PDRs4All II: JWST's NIR and MIR imaging view of the Orion Nebula

The JWST has captured the most detailed and sharpest infrared images ever taken of the inner region of the Orion Nebula, the nearest massive star formation region, and a prototypical highly irradiated dense photo-dissociation region (PDR). We investigate the fundamental interaction of far-ultraviolet photons with molecular clouds. The transitions across the ionization front (IF), dissociation front (DF), and the molecular cloud are studied at high-angular resolution. These transitions are relevant to understanding the effects of radiative feedback from massive stars and the dominant physical and chemical processes that lead to the IR emission that JWST will detect in many Galactic and extragalactic environments. Due to the proximity of the Orion Nebula and the unprecedented angular resolution of JWST, these data reveal that the molecular cloud borders are hyper structured at small angular scales of 0.1-1"(0.0002-0.002 pc or 40-400 au at 414 pc). A diverse set of features are observed such as ridges, waves, globules and photoevaporated protoplanetary disks. At the PDR atomic to molecular transition, several bright features are detected that are associated with the highly irradiated surroundings of the dense molecular condensations and embedded young star. Toward the Orion Bar PDR, a highly sculpted interface is detected with sharp edges and density increases near the IF and DF. This was predicted by previous modeling studies, but the fronts were unresolved in most tracers. A complex, structured, and folded DF surface was traced by the H2 lines. This dataset was used to revisit the commonly adopted 2D PDR structure of the Orion Bar. JWST provides us with a complete view of the PDR, all the way from the PDR edge to the substructured dense region, and this allowed us to determine, in detail, where the emission of the atomic and molecular lines, aromatic bands, and dust originate.


Introduction
Massive stars dominate the evolution of the galaxy through the injection of radiative and mechanical energy into their natal molecular cloud and surrounding interstellar medium (ISM). This feedback stirs up and heats the gas and limits molecular cloud lifetimes through photo-ionization and photo-evaporation, inhibiting future star formation. Feedback can also trigger star formation as gas is swept up in dense and massive shells. Hence, feedback is closely tied to the star formation efficiency of molecular gas (e.g., Elmegreen 2011;Hopkins et al. 2014).
Our understanding of stellar feedback is directly linked to studies of photo-dissociation regions (PDRs). PDRs are the regions where far-ultraviolet (FUV; 6 eV < hν < 13.6 eV) radiation from massive stars dominates the thermal processes and chemistry (see reviews by Hollenbach & Tielens 1997;Wolfire et al. 2022). PDRs separate the gas ionized by a star from the molecular cloud in which the star was born. Hence, stellar radiative energy is mainly deposited in a PDR, while the mechanical energy is transmitted through the PDR layer to the interiors of molecular clouds in the form of shock waves and/or turbulence. The "classic" PDRs are at the interface between the HII region and the molecular cloud, extending into the deeper molecular layers (e.g., Tielens & Hollenbach 1985b,a). PDRs however, appear in many other environments including reflection nebulae (Sheffer et al. 2011;Peeters et al. 2017), planetary nebulae (Bernard-Salas & Tielens 2005), surfaces of pillars and globules Schneider et al. 2021), the diffuse ISM (Wolfire et al. 1995(Wolfire et al. , 2003, and in protostellar and protoplanetary disks (Gorti & Hollenbach 2002;Vicente et al. 2013;Champion et al. 2017; see reviews by Öberg & Bergin 2021;Winter & Haworth 2022). In fact, most of the non stellar baryons in galaxies are in PDRs (Hollenbach & Tielens 1999). Thus, understanding the physics and chemistry of PDRs is critical for understanding the star formation history of the Universe.
With their wavelength coverage extending well into the midinfrared, the instruments on JWST are well suited to study the physical and chemical characteristics of PDRs. The warm (T ≃ 100 − 1000 K) and dense (n ≳ 10 3 cm −3 ) PDR gasmainly heated through photoelectric emission of electrons from small grains and molecules (Bakes & Tielens 1994;Weingartner & Draine 2001;) -is bright in the pure rotational transitions of H 2 , the mid-and far-infrared fine-structure transitions of atomic ions and neutrals (e.g., Si + , Fe + , C + , O), rotational transitions of CO and its isotopes, and rotational transitions of small radicals. Carbon is singly ionized in the PDR surface layers and the cascade generated by electron recombinations will produce a rich spectrum of C 0 lines (Cesarsky 1982;Walmsley et al. 2000). In addition, the strong FUV field produces bright fluorescence in the near-infrared ro-vibrational transitions of H 2 , atomic transitions of O and N, and the Aromatic Infrared Bands (AIBs), generally attributed to the emission of vibrationally excited polycyclic aromatic hydrocarbon molecules (PAHs) (e.g., Black & van Dishoeck 1987;Marconi et al. 1998a;Martini et al. 1999;Peeters et al. 2004;Tielens 2008;Habart et al. 2011). The adjoining ionized gas will show bright IR line emission produced by collisional excitation of fine-structure levels (e.g., Fe + , Ar + , Ar 2+ , Ne + , Ne 2 +, S 2+ , S 3+ ) and by recombination lines from HI and HeI (Martín-Hernández et al. 2002;Rubin et al. 2007). The spatial resolution of JWST (∼ 0.1 − 1 ′′ ) exceeds that of all other space telescopes over the same wavelength range 1 and is similar in spatial resolution to that of the Atacama Large Millimeter Array (ALMA) at submillimeter wavelengths. JWST has many IR filters centered on gas lines, molecular spectroscopic patterns as well as the continuum due to interstellar dust (emission and scattering). JWST emission line and continuum images of a PDR thus carry key information relevant to our understanding of the morphological impacts of stellar feedback, and JWST observations enable us to probe at unprecedented resolution how a molecular cloud is being disrupted by strong stellar UV radiation, winds, outflows and jets.
The focus of this article is on JWST NIRCam and MIRI images of the Orion Nebula complex carried out as part of the PDRs4All Early Release Science (ERS) program . The Orion Nebula complex is a nearby site of active star formation exhibiting many feedback processes and PDR interfaces (e.g., Pabst et al. 2019). The prototypical highly irradiated dense PDR in this nebula is usually referred to as the "Bright Bar" or "Orion Bar" (e.g., Elliott & Meaburn 1974;O'Dell & Yusef-Zadeh 2000). In the following, we refer to it as the "Bar." The ionizing and FUV radiation from the Orion Trapezium Cluster shines directly on the face of the Bar. At the outer layers, the ionized gas recombines at the ionization front (IF) and the gas becomes neutral hydrogen. This corresponds to the edge of the neutral PDR. The gas remains atomic (e.g., van der Werf et al. 2013;Henney 2021) until the H 2 dissociation front (DF), where the molecular hydrogen abundance increases rapidly. Over nearly 40 years, the Bar has been the target of many studies to elucidate the physical and chemical characteristics of PDRs (Parmar et al. 1991;Tauber et al. 1994;Young Owl et al. 2000;Bernard-Salas et al. 2012;Goicoechea et al. 2016;Parikka et al. 2018;Kaplan et al. 2021) and provides therefore a widely-used template for the observational signature of the interaction of stars with their environment, both in the Milky Way and galaxies out to high redshifts (Stacey et al. 2010;Vallini et al. 2018;Wolfire et al. 2022). With its wide wavelength coverage, high sensitivity, multiple filters and high spectral resolution resulting in large line-to-continuum ratios, JWST has the potential to provide a coherent vision of the structure of the Bar. Its structure includes the extended atomic layers (often called the "inter-clump" medium) and the thin emission layers of dense warm gas associated with the DF as well as the illuminated surfaces of large dense clumps.
The article is organized as follows. In Sect. 2, we describe the main physical characteristics of our target inferred from previous studies. The observations, data reduction and the fractional contributions of line, AIB, and continuum emission to our NIRCam images are described in Sect. 3. In Sect. 4, we describe the structures observed by NIRCam and MIRI within the inner region of the OrionNebula. In Sect. 5, we focus on the Bar as a template region to understand the structure and morphology of a strongly irradiated PDR. The complex transition from the IF, the PDR DF to the molecular cloud is studied, and we determine in detail the origin of the atomic and molecular lines, aromatic bands, and dust emission. In Sect. 6, we describe the photoevaporating proto-planetary disks observed in the whole NIRCam fields. A summary and conclusions are given in Section 7. In Appendix A, we show the template NIRSpec spectra presented in Peeters et al. (2023) in the wavelength domain of NIRCam filters, illustrating the variation of the contribution of different lines into each imaging band. Appendix B provides NIRCam and ground based images of the Bar. In Appendix C, we describe the structures observed within the NIRCam fields north of the Dark Bay, north of M42 and in M43 (Fig. 1).

