betaFeedback

WorksPhysics and Astronomy

Structure and vibrational properties of methane up to 71 GPa

Abstract

Single-crystal synchrotron X-ray diffraction, Raman spectroscopy, and first principles calculations have been used to identify the structure of the high-pressure (HP) phase of molecular methane above 20 GPa up to 71 GPa. The structure of HP phase is trigonal R3, which can be represented as a distortion of the cubic phase B, previously documented at 7-15 GPa and confirmed here. The positions of hydrogen atoms in HP phase have been 19 obtained from first principles calculations. The molecules occupy four different crystallographic sites in phases B and eleven sites in the HP phase, which result in splitting of molecular stretching modes detected in Raman spectroscopy and assigned here based on 22 a good agreement with the Raman spectra calculated from the first principles.

Introduction

Methane represents a rare and fundamental example of hydrogen-bearing molecular compound with no hydrogen bonds unlike other simple molecules HF, H2O\mathrm{H_2O}, H2S\mathrm{H_2S}, and NH3\mathrm{NH_3}. Thus, the crystal structure formation in methane solids is expected to be mainly determined by the steric effects and van der Waals bonds, suggesting that its behavior at high pressure may be similar to the rare gas solids, for example Ar [1], which persists in a face-centered cubic (fcc) structure to very high pressures. Furthermore, the high-pressure behavior of methane is of interest to explore possible deviations from simple closed packed structures and related to this molecular orientational ordering, where the hydrogen atoms find the most energetically favorable stable positions. The fate of methane at even higher compression states in the regime of molecular proximity (i.e. shortened intermolecular H-H distances) is of great interest as it can be considered as a hydrogen dominant alloy [2] with the relevance to high-temperature superconductivity [3]. In addition, the behavior of methane at high pressures and temperatures may be of interest because it is believed to be a valued component of the deep planetary interiors of “icy” giant planets such as Uranus and Neptune [4]. Such hot molecular ices of water and ammonia form superionic crystals with diffusive protons at extreme pressure-temperature (P-T) conditions of relevance for these planets [5, 6, 7], while methane tends to chemically react yielding oligomeric hydrocarbon compounds [8] and dissociate at higher P-T conditions yielding diamond and hydrogen [9, 10, 11].

Similar to Ar, methane crystallizes in phase I at about 1.5 GPa upon compression at room temperate forming a face-centered cubic (fcc) closed packed structure of the molecular centers at the carbon atoms [12], and this crystal is plastic - molecules are freely rotating. However, unlike in noble gases, this structure transforms at 5.4 GPa to a complex rhombohedral structure A with 21 molecules in the unit cell solved by a combination of single crystal neutron and X-ray diffraction (XRD) techniques [13]. The structure is rhombohedrally distorted from the fcc methane I, with the carbon atoms occupying nine different crystallographic positions, and some of them may be associated with orientationally ordered molecules. However, the Raman and IR spectra do not show the obvious splitting of the major ν1\nu_1 and ν3\nu_3 C-H stretching modes [14].

Further compression leads to a transition to phase B with a very large unit cell containing 58 molecules, which has been found cubic as solved recently by Maynard-Casely et al. [15] for the carbon subsystem by combining powder and single-crystal XRD. The phase B shows a splitting of ν1\nu_1 and ν3\nu_3 stretching C-H modes in IR and Raman spectra [1, 14, 16, 17], suggesting that it possesses some elements of orientational order (e.g. Refs. [14, 16]). The transition from phase A to phase B at about 8 GPa is very sluggish and phase B can be missed if fast compressed yielding a simple cubic phase pre-B [16], where the Raman active ν1\nu_1 and ν3\nu_3 C-H stretching modes are not split suggesting that methane molecules are orientationally disordered. Phase B is reported to transform to a high pressure phase 1 (HP1) at about 25 GPa manifested by anomalies in IR spectra, vibrational frequencies [18], and specific volume [19], while no drastic structural changes were detected in the powder diffraction patterns. At higher pressures, Raman spectroscopy observations suggest further structural modifications to HP2 and HP3 phases reported up to 62 GPa; however, the transition pressures to the subsequent high-pressure phases HP2 and HP3 are contradictory [16, 17]. The structure of the HP phase(s) has not been solved; it has been postulated to be cubic (e.g. Ref. 20), but no detailed structural information is available. The XRD experiments have been extended to 202 GPa and the volume has been determined based on a cubic symmetry, even though a major transformation has been detected and a volume discontinuity was found at 94 GPa [20]. Methane was reported to persist in a molecular crystal structure beyond 200 GPa [20, 21].

In this paper, we examine the structure of methane in the pressure range between 7 and 71 GPa using single-crystal XRD, Raman spectroscopy, and first-principles theoretical calculations. Our single-crystal XRD data directly determine the structure of the carbon subsystem in agreement with the previous report in phase B [15], while we find that HP phase can be viewed as rhombohedrally distorted phase B (space group R3R3) with 87 molecules in the unit cell (29 in the primitive unit cell). We applied ab initio random structure search [22] of the most stable structures with and without any restriction of the symmetry. The results yield the structure of HP phase in agreement with XRD data, and determine the hydrogen positions in HP phase. The Raman spectrum of this phase calculated from the first principles is in a qualitative agreement with the experiment demonstrating that the C-H stretching mode splitting is due to a variation in the crystal field at different crystallographic sites.

