Weak, extended water vapor emission in the Horsehead nebula

TL;DR

The paper tackles the problem of understanding water vapor distribution in the Horsehead PDR under relatively low UV illumination. It introduces a PDR wrapper that projects 1D Meudon PDR outputs onto a curved 2D geometry, enabling direct comparison with edge-on observations and spatial emission profiles. Using a grid of isobaric PDR models and a comprehensive chemical network, the study constrains the thermal pressure to ~$P_{ m th} \\approx 4.1 \\\times 10^{6}$ K cm$^{-3}$ and a curvature scale of ~$R_{ m C} \\approx 0.057$ pc for the C$^+$/C/CO tracers, but finds that the ground-state o-H$_2$O 557 GHz line is overpredicted by at least a factor of ~7, indicating missing radiative-transfer or surface-chemistry physics. The authors propose a two-component model in which a low-density warm envelope scattering water photons can account for the observed extended H$_2$O emission, highlighting the importance of non-local radiative effects in PDRs and the need for improved grain-surface chemistry models. Overall, the work demonstrates that geometry and radiative transfer critically influence inferred water abundances in PDRs and motivates higher-resolution observations and refined surface chemistry in future studies.

Abstract

We analyzed archival Herschel observations of water vapor emission toward the Horsehead photon dominated region (PDR), along with supporting ground-based and airborne observations of CO isotopologues and fine structure lines of ionized and atomic carbon to determine the distribution and abundance of water vapor in this low-UV illumination PDR. Water emission in the Horsehead nebula is very weak and, surprisingly, extends outward beyond other PDR tracers such as $^{12}$CO or [CI] 609 $μ$m, reaching as far out as [CII] 158 $μ$m. We model the observations using a newly developed PDR wrapper that takes into account the geometry of this region. PDR modeling of the molecular and atomic lines studied here provides strong constraints on the thermal pressure, but not on the UV illumination. Maximum model line intensities %typically agree to within ~40\% with the observations. and spatial profiles are well reproduced, except for CO isotopologues, where the increase on the illuminated side of the PDR is steeper than observed. Water vapor abundance in the model reaches $3.6 \times 10^{-7}$ at $A_V \sim 3$ mag. However, the ground state $o$-H$_2$O 557 GHz line is systematically overestimated by the models by at least a factor of 7 for any values of the model parameters. This line has a very high optical depth and the emergent line intensity is sensitive to radiative transfer effects such as line scattering by water molecules in a low-density halo surrounding the dense PDR and the assumed microturbulent line width. A more accurate model of the water surface chemistry is required.

Figures (16)

• Figure 1: (Left panel) SPIRE 350 $\mu$m image of the Horsehead PDR. The white circle shows the FWHM SPIRE beam ($25.2^{\prime\prime}$). Black circles show the locations at which 557 GHz $o$-H$_2$O spectra were obtained, with the circle size corresponding to the FWHM HIFI beam ($38.1^{\prime\prime}$). (Right panel) Spectra of the 557 GHz $o$-H$_2$O and CO (2--1) lines across the PDR (black and blue curves, respectively), labeled by the right ascension offsets with respect to the reference position. The CO 2--1 spectra have been scaled down by a factor of 200. The bottom-right panel shows spectra of the ground state $o-$ and $p-$H$_2$O lines at the central position
• Figure 2: Integrated line intensities of atomic and molecular tracers computed over 9.2--11.8 km s$^{-1}$ velocity range across the PDR as a function of right ascension offset at $\Delta \delta = 0$. (Upper) [C ii], [C i], and H$_2$O (black, blue, and magenta, respectively). (Lower panel) CO (2--1), C$^{18}$O (2--1), CO (4--3), and H$_2$O (black, blue, red, and magenta, respectively). Line intensities have been normalized to their maxima toward the PDR, with the corresponding scaling factors as labeled. The data are shown at the native resolution of the images, with the corresponding FWHM beam sizes marked for each tracer. The size of the squares corresponds approximately to $\pm 1 \sigma$ H$_2$O observational uncertainties.
• Figure 3: Same as Figure \ref{['fig:strip']}, except that all tracers are convolved to the $38.1^{\prime\prime}$ resolution of the HIFI H$_2$O spectra.
• Figure 4: Illustration of the geometry of one-dimensional PDR models made with the Meudon PDR code. We see the two-component radiation field composed of a beamed part representing the stellar radiation field, and an isotropic part representing the ISRF. This configuration yields the typical chemical stratification with the ionization front denoting the beginning of the PDR, followed by the dissociation front, or H/H$_2$ transition, and the C$^+$/C/CO transition deeper in the cloud. H$_2$O appears even deeper.
• Figure 5: Illustration of the geometry of the PDR wrapper. (Left panel) Image of the Horsehead nebula as seen by the Euclid space telescope (Credits: ESA/Euclid/Euclid Consortium/NASA, image processing by J.-C. Cuillandre (CEA Paris-Saclay), G. Anselmi). Notice the reference frame, with the $x-$axis corresponding to the direction of the lines of sight. We rotate the nebula so that the $x-$axis points to the right. (Middle panel) Artistic interpretation of the rotation of the Horsehead nebula, generated using OpenAI’s DALL·E model based on the first panel image. (Right panel) Geometry of the wrapping 2D model, with the star radiation field coming from the top, and examples of Meudon PDR model output over-plotted as red vertical lines. Each of these lines correspond to a PDR model as represented on Fig. \ref{['fig:Meudon_edge-on_geometry']}. Notice that we use the same model for each position along the cut of the cloud. The multiple lines of sight correspond to different pixels along a vertical one-dimensional cut in the Horsehead nebula. We emphasize that only the lines of sight shown as solid lines are correctly accounted for by the model, as the central section of the nebula outside of the crescent-shaped area is not included in the computations. The geometry is assumed circular, with a radius of curvature that remains to be determined.