Photoevaporation Can Reproduce Extended $\mathrm{H_2}$ Emission from Protoplanetary Disks Imaged by JWST MIRI

TL;DR

This study develops the first photoevaporative $H_2$ wind model tailored for direct JWST comparisons by combining radiation-hydrodynamics with chemistry and non-LTE line post-processing. The synthetic $H_2$ pure rotational maps reproduce the observed X-shaped morphologies, with extents of ~50–300 au and semi-opening angles ~37°–50°, consistent with Tau 042021 and SY Cha, while lower-$J$ line fluxes align well with observations under modest dust obscuration. The results show that photoevaporation is a viable mechanism to explain key wind features, though conclusive wind origins require source-specific modeling and consideration of pumping processes that may boost higher-$J$ emission. The work emphasizes that morphology alone cannot definitively distinguish photoevaporative winds from MHD winds and advocates using energetics, line ratios, and stellar UV/X-ray properties as robust diagnostics, while outlining needed refinements for more precise comparisons and broader applicability.

Abstract

Understanding dispersal of protoplanetary disks remains a central challenge in planet formation theory. Disk winds, driven by magnetohydrodynamics (MHD) and/or photoevaporation, are now recognized as primary agents of dispersal. With the advent of James Webb Space Telescope (JWST), spatially resolved imaging of these winds, particularly in H2 pure rotational lines, has become possible, revealing X-shaped morphologies and integrated fluxes of $\sim 10^{-16}$-$10^{-15}{\rm \,erg\,s^{-1}\,cm^{-2}}$. However, the lack of theoretical models suitable for direct comparison has limited interpretation of these features. To address this, we present the first model of photoevaporative \ce{H2} winds tailored for direct comparison with JWST observations. Using radiation hydrodynamics simulations coupled with chemistry, we derive steady-state wind structures and post-process them to compute H2 level populations and line radiative transfer, including collisional excitation and spontaneous decay. Our synthetic images reproduce the observed X-shaped morphology with radial extents of $\gtrsim 50$-$300{\rm \,au}$ and semi-opening angles of $\sim 37^\circ$-$50^\circ$, matching observations of Tau 042021 and SY Cha. While the predicted line fluxes are somewhat lower than the observed values. These results suggest that photoevaporation is a viable mechanism for reproducing key features of observed H2 winds, including morphology and fluxes, though conclusive identification of the wind origin requires source-specific modeling. This challenges the reliance on geometrical structures alone to distinguish between MHD winds and photoevaporation. Based on our findings, we also discuss alternative diagnostics of photoevaporative winds. This work provides a critical first step toward interpreting spatially resolved H2 winds and motivates future modeling efforts.

Figures (19)

• Figure 1: Stellar spectrum adopted in our fiducial model. The vertical dashed lines mark the lower energy limits of the FUV and EUV bands. The lower limit of the X-ray band is $100\,\mathrm{eV}$. (See also Table \ref{['tab:fiducialmodel']} for the integrated luminosities.) The thin green line shows the photospheric emission, which overlaps with the blue line below $6\,\mathrm{eV}$, for reference.
• Figure 2: Snapshot of H2 density (top), gas temperature (second), H2 abundance (third), and poloidal velocity (bottom). Cyan contours indicate isosurfaces for each quantity: $n_{\rm H2} = 10, 10^2, 10^3, 10^4, 10^5, 10^6, 10^7 \,\mathrm{cm}^{-3}$; $T = 200, 500, 1000, 2000, 3000, 5000 \,\mathrm{K}$; $y_{\rm H2} = 10^{-4}, 10^{-3}, 10^{-2}, 10^{-1}$; and $v_{\rm p} = 1, 2, 5, 10, 20, 30 \,\mathrm{km}\,\mathrm{s}^{-1}$. In all panels, the white and green dashed contours denote the H2 dissociation front ($y_{\rm H2} = 0.25$) and the isothermal sonic surface, respectively. White arrows indicate the velocity field (direction only, not its magnitude), and are omitted where $T < 100 \,\mathrm{K}$. In the H2 abundance map, the gray dotted contour marks the HI ionization front ($y_{\rm HII} = 0.5$).
• Figure 3: Physical properties along streamlines. (a) H2 abundance map (same as Figure \ref{['fig:hydro']}) with selected streamlines shown in gray. Three representative streamlines are marked in blue, red, and cyan. (b)--(f) Profiles of cylindrical radius, densities, temperature, poloidal velocity, H2 abundance along the three streamlines, using consistent color coding. The horizontal axis indicates the distance along each streamline, with $s=0$ defined at the base. In panel (c), solid and dashed lines show H2 and hydrogen nucleus densities, respectively.
• Figure 4: Specific rates of major heating and cooling processes along the blue, red, cyan streamlines in Figure \ref{['fig:streamlines']}, shown from left to right. Labels indicate: EUV photoionization heating ("EUV"); X-ray photoionization heating ("X-ray"); line cooling from OI, H2, and CO; and adiabatic cooling ("adi"). (See Figure \ref{['fig:streamlines_heatcool_detailed']} in Appendix for the heating and cooling rates of other processes.)
• Figure 5: Intrinsic morphologies of the H2 pure rotational lines (S(1)--S(9)) for a disk viewed edge-on ($i=90^\circ$) without disk obscuration. The colors indicate the frequency-integrated intensities, with individual color bars scaled independently in each panel, Right ascension and declination offsets are shown assuming a source distance of $d = 140\,\mathrm{pc}$.