Experimental and theoretical procedures

Methane gas was loaded in a mixture with He gas at approximately 1:1 mole ratio compressed to 0.15 GPa at room temperature in a BX90 diamond anvil cell [23] (Figure S1 of the Supplemental Material [24]). In order to obtain good-quality single crystals of phase B of methane, the gas mixture was slowly compressed at 297 K through the methane solidification point at 1.5 GPa and then to 8 GPa. At this pressure, excellent single crystals of phase B were grown overnight judging on Raman observations of a splitting of the ν1\nu_{1} C-H stretch mode (Figure S2 of the Supplemental Material [24]). These crystals yield very sharp Bragg reflections with typical rocking-curve width of 0.64º at 7.1 GPa (Figure 1). Two samples (#1 and #2) were compressed up to 39 GPa and 71 GPa respectively and concomitant XRD and Raman spectroscopy measurements (Figure 1, Figure 2 and Fig. S2 of the Supplemental Material [24]) were taken at selected pressure points. All the experiments were performed at room temperature (295(2) K).

Reconstructed reciprocal lattice planes of methane at 7.07, 35.6, and 63.2 GPa

Figure 1. Reconstructed reciprocal lattice plane of methane visualized by CrysAlisPro software package[25] at 7.07 GPa (left), 35.6 GPa (middle) and 63.2 GPa (right). D stands for diamond reflections. The observed diffraction spots from the sample have been indexed and used to determine the structure of phases B (left panel) and high-pressure (HP) phase (middle and right panel) (Table S1 and Table 1, Table 2, respectively).

Raman spectra of the C-H stretching modes of methane in HP phase at various pressures

Figure 2. Raman spectra of the C-H stretching modes of methane in HP phase at various pressures. The dots are the data and the lines are the fits to the Voigt oscillator model with eight peaks; variable color lines show the individual ν1\nu_{1} and ν3\nu_{3} components.

Sample chamber with methane crystals surrounded by He medium

Figure S1. Sample chamber with methane crystals surrounded by He medium.

Raman spectrum of methane at various pressures

Figure S2. Raman spectrum of methane at various pressures. Methane was grown in phase B, identified via splitting of the symmetric C-H stretch mode ν1\nu_1 at 3010 cm−1^{-1}. Methane is in phase B at 8 and 15 GPa and transforms to HP1 phase at higher pressures. Asterisk marks the appearance of the ν1\nu_1 further peak splitting, manifesting the transition to HP1 phase. The collection time is 10 s and the laser power is 5 mW except the spectrum measures at 8 GPa, where 0.5 mW laser power was used. The excitation wavelength is 488 nm for 8 GPa.

X-ray diffraction was performed at the beamlines 16-ID-B (APS, Argonne, USA) and P02.2 (Petra III, DESY, Hamburg, Germany). The following beamline setups were used. P02.2: λ=0.289 A˚\lambda= 0.289\,\mathring{A}, beam size ∼2×2 μm2\sim2 \times2\,\mu\mathrm{m}^{2}, Perkin Elmer XRD 1621 detector; 16-ID-B: λ=0.344 A˚\lambda= 0.344\,\mathring{A}, beam size ∼3×3 μm2\sim3 \times3\,\mu\mathrm{m}^{2}, Pilatus 1M detector. For the single-crystal XRD measurements samples were rotated around a vertical ω\omega-axis in a range ±22∘\pm22^\circ. The diffraction images were collected with an angular step Δω=0.5∘\Delta\omega= 0.5^\circ and an exposure time of 2-10 s/frame. For analysis of the single-crystal diffraction data (indexing, data integration, frame scaling and absorption correction) we used the CrysAlisPro software package[25]. To calibrate an instrumental model in the CrysAlisPro software, i.e., the sample-to-detector distance, detector’s origin, offsets of goniometer angles, and rotation of both X-ray beam and the detector around the instrument axis, we used a single crystal of orthoenstatite ((Mg1.93Fe0.06)(Si1.93,Al0.06)O6( \mathrm{Mg}_{1.93}\mathrm{Fe}_{0.06})(\mathrm{Si}_{1.93},\mathrm{Al}_{0.06})\mathrm{O}_{6}, PbcaPbca space group, a=8.8117(2)a = 8.8117(2), b=5.18320(10)b = 5.18320(10), and c=18.2391(3) A˚c = 18.2391(3)\,\mathring{A}). The same calibration crystal was used at both beamlines. The structure was solved with the ShelXT structure solution program[26] using intrinsic phasing and refined with the Jana 2006 program [27]. Raman spectra were collected using the GSECARS Raman System[28] with the excitation wavelength of 532 nm and at the Earth and Planets Laboratory of Carnegie (EPL) using a 488 nm excitation line (e.g. (e.g. Ref. 29)). Pressure was determined based on the shift of the R1 ruby fluorescence line [30].

Ab Initio Random Structures Searches (AIRSS) [31] were performed using the CASTEP code [22] to explore the structures of the HP methane with the rhombohedral symmetry (the unit cells with up to 87 atoms were explored), with a subsequent relaxation using density functional theory (DFT) calculations within the CASTEP code [22]. Raman spectra were computed using density functional perturbation theory (DFPT) + Finite differences method, using the default CASTEP 18 NC pseudopotentials with a 1250 eV energy cutoff, and a similarly dense k-point grid.

