Hypervalent hydridosilicate in the Na–Si–H system

Hydrogenation reactions at gigapascal pressures can yield hydrogen-rich materials with properties relating to superconductivity, ion conductivity, and hydrogen storage. Here, we investigated the ternary Na–Si–H system by computational structure prediction and in situ synchrotron diffraction studies of reaction mixtures NaH–Si–H2 at 5–10 GPa. Structure prediction indicated the existence of various hypervalent hydridosilicate phases with compositions NamSiH(4+m) (m = 1–3) at comparatively low pressures, 0–20 GPa. These ternary Na–Si–H phases share, as a common structural feature, octahedral SiH6 2− complexes which are condensed into chains for m = 1 and occur as isolated species for m = 2, 3. In situ studies demonstrated the formation of the double salt Na3[SiH6]H (Na3SiH7, m = 3) containing both octahedral SiH6 2− moieties and hydridic H−. Upon formation at elevated temperatures (>500°C), Na3SiH7 attains a tetragonal structure (P4/mbm, Z = 2) which, during cooling, transforms to an orthorhombic polymorph (Pbam, Z = 4). Upon decompression, Pbam-Na3SiH7 was retained to approx. 4.5 GPa, below which a further transition into a yet unknown polymorph occurred. Na3SiH7 is a new representative of yet elusive hydridosilicate compounds. Its double salt nature and polymorphism are strongly reminiscent of fluorosilicates and germanates.


