental bond lengths of single (1.45 Å) and double (1.20 Å) bonds.28 Thus, the anion is distorted from the ideal D5h symmetry. Calculated MBO values range between 1.4 and 1.5. It is worth mentioning that such a seven-π-electron five-membered ring does not obey Hückel’s rule; thus, the [N5]2− anion is intrinsically unstable. It is only stabilized in an extremely strained crystal under high pressure. Phonon calculations suggested that the CaN5 compound is

N−N distances of 1.35 Å. The calculated MBO values in the [N6]4− anion are between 1.17 and 1.36, which are stronger than the single bond (1.0) but weaker than the double bond (2.0). Recently, a similar [N6]4− anion has been identified in the high-pressure structures of XeN6.26 The N6 molecule and its anions are fundamentally interesting in line with the significant interest in potential energy storage applications. The ground-state N6 molecule has an open-chain diazide structure that is consistent with a N3−−N3+ adduct. The [N6]4− anion, on the other hand, adopts a ring structure but has to distort D

DOI: 10.1021/acs.inorgchem.6b00948 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

sharing in the chain. CaN4 is promising for this proposal because of the relatively low ionization potential of calcium. Electrons acquired from calcium in the CaN4 crystal reduce the amount of electrons shared by nitrogen atoms, which ultimately reduces the bonding order in the chain. In addition, the anonized nitrogen chains also interact favorably with Ca2+, which is expected to be more stable than their neutral counterparts.

unstable at ambient pressure. Thus, this serves as an interesting example that common chemical intuition, which was developed at normal conditions, can break down at extreme conditions. Because [N5]2− has one additional electron compared to its aromatic counterpart,21,22 CaN5 has a metallic ground state (Figure 5c). The electronic DOS at the Fermi level consists primarily of nitrogen p states with moderate mixing with the s states. The CaN4 phase is predicted to form at pressures above 5 GPa (Figure 2), initially in the P4/mbm structure with dinitrogen units (Figure 3o), and then transform to the P41212 structure (Figure 3p) at 19 GPa. The P41212 phase is particularly interesting: it is the only CaN structure discovered so far that contains a nonmolecular form of nitrogen. In this structure, the nitrogen atoms form one-dimensional bent chains that extend infinitely into the crystal (Figure 6a). The smallest



CONCLUSION In the present study, we carried out swarm intelligence structure searches to investigate the possible formation of new calcium polynitrides at ambient and high pressures. The goal is to discover stable stoichiometries and structures of CaNx containing energetic polynitrogen forms and the conditions for their formation. Along with the known Ca3N2, CaN2, and Ca2N, we predicted five new stoichiometries, namely, Ca2N3, CaN, CaN3, CaN4, and CaN5, to also become stable provided suitable pressure conditions. In particular, the CaN phase is predicted to be stable already at ambient pressure, which may be synthesized using CaN2 and Ca2N as precursors. In the predicted crystals, the calcium atoms act as electron donors and modify the amount of electron sharing in the N−N bond. By adjustment of the ratio of calcium in the stoichiometry, a series of polynitrogen forms were achieved in the crystalline CaNx phases, including small clusters made of a few atoms (N2, N4, N5, and N6) to extended chains (N∞). Because of the large energy difference between the single and triple nitrogen bonds, dissociation of the CaNx crystals with polynitrogens is expected to be highly exothermic, which may potentially lead to applications as energy carriers. The present study provides new insight into the understanding of polynitrogens and encourages experimental exploration into the synthesis of these promising materials.

Figure 6. Calculated electronic properties of the P41212 phase of CaN4 at 20 GPa. (a) Isolated one-dimensional N∞ chain in the P41212 structure. (b) ELF map shown on the N∞ plane of the P41212 structure. (c) Electronic band structure and projected DOS of the P41212 structure.



ASSOCIATED CONTENT

S Supporting Information *

repeating unit of this chain contains four nitrogen atoms in a nominal [N4]2− state. In its Lewis structure, an extending [N4]2− unit should have three single bonds and one double bond, i.e., [−NNNN−]2−, which is likely to be stabilized by several resonance structures. The calculated MBOs of the four independent bonds in the chain are similar, i.e., 1.29, 1.38, 1.38, and 1.29, respectively, consistent with the resonance nature of the structure. The four calculated independent bond distances in the chain are 1.33, 1.32, 1.32, and 1.33 Å, respectively. The electron localization function (ELF)29 map (Figure 6b) also confirms the nearly equal bonding nature of [N4]2−. The band structure of the P41212 phase (Figure 6c) reveals that this structure is metallic, with significant entanglements between the states and a continuous DOS in the energy space, which is consistent with the polymerization of nitrogen. In our recent study, we predicted a similar chain structure of nitrogen in a high-pressure phase of CsN2, where the smallest repeating unit was [N8]4−. The two chain structures in CaN4 and CsN2 have similar energy content, in terms of the ratio of single bonds; however, the pressure required to synthesize CaN4 is much lower than that for CsN2, by nearly 20 GPa, and it is well within the current capability of high-pressure techniques. Prior to these studies, polymeric nitrogen phases were primarily studied in pure nitrogen. The disadvantage of pure nitrogen chains is that the ratio of single bonds to double bonds is fixed at 1:1. The present study suggests that it is possible to increase the content of the single bonds by reducing the amount of electron

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.inorgchem.6b00948. Additional crystal structure information, electronic properties, phonon dispersion curves, band structures, and electronic DOS for various stoichiometries under different pressures (PDF)



AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected] (F.P.). *E-mail: [email protected] (T.G.). *E-mail: [email protected] (Y.Y.). Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS We thank the Natural Science Foundation of China (Grant 11304141), China Postdoctoral Science Foundation (Grant 2016M590033), and the Natural Sciences and Engineering Research Council of Canada for a Discovery Grant, and work at the Carnegie Institution of Washington was supported by EFree, an Energy Frontier Research Center funded by the DOE, Office of Science, Basic Energy Sciences, under Award DE-SC-0001057 (salary support for H.L). The infrastructure E

DOI: 10.1021/acs.inorgchem.6b00948 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

(28) Eremets, M. I.; Gavriliuk, A. G.; Trojan, I. A.; Dzivenko, D. A.; Boehler, R. Nat. Mater. 2004, 3, 558−563. (29) Becke, A. D.; Edgecombe, K. E. J. Chem. Phys. 1990, 92, 5397.

and facilities used at the Carnegie Institution of Washington were supported by NNSA Grant DE-NA-0002006. This work is also sponsored by the Program for Science and Technology Innovation Research Team in University of Henan Province (Grant 13IRTSTHN020), the Program for Science and Technology Innovation Talents in University of Henan Province (Grant 17HASTIT015), the Key Scientific and Technological Project of Henan Province (Grant 152102210307), and the Open Project of the State Key Laboratory of Superhard Materials (Jilin University; Grant 201602). Some of the calculations were performed at the High Performance Computation Training and Research Facilities at the University of Saskatchewan.



REFERENCES

(1) Papaconstantopoulos, D. A.; Pickett, W. E.; Klein, B. M.; Boyer, L. L. Phys. Rev. B: Condens. Matter Mater. Phys. 1985, 31, 752−761. (2) Raza, Z.; Pickard, C.; Pinilla, C.; Saitta, A. Phys. Rev. Lett. 2013, 111, 235501. (3) Knittle, E.; Wentzcovitch, R. M.; Jeanloz, R.; Cohen, M. L. Nature 1989, 337, 349−352. (4) Meng, Y.; Mao, H.-K.; Eng, P. J.; Trainor, T. P.; Newville, M.; Hu, M. Y.; Kao, C.; Shu, J.; Hausermann, D.; Hemley, R. J. Nat. Mater. 2004, 3, 111−114. (5) Bruls, R. J.; Hintzen, H. T.; Metselaar, R. J. Mater. Sci. 1999, 34, 4519−4531. (6) Parkin, I. P.; Nartowski, A. M. Polyhedron 1998, 17, 2617−2622. (7) Kobashi, M.; Okayama, N.; Choh, T. Mater. Trans., JIM 1997, 38, 260−265. (8) Lorenz, H.; Kühne, U.; Hohlfeld, C.; Flegel, K. J. Mater. Sci. Lett. 1988, 7, 23−24. (9) Schneider, S. B.; Frankovsky, R.; Schnick, W. Inorg. Chem. 2012, 51, 2366−2373. (10) Braun, C.; Börger, S. L.; Boyko, T. D.; Miehe, G.; Ehrenberg, H.; Höhn, P.; Moewes, A.; Schnick, W. J. Am. Chem. Soc. 2011, 133, 4307−4315. (11) Römer, S. R.; Schnick, W.; Kroll, P. J. Phys. Chem. C 2009, 113, 2943−2949. (12) Hao, J.; Li, Y. W.; Wang, J. S.; Ma, C. L.; Huang, L. Y.; Liu, R.; Cui, Q. L.; Zou, G. T.; Liu, J.; Li, X. D. J. Phys. Chem. C 2010, 114, 16750−16755. (13) Keve, E. T.; Skapski, A. C. Inorg. Chem. 1968, 7, 1757−1761. (14) Gregory, D. H.; Bowman, A.; Baker, C. F.; Weston, D. P. J. Mater. Chem. 2000, 10, 1635−1641. (15) Wang, Y.; Lv, J.; Zhu, L.; Ma, Y. Comput. Phys. Commun. 2012, 183, 2063−2070. (16) Wang, Y.; Lv, J.; Zhu, L.; Ma, Y. Phys. Rev. B: Condens. Matter Mater. Phys. 2010, 82, 094116. (17) Kresse, G.; Furthmüller, J. Phys. Rev. B: Condens. Matter Mater. Phys. 1996, 54, 11169. (18) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett. 1996, 77, 3865. (19) Blöchl, P. E. Phys. Rev. B: Condens. Matter Mater. Phys. 1994, 50, 17953. (20) Monkhorst, H. J.; Pack, J. D. Phys. Rev. B 1976, 13, 5188−5192. (21) Peng, F.; Yao, Y.; Liu, H.; Ma, Y. J. Phys. Chem. Lett. 2015, 6, 2363−2366. (22) Peng, F.; Han, Y.; Liu, H.; Yao, Y. Sci. Rep. 2015, 5, 16902. (23) Heyns, A. M.; Prinsloo, L. C.; Range, K.-J.; Stassen, M. J. Solid State Chem. 1998, 137, 33−41. (24) Mayer, I. Chem. Phys. Lett. 1983, 97, 270−274. (25) Wang, H.; Yao, Y.; Si, Y.; Wu, Z.; Vaitheeswaran, G. J. Phys. Chem. C 2014, 118, 650−656. (26) Peng, F.; Wang, Y.; Wang, H.; Zhang, Y.; Ma, Y. Phys. Rev. B: Condens. Matter Mater. Phys. 2015, 92, 094104. (27) Duan, H.; Gong, Z.; Cheng, J.; Zhu, W.; Chen, K.; Jiang, H. J. Phys. Chem. A 2006, 110, 12236−12240. F

DOI: 10.1021/acs.inorgchem.6b00948 Inorg. Chem. XXXX, XXX, XXX−XXX