MMNRAS , 1–29 (2018) Preprint 19 April 2018 Compiled using MNRAS L A TEX style file v3.0
Layers in the Central Orion Nebula
C. R. O’Dell (cid:63)
Department of Physics and Astronomy, Vanderbilt University, Box 1807-B, Nashville, TN 37235 USA
Accepted 2018 April 6. Received 2018 March 25; in original form 2018 February 27
ABSTRACT
The existence of multiple layers in the inner Orion Nebula has been revealed usingdata from an Atlas of spectra at 2 (cid:48)(cid:48) and 12 km s − resolution. These data were some-times grouped over Samples of 10 (cid:48)(cid:48) × (cid:48)(cid:48) to produce high Signal to Noise spectra andsometimes grouped into sequences of pseudo-slit Spectra of 12 . (cid:48)(cid:48) (cid:48)(cid:48) width for highspatial resolution studies. Multiple velocity systems were found: V MIF traces the MainIonization Front (MIF), V scat arises from back-scattering of V
MIF emission by parti-cles in the background Photon Dissociation Region (PDR), V low is an ionized layerin front of the MIF and if it is the source of the stellar absorption lines seen in theTrapezium stars, it must lie between the foreground Veil and those stars, V new , [O III] may represent ionized gas evaporating from the Veil away from the observer. Thereare features such as the Bright Bar where variations of velocities are due to changingtilts of the MIF, but velocity changes above about 25 (cid:48)(cid:48) arise from variations in ve-locity of the background PDR. In a region 25 (cid:48)(cid:48) ENE of the Orion-S Cloud one findsdramatic changes in the [O iii ] components, including the signals from the V low , [O III] and V MIF , [O III] becoming equal, indicating shadowing of gas from stellar photons of > Key words: H ii regions – ISM:atoms – ISM:dust,extinction – atomic processes –radiation mechanisms:general – ISM:individual objects:Orion Nebula (NGC 1976) The general nature of the Orion Nebula and its associatedOrion Nebula Cluster (ONC) of young stars is now well es-tablished (Muench et al. 2008; O’Dell et al. 2008). The visualwavelength images are brightest within a region designatedas the Huygens Region of about 5 (cid:48) diameter lying in theNE corner of a 24 (cid:48) × (cid:48) region designated as the ExtendedOrion Nebula (EON), (G¨udel et al. 2008). The goal of thestudy reported on here is to determine the nature of large-scale features that appear only in high velocity resolutionspectroscopy and establish what these features tell us aboutthe true nature of the Orion Nebula. The Huygens Region is dominated by emission from a thinlayer of ionized gas on the facing surface of the host Orion (cid:63)
E-mail: [email protected] Throughout this paper we frequently use capital letters for theabbreviations for common directions, such as northeast. When thefull spelling is used and capitalized, it indicates a region, when itis not, it indicates a direction.
Molecular Cloud (OMC). The dominant emitting volumeis designated as the Main Ionization Front (MIF) and itis stratified in ionization, become more highly ionized andof lower density away from the actual ionization boundary.The boundary between the MIF and the OMC is the MainIonization Boundary (MIB), the region where ionization ofhydrogen stops. On the other side of the MIB lies a denselayer of gas, dust, and molecules known as the Photon Dis-sociation Region (PDR) recently imaged at 11 . (cid:48)(cid:48) ii ] emission (Goicoechea et al. 2015) that arises slightlywithin the PDR. Variations in surface brightness occur be-cause of increasing distance from the dominant ionizing star θ Ori-C, which lies between the MIF and the observer, andlimb-brightening effects in tilted regions of the MIF.In the foreground of the ionized volume there are two ir-regular layers of primarily neutral gas (van der Werf & Goss1989; van der Werf et al. 2013; Abel et al. 2016) known as theVeil. The dust component of the Veil (O’Dell & Yusef-Zadeh2000) (henceforth OY-Z) accounts for most of the opticalextinction. This extinction further modifies the appearanceof the Huygens Region. The region of highest extinction liesto the east of θ Ori-C and is commonly called the DarkBay. c (cid:13) a r X i v : . [ a s t r o - ph . GA ] A p r C. R. O’Dell
About 60 (cid:48)(cid:48) southwest of θ Ori-C there is an imbeddedgroup of stars within a dense region designated as Orion-S(O’Dell et al. 2008) that is the source of multiple collimatedoutflows from young stars (O’Dell et al. 2015) (henceforthO15). The nature of the Orion-S feature is not exactly un-derstood but certainly includes a neutral cloud of gas thatwe refer to in this paper as the Orion-S Cloud. The pres-ence of molecular and neutral hydrogen lines in absorptionmeans that there is an ionized region beyond it. There areno optical features that can be attributed to the Orion-SCloud, but about 25 (cid:48)(cid:48)
ENE from the centre of this cloudis the brightest part of the Huygens Region, having mul-tiple structures and rapid velocity and brightness changes.We designate this as the Orion-S Crossing. This Crossing isnear the centre of imbedded young stars and the numerouscollimated outflows from them.O’Dell et al. (2009b) and O’Dell & Harris (2010) ar-gue that Orion-S is a free-floating cloud within the cavityof the concave MIF. This would be an isolated remnant ofdense gas and dust within the OMC. An isolated dense cloudwould cast a radiation shadow in ionizing Lyman Continuum(LyC) photons and thus could represent the tip of a ionizedpillar such as seen in NGC 6611. Isolated cloud or pillarwill depend upon the strength of the LyC radiation field inthe shadowed region. If the diffuse LyC radiation field thatis formed by recombining hydrogen ions is weak, then theshadowed region will have an ionization boundary and wewould see a pillar. If this diffuse LyC field, supplemented byphotons from θ Ori-A, is strong, then we will see an iso-lated cloud. There are no optical features that indicate theboundaries of a pillar seen edge-on and the isolated cloudmodel is the more likely. More creative but probably lesslikely Troland et al. (2016) argue that it possible that theOrion-S feature actually lies beyond the MIF and its back-ground radio continuum arises in an otherwise unobservedionized region beyond the PDR. In any event, its NE bound-ary forms the brightest feature within the Huygens Regionthrough proximity to θ Ori-C and limb-brightening. Hence-forth we will refer to this bright region as the Orion-S Re-gion.About 110 (cid:48)(cid:48) to the SE of θ Ori-C there is a long linearfeature known as the Bright Bar, again being bright becauseof limb-brightening. This was most recently studied in theoptical by O’Dell, Ferland, & Peimbert (2017a) (henceforthO17a) and in the radio by Goicoechea et al. (2016).
In the current study we investigate the detailed structure ofregions of the Huygens Region, the Veil, and a recently es-tablished (Abel et al. 2016) layer of ionized gas lying between θ Ori-C and the Veil that we call, the V low component. Henceforth in this paper directions such as southwest will beabbreviated to use combinations of upper case letters, e.g. SW. In this paper all velocities are in km s − unless otherwise statedand all radial velocities are in the Heliocentric velocity system.To convert Heliocentric radial velocities to the Local Standardof Rest (LSR) system, subtract 18.1 km s − . When angular dis-tances are converted to linear distances, we have adopted a dis-tance of 388 ± Our approach is one of trying to explain variations invelocity according to their individual features and regions.This is in contrast with the earliest studies that sought tocharacterize variations in velocities as being due to featuresof turbulence. The most recent and arguably best study us-ing the statistical approach is that of Arthur, Medina, andHenney (2016), where they conclude that turbulence domi-nates the velocities between scales of 8 (cid:48)(cid:48) and 22 (cid:48)(cid:48) and thatthe emitting gas is confined to a thick shell. The idea of thegas being primarily in a layer is confirmed in the study re-ported here, but we establish that important variations invelocity occur locally within 25 (cid:48)(cid:48) due to variations in tilt ofthe emitting layer.The region closest to the MIB is a thin layer of H + +He o that gives rise to the [N ii ] emission and has a characteristice − thickness of 0.0012 pc, corresponding to 1.0 (cid:48)(cid:48) , from theCloudy models for the central region of the Huygens Re-gion (O17a). The same set of calculations give an [O iii ] e − thickness of 0.026 pc, corresponding to 14 (cid:48)(cid:48) . In the seminalOrion study of Baldwin et al. (1991), they derived the thick-ness of the ionized hydrogen zone as 0.08 pc, correspondingto 43 (cid:48)(cid:48) , from their extinction corrected Pa11 surface bright-ness. Given that the hydrogen emission arises from both theHe o +H + and He + +H + zones and the method of calculationdoes not account for the increase in density within bothzones, the models and the derived values are compatible.The surface brightness of the MIF will decrease with in-creasing spatial separation from θ Ori-C, but limb bright-ening will enhance the local brightness of a tilted region.For a fixed size of the tilted region, more of the [N ii ] emit-ting region will be seen edge-on and the [N ii ]/[O iii ] surfacebrightness ratio will be increased, being a maximum near the[N ii ] boundary. The thinner nature of the [N ii ] layer makesit the more useful measure of what is happening in a tiltedregion and [N ii ] is usually preferred in tracing conditionswithin the MIF. Outside of a tilted region the [N ii ]/[O iii ]ratio depends on many local factors, in particular, the illu-mination by θ Ori-C.In addition to high velocity features arising from out-flows from young stars within the ONC, variations of theobserved radial velocity (V obs ) across the face of the Huy-gens Region are well known. In the case of material flowingaway from an ionization front (Henney, Arthur, and Garc´ıa-D´ıaz 2005) the gas will have a characteristic flow velocityaway from the underlying PDR; where, in the case of the[N ii ] emitting layer, the material receded about 300 yearsbefore. If the underlying OMC velocity was constant acrossthe nebula and the MIF lay in the plane of the sky (hence-forth simply the sky), then the observed radial velocitieswould be the velocity of the OMC blue shifted by the flowvelocity for each ion (the material is accelerated during theflow, so that the evaporative flow velocity for the [O iii ] emis-sion is greater than that for [N ii ] emission).If there were no significant differences of velocities of thePDR, the differences in the observed optical line velocitiesat different lines-of-sight will reflect the tilt of the surfaceof the MIF. This means that a MIF surface tilted perpen-dicular to the sky will have no component of radial velocity used frequently. The Kounkel et al. (2017) distance gives a scaleof 0.01 pc = 5 . (cid:48)(cid:48)
32. MNRAS , 1–29 (2018) ayers in the Central Orion Nebula due to photo-evaporation flow and the observed velocity willbe that of the PDR. However, there may be large variationsin the radial velocity of the OMC gas, which means thatthe PDR velocity is not constant. Evidence for this is givenin the statistical study of the radial velocities by Arthuret al. (2016), who conclude that density variations withinthe OMC lead to much of the velocity and brightness vari-ations seen in ionized gas. One goal of the present study isto determine where the radial velocities change because ofdifferences of tilt and the velocity of the underlying gas.In this analysis we often need to refer to a character-istic value of V PDR . When this is necessary we will use theresults for the entire Huygens Region. The recent study of[C ii ] by Goicoechea et al. (2015) gave an average velocity of27.5 km s − with a characteristic Full Width at Half Maxi-mum (FWHM) line width of about 5 km s − . Examinationof their velocity channel images indicates no radial velocitychanges within the line’s FWHM that cannot be ascribedto tilted or similarly peculiar regions, therefore our assump-tion of a constant V PDR is useful at the level of the [C ii ]study’s angular resolution and the radial velocity is V([C ii ])= 27.5 ± − .The average velocity of molecules more massive thanH O (but excluding CO) as compiled in Table 3.3.VII ofGoudis (1982), is V ave = 25.9 ± − , while his tab-ulation for the bright CO lines (which must arise furtherinto the PDR than the [C ii ] emitting layer) yields V CO = 27.3 ± − . In this study we will adopt V PDR =Vco = 27.3 ± − , which is consistent with the otherPDR values and is similar to the velocities of the ONC starsof 25 ± − (Sicilia-Aguilar et al. 2005) and 26.1 ± − (F´ur´esz et al. 2008). In our modeling we assume forthe reference value of V PDR ± − , which is withinthe probable error of the molecular cloud velocity V OMC of25.9 ± − and the [C ii ] velocity of 27.5 ± − . In Section 2 we describe the observational data used.The visual images were all from the Hubble Space Tele-scope (HST), the [C ii ] images and spectra were from theHerschel/HIFI instrument observations (Goicoechea et al.2015), and the visual range spectra from a Spectral Atlas(Garc´ıa-D´ıaz et al. 2008) (henceforth the Atlas). In Sec-tion 3 we use spectroscopic data of large areas and highsignal to noise ratio (S/N) to demonstrate the nature of redand blue velocity components on the shoulders of the MIFemission and establish their origins. The related Appendix Aestablishes how well one can identify these components. Re-gions where variations in velocity are primarily produced bychanges in the tilt of the MIF are discussed in Section 4. Thesignificant variations in the vicinity of the Orion-S Cloudare discussed in Section 5. Section 6.1 treats a region nearthe outer Bright Bar. Section 6.2 discusses a region of highlocalized extinction. Section 6.3 considers a profile of spec-tra that cross both the Bright Bar and the Dark Bay. Theinterpretation of the velocity systems is then discussed inSection 7. This paper addresses many features in the Huygens Region,some of which have been noted before and named. Some-times the names of individual features have evolved as dif-ferent investigators emphasized different aspects of the ob-jects. Some of the features are newly recognized and we havetried to use simple but descriptive names for them. In orderto help the reader make their way through this paper, wegive below a simple glossary of terms.
Atlas.
The compilation of high resolution spectra pre-pared by Garc´ıa-D´ıaz & Henney (2007).
Boldface numbers.
These are to designate Slit Spec-tra in a profile, e.g. , – . Dark Arc.
An arcuate feature within the Orion-SCrossing that appears as darker than its surroundings. He o +H + zone. The He o +H + layer that is closest tothe Main Ionization Front and produces the [N ii ] emission. He + +H + zone. The He + +H + layer that overlies theHe o +H + zone and produces the [O iii ] emission. Line.
A series of 10 . (cid:48)(cid:48) × . (cid:48)(cid:48) MIF.
The boundary between gas ionized by high energystellar photons and the Photon Dominated Region.
NE-Region.
A grouping of 75 10 . (cid:48)(cid:48) × . (cid:48)(cid:48) Orion-S Cloud.
