Raman spectroscopy and x-ray diffraction of ${\mathit{sp}}^3\mathrm{CaC}{\mathrm{O}}_3$ at lower mantle pressures

The exceptional ability of carbon to form ${\mathit{sp}}^2$ and ${\mathit{sp}}^3$ bonding states leads to a great structural and chemical diversity of carbon-bearing phases at nonambient conditions. Here we use laser-heated diamond-anvil cells combined with synchrotron x-ray diffraction, Raman spectroscopy, and first-principles calculations to explore phase transitions in $\mathrm{CaC}{\mathrm{O}}_3$ at $P>40\mathrm{GPa}$. We find that postaragonite $\mathrm{CaC}{\mathrm{O}}_3$ transforms to the previously predicted $P{2}_1/c\mathrm{CaC}{\mathrm{O}}_3$ with ${\mathit{sp}}^3$-hybridized carbon at 105 GPa ($\sim 30\mathrm{GPa}$ higher than the theoretically predicted crossover pressure). The lowest-enthalpy transition path to $P{2}_1/c\mathrm{CaC}{\mathrm{O}}_3$ includes reoccurring ${\mathit{sp}}^2$ and ${\mathit{sp}}^3\mathrm{CaC}{\mathrm{O}}_3$ intermediate phases and transition states, as revealed by our variable-cell nudged-elastic-band simulation. Raman spectra of $P{2}_1/c\mathrm{CaC}{\mathrm{O}}_3$ show an intense band at $1025\mathrm{c}{\mathrm{m}}^{-1}$, which we assign to the symmetric C-O stretching vibration based on empirical and first-principles calculations. This Raman band has a frequency that is $\sim 20\%$ lower than the symmetric C-O stretching in ${\mathit{sp}}^2\mathrm{CaC}{\mathrm{O}}_3$ due to the C-O bond length increase across the ${\mathit{sp}}^2-{\mathit{sp}}^3$ transition and can be used as a fingerprint of tetrahedrally coordinated carbon in other carbonates.

Figure 1

(a) X-ray diffraction (XRD) of $\mathrm{CaC}{\mathrm{O}}_3$ before heating, at $T$, and after heating at 105 GPa (with background). (b) Le Bail fit of the theoretically predicted $P{2}_1/c\mathrm{CaC}{\mathrm{O}}_3$ (red line) to the experimentally observed XRD pattern (black crosses). The thin black line is the difference curve. The corresponding rectangular diffraction image is shown in the top part. Asterisks and black boxes mark some of the diffuse peaks of remnant $\mathrm{CaC}{\mathrm{O}}_3$. X-ray energy is 42 keV.

Figure 2

Structures of (a) $P{2}_1/c\mathrm{CaC}{\mathrm{O}}_3$ and (b) $C{222}_1\mathrm{CaC}{\mathrm{O}}_3$ with outlined $\mathrm{C}{\mathrm{O}}_4$ tetrahedra. Calcium atoms are shown in blue, carbon is in brown (inside the polyhedra), and oxygen is in red. Black arrows show the distinct chirality of $\mathrm{C}{\mathrm{O}}_4$ tetrahedra chains in the crystal structures.

Figure 3

Pressure-volume relations for Pmmn $\mathrm{CaC}{\mathrm{O}}_3$ (black dots) and $P{2}_1/c\mathrm{CaC}{\mathrm{O}}_3$ (blue dots). Black line is a 300 K third-order Birch-Murnaghan equation of state (EOS) of Pmmn $\mathrm{CaC}{\mathrm{O}}_3$ (postaragonite) fitted to the $P-V$ data collected here. Best fits were obtained using the previously reported postaragonite $V_0$ value $(49.15{\AA }^3/\mathrm{f}.\mathrm{u}.)$ [18] in combination with $K^{\prime }$ in the range of 4.5–4.7. Fixing $V_0$ to the reported value is appropriate because of the larger number of $P-V$ measurements in the previous study. Corresponding EOS parameters are given in the bottom left corner. Pressure uncertainty σ is assumed to be 0.5, 1, and 1.5 GPa for $P<70$, 80–100, and $>100\mathrm{GPa}$, respectively.

Figure 4

(a) Raman spectra of $\mathrm{CaC}{\mathrm{O}}_3$ at 105 GPa collected with 488-, 532-, and 660-nm excitations. The gray curve is the spectrum of postaragonite $\mathrm{CaC}{\mathrm{O}}_3$ collected outside of the heated region. Black vertical bars are computed Raman modes of $P{2}_1/c\mathrm{CaC}{\mathrm{O}}_3$ (bottom) and postaragonite $\mathrm{CaC}{\mathrm{O}}_3$ (top) corrected upwards in frequency by $1.5\%$ and $0.5\%$, respectively. The height of the bars is proportional to the band intensity. The peak indicated by the question mark deviates significantly from the $P{2}_1/c\mathrm{CaC}{\mathrm{O}}_3$ model and may be due to the unheated $\mathrm{CaC}{\mathrm{O}}_3$, its yet unidentified phase, or minor nonmolecular $\mathrm{C}{\mathrm{O}}_2$ formed upon $\mathrm{CaC}{\mathrm{O}}_3$ thermal decomposition on Pt chunks. (b) Experimental spectrum of $\mathrm{CaC}{\mathrm{O}}_3$ laser heated at 105 GPa in comparison with the theoretical spectra of $P{2}_1/c$ and $C{222}_1\mathrm{CaC}{\mathrm{O}}_3$ at 105 GPa as computed by LDA-DFT. Gray areas are guides to compare the computed spectra with experiment.

Figure 5

(a) Raman spectra $(\lambda =488\mathrm{nm})$ of $\mathrm{CaC}{\mathrm{O}}_3$ collected on decompression after laser heating at 105 GPa. Gray areas show characteristic Raman bands of $P{2}_1/c\mathrm{CaC}{\mathrm{O}}_3$. (b) Pressure dependencies of the experimentally observed (squares) and computed (lines) Raman bands of $P{2}_1/c\mathrm{CaC}{\mathrm{O}}_3$ (frequency is corrected upwards by $1.5\%$). $P{2}_1/c\mathrm{CaC}{\mathrm{O}}_3$ is preserved down to $P=57\mathrm{GPa}$. The error bar for experimental measurements is not shown as it is smaller than the symbols (black squares).

Figure 6

Mechanism of the $\mathrm{Pmmn}\left(\mathrm{postaragonite}\right)\to P{2}_1/c$ transition of $\mathrm{CaC}{\mathrm{O}}_3$ at 100 GPa. Structures of initial postaragonite phase; transition states $\mathrm{T}{\mathrm{S}}_1, \mathrm{T}{\mathrm{S}}_2$, and $\mathrm{T}{\mathrm{S}}_3$; intermediate phases $\mathrm{I}{\mathrm{P}}_1$ and $\mathrm{I}{\mathrm{P}}_2$; and final $P{2}_1/c$ of $\mathrm{CaC}{\mathrm{O}}_3$ are shown (for clarity, we highlighted $\mathrm{C}{\mathrm{O}}_4$ tetrahedra). The evolution of the five shortest C-O distances is shown across the proposed transition path.