Results and discussion

At pressures below ∼16\sim16 GPa, the diffraction pattern of CH4\mathrm{CH}_4 (Figure 1) can be indexed with a body-centered cubic unit cell of phase B (Table S1 and Figure S3 [24]). Structure solution and refinement is in agreement with the model reported by Maynard-Casely et al. [15]. Due to weak scattering of hydrogen atoms, we were able to refine the positions of carbon atoms only. The unit cell contains four independent atomic sites: C1, C2, C3, and C4 occupying the positions 2a, 24g, 8c and 24g, respectively. Therefore, there are 58 CH4\mathrm{CH}_4 molecules in the unit cell.

Crystal structure of methane-B

Figure S3. Crystal structure of methane-B. Black, red, green and blue balls show the positions of C1, C2, C3 and C4 atoms respectively (Table S1).

Pressure (GPa)7.0715.2
Crystal data
Crystal system, space groupCubic, I4ˉ3mI\bar{4}3m
Temperature (K)293
aa (Å)11.942 (2)11.439 (2)
VV (Å3^{3})1703.0 (5)1496.9 (5)
ZZ58
Radiation typeX-ray, λ=0.34453\lambda= 0.34453 Å
μ\mu (mm−1^{-1})0.03
Data collection
DiffractometerGP @ ID16-B
Absorption correctionMulti-scan
No. of measured, independent and observed [I>3σ(I)I > 3\sigma(I)] reflections818, 323, 232728, 279, 186
RintR_{\mathrm{int}}0.0720.054
(sin⁡θ/λ)max⁡(\sin\theta/\lambda)_{\max} (Å−1^{-1})0.7560.744
Refinement
R[F2>2σ(F2)]R[F^{2} > 2\sigma(F^{2})], wR(F2)wR(F^{2}), SS0.106, 0.126, 3.160.108, 0.124, 3.52
No. of reflections323279
No. of parameters1717
Δρmax⁡\Delta\rho_{\max}, Δρmin⁡\Delta\rho_{\min} (e Å−3^{-3})0.54, -0.600.31, -0.32
Crystal structure
Atomic coordinates and Wyckoff positionsC1 - 2aa (0 0 0) C2 - 24gg (0.0895(3) 0.0895(3) 0.2801(4)) C3 - 8cc (0.3211(4) 0.3211(4) 0.3211(4)) C4 - 24gg (0.1443(3) 0.1443(3) 0.5368(4))C1 - 2aa (0 0 0) C2 – 24gg (0.0893(3) 0.0893(3) 0.2808(5)) C3 – 8cc (0.3216(4) 0.3216(4) 0.3216(4)) C4 – 24gg (0.1456(3) 0.1456(3) 0.5354(5))
Atomic displacement parameters, UisoU_{\mathrm{iso}} (Å2^{2})C1 - 0.031(2) C2 - 0.0346(17) C3 - 0.0331(14) C4 - 0.0365(17)C1 – 0.027(3) C2 – 0.032 (2) C3 – 0.0256(15) C4 – 0.035(2)

Table S1. Crystal data, refinement and crystal structure details of Methane-B phase

Further compression leads to a noticeable lattice distortion at above 18 GPa and the unit cell can no longer be considered cubic. This distortion is accompanied by the appearance of new modes in the Raman spectrum (Figure 2 and Figure S2 of the Supplemental Material [24]) in agreement with the previous observations [16, 17] at similar pressure conditions. The most plausible indexing of the diffraction patterns of a methane high-pressure phase (HP) above 18 GPa can be performed using a rhombohedral unit cell (Table 1, Table 2). Symmetry-lowering transitions from cubic I4ˉ3mI\bar{4}3m to rhombohedral symmetry may follow Γ4\Gamma_4 or Γ5\Gamma_5 irreducible representations with the following transformation of the lattice parameters:

Pressure, GPa# reflectionsR3R3 model: # parametersR3R3 model: R1R_1R3mR3m model: # parametersR3mR3m model: R1R_1
22.4381410.127280.1303
27.4474410.138280.1422
35.6280410.1020280.1033
39.0198410.1415280.1436

Table 1. Comparison of agreement factors between R3R3 and R3mR3m models of the HP methane