M42 and the Bar
The Orion HII region, M42, -the nearest site of massive star formation -is illuminated by the Trapezium stars for which the O7-type star, θ 1 Ori C dominates (Sota et al. 2011), being the most massive and luminous member of the Trapezium cluster at the heart of the Orion Nebula (e.g., O'Dell 2001b). θ 1 Ori C has created a concave blister of ionized gas on the surface of the underlying Orion Molecular Core 1 (OMC-1) (see Wen & O'Dell 1995;O'Dell 2001b, and references therein), with the brightest portion called the Huygens Region (Fig. 1). The electron density varies across this region from a central high of almost 10 4 cm −3 to 3×10 3 cm −3 in the outer regions, while the electron temperature is usually about 9000 K (Weilbacher et al. 2015). Between the ionized gas and OMC-1 lies a PDR that is face-on to our line of sight. In the region of the Bar, the ionized atomic layer and its PDR are tilted almost along the line of sight (∼4 (1-8) degrees, Walmsley et al. 2000;Pellegrini et al. 2009;Salgado et al. 2016;Peeters et al. 2023), forming an escarpment in the Main Ionization Front (MIF) and due to projection effects, producing one of the optically brightest features of the Huygens Region. This also provides the ability to probe without overlapping the multiple layers of the PDR Hogerheijde et al. 1995, c.f. Fig. 5). The Bar is a strongly UV irradiated PDR viewed nearly edge-on.
The gas density (n H ) in the ambient molecular cloud is estimated to be n H = 0.5 − 1.0 × 10 5 cm −3 from a variety of IR and submillimeter gaseous emission lines (Tielens & Hollenbach 1985a;Bernard-Salas et al. 2012;Goicoechea et al. 2017). In addition, much denser small structures and molecular condensations ("clumps") are embedded in the Bar (n H ≳ 10 6 cm −3 ; Lis & Schilke 2003;Goicoechea et al. 2016;Joblin et al. 2018;Cuadrado et al. 2019). The Far-UV (FUV) radiation field incident on the PDR of the Bar is G 0 = 2.2 − 7.1 × 10 4 in Habing units (1.6 × 10 −3 erg cm −2 s −1 ; Habing 1968) as derived from UV-pumped IR-fluorescent lines of OI by Marconi et al. (1998b) and Peeters et al. (2023). Given the stellar characteristics of θ 1 Ori C and the far-IR surface brightness of the Bar, this places θ 1 Ori C at a physical distance of 0.33 pc from the far-IR dust emission (Salgado et al. 2016). The projected distance between the star and the ionization front (IF) is about 0.2 pc (c.f., Fig. 5). Beyond the IF, where the hydrogen gas converts from ionized to neutral, only FUV photons with energies below 13.6 eV penetrate the cloud. This corresponds to the edge of the neutral PDR but note that species with low ionization potentials (e.g., C, S, Fe) are still ionized in Article number, page 3 of 51 A&A proofs: manuscript no. main the surface layers of PDRs. Deeper in (∆A V ≃ 1 − 2 mag), the radiation field has been sufficiently attenuated by dust extinction that hydrogen goes from atomic to molecular. At a depth of ∆A V ≃ 2 − 4 mag, the carbon balance shifts from C + to C 0 to CO (Tielens & Hollenbach 1985b). This chemical stratification has been verified by an important series of infrared to radio observations (e.g., Jansen et al. 1995;Goicoechea et al. 2016). Relative to the molecular gas in OMC-1, the PDR gas flows through the ionization front at ≃ 1 km/s (Pabst et al. 2019) and once ionized accelerates away at about 7±4 km/s for the [N II] emitting layer close to the ionization front and 12±4 km/s for the [O III] emitting layer further out (O'Dell et al. 2020) as it joins the general expansion of the nebula. Close to θ 1 Ori C there is a low density bubble of gas of diameter of 0.2 pc shaped by its stellar wind (O'Dell et al. 2009(O'Dell et al. , 2020. This wind-blown cavity is open to the southwest and feeds the region of the Extended Orion Nebula (EON,Figure 1) where hot shocked gas has been detected (Güdel et al. 2008). A layer of ionized gas covers much of the Huygens Regions (García-Díaz & Henney 2007;O'Dell et al. 2020) and outside of this is the atomic layer of gas known as the Veil, one portion of which is a hemispherical bubble best described as the Outer Shell, which was discovered by Pabst et al. (2019) and expands at about 13 km/s away from the OMC. The Veil has a column density of N H = (2 − 6) × 10 21 cm −2 , depending on the direction, and obscures the Huygens Region by 1-3 magnitudes of visual extinction (O'Dell et al. 1992;O'Dell & Yusef-Zadeh 2000;Weilbacher et al. 2015). The physical characteristics of the Bar are summarized in Table 2.
Assuming that the background PDR directly behind the Trapezium stars is a face-on PDR, the incident FUV field is estimated to be 10 5 Habings from the observed far-IR surface brightness (Tielens & Hollenbach 1985a); a factor of ≃ 2 − 5 higher than the FUV field incident on the Bar. The gas density in this PDR is estimated to be slightly higher (10 5 cm −3 ) than in the Bar (Tielens & Hollenbach 1985a), concomitant with the higher density of the ionized gas (Weilbacher et al. 2015).
The wide field of view of the JWST images include the nearby low-ionization HII region M43 (NGC 1982) powered by HD 37061 (also known as NU Ori), a B0.5V star (Fig. 1). M43 lies to the northeast of M42 and this object has not been well studied but we include analysis of this JWST data in this article as well. M43 is seen to be shielded from illumination by θ 1 Ori C by the northeast portion of the wall bounding M42.
M42 exhibits several high-velocity features, including microjets, large scale Herbig-Haro flows, and wind driven shocks (e.g., Bally et al. 2000;O'Dell 2001b). Protostellar jets and outflows emanate from dust-enshrouded, nascent stars. Shocks are formed where the collimated flows interact with the nebula's ambient ionized gas and the neutral foreground veil. Additionally, uncollimated flows from the low-mass accreting stars and the stellar wind from θ 1 Ori C produce shocks. M42 further shows structures resulting from embedded sources of outflow in the BN-KL and Orion-S regions.
As this summary demonstrates, many questions remain on: (i) the detailed geometry of this highly irradiated and very structured PDR; (ii) the best tracers of the different physical zones (H + , H 0 , H 2 , C + , C, CO); (iii) the physical and chemical conditions in these different zones, particularly at the ionization and dissociation fronts; (iv) the relationship of the various components (interclump, clumps, proplyds, winds & jets) populating this region. JWST, with its high spatial resolution, can uniquely address these issues and thereby provide valuable insight in the physical and chemical processes taking place in FUV irradiated, interstellar material.

Observations
We provide in this section a summary of the main parameters of imaging observations obtained within the ERS project #1288 "Radiative Feedback from Massive Stars as Traced by Multiband Imaging and Spectroscopic Mosaics," based on NIRCam and MIRI observations . The details of the observations can be retrieved from STScI using the Astronomer Proposal Toolkit (APT) under program ID 1288.
The telescope orientation (V3 Position Angle) was left unconstrained between 260 and 270 degrees. At the time of observation, 10-11th of September 2022, the telescope was oriented at about 265 degrees, resulting in the pointing illustrated in Fig.  2.
For NIRCam observations, we refer in particular to the data obtained under Observation 1 "NIRCam Orion Bar Imaging,", whereas for MIRI we refer to Observations 15 and 16 "MIRI Orion Bar Imaging." These two MIRI observations were executed with NIRCam in parallel on September 11th for the F1500W and F2550W filters and September 18th for the F770W and F1130W filters, with a difference between these two dates for the telescope orientation on the ecliptic plane of about 3 degrees. Due to the V3 orientation, the parallel NIRCam field observed covers the northern part of the Orion Nebula, near the M43 (NGC 1982) region. A second set of background images, Observations 14 and 17, were taken at an offset field about 2 degrees west of the Bar using again MIRI and NIRCam in parallel with the same parameters adopted for Observation 15 and 16, respectively. However, Observations 14 could not be achieved due to Fine Guidance Sensor (FGS) loss of fine guidance control, the background observations in the corresponding MIRI and NIRCAM filters are therefore not available. An overview of the imaging filter selection is given in Table 1.

NIRCam imaging
The selected NIRCam filters cover i) the 3.3-3.4 µm Aromatic Infrared Bands (AIBs), ii) the ro-vibrationally and rotationally excited lines of H 2 1-0 S(1) at 2.12 and H 2 0-0 S(9) at 4.69 µm, tracing the dissociation front, iii) the [FeII] line at 1.64 µm, tracing the ionization front, and iv) the Paschen Pa α and Brackett Br α atomic hydrogen lines, tracing the H ii region. Each filter was paired with a reference filter centered on an adjacent wavelength for subtraction of the underlying continuum emission. We mapped the PDR region with a single pointing using a 4 point primary dither (Fig. 2). To avoid saturation, we used the RAPID readout mode with two groups per integration and two integrations per exposure. With 4 dithered pointings, this corresponds to 8 total integrations. The integration time per pixel from the reset to the second sample was about 21.47s, corresponding to a total exposure time of 171.788 s. The minimum resulting signalto-noise ratio on the extended emission toward the Bar template regions covered by NIRSpec is given for all filters in Table 1.

