Mapping water ice with infrared broadband photometry

TL;DR

The study tackles the challenge of mapping interstellar ices on Galactic scales by introducing the Ice method, which infers the ice extinction from broadband infrared photometry. It defines two metrics, the ice color excess $Λ(m_1 - m_2)$ and a reddening-free index $Q'$, and identifies optimal passbands, notably the $W_1 - I_1$ combination, to minimize systematic uncertainties. An empirical calibration using 56 literature sources establishes a tight relation between the peak ice optical depth $τ_{3.0}^{\max}$ and $Λ(W_1 - I_1)$, yielding the robust form $τ_{3.0}^{\max} = -3.51 \ln[1 - 0.92 Λ(W_1 - I_1)]$, with explicit handling of error propagation and extinction-law dependencies. The method leverages archival Spitzer and WISE data to enable large-scale, photometric ice mapping, providing a powerful complement to spectroscopic surveys and enabling new studies of how ice formation depends on the environment across the Galaxy. While extinction-law variations and overlapping clouds pose challenges, the Ice approach offers a scalable pathway to construct high-resolution maps of the icy component of the ISM and to situate detailed JWST findings in a Galactic context.

Abstract

Interstellar ices play a fundamental role in the physical and chemical evolution of molecular clouds and star-forming regions, yet their large-scale distribution and abundance remain challenging to map. In this work, I present the ice color excess method, which parametrizes the peak optical depth ($τ_{3.0}^{\mathrm{max}}$) of the prominent 3$μ$m absorption feature, which is predominantly caused by the presence of solid H$_2$O. The method builds on well-established near-infrared color excess techniques and uses widely available infrared broadband photometry. Through detailed evaluation of passband combinations and a comprehensive error analysis, I construct the ice color excess metric $Λ(W_1 - I_1)$. This parameter emerges as the optimal choice that minimizes systematic errors while leveraging high-quality, widely available photometry from Spitzer and WISE data archives. To calibrate the method, I compile from the literature a sample of stars located in the background of nearby molecular clouds, for which spectroscopically measured optical depths are available. The empirical calibration yields a remarkably tight correlation between $τ_{3.0}^{\mathrm{max}}$ and $Λ(W_1 - I_1)$. This photometric technique opens a new avenue for tracing the icy component of the interstellar medium on Galactic scales, providing a powerful complement to spectroscopic surveys and enabling new insights into the environmental dependence of the formation and evolution of icy dust grains.

Figures (12)

• Figure 1: Infrared passbands and spectral energy distributions (SEDs). The upper panel illustrates transmission curves for the infrared passbands $K_S$, $W_1$, $I_1$, $L'$, $I_2$, and $W_2$. The lower panel presents the SEDs of three distinct sources: the light blue line represents an X-SHOOTER spectrum of source 15 (refer to Table \ref{['tab:master']}), the solid red line corresponds to the source NIR38, and the black line depicts Vega. All spectra have been rescaled to facilitate a better comparison.
• Figure 2: Extinction curves normalized to their value at the effective wavelength of the $W_1$ passband. The different lines correspond to various dust models and $R(V)$ values, as indicated in the legend. WD01 refers to the models of Weingartner2001, and D03 to those of Draine2003. The curves illustrate how both the choice of dust model and $R(V)$ parameter affect the wavelength dependence of extinction.
• Figure 3: Intrinsic colors for dwarfs (blue symbols) and giants (red symbols) as a function of spectral type, derived from synthetic photometry of IRTF spectra. Solid lines indicate power-law fits to the intrinsic colors for dwarfs (blue) and giants (red), respectively. The upper panel shows intrinsic $K_S - W_1$ color, while the lower panel displays $W_1 - I_1$. For both color indices, significant deviations from 0mag are observed only for late spectral types (later than approximately M0).
• Figure 4: Normalized absorption profiles of the 3µm feature toward two background stars published by Boogert2011McClure2023 and one protostar Rocha2025, compared with laboratory ice spectra from Hudgins1993. The observed profiles (red, blue, and green) were extracted from published data and interpolated onto a common wavelength grid. Laboratory spectra correspond to pure water ice (dotted black) and a strong ice mixture (solid black) at 10K. All profiles are normalized to a peak optical depth of 1 for a direct comparison of the profile shape.
• Figure 5: Standard deviation (solid lines) and maximum difference (dashed lines) in $K_S-W_1$ (black) and $W_1-I_1$ (red) colors as a function of the peak optical depth of the 3µm feature, calculated for five absorption profiles (see Fig. \ref{['fig:ice_profiles']}). The $K_S-W_1$ color is significantly more sensitive to variations in the absorption profile, while $W_1-I_1$ remains comparatively stable, enabling robust measurements of $\tau_{3.0}^{\max}$.