CO2 fluid inclusions are popular in the mantle‐derived rock. CO2 Raman densimeter is widely used for estimating depth provenance and magma plumbing system. However, in order to obtain precise CO2 density, we should also measure CO2 temperature simultaneously because the densimeter has temperature dependency. In this study, we measured CO2 Raman spectra with densities of 0.8–1.0 g/cm3 at temperatures of 15, 25, 35, 45, and 55°C using a high‐pressure optical cell. We propose a new equation relating hot bands to the Fermi diad intensity ratio, temperature, and distance between the Fermi diad (delta, cm−1), which has higher accuracy than those of previous studies (±3.9–4.7°C) across all measurement conditions. The change in temperature engenders thermal expansion or shrinkage of mineral, resulting in change in CO2 density of fluid inclusion. Simultaneous measurement of both density and temperature of CO2 will be a probe for elastic property of minerals.

Unintended local temperature enhancement by excitation laser might change Raman spectral features and potentially lead to misinterpretation of the data. To evaluate robustness of Raman CO2 densimeters in the presence of laser heating, we investigate the relation between temperature (T, °C), density (ρ, g/cm3), and Fermi diad split (Δ, cm−1) using a high‐pressure optical cell at 23°C to 200°C and 7.2–248.7 MPa. Results indicate that Δ decreases concomitantly with increasing temperature for a constant density in all density regions investigated. This result suggests that the density estimated based on Δ might be underestimated if the fluid is heated locally by the laser. Combining results of earlier studies with those of the present study indicates that the temperature dependence of Δ (|(∂Δ/∂T)ρ|) has a maximum value around 0.6–0.7 g/cm3. Consequently, at very high densities such as 1.1–1.2 g/cm3, |(∂Δ/∂T)ρ| is small. Thus, Δ at such densities is less affected by laser heating. However, at densities below approximately 0.7 g/cm3, although |(∂Δ/∂T)ρ| becomes smaller at lower densities, the relative density decrease becomes larger even for a small density decrease because the density itself becomes smaller. Therefore, at such densities, a density decrease of more than 10% was observed for some fluid inclusions, even at typical laser powers for inclusion analysis. Finally, to accurately estimate the density even in the presence of laser heating, we show that it is effective to estimate the intercept Δ from the correlation between Δ and laser power and substitute it into Δ–ρ relations.

No abstract

We investigated the relation between Sr# [= 100 Sr/(Sr + Ca) in mol] of calcites and their Raman spectra for Sr‐doped calcite samples with Sr# of 0–13.2. We selected 3 major Raman peaks observed at 150–155 cm−1 (peak A), 274–282 cm−1 (peak B), and 710–712.5 cm−1 (peak C). With increasing Sr#, the peaks shifted monotonously to lower wavenumbers. We obtained a linear expression for peak B as Sr# = −1.734 ΔνB + 76.93, where ΔνB is a separation between peak B and a neon emission line of 237.07 cm−1 in Raman shift. The present spectroscopic analysis can ascertain Sr# of calcite with precision of ±1.26 in the range of Sr# of 0–13.2.

We improved Raman spectroscopic densimetry for a CO2 fluid using a micro-Raman spectrometer with high spectral resolution. For precise determination of CO2 density, we obtained Δ, the separation of wavenumbers between two main peaks of the CO2 Raman spectrum, with a precision of ±0.006 cm−1 (1σ). This precision will be enhanced to ±0.003 cm−1 by increasing counts, corresponding to a density precision of ±0.0009 g cm−3, which is approximately one order of magnitude less than that reported in earlier studies.