|Home | About | Journals | Submit | Contact Us | Français|
Ninefold coordination of hydrogen is very rare, and has been observed in two different hydride complexes comprising rhenium and technetium. Herein, based on a theoretical/experimental approach, we present evidence for the formation of ninefold H- coordination hydride complexes of molybdenum ([MoH9]3−), tungsten ([WH9]3−), niobium ([NbH9]4−) and tantalum ([TaH9]4−) in novel complex transition-metal hydrides, Li5MoH11, Li5WH11, Li6NbH11 and Li6TaH11, respectively. All of the synthesized materials are insulated with band gaps of approximately 4 eV, but contain a sufficient amount of hydrogen to cause the H 1s-derived states to reach the Fermi level. Such hydrogen-rich materials might be of interest for high-critical-temperature superconductivity if the gaps close under compression. Furthermore, the hydride complexes exhibit significant rotational motions associated with anharmonic librations at room temperature, which are often discussed in relation to the translational diffusion of cations in alkali-metal dodecahydro-closo-dodecaborates and strongly point to the emergence of a fast lithium conduction even at room temperature.
Although exhibiting the simplest form, hydrogen has an exceptionally rich chemistry, forming various chemical bonds in materials1. The flexibility of hydrogen is the source of many interesting functionalities in hydrides, such as hydrogen storage2, superconductivity3,4,5,6, fast ionic conductivity7, magnetism8 and metal–insulator transition9,10; thus, materials containing a large amount of hydrogen are promising candidates not only for improving material performance but also for the emergence of latent functionalities.
Complex transition metal hydrides represent a class of hydrogen-rich materials, wherein a number of H atoms covalently bind to a transition metal element (T) to form hydride complexes with a remarkably rich variety of H-coordination modes ranging from twofold to ninefold1,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25. These complexes are stabilized by charge transfer from electropositive counterions, such as alkali and alkaline earth metals, to form insulating hydrides1,20,23. In general, the H-coordination number increases from right to left along a given period because of the trend of increasing atomic size, which allows more hydrogen binding. At the same time, despite research spanning the past several decades, the elements capable of forming hydride complexes are thought to be limited to those in Groups 7–1218. This limitation has so far precluded further discovery of hydride complexes with high H-coordination. For example, the presently known members with the ninefold H-coordination are limited to those comprising the Group 7 elements technetium ([TcH9]2−)11 and rhenium ([ReH9]2−)12,13,14,16,17. However, we recently reported the formation of the first Group 6 hydride complex [CrH7]5− in Mg3CrH825, thereby demonstrating that there is no a priori reason for this limitation and paving the way for the discovery of higher H-coordination hydride complexes comprising unexplored elements to the left of Group 7.
To demonstrate the feasibility of this strategy, we have experimentally examined the formation of novel transition metal hydride complexes of Group 5 and Group 6 elements (Mo, W, Nb and Ta) based on a first-principles prediction. Lithium was selected as a counterion because of its low electronegativity, which is expected to thermodynamically stabilize the complex hydride phases1,20,23. The crystal structures and electronic structures of the obtained samples were investigated using a combination of experimental and ab initio approaches. Finally, we discuss the prospects for the emergence of H-driven functionalities in the obtained materials.
We first performed a ground-state structure search of Li–Mo–H and Li–Nb–H systems using first-principles calculations. Figure 1 shows the obtained most stable structures found with stoichiometries of Li5MoH11 and Li6NbH11. In Li5MoH11, nine H atoms (H1) form a tricapped trigonal prism around the Mo atoms sitting at the (0, 0, 0) and (0, 0, 1/2) coordinates of the hexagonal lattice and two isolated H atoms (H2) are located at the approximate centres of the trigonal prisms formed by the six Mo atoms. The four Li atoms coordinate to the H2 atoms, forming a two-dimensional network of corner-sharing tetrahedra between the layers of MoH9 units. Li6NbH11 has a structure similar to that of Li5MoH11; however, it has an additional Li site between the NbH9 tricapped trigonal prisms along the c-axis, which accommodates the 6:1 stoichiometry between Li and Nb.