Pressure (GPa)22.427.435.639.0
Crystal data
Crystal system, space grouptrigonal, R3R3trigonal, R3R3trigonal, R3R3trigonal, R3R3
ZZ87878787
aa, cc (Å)15.735(3), 9.660(2)15.396(3), 9.481(2)15.176 (8), 9.11 (7)14.918(8), 9.14(7)
VV (Å3^3)2071.3(7)1946.1(7)1818(13)1762(14)
Radiation typeX-ray, λ=0.34453\lambda= 0.34453 ÅX-ray, λ=0.34453\lambda= 0.34453 ÅX-ray, λ=0.34453\lambda= 0.34453 ÅX-ray, λ=0.2891\lambda= 0.2891 Å
μ\mu (mm−1^{-1})0.020.020.020.02
Data collection
DiffractometerGP @ ID16-BGP @ ID16-BGP @ ID16-BGP @ P02.2
Absorption correctionMulti-scanMulti-scanMulti-scanMulti-scan
No. of measured, independent and observed [I>3σ(I)I > 3\sigma(I)] reflections676, 663, 381640, 627, 474475, 365, 280792, 587, 198
RintR_{\mathrm{int}}0.1110.0260.0390.059
(sin⁡θ/λ)max⁡(\sin\theta/\lambda)_{\max} (Å−1^{-1})0.7520.7680.5860.664
Refinement
R[F2>2σ(F2)]R[F^2 > 2\sigma(F^2)]0.127, 0.136,0.138, 0.150,0.102, 0.117,0.141, 0.149,
wR(F2)wR(F^2), SS3.145.703.611.60
No. of reflections663627365587
Δρmax⁡\Delta\rho_{\max}, Δρmin⁡\Delta\rho_{\min} (e Å−3^{-3})0.32/-0.280.36, -0.370.25, -0.190.23, -0.20
Crystal structure
C1_1
x0000
y0000
z0.02(12)0.02(8)0.000(17)0.013(9)
UisoU_{\mathrm{iso}}0.09(11)0.090(12)0.049(8)0.033(4)
C2_1
x0.0618(8)0.0644(9)0.0623(8)0.0635(9)
y0.1264(8)0.1264(9)0.1234(9)0.1251(8)
z0.33(12)0.33(8)0.340(16)0.340(8)
UisoU_{\mathrm{iso}}0.045(3)0.035(3)0.037(4)0.050(3)
C2_2
x0.4816(9)0.4801(9)0.4844(8)0.4863(8)
y0.1468(9)0.1451(9)0.1485(8)0.1476(7)
z0.17(12)0.17(8)0.165(15)0.174(8)
UisoU_{\mathrm{iso}}0.041(3)0.028(3)0.028(3)0.046(3)
C2_3
x0.6663(9)0.6653(9)0.6632(7)0.6622(7)
y0.1458(9)0.1477(9)0.1483(7)0.1495(7)
z0.17(12)0.17(8)0.162(16)0.166(8)
UisoU_{\mathrm{iso}}0.039(3)0.028(3)0.029(3)0.041(3)
C2_4
x0.1216(7)0.1230(8)0.1256(7)0.1245(7)
y0.2456(7)0.2449(8)0.2489(6)0.2506(7)
z0.09(12)0.09(8)0.101(16)0.093(9)
UisoU_{\mathrm{iso}}0.041(3)0.035(3)0.034(3)0.046(3)
C3_1
x0000
y0000
z0.67(12)0.67(8)0.666(16)0.670(10)
UisoU_{\mathrm{iso}}0.016(4)0.014(4)0.015(4)0.049(6)
C3_2
x0.2388(6)0.2376(7)0.2361(6)0.2339(7)
y0.1208(6)0.1174(7)0.1178(6)0.1175(7)
z0.14(12)0.14(8)0.152(15)0.156(8)
UisoU_{\mathrm{iso}}0.018(2)0.014(2)0.019(3)0.039(3)
C4_1
x0.4644(8)0.4595(8)0.4648(7)0.4635(7)
y-0.0756(7)-0.0774(8)-0.0704(6)-0.0730(7)
z0.24(12)0.24(8)0.244(16)0.239(8)
UisoU_{\mathrm{iso}}0.047(3)0.038(3)0.036(3)0.047(3)
C4_2
xx0.0092(9)0.0123(10)0.0065(9)0.0062(7)
yy0.3087(10)0.3130(11)0.3133(9)0.3104(7)
zz0.33(12)0.33(8)0.339(16)0.337(8)
UisoU_{\mathrm{iso}}0.050(4)0.049(4)0.036(3)0.038(3)
C4_3
xx0.3005(8)0.2991(10)0.3080(9)0.3070(7)
yy0.3121(9)0.3081(11)0.3116(9)0.3122(7)
zz0.33(12)0.33(8)0.350(16)0.342(8)
UisoU_{\mathrm{iso}}0.047(4)0.043(4)0.037(3)0.038(3)
C4_4
xx0.2294(8)0.2239(10)0.2251(8)0.2230(8)
yy0.4529(8)0.4531(10)0.4496(9)0.4481(8)
zz0.19(12)0.18(8)0.166(15)0.154(8)
UisoU_{\mathrm{iso}}0.038(3)0.039(3)0.030(3)0.040(3)

Table 2. Crystal data, refinement and crystal structure details of Methane HP phase

(aRbRcR)=(−1100−111/21/21/2)(acbccc)\begin{pmatrix} a_R \\ b_R \\ c_R \end{pmatrix} = \begin{pmatrix} -1 & 1 & 0 \\ 0 & -1 & 1 \\ 1/2 & 1/2 & 1/2 \end{pmatrix} \begin{pmatrix} a_c \\ b_c \\ c_c \end{pmatrix}

where aRa_R, bRb_R, cRc_R are lattice vectors of the rhombohedral unit cell (hexagonal setting), while aca_c, bcb_c, ccc_c lattice vectors of the cubic unit cell. The Γ4\Gamma_4 distortion leads to the space group symmetry R3mR3m (#160), while Γ5\Gamma_5 - to R3R3 (#146). We should note that at pressures above 30 GPa we see a slight departure of the lattice parameters from RR-centered hexagonal lattice (a=ba=b, α=β=90\alpha=\beta=90, γ=120∘\gamma=120^\circ) suggesting a further monoclinic distortion. However, the degree of this distortion is not reproducible between two experiments (Figure S4). Therefore, we conclude that this effect is related to the development of non-hydrostatic stresses in the sample, which is unavoidable at very high pressures and especially for very soft crystals like methane. We therefore treat the high-pressure phase of methane as a single R3R3 phase.