MIRI imaging
Data were obtained in i) the 7.7 and 11.3 µm filters including AIBs which, when combined, could provide a proxy for PAH ionization (e.g., Joblin et al. 1996;Galliano et al. 2008) and Article number, page 4 of 51 ii) the 15 and 25 µm filters dominated by continuum emission tracing warm dust in the H ii and neutral region, similarly to the corresponding WISE, Spitzer, and IRAS filters. We obtain a 3 × 3 mosaic using a three point dither pattern (3-POINT-MIRI-Article number, page 5 of 51 A&A proofs: manuscript no. main F770W-WITH-NIRCam) (Fig. 2). To prevent saturation given the brightness of the Bar, we use the FASTR1 readout pattern and the SUB128 imaging subarray.

NIRCam parallel observations
We obtained parallel NIRCam observations with the on-source MIRI imaging. The adopted filters cover i) the 3.3 and 3.4 µm AIB, ii) the vibrationally excited line of H 2 1-0 S(1) at 2.12 µm, and iii) the Pa α and Br α lines. Also in this case, each filter was paired with a reference filter to estimate and be able to subtract the underlying continuum emission. The pointings, number of dithers, and dither pattern were set by those of the primary observations (on-source MIRI imaging). To accommodate the brightness of the Bar, we used the BRIGHT2 readout mode with two groups per integration, one integration per exposure, and the three dithered positions. The effective exposure time in this case results is 1159.569 s, corresponding to 580 s with correlated double sampling.

NIRCam
Given the evolving nature of the automatic pipeline producing the data available on MAST, we have chosen to reduce our observations starting from the original _uncal files, that is those produced by the preliminary Stage 0, adopting the latest available development version of the pipeline at that time 1.7.3. Stage 1 corrects instrument signatures that need to be treated at the level of individual groups before ramp fitting, such as for dark current, nonlinearity detector response, and cosmic ray (CR) events. Given the intensity of the signal, we turned off the suppress_one_group option of the ramp fitting step to recover signal saturated after the very first sample. Stage 2 operates on the count rate images produced by Stage 1 removing the background -if dedicated files are provided-, calibrating each exposure individually to produce images in physical units of MJy/sr and rectifying the images for final combination with other images. The last stage combines the rectified exposures from multiple exposures, performing outlier pixel removal and astrometric alignment to produce the final products, i.e. drizzled and mosaicked images with their associated data catalogs and segmentation maps. We have chosen to independently combine the different dithers for each module and each filter. We have also tested the impact of different parameters. First, we have tried bypassing the outlier detection step. This step allows finding some bad pixels and cosmic-rays not corrected in the previous stage. Without this step, the final images show a strong "salt and pepper" pixeling effect. On the short wavelength filters (F162M, F164N, and F212N), we detected some artefacts identified as wisps. They are due to straylight and are located at the same positions in each detector. The wisps are mainly on the B4 and A3 detectors and their positions depend on the filter, the detector, and the observation. To correct them, we selected polygons encompassing the wisps by hand. Then, we flagged these polygons as DO_NOT_USE in the DQ array. The wisps were removed because they were not considered further but as a consequence the noise was higher in these areas in the mosaic, because there is less information. We corrected for the 1/f noise by subtracting the median value of each row and then each column before the JWST pipeline stage 3 process. We first took care to apply a mask on all the bright sources when computing the median values for row and column subtractions and carefully inspected the results for any unintended consequences.
Finally, we improved the world coordinate system (WCS) alignment between the different mosaics by detecting unsaturated point sources in all detector images and comparing their positions with the astrometry from the Gaia DR3 catalog. The astrometry correction was at most of 0.3 ′′ . To get the fluxcalibrated line maps continuum subtracted in erg cm −2 s −1 sr −1 , the maps (in MJy/sr) are multiplied by 10 6 ×10 −23 ×∆ λ×c/(λ c ) 2 with ∆ λ and λ c the bandwidth and pivot wavelength of the filters from the NIRCam manual.

MIRI
As for the NIRCam observations, we reduced our observations starting from the raw data using the latest version of the pipeline and reference files available at that time (February 2023). The inter-pixel capacitance (IPC) step, correcting for the interpixel capacitance, was skipped in stage 1 as recommended by the instrument team because of poorly defined deconvolution kernels. Similarly, the Reset Switch Charge Decay (RSCD) step correcting for some ideal detector and readout effects was skipped because of the low number of groups per integration in our observations. Dedicated backgrounds were observed for the F1500W and F2550W filters and were subtracted from the data in stage 2 of the pipeline. As mentioned in Section 3.1, background observations in the F770W and F1130W filters were not successful due to FGS loss of fine guidance control. The tweakreg step in stage 3 used to improve the alignment of the different input images was skipped because of the lack of point-like sources in the different filters, as well as the outlier detection step as we observed a deterioration of the final mosaics when applying this correction. For the F2550W filter, as most of the groups reached the saturation level for the pixels along the Bar, we skipped the jump step because of poor performance.
Moreover, we observed strong edge-brightening effects in the final mosaics attributed to straylight features in the SUB128 array flat fields (the SUB128 array is located at the top left of the detector), stronger at longer wavelengths (F1500W and F2550W filters). A first solution was to flag the affected left columns and top rows as DO_NOT_USE (about 15 rows and columns) so that they are not used in the final mosaic. For the F2550W a region around the Lyot mask has also been flagged. The number of rows/columns to flag has been selected for each filter as a trade off between having sufficient pixels for the overlap and remov-ing pixels with poor quality. A better solution, used in this article, is to use the mosaics for flat field estimations with backprojection (projection of the computed mosaic back to the detector at the central position). Then the mosaics are recomputed using the estimated flat fields. Repeating this process with an additional deflagging of the affected columns/rows in each iteration, significantly improved signal recovery.

Line, AIBs, and continuum emission contributions to imaging bands in the NIRCam filters
Here we summarize our results regarding the fractional contributions of line, AIBs, and continuum emission to our NIRCam images (for a complete description see the Science Enabling Product 4 article, Chown et al. 2023). We computed synthetic NIR-Cam images from the stitched and co-added NIRSpec mosaic from Peeters et al. (2023). We did this by applying Equation 5 of Gordon et al. (2022) to each spaxel. This approach is similar to what has been done with Spitzer/IRS data 2 . We call these images "synthetic images" in order to distinguish them from images observed by NIRCam. We then decomposed each spaxel into line and continuum components (based on local baseline fit around prominent lines), and then calculated synthetic images from each of the line and continuum components. The fractional contribution of emission component i to imaging band j, is given by the ratio of the synthetic image of component i in band j to the synthetic image in band j of the total spectrum. Note that this measurement relies solely on NIRSpec data, and so it is not affected by any differences from NIRSpec and NIRCam data (such as flux calibration, resolution, etc.) -we investigate such differences in Chown et al. (2023). is well-correlated with the true Br α line intensity. In the Bar (on-target) observations, the F480M filter can be used to trace the continuum underlying F405N. F480M traces the continuum well with a small contribution from an H i line or an H 2 line ( Fig. A.3). F405N-F480M flux is overall a good measure of Br α intensity, but suffers from more scatter than the F405N vs. Br α correlation. -H 2 1-0 S(1) 2.12 µm: The fractional contribution of continuum emission in F212N is higher than that in F210M over a large area. This is due to bright H i and He i emission lines that fall in the F210M filter but fall outside of the F212N filter ( Fig. A.4). This results in a negative flux in the F212N-F210M image in a significant fraction of area closer to the exciting sources. Note also that the He lines that are close to H 2 2.12 µm contribute to F212N at a comparable degree to the H 2 line except in regions that are sufficiently far away from the exciting sources. A detailed analysis of a full set of lines will be presented in the science enabling product and the associated article (Chown et al. 2023).

Morphology of the Orion Nebula inner region
With their high angular resolution and unparalleled sensitivity, NIRCam and MIRI unveil the structures at very small scales of the Orion Nebula (0.1 to 1 ′′ from 2 to 25 µm, equivalent to 2 × 10 −4 to 2×10 −3 pc at the Orion distance of 414 pc). It displays an incredible richness of substructures, as well as previously hidden stars and even background galaxies. In this section, we present several prominent features arising in the images within the inner region of the Orion Nebula (M42), i.e. the fields centered on the Bar. Fig. 3 shows composite NIRCam images in three selected filters (F187N, F335M, and F470N). The F187N filter captures the distribution of ionized gas via the bright Pa α, F335M traces mostly emission from the AIB 3.3-3.4 µm aromatic and aliphatic CH stretching mode bands and F470N traces the dust continuum and the high excited H 2 0-0 S(9) pure rotational line. Fig. B.1 shows composite MIRI images in the three selected filters F770W, F1130W, F1500W. These filters image the 7.7 and 11.3 µm aromatic bands and the continuum emission from hot/warm dust at thermal equilibrium. The emission at 15 µm is mainly produced by very small carbonaceous grains whereas at 25 µm slightly larger grains can contribute (e.g., Compiègne et al. 2011). In Appendix B, all the images obtained in the filters listed in Table 1, and for some gas lines continuum subtracted, are presented.
A schematic view of the Bar inferred from both these JWST observations and the literature on previous observations from visible to millimeter is presented in Fig. 5. We are viewing the main ionization front almost edge-on along the Bar.