The calculated standard heats of formation are −510kJ/mol for Li5MoH11 and −637kJ/mol for Li6NbH11, which are much lower than those of the conceivable decomposed products, 5LiH+Mo+3H2 and 6LiH+NbH0.65+2.175H2, respectively; therefore, the materials could potentially be synthesized via the following reactions: 5LiH+Mo+3H2→Li5MoH11 and 6LiH+NbH0.65+2.175H2→Li5MoH11, with the enthalpies of reaction ΔHr of −105 and −91kJ/mol, respectively (note that we used −81kJ/mol26 for LiH and −60kJ/mol for NbH0.6527). We observed that the isostructural 5d analogues Li5WH11 and Li6TaH11 also have higher thermodynamic stabilities of −529 and −651kJ/mol with respect to the conceivable decomposition products 5LiH+W+3H2 and 6LiH+TaH0.65+2.175H0.65, respectively (note that the enthalpies of reaction are estimated to be −128 and −117kJ/mol, respectively, using −51kJ/mol for TaH0.6528). The results are summarized in Table 1.
Based on the obtained structures, we discuss the electronic structures. Figure 2 shows the calculated electronic densities of states (DOSs) and the H s and T spd projections. We checked for magnetic states and observed no magnetic ordering in any of the materials.
The most important feature of the electronic structures is that the 1s states of the nine H1 atoms forming the TH9 tricapped trigonal prisms reach the Fermi level (EF) in all materials, as observed in the third panels from the top in Fig. 2a–d. These states strongly hybridize with the T spd states to form the σ-bonds, which is evident from the H1 s and T spd characters of both the valence and conduction bands. These results are consistent with a previous electronic structure calculation by Singh et al.29 for BaReH9, which contains the hydride complex ReH9. Because all nine bonding states are fully occupied by 18 electrons per formula unit, the TH9 units can be considered as the hydride complexes [MoH9]3−, [WH9]3−, [NbH9]4− and [TaH9]4−. The 1s states of the isolated H2 atoms lie immediately below EF without noticeable hybridization with any other states, as observed in the second panels from the top in Fig. 2a–d. Thus, these atoms can be considered as H− ions. The electropositive Li states occur far above EF (not shown) and donate one electron per atom to the hydride complexes to become Li+ cations. Thus, we conclude that the ionic configurations of the current materials are [Li+]5[MoH9]3−[H−]2, [Li+]5[WH9]3−[H−]2, [Li+]6[NbH9]4−[H−]2 and [Li+]6[TaH9]4−[H−]2. The band gaps are approximately 4eV in all of the materials.
We experimentally examined the possibility of the formation the theoretically predicted complex transition metal hydrides using a high-pressure and high-temperature technique. Powdered mixtures of LiH and Group 5 and Group 6 transition metals were hydrogenated under 5GPa at 923–1023K for 24–48h. The colours of the samples recovered at ambient pressure and temperature were yellow for Li–Mo–H, greenish yellow for Li–W–H, light grey for Li–Nb–H and white for Li–Ta–H, implying the insulating character, as is typical with complex transition-metal hydrides.
The recovered samples were characterized by Raman spectroscopy, and the results are illustrated in Fig. 3a–d (black lines) along with the density functional theory (DFT) spectra (red lines) calculated using a method reported in the literature30. Because the primitive cells of Li5MoH11/Li5WH11 and Li6NbH11/Li6TaH11 comprise 34 and 36 atoms, respectively, there are a total of 99 (10A1+6A2+6B1+11B2+17E2+16E1) and 105 (11A1+6A2+6B1+12B2+18E2+17E1) zone-centre optical phonon modes, respectively, as represented by the tick marks in Fig. 3a–d. Among them, 76 modes (10A1+17E2+16E1) of Li5MoH11/Li5WH11 and 81 modes (11A1+18E2+17E1) of Li6NbH11/Li6TaH11 are Raman active. These modes provide an intense and asymmetric band at approximately 1660–1860cm−1 and a weak band at approximately 1000cm−1 in all the DFT spectra, which originate from the T–H stretching and H–T–H bending vibrations, respectively. The experimental spectra exhibit very similar features to those obtained by the DFT calculations, strongly supporting the formation of the theoretically predicted ninefold H-coordination hydride complexes.