Maximum angle deviation from the ideal angles of the hexagonal lattice

Figure S4. Maximum angle deviation from the ideal angles of the hexagonal lattice (α=β=90∘\alpha= \beta= 90^\circ, γ=120∘\gamma= 120^\circ).

We have tested both R3R3 and R3mR3m models (Table 1). Although the R3R3 model has smaller agreement factors, they are not significantly different from those of the R3mR3m model. We have performed a Hamilton significance test in order to check if the improvement of the agreement factors for the R3R3 model may be considered significant [32]. For this test, both models were refined using the same set of reflections averaged based on the R3R3 symmetry. Based on the Hamilton test, the R3R3 model is preferable with the 75-90 % confidence level depending on the pressure point. We, therefore, cannot completely reject the R3mR3m model based on the XRD data, but can conclude that the R3R3 model is more likely (Table 1).

In order to complete the structure solution and determine the positions of hydrogen atoms, theoretical structure optimizations have been performed at various pressures (Figure 3). First, we performed ab initio Random Structure Search (AIRSS) using the CASTEP code [22]. All calculations used the PBE functional together with the DFT-D dispersion correction scheme of Tkatchenko and Scheffler [33]. Among a plethora of phases marginally different in enthalpies, two structures emerged having R3R3 symmetries and 29 molecules in the primitive unit cell (R3R3-I and R3R3-II on the Figure 2(a)) in agreement with the experimental HP phase (Figure 4). Both calculated R3R3 structures have equivalent carbon substructures and slightly different orientations of CH4CH_4 molecules centered at the C4 atoms. According to the theoretical calculations both R3R3 structures are not the most energetically favorable at 0 K. However, if temperature effects are taken into account, the calculated Gibbs free energies show that the R3R3-II phase becomes the most favorable above 260 K at 25 GPa, which perfectly agrees with the experiment (Figure 3(b)). A monoclinic distortion noticed experimentally is found to be energetically unfavorable. Below we will mainly discuss the most energetically favorable R3R3 phase (R3R3-II, Table 3).

Relative enthalpy and Gibbs free energy of methane phases

Figure 3. (a) Theoretically calculated relative enthalpy of the most stable methane phases. The results for HP phase (R3R3-IIII) are shown by a thick red line. Other orientationally ordered phases are shown by thin lines of various colors. The vertical dashed line correspond to the pressure at which the temperature effects were taken into account (b) Relative Gibbs free energies of the most stable methane phases as a function of temperature at 25 GPa. The symbols at T=0T=0 K represent relative classical enthalpies as computed with norm-conserving pseudo potentials.

Crystal structure of HP-CH4 at 35 GPa

Figure 4. Crystal structure of HP-CH4_4 at 35 GPa. Black, red, green and blue balls represent the positions of C1, C2, C3, and C4 carbon atoms, respectively; small pinkish circles are hydrogen atoms. Left panel shows C atoms only; gray lines connect the nearest C atoms with a cutoff of 3.08 Å.

Table 3. Crystal structure of HP-CH4_4 at 25 GPa optimized by calculations (R3R3, aa = 15.3315, cc = 9.5221 Å).

Sitexxyyzz
C1_100-0.9996
H1_1_1-0.365771-0.622456-0.628642
H1_1_200-0.11349
C2_1-0.131193-0.066507-0.68994
H2_1_1-0.181504-0.146561-0.695142
H2_1_2-0.083784-0.042373-0.781794
H2_1_3-0.085473-0.048146-0.596947
H2_1_4-0.174924-0.029387-0.685741
C2_20.1482450.474857-0.519002
H2_2_10.0792760.476659-0.509461
H2_2_20.2054560.534268-0.456087
H2_2_30.1363050.402921-0.484066
H2_2_4-0.162001-0.181677-0.294102
C2_3-0.1830060.005534-0.191981
H2_3_1-0.255161-0.062329-0.199231
H2_3_20.144078-0.273369-0.461258
H2_3_3-0.129179-0.010913-0.144596
H2_3_4-0.1571890.036738-0.29511
C2_4-0.252732-0.13696-0.448695
H2_4_1-0.285273-0.116555-0.534236
H2_4_2-0.300828-0.15757-0.357472
H2_4_3-0.243439-0.199203-0.480598
H2_4_4-0.180176-0.073615-0.422955
C3_100-0.346612
H3_1_10.044147-0.032218-0.309195
H3_1_200-0.460331
C3_2-0.096727-0.545671-0.549019
H3_2_10.1730630.094993-0.812157
H3_2_2-0.079162-0.473773-0.593158
H3_2_3-0.447561-0.266257-0.299235
H3_2_40.2999050.127484-0.823554
C4_1-0.207481-0.422449-0.598684
H4_1_1-0.274275-0.422628-0.563519
H4_1_2-0.226509-0.496916-0.631667
H4_1_3-0.173923-0.37107-0.685189
H4_1_4-0.154241-0.399094-0.513821
C4_2-0.322138-0.356324-0.361957
H4_2_1-0.0473610.238783-0.656792
H4_2_20.0832510.322741-0.659402
H4_2_3-0.33452-0.297433-0.322463
H4_2_4-0.323976-0.355022-0.47503
C4_3-0.316752-0.01339-0.682065
H4_3_1-0.3082460.013518-0.575783
H4_3_20.2818390.313491-0.391824
H4_3_3-0.2522040.038722-0.743068
H4_3_40.3423940.246027-0.349393
C4_40.2144130.114945-0.509708
H4_4_10.2284180.190425-0.496222
H4_4_20.1414920.062025-0.467591
H4_4_30.2723120.106787-0.457289
H4_4_40.2147460.101117-0.620706

Table 3.