A molecule rich region seen in absorp-tion lines in the radio continuum of the nebula.
Orion-S Crossing.
A region of rapid changes lyingabout 25 (cid:48)(cid:48)
ENE of the Orion-S Cloud.
Orion-S region.
A broader region including both theOrion-S Cloud and the Orion-S Crossing.
OriS-IF.
The ionized region facing the observer on thesurface of the Orion-S Cloud.
Profile.
A series of Slit Spectra chosen to cross a featureat a diagnostically useful angle.
Sample.
An area over which spectra in the Atlas havebeen averaged.
Slit Spectrum or Spectrum.
A spectrum composedof Atlas spectra selected to lie closest to a line on the sky.This simulates the results from a long-slit spectrum.
Southwest Spoked Feature.
An unusual structurethat appears to be moving past the Bright Bar to the NNW.
SW Cloud.
A discrete cloud of foreground material,associated with the Red Fan feature pointed out by Garc´ıa-D´ıaz & Henney (2007).
Velocity Components.
Discrete features identified bythe deconvolution of a spectrum. V blue . Infrequent velocity components lying to the blueof the V low components V low . A system of velocity components blue shifted withrespect to V
MIF . V MIF . Usually the strongest component of a spectrumand attributed to emission by a layer of material close to theMain Ionization Front. V new , [O III] . A newly recognized but difficult to detect[O iii ] velocity component lying between V
MIF and V scat . V scat . A weak component redshifted with respect toV
MIF and attributed to dust back-scattering.
MNRAS000
MNRAS000 , 1–29 (2018)
C. R. O’Dell
Figure 1.
This 233 (cid:48)(cid:48) × (cid:48)(cid:48) image of the Huygens Region is areduced resolution sample from a larger mosaic (O’Dell & Wong1996). Blue indicates [O iii ], green indicates H α , and red indicates[N ii ] emission. Major features of the nebula plus objects and re-gions discussed in this article are labelled. As in all images in thispaper, north is at the top and west to the right.The irregular formlabelled as the Orion-S Cloud is the outer contour of the 21-cmabsorption line of H i (van der Werf et al. 2013). We have been able to draw upon published images and spec-tra of the Huygens region in this study. The images weremade with the HST and are better than 0 . (cid:48)(cid:48) (cid:48)(cid:48) . Fortu-nately, the small velocity shifts caused by strong tilts in theMIF are well defined in the spectra. We have used a wide variety of Hubble Space Telescopeimages made in narrow-band filters that isolate individualemission lines. We present in Fig. 1 a low resolution replica-tion of an early (O’Dell & Wong 1996) mosaic of WFPC2camera images. This illustrates quite well both the large andsmall scale variations in the ionization in the nebula. Thebrightest part of the Huygens Region lies in the Orion-SCrossing and in the [N ii ] line is rivaled in surface brightnessby the Bright Bar. In this figure we show the position of thetwo stars discussed below ( θ Ori-C and θ Ori-A), the for- mer being the dominant source of ionizing photons withinthe Bright Bar and the latter dominant outside the BrightBar (O’Dell, Kollatschny, and Ferland 2017b).
The spectra we use are from the compilation of Garc´ıa-D´ıazet al. (2008), where a compilation of north-south orientedslits at spacings of 2 (cid:48)(cid:48) in Right Ascension covering muchof the Huygens Region are given. This Spectroscopic At-las (henceforth the Atlas) is made from spectra of about10 km s − spectral resolution and is calibrated to 2 km s − accuracy and presented in steps pf 4 km s − . It is quite com-plete in the H α iii ] 500.7 nm, and [N ii ] 658.4nm lines, but has less complete coverage in [O i ] 630.0 nm,[S ii ] 671.6 nm+673.1 nm, and [S iii ] 631.2 nm. The spectrawere sampled along the original north-south slits in steps of0 . (cid:48)(cid:48)
53. We have used only the [N ii ] and [O iii ] emission linesbecause the large width of the H α line caused by thermalbroadening precluded study of small velocity differences.In order to characterize the spectra, we performed de-convolution of each using the IRAF task ‘splot’. A discussionof the accuracy of the results of using ‘splot’ is discussed inAppendix A.We made several types of samples of the spectra in thisstudy, some of large areas and others that simulate subject-specific slit Spectra. The idea in each case was to increasethe S/N above that in individual Atlas spectra. This en-hanced the visibility of faint features on the shoulder of thestrong emission lines. When we created a pseudo-slit Spec-trum along a non-north-south angle, we call that a Spec-trum. The spectra were calibrated by taking the average signal ofspectra in a 10 (cid:48)(cid:48) × (cid:48)(cid:48) region beginning 5 (cid:48)(cid:48) west of θ Ori-Cand comparing the same region with Hubble Space Telescopefilter images that had been calibrated using the techniqueand reference numbers in ? . For convenience, we work inunits of 100,000 original Atlas units. Conversion of our At-las units to the surface brightness units of ergs cm − sec − steradian − is found by multiplying the 500.7 nm instrumen-tal units by 0.0614 and multiplying the 658.4 nm instrumen-tal units by 0.00782. In order to derive the velocity components in spectra from arelatively simple region of the nebula, which is also a regionused in many other spectroscopic studies of the HuygensRegion, we first derived high S/N spectra. These spectrawere then studied for patterns in their velocity components,with multiple components being identified.
MNRAS , 1–29 (2018) ayers in the Central Orion Nebula Figure 2.
Like Fig. 1 but excluding most of the labels and show-ing the white line orthogonal grid indicating the positions of10 . (cid:48)(cid:48) × . (cid:48)(cid:48) θ Ori-C at its centre. For convenience, thecorresponding box numbers along the north-south axis are givenon the right side. These are called Lines in the text. The irregularblack line indicates the Samples included in the NE-Region that isused to characterize the different velocity components (Section 3)
We averaged the spectra in the Spectroscopic Atlas in boxesof 10 . (cid:48)(cid:48) × . (cid:48)(cid:48)
15 with the reference Sample (X=0, Y=0) cen-tred on θ Ori-C. These are shown in Fig. 2. Within this ar-ray we identify a region designated as the NE-Region. Thisregion represents a less complex area within the HuygensRegion as it does not include features like the Dark Bay, theOrion-S Cloud, and the Bright Bar.
Study of the deconvolution of the spectra of the boxes withinthe NE-Region indicated that there were patterns in the ve-locity components. Analysis of these patterns indicate thatthere are multiple components, each having similar charac-teristics of radial velocity and signal. The criteria we used inthe assignment of velocity components are given in Table 1 and the frequency of their occurrence is shown in Fig. 3. Notall Samples showed every component.Usually the V
MIF component was strongest and is asso-ciated with emission from the evaporating gas near the MIF.The largest velocity V scat component was discovered in ear-lier studies and is discussed in Section 3.4.1. The V new , [O III] component lying between the V scat and V MIF components ispresent in only the [O iii ] spectra and is significantly strongerthan the V scat component, although difficult to deconvo-lute because of the small separation in velocity from theV
MIF component. Although seen in the high resolution [O iii ]study of Casta˜neda (1988), it is elaborated in this studyand is discussed in Section 5.2.4. The V low component hasa lower velocity than the V
MIF component and is also a dis-covery of this study. In earlier studies the term V blue wasoften used, but now we see that there are two low velocitysystems, the frequent V low and the more rare V blue . V low is sometimes stronger that the V
MIF emission and is dis-cussed in Section 3.4.3. The lowest velocity and weak V blue components have probably been discovered before and arediscussed in Sections 5.2.5 and 5.2.6 . All of the componentsare present in both [N ii ] and [O iii ] with the exception ofV new , [O III] , as noted. Each velocity component in the NE-Region has its own setof characteristics and are never the same for both ions.The frequency of occurrence and the average velocities areshown in Fig. 3. In addition we give the average ratio ofthe signal of each as compared with the MIF component(S component /S mif ). This ratio was not calculated where thepresumed MIF component has been assigned to the V wide class. V MIF average velocities are different, with [O iii ] clearlymore blue shifted than [N ii ]. Interpreting the displacementof both from V OMC = 27.3 ± − as due to photo-evaporation from the flat face of the PDR, this meansthat the characteristic photo-evaporation velocity is about 4km s − for [N ii ] and 8 km s − for [O iii ]. The relative magni-tude of these values is what is expected from gas acceleratingaway from the MIF, but greater than predicted in the besttheoretical model (Henney, Arthur, & Garc´ıa-D´ıaz, 2005).The wider distribution of V MIF , [O III] is consistent with theexpectation that it arises from a thicker emitting layer. V wide occurs in a fraction of the Samples. We find thatthe average velocities are slightly bluer than the V MIF com-ponent (-1.4 km s − for [N ii ] and -2.6 km s − for [O iii ]).This is probably due to the line broadening arising from theMIF component being blended with a lower velocity com-ponent. A more quantitative analysis is not possible. Themagnitude of the velocity shift and increase of FWHM ofa composite (treated as a single line) formed from two sep-arate lines is complex (O17a) as they depend on the rela-tive strength of the two components and their velocity dif-ferences. The different break-point for the two ions (18.0km s − for [N ii ] and 16.0 km s − for [O iii ]) reflect that the[N ii ] lines are generally slightly broader. V scat is a common feature in both ions with a widespread of velocities. This is what is expected from the weak-ness of this component, although this is balanced in part bythe large velocity separation from the MIF component. The MNRAS000
MIF emission and is dis-cussed in Section 3.4.3. The lowest velocity and weak V blue components have probably been discovered before and arediscussed in Sections 5.2.5 and 5.2.6 . All of the componentsare present in both [N ii ] and [O iii ] with the exception ofV new , [O III] , as noted. Each velocity component in the NE-Region has its own setof characteristics and are never the same for both ions.The frequency of occurrence and the average velocities areshown in Fig. 3. In addition we give the average ratio ofthe signal of each as compared with the MIF component(S component /S mif ). This ratio was not calculated where thepresumed MIF component has been assigned to the V wide class. V MIF average velocities are different, with [O iii ] clearlymore blue shifted than [N ii ]. Interpreting the displacementof both from V OMC = 27.3 ± − as due to photo-evaporation from the flat face of the PDR, this meansthat the characteristic photo-evaporation velocity is about 4km s − for [N ii ] and 8 km s − for [O iii ]. The relative magni-tude of these values is what is expected from gas acceleratingaway from the MIF, but greater than predicted in the besttheoretical model (Henney, Arthur, & Garc´ıa-D´ıaz, 2005).The wider distribution of V MIF , [O III] is consistent with theexpectation that it arises from a thicker emitting layer. V wide occurs in a fraction of the Samples. We find thatthe average velocities are slightly bluer than the V MIF com-ponent (-1.4 km s − for [N ii ] and -2.6 km s − for [O iii ]).This is probably due to the line broadening arising from theMIF component being blended with a lower velocity com-ponent. A more quantitative analysis is not possible. Themagnitude of the velocity shift and increase of FWHM ofa composite (treated as a single line) formed from two sep-arate lines is complex (O17a) as they depend on the rela-tive strength of the two components and their velocity dif-ferences. The different break-point for the two ions (18.0km s − for [N ii ] and 16.0 km s − for [O iii ]) reflect that the[N ii ] lines are generally slightly broader. V scat is a common feature in both ions with a widespread of velocities. This is what is expected from the weak-ness of this component, although this is balanced in part bythe large velocity separation from the MIF component. The MNRAS000 , 1–29 (2018)
C. R. O’Dell
Table 1.
Criteria for Identifying Velocity ComponentsIon Component Criteria[N ii ] V scat V obs , [N II] ≥ − , the single red component.[N ii ] V MIF , [N II] The strongest component with V obs , [N II] ≥
15 and FWHM ≤ − .[N ii ] V wide , [N II] A V
MIF , [N II] component with FWHM ≥ − .[N ii ] V low Velocity range 0 – 15.00 km s − . The longer of two blue components when there aretwo or S(obs,[N ii ])/S(mif,[N ii ]) ≥ ii ] V blue Velocity range -10 – 1 km s − . The shorter of two blue components when there aretwo and S(obs,[N ii ])/S(mif,[N ii ]) ≤ iii ] V scat Longer of two red components, when there are two or S(obs,oiii)/S(mif,[O iii ]) ≤ iii ] V new , [O III] Shorter of two red components or S(obs,oiii)/S(mif,[O iii ]) ≥ iii ] V MIF , [O III] The strongest component with V obs , [O III] ≥ − and FWHM ≤ − .[O iii ] V wide , [O III] A V
MIF , [O III] component with FWHM ≥ − .[O iii ] V low Velocity range 0 – 15.00 km s − . Longer of two blue components when there are twoor S([O iii ])/S(mif,[O iii ]) ≥ iii ] V blue Velocity range -10 – 6 km s − . Shorter of two blue components when there are twoor S([O iii ])/S(mif,[O iii ]) ≤ probable nature of this component as light back-scatteredby the PDR is discussed in Section 3.4.1. V new , [O III] only occurs in [O iii ]. Its velocity is aboutmid-way between that of the much stronger V MIF , [O III] andthe much weaker V scat , [O III] components. A similar strength[N ii ] component would be more difficult to detect becauseof the smaller separation of V MIF , [N II] and V scat , [N II] , butwe have found no indication of a V new , [O III] component in[N ii ] and it is probably simply absent. The nature of theV new , [O III] component is discussed in Section 7.2.5. V low occurs in both [N ii ] and [O iii ], being much morenumerous in [N ii ]. Average signal ratios are not shown inFig. 3 because of the wide range of values. In the case of [N ii ]0.61 of the ratios occur between the lower limit of 0.05 and0.20, with 0.14 occurring between 0.30 and the maximum of0.44. In the case of [O iii ] 0.52 of the ratios occur betweenthe lower limit of 0.07 and 0.20, with 0.41 occurring between0.30 and the maximum of 1.68. These numbers indicate thatthe V low , [N II] values mostly clump within the low range ofratios with a small fraction near the highest ratios. Thiscontrasts with the V low , [O III] values where the distributionsare more nearly equal. V blue components are rarer than the V low components.We have started their identification at -10 km s − , with theassumption that more negative velocities are the results ofoutflows from discrete objects. In the case of the intrinsicallymost common type of young stellar object (bipolar outflowsfrom sources beyond the PDR), we selectively see the blueshifted components (O’Dell 2001). This component is al-ways weaker than the V low components.The V blue and V low components may be part of a singletype of component lying to the blue of the MIF compo-nents. In Table 1 they have been distinguished by pairs ofcriteria that overlap. The V blue velocity range is -10 km s − –1 km s − for [N ii ] and -10 km s − – 6 km s − for [O iii ] withsignal ratio criteria maxima of 0.05 for [N ii ] and 0.07 for[O iii ]. The V low velocity range is 0 km s − – 15 km s − forboth [N ii ] and [O iii ] with signal ratio minima of 0.05 for[N ii ] and 0.07 for [O iii ]. The break-point between the cri-teria was determined by examining the data and identifying natural divisions. If there is but a single V ≤
15 km s − component for each ion, we used the signal ratio limits.We consider it most likely that the V low and V blue com-ponents are separate systems, with the < V low > for both ionsbeing indistinguishable with a combined average of 6.4 ± − , but that there may be different velocities for [N ii ]and [O iii ]. Further evidence for these being separate systemsare presented in Section 5.2.5 and Section 5.2.6. We have sought to determine if there are relations be-tween the several weaker velocity components with the MIFby comparing their velocities. When there are clear corre-lations between the component and MIF velocities thereis likely to be a physical relationship. We will see thatthere are correlations for V scat but none for V low , andV blue . There may be a correlation between V
MIF , [O III] andV new , [O III] (Section 7.2.5). scat Component
The first study to report the V scat component (O’Dell, Wal-ter, and Dufour 1992) attributed it to back-scattering fromthe high density and dusty background PDR. That interpre-tation is strengthened by the fact that the continua in neb-ular spectra near θ Ori-C are much stronger than expectedfrom atomic processes and this is probably due to back-scattering of stellar continuum from the Trapezium stars(Baldwin et al. 1991).The PDR is red shifted with respect to the emitting lay-ers of the MIF (actually, the emitting layers are physicallyblue shifted with respect to the nearly stationary PDR) andthe scattered light from the PDR appears at about twicethe velocity difference between the emitting layers and thePDR. In a theoretical study Henney (1998) put this in-terpretation onto more solid ground when he modeled thecase where light is scattered from a moving layer, coveringall the possibilities (red and blue shifted scattering layers
MNRAS , 1–29 (2018) ayers in the Central Orion Nebula -0.7+3.3 km s-1Ratio 0.02+0.01 2462462462462462.55.07.5VscatVmifVwide Vlow0 10 30 40V (km s-1)2010.05.010.05.010.05.010.05.010.05.0 39.8+3.5 km s-1Ratio 0.06+0.0323.2+3.6 km s -1 -1 -1 [N II] 0 10 30 40V (km s-1)20 Figure 3.