ionized gas, crenellations, bow-shocks and YSOs
In Fig. 3, the ionized gas (in blue) comes from the MIF extending from the Trapezium grouping of stars to the Bar. Beyond the Bar, the MIF is primarily photoionized by Θ 2 Ori A (O'Dell et al. 2017a). Due to intense ultraviolet and ionizing radiation, hot and ionized gas is photoevaporating away from the MIF. H i and He i lines observed toward the Huygens Region at visible-wavelengths are blue-shifted, by ≃10 km s −1 , with respect to the molecular gas emission ). This velocity difference approximately agrees with that inferred from observations of H and He radio recombination lines (e.g., Cuadrado et al. 2019).
In the NIRCam images, ionized gas flows from the IF of the Bar (see panel G in Figs. 3 and 4) are not easy to discern because they are seen in the foreground of the MIF emission. The NIR-Cam and MIRI images do not show any AIBs and H 2 emission that would be associated with the photoevaporating flows from the IF of the Bar. The AIBs and H 2 emission in front of the Bar most likely originate from the surface of the OMC-1. This surface is perpendicular to the line of sight and is illuminated by the Trapezium cluster, making it a face-on PDR 3 .
On the background OMC-1 surface, several structures as shown in panels A and B in Fig. 4 are spatially resolved. These types of features are not the only ones and several of them were observed at visual wavelengths with the HST. They were called "Crenellations" by O'Dell et al. (2015). The interaction of collimated jets and outflows from protostars inside the molecular cloud likely drives the shocks that create these structures at the surface of the cloud (see also Kavak et al. 2022b,a). In order to study whether some of these structures are "apparent" structures produced by extinction variations or the edges of dense molecular gas structures, high angular resolution molecular line tracers are needed.
One of the largest outflows with highly blue-shifted features (e.g., O'Dell et al. 1997;O'Dell & Henney 2008) is HH 203 and 204, localized southwest of θ 2 Ori A, are well seen in the Pa α, Br α and [FeII] line NIRCam images and in the continuum filters at shorter wavelengths (panel F in Figs. 3 and 4 and Figs. in Appendix B). HH 203 is a well defined bow-shock with low ionization characteristics at its end. It is driven by a high-velocity jet that emerges into the area ionized by θ 1 Ori C or the nearest (in the plane of the sky) star θ 2 Ori A. HH 204 is almost at the same position angle as HH 203 but shows a different structure. Its top is a flocculant structure (that includes low ionization characteristics) and no jet is visible. These structures are not detected in H 2 line emission ( Fig. 9 shows the H 2 emission in the map and the spatial profile towards the cut 5 which indicate that the H 2 emission is associated with the Bar and not with the bow shock of HH 203). This supports the suggestion that these flows interact with the ambient ionized gas of the nebula and not the molecular gas (e.g., O'Dell & Henney 2008).
NIRCam images in the F470N and F480M filters reveal for the first time strong emission in the surroundings of the bright star θ 2 Ori A (at about ∼ 8 ′′ or 0.015 pc from the star for a dis-3 The emission from this background face-on PDR was previously observed with Herschel in other PDR tracers, especially in high-J CO, CH + lines (Parikka et al. 2018 Pellegrini et al. 2009). It displays the main features discussed in detail in the core of this article, and inferred from both the imaging observations (this work) and NIRSpec spectroscopic observations . We note that for clarity the dimensions perpendicular to the Bar are not to scale, the true spatial scales being explicitly given in the annotations. The adopted distance to the Bar is 414 pc. In addition, the sketch does not include foreground material, which includes a layer of ionized gas (O'Dell et al. 2020) and, closer to the observer, layers that are grouped together under the designation as the Veil (e.g., van der Werf et al. 2013;Pabst et al. 2019Pabst et al. , 2020.    Menten et al. (2007). Großschedl et al. (2018) suggested that the about 10 pc difference compared to the literature value can be seen as an estimate of remaining systematic uncertainties. In Appendix A of Kuhn et al. (2019), they compare their distance estimate to one from Kounkel et al. (2018) and discuss effects of the threedimensional structure of Orion A. Moreover, one should note that these distances correspond to the distances of the stars and not the one of the molecular cloud and the Bar. New Gaia releases (in combination with new trigonometric parallax) will provide even more accurate distance determinations. For simplicity, we choose to assume the distance of 414±7 pc from Menten et al. (2007) in order to remain consistent with . The difference in distance values has anyway no important implications for the results presented in this article. tance of 414 pc). These previously unknown features are also visible in F277W, F300M, F323N and F335M filters. A stellar wind from the θ 2 Ori A star most likely forms these features. A bow wave around the θ 2 Ori A star is probably moving into the MIF. The dynamical and radiative impact of θ 2 Ori A is influencing and complicating its nearby environment. O'Dell et al. (2017a) showed that foreground objects near this location are illuminated by θ 2 Ori A.
A very bright substructure located in the southern part of the Bar shows a very particular structure ∼1000 au in size for a distance of 414 pc (see panel C in Figs. 3 and 4). This structure is also visible with ALMA (see Fig. B.5 where it appears as a globule) and might correspond to the surroundings of a protostellar source embedded in the highly irradiated environment of the Bar (Goicoechea et al. in prep). No MIRI point source is detected at this position, probably because the extended emission from the Bar is very bright and dominates the continuum emission. This indicates that this YSO is most likely too cold to emit strongly at mid-infrared wavelengths. The structures around this YSO are very bright in the H 2 0-0 S(9) and 1-0 S(1) lines in the NIRCam and Keck ) maps (see Fig. B.5) but also in the AIB emission (see panel C in Fig. 4). This emission likely arises from a combination of irradiated shocks from the outflow and PDR emission. Finally, several bright emission features associated with embedded proplyds are detected in the Bar and in front of it (e.g., panel G in Figs. 3 and 4). The proplyds detected within the inner region of the Orion Nebula are discussed in Sect. 6.

Crenellations on the Bar
One of the most striking features from the NIRCam and MIRI images, and in particular in filters probing AIBs and H 2 emission, is that the molecular cloud border in the background and the Bar appears hyper structured, most likely turbulent (e.g., Goicoechea et al. 2016). In the Bar, lots of patterns are apparent such as crenellated structures or ridges (e.g., panels D, E in Figs. 3 and 4). The upper west corner of panel E in Fig. 4 coincides with a region where the Bar has no sharp boundary, a region labeled "SW-Gap" in Figure 24 of O'Dell et al. (2015). In this zone, a number of crenellated structures were seen with HST. These features are both detected in the NIRCam filters dominated by the ionized gas and continuum below 3µm. As suggested by O'Dell et al. (2015), these structures are very likely bow shocks forming in the tilted portion of the Bar or the foreground Veil. On the other hand, in the regions further inside the Bar (e.g., panel D or panel E from upper east corner to west corner), the structures seen in the NIRCam and MIRI filters probing AIBs and H 2 lines emission likely correspond to the edges of dense molecular gas inside the Bar. Most of these structures are in fact well seen in the submm HCO + (J=4-3) emission (with ALMA, see Sect.5) which is sensitive to molecular gas density variations.
tures. The F335M filter is essential for bringing out the hightextured structure of the UV irradiated molecular cloud surfaces. The aromatic emission traces a combination of cloud density and strength of the local FUV field (see Sect. 5.1). The F335M filter provides one of the highest resolution views of the outer molecular layer of the PDRs available with JWST. Intricate fine details of how interstellar matter is structured at small scale is thus revealed. At the cloud edge, AIB emission is more restricted to the atomic H layers of the PDR than the H 2 line emission (see Sect. 5). With the high incident FUV radiation field on the Bar, molecules in general are expected to survive longer in the shielded environment offered by the dense Bar or OMC-1. However, emission in highly rotationally and ro-vibrationally excited H 2 lines require FUV photons to be pumped. These lines are therefore observed at the photo-dissociation front more specifically at the H 0 /H 2 transition where atomic hydrogen becomes molecular. Consequently, the subtracted continuum H 2 line images (e.g., the H 2 0-0 S(9) and 1-0 S(1) lines at 4.69 and 2.12 µm) highlight the irradiated edges of dense molecular structures. A detailed comparison between the atomic and molecular phase tracers across the Bar PDR is given below in Sect. 5.

Transition from ionization front to H 0 /H 2 dissociation front
In this section, we focus on the Bar, a prototypical highly irradiated dense PDR. Figs. 7 to 11 and B.5 show close-up maps and surface brightness profiles of the Bar viewed edge-on. A gradual structure is evident when moving away from the excitation source as the ionization front, the AIB and H 2 emission layers appear successively, in agreement with previous studies (e.g., ). However, instead of a smooth PDR transition, multiple ridges in AIB and H 2 emission are spatially resolved for the first time. In addition JWST unambiguously reveals very sharp edges (on scales of ∼1 ′′ or 0.002 pc) and rich small-scale structures (with typical widths of ∼ 0.5 − 1 ′′ or ∼0.001-0.002 pc). This is in agreement with previous high spatial resolution ALMA HCO + emission maps (Goicoechea et al. 2016) which showed a highly sculpted interface. Along with analysis of high-J CO and CH + Herschel observations and H 2 pure rotational lines from ISO, Joblin et al. (2018) concluded that the emission of these tracers arises from a thin (a few 10 −3 pc), high thermal pressure (P∼10 8 K cm −3 ) layer at the surface of the molecular region. However, the corresponding structures were unresolved in most tracers until now.
In the following, we use the notations of the scheme in Fig. 5 and Table 2 for the physical quantities related to the Bar PDR. In particular, the assumed distance to the Bar is 414 pc (Menten et al. 2007). This value is slightly higher or similar to more recent estimates using Gaia observations (see annotations in Table 2. We use the notation l los PDR the length of the PDR along the line of sight toward us and d PDR for the depth in the PDR from the IF. N los H and A los V are the column density and visual extinction along the line of sight toward us. N H and A V are the column density and visual extinction in the UV illuminating direction (perpendicular to the Bar).