The unit cell of HP-CH4\mathrm{CH}_{4} (Figure 4 and Figure S5 of the Supplemental Material [24]) contains 11 symmetry independent carbon atoms with atoms C1_1\mathrm{C1\_1} and C3_1\mathrm{C3\_1} occupying positions 3a3a, while the rest occupying positions 9b9b. Here, the C atoms are named following the numbering for phase B [15] to show the relation between B and HP phases, i.e. the atom C3_1\mathrm{C3\_1} is generated from the atom C3 of the B-phase.

Crystal structure of HP CH₄ at 30 GPa

Figure S5. Crystal structure of HP CH4\mathrm{CH}_4 at 30 GPa. Black, red, green and blue balls represent the positions of C1\mathrm{C}_1, C2\mathrm{C}_2, C3\mathrm{C}_3 and C4\mathrm{C}_4 atoms respectively (Table S4).

Figure 4 shows the main motif of the most energetically favorable R3R3 structure. The C1_1\mathrm{C1\_1} and C3_1\mathrm{C3\_1} atoms that lie on the three-fold axis possess a distorted tetrahedral icosi-octahedral environment. This polyhedron can be understood in the following way (see bottom of the Figure S5 [24]): the atom C3_1\mathrm{C3\_1} is coordinated by two six-membered rings, one of which is planar and capped, while the other is in a chair arrangement capped by a triangle. The packing of HP-CH4_4 is primarily defined by intermolecular H-H interactions. The histogram of H-H distances has two distinct peaks (Figure S6 [24]). The sharp peak at ~1.75 Å corresponds to intramolecular H-H distances and almost does not shift with pressure as the molecules remain almost rigid under compression. The broad maxima at ~2.05 and ~1.95 Å at 20 and 30 GPa, respectively, correspond to closest intermolecular H…H contacts. These distances are in a good agreement with shortest intramolecular H…H contacts in compressed hydrogen [34].

Distribution of H-H distances in calculated CH₄ HP structures at various pressures

Figure S6. Distribution of H-H distances in calculated CH₄ HP structures at various pressures. Distances are given in Å, vertical axis represents the number of occurencies.

The ν1\nu_1 and ν3\nu_3 C-H stretching Raman modes become composite (see also Refs. [16, 17]) in phases B and HP suggesting either site symmetry or vibrational splitting. The Raman spectra collected at 25 GPa in HP phase show that each of the ν1\nu_1 and ν3\nu_3 multifold can be well represented by 4 peaks (Figure 2, Figure 5). On the other hand, theoretical calculations of the Raman activity of HP phase performed in this work show a larger number of components, which correspond to various molecular sites as the analysis for the ν1\nu_1 modes shows (Figure 5). However, the splitting pattern and value are similar to the experimental; moreover, some theoretically calculated peaks are very close to each other making them difficult to resolve in the experiment; this likely results in broadened peaks, which consist of many components. A prominent feature of the ν1\nu_{1} Raman spectra is a well split off low frequency peak, that theory predicts is due internal vibrations of molecules involving C1 and C3 carbon in 3a sites, which are characterized by relatively more uniform in length and overall longer intermolecular distances (Figure S7 [24]). This distinction results in a well-separated low-frequency ν1\nu_{1} component, which can be also seen in phase B with similar carbon site geometry save the lattice distortion.

Figure 5. Raman spectra of ν1\nu_{1} (main panel) and ν3\nu_{3} (insets) stretching modes. The top panels is the spectrum measured at 25 GPa (488 nm excitation). Crosses are the data and gray line is the best fit to the Voigt oscillator model; variable color lines show the individual ν1\nu_{1} and ν3\nu_{3} components. The bottom panel shows the theoretically computed spectra at the same pressure. The labels (carbon atom sites) show the assignment of the peaks to the molecules of different kind. The frequency axes in the top and the bottom panels are scaled for the best correspondence of the experimental and theoretical peaks.

Average C-C distances of C1-C4 methane molecules to the next neighbors as a function of pressure

Figure S7. Average C-C distances of C1-C4 methane molecules to the next neighbors as a function of pressure.

The unit-cell parameters and symmetry of phases B and HP obtained here experimentally and theoretically provide reliable information on the P-V equation of state of methane (Figure 6). Our results agree well with the previous determinations based on single-crystal [13, 15] and powder [19, 20, 35] XRD. The cubic unit cell of phase B shows a distortion in the HP phase, which first appears small but then increases with pressure (Figure S4 [24]). Please note that pre-B phase, which appears on fast compression of phase A, was indexed as a simple cubic lattice [19, 20]; this phase is not examined with XRD here (but was detected by Raman spectroscopy to coexist with HP phase in the experiment #2 up to the highest pressure investigated) but it is clearly different from HP phase of this study. HP and pre-B methane phases have very close volumes per the formula unit (Figure 6) if one assumes that pre-B has 21 molecules in the unit cell as in phase A. Unlike previously reported modifications of HP phase based on changes in a number of ν1\nu_1 and ν3\nu_3 Raman components [16, 17], our direct structural data (Figure 1, Table 1-Table 2) and analysis point out into the sole HP phase up to 71 GPa (cf. Refs. [16, 17]), where the rhombohedral distortion gradually increases with pressure (Figure S4 [24]). Consistently, we also find (by fitting the spectra using an oscillator models, Figure 2, Figure 5) that the number of C-H Raman components remains the same (cf. Refs. [16, 17]) and their frequencies (Figure 7 [24]) agree well with the results of previous Raman investigations [16, 17]. Previous Raman investigations [16, 17] reported minor changes in the spectra (peak splitting and changes in peak intensities) and in the PP dependence of the C-H stretch frequencies. However, these changes do not signify the phase transitions. The Raman peaks may split because they consist of a number of components, which have different pressure slopes so they become separated with pressure, while their relative intensity may change due to a number of reasons including the mode coupling and modification in the orientational order.