Histograms of the velocity components with the NE-Region are shown and are discussed in Section 3.3. The red lineindicates the average velocity of the PDR. in the foreground, red and blue shifted scattering layers inthe background). In each case there were upper limits to thedisplacement of the scattered light, but the flux distributionat velocities less that the maximum varied significantly, de-pending on the geometric model adopted. A separate papermodeling the predicted polarization of scattered light Hen-ney (1994) agreed with the limited observations available(Leroy & Le Borgne 1987).We present here the most complete test of the inter-pretation of the red component of the emission line profilesas scattered light. This is done by using the high S/N NE-Region spectra. After establishing its origin, the red compo-
Vmif (km s-1)[N II]Vscat VblueVlow
No slope
High Vscat
Figure 4.
This figure illustrates the relation and lack of relationof V
MIF , [N II] with V scat , [N II] , V low , [N II] , and V blue , [N II] . Theblack symbols are from large Samples and the blue symbols areindividual Spectra from outside this region. The black lines indi-cate the best fitting slope 1.0 for V scat , [N II] and V low , [N II] .Thered sloped line is the least squares fit of slope 0.8 for V low , [N II] .Only < V blue , [N II] > values are given as there is no obvious cor-relation with V MIF , [N II] . The ellipse surrounds the set of datanot included in the fitting of the V scat , [N II] data, as describedin Section 3.4.1. The velocity difference labels refer to the differ-ence in velocity from V MIF , [N II] of the V scat , [N II] and V low , [N II] components slope 1.00 fits. nent can be used to inform the discussion of individual areaswithin the Huygens Region.In the simplest model (the ionized surface lies in thesky) where the back-scattered component arises from onlythe column passing from the observer to the PDR, the V scat component’s displacement will be 2 × V evap , where V evap isthe photoo-evaporation velocity away from the PDR. How-ever, the study of Henney (1998) demonstrated that theback-scattered light will originate in emission outside theobserved column, hence have different relative velocities.The result is that the back-scattered component will havea distorted line profile, whose peak displacement will be nogreater than 2V evap and can be expressed as 2A scat × V evap ,where A scat is a correction factor for the distorted line pro-file. A scat will be one for the simplest geometry and lessthan one for more realistic cases. The expected relation ofthe velocities will beV scat = V MIF + 2A scat × V evap . (1)V evap is different for various ions and is theoreticallyexpected to be lowest for emission occurring near the MIF(e.g. [N ii ]) and greater for emission rising further away (e.g.[O iii ]) (Henney et al. 2005). Indeed, this expectation is whatgave rise to recognition of the blister model for the HuygensRegion (Zuckerman 1973) Ballick, Gammon, and Hjellming(1974). A scat can also be different for the two observed ionsbecause of the [O iii ] emission arising further from the MIF.There may even be sample to sample differences in A scat within a specific ion because the distance (emitting layer toPDR) can vary across the nebula. However, this variation isexpected to be less for [N ii ] because that emission is concen- MNRAS000
MIF , [N II] with V scat , [N II] , V low , [N II] , and V blue , [N II] . Theblack symbols are from large Samples and the blue symbols areindividual Spectra from outside this region. The black lines indi-cate the best fitting slope 1.0 for V scat , [N II] and V low , [N II] .Thered sloped line is the least squares fit of slope 0.8 for V low , [N II] .Only < V blue , [N II] > values are given as there is no obvious cor-relation with V MIF , [N II] . The ellipse surrounds the set of datanot included in the fitting of the V scat , [N II] data, as describedin Section 3.4.1. The velocity difference labels refer to the differ-ence in velocity from V MIF , [N II] of the V scat , [N II] and V low , [N II] components slope 1.00 fits. nent can be used to inform the discussion of individual areaswithin the Huygens Region.In the simplest model (the ionized surface lies in thesky) where the back-scattered component arises from onlythe column passing from the observer to the PDR, the V scat component’s displacement will be 2 × V evap , where V evap isthe photoo-evaporation velocity away from the PDR. How-ever, the study of Henney (1998) demonstrated that theback-scattered light will originate in emission outside theobserved column, hence have different relative velocities.The result is that the back-scattered component will havea distorted line profile, whose peak displacement will be nogreater than 2V evap and can be expressed as 2A scat × V evap ,where A scat is a correction factor for the distorted line pro-file. A scat will be one for the simplest geometry and lessthan one for more realistic cases. The expected relation ofthe velocities will beV scat = V MIF + 2A scat × V evap . (1)V evap is different for various ions and is theoreticallyexpected to be lowest for emission occurring near the MIF(e.g. [N ii ]) and greater for emission rising further away (e.g.[O iii ]) (Henney et al. 2005). Indeed, this expectation is whatgave rise to recognition of the blister model for the HuygensRegion (Zuckerman 1973) Ballick, Gammon, and Hjellming(1974). A scat can also be different for the two observed ionsbecause of the [O iii ] emission arising further from the MIF.There may even be sample to sample differences in A scat within a specific ion because the distance (emitting layer toPDR) can vary across the nebula. However, this variation isexpected to be less for [N ii ] because that emission is concen- MNRAS000 , 1–29 (2018)
C. R. O’Dell
Vmif (km s-1) [O III]Vscat VblueVlowVnew
No slope
Figure 5.
Like Fig. 4 except for [O iii ] data. It also shows thedata for the V new , [O III] components. Unlike Fig. 4 all of thedata points are now used since the wider scatter may hide anypeculiar clumping.i The red diamonds with central black mark-ings are large Samples with high S(V low , [O III] )/S(V MIF , [O III] )values (0.20 – 1.68), while the average of the other Samples is0.10 ± scat , [O III] data points is a slope1.00 line fit to all of the data. The red sloped line is the leastsquares fit of slope 0.6 for V low , [O III] while the black sloped lineis a slope 1.00 fit. V new , [O III] , and V blue , [O III] do not show awell defined correlation with V MIF , [O III] and only their veloci-ties are shown. The velocity difference labels refer to the differencein velocity from V MIF , [O III] of the V scat , [O III] and V low , [O III] components slope 1.00 fits. trated to near the MIF, whereas the [O iii ] emission arisesfrom throughout the He + +H + zone.These expectations are realized in the upper portions ofFig. 4 and Fig. 5. There we see that < V scat - V MIF > is 17 ± − for [N ii ] and 19 ± − for [O iii ]. As expected,there is greater spread and uncertainty for [O iii ]. In thecalculation of the [N ii ] difference we have not included thepoints within the ‘High V scat ’ region because they are ob-viously anomalous, reflecting basically different conditionsthan in the other Samples. These differences would corre-spond to 2A scat × V evap .If V MIF varies because of the local tilt, one would ex-pect that the lowest values of V
MIF to correspond to wherethe MIF is nearly in the sky and would have the value V
MIF = V
PDR - V evap and the back-scattered component would beV scat = V
PDR + 2A scat × V evap . The largest value of V MIF would be when one views the MIF edge-on, thus remov-ing the line-of-sight (LOS) component of V evap and V
MIF = V
PDR . At that point the LOS component of the back-scattered light will also have been removed and V scat willhave dropped to V
PDR . These considerations mean that thepoints for the MIF component in Fig. 4 and Fig. 5 shouldhave a negative slope, whereas the actual slopes are positive.There are several other features to be noted in Fig. 4 andFig. 5. In the [N ii ] figure we see that the highest V MIF valuesnear 27.5 km s − equal that of the PDR, but these occurat Samples -90 (cid:48)(cid:48) ,Line15 and -80 (cid:48)(cid:48) ,Line15 where there is nochange of ionization, as expected when viewing an ionizationfront edge-on. The strength of [O iii ] in Fig. 2 is due to the high ionization Big Arc East (Garc´ıa-D´ıaz & Henney 2007)which is not a product of tilt of the MIF.The positive tilts in the figures and the occurrence of themaximum value of V MIF , [N II] at a region that does not showthe ionization changes of a tilted MIF argue that at largescales the variations in V MIF are primarily due to variationsin velocity of the underlying MIF.The good agreement of V scat - V
MIF with the expecta-tions for back-scattering from the PDR confirms anew theinterpretation commonly applied to the V scat velocity com-ponent. It also explains the variation in V
MIF , [N II] valuesseen in Fig. 3 as being due to variations in V PDR , since itis established (Goicoechea et al. 2015) that the [C ii ] emis-sion that traces the PDR has a FWHM of about 5 km s − .This agreement means that we can use similar observationsof smaller areas of the Huygens Region to evaluate the lo-cal conditions. There is one region where the V scat - V MIF values for [O iii ] are quite different from other regions and isdiscussed in Section 7.2.2. low components real?
Are there deconvolution problems?
The [O iii ] obser-vations are more likely to be affected by deconvolution prob-lems because of the smaller separation from the MIF com-ponents. However, the model calculations discussed in Ap-pendix A indicate that the derived values should be real forboth [N ii ] and [O iii ]. Are the V low components due to blueshiftedshocks?
The NE-Region contains 75 Samples, 60 of whichshow V low , [N II] components and 27 show V low , [O III] com-ponents. The V low , [N II] components are evenly distributed,except that none appear in Line 21. The NE-Region is highlyionized and the broad distribution of the V low , [N II] compo-nents argue that they are not due to low ionization shocks.In contrast, most of the V low , [O III] components fall in theleft Samples in the lower half of the NE-Region, with almostall of those with S(V low , [O III] )/S(V MIF , [O III] ) ≥ low , [O III] )/S(V MIF , [O III] ) occurwhere Doi et al. (2004) detected a blue shifted [O iii ] featurenear our V low , [O III] values that they call the ‘Big Arc East’.We conclude that the strongest V low , [O III] components aredue to blueshifted shocks, but that the weaker V low , [O III] and the V low , [N II] components have another origin. low , [N II] and V low , [O III] components asscattered light The V low , [N II] velocity component is probably correlatedwith the MIF emission, as we see in the lower portion ofFig. 4. V low , [N II] agrees well with a slope 1.00 line. Themagnitude of the velocity differences is similar to that ofthe V scat components, with < V MIF , [N II] - V low , [N II] > = 18km s − .In Fig. 5 we see that there is only a hint of a correla-tion of V MIF , [O III] and V low , [O III] . This has greater scatteraround a slope of one and the best fitting slope is 0.64. Thislack of a good correlation for the [O iii ] velocities may nottell us much since the emission producing the V MIF , [O III] layer is more diffuse. MNRAS , 1–29 (2018) ayers in the Central Orion Nebula A blue shifted component can arise when there is a fore-ground layer of dust sufficiently optically thick in visiblelight and is blue shifted with respect to the MIF emission.We designate the velocity of the scattering layer as V front and the velocity of its scattered light as V low . The definingequation will beV low = V
MIF − A low × (V MIF − V front ) , (2)where A low is again a scaling factor related to the maxi-mum velocity shift expected by the simplest version of thescattering model.If one assumes that A low is near the expected valueof one, then equation 2 says that the scattered componentwould appear at the velocity of V front along each line ofsight. The observed linear relation would argue that theV MIF , [N II] and V front are directly related, with the averagedV front varying from about 0 – 10 km s − . The V low com-ponent would then be expected to show the same spectrumas the MIF, except without the velocity variation caused bydifferent velocities of photo-evaporation. In the case of theNE-Region Samples, Figure 3 shows the average V low , [N II] (5.7 ± − ) and V low , [O III] (8.5 ± − ) valuesare indistinguishably the same.In summary, we can say that the V low velocities areconsistent with a foreground scattering layer whose velocityis closely linked to that of the underlying MIF. However, inthe next section we demonstrate that another interpretationis more likely. low layer as ionized gas One must consider the possibility that the V low layer invokedas a simple scattering layer could also be ionized. If that isthe case, then the V low component could be a combinationof forward scattered light from the MIF and locally emittedradiation.We can determine if a blue shifted scattering layer isconsistent with the V low layer being ionized by calculatingif such a scattering layer is optically thick to the LyC. Toproduce scattered light at about 10% the level of the MIFemission requires an optical depth of about 0.1. The productof the ratio of the number of dust particles to hydrogenatoms times the ratio of the scattering cross section of agrain to the absorption cross of hydrogen at the start of theLyC is small (Osterbrock & Ferland 2006). This means thata scattering layer sufficiently optically thick to produce thescattered light would be very optically thick to LyC ionizingradiation unless the dust/gas ratio there is much lower thanfor the general interstellar medium. Under the assumptionof a normal gas/dust ratio, the gas in the scattering layerwould be ionization bounded and a bright ionization frontfacing θ Ori-C would be formed. Another way of saying thisconclusion is that a layer of gas and dust in the foregroundof the MIF is much more likely to produce locally producedemission lines rather than scattering emission from the MIF.If there is a low column density of gas and there is a highLyC radiation field, the foreground ionized layer would becompletely ionized (the gas bounded case) and some LyCradiation would penetrate and ionize the Veil. This is themodel assumed by Abel et al. (2016). However, if the col-umn density of gas is high and there is a low LyC radiation field, the foreground material would be only partially ionized(the ionization bounded case) and LyC radiation would notpenetrate into the Veil. The presence of [O i ] emission in theV low layer (as discovered in this study, Section 7.2.3) arguesfor the ionization bounded case and would necessitate newmodels to explain the ions and molecules observed in theVeil.There is other evidence for a blue shifted ionized layer.In Section 7.1.6 we discuss the results of optical wavelengthabsorption lines formed in the Trapezium stars at veloci-ties similar to the V low components. In addition, in Sec-tion 7.1.10 we summarize the results from UV absorptionlines found in θ Ori-B at similar velocities. If truly linkedto the layer producing the blue shifted V low emission, thislayer must be between the observer and the Trapezium stars.In subsequent sections of this paper we try in part toclarify this issue by the study of detailed regions. In anyevent, as pointed out in Abel et al. (2016), the V low materialis going to be colliding with Veil Component B in 30,000 to60,000 years.