Spatial distribution of the AIB emission
We first analyze the surface brightness profiles of the filters probing the AIBs and continuum emission along the direction going away from the sources of UV illumination. In order to probe variations across the PDR, several cuts perpendicular to the Bar from southwest to northeast are shown in Figs. 7-8. This allows us to probe the flux variations along the entire illuminated interface. An approximate position of the IF has been marked in maps and profiles as a vertical line, which corresponds to the emission peak of the [FeII] 1.64 µm line and the rise of the AIBs and gas lines in the atomic zone (e.g., [OI] 63 and 145 µm). The different components observed in the AIB emission profiles in the different PDR zones are annotated (Fig. 6) and discussed next. Our high spatial resolution JWST study permits filling in the near IF portion of the Bar's PDR that the study of Henney (2021) based on Herschel data did not allow (c.f. his Figure 5).

Steep density rise at the IF
At first glance, the spatial distribution of the AIB filter emission profiles follows the same trend as we move across the Bar, a steep increase at the IF position, followed by a slower decrease (with a typical scale of 10 ′′ or ∼ 0.02 pc, see Figs. 6, 7 and 8). At first order, the AIBs surface brightness is proportional to the column density of the band carriers and to the local FUV flux strength, I ∝ N los H × [C/H] × G 0 × e −τ FUV with N los H = n H × l los PDR and [C/H] their carbon abundance. The column density is along the line of sight to us while the opacity τ FUV is in the FUV illuminating direction, perpendicular to the Bar. The initial brightness increase is expected due to a large increase in dust column density at the IF, while the slower decrease is expected due to the extinction of the incident UV irradiation field. What is noteworthy is that the emission spatial profiles show an extreme climb over few sub-arcsecs just after the IF. The emission peak arises 1 ′′ to 2.5 ′′ (0.002 to 0.005 pc) from the IF depending on the position in the Bar (see Fig. 7). Such a sharp density rise is expected due to the sharp decrease in gas temperature at the IF if the thermal pressures in the ionized and neutral regions are of similar magnitude, as discussed below. In addition, the extreme rise of the emission profile could put strong constraints on the tilt angle θ of the PDR between the plane of the irradiated surface and the line of sight. This is discussed in  and will be investigated in more detail in a future work.

FUV dust extinction and density in the atomic PDR
Just after the rise, one can clearly observe in the AIB profiles a rapid decrease which must result from the FUV radiation field decreasing with depth inside the PDR as a consequence of the dust extinction. Fluxes in filters centered on AIBs are most likely dominated by the emission produced by (sub)nanometric particles including large molecules (PAHs) stochastically heated, which is proportional to the FUV radiation field strength. To visualize the decay of the FUV flux, the curve exp(−τ FUV )=exp(−σ H N H ) with σ H the dust FUV extinction cross-section per proton, and N H the column density from AIB emission peak (along the UV illuminating direction, perpendicular to the Bar), is plotted on Fig. 7 (dotted line). The parameters employed to reproduce the initial part of the observed decay are given in Table 2. We assume σ H , R V and A V /N H in agreement with the extinction curve measured in the Orion Nebula by Cardelli et al. (1989) and as refined by Blagrave et al. (2007). These values differ from that of the average measured in the ISM (R V = 3.1 and A V /N H = 5.3 × 10 −22 mag cm −2 ) and lead to an increased penetration of FUV photons compared to the dust found in the diffuse ISM 4 . Then, in order to compute the column A&A proofs: manuscript no. main density N H , we assumed after the initial rise in density (i.e., after the emission peak) a constant density n H in the atomic PDR as in Arab et al. (2012) and Schirmer et al. (2022). N H is then given by n H × d PDR with d PDR the distance from the emission peak. The density n H is adjusted to reproduce the initial part of the observed decay (d PDR ∼0.002-0.01 pc). We derive n H =(5-10)×10 4 cm −3 (see Fig. 7). The density value range we derived in the atomic PDR is in agreement with the location of the H 0 /H 2 transition obtained with NIRCam (see Table 2) as discussed in Sect. 5.2.3 as well as estimates from atomic gas FIR lines (e.g., The density we derived in the atomic PDR is significantly higher (factor 10-20) than the electron density n e derived at the IF which is about ∼ 5 × 10 3 cm −3 (see Figs. 26-27 in Weilbacher the THEMIS dust evolution model (Jones et al. 2017). Despite the poor resolution, a radiative transfer model of the Bar including the evolution of dust grains (size, abundance, properties) allowed them to highlight a strong depletion of the subnanometric grain population and a size distribution shifted towards larger particles compared to the diffuse ISM. A more recent study with JWST data confirms this result (El Yajouri et al. in prep.). et al. 2015). In a D-critical IF 5 (that can be expected in a blister HII region), the pressure in the neutral gas at the close-back of the IF is expected to be a factor of 2 higher than in the ionized gas in the close-front of the IF. A strong rise in density is then expected to compensate for the much stronger temperature decrease between the ionized and neutral region. In that case, a pressure of about ∼ 2 × 10 8 K cm −3 is expected in the neutral atomic region 6 , and given our density estimate, this would mean a temperature of a few 10 3 K.

Extended emission and secondary peaks towards the molecular region
Another important characteristic of the observed emission profiles is that when entering the PDR, the emission decays to a non zero plateau that extends into the molecular region (Figs. 6 and 7). This extended emission is most likely due to irradiated 5 In the usual classification of IF types (first established by Kahn 1954;Axford 1961, and summarized in usual textbooks, e.g., Spitzer 2004; Draine 2011), a D-critical front travels at subsonic speed with respect to the neutral gas, while the ionized gas is expelled at the speed of sound with respect to the front. 6 The electron temperature T e derived at the IF is of the order of ∼ 9 × 10 3 K (see Figs. 24-25 in Weilbacher et al. 2015) which gives a thermal pressure of P th/k = 2 × n e × T e ∼ 10 8 K cm −3 in the neutral gas.  atomic material along the line of sight located in front of the molecular region, in the foreground face-on PDR surface layer (as seen in geometry on Fig. 5). This emission may originate in the flattened region that is still illuminated directly by the ionizing stars. The MIF turns up at the Bar and then continues a flatter rise further away (as illustrated in the lower part of Figure 13 in O'Dell & Harris 2010). A detailed Spitzer study revealing what is behind the Bar and supporting this geometry is published by Rubin et al. (2011). For AIBs and dust continuum emission, the further inside the Bar, the greater the face-on PDR contribution with the FUV radiation in the Bar being rapidly attenuated.
Furthermore, in the decaying part of the profile, several secondary peaks are visible ( Fig. 6 and 7). This might be associated with multiple irradiated ridges with varying densities. These ridges are located after the main edge in the FUV illuminating direction. The sub-peaks in the region where the hydrogen is mostly molecular (d PDR > 0.02 pc) spatially and individually coincide with strong H 2 line emission peaks (Fig. 9), hinting that these AIB emission peaks arise from the material at or close to the DF. Some AIBs sub-peaks are very pronounced (such as the one at d PDR = 0.04 pc, Fig. 7, panel b). For those, the AIB emission sub-peak is observed slighlty shifted (by ∼ 0.2 ′′ ) from the H 2 emission peak (i.e. H 2 is observed closer to the Trapezium). In order to investigate this shift in detail, radiative transfer calculations are required.

Excited dense molecular gas
In this section, we analyze the distribution of the excited dense molecular gas traced by H 2 emission in order to probe the gas physical structure and the location of the key chemical transitions occurring in the molecular PDR. With JWST, we are for the first time able to spatially resolve the emission profiles of both the high rotationally and vibrationally excited H 2 lines (see Figs. 9,B.5 and 11). Our NIRCam observations show a very good agreement with Keck/NIRC2 observations  in terms of vibrationally excited line distribution and intensity. In the following, we examine the highly structured H 0 /H 2 dissociation fronts towards the Bar as well as the remarkably similar spatial distribution between the highly rotationally (0-0 S(9)) and vibrationally (1-0 S(1)) excited H 2 lines and the HCO + line J=4-3 emission observed with ALMA. The spatial distribution of the emission lines as a function of the geometry of the DF surface layer, density variation with depth into the PDR and extinction along the line of sight is now discussed in detail.