Figure 3e–h shows X-ray diffraction (XRD) profiles of the recovered samples (black lines) and the profiles simulated based on the DFT structures (red lines). Although the samples contain small amounts of impurities, the Bragg peaks were well indexed to a similar hexagonal lattice with the lattice parameters listed in Table 2. While all of the Bragg peaks are considered to primarily originate from diffraction by the transition metals due to the small atomic scattering factors of H and Li, we successfully reproduced the profiles by placing transition metals at the (0, 0, 0) and (0, 0, 1/2) coordinates, which is similar to the DFT structures. The low concentrations of transition metals suggest the existence of large amounts of Li and H atoms along with transition metals.
The DFT lattice parameters summarized in Table 2 are slightly smaller than those determined from the XRD profiles measured at room temperature (4–10% smaller in volume), implying the presence of finite-temperature effects. As such, we examined the dynamics of the room-temperature phase of Li5MoH11 as an example using both first-principles molecular dynamics (FPMD) calculations in the isothermal–isobaric (NPT) ensemble at 298K and neutron diffraction (ND) measurements at room temperature (strictly, ND measuremts were performed on the deuteride analogue Li5MoD11).
In FPMD calculations, the two isolated H atoms (H2) isotropically fluctuate around their original positions determined by DFT calculations at 0K, whereas the nine H atoms in the MoH9 units (H1) rotate around the Mo atoms, generating spherical shell-like distribution, as illustrated in Fig. 4a. This type of rotational motion associated with anharmonic librations is acknowledged as leading to phase transitions in borohydrides31 and is thus expected to promote the volume expansion at room temperature. In fact, we observed a ~4% expansion (285Å3) relative to the ground-state volume obtained by DFT calculations at 0K (274Å3), as shown in Fig. 4b. Additionally, the FPMD trajectory reasonably reproduces the experimental ND profile in Fig. 4c, as observed by the comparison of the simulated profile from the trajectory (red line in Fig. 4d) with the Rietveld refinement fit (black line in Fig. 4d), where the P6cc structure (No. 184) with nine disordered D atoms surrounding the Mo atoms was employed, as shown in Fig. 4e, suggesting the adequacy of the current discussion. The experimentally determined structural parameters are summarized in Table 3.
We have examined the possibility of the formation of novel transition metal hydride complexes with high H-coordination using a combined theoretical/experimental approach. All experimental data strongly point to the successful syntheses of a series of theoretically predicted complex transition metal hydrides Li5MoH11, Li5WH11, Li6NbH11 and Li6TaH11 containing ninefold H-coordination hydride complexes and two H− anions. Our theoretical calculations reveal the σ-bonds of the nine H atoms forming the hydride complexes and the transition metals reach EF in all synthesized materials.
Generally, the electronic structures of complex transition-metal hydrides can be viewed as comprising low-lying H 1s- and high-lying T spd-derived states, as schematically depicted in Fig. 5. The ligand field induced by the H-coordination divides the T spd states into the nonbonding and antibonding states (σ*-bands), and EF falls in the ligand-field gap to accommodate the so-called 18-electron rule. Therefore, the valence bands around EF are normally dominated by the T spd-derived nonbonding states (strictly, the T d-derived nonbonding states); however, we emphasize here that the contribution from the H 1s-derived σ-bands in the valence bands increases with increasing H-coordination number, pushing the T spd-derived states out of the valence bands and reaching EF at ninefold coordination (Fig. 5). This is a novel strategy for obtaining solid-state covalently bonded hydrides with the H 1s-derived states reaching EF.
As is well known, it is indispensable to have ‘σ-bands’ of light elements at EF to achieve high-critical-temperature (high-Tc) superconductivity within the Bardeen–Cooper–Schrieffer (BCS) theory32. For example, the relatively high Tc observed in MgB233 is primarily attributed to the high-lying two-dimensional σ-bands derived from the in-plane B–B bonds, the holes of which strongly couple with the bond stretch phonons34. Conversely, in all of the materials synthesized in this study, nine of the lightest H atoms form strong σ-bonds with transition metal elements, providing a DOS just below EF. In this context, the occurrence of stronger electron–phonon coupling may be expected if the band gaps close under compression. From our preliminary calculations, we observed that there is at least one stable metallic phase of Li5MoH11 above 94GPa (see Supplementary information).
Besides the electronic feature, these materials have a characteristic phonon property such that the ninefold hydride complexes exhibit rotational motions associated with anharmonic librations at room temperature. This type of motion has never been observed in the complex transition metal hydrides, but has often been discussed in relation to the occurrence of fast ion conduction in alkali metal dodecahydro-closo-dodecaborates such as Na2B12H12. In this material, the first-order phase transition from the low-temperature monoclinic to the high-temperature cubic phase occurs near 520K, which is accompanied by an increase of two orders of magnitude in the rate of reorientational jumps of the icosahedral [B12H12]2− anions with the fast translational diffusion of Na+ ions35. Although we observed no obvious translational diffusion of Li+ ions in our FPMD calculations because of the limited simulation time (12ps) and the unit cell size (136 atoms), the significant rotational motions of [MoH9]3− ions, as observed in Fig. 4a, suggest the potential fast lithium ion conduction even at room temperature. It should be noted that given the quantum mechanical tunnelling of the H atoms, the rotational motions of hydride complexes may be sustained at extremely low temperature at which thermal diffusion is supressed.
Although the hydride complexes with ninefold H-coordination were discovered in 196418, they are very rare and their properties have not been fully clarified. Therefore, our findings will facilitate the discovery of latent functionalities, such as those described in this study and will lead to renewed interest in complex transition metal hydrides.
The ground-state crystal structures of the Li–Mo–H, Li–W–H, Li–Nb–H and Li–Ta–H systems were explored using the structures of existing compounds as references. This exploration was performed using first-principles DFT calculations as implemented in the Vienna ab initio simulation package (VASP)36,37. We used a plane-wave basis and the projector augmented wave method38,39 within the generalized gradient approximation with the Perdew–Burke–Ernzerhof exchange-correlation functional40. For Li5MoH11, we assumed that the structure has a metal framework similar to that of CaNi5 because of the similarity in their metal stoichiometries (Ca and Ni were replaced by Mo and Li, respectively, in the doubled CaNi5 structure along the c-axis). We used nine H atoms (18H atoms per formula unit) to coordinate the Mo atoms in the fashion of a tricapped trigonal prism, and placed two H atoms (four H atoms per formula unit) in the largest tetragonal site comprising four Li atoms. For Li6NbH11, an additional Li atom (two Li atoms per formula unit) was placed at the centre of the largest trigonal antiprism comprising six H atoms in Li5MoH11 to accommodate the 6:1 stoichiometry between the Li and transition metals. We assumed that Li5WH11 and Li6TaH11 have isostructures to Li5MoH11 and Li6NbH11, respectively. Next, full structure relaxations, including those for the lattice parameters, shapes and internal coordinates, were performed. Subsequently, phonon calculations were performed to verify that the relaxed structures were at the true minimum. When imaginary phonon frequencies were observed, we slightly displaced the atoms along the directions of the eigenvectors of the imaginary modes and further relaxed the structures to eliminate them. This procedure was performed until the ground state was reached. This was done using well-converged plane-wave basis sets with a cut-off energy of 800eV. An 8×8×4 grid was used for the k-point sampling of the Brillouin zone.
The room-temperature phase of Li5MoH11 was examined using FPMD calculations in the NPT ensemble at 298K with a 2×2×1 supercell (136 atoms) of the primitive cell. We employed a well-converged cut-off energy of 600eV with k-point sampling at the Γ-point. The simulation temperature was set to 298K, which was controlled by a Langevin thermostat41 with a friction coefficient of 10ps−1 for all atoms. The simulation pressure was controlled by a Parrinello–Rahman barostat42,43, where a friction coefficient of 10ps−1 were used for the lattice degrees of freedom. The total simulation time was 12ps with a timestep of 0.5fs.
The integrated time-of-flight (TOF) ND peak intensity I(Kn) at the FPMD step n was calculated from the FPMD trajectory using the following equation:
where Kn is the reciprocal lattice vector, L(Kn) is the Lorentz factor, bj is the neutron-scattering length of atom j, and rj,n is the position of atom j. The ND profile was obtained by integrating the discrete ND peaks broadened by a Gaussian function with a width of 0.02Å over a range from 0.8fs to 12ps, where the unit cell volume is well converged (see Fig. 4b).
High-purity metal powders of Mo (99.99%, Sigma-Aldrich), W (99.99%, Sigma-Aldrich), Nb (99.9%, Mitsuwa Chemicals Co., Ltd.) and Ta (99.9%, Mitsuwa Chemicals Co., Ltd.) were mixed with LiH (95%, Sigma-Aldrich) at molar ratios of 1:6 (Li–Mo–H and Li–W–H) and 1:8 (Li–Nb–H and Li–Ta–H) by mechanical milling at 400rpm for 4h under a 0.1-MPa Ar atmosphere. The mixtures were compacted into pellets (diameter and thickness=1.0mm) and encapsulated in sample capsules comprising pyrolytic boron nitride. The capsules were loaded together with a hydrogen source (AlH3) into NaCl capsules, which were used to seal the hydrogen fluid that evolved from AlH3. The NaCl capsules were pressurized to 5GPa at room temperature using a multi-anvil high-pressure apparatus, followed by subsequent heating at 923K (Li–Nb–H), 973K (Li–Mo–H) and 1023K (Li–W–H and Li–Ta–H) for 24h (Li–W–H and Li–Ta–H) and 48h (Li–Mo–H and Li–Nb–H). A similar procedure was used to synthesize the deuteride analogue Li5MoD11, where LiD and AlD3 were used in place of LiH and AlH3.
Neutron scattering experiments of Li5MoD11 (41mg) were performed for an exposure time of 8h at room temperature under He on the NOVA high-intensity total diffractometer installed at the 500kW spallation neutron source at the Materials and Life Science Experimental Facility (MLF) Japan Proton Accelerator Research Complex (J-PARC). The sample was filled in a cylindrical sample container with an external diameter of 3.0mm and thickness of 0.1mm. Rietveld analysis was performed using the General Structure Analysis System (GSAS) software and the graphical interface EXPGUI (version 1.80)44 on the ND data, with d=0.25Å–6.16Å obtained at room temperature from detectors of medium resolution (Δd/d~0.6%) at the scattering angle 2θ=72°–108°. The profile function developed by Von Dreele et al.45 was used for performing Rietveld analysis. The background was modeled using the 10-terms Chebyschev polynomial function model in GSAS. The distances of Mo–D in the complex anion [MoD9]3− and Li–D in the tetrahedral site in Li5MoD11 were soft constrained to 1.75Å. The constraints were weighted such that they imposed only a minor deterioration of the fit to the experimental data. ND peaks of an unknown phase were indexed by a monoclinic unit cell with a≈9.79Å, b≈7.03Å, c≈5.55Å and β≈105.7° using the indexing programs TREOR9746 and PIRUM47. Based on hkl extinction rules, the space group for the unknown phase was predicted to be P21/m (No. 11). Because it was not possible to determine the atomic positions in the unknown phase, Le Bail fitting was simultaneously performed for the unknown phase in Rietveld analysis.
How to cite this article: Takagi, S. et al. Formation of novel transition metal hydride complexes with ninefold hydrogen coordination. Sci. Rep. 7, 44253; doi: 10.1038/srep44253 (2017).
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
We are grateful for helpful discussions with K. Aoki, technical support from R. Hiraoka and H. Omiya, and the use of the SR16000 supercomputing resources at the Center for Computational Materials Science of the Institute for Materials Research, Tohoku University. The neutron scattering experiment was approved by the Neutron Scattering Program Advisory Committee of IMSS, KEK (Proposal No. 2014S06). The high-pressure hydrogenation technique was developed by using synchrotron radiation X-rays at BL14B1, SPring-8 (Proposal No. 2014B3602). This work was supported by JSPS KAKENHI (Grant Numbers 16H06119, 25220911, 16K06766 and 25420725), the Photon and Quantum Basic Research Coordinated Development Program by MEXT and Collaborative Research Center on Energy Materials in IMR (E-IMR).
The authors declare no competing financial interests.
Author Contributions S.T. conceived this study, performed the theoretical calculations, analysed the data and wrote the manuscript. Y.I. prepared the samples with immense help from H.S. and performed the XRD and Raman measurements. T.S. analysed the XRD data and prepared AlH3. K.I. and T.O. performed the ND measurements. T.I. provided significant help in analysing the data. K.M. calculated the theoretical Raman spectra. S.O. designed and conducted the project. All authors commented on the final manuscript and conclusions of this work.