Lattice parameters and volumes per formula unit of methane as a function of pressure

Figure 6. Lattice parameters (a) and volumes per formula unit (b) of methane as a function of pressure. Note two vertical scales in the panel (a) for phases B (blue) and HP (red) which are scaled by a factor of 3/2\sqrt{3}/2 reflecting the difference in the unit cell dimensions. Also, the parameter aa of the HP phase is divided by a factor of 1.633. Our experimental and theoretical volumes (b) are compared to previous experiments in phases I, A, B, and HP [12, 13, 15, 19, 20]. The unit cell volume of Ref. [20] was assumed to contain 21 formula units. The lines are guides to the eye.

Raman frequencies of the C-H stretching modes of methane as a function of pressure

Figure 7. Raman frequencies of the C-H stretching modes of methane as a function of pressure. Vertical dashed lines mark the phase transitions following Ref. [16]. The symbols are the

The unit cell volume changes smoothly through the B-HP transition suggesting a very small or no discontinuity. The theoretically computed lattice parameters and volumes are smaller than those determined theoretically, but the discrepancy is within 3%, which commonly is considered acceptable. However, this small difference may indicate that HP methane may still be partially disordered, which also corroborate with smaller (by some 20%) observed values of ν1\nu_1 and ν3\nu_3 multifolds splitting compared to the computed ones (Figure 5).

results of two experiments of this work (solid line is the guide to the eye), while dashed and dotted lines represent the results of previously published works [16, 17].

Conclusions

Our concerted experimental and theoretical investigation determined the structure of the high-pressure HP phase of methane including the positions of hydrogen atoms. The structure is rhombohedral (space group R3R3) with 87 molecules in the unit cell. This extraordinary complex structure of approximately spherical molecules is in a drastic contrast with simple fccfcc (hcphcp) structures of noble gas solids. Moreover, the theoretically predicted orientationally ordered structures with smaller unit cell and with well elaborated intermolecular coupling schemes also do not capture the real structure. This indicates that the observed structure represents a balance between enthalpy and entropy terms thus suggesting that some molecular sites remain orientationally disordered; this makes this structure analogous to clathrates, where host and guest molecules are the same.

Our direct single-crystal XRD results suggest that there is only one high-pressure phase (HP) of methane between 20 and 71 GPa in contrast to previous works, which suggested three HP phase based on the results of Raman spectroscopy. HP phase investigated here can coexist with a cubic pre-B phase investigated previously, the detailed structure of which still remains unknown.

Acknowledgements:

Parts of this research were carried out at the Extreme Conditions Beamline (P02.2) at DESY, a member of Helmholtz Association (HGF). Portions of this work were performed at HPCAT (sector 16) of the Advanced Photon Source (APS), Argonne National Laboratory. HPCAT operations are supported by DOE-NNSA’s Office of Experimental Sciences. Concomitant Raman spectroscopy experiments were performed at GeoSoilEnviroCARS (The University of Chicago, Sector 13), Advanced Photon Source (APS), Argonne National Laboratory. GeoSoilEnviroCARS is supported by the National Science Foundation - Earth Sciences (EAR - 1634415) and Department of Energy GeoSciences (DE-FG02-94ER14466). The Advanced Photon Source is U.S. Department of Energy (DOE) Office of Science User Facility operated for the DOE Office of Science by Argonne National Laboratory under Contract No. DE-AC02-06CH11357. CJP acknowledges financial support from the Engineering and Physical Sciences Research Council [Grant EP/P022596/1]. MB research was sponsored by the Army Research Office and was accomplished under the Cooperative Agreement Number W911NF-19-2-0172.