Highly structured H 0 /H 2 dissociation fronts
In Figs. 9, 10 and 11 the NIR H 2 line emission (delineating the H 0 /H 2 transition) show several bright ridges which are spatially resolved and small scale structures. The H 0 /H 2 fronts appear highly structured with several ridges and the emission rise is extremely sharp with a width of 0.5 to 1 ′′ (0.001-0.002 pc or 200-400 AU). The ridges run parallel to the Bar but a succession of bright substructures is also observed from the edge towards Article number, page 15 of 51 A&A proofs: manuscript no. main the molecular region. This is particularly clear in the southwest part of the Bar which corresponds to the upper part of the map displayed in Fig. 9. In this area, the structure of the Bar is very complex and irregular. The H 2 emission ridges appear in an area that starts at about 10 ′′ from the IF (d PDR =0.02 pc) and up to 20 ′′ (d PDR =0.04 pc) as shown in Fig. 9. We interpret the three main ridges that appear as three edge-on portions of the DF surface which are successively more and more distant from the IF. These edge-on portions of the DF, denoted DF1, 2, 3 thereafter, are located at a projected distance from the IF of d PDR ∼0.02 pc, 0.027 pc, and 0.038 pc respectively, as indicated by vertical dashed lines in Figs. 9 and 10. The DF2 is the one which coincides best with the average position of H 2 emission ridges all along the Bar.
A terraced-field-like structure. A terraced-field-like structure with several steps seen from above as shown in Fig. 5 can explain the succession of H 2 ridges across the Bar. In that geometry, each H 2 emission ridge corresponds to a portion of the DF seen edge-on, i.e., a step. Since highly rotationally and rovibrationally excited H 2 emission profiles are sensitive to the gas density, the very narrow and bright ridges must be due to irradiated dense material. For low density interface, the H 2 emission is spatially more extended and weaker. In the isobaric hypothesis, as discussed previously, the gas density rises as the gas cools at the DF (e.g., Allers et al. 2005;Joblin et al. 2018).
Additional evidence for the terraced-field-like structure comes from the difference in visual extinction A los VBar along the line of sight (see Fig. 5 and Sect. 5.2.2; Peeters et al. 2023), as well as the comparison between NIR and millimeter data (Sect.

5.2.4), showing that A los
VBar is higher for the DF1 than for the DF2 and the DF3 dissociation fronts. Furthermore, DF1 remains visible in the NIRCam filter F335M but is no longer discernible in the F210M filter (see Fig. 10). This last filter is most likely due to the dust scattered light. The fact of not seeing DF1 in F210M confirms that there is more material along the line of sight at the DF1 position.
An additional morphological point to highlight is the contrast between a relatively smooth and unstructured IF (see the Pa α line map in Fig. 9) and a complex, structured, folded DF surface as traced by the H 2 0-0 S(9) line in Fig. 9). Moreover, the southwest part of the Bar which corresponds to the upper part of the map displayed in Fig. 9 is much more structured than the other regions of the Bar. In the northeast, a single main DF is observed. This could be related to previous ground-based observations of the molecular condensations deeper inside the PDR which showed that the northeast part of the Bar has a main condensation while the southwest part is fragmented into several components (e.g., Lis & Schilke 2003;Lee et al. 2013).
Physical origin of the terraced-field-like structure. These structures may result from pre-existing high density structures shaped by the high FUV field inducing a compression. The density contrasts increase due to compression. Another potential explanation is that the region exposed to stellar winds and protostellar outflows make it especially turbulent. On larger spatial scales, SOFIA observations of C + reveal that stellar winds and protostellar outflows shape the molecular cloud and also inject mechanical energy . Regularly spaced ridges that run parallel to the photo-dissociation front could also sug-gest that large-scale magnetic fields are dynamically-important (Mackey & Lim 2011) and raise the question if they could be associated with magnetic-driven density peaks. If the gas thermal pressure is very large then one needs strong magnetic fields for such an hypothesis to be dynamically relevant (e.g., for P th ∼2 10 8 K cm −3 , one needs 800 µG; Pellegrini et al. 2009;Goicoechea et al. 2016). SOFIA HAWC+ observations of the dust polarization toward the Bar reveal a magnetic field strength of ∼ 300 µG (Chuss et al. 2019;Guerra et al. 2021). However, high angular observations of the dust polarization are needed to confirm its relevance. Berné et al. (2010) observed similar ridge-like structures, coined the "Ripples," in the western part of the Orion Nebula, also at the interface between the molecular cloud and the HII region. They interpreted the formation of these structures as the result of the Kelvin-Helmholtz (KH) instability occurring due to gas shearing at the interface. Interestingly, the spatial extension of these structures is about 10 times larger (∼ 0.1pc) than that of the ridges observed in the Bar (∼ 0.01 pc, Fig. 6). This can be explained by the larger density of the molecular gas in the Bar (close to 10 5 cm −3 at the DFs, Table 2) as compared to the density in the Ripples (closer to 10 4 cm −3 ). The fact that the ridges in the Bar appear less well-aligned than in the Ripples also suggests that the gas is in a more turbulent phase, which could correspond to the decay of the KH instability (Berne & Matsumoto 2012).

Extinction attenuation of the H 2 NIR lines along the line of sight
The intensity variations in the different H 2 1-0 S(1) at 2.12 µm emission peaks ranging from ∼ 2 to ∼ 10 × 10 −4 erg s −1 cm −2 sr −1 (Fig. 11) may result from a combination of effects due to the local gas densities, geometry (length of the edge-on portion along the line of sight) and dust extinction along the line of sight.
For the H 2 emission in the PDR, extinction along the line of sight due to the dust in the Bar itself between the ionized gas and the region of excited H 2 may significantly attenuate the NIR emission. The extinction might be variable depending on the sightline and the density of the region crossed. From the radio and NIR H 2 line maps, Walmsley et al. (2000) suggests, in fact, that extinction can vary rapidly as a function of position in the Bar. Precise spatial estimates of the internal PDR extinction are thus required. This is possible with the high angular resolution near-IR line maps we obtained with JWST that constraint in detail how dust extinction affect the apparent morphology of the NIR H 2 line emission and how the matter is distributed along the line of sight.
Extinctions towards the edge-on DFs. An effective way to measure extinction is comparing the observed-to-theoretical H 2 ro-vibrational line flux ratios from pairs of lines arising from the same upper level that are separated in wavelength. The dust absorption cross-section rapidly drops with increasing wavelength. Although NIRCam maps can give an overall view of the extinction across the entire front, we cannot use the maps in the F212N filter (centered on H 2 1-0 S(1) line at 2.12 µm) and the F323N filter (centered on 1-0 O(5) line at 3.23 µm with same upper level) since this last filter is dominated by aromatic band emission (see Sect. 3.3). NIRSpec observations from Peeters et al. (2023) where line intensity can be measured without being contaminated by bands were used. The extinction map and profile derived along the line of sight, A los VBar , in the NIRSpec field is shown in Peeters et al. (2023). A los VBar is found equal to 10-12 on DF1 and decreases to 5-3 on DF2 and DF3. This shows that DF1 is farther along the line of sight and is in agreement with the stepwise structure (see Fig. 5) with the column density along the line of sight increasing for the first DFs which are more distant from the observer (but closer in projected distance from the IF).
NIR H 2 line intensity variations due to extinction. Due to extinction effects along the line of sight, the H 2 1-0 S(1) line at 2.12 µm is significantly attenuated compared to the 0-0 S(9) line at 4.69 µm at the DF1 (by about 50%, see the line profiles shown in Fig. 11 towards the NIRSpec cut). Comparing the NIR-Cam H 2 0-0 S(9) 4.69 µm and Keck 1-0 S(1) 2.12 µm line maps, one can see that the latter is systematically weaker all along the DF1 on the entire southern part of the Bar (Fig. B.5). However, the extinction alone cannot account for the intensity variations observed between the DFs. The H 2 at 4.69 µm, which is little attenuated by the dust extinction effects (decrease less than 10-20% for A V = 10), is in fact twice as high at DF3 than at DF1. This intensity variation must result from geometrical or density effects.

