The Colors of Ices: Measuring ice column density through photometry

TL;DR

The study demonstrates that JWST photometry, when combined with laboratory-based ice opacities through the icemodels toolkit, can quantify interstellar ice column densities without spectroscopy. By applying this method to Galactic Center clouds and validating against NIRSpec data, the authors detect CO, H$_2$O, CO$_2$ ices and find evidence for CH$_3$OH-related absorption in F356W, with GC ices showing a higher H$_2$O fraction and overall higher CO ice abundance than local clouds. They infer that a substantial fraction of carbon is frozen in CO ice, implying an elevated metallicity in the GC (Z$_{GC} \\gtrsim 2.5\\,Z_\\odot$) and a metallicity-ice abundance relation [CO$_{ice}$/H$_2$] ≈ 0.23 (Z/Z$_\\odot$) − 4.27. The work shows photometric ice measurements can probe the cold ISM metallicity structure and offers a scalable approach to map ices across many sightlines with JWST. Together, these results highlight enhanced ice-driven chemistry and a strongly metal-enriched GC environment, with broad implications for cloud evolution and star-formation conditions.

Abstract

Ices imprint strong absorption features in the near- and mid-infrared, but until recently they have been studied almost exclusively with spectroscopy toward small samples of bright sources. We show that JWST photometry alone can reveal and quantify interstellar ices, and we present a new open-source modeling tool, icemodels, to produce synthetic photometry of ices based on laboratory measurements. We provide reference tables indicating which filters are likely to be observably affected by ice absorption. Applying these models to NIRCam data of background stars behind \refereeseveral Galactic Center (GC) clouds \referee(dust ridge clouds A [the Brick], C, and D), and validating against NIRSpec spectra of Galactic disk sources, we find clear signatures of CO, H$_2$O, and CO$_2$ ices and evidence for excess absorption in the F356W filter likely caused by CH-bearing species such as methanol. The ice ratios differ between the Galactic disk and Center, with GC clouds showing a higher H$_2$O fraction. \refereeA large ice abundance \refereeis observed in CO, H2O, and possibly complex molecules, \refereewhich implies that there is substantial freezeout and therefore potential for ice-phase chemistry in non-star-forming gas. Accounting for all likely ices, we infer that $>25%$ of the total carbon is frozen into CO ice in the GC, which exceeds the entire solar-neighborhood carbon budget. By assuming the freezeout fraction is the same in GC and disk clouds, we obtain a metallicity measurement indicating that $Z_GC\gtrsim2.5Z_\odot$. These results demonstrate that photometric ice measurements are feasible with JWST and capable of probing the metallicity structure of the cold interstellar medium.

Figures (27)

• Figure 1: Example JWST NIRSpec archival spectra with the filters used in project 2221 and emphasized throughout this work, F182M, F212N, F405N, F410M, and F466N, shown. HOPS-383 (left) from program 5804 is a Class 0 YSO Megeath2012. IRAS16293 slit 74 from program 3222 is a star behind the IRAS16293 molecular cloud. The legend shows the synthetic magnitudes and colors computed from the spectrum. The title provides the source of the spectrum, including the source name from the project(HOPS-383, left, and spectrum ID 74, right), the program ID (5804, left, and 3222, right), and the grating used (PRISM, left, and G235M+G395M, right). The second line of the title provides the star's ICRS coordinates.
• Figure 2: F466N transmission spectrum (grey, arbitrary scaling) and ice effective opacity $\kappa_{eff}$ (Eqn \ref{['eqn:kappaeff']}) as a function of wavelength across the F466N band. The colored curves show single-ice opacities for CO at 25 K Gerakines2023, $\mathrm{H}_2\mathrm{O}$ at 25 K Mastrapa2009, and OCN$^-$ (with contaminants) from Novozamsky2001 as described in the legend.
• Figure 3: (top) Selected laboratory model ice opacities overlaid on the F405N, F410M, and F466N filter transmission profiles. The same opacity curves as in Figure \ref{['fig:F466Nplusopacities']} are shown, and CO$_2$ at 8 K from Gerakines2020 is added. (bottom) Selected model ice opacities overlaid on the wide-band filters F356W and F444W. These opacity curves are shown on all remaining NIRCam filters in Appendix \ref{['sec:moreabsorptionplots']}.
• Figure 4: The distributions of Brick sources in the [F405N]$-$[F466N] and [F410M]$-$[F466N] vs [F182M]$-$[F212N] color-color spaces compared to the intrinsic estimated colors of ice mixtures as described in § \ref{['sec:colors']} and indicated with curves of different colors as shown above. These diagrams show photometry from narrow- and medium-band filters used in JWST program 2221. The model curves show the effect of adding the labeled ice mixtures to the absorber starting at A$_V=17$ with a fixed CO/H$_2=2.5\times10^{-4}$ up to a maximum column density of N(H$_2)=5\times10^{22}$$\mathrm{cm}^{-2}$. As indicated in the legend, synthetic photometry has been computed for background stars behind local clouds in the I16923 observations from JWST project 3222 and the Serpens cloud from project 1611, and for YSOs from the Orion region from project 5804. These are plotted as red x's, blue y's, and purple +'s, respectively.
• Figure 5: The distributions of Brick, Cloud C, and Cloud D sources in the [F405N]$-$[F466N] and [F410M]$-$[F466N] vs [F182M]$-$[F212N] color-color spaces compared to the intrinsic estimated colors of ice mixtures as described in § \ref{['sec:colors']} and indicated with curves of different colors as shown above. The legend and models are the same as in figures \ref{['fig:colorcolor']} and \ref{['fig:ccd_f405nmf410m']}.