Unlocking accretion rate diagnostics for high-mass protostars using JWST/MIRI HI lines

Abstract

While many aspects of high-mass star formation have been investigated, the accretion onto the central protostars is one of the most fundamental but less explored physical properties. JWST/MIRI offers a unique opportunity to explore tracers of accretion at less-extincted wavelengths (5 to 27 um) than those studied so far. We probe the MIRI (MRS/IFU) capability to detect and resolve atomic Hydrogen (HI) emission lines in such embedded objects, to subsequently estimate accretion luminosities (Lacc) and accretion rates (Macc) for the first time in a sample of high-mass star forming regions at different evolutionary stages. We use dereddened HI line luminosities as tracers of accretion by applying existing line-to-accretion-luminosity relations (Lacc-calibrations). As they were originally established for low-mass Class II objects, we assess their applicability on our sample prior to estimating Macc. The infrared continuum reveals, at much higher spatial resolution than before, the location of new protostars, toward which we detect a handful of HI lines. While a few lines are secure detections, many are tentative. The most commonly detected line is HI 7-6, followed by HI 8-6 and HI 6-5. Assuming that their line fluxes are dominated by accretion, we find that two of the three existing Lacc-calibrations predict excessively high Lacc that largely exceed the corresponding L_bol, and that the third Lacc-calibration still overpredicts Lacc for some sources. Considering the given uncertainties, estimated accretion rates are only tentative. This work demonstrates the great potential of JWST/MIRI to probe HI line emission originated in the innermost regions of high-mass protostars, setting the ground floor for further investigations into accretion. While this project had the ambitious goal of robustly quantifying Macc, we have shed light on what outstanding methodological challenges remain.

Figures (14)

• Figure 1: Overview of the three less evolved regions of our sample: G28IRS2, G28P1, and G28S. In this case all of them belong to the same IRDC G28. Left: Spitzer (8 $\mu$m) view of the whole IRDC G28. The three target subregions are depicted with millimeter contours (white) from ALMA for G28IRS2 Molinari2025 and G28P1 Zhang2015, and from SMA for G28S Feng2016a, with cyan boxes highlighting the field of view of our MIRI observations, although slightly magnified for better visualization. Middle and right panels: zoom-in to the target subregions. They show the MIRI continuum at 14 $\mu$m, whereas cyan contours trace the continuum at 5 $\mu$m (no IR source was detected toward G28S). The ellipses in the bottom right corners represent beam sizes, grey for the 14 $\mu$m image, and cyan for the 5 $\mu$m contours.
• Figure 2: Overview of the three more evolved protostellar regions of our sample: IRAS18089, G31, and IRAS23385. Top panels: they show each region as seen by Spitzer (3.6 $\mu$m) or WISE (12 $\mu$m), with white contours depicting millimeter data from ALMA for IRAS18089 (1.2 mm; Sanhueza2021) and G31 (1.3 mm; Beltran2018), or NOEMA for IRAS23385 (1 mm; Beuther2018). The cyan boxes highlight the field of view of our MIRI observations. Bottom panels: MIRI integrated images at 14 $\mu$m for each region covering the corresponding upper panels. The cyan contours represent the continuum at 5 $\mu$m.
• Figure 3: Full MIRI spectrum of the G28IRS2 protostar (black). It shows several absorption features, where the broader one at 9.7 $\mu$m is produced by silicate grains. Blackbody components (3 in this case) absorbed by silicates and H$_2$O are fitted to reproduce the overall continuum. The blue curve represents the fit obtained by considering those spectral ranges that are predominantly absorbed by silicates and H$_2$O (in red, also indicated by the R1-R4 ranges). The green curve represents blackbody emission that would only be absorbed by silicates, while the dotted magenta curve shows the modeled blackbody emission unaffected by absorption. The black dashed curve above the main silicate absorption feature at 9.7 $\mu$m is an interpolation of the silicate-absorbed continuum from a five-order polynomial fit. Here, the red vertical line highlights the resulting depth of the feature that determines $\tau_{9.7}$, whose uncertainty is given by other two polynomials (order 1 and 4) fits passing above and below (grey curves). The black dashed line at R4 indicates the location of the secondary silicate absorption feature.
• Figure 4: The three most commonly detected H i lines across our sample. We show them using two of our regions. Left: line moment 0 maps. The channels used to integrate the emission are those encompassed by the line Gaussian fits in the right panels. Red apertures enclose the map area used to extract the spectra at the protostars shown in right panels (upper insets), while the broader grey aperture (minus the central red aperture) is used to extract the spectra of the local background (right panels, lower insets). Right: Gaussian curves fit the line emission for both protostar (red, upper inset) and backgrounds (grey, lower insets). We reproduce the Gaussian fits of the background in upper insets to allow comparisons with the protostar Gaussian fit. For some of the lines (middle row) a second Gaussian is required for fitting another nearby emission line. Other lines detected are shown in Figures \ref{['Fig:map_spec2']}, \ref{['Fig:map_spec3']} and \ref{['Fig:map_spec4']}.
• Figure 5: Line ratios normalized to the Hu$\alpha$ line, compared to the local excitation (KF11) and Case B models Storey-Hummer1995. Error bars are omitted, since the overall uncertainty is dominated by the discrepancy between the two extinction curves (triangle and circle markers). Solid markers represent detections with SNR $>5$, while open markers are detections with $3<$ SNR $<5$. Panels display models for temperatures of 10000 K (left) and 7500 K (right), and for the KF11 model we also include different density values ranging from n$_H = 10^9$ to $10^{12}$ cm$^{-3}$. We show the Case B model for a single electron density value $N_{\rm e}$ (in cm$^{-3}$) as this parameter appears to not change predictions significantly. Lines in the x-axis are organized by increasing wavelength.