Highly rotationally and ro-vibrationally excited H 2 lines profiles
Here, we compare the surface brightness profiles of several rotationally and ro-vibrationally excited H 2 lines, [FeII] 1.664 µm, Pa and Br α lines and AIBs measured with NIRCam and NIR-Spec (see Figure 11). A very good agreement between NIRCam and NIRSpec is found in terms of line distribution and intensity.
The spatial emission profiles of H 2 lines agree in remarkable detail. The H 2 1-0 S(1) (v=1, J=3, E u =6951 K), 0-0 S(9) (v=0, J=11, E u =10261 K), 2-1 S(1) (v=2, J=3, E u =12550 K) line emission show the same spatial behavior with a strong increase at the edge-on DF1, DF2 and DF3. The emission peaks of the different H 2 lines at the edge-on DFs spatially coincide at the spatial resolution of our observations (Figure 11). The H 2 line profiles follow each other very well for d PDR >0.01 pc, with small line ratio variations. Along the NIRSpec cut, the most significant line ratio variation with a strong excess of the 1-0 S(1) line is observed on the irradiated disk d203-506. This is due to a density increase ). For dense highly irradiated conditions, collisional population of the v=1 J=3 level becomes competitive, and the 1-0 S(1) / 2-1 S(1) lines ratio is thus expected to increase from a pure radiative cascade value (∼2) to a collisional excitation value (of the order of 10). In the Bar, the 1-0 S(1) / 2-1 S(1) line ratio is on the order of ∼5. This ratio varies little between the different edge-on DFs. Gas density of the H 2 emission zone must remain comparable along the folded DFs surface.
Background H 2 emission toward the ionized and atomic region. Along the profiles, the H 2 emission seen in projection in front of the Bar and in the atomic region mostly comes from the surface of the OMC-1 in the background, not from the Bar itself. This is demonstrated by several points as explained below. First, NIRCam and NIRSpec emission line profiles are flat and at the same level of intensity in the ionized and atomic regions. Second, the NIRSpec H 2 excitation diagrams are very similar in the ionized and atomic region . The emission in the atomic region itself is predicted by PDR models to be very weak. For example, for an isobaric model (with the Meudon PDR code, Le Petit et al. 2006) with P = 5 × 10 7 − 10 8 K cm −3 (corresponding to n ∼5 10 4 cm −3 in the atomic region and n ∼ 10 5 cm −3 in the zone where the H 2 abundance increases sharply and high-J and v H 2 lines emit), the predicted H 2 0-0 S(9) line emissivity is on average 100-50 times lower in the atomic region than at the H 2 peak.
Article number, page 17 of 51 A&A proofs: manuscript no. main Density in the H 2 emission zone compared with that estimated in the atomic region. By fitting the intensities of a hundred H 2 lines measured with NIRSpec to the grid of Meudon PDR models Peeters et al. (2023) found that densities are about n H = (3.5 − 10) × 10 4 cm −3 in the H 2 emission zone in the Bar, and similar towards the face-on background OMC-1 PDR. This density is of similar magnitude with density estimates from AIB emission profiles in Sect. 5.1.2. Density must be roughly constant from the PDR edge (where AIB emission peak) to the beginning of the ro-vibrationally excited H 2 emission layer where the density and H 2 abundance starts to increase sharply. This is consistent with the average atomic PDR density derived from the observed location of the H 0 /H 2 transition. The H 0 /H 2 is predicted by PDR models to be displaced inward from the IF by A V ∼ 1 for the Bar physical conditions (i.e., high G 0 /n H regime). Using A V /N H = 3.5 × 10 −22 mag/cm −2 , this translates into an average atomic PDR density of 1 mag /(3.

Spatial distribution between the H 2 and HCO + J=4-3 emission
Common substructures in the H 2 and HCO + J=4-3 lines. Figs. 10 and B.5 compare the H 2 maps to HCO + J=4-3 line map across the same field of view. Most of the substructures are common to both maps and show a very similar distribution. The overall spatial coincidence between the H 2 and HCO + line emission shows that they both come from the edge of dense structures and that they are chemically linked. Because of its high dipole moment, and thus critical density for collisional excitation (a few 10 6 cm −3 for optically thin emission), the HCO + J=4-3 rotational line is a good indicator of dense molecular gas. Thus, some of the densest portions of the Bar lie very near the DFs. Detection of both bright HCO + and CO emission by ALMA towards the H 2 vibrational emission layers suggests that the C + /CO transition nearly coincides with the H 0 /H 2 transition (in agreement with Goicoechea et al. 2016). In dense PDRs, the reaction between vibrationally excited H 2 molecules and C + ions becomes exothermic and leads to the formation of CH + . Fast exothermic reactions with H 2 subsequently lead to the formation of CH + 3 . This key hydrocarbon ion reacts with abundant oxygen atoms and enhances the HCO + abundance in the H 2 emitting PDR layers (Goicoechea et al. 2016). We estimated that the average offset between H 2 and HCO + is less than 1 ′′ , about ∼ 0.6 ′′ (or 0.0012 pc). This is close to the distance between the H 0 /H 2 and C + /C/CO transition as predicted by high-pressure isobaric stationary PDR models (Joblin et al. 2018).
Bright emission from the surface of the molecular condensations. The small H 2 and HCO + J=4-3 structures localized at the DF are in general shifted by a about ∼10-20 ′′ relative to the center of the bigger (5 ′′ − 10 ′′ ) molecular condensations seen more inside the molecular cloud (Young Owl et al. 2000;Lis & Schilke 2003). However, some bright H 2 fronts detected in the northwest end of the Bar (sixth cut in Fig. 9) and between the center and the southwest end of the Bar (zone C in Figs. 3 and 4) correspond to the irradiated superficial layers of the bright cold cores detected in CS J=2-1 (cores denoted 3 in the north and 1 in the center-south respectively in Lee et al. 2013). These starless cores are fragmented into 3-5 components, and their fragments are embedded in larger filamentary structures. Some of the clumps are likely collapsing to form a low-mass star (Lis & Schilke 2003). JWST observations could therefore provide very strong constraints on the external boundary conditions (n H , T g ) of these molecular condensations.
Substructures located further along the line of sight. The HCO + J=4-3 substructures found at DF1 have emission velocity (v LS R = 8 − 9 km s −1 ) more consistent with emission from the background OMC-1 than from the Bar (v LS R = 10.5 km s −1 ). These structures may thus be located further along the line of Article number, page 19 of 51 A&A proofs: manuscript no. main sight. This is consistent with the stepwise structure (Fig. 5) and estimates of extinction from H 2 lines (see Sect. 5.2.2). Towards the HCO + J=4-3 and H 2 common substructures, those found at DF1 are accordingly faintly visible in the H 2 ro-vibrational emission at 2.12 µm (Figs. 11 and B.5), since they are affected by extinction along the line of sight.

Photoevaporating protoplanetary disks
In this section, we describe the photoevaporating proto-planetary disks observed in the whole NIRCam fields. The NIRCam images of the Orion Nebula show, in several passbands, numerous spatially resolved externally illuminated protoplanetary disks surrounding young stars, also known as proplyds (O'Dell et al. 1993). They are mostly found in the M42 region, clustering around the Trapezium stars and θ 2 Ori A, south of the Bar. A couple of proplyds were also identified in M43, located nearby NU Ori (HD 37061), the B0.5V star creating the almost spherical HII region of M43 (see Appendix C). Proplyds inside the HII region typically show bright heads of ionized gas (ionization front) and tails pointing directly away from the brightest OB stars. The proplyd family (as defined in O'Dell et al. 1993) also include YSO's seen in silhouette against the background HII region. The disks appear as dark ellipses on top of a bright nebula background and are called pure silhouettes.
The Hubble Space Telescope was the first observatory to spatially resolve proplyd anatomy and show detailed structure of each one of its components: ionized cocoon, embedded photoevaporating disk and jets/outflows. Nearly 200 proplyds in the Orion Nebula were discovered in HST images, mostly using narrow-band filters centered on the Hα λ6563 and forbidden lines of [NII] (Prosser et al. 1994;O'Dell & Wen 1994;O'Dell & Wong 1996;McCaughrean & O'Dell 1996;Bally et al. 1998Bally et al. , 2000O'Dell 2001a;Smith et al. 2005). The most complete catalog of proplyds found in M42 and M43 is presented in Ricci et al. (2008). The atlas is based on HST/ACS/WFC observations obtained under the Treasury Program on the Orion Nebula (PI: M. Robberto, GO 10249) using B, V, I, z passbands and the Hα narrow-band filter. They compile proplyds from previous studies and find new ones, including disks identified by their bipolar nebulae and jets, if closer than 1 ′′ to the stellar source. They exclude HH objects, bow-shocks and elongated jets, but include candidate background galaxies or filaments. The catalog lists 178 bright proplyds with tails, 28 silhouettes, 8 bipolar nebulae and 5 jets, on a total of 219 sources.
In order to identify known proplyds, and possibly find new ones, we cross-matched the sources in the NIRCam images with the Ricci et al. (2008) catalog. We used the F187N narrow-band filter, centered on the Paα emission line, because proplyd ionization fronts and tails are better traced in hydrogen recombination lines. In addition, the JWST diffraction limit at this wavelength (1.87 µm) is nearly the same as HST at 0.656 µm or Hα, that is ∼ 0.07 ′′ , allowing for a direct comparison of proplyd morphology in the optical and in the near-IR. Narrow-band filters centered on emission lines are always preferred to observe proplyd structure because they cancel most of the continuum from the central star. Nevertheless, and due to the extreme sensitivity of JWST/NIRCam instrument, most proplyds remain unseen, obscured by the bright "snow-flake" shape PSF of the JWST, even in the narrow-band filters. The best targets are proplyds with nearly edge-on disks that still remain optically thick at near-IR wavelengths and hence are able to cover the young star. A few proplyds show wind-wind arcs and a couple show collimated jets. A detailed analysis of a few of these objects will be the subject of another article.