Detailed description of experiments and data analysis
All steps of sample preparation were performed in a glove box under argon atmosphere.Powdered NaH (Sigma Aldrich, 90%) and powdered Si (325 mesh, 99.999% (metals basis)), Thermo Scientific) were carefully mixed at a molar ratio of 1:1 and 2:1 (NaH:Si) and compressed into pellets with an outer diameter (OD) of 2 mm and 0.9 -1.1 mm height.Ammonia borane (BH 3 NH 3 , Sigma Aldrich, 90%) was used as hydrogen source since it has a well-defined decomposition behavior at high pressures and produces chemically inert BN as residual (Nylén et al., 2009).The amount of BH 3 NH 3 used for each sample corresponded to a approx.4× molar excess of H 2 with respect to Si. NaH/Si sample pellets were sandwiched between pelletized BH 3 NH 3 and sealed inside NaCl capsules.The salt capsules had 3.0 mm OD and ~3.6 mm height.
High pressure experiments were performed at starting pressures of approx.5 and 9 GPa and employed 14/7 multianvil assemblies.A detailed description of the 14/7 setup is provided elsewhere (Vekilova et al., 2023).In addition, 2 mm OD amorphous SiBCN rods and ~5.5 mm wide rectangles made of either MgO or amorphous BCN epoxy (TU Freiberg) were used as X-ray windows in the octahedra and gaskets, respectively, along the beam direction.Filled MgO octahedra were positioned between eight truncated tungsten carbide cubes (32 mm, grade TF09, Fujilloy Co. Ltd.) fitted with pyrophyllite gaskets.Assemblies were compressed at a rate of 1 bar/min oil pressure (~3.2-3.7 GPa/h) to the target pressures and heated in the Voggenreiter-built modified-cubic press at beamline ID06-LVP, ESRF (Guignard and Crichton, 2015).The heating was performed at various rates, ranging from ~10-20 °C/min (comparatively fast) at T < 550 °C to ~5 °C/min (comparatively slow) at higher temperatures.The heating was arrested each time the release of hydrogen from BH 3 NH 3 was expected or the growth of ternary Na-Si-H materials was detected.Pressure was estimated in situ from powder X-ray diffraction (PXRD) patterns using the equation of state (EOS) of NaCl by Matsui et al (Matsui et al., 2012).Temperature was evaluated from power -T calibration curves obtained by reproducing in situ runs offline using analogous 14/7 setups equipped with central C-type thermocouple.The effect of pressure on the thermocouple EMF in the studied range of p, T is expected to be negligible (Nishihara et al., 2020).
Angle-dispersive PXRD patterns were collected continuously in 1.27 -15.26° 2θ range at a constant wavelength (λ = 0.233933 Å, E ≈ 53.0 keV), selected by the Si111 double-crystal monochromator from the emission of a U18 cryoundulator at ~6 mm magnetic gap.Data were acquired using the Pilatus3X-900 kW CdTe high-resolution 2D detector which was built and developed by DECTRIS specifically for the ID06-LVP beamline.Acquisition rate per dataset typically varied from every 60 sec during the compression to every second during the heating.Sample-to-detector distance and detector offset were calibrated using LaB 6 -SRM660a (NIST).The in situ data were integrated, visualized and manipulated using the Fit2D software (Hammersley, 2016).Indexing of the powder patterns was performed using DICVOL and TAUP algorithms within the CRYSFIRE package (Shirley, 2004).For least-squares refinement of unit cell parameters the UnitCell software was employed (Holland and Redfern, 1997).
2. Specific description of a 9 GPa run (2:1 NaH:Si composition) The sample was heated upon reaching target pressure (~8.6 GPa).The plot showing temperature variations during the experiment is provided in Figure S1.At ~300 °C hydrogen release from NH 3 BH 3 source was expected to be completed.During a ~12 min temperature dwell at ~380 °C a set of broad low intensity reflections appeared.The peaks could be indexed to a primitive hexagonal unit cell (a ≈ 4.75 Å, c ≈ 7.65 Å).However, this evaluation is rather approximate due to the weak intensity and diffuse nature of the reflections.Above ~560 °C the hexagonal phase was promptly replaced by another new set of reflections which were indexed to a primitive tetragonal unit cell (with a highest applicable space group being P4/mbm (№127).Shortly after the beginning of the growth (~610 °C) the lattice parameters of the tetragonal phase were estimated to a ≈ 6.49 Å, c ≈ 4.73 Å, while at highest temperature (~850 °C) they corresponded to a ≈ 6.59 Å, c ≈ 4.78 Å. Symmetry and cell parameters of the tetragonal phase strongly hinted at the isostructural relation to the double salts with K 3 SiF 7type structure, suggesting Na 3 SiH 7 composition.Interestingly, even upon heating to 850 °C the growth of the tetragonal Na 3 SiH 7 was very sluggish while noticeable amounts of unreacted NaH and Si were still present, which is attributed to a poor reactivity of elemental Si.Also, at this temperature the diffraction pattern of Na 3 SiH 7 became noticeably textured.The sample was then slowly cooled (~5 °C/min) from ~850 °C to ~715 °C, and then at ~25 °C/min to RT. Upon cooling below 130 °C a (second-order) continuous phase transition of tetragonal Na 3 SiH 7 to an orthorhombic Pbam polymorph was observed.The estimated pressure after cooling was ~6.2 GPa, and the cell parameters of the Pbam phase were approximated as: a ≈ 9.165 Å, b ≈ 9.289 Å, c ≈ 4.758 Å.
The sample was decompressed over the course of ~12 hours.The phase transition of Pbam-Na 3 SiH 7 to a yet unidentified low pressure (LP) polymorph started below ~4.5 GPa.The estimated cell parameters of Pbam phase at the onset of transformation were: a ~ 9.29 Å, b ~ 9.42 Å, c ~ 4.81 Å.Shortly after, at ~4.2 GPa, the pressure dropped rapidly to ~0.5 GPa, which obscured further characterization of LP phase.The sample was recovered at ambient conditions, however, its PXRD characterization was not conclusive.It is not clear whether Na 3 SiH 7 is recoverable to ambient conditions.Figure S1.Temperature variations as a function of time for the NaH-Si-H 2 system at ≈9 GPa (2:1 NaH:Si ratio).Filled circles and dropped dashed lines mark the temperature and time points when the growth of tetragonal (P4/mbm) Na 3 SiH 7 and its transition to the orthorhombic (Pbam) polymorph were detected.Orange and blue colors correspond heating and cooling, respectively.