Supplemental Material

References

  1. [1]P. Hebert, A. Polian, P. Loubeyre and R. Le Toullec, Physical Review B 36 (17), 9196-9201 (1987).
  2. [2]N. W. Ashcroft, Physical Review Letters 92 (18), 187002 (2004).
  3. [3]E. Snider, N. Dasenbrock-Gammon, R. McBride, M. Debessai, H. Vindana, K. Vencatasamy, K. V. Lawler, A. Salamat and R. P. Dias, Nature 586 (7829), 373-377 (2020).
  4. [4]W. B. Hubbard, Planetary interiors. (Van Nostrand Reinhold, New York, N.Y., 1984).
  5. [5]C. Cavazzoni, G. L. Chiarotti, S. Scandolo, E. Tosatti, M. Bernasconi and M. Parrinello, Science 283 (5398), 44-46 (1999).
  6. [6]S. Ninet, F. Datchi and A. M. Saitta, Physical Review Letters 108 (16), 165702 (2012).
  7. [7]V. B. Prakapenka, N. Holtgrewe, S. S. Lobanov and A. Goncharov, arXiv:2007.07715v1 [cond-mat.mtrl-sci] (2020).arxiv.org/abs/2007.07715
  8. [8]S. S. Lobanov, P.-N. Chen, X.-J. Chen, C.-S. Zha, K. D. Litasov, H.-K. Mao and A. F. Goncharov, Nature Communications 4 (1), 2446 (2013).
  9. [9]F. Ancillotto, G. L. Chiarotti, S. Scandolo and E. Tosatti, Science 275 (5304), 1288-1290 (1997).
  10. [10]D. Kraus, J. Vorberger, A. Pak, N. J. Hartley, L. B. Fletcher, S. Frydrych, E. Galtier, E. J. Gamboa, D. O. Gericke, S. H. Glenzer, E. Granados, M. J. MacDonald, A. J. MacKinnon, E. E. McBride, I. Nam, P. Neumayer, M. Roth, A. M. Saunders, A. K. Schuster, P. Sun, T. van Driel, T. Döppner and R. W. Falcone, Nature Astronomy 1 (9), 606-611 (2017).
  11. [11]G. Gao, A. R. Oganov, Y. Ma, H. Wang, P. Li, Y. Li, T. Itaka and G. Zou, The Journal of Chemical Physics 133 (14), 144508 (2010).
  12. [12]R. M. Hazen, H. K. Mao, L. W. Finger and P. M. Bell, Applied Physics Letters 37 (3), 288-289 (1980).
  13. [13]H. E. Maynard-Casely, C. L. Bull, M. Guthrie, I. Loa, M. I. McMahon, E. Gregoryanz, R. J. Nelmes and J. S. Loveday, The Journal of Chemical Physics 133 (6), 064504 (2010).
  14. [14]R. Bini, L. Ulivi, H. J. Jodl and P. R. Salvi, The Journal of Chemical Physics 103 (4), 1353-1360 (1995).
  15. [15]H. E. Maynard-Casely, L. F. Lundegaard, I. Loa, M. I. McMahon, E. Gregoryanz, R. J. Nelmes and J. S. Loveday, The Journal of Chemical Physics 141 (23), 234313-234313 (2014).
  16. [16]H. Hirai, K. Konagai, T. Kawamura, Y. Yamamoto and T. Yagi, Chemical Physics Letters 454 (4-6), 212-217 (2008).
  17. [17]P.-N. Chen, C.-S. Zha, X.-J. Chen, J. Shu, R. J. Hemley and H.-k. Mao, Physical Review B 84 (10), 104110 (2011).
  18. [18]R. Bini and G. Pratesi, Physical Review B 55 (22), 14800-14809 (1997).
  19. [19]S. Umemoto, T. Yoshii, Y. Akahama and H. Kawamura, Journal of Physics: Condensed Matter 14 (44), 10675-10678 (2002).
  20. [20]L. Sun, W. Yi, L. Wang, J. Shu, S. Sinogeikin, Y. Meng, G. Shen, L. Bai, Y. Li, J. Liu, H. k. Mao and W. L. Mao, Chemical Physics Letters 473 (1-3), 72-74 (2009).
  21. [21]L. Sun, Z. Zhao, A. L. Ruoff, C.-S. Zha and G. Stupian, Journal of Physics: Condensed Matter 19 (42), 425206 (2007).
  22. [22]S. J. Clark, M. D. Segall, C. J. Pickard, P. J. Hasnip, M. I. J. Probert, K. Refson and M. C. Payne, Zeitschrift für Kristallographie - Crystalline Materials 220 (5-6), 567-570 (2005).
  23. [23]I. Kantor, V. Prakapenka, A. Kantor, P. Dera, A. Kurnosov, S. Sinogeikin, N. Dubrovinskaia and L. Dubrovinsky, Review of Scientific Instruments 83 (12), 125102 (2012).
  24. [24](See Supplemental Material for Methods, Supplemental Figures (S1-S5) and Tables (S1-S5), and References (28-33).).
  25. [25]V. CrysAlisPro Software System, Rigaku Oxford Diffraction, Oxford, UK. Oxford, UK 2014.
  26. [26]G. M. Sheldrick, Acta Crystallographica Section A Foundations and Advances 71 (1), 3-8 (2015).
  27. [27]V. Petricek, M. Dusek and L. Palatinus, Zeitschrift für Kristallographie 229 (5), 345-352 (2014).
  28. [28]N. Holtgrewe, E. Greenberg, C. Prescher, V. B. Prakapenka and A. F. Goncharov, High Pressure Research 39 (3), 457-470 (2019).
  29. [30]H. K. Mao, J. Xu and P. M. Bell, Journal of Geophysical Research: Solid Earth 91 (B5), 4673-4676 (1986).
  30. [31]C. J. Pickard and R. J. Needs, Journal of Physics: Condensed Matter 23 (5), 053201 (2011).
  31. [32]W. C. Hamilton, Acta Crystallographica 18 (3), 502-510 (1965).
  32. [33]A. Tkatchenko and M. Scheffler, Physical Review Letters 102 (7), 073005 (2009).
  33. [34]V. Labat, P. Gonzalez-Morelos, R. Hoffmann and N. W. Ashcroft, The Journal of Chemical Physics 136 (7), 074501 (2012).
  34. [35]I. Nakahata, N. Matsui, Y. Akahama and H. Kawamura, Chemical Physics Letters 302 (3), 359-362 (1999).

Paper details

vlnc:W_t8trydqw
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents