All six SiO-to-other-molecule emission ratios were lower in clouds with a larger density-time product, according to a survey of the Milky Way's Central Molecular Zone. The product combines each cloud's average H2 density with its dynamical time, while the ratios compare SiO intensity with six reference molecular lines.
The clearest associations were SiO/H13CN and SiO/CS. Their Spearman rank-correlation coefficients were -0.82 and -0.70, respectively, and both had 0.10 dex of residual scatter in log-log fits. The negative values indicate that the ratios tended to fall as the density-time product rose, but they remain observational correlations rather than causal effects.
From a strip of the CMZ to a cloud catalog
The observations were made with the Nobeyama Radio Observatory's 45 m telescope across a 3.5-by-0.5-degree region, at 20-arcsecond resolution, in eight molecular lines. The team released calibrated FITS cubes, integrated-intensity maps, analysis masks, the SiO cloud catalog and cloud-integrated line-intensity tables through Zenodo.
SCIMES identified 258 SiO-emitting clouds. The researchers measured all eight observed line intensities inside each cloud's SiO-defined mask, so the ratios compared emission from the same cataloged structures.
Most ratio analyses used 255 clouds after three exclusions. The SiO/H13CO+ comparison used 243, after 12 clouds with nonpositive intensities were removed, while CH3OH was excluded because its emission was weak.
Turning line emission into cloud properties
To put the clouds on the density-time axis, the researchers estimated H13CN-based masses using a 5 K excitation temperature, local thermodynamic equilibrium, an optically thin treatment and an assumed H13CN-to-H2 abundance of 4 x 10^-10. A spherical cloud volume was used to calculate density, while dynamical time was estimated as cloud size divided by velocity dispersion.
An optical-depth check, used to test the thin-emission assumption, found a median effective H13CN optical depth of about 0.19, with roughly 90% of clouds below 0.5. The study therefore treated the line as optically thin for most of the catalog, while noting that these are effective cloud-level estimates tied to the adopted excitation and radiative-transfer assumptions.
Those calculations produced 4.0 million solar masses of H13CN-traced gas in the SCIMES clouds, compared with about 28 million solar masses across the full survey area. The catalog therefore represented about 15% of the area-wide mass estimate. Among clouds with valid mass estimates, the median average H2 density was about 1,900 per cubic centimeter and the median dynamical time about 150,000 years.
The combined measure carried the clearest signal
For all six ratios, correlations with the density-time product were tighter than correlations with average density or dynamical time alone. For most ratios, the product's reported coefficient was 0.1 to 0.2 stronger than the density-only coefficient, while dynamical-time-only correlations had absolute coefficients no greater than 0.16.
The other four ratios followed the same downward direction but with more scatter. SiO/H13CO+ had rs=-0.64 with 0.35 dex scatter; SiO/HCN had rs=-0.37 with 0.25 dex; SiO/HCO+ had rs=-0.44 with 0.32 dex; and SiO/SO had rs=-0.67 with 0.15 dex. Weak H13CO+ emission was identified as a likely source of extra scatter in that comparison.
The density-time product also showed a spatial pattern: it was larger toward central longitudes and anticorrelated with absolute longitude offset from Sgr A*, with rs=-0.68.
A suggestive pattern, with clear limits
The authors interpret the overall decline as qualitatively consistent with postshock depletion of gas-phase SiO back onto dust grains. They also regard the stronger performance of the density-time product as support for that interpretation. But the reported coefficients are observational correlations, not causal effects, so the survey cannot establish that the product causes SiO depletion.
The average density is a volume average over each SCIMES mask, not a direct measure of the densest substructure. The mass and density scale also depends on assumptions about excitation, abundance, optical depth and geometry, making the density-time product a proxy rather than a direct measurement of chemical depletion.
The document is labeled arXiv:2608.25611v1 [astro-ph.GA] and dated 26 Aug 2026. The study was supported by JSPS KAKENHI grants JP19K14768, JP24K17091 and JP20H00178.
Paper data and sources
Original title: The Nobeyama 45 m Survey of Shocked Molecular Gas in the Central Molecular Zone. I. Survey Data, Cloud Catalog, and SiO Line-Ratio Trends
Authors: Shunya Takekawa, Shiho Tsujimoto, Tomoharu Oka, Rei Enokiya
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-08-26
DOI: Not available
Original paper · Full text