In Section 3 we characterized the properties of a section ofthe Huygens Region that is arguably the least complex andadded to the diagnostic figures the results from individualSamples, all of the data supporting the model that varia-tions in V
MIF are primarily due to velocity changes in thePDR. However, there are regions of known or suspected hightilts of the MIF and these are studied in this section. In thefollowing section we describe the expected variations in theradial velocity across a tilted MIF. This is followed by appli-cation of this model to two regions through the use of highspatial resolution composite Spectra.The first highly tilted region in the MIF to be recognizedwas the Bright Bar (sometimes called the Orion-Bar). Morerecently the NE Orion-S Region was recognized to be simi-lar by Mesa-Delgado et al. (2011) (henceforth M-D). OY-Zpointed out that there are similar, but less pronounced fea-tures called the East Bright Bar and the E-W Bright Bar.These objects were confirmed in the low ionization radial ve-locity study of Garc´ıa-D´ıaz & Henney (2007), who also dis-covered another similar object designated as the Near EastBright Bar.
In order to test our basic assumption that variation inV
MIF , [N II] values are primarily due to variations in the tiltof the MIF, we have conducted a thorough examination oftwo regions having strong published evidence for their beinglocally highly tilted. In order to confirm our interpretation ofvariations in V MIF , [N II] as due to tilt, one needs supportingdata that can be supplied by variations of surface brightnessand ionization.In Fig. 6 we illustrate the expected variations in theobserved radial velocity as the observer’s LOS traces acrossa raised feature in the MIF. Proceeding from the left to rightin this figure the sequence is as follows. At position A V obs will be V OMC - V evap . Between points A and B as the MIF
MNRAS000
MNRAS000 , 1–29 (2018) C. R. O’Dell tilts toward the observer the V evap component decreases asthe cosine of the tilt, thus raising V obs . V obs is a maximumat point B, where the MIF is perpendicular to the observer’sLOS. V obs then decreases to V
OMC - V evap as one reachesthe crest of the raised feature at point D. The pattern inV obs after point D simply mirrors that between A and D.If one is observing an escarpment, i.e. a feature that risesfrom points A through D and then remains flat at the highestlevel, then one will see only a single velocity peak followedby a drop back to V
OMC - V evap . If the MIF continues toslope towards the observer beyond the point of maximumtilt, then V obs will slowly decrease from the maximum value.If V obs continues near the peak velocity, the MIF surfacemust continue to rise, albeit slower.What the observer can expect to see is determined bythe conditions of illumination of the MIF by the LyC pho-tons that create the MIF. If θ Ori-C were located in thelower left in the figure, then the regions to the right of pointD would be shadowed from ionizing photons and there wouldbe no emission. In fact, since θ Ori-C is the dominant ioniz-ing star in the inner Huygens Region, the physical shape ofthe MIF will automatically be determined by the interactionof the radiation from that star and the underlying densityvariations of the OMC. This is what produces the concavityof the Huygens Region. These considerations mean that inthe outer part of the Huygens region we would only expect tosee single-peak velocity features. In contrast, in the regionsof the nebula where the MIF is illuminated from above, thenone can see a double-peak feature.A similar velocity pattern would also be seen if the fea-ture of the MIF is a depression, rather than an elevation. Theimportant difference is in the illumination. If θ Ori-C werelocated in the upper left in the figure, then the region A andC (and possibly on to D) would be in shadow and one wouldonly see the velocity features arising from the illuminatedfurther portions. Again, an illumination from above wouldallow seeing a double-peak feature. The displacement of thepeak of the LOS emission from the direction of the maxi-mum velocity can be very useful in distinguishing whetherthe velocity features are caused by local elevations or localdepressions.
We created Spectra of the Bright Bar along a 53 ◦ positionangle (PA) as shown in Fig. 7. These were made by samplingamong the nearest spectra in the Atlas’s rectilinear grid,forming Spectra along lines with steps of 2 . (cid:48)(cid:48) (cid:48)(cid:48) .Henceforth we will use integer numbers in boldface typeto indicate Spectrum numbers unless it is obvious that thisis not the case. Earlier conclusions that the Bright Bar is a highly tiltedionization front (Balick et al. 1974; Tielens et al. 1993; Pel-ligrini et al. 2009; Shaw, et al. 2009; Mesa-Delgado et al.2011; Ossenkopf et al. 2013; O’Dell et al. 2017a) or an edge-on view of a curved ionization front (Walmsley et al. 2000)is confirmed and refined in this study. Our study is com-
VOBSERVED
VOMCVOMC -VEVAP
RAISED FEATURE
A B C D E FA B C D FE
MIF
Figure 6.
This cartoon describes the relation between the angleof the MIF and the observed radial velocity along lines of sightcrossing a raised feature. It is assumed that the photo-evaporationflow (V
EVAP ) is always perpendicular to the MIF. The radial ve-locity of the Orion Molecular Cloud is V
OMC and the significanceof the positions A through F are discussed in Section 4.1. plementary to that of M-D, who used 2-D spectrophotome-try in the SE thin line white box shown in Fig. 7 to deter-mine the electron temperature (T e ) and density (n e ) and theionization conditions. A multi-slit study complementary tothe present investigation was described by O17a, where theyused shorter slits with spacings of 1 (cid:48)(cid:48) across the SE heavywhite box region shown in Fig. 7. The latter study employedthe same set of spectra that we use, in addition to MUSE(Weilbacher et al. 2015) line ratios to derive T e and n e , find-ing similar results as M-D and confirming the edge-on viewinterpretation of the Bright Bar. Our current study differsfrom O17a in that it uses longer (37 (cid:48)(cid:48) instead of 24 (cid:48)(cid:48) ) andwider (2 . (cid:48)(cid:48) . (cid:48)(cid:48)
0) slits, thus providing a higher S/N.The higher quality of the slit spectra have allowed study ofboth [N ii ] and the lower signal [O iii ], including their redand blue components, at a degraded but acceptable spatialresolution. We can use the signal and velocity information from ourseries of slits shown in Fig. 7. The results for the individualSpectra are shown in Fig. 8 and averages of selected groupsof spectra are shown in Table 2 and Table 3. We discuss theindividual Spectra in the order of passage across the BrightBar.There is a local maximum in V
MIF , [N II] at , whichindicates the direction of a tilted region in the MIF priorto reaching the Bright Bar. In full HST resolution images(O’Dell & Wong 1996) one sees local [N ii ] peaks that alsoindicate tilted structures within the MIF.In Fig. 7 one sees that falls onto the bright regionusually identified as the Bright Bar. In Fig. 8 the peak of thesignal is at while the signal is only slightly less at . This MNRAS , 1–29 (2018) ayers in the Central Orion Nebula Figure 7.
Like Fig. 1 and Fig. 2 but excluding some labels and in-cluding the individual artificial slits (Bright Bar Profile) crossingthe Bright Bar feature and the areas covered in the other profilesthat intersect in the Orion-S Crossing (large white line circle).The slit numbers for the profiles intersecting the Orion-S Cross-ing are given in Fig. 9 and those in the Red Fan Profile are givenin Fig. 13. The black box indicates the most thoroughly studiedregion in ? . The two thin white line boxes show the regions stud-ied by Mesa-Delgado et al. (2011) while the wide white line boxesshow the Bright Bar and Orion-S NE-Regions studied by O’Dell,Ferland, and Peimbert (2017a). Much of the region pictured isshown as a ratio image in Fig. 9, where detailed features of thecentral slit sequences are shown. means that the features near these positions characterize theBright Bar. In Fig. 8, where the expected accuracy of thedetermination is about 1 km s − , one sees a clear V MIF , [N II] peak of 24.0 km s − for . Following the logic explained inSection 4.1, this means that lies along the LOS wherethe MIF is most tilted. Beyond V MIF , [N II] decreases.The average of – for [N ii ] is < V obs > = 18.3 ± . − and the average of Samples that lie beyond (-90 (cid:48)(cid:48) ,Line 7; -80 (cid:48)(cid:48) ,Line7;-90 (cid:48)(cid:48) ,Line6;-80 (cid:48)(cid:48) ,Line6;-90 (cid:48)(cid:48) ,Line 5;-80 (cid:48)(cid:48) ,Line 5) is < V MIF , [N II] > = 17.5 ± − . The de-crease in V MIF , [N II] outside of indicates that the MIFflattens in this region.We have assigned different velocity components of indi-vidual Spectra according to the criteria for the NE-RegionSamples as described in Table 1. The results are presentedin Table 2 and Table 3 (signal ratios). The sampled regions Bright Bar
Black-[NII]Red-[OIII]642 SMIFSlow
Spectrum Number
Bright Bar V new (?) Figure 8.
The upper panel shows the signal for both the MIFcomponents (heavy box symbol) and the S(V low ) components (di-amonds) plotted versus the Spectrum number (increasing fromNW to SE) for the Bright Bar Spectra shown in Fig. 7. TheSpectrum spacing is 2 . (cid:48)(cid:48)
7. The letter W is superimposed on [N ii ]MIF components with FWHM >
18 km s − and [O iii ] MIF com-ponents with FWHM >
16 km s − . The meaning of these sym-bols remains the same throughout the remainder of this paper.As new symbols are employed their meanings are presented uponfirst use, then used without explanation subsequently. The lowerpanel shows the velocities of the MIF, V scat , and V low compo-nents. The upper small boxes correspond to V scat . selected are thought to be representative of the regions im-mediately before and after the Bright Bar.The V scat , [O III] values for and are much lower thanthe other V scat , [O III] values (27.0 km s − and 28.3 km s − re-spectively). These probably belong in the V new , [O III] groupin spite of their low signals compared to the V MIF , [O III] component (0.06 and 0.048 respectively).It is not clear where the V low , [N II] components for and belong within the NE-Region classification. Their ve-locities (-1.7 km s − and -1.6 km s − ) are similiar to thosein and (-1.4 km s − and -2.1 km s − ). However, theirratios (0.07 and 0.11) are notably higher than in and (0.05 for both).The location of the peak signals give us some indicationof the thickness of the emitting layers. The fact that thepeak signal for [N ii ] lies about one step (2 . (cid:48)(cid:48)
7) inside the peakvelocity indicates that this must be about the thickness ofthe projected [N ii ] zone, which would be larger than theemitting layer if the tilt is not exactly 90 ◦ . The [O iii ] signal MNRAS000
7) inside the peakvelocity indicates that this must be about the thickness ofthe projected [N ii ] zone, which would be larger than theemitting layer if the tilt is not exactly 90 ◦ . The [O iii ] signal MNRAS000 , 1–29 (2018) C. R. O’Dell
Table 2. < V(component) > for Bright Bar Spectra*Component Spectrum Range < V(component) > V low , [N II] ± ± low , [O III] ± ± MIF , [N II] ± ± MIF , [O III] ± ± scat , [N II] ± ± scat , [O III] ± ± − . Table 3. < S(component)/S(MIF) > for Bright Bar SpectraComponent Spectrum Range < S(component)/S(MIF) > V low , [N II] ± ± low , [O III] ± ± scat , [N II] ± ± scat , [O III] ± ± behaves very differently, monotonically decreasing across theregion of maximum tilt and shows a weak rise near , whichis just inside of the peak [O iii ] V MIF at . The displacementof the [O iii ] features of about three steps suggest that the[O iii ] emitting layer is about 8 (cid:48)(cid:48) thick.These results can be compared with the nearby sam-ple studied in O17a. The O17a region does not include thestructured features that produce the local V MIF , [N II] featureat . However, one sees a similar displacement of the peaksignals of [N ii ] and [O iii ] and peaks in V MIF . Again, theV
MIF , [N II] and V MIF , [O III] drop outside of the Bright Bar,but the region sampled does not go out far enough to es-tablish that one is seeing the flattening of a previously steepregion of the MIF.Because of the lower S/N, O17a could not measureV low , [O III] , although V low , [N II] was measured. In that studythe lowest velocity components are called V(blue), but withthe recognition in the present study of a difference of theblue and low components, they should be considered to bepart of the V low components. Including all of our new [O iii ]data into the V low , [O III] system requires lowering the lowestvelocity of this system to -1.5 km s − (from 0 km s − ). For < S(V low , [O III] )/S(V MIF , [O III] ) > = 0.06 ±
13 — 17 is 0.10 ± low , [N II] components fall within the lower limit of the NE-Regiondefinition of 0.05, there is a noticeable difference on the twosides of the Bright Bar since < S(V low , [N II] )/S(V MIF , [N II] ) > is 0.09 ± and 0.15 ± .
