Reply to Datchi et al.: Recovered phase CO₂-V at low temperature and a newly predicted 3D-extended CO₂ phase

In their letter, Datchi et al. (1) question (i) the recoverability of CO₂-V at low pressure and (ii) the identification of CO₂-V as a Pna2₁ structure in our paper (2). We argue below that both of Datchi et al.’s (1) claims are incorrect.

Regarding Datchi et al.’s (1) first claim, it is well known that polymeric phase V of CO₂ (CO₂-V) is metastable at low pressures below 10 GPa, where phase I is a thermodynamically stable phase. As such, in this pressure region CO₂-V transforms back to phase I rather rapidly, especially below 3–5 GPa. This point is highlighted in figure 5A of our paper (2), showing a rapid conversion of CO₂-V below 10 GPa, as indicated by the rapid reduction in Raman intensity of its signature vibron at ∼700 cm⁻¹. This result clearly shows that nearly no phase V remain at 4 GPa at ambient temperature in a time-scale of experiments (seconds to minutes). In contrast, the CO₂-V phase at 180 or 191 K at ambient pressure is quite different, showing “kinetically” controlled stability over hours to a day. Note that there is absolutely no change in the intensity of the 650-cm^–1 vibron of phase V, compared with that at 11 GPa. Thus, its rapid disassociation (or sublimation) occurs above ∼200 K; the melting point of dry ice phase-I at ambient pressure suggests that the apparent “low” melting temperature of CO₂-V is likely of a kinetic origin, not the melting of extended CO₂-V. The peak at low frequency (below 200 cm⁻¹) may well be coming from solid phase I at 191 K at ambient pressure, although the observed peaks do represent some different intensities and peak positions. On the other hand, this is not at all unusual, because CO₂-V is always surrounded by untransformed CO₂-I. The important aspect of this finding is the fact that CO₂-V do exist “happily” with CO₂-I at ambient pressure and low temperature without CO₂-V being transformed to CO₂-I, again underscoring the kinetically controlled metastability.

As to Datchi et al.’s (1) second claim, in our paper (2) we did not present the Pna2₁ as the conclusive structure of CO₂-V, but suggested it as a potential new structure for extended CO₂. We have also pointed out that there are some differences in the calculated X-ray pattern (3) from those of experimental CO₂-V in our paper (2). Nevertheless, we have emphasized the presentation of a new potential Pna2₁ structure for CO₂ in a 3D network with fourfold sp³ carbon atoms. Therefore, we just marked Pna2₁ as CO₂-V'. Our predictions may provide insights into many potentially possible metastable phases in local energy minima, just like the CO₂-V_TR (P2₁2₁2₁). Moreover, our computed Raman spectrum of Pna2₁ structure in Fig. 1 disagrees with the calculation reported from Datchi et al. (1). In addition, the citation of experimental Raman spectrum of phase V in figure 1 of Datchi et al. (1) is incorrect, because in Yoo et al. (3) there is no Raman spectrum of phase V.

Fig. 1.

The computed Raman spectra of Pna2₁ structure along with previous calculation of Datchi et al. (1) at 48 GPa. The black line is taken from Datchi et al. (1) and the red line is from this response.