Specific description of a 5 GPa run (2:1 NaH:Si composition)
The pressure before the start of the heating corresponded to ~5.1 GPa.The plot showing temperature variations during the experiment is provided in Figure S2.The sample was annealed at ~430 °C for approx.20 min for equilibration after the hydrogen release.Afterwards the sample was heated to ~515 °C, and the heating was arrested for ~ 70 min.No growth of tetragonal Na 3 SiH 7 was detected during this time.It was realized that at this point pressure had decreased to ~4.5 GPa.During the next 2 hours the pressure was increased to ~6.9 GPa (while keeping the power to the heater constant).The growth of tetragonal phase was seen at 510 °C above 5.2 GPa (cell parameters at the start of the growth were approx.: a ~ 6.65 Å, c ~ 4.83 Å).At ~6.9 GPa and ~490 °C its cell parameters corresponded (approximately) to a ~ 6.55 Å, c ~ 4.77 Å (the reflections were still very weak, which hindered more precise evaluation).Afterwards the temperature was carefully increased to ~840 °C while the Na 3 SiH 7 growth proceeded very sluggishly, very similar to the 9 GPa experiment.At ~770 °C the cell parameters of tetragonal Na 3 SiH 7 were refined to a = 6.59134(15)Å, c = 4.78027(15) Å.The weight percentage of Na 3 SiH 7 , NaH and Si at this point was estimated to be ~17.6 %, ~49.9% and ~32.5%, respectively.
From ~840 °C the sample was cooled at a rate of ~20 °C/min.The phase transition to Pbam-Na 3 SiH 7 polymorph occurred below 120 °C.The pressure corresponded to ~5.2 GPa after cooling, and cell parameters of Pbam phase at this pressure were estimated to be a ≈ 9.227 Å, b ≈ 9.357 Å, c ≈ 4.757 Å.The sample was decompressed over the course of ~9 hours.The transition to low pressure polymorph could be observed more clearly, as compared to the 9 GPa experiment.The onset of transformation started below ~4.3 GPa.Cell parameters of the Pbam phase at this point corresponded to a ≈ 9.28 Å, b ≈ 9.41 Å, c ≈ 4.79 Å.
The intermediate hexagonal phase seen in the 9 GPa run was not observed at 5 GPa.However, the additional unidentified phase, growing in parallel with P4/mbm-Na 3 SiH 7 (marked in Figure 3, main text) is also observed in the 5 GPa experiment, see Figure S4, and, as in the 9 GPa run, the compound decomposed above ~830 °C.
The 5 GPa run was reproduced using a sample of a different composition (1:1 NaH:Si molar ratio).However, changing ratio of the components did not have any significant effect on the reaction outcome or its kinetics.Instead of slow cooling the tetragonal Na 3 SiH 7 was temperature-quenched which resulted in the instantaneous transformation to the Pbam phase.
Figure S2.Temperature and pressure variations as a function of time for the NaH-Si-H 2 system in a 5 GPa run (2:1 NaH:Si ratio).Filled circles and dropped dashed lines mark the growth of tetragonal (P4/mbm) Na 3 SiH 7 and its transition to orthorhombic (Pbam) polymorph.Orange and blue colors correspond heating and cooling, respectively.

Strucure characterization of P4/mbm-Na 3 SiH 7
Le Bail fitting (Le Bail et al., 1988) and Rietveld refinement (Rietveld, 1969) against the in situ data were performed in Jana2006 (Petříček et al., 2014).PXRD patterns used for the analysis were acquired during the 5 GPa (NaH:Si=2:1) run after the second pressure increase (temperature dwell at ~770 C°, cf. Figure S2) where the pressure corresponded to ~7.2 GPa according to the EOS of NaCl (Matsui et al., 2012).At this point of the experiment the content of tetragonal Na 3 SiH 7 sample was sufficient for fitting while the texture due to active recrystallization was not yet significant.The data were prepared for analysis by averaging ~130 sequential patterns in order to diminish possible effect of texture on the intensities as well as to improve signal-to-noise ratio.In addition, the observed intensities were corrected for background by subtracting the lowest non-zero value from the I obs column.The structural model for the tetragonal Na 3 SiH 7 used in the Rietveld refinement (SG №127, P4/mbm) corresponded to the calculated structure relaxed at 10 GPa (Table S5).The refined parameters included background (10 th degree Chebyshev polynomials in combination with manually assigned points), zero shift, unit cell dimensions, peak profile parameters (corresponding to pseudo-Voigt function), scale factor, atomic coordinates for Na atoms at 4h Wyckoff site and ADPs for Na and Si atoms.The ADPs of Na atoms in Na 3 SiH 7 structure were constrained together.The atomic coordinates and ADPs of the hydrogen atoms remained fixed during the refinement.The U iso (H) were assigned an arbitrary value of 0.038 Å 2 (B iso ≈ 3.0 Å 2 ).In addition, the following phases were included in the refinement: NaH (SG №225), Si (SG №227), NaCl (SG №225).Peaks originating from assembly materials, some of the NaCl peaks and the zero-intensity regions arising from the gaps between the detector chips were excluded from the refinement.Results of the Rietveld analysis for the tetragonal Na 3 SiH 7 are shown in Tables S1 and S2 as well as in Figure S3.S1-2).Splitting of its reflections upon transition to Pbam phase is visible on cooling below 120 °C.Peak of an additional unidentified phase growing in parallel with Na 3 SiH 7 and disappearing above ≈830 °C is marked with a blue asterisk.This phase is also seen in the 9 GPa experiment, cf. Figure 3.

II. Theoretical calculations.
Figure S5.Phonon dispersions for Na m SiH (4+m) compounds at 0 and 10 GPa (and zero Kelvin).At ambient pressure Pbam Na 3 SiH 7 has a negative formation enthalpy (with respect to 3NaH + Si + 2H 2 ) but is not dynamically stable (not shown).At ambient pressure Na 2 SiH 6 (as P6 3 mc) is slightly above the convex hull and, in addition, are not dynamically stable.

Figure S1 .
Figure S1.Temperature variations as a function of time for the NaH-Si-H 2 system at ≈5 GPa.

Figure S2 .
Figure S2.Temperature variations as a function of time for the NaH-Si-H 2 system at ≈9 GPa.

Figure S7 .
Figure S7.Electronic band structure and density of states of P4/mbm-Na 3 SiH 7 at ambient pressure.

Figure S3 .
Figure S3.PXRD patterns of P4/mbm-Na 3 SiH 7 (orange, magnified ×1.2), silicon (purple) and sodium hydride (blue) simulated from the results of Rietveld fit of the corresponding phases to the PXRD data (black) collected in situ at ID06-LVP, ESRF, during the 5 GPa run (2:1 NaH:Si starting mixture) upon temperature dwell at ~770 °C (λ = 0.233933 Å).Diffraction peaks corresponding to assembly (MgO), most of NaCl reflections, artifacts and zerointensity regions arising from the gaps between the detector chips, and a weak peak of an unidentified impurity (marked with an asterisk) were excluded from the refinement.Excluded regions are marked by gray rectangles in the difference plot.

Figure S4 .
Figure S4.Compilation of PXRD patterns (λ = 0.233933 Å) acquired during the hydrogenation of 2NaH:1Si mixture in 5 GPa experiment (observed intensities are shown on logarithmic scale).Overlain is the simulated pattern of P4/mbm-Na 3 SiH 7 at ≈700 °C (shown in red, linear scale) based on the refined model (TableS1-2).Splitting of its reflections upon transition to Pbam phase is visible on cooling below 120 °C.Peak of an additional unidentified phase growing in parallel with Na 3 SiH 7 and disappearing above ≈830 °C is marked with a blue asterisk.This phase is also seen in the 9 GPa experiment, cf.Figure3.

Table S1 .
Results of the Rietveld refinement of the P4/mbm-Na 3 SiH 7 structure at ≈7.2 GPa,

Table S2 .
Fractional coordinates and atomic displacement parameters obtained for the P4/mbm-Na 3 SiH 7 at ≈7.2 GPa, ≈770 °C.Hydrogen atom positions correspond to the DFToptimized structure, see TableS5d.The coordinates and ADPs of H atoms remained fixed during the refinement.

Table S4 .
Relaxed structures for P6 3 mc and P3m1 Na 2 SiH 6 at various pressures