03 for
14 — 17 . Asnoted previously (Section 2.1), the region outside the Bright
Figure 9.
This 194 (cid:48)(cid:48) × (cid:48)(cid:48) image of the central Huygens Regiondepicts a full resolution HST image in the F658N ([N ii ]) filterdivided by the image in the F502N ([O iii ]) filter. The smallerwhite boxes indicate regions studied in detail in O&Y-Z, M-D,and O17a. The large white box indicates the region to the SW ofthe Orion-S Cross that is discussed in Section 5.2.7. The positionof three sequences of Spectra are shown, together with their slitnumbers. . The Orion-S Cloud is indicated to the SW of θ Ori-C(marked by a white dot) by the outermost 21-cm line absorptionprofile from the study of van der Werf et al. (2013). The approx-imate outline of the High Ionization Arc (O’Dell et al. 2009b) isshown with the dashed white line.
Bar is ionized by θ Ori-A, thus it is not surprising that theV low , [N II] system is different there.As in O17a we again see that V low , [N II] drops abruptlywhen crossing the Bright Bar feature and now we note thatthe same is true for V low , [O III] . The significantly differentlow values of V low , [N II] at and , are probably associatedwith the secondary regions of tilt associated with the dropthere in V MIF , [N II] , that follows the velocity peak at . Thisregion was not included in the O17a study.We can summarize this section by saying that in thissection we see strong evidence for variations in V MIF , es-pecially V
MIF , [N II] , as being due to structure (varying tilt)within the MIF. The scale of these changes are of a few slitspacings (three would the 8 . (cid:48)(cid:48)
1, slightly less than the size ofthe NE-Region Samples)
The NE part of the Orion-S Region is the second regionwith a well established highly tilted structure (M-D). It isthe brightest part of the Huygens Region in H α radiation.This is due to a combination of two facts: the region has a MNRAS , 1–29 (2018) ayers in the Central Orion Nebula W NE-SW
Black-[NII]Red-[OIII]24135 M-D
Spectrum Number
NE-SW u Figure 10.
Like Fig. 8 except now for the NE-SW Spectra shownin Fig. 9. These have an average spacing of 3 . (cid:48)(cid:48) high LyC flux due to the proximity to θ Ori-C and thereis limb-brightening due to viewing an ionized layer nearlyedge-on. Its structure was very accurately evaluated usingmonochromatic images of multiple ions in the M-D study;however, they did not consider variations in V obs and wasover a limited range of positions (16 (cid:48)(cid:48) × (cid:48)(cid:48) Field of View(FOV)). In this section we will expand on the M-D study ofthis region.We made a series of Spectra described in detail in Sec-tion 5.1 passing through the M-D FOV. The features in theirstudy were collectively called ‘NE Orion-S’. The positions ofour Spectra crossing the NE Orion-S feature are shown inFig. 7. Our average Spectrum spacing was 3 . (cid:48)(cid:48)
6, being slightlylarger in the negative value slits and closer where condi-tions were changing rapidly. The resulting spatial resolutionis poorer than the 1 (cid:48)(cid:48) spaxels of M-D, although they over-sampled their poor resolution images. The results from ourSpectra are shown in Fig. 10, with the slits crossing the M-D FOV indicated by heavy dashed lines. These are drawn slightly wider than the M-D FOV because of the effects ofspatial resolution.Within the area of overlap of our Spectra and the M-D FOV, there is excellent agreement of our S([N ii ]) andS([O iii ]) distributions with those of M-D. – cross theM-D FOV. Our velocity information for [N ii ] indicates thatthe MIF tilts up beginning at and reaches a maximumtilt at , then slowly decreases in tilt to the SE, with atemporary increase in tilt at . The change of velocity(7.5 km s − ) is much greater than in the Bright Bar (4.2km s − ). The similarity continues in that the two peaks in[N ii ] Signal ( and ) are displaced towards θ Ori-C fromthe maximum tilt Spectra, although in this case by largeramounts. The greater change in velocities indicates that ei-ther V evap , [N II] is much greater in this part of the nebula(entirely reasonable because of the higher illumination byLyC photons) or the NE Region is tilted almost edge-on.In the case of [O iii ], we see evidence for a maximum tiltat and the corresponding peak in Signal occurs at . Afterthe peak at , the V obs , [O III] slowly decreases, with a slightrise at . The peculiar nature of the S [O III] and V obs , [O III] values in – are discussed in Section 7.3.These results are similar to those concluded in M-D,where they demonstrated that this region has many of theionization features associated with seeing an edge-on ioniza-tion front. These included an elongated peak in n e insidean elongated increase in T e , in addition to a progression ofincreasing ionization towards θ Ori-C (both are featuresseen in the Bright Bar). The difference in our two studies isthat M-D drew only on spectrophotometric data and placedthe region of maximum tilt in the SW corner of their FOV,whereas we use velocity data to present evidence that thepeak occurs about 6 (cid:48)(cid:48) further to the SW.For the purposes of this study, the important result hereis that the NE Orion-S feature velocity changes over scalesof about 25 (cid:48)(cid:48) can be primarily attributed to variations in thetilt of the MIF.
In Section 1.1 we summarized the properties of the Orion-SCloud and the evidence for its being a high molecular densityregion with active star formation. Through the presence ofneutral hydrogen and molecular absorption lines seen in theradio thermal continuum, we know that it has ionized gasboth on the nearer side facing the observer and on the farside.Some comments here on the adopted nomenclature areappropriate. The Orion-S Cloud will often be called the”Cloud” (not to be confused with the background OrionMolecular Cloud-the OMC). The larger region near theCloud and most visible in optical radiation we will call the”Orion-S Region”. The smaller area within the Orion-S Re-gion where there is an intersection of the axis of a series ofSpectra will be called the ”Orion-S Crossing”. The sequencesof Spectra will be called ”Profiles”.It is debatable if the emission arising from the ob-server’s side of the optically thick Cloud should be called theMIF. More properly, the layer that produces the continuumagainst which the radio absorption lines should be called the
MNRAS000
MNRAS000 , 1–29 (2018) C. R. O’Dell
MIF. For this reason we designate the ionized layer on theobserver’s side of the Cloud as the Orion-S Ionization Front(OriS-IF).There are many regions of small-scale structures in theOrion-S Region and this has driven the decision to use fairlynarrow spectra, which in turns means that the S/N ratios ofthe Spectra are less than in the larger Samples used in thestudy of the NE-Region. The exceptions to this are the verybright regions in the NE corner of the Orion-S Cloud. Thisprobably accounts for our not seeing some features presentin the NE-Region Samples.At the Orion-S Cloud we know that the observed radi-ation will be dominated by the OriS-IF and clockwise fromthe SE through the north of the cloud the radiation will befrom the MIF. Beyond the Cloud the observed radiation cancontinue to be from the OriS-IF, can be dominated by theMIF, or can be a combination of both. We will first summa-rize the information from each of the profiles, then combinethis information with that from a large Sample SW of thecrossing of the profiles (Section 5.2.7).The Northeast-Southwest (NE-SW), East-West (E-W),and South-North (S-N) Profiles naturally divide into threesections. In NE-SW and E-W the first section representswhat resembles MIF emission, the second clearly dominatedby OriS-IF emission, and a third within which it is moreambiguous about the type of emission that dominates.Division of the S-N Profile is less clear. In the Northsection both the [N ii ] and [O iii ] lines are a product of theMIF and then transitions to the OriS-IF area. In the Southsection of this profile [N ii ] emission appears to be from theMIF, but V MIF , [O III] drops to velocities associated with theV low system. In order to fully map the Spectra of the Orion-S Region, wecreated three sequences of Spectra, all passing through ornear the Dark Arc feature that lies near the NE boundaryof the Orion-S Cloud. One sequence was perpendicular to themultiple linear features on the NE boundary of the Orion-Scloud (PA = 124 ◦ , O15) discussed in Section 4.4. The othertwo sequences are north-south and east-west in orientation. NE-SW Spectra were created in a fashion similar tothose for the Bright Bar. However, in this case the averagespacing of the slits was close to 3 . (cid:48)(cid:48) (cid:48)(cid:48) .The spacing between slits -3 and -4 was 5 . (cid:48)(cid:48) ◦ , which is perpendicular to theorientation of the local highly tilted ionization front (O15).The width was selected to match the 16 (cid:48)(cid:48) × (cid:48)(cid:48) region studiedwith a multi-aperture spectrograph by Mesa-Delgado et al.(2011).The axis of this series of slits lies in the direction of theopening of the C-shaped high ionization arc that surrounds θ Ori-C (O’Dell et al. 2009b, 2017a). As discussed in O’Dellet al. (2009b), this well defined feature is best seen in imagesthat are the ratio of high and low ionization emission lines.Therefore, we show in Fig. 9 a ratio image in [N ii ] over[O iii ]. It appears that the Orion-S Cloud either lies in frontof the high ionization shell or interrupts it. E-W Spectra were created from a sequence of 25 slitswith a spacing of 4 . (cid:48)(cid:48)
0. The easternmost slit is 9 . (cid:48)(cid:48) θ Ori-C and the centre of the sequence is 25 . (cid:48)(cid:48) θ Ori-C. The slit heights are 12 . (cid:48)(cid:48) S-N Spectra were created with a width of 8 . (cid:48)(cid:48)
0, centred35 . (cid:48)(cid:48) θ Ori-C. The top slit ( ) is 81 . (cid:48)(cid:48) θ Ori-C and the bottom slit ( ) is 75 . (cid:48)(cid:48) θ Ori-C.
Red Fan Spectra were made from the same north-south Atlas spectra as the S-N spectra, except now the slitsrun from -14 at 122 . (cid:48)(cid:48) θ Ori-C to at 74 . (cid:48)(cid:48) θ Ori-C.The results of the NE-SW Spectra are shown in Fig. 10.To these we add the results for the E-W Spectra in Fig. 11and for the S-N Spectra in Fig. 12. In this section we willfirst analyze the results from [N ii ], as it arises from closeto an ionization front, and then [O iii ] which can arise fromfurther from an ionization front or even in fully ionized re-gions not immediately associated with an ionization front.We will see that there are [O iii ] components not seen in[N ii ]. We first discuss the results for sections of the Profileslying outside the Orion-S Crossing and Samples near theProfiles (Section 5.2). We then discuss the Profile sectionswithin the Orion-S Crossing (Section 5.3). ii ] in theprofiles The [N ii ] emission must arise from a thin layer that is closeto either the MIF or the OriS-IF. This makes it useful indeveloping a picture of the 3-D structure of the HuygensRegion.We see in Fig. 10, Fig. 11, and Fig. 12 that near the endof the profiles V MIF , [N II] values are generally below thosefor the NE-Region (23.2 ± − ) with the exceptionof the East end of the E-W Profile. This argues that theextreme regions, which are well beyond the Orion-S Crossingare similar to the NE-Region, but are either flatter or have alower V PDR . The NE-SW Profile shows a S(V
MIF , [N II] ) peakand a V MIF , [N II] peak of 24.1 km s − at , which is locatedjust beyond the M-D FOV and on the northeast edge of theDark Arc. These peaks are almost certainly associated withthe OriS-IF, while the slow decrease in V MIF , [N II] to the SWmay be due to an increase of tilt of the OriS-IF, a transitionto a mix with MIF emission, or a change in the underlyingV PDR (or a mix of these factors).The E-W Profile shows its strongest signal peak at where it is established on the rise of the MIF and possibletransition to the OriS-IF. A second signal peak is at ,where it crosses the series of high velocity features associatedwith HH 269. The smooth decrease of V MIF , [N II] to the Westis subject to the same multiple possibilities as in the SWportion of the NE-SW Profile.The S-N Profile is the most structured of this series. Itsfirst sharp signal and velocity peak at occurs within theOrion-S Crossing at the passage of the HH 269 associatedfeatures. The second signal peak at - within a regionwhere V MIF , [N II] is decreasing but where this decrease isalmost stopped. The signal and velocity peaks at – occur immediately before where the profile crosses the innerboundary of the High Ionization Arc. The signal peak at MNRAS , 1–29 (2018) ayers in the Central Orion Nebula Black--[NII]Red-[OIII]
East-West
Spectrum Number
East-West
Figure 11.
The upper panel gives the Signal results from theseries of Spectra in the E-W sequence shown in Fig. 9. The lowerpanel gives the velocities. The spacing of the Spectra is constantat 4 . (cid:48)(cid:48)
0. Again the Orion-S crossing is shown. is in the low ionization portion of the ionization stratifiedHigh Ionization Arc.Well to the NE of in the NE-SW Profile the ob-served radiation must originate in the MIF. This group isthe closest to the dominant ionizing star θ Ori-C and thusprobably subjected to the greatest radiation field and stellarwind. It provides a useful region of reference to the others.The distance between the MIB and θ Ori-C in the fore-ground has been determined from the extinction correctedH β surface brightness (Baldwin et al. 1991) to be 0.23 ± (cid:48)(cid:48) ± (cid:48)(cid:48) .These profiles all support the idea that the Orion-SCrossing region is in a region closer to the observer than theMIF in the NE-Region Samples but do not indicate quanti-tatively their relation to the MIF. Unfortunately, the datado not indicate if the OriS-IF is dominant to the SW. iii ] in theprofiles The characteristics of the Cloud region are quite different in[O iii ]. This is not surprising because in photo-evaporationflow the [O iii ] emission arises further from the MIB thanthe [N ii ] emission. In addition, gas not directly linked tophoto-evaporation flow is expected to be fully ionized be-cause of proximity to θ Ori-C and there will be no accom-
W W WW WW WW WW W WW W
South-North
WW WW WW WWW WW W
Spectrum Number
Black--[NII]Red-[OIII]
South-North
Vnew u Figure 12.
The upper panel gives the Signal results from theseries of Spectra in the S-N sequence shown in Fig. 9 are shown.The spacing of the Spectra is constant at 3 . (cid:48)(cid:48)
2. Again the Orion-Scrossing is shown. The lower panel gives the Velocities. The hor-izontal bar symbol depicts the data for the V new , [O III] velocitysystem seen in this Profile and the more crowded upper panelshows S(V new , [O III] ) as horizontal bars. panying [N ii ] features. This means that very different radialvelocities can be encountered. In this section we restrict ourdiscussion to the regions outside the Orion-S Crossing areabecause the [O iii ] emission is quite different. A separate sec-tion (5.3) will discuss the Orion-S Crossing in both [N ii ] and[O iii ].Outside of the Orion-S Crossing area all of the strongest[O iii ] components clearly belong to the V MIF , [O III] systemas defined in Table 1. However, there are revealing fea-tures. In Fig. 10 we see that SW of the Orion-S Crossingthe strongest component has V obs , [O III] < − , butin the case of the lowest values the line is unusually broad(FWHM >
18 km s − ) and it is likely to be a blend with aV low , [O III] component. In Fig. 11 we see that the V MIF , [O III] values in the East fit into the pattern established in theNE-Region, but in the West the values are slightly higherand almost the same as V MIF , [N II] . We see in Fig. 11 thatS(V MIF , [O III] ) has dropped to the level of the almost con-stant S(V low , [O III] ) values.Taken together, these results for outside the Orion-SCrossing indicate that the V MIF , [O III] components are verysimilar to those in the NE-Region, with the exception of theS-N South Region, and probably reflect similar conditions. MNRAS000
18 km s − ) and it is likely to be a blend with aV low , [O III] component. In Fig. 11 we see that the V MIF , [O III] values in the East fit into the pattern established in theNE-Region, but in the West the values are slightly higherand almost the same as V MIF , [N II] . We see in Fig. 11 thatS(V MIF , [O III] ) has dropped to the level of the almost con-stant S(V low , [O III] ) values.Taken together, these results for outside the Orion-SCrossing indicate that the V MIF , [O III] components are verysimilar to those in the NE-Region, with the exception of theS-N South Region, and probably reflect similar conditions. MNRAS000 , 1–29 (2018) C. R. O’Dell
In addition, in the west area of the E-W Profile where thestrength of the V low , [O III] and V MIF , [O III] components areabout the same, V MIF , [N II] and V MIF , [O III] are about thesame, indicating that the photo-evaporation model does notapply in this region. The behavior of the V MIF , [O III] widecomponents in the SW portion of the NE-SW Profile alsoindicates that S(V MIF , [O III] ) and S(V low , [O III] ) are aboutthe same there also. A region to the SW of the Orion-SCrossing composed of low resolution Samples is discussed inSection 5.2.7. scat components in the line profiles We see in Fig. 10, Fig. 11, and Fig. 12 that in the pro-files the difference in velocities between the V scat and V
MIF components are very similar in both ions to those in theNE-Region.This is not true in both ions for the S-N Profile SouthRegion ( – ). The separation there in [N ii ] ( < V scat , [N II] -V MIF , [N II] > ) is 17.8 ± − , comparable to 17 ± − in the NE-Region. The [O iii ] velocities are morecomplex. For [O iii ] < V scat , [O III] -V MIF , [O III] > = 31 ± − and < V new , [O III] -V MIF , [O III] > = 17 ± − ,while V scat , [O III] - V MIF , [O III] = 20 ± − for the NE-Region. While the V new , [O III] separation more closely agreesthe NE-Region results, the signal of V new , [O III] is not strongenough to produce the observed S(V scat , [O III] ). In addition,the most likely interpretation of the V new , [O III] componentplaces it far away from the MIF (Section 7.2.5. Because ofthe peculiar shift in V MIF , [O III] , it is probably not wise todraw conclusions based on the V scat , [O III] V MIF , [O III] differ-ences in the S-N Profile South Region. iii ] The new [O iii ] velocity system, designated as V new , [O III] ,seen in the NE-Region is also present in the S-N Profile.The average values for the S-N Profile are 27.3 ± − and the signal ratios are low in both the S-N South andNorth regions, about 0.01. It is seen in ten Spectra withinthe Orion-S Crossing, being unusually strong in . low components in the line profiles Many of the spectral components in the Profiles clearly fallinto the V low and V blue categories as defined in Table 1.Since these criteria depend upon both the velocity rangeand the strength of the signal relative to the MIF com-ponent, the boundaries are sometimes uncertain, especiallywhen the MIF component is wide. However, some clear pat-terns emerge.We find the V low , [N II] component in all of the pro-files and also consider it here only where it appearsoutside the Orion-S Crossing area. In the NE-SW Pro-file it is at -4 – -2 ( < V low , [N II] > = 1.8 ± − )then appears again after the Orion-S Crossing in – ( < V low , [N II] > = 1.6 ± − ). In the E-W Pro-file it is at – ( < V low , [N II] > = 3.1 ± − ),then reappears after the Orion-S Crossing at – ( < V low , [N II] > = 1.0 ± − ). In the S-N Profile it is at – ( < V low , [N II] > = 3.3 ± − ), except for , then reappears north of the Orion-S crossing at – , , , and ( < V low , [N II] > = 3.6 ± − ). All ofthese averages fall within the dispersion in the NE-RegionV low , [N II] values, where the average is 6.2 ± − .We also find the V low , [O III] component in all ofthe profiles and we consider it in this section onlywhere it appears outside the Orion-S Crossing area. Inthe NE-SW profile it is at -3 and – , with < V low , [O III] > = 2.3 ± − . In the E-W profile it isin – , with < V low , [O III] > = 6.3 ± − . In theS-N profile it appears in , – , and – , with < V low , [O III] > = 5.4 ± − . Except from the NE-SWProfile (only four occurrences), all of these averages fallwithin the dispersion in the NE-Region V low , [O III] values,where the average is 6.9 ± − .In summary, we can say that V low , [O III] is clearlypresent to the SW, west, and north of the Orion-S Crossingarea and is present in the NE and north. It is compara-ble in signal to the MIF component to the west. V low , [N II] is clearly present to the NE, SW, west, and north of theOrion-S Crossing. When both V low components fall in thesame region, < V low , [O III] > is always larger, by 0.7 to 5.3km s − . As an overall group the Profile V low values are sim-ilar to V low values NW of the Bright Bar. blue components in the line Profiles In the Profiles, the V blue components are less frequentthan V low components, as was the case for the NE-Region.They are almost never seen in the Orion-S Crossing re-gion and outside of there the V blue , [N II] components (37)are more frequent than the V blue , [O III] components (19).Their average velocities give weak evidence that the [N ii ]component ( < V blue , [N II] > = -5.2 ± − > ) is moreblueshifted than the [O iii ] component ( < V blue , [O III] > = -0.9 ± − ). These properties are discussed in Sec-tion 7.2.4. In order to verify the trends in the Spectra from the Profilesin the region to the SW, South, and West of the Orion-SCrossing we employed a group of our Samples (40 (cid:48)(cid:48) – 80 (cid:48)(cid:48) ,Lines 10 – 13) . The boundary is shown in Fig. 9 and wasselected to avoid contamination by the high velocity flowsassociated with HH 269 and to cover the area of the Orion-SCloud. The results of the deconvolutions are given in Table 4,where the criteria for assignment of velocity components es-tablished in the NE-Region (given in Table 1) were used.A comparison of the results in Table 4 with the Profilesbracketing and crossing this grouping of Samples indicatesthat their properties extend across the Orion-S Cloud. TheV low , [O III] component agrees with the values at – in the NE-SW Profile, indicating that our interpretation ofthe highest slits as blends of V MIF , [O III] and V low , [O III] iscorrect. A notable difference between the results of the Pro-files and these Samples is that V blue , [N II] is much less fre-quent across the Orion-S Cloud and the V blue , [O III] is notdetected. These properties indicate that the Orion-S Cloudeither intercepts the layer producing the blue componentsor is preventing its ionization. MNRAS , 1–29 (2018) ayers in the Central Orion Nebula Table 4.
Properties of Region SW of the Orion-S Crossing*Component < Velocity > S(component)/S(MIF)V scat , [N II] ± ± MIF , [N II] ± low , [N II] ± ± blue , [N II] ** -12.5 ± ± scat , [O III] ± ± MIF , [O III] ± low , [O III] -1.0 ± ± − . Samples: 50 (cid:48)(cid:48) – 80 (cid:48)(cid:48) ,in Lines 11 – 13.** Four Samples only. The FWHM of the MIF components were different, withall of the V
MIF , [N II] components having FWHM ≤ − , while 11 of the 12 V MIF , [O III] components hadFWHM ≥ − . Therefore, the V MIF , [O III] lineswould be considered wide in the Profiles.A previously unknown [N ii ] velocity component wasfound in all four of the Line 11 Samples. It has anaverage velocity of 31.4 ± − and is strong( < S(component)/S(MIF) > =0.16 ± (cid:48)(cid:48) ,Line 13 Sample a strong(S([O iii ])/S(MIF) > =0.43) component was seen at3.7 km s − . The strength and location indicates that thiscomponent is associated with HH 1131 . The Orion-S Crossing also contains the curious feature des-ignated (OY-Z) as the Dark Arc. It is highly visible, evenin the reduced resolution image in Fig. 1 and more clearlyin the full HST resolution F658N over F502N images shownas Fig. 9 and in various color depictions such as Fig. 20 inBally, O’Dell, and McCaughrean (2000). Although in singlefilter images it appears as a dark feature, it has little if anyextinction as determined by OY-Z from the ratio of surfacebrightness in the H α optical line and the 20.46 cm radio con-tinuum. The surface brightness in [O iii ] drops the least ofthe strong optical emission lines, probably due to the [O iii ]emission arising from an ionized layer larger that the physi-cal size of the Dark Arc feature itself. This is what makes itleast visible in this ion.This feature is best evaluated using the NE-SW Profilebecause it crosses almost perpendicular to the east side ofthe feature. It’s NE boundary falls between and , whilethe maximum height along the NE-SW Profile occurs at .This means that it falls on the rapidly descending OriS-IFin the direction of θ Ori-C.The SE portion of includesa portion of the peculiar feature we designate here as theDark Box (5:35:14.0 -5:23:59, 4 . (cid:48)(cid:48) × . (cid:48)(cid:48)
6, orientation PA =107.5 ◦ ), which has many of the characteristics of the DarkArc.OY-Z present profiles of the Dark Arc along the short-axis of the small Box in Fig. 9. The HST images are morethan 20 × better resolution than our Spectra, so they becomethe primary source of information about the nebula near theDark Arc. The NE portion of the HST sample falls near . In the NE of the HST sample there is a peak in [N ii ] adjacentto a narrow peak of [O i ] emission away from θ Ori-C andimmediately on the NW edge of the Dark Arc.Within the dark edge [O i ] and [N ii ] emission dropsharply ([N ii ] about 50%), whereas H α only drops about30% and [O iii ] only about 20%. All have recovered fromthe effects of the Dark Arc by about 3 (cid:48)(cid:48) from the dark edge(about ).O15 point out that the series of shocks designated asHH 1127 arise from either of the high extinction sourcesMAX 46 or COUP 602, which lie north of the Dark Arc.These shocks appear at the NW edge of the Dark Arc. Thisstrongly argues that the space south of the Dark Arc’s edgeis open, rather than being dark because it is beyond an ion-ization front. The MIB continues to the SW, but there isa brief space where it is only illuminated by scattered LyCphotons, rather than directly by θ Ori-C. The region illumi-nated by scattered LyC photons will be only about one-tenthas bright as an adjacent directly illuminated region and thisis what produces the Dark Arc. The feature is a curved ridgeon the descending OriS-IF. The cause of this ridge is uncer-tain. Near its centre of symmetry there is an unidentifiedsource producing a rapidly moving series of [O iii ] shocks di-rected towards the Dark Arc (O15) and the momentum ofthis flow of material may produce the local curved ridge.
The Orion-S Crossing is arguably the most complex areawithin the Huygens Region. It physically lies above the MIFlevel of the sub- θ Ori-C region and contains the high densityOrion-S Cloud. There is a centre of imbedded young starsnear the peak of the rise. In this section we will draw heavilyupon Fig. 10, Fig. 11, and Fig. 12. In this area we see thatboth V
MIF and V low systems are present for both [N ii ] and[O iii ].[N ii ] is the less complex ion. We see an increase in thestrongest (MIF) component by 4 – 5 km s − above the valuesin the adjacent regions outside of the Crossing. S(V MIF , [N II] )has a double peak in each profile. One of the peaks occurswithin the Crossing in the NE-SW Profile and the E-W Pro-file, while the peaks in the S-N Profile straddle the two sidesof the Crossing. The coincidence of velocity and signal max-ima at NE-SW , E-W – , and S-N indicate thatthese are highly tilted regions, as well they should be as onesees in Fig. 9 that these Spectra coincide and lie on the SWside of the M-D region. NE-SW , E-W and the almost co-incident S-N – ) have lower velocities, indicating thatthese are not as highly tilted.The V low , [N II] values appear in all of the Crossing Pro-files. They appear at velocities about 4.4 ± − and4.3 ± − higher than the regions outside the Cross-ing in the NE-SW and E-W Profiles respectively, while theS-N Crossing V low , [N II] values are about the same as theadjacent regions. However, if one only considers the fourV low , [N II] values nearest the S-N signal maximum andthe two regions closest to the Crossing – and – ,there is a small velocity difference (about 1 km s − ) of theCrossing and background values.Better defined than the small V low , [N II] velocity differ-ences (well defined at about 4 km s − for V MIF , [N II] and MNRAS000
MIF and V low systems are present for both [N ii ] and[O iii ].[N ii ] is the less complex ion. We see an increase in thestrongest (MIF) component by 4 – 5 km s − above the valuesin the adjacent regions outside of the Crossing. S(V MIF , [N II] )has a double peak in each profile. One of the peaks occurswithin the Crossing in the NE-SW Profile and the E-W Pro-file, while the peaks in the S-N Profile straddle the two sidesof the Crossing. The coincidence of velocity and signal max-ima at NE-SW , E-W – , and S-N indicate thatthese are highly tilted regions, as well they should be as onesees in Fig. 9 that these Spectra coincide and lie on the SWside of the M-D region. NE-SW , E-W and the almost co-incident S-N – ) have lower velocities, indicating thatthese are not as highly tilted.The V low , [N II] values appear in all of the Crossing Pro-files. They appear at velocities about 4.4 ± − and4.3 ± − higher than the regions outside the Cross-ing in the NE-SW and E-W Profiles respectively, while theS-N Crossing V low , [N II] values are about the same as theadjacent regions. However, if one only considers the fourV low , [N II] values nearest the S-N signal maximum andthe two regions closest to the Crossing – and – ,there is a small velocity difference (about 1 km s − ) of theCrossing and background values.Better defined than the small V low , [N II] velocity differ-ences (well defined at about 4 km s − for V MIF , [N II] and MNRAS000 , 1–29 (2018) C. R. O’Dell less well defined as somewhat less for V low , [N II] ), there is abig difference in the S(V low , [N II] ) as one enters the Orion-SCrossing. In each Profile, the crossing values are about 3 ormore times stronger than the adjacent regions.The [O iii ] properties are more complex. The V MIF , [O III] components within the Crossing are higher for both the NE-SW (4.5 ± − ) and E-W (4.0 ± − ) Profiles.In both of these Profiles, the S(V MIF , [O III] ) values drop wellbelow an interpolation across the nearby regions and theV low , [O III] components have become comparable in signal.In the NE-SW Crossing V low , [O III] is 2.6 ± − greaterthan the adjacent values and in the E-W Profile it is 0.9 ± − greater.The S-N Profile in [O iii ] is more difficult to determine(Fig. 12) because of the ambiguity in assignment to differentvelocity systems. South of the Crossing the strongest signalhas been assigned to V MIF , [O III] , which has only a slightlyhigher velocity than the much weaker V low , [O III] north of theCrossing. The V new , [O III] components south of the Crossinghave similar velocities to the V MIF , [O III] components northof the Crossing, but they are much weaker than the S-NProfile South Region V MIF signals. The velocity differenceabove the local values is impossible to determine becausethis is where the average V
MIF , [O III] drops from 13.5 ± − north of the Crossing to 7.2 ± − south of theCrossing. The complexity of the core of the Crossing – is peculiar in its [O iii ] signal. Average S(V MIF , [O III] )has decreased to 0.81 ± new , [O III] )has increased to 0.36 ± , where the V new , [O III] and V MIF , [O III] ve-locites are 5.1 km s − and 15.2 km s − and the signal ratioS(V new , [O III] )/S(V MIF , [O III] ) = 0.27. Both components areunusually narrow, with FWHM(V MIF , [O III] ) = 8.5 km s − and FWHM(V new , [O III] ) = 9.2 km s − , much narrower thatthe usual value of about 14 km s − .Proceeding from north to south we can say that in thecore of the Crossing that the V low , [O III] component has dis-appeared, not to be seen again to the south. We can alsosay that the S(V MIF , [O III] ) drops markedly in the Core re-gion never to be as strong as it was on the θ Ori-C(north)side of the Cross. The fact that S(V new , [O III] ) temporarilyincreases in the core of the Crossing (it is seen in ten Spec-tra in the Crossing, being unusually strong in . indicatessome interaction with the layer producing V new , [O III] . In order to more completely investigate regions outside theOrion-S Crossing, we also studied a region to the south ofthe S-N Profile, a high extinction feature to the SW, and aprofile along the eastern boundary of the Atlas.
Near the Bright Bar region south of the Orion-S Cloud liesa unique structure. The Southwest Spoked Feature (SWSF)is the leading edge of an incomplete parabolic arc of brightgas pointed to the NNW. Within this arc are multiple radialspokes. The feature has been most extensively described inO15 and is well illustrated in Fig. 24 of that publication andin Fig. 13. A thin section of the Bright Bar crosses the arc.
SWSF
Figure 13.
Like Fig. 9 this is a ratio image except now of a129 (cid:48)(cid:48) × (cid:48)(cid:48) FOV centred 37 . (cid:48)(cid:48) . (cid:48)(cid:48) θ Ori-C.The white sequence shows the slit locations of the Red Fan Profileas shown in Fig. 7 The black box indicates the three low spatialresolution Samples used to create high S/N spectra of the RedFan. The SW Group of shocks are discussed in O15.
To the NW of the SWSF lies a broad low ionization featurefirst identified because of its high velocity by Garc´ıa-D´ıaz &Henney (2007) and was designated there as the Red Fan.We consider in this section a series of slit Spectra (the RedFan Profile) that lie south of the S-N Profile, having thesame widths and spacings. Their location is shown in theSW portion of Fig. 7 and Fig. 13.The greatest variations of velocities in the present studyoccur in the Red Fan Profile. In addition, we see features notfound in other profiles. Fig. 14 gives the results for this pro-file, using the same symbols as before, i.e. an assignment tothe previously recognized velocity systems has been made.This is straightforward for [N ii ], where we see V MIF , [N II] ,V low , [N II] , and V scat , [N II] and all these components are con-tinuations from the southern portions of the S-N Profile.However, no V blue , [N II] components were seen, nor were anyV scat , [N II] components across the Red Fan feature. In addi-tion to the Red Fan Profile, Fig. 14 also shows the resultsfrom a region of grouped low resolution Samples at 40 (cid:48)(cid:48) –60 (cid:48)(cid:48) ,Line 5 and 40 (cid:48)(cid:48) – 50 (cid:48)(cid:48) ,Line 4 (these cover the brightestpart of the Red Fan).With a maximum [N ii ] velocity of 32 km s − , it appearsthat the Red Fan feature is the result of the MIF movingabout 17 km s − greater velocity in [N ii ] than the MIFimmediately to its north and 9 km s − greater than typicalfor the NE-Region and is rapidly moving into the host OMC.The strongest [N ii ] component reaches an unprece-dented high of 35.0 ± − at the centre of the RedFan feature. This indicates a motion of 8 ± − towardsthe OMC at 27.3 ± − . This variation in velocity isalmost certainly due to motion in a discrete cloud, ratherthan a changing tilt because the velocity maximum occurswhere there is a S(V MIF , [N II] ) minimum. MNRAS , 1–29 (2018) ayers in the Central Orion Nebula Red Fan Profile
200 -10 -5 0 5 10
Spectrum Number
Black--[NII]Red-[OIII]
Red Fan Profile
Red Fan
Red Fan
Figure 14.
Like Fig. 12 except showing the results for the RedFan Profile and with the strongest components that are domi-nated by the Red Fan Cloud are shown as circled squares (Sec-tion 6.1). The results for the high S/N sample using the results inthe large box sampling the Red Fan, as shown in Fig. 13, are pre-sented. The large Sample results are depicted with open symbols,with the circle indicating the results for the [N ii ] component thatfalls at the position where V low , [N II] changes abruptly. In both the North and South Regions of the Red FanProfile, the V low , [N II] component has a separation fromV MIF , [N II] similar (about 18 – 19 km s − ) to that in theother profiles , even as V MIF , [N II] changes. The separationnarrows as one moves towards the brightest part of the RedFan. This smooth change breaks down as one crosses thebrightest part of the Red Fan. There one sees that theV low , [N II] component abruptly jumps to a lower velocity,which then continues to mimic V MIF , [N II] with the originalseparation. The high S/N large Sample reveals two V low , [N II] components of very similar strengths, one at the projectionof the north to south pattern and the other at the projectionof the south to north pattern.The broad rise in V MIF , [N II] over -7 – +7 probablydoes not actually arise from the MIF, rather, from a dis-crete cloud moving towards the OMC that we now desig-nate as the Red Fan Cloud (RFC). The large change in theV low , [N II] values that occur at – probably indicates thatthe RFC interferes with the layer that usually produces theV low , [N II] system. The peaks in S(V MIF , [N II] ) at -6 – -5 and – would indicate where we are observing along the edgeof the RFC. The argument for a discrete cloud is strength- Table 5.
Average Characteristics in the Southwest Cloud and anearby comparison region*Component Comparison Southwest CloudV low , [N II] -1.7 ± ± low , [N II] )/S(MIF,[N ii ]) 0.07 ± ± MIF , [N II] ± ± scat , [N II] ± ± scat , [N II] )/S(MIF,[N ii ]) 0.14 ± ± low , [O III] -1.0 ± ± low , [O III] )/S(MIF,[O iii ]) 0.07 ± ± MIF , [O III] ± ± scat , [O III] ± ± scat , [O III] )/S(MIF,[O iii ]) 0.09 ± ± MIF , [N II] )/S(V MIF , [O III] ) 1.07 ± ± low , [N II] )/S(V low , [O III] ) 0.93 ± ± ened by the lack of a V scat , [N II] component across the RFC,indicating that there is not a nearby back-scattering layer.Patterns in the [O iii ] velocities are similar to those in[N ii ]. However, in the middle of the RFC, the [O iii ] compo-nents are about 9 km s − lower velocity. The V low , [O III] com-ponents are about 16 km s − lower than the V MIF , [O III] com-ponents, only slightly smaller than in other regions withinthe Huygens Region.The sense of the RFC velocities (V MIF , [O III] less thanV MIF , [N II] ) indicates that we are seeing flow towards theobserver from a cloud moving even faster than the centralV MIF , [N II] value. The RFC would be ionized by θ Ori-C. As indicated by the low S(V
MIF , [O III] )/S(V MIF , [N II] )values in the centre direction, the RFC is very low ion-ization. As encountered previously, in the RFC centrethe S(V MIF , [O III] ) and S(V low , [O III] ) are about the same,whereas the S(V MIF , [N II] ) component remains much largerthan S(V low , [N II] ). Fig. 27 of Weilbacher et al. (2015) showsthat the [S ii ] derived electron density rises to about 2000cm − along the NW edge of the SWSF and to about 1600cm − at the centre of the RFC. In order to further clarify the effects of obscuring foregroundfeatures, we isolated a Sample on the high extinction South-west Cloud (O-YZ) and compared it with a nearby regionthat is not heavily obscured. The expectation is that the[O iii ] components will be more affected than [N ii ] compo-nents by the selective reddening of extinction. Our South-west Cloud group was made from Samples: 60 (cid:48)(cid:48) ,Line 8;70 (cid:48)(cid:48) ,Line 8; 80 (cid:48)(cid:48) ,Line 8. The comparison group was: 30 (cid:48)(cid:48) ,Line8; 40 (cid:48)(cid:48) ,Line 8; 30 (cid:48)(cid:48) ,Line 9;40 (cid:48)(cid:48) ,Line 9. The results are shownin Table 5, where the classification criteria of Table 1 wereused. No V blue components were seen.We see that the components fall into the usual sys-tems of V scat , V MIF , and V low . The small differencesin velocity don’t reveal anything about the effects ofthe Southwest Cloud. However, the big difference ofS(V
MIF , [N II] )/S(V MIF , [O III] ) (1.07 ± ± ii ] and [O iii ] predominantly arisefrom beyond the obscuring Southwest Cloud. The difference MNRAS000
MIF , [N II] )/S(V MIF , [O III] ) (1.07 ± ± ii ] and [O iii ] predominantly arisefrom beyond the obscuring Southwest Cloud. The difference MNRAS000 , 1–29 (2018) C. R. O’Dell
Dark BayCentralRegionOutside Region
Dark BayCentralRegionOutside Region0.51.51.040 Sample Number
Black--[NII]Red-[OIII] -90 o Column
Black--[NII]Red-[OIII] -90 o Column
Figure 15.
Like Fig. 12 except now for the series of Samplesat -90 (cid:48)(cid:48) as shown in Fig. 2. The details are described in Sec-tion 6.3. Like the earlier figures, W indicates the velocity of aMIF component with FWHM([N ii ]) larger than 18.0 km s − orFWHM([O iii ]) larger than 16.0 km s − , these being Samples thatmay be a blend with another component. For clarity in this fig-ure, these Samples are not denoted with a W in the upper panel.The highest S([N ii ]) values lie off-scale and their values are indi-cated. The locations of the Bright Bar, the Central Region, andthe Dark Bay are indicated. in S(V low , [N II] )/S(V low , [O III] ) indicates that the SouthwestCloud also lies in front of the layer producing both V low components. The fact that the difference in this ratio iseven greater than for the MIF components suggests thatproportionately more of the V low , [O III] component is beingobscured by the Southwest Cloud. (cid:48)(cid:48) east In order to further illuminate the large scale variations inproperties of the velocity systems we created a profile fromthe series of low spatial resolution Samples all with cen-tres 90 (cid:48)(cid:48) east of θ Ori-C (Fig. 2). The results are shown inFig. 15. Lines – include portions of HH 203 and HH 204but the velocities of those shocks are so large that they didnot interfere with identification of the components arisingfrom the nebula.The Samples to the south of the Bright Bar are labelledas the ”Outside Region” in Fig. 15 and represent regionsprimarily ionized by θ Ori-A. There we confirm the re- sults found in Section 4.3.1 that the V
MIF , [N II] , V MIF , [O III] ,V scat , [N II] , V scat , [O III] , V low , [N II] , and V low , [O III] compo-nents are present (although V low , [O III] is seen only in Line5) and no V blue components are found. [N ii ] emission domi-nates, with S(V MIF , [N II] )/S(V MIF , [O III] ) = 3.1 ± ± ± ± − ). These differencesprobably reflect the different radiation field seen in the tworegions.The Central Region (Lines – ) shows thesame velocity systems as in the Outside Region, butnow one commonly sees the V blue , [N II] component andalso the V blue , [O III] component at the Line (16) clos-est to the Big Arc East (Garc´ıa-D´ıaz & Henney2007). The velocities ( < V MIF , [N II] > = 24.2 ± − and < V MIF , [O III] > = 18.6 ± − ) and signals(S(V MIF , [N II] )/S(V MIF , [O III] ) = 1.1 ± iii ] emission (signal ratio changes from2.4 to 4.5). The separation velocity of [N ii ] and [O iii ] ap-pears to be different from that in the Central Region andchanges from 11.9 km s − to 2.9 km s − over Lines – ,but the significance of this is not clear since one or both com-ponents may be the results of blends of the V MIF and V low components. Certainly the V low , [N II] component is presentin the direction of the Dark Bay, at a much reduced signal,which places its source between the distance of the Dark Baymaterial (a part of the Veil) and θ Ori-C.