M42 region
The NIRCam images were divided into module A, covering the North of the Dark Bay and corresponding to the northeast region of the Trapezium stars, and module B, covering the southeast region of the Trapezium stars, the Bar and θ 2 Ori A. The M42 south field of view includes 62 known proplyds but only 34 of these were identified as having extended proplyd structure in the F187N image: 29 bright proplyds and 5 pure silhouettes. They are shown in Fig. 12. The giant proplyd 244-440 is shown in Fig. 13 due to its much larger size. No new proplyds were found.
The field of view north of the Dark Bay includes 16 sources from the Ricci et al. (2008) catalog, of which 12 are bright proplyds, 3 are pure silhouettes, and one is a bipolar reflection nebula. Only 7 of the 12 proplyds show extended structure, and only one silhouette of the 3 is visible in the F187N image because it is an edge-on disk. The other proplyds are simply point-like sources. The proplyd 215-106 shows a bright spot in the PSF that can be associated to a jet or a fainter smaller companion, requiring further analysis. The reflection bipolar nebula 208-122 does not appear in the F187N image but it reveals instead a close binary. The reflection polar nebulae and a dark disk are clearly seen in the F164N image centered at the [FeII] line at 1.64 µm. These objects are shown in Fig. C.12. We also find 3 new proplyd candidates (shown in Fig. C.13), very faint and small, and named them 171-212, 180-218 and 234-104, following O'Dell & Wen (1994) coordinate-based naming convention. 171-212 is a small proplyd in the north with a faint tail. 234-104 shows a faint ionized cusp in Paα but not a tail. It lies in the northeast, at a distance greater than 140 ′′ from the Trapezium. Both objects face the Trapezium stars. 180-218 lies along what seems to be a faint ionized filament or shock front. A round cusp surrounding a star is seen in Paα and Brα hydrogen recombination lines, but not in other line tracers. It also shows no tail. The cusp can be the ionization front or a bow-shock caused by wind-wind interaction, requiring further analysis. The 3 objects are proplyds in nature because they fill in the proplyd criteria which are a disk and/or envelope that is being photoevaporated by external UV radiation. They show an ionization front and sometimes a tail and a bow-shock and have already a YSO forming inside. They differ from Evaporating Gaseous Globules (EGGs) which do not have yet a star forming and are just condensations of dense gas experiencing external photoevaporation. The 3 objects are not part of the 2MASS catalog of point sources and hence they must be very-low mass YSOs.

M43 region
The NIRCam images obtained in the parallel mode were divided into module A, covering M43, and module B, covering a region in the north of M42. The NIRCam images of M43 include 4 proplyds from Ricci et al. (2008) catalog but only one of them shows extended structure. That is proplyd 332-1605 pointing directly to the ionizing star NU Ori, at 27 ′′ to the west, and showing a long tail, with a head-to-tail extension of 11.6 ′′ or 4 600 au at 414 pc, measured in the F187N NIRCam image. That is nearly 10 times larger than the proplyd HST10 which makes it a giant proplyd candidate. This proplyd was first discovered in HST/WFPC2 parallel images (PI: Rubin,GO 6065)   . The tiles are 5 ′′ × 5 ′′ with north up and east to the left. Some images suffer from instrumental effects such as the diffraction pattern of bright stars and uncorrected cosmic ray events, affecting particularly the edge of the images because of lack of redundancy. and [SII] images but not in [OIII], and no tail. This object was also imaged with HST/ACS in Hα by Ricci et al. (2008) who confirmed its tailess structure. The fact that we see a long tail in Paα and not in Hα is consistent with the low ionizing power of the star NU Ori (B0.5V) when compared to the Trapezium stars (O7V for θ 1 Ori C) or θ 2 Ori A (O9.5V). A new proplyd candidate is found in the F187N image with a prominent jet rendered visible by a chain of knots or HH objects. The ionization front and the knots are also visible in the HST/ACS image in Hα, but not the central star. The powerful jet and the fact that it is not visible in the optical means this object must be very young, still embedded in its circumstellar envelope of gas and dust and experiencing high accretion, that is, it is still a protostar. This object is part of the 2MASS catalog of point sources. We named it 269-1713 following O'Dell & Wen (1994) coordinate-based naming convention. The new proplyd is located at 97 ′′ to the southwest of the ionizing star, NU Ori. Fig. C.14 and Fig. C.15 show respectively the giant proplyd 332-1605 and the new proplyd candidate 269-1713 in NIRCam Paα vs. HST/ACS Hα (PI: J.Bally, GO 9825) images. In the NIRCam images of M43 and M42 north we find numerous extended, elliptical and diffuse objects, sometimes with spiral arms. These are background galaxies.

Conclusions
The JWST/NIRCam and MIRI imaging observations of the Orion Nebula allow us to probe the global fundamental structure and small-scale structures of an interstellar cloud strongly illuminated by UV radiation. We have access to the multiple scales of the nebula with resolution of 0.1 to 1 ′′ from 2 to 25 µm, equivalent to ∼ 2 × 10 −4 to 2 × 10 −3 pc or 40 to 400 au at 414 pc, over a field of view of 150 ′′ and 42 ′′ , equivalent to ∼0.3 and 0.08 pc (at the Orion distance of 414 pc) for NIRCAM and MIRI im-Article number, page 21 of 51 A&A proofs: manuscript no. main ages centered on the Bar. Our main results can be summarized as follows.
-One of the most striking features observed in all our NIR-CAM and MIRI images is that the molecular cloud borders appear structured at small scales. Numerous patterns are observed, such as ridges, waves and globules. This highlights a very intricate irradiated cloud surface (most likely turbulent) and sub-dense structures at such small scales that they were inaccessible to previous IR observations. Several bright emission features associated with the highly irradiated surroundings of dense molecular condensations, embedded young star and photoevaporated protoplanetary disks are detected in the extended PDR layers. -The observations spatially resolve the transition from the ionization front, the dissociation front to the molecular cloud of the prototypical highly irradiated extended dense Orion Bar PDR. This allows us to study the PDR along all its fronts and to spatially resolve the FUV radiation penetration scales inside the molecular cloud. A progressive structure is evident in agreement with previous studies. However, instead of a smooth PDR transition, JWST unambiguously reveals a highly sculpted interface with very sharp edges and multiple ridges. -The spatial distribution of the AIB emission reveal a very sharp illuminated edge at the IF (on scales of 1 ′′ or 0.002 pc) with a strong density rise in the neutral zone. This is expected due to the sharp decrease in gas temperature at the ionization front if the thermal pressures in the ionized and neutral region are of similar magnitude. The density we derived in the atomic region (n H ∼ (5 − 10) × 10 4 cm −3 ) is much higher (factor 10-20) than the electron density previously derived at the IF. Behind the sharp PDR edge, an extensive warm layer of neutral material, essentially atomic with strong emission from the AIBs, is observed up to the H 0 /H 2 dissociation front at 10-20 ′′ or 0.02-0.04 pc from the IF.
-In contrast to the IF, a very complex, structured and folded H 0 /H 2 dissociation front surface is traced by the H 2 lines. This is particularly apparent in the southwestern part of the Bar. A terraced-field-like structure with several steps seen from above can explain the succession of H 2 ridges across the Bar. In that geometry, each observed H 2 emission ridge corresponds to a portion of the DF seen edge-on. -The line spatial profiles of the highly rotationally and rovibrationally excited H 2 agree in remarkable detail. Physical conditions must be comparable along the folded DFs surface. Very thin and bright H 2 emission layers (∼ 10 −3 pc) are spatially resolved at the irradiated surface of the dense molecular regions. The highly excited H 2 emission arises from the very thin zone where the gas density and H 2 abundance starts to increase sharply. -A remarkable agreement in the spatial distribution between the rotationaly and rovibrationaly excited H 2 and ALMA HCO + J=4-3 emission maps is observed. This indicates that they both come from the edge of dense structures and that they are chemically linked. Some of the densest portions of the Bar lie very near the DFs. This is in agreement with previous analysis of ALMA and Herschel observations. However, the small structures were unresolved in most tracers until now. JWST observations provide very strong constraints on the external boundary conditions of the dense molecular condensations. -In M42, several outflows interacting with the ambient ionized gas of the nebula or the molecular gas are detected. Crenellated structures and various arches which are most likely bow-shocks are observed. Regions exposed to stellar winds and protostellar outflows might be especially turbulent. -Numerous proplyds are identified in the NIRCam images of M42 and M43. Nevertheless, many remain unseen, obscured by the bright snow-flake PSF of the very sensitive JWST. The best observed targets are proplyds with nearly edge-on disks that remain optically thick at near-IR wavelengths and thus are able to cover the young star. For these proplyds, the NIR-Cam instrument offers a unique opportunity to study proplyd morphology in the near-IR with a spatial resolution comparable to the HST in the optical. We find 4 new proplyds identified in the F187N images, 3 located at the northeast from the Trapezium stars and one in the M43 region. They were named 171-212, 180-218, 234-104 and269-1713, following O'Dell &Wen (1994) coordinate-based naming convention.
The JWST ERS program on the Orion Bar PDR  gives access to IFU spectroscopy with NIRSpec and MIRI which will be published in other articles. IFU spectroscopy provides insight into the local gas physical conditions (temperature, density, and pressure), the dust properties and the chemical composition of the warm, very structured irradiated medium. It will be possible to probe the dust properties and physical conditions in the dense substructures detected with NIRCam and MIRI images described in this article. To determine the pressure and density variations at the PDR edge, future detailed spatial studies of both the H 2 pure rotational and rovibrational lines will be carried out Van De Putte et al. 2023). These constraints on the physical conditions may allow us to better understand the dynamical effects in PDRs, such as compression waves and photo-evaporative flows.