We see a variety of features in each of the regions inves-tigated. Many of the regions share properties and here wesummarize and discuss the regions separately. Appendix Ahas established the validity of the identified velocity sys-tems and gives the relation of the intrinsic relative signalsto what is derived from using IRAF task splot. Assignmentto a velocity system was based on the velocity relative tothe strongest (outside the Orion-S Crossing) MIF compo-nent and the signal strength relative to the MIF component.There are of course regions where assignments are not clear,near the borders of the velocity and relative strength crite-ria.
This study is only the most recent of high velocity resolutionstudies that cover much or most of the Huygens Region.Their techniques have varied, but we have been able to drawon them in the discussion of our results. Unfortunately themuch more numerous lower velocity resolution and single slitsetting studies are of little value for this investigation.
MNRAS , 1–29 (2018) ayers in the Central Orion Nebula The multi-slit observations of Wilson et al. (1959) was thefirst complete survey of radial velocities, covering the in-ner portions of the Huygens Region in [O ii ], [O iii ], and H β at a velocity resolution of about 10 km s − . Since the de-tector was a photographic emulsion, it was not possible toget precise relative component strengths, but it could givegood values of V MIF and the blue component, where seen,that we now describe as the V low component. These wereseen in both the 372.6 nm [O ii ] and the [O iii ] line. Theaverage velocity of the [O ii ] V MIF component was 23.3 ± − and for the [O ii ] V low component 11.1 ± − ,both calculated for the 202 most certain samples. The [O ii ]emission should arise from the same He o +H + zone as the[N ii ] emission. The average V MIF for [O ii ] agrees well withV MIF , [N II] = 23.2 ± − for the NE-Region, and thegreat uncertainty in the average V low for [O ii ] means that itfalls into the range of the NE-Region (V low , [N II] = 6.2 ± − ). The Fabry-Perot study of Deharveng (1973) covered a largerarea than the Wilson et al. (1959) study. Her area was lo-cated in the northern part of the EON and used a similarvelocity resolution in [N ii ]. An absolute velocity calibrationwas not made and the paper is limited to reporting line split-ting in three areas bordering the Huygens Region. This linesplitting was not seen in the inner, much more densely ex-posed regions, and it is not clear that their absence there isphysical or due to emulsion saturation. The three regions ofdetected line splitting were: A , a trapezoid starting at the south side of the Dark Bayand proceeding south to the FOV limit at about 159 (cid:48)(cid:48) south.The top is 107 (cid:48)(cid:48) wide and lies immediately south of the DarkBay, the bottom is 196 (cid:48)(cid:48) wide. The average split is 14.2 ± − . This area runs along the inside of the Rim Feature(SW of θ Ori-C) designated in O’Dell & Harris (2010) thatextends from the SSW to the NNE of θ Ori-C and boundsthis portion of the EON. B , an elliptical form (200 (cid:48)(cid:48) × (cid:48)(cid:48) ) with the long axis orientedtowards PA = 258 ◦ , centred 45 (cid:48)(cid:48) W,187 (cid:48)(cid:48)
N. This runs alongthe inside of the north portion of the Rim Feature. The av-erage split is 12.1 ± − . C , an L shaped form varying from 73 (cid:48)(cid:48) –104 (cid:48)(cid:48) wide. Itstarts at 160 (cid:48)(cid:48) W,53 (cid:48)(cid:48)
N , goes to 375 (cid:48)(cid:48)
W,0 (cid:48)(cid:48)
N then ends at643 (cid:48)(cid:48)
W,324 (cid:48)(cid:48)
N. The average split is 14.6 ± − . Thisstarts within the central cavity of the Huygens Region westof the Orion-S Crossing and extends in the direction of alarge concave form defining the limit of the EON to theNW.Adopting the NE-Region V MIF , [N II] of 23.2 km s − , theblue components are at A ± − , B ± − , C ± − . These are to be compared with the NE-Region < V low , [N II] > = 6.2 ± − . θ Ori-C and θ Ori-A byCasta˜neda (1988)
Casta˜neda (1988) mapped the regions close to θ Ori-Cand θ Ori-A in [O iii ] in [O iii ] at twice the velocity res-olution of the Atlas. He used a 3 (cid:48) long slit centred on thestars or using combinations of Trapezium stars, producinggood mapping to distances of about 90 (cid:48)(cid:48) from the alignmentstars. One of these slits lies within the Orion-S Crossing Re-gion. The lines were submitted to deconvolution into multi-ple components, much as done in the present study. Multi-ple velocity systems were identified, including what we nowdesignate as V
MIF , [O III] (his system A), V low , [O III] (his sys-tem B), and V scat , [O III] (his system C). There are also com-ponents that are probably attributable to V blue , [O III] andV new , [O III] . Looking through his many tables of data it ap-pears that V low , [O III] is on the average several km s − higherthan for the NE-Region Samples. The advantages of usingCCD detectors with their linearity of response and large dy-namic range were employed in all subsequent optical andUV studies. In their original study van der Werf & Goss (1989) mappedthe entire Huygens Region in the 21-cm line seen in absorp-tion against the radio continuum emission from the back-ground MIF. The amount of LOS H i generally increasedtowards the NE and the line saturated in the vicinity of theDark Bay. Three velocity systems were identified, very clearsystems at 24 km s − (designated as A) and 21 km s − (des-ignated as B) and a less well defined system C at 16 km s − .The values for A and B were later refined to 23.2 km s − and 19.5 km s − by Troland et al. (2016). O’Dell & Wen (1992) mapped the vicinity of θ Ori-C in[O i ] in the 630.0 nm line, much in the manner of Casta˜neda(1988). In studies of [O i ] the location of the foreground nightsky line can be a serious problem. Most of their observationswere made about 1988 November 14, when the heliocentricvelocity correction was +12.8 km s − , which means that thenight sky component was at that velocity, while the observednebular component was around 26.7 ± − . As shownin Fig. 2a and Fig. 2b of their paper, the proximity of thenight sky line (at a separation of about 14 km s − ) onlyallows measurements of the MIF component, although thepresence of a V scat component is obvious. Later observations(1991 January 1) at a heliocentric velocity correction of -8.8km s − clearly shows (their Fig. 2c) V low , V MIF , and V scat components.
In this high resolution (FWHM = 3.3 km s − ) O’Dell et al.(1993) measured the He i line at 388.9 nm and the Ca ii lineat 393.4 nm in the four brightest Trapzezium stars and in MNRAS000
In this high resolution (FWHM = 3.3 km s − ) O’Dell et al.(1993) measured the He i line at 388.9 nm and the Ca ii lineat 393.4 nm in the four brightest Trapzezium stars and in MNRAS000 , 1–29 (2018) C. R. O’Dell θ Ori-A. The He i line arises from a state populated by re-combination of ionized helium that has no permitted tran-sitions to lower energy levels. Thus it is a measure of thecolumn density of He + . In contrast, the ground state Ca ii line arises from a trace state of ionization, most of the cal-cium being Ca iii since ionization of Ca ii requires only 11.9eV and the next stage of ionization requires 47.3 eV photonsand those are infrequent in θ Ori-C radiation. They foundan average He i velocity of 2.1 ± − for the Trapez-ium stars, which is most likely associated with the V low sys-tem. Multiple components were seen in Ca ii . The strongestwere seen in all of the Trapezium stars and were at 7.5 ± − , 22.0 ± − , and 30.9 ± − . A separatestudy by Hobbs (1978) showed strong Na i components at-16.8 km s − , 6.0 km s − , 17.4 km s − , 22.3 km s − , and31.1 km s − in θ Ori-C. Similar to Ca ii , Na i is a minor ion-ization stage since it requires only 5.1 eV photons to ionizeit to Na ii and the next stage of ionization requires 50.9 eVphotons, which are again infrequent in θ Ori-C radiation.
Doi et al. (2004) mapped a 3 (cid:48) × (cid:48) region with north-southslits in H α , [O iii ], and [N ii ]. These data were then abso-lute velocity corrected later and incorporated in the Atlasprepared by Garc´ıa-D´ıaz & Henney (2007). Seeing limitedsamples in declination were submitted to deconvolution us-ing splot and from this a map of velocities was prepared,being presented in their Fig. 2 in velocity bands. It shouldbe noted that their velocities are presented in terms of dis-placement from a systemic velocity of 18 ± − . Thepurpose of the study was to detect new features of high rel-ative velocities and many were found, including the Big ArcEast, the Big Arc South, HH 512, HH 725, and HH 726.With the exception of the Big Arc components, all of thesenew features were attributed to shocks arising from outflows. In their study of the low-ionization lines from the AtlasGarc´ıa-D´ıaz & Henney (2007) worked with the [O i ] 630.0nm line, the red [S ii ] 671.7 nm and 673.1 nm doublet, andthe [S iii ] 631.2 nm line, smoothing the data to give a uni-form FWHM of 12 km s − . Rather than deconvolution ofthe spectra, their analysis was based on identifying velocitybands of about 10 km s − that sampled different portions ofthe typical line profiles. This technique was well suited forthe identification of multiple velocity features throughoutthe Huygens Region. These are summarized nicely in theirFig. 7. Of particular interest to the present study is theregion NW of the Bright Bar near θ Ori-A that they des-ignate as the Southeast Diffuse Blue Layer. There the [S ii ]lines are split, with the longer component (which we asso-ciate with V MIF ) at about 20 – 25 km s − and the weakerand shorter component at about -2 km s − (which we asso-ciate with V low ). They were able to determine the densityof this feature as about 400 cm − .A closer examination of a selected set of [O i ] spectra inthe Atlas indicates that [O i ] is present in nearby regions, incontrast with its being absent in the Southeast Diffuse Blue Layer , as reported in Garc´ıa-D´ıaz & Henney (2007). Theproblem is not one of presence in one area and not the other,rather, some of the Atlas data are more suitable for lookingfor an [O i ] V low component. As presented in Section 7.1.5,a V low nebular component is seen in the O’Dell & Wen(1992) spectrum with a heliocentric velocity correction of -8.8 km s − . The bulk of the Atlas observations in [O i ] weremade on 2003 December 14 when the heliocentric velocitycorrection of was -0.4. In processing the data they identifiedthe strong night sky component and subtracted it beforeinclusion in the Atlas. The separation would have been 25km s − and with a FWHM of 12 km s − they should havebeen able to see the blue component, however the S/N in theregion of the V low component is quite low, therefore V low in[O i ] was reported as not present.In the search for an [O i ] V low component three observa-tions (-28 . (cid:48)(cid:48)
8, -11 . (cid:48)(cid:48)
4, -4 . (cid:48)(cid:48) − . In order to isolate a region free ofthe effects of either the Big Arc or the Bright Bar we choseto group the slit spectra over the range of 45 (cid:48)(cid:48) north andsouth from θ Ori-C. In order to avoid the effects of system-atic correction for the night sky lines, we worked with cali-brated spectra before the night sky line correction was made.These were kindly provided by W. H. Henney, a coauthorof the Atlas. Deconvolution of the averaged spectra revealedV
MIF , V scat , and V low components, in addition to the nightsky component. The V scat component is very weak and thevelocity is quite uncertain. Although the relative signal ofthe V low component is weak, it is still at a believable level.This averaged spectrum is very similar to Fig. 2c in O’Dell& Wen (1992), except that their figure includes the nightsky component. [O i ] is discussed further in Section 7.2.3. In their higher spatial resolution (7 . (cid:48)(cid:48) × . (cid:48)(cid:48)
7) and higher S/Nstudy van der Werf et al. (2013) refined their earlier study(van der Werf & Goss 1990) of high negative velocity 21-cmH i absorption line features. They established that several re-gions are associated with emission line features, in particularthe region NW of θ Ori-A (3.5 – 7.3 km s − ) that overlapswith Southeast Diffuse Blue Layer of Garc´ıa-D´ıaz & Henney(2007) and a region to the NW of θ Ori-C (-3.0 – +7.3km s − ) that they associate with HH 202. Our N-S ProfileSpectra that cross this feature – have an average ofboth [N ii ] and [O iii ] of 3.8 ± − . This most recent (Abel et al. 2016) of a series of four pa-pers is primarily concerned with conditions in the partiallyionized Veil as determined by a host of absorption lines,extending the list to include H . The UV absorption linesinclude P iii (4.8 ± − ) and S iii (4.5 ± − ),both of which are expected to arise in a He o +H + zone. Theirvelocities suggest an association with the V low layer as theyare unlikely to be part of the higher velocity Veil. They alsopresent emission line velocities for the MIF over a 40 square MNRAS , 1–29 (2018) ayers in the Central Orion Nebula arc second area and for V low components of [N ii ] (1.8 ± − ) and [O iii ] (0.9 ± − ). All of the availableinformation was used in creating models of the physical con-ditions and locations of Veil components A and B. Each of the velocity components have their own character-istics in the quiescent regions, which we define here to meanthose outside of the Orion-S Crossing, the Red Fan, theBright Bar and the -90 ◦ Profile. The quiescent regions areexpected to show features related to most of the HuygensRegion.
MIF
Components
The profiles NE-SW (Northeast and Southwest Regions), E-W(East Region), and S-N (North Region) all resemble theNE-Region in that V
MIF , [N II] - V MIF , [O III] is about fourto five km s − . This is what is expected from models of aphoto-evaporation front, although the observed magnitudesof the velocities are more than those expected from the mod-els (Henney et al. 2005). The V MIF , [N II] components in theNE-Region are on the average a few km s − higher than inthe other regions. The most anomalous region is in the Southportion of the S-N Profile. There we see that V MIF , [N II] issimilar to the North portion of that profile and the non-NE-Region Samples. However, V MIF , [O III] drops to veloci-ties characteristic of the V low , [O III] components. It is as iffew photons of greater than 24.6 eV are reaching the layerassociated with the MIF.As established in Section 3.4.1, the comparison of V scat and V MIF components indicate that the usual variationsin V
MIF are caused by changes of velocity of the underly-ing PDR just beyond the MIB. Only small scale variations( < (cid:48)(cid:48) ) can be explained by changes in the tilt of the ionizedlayers near the MIB. scat Components
The V scat components satisfy the predictions of this beingdue to back-scattering from dust in the PDR. Within theNE-Region the V scat components are about the same aver-age velocities, but when one examines Fig. 4 and Fig. 5 onesees the velocity differences 17 km s − and 19 km s − re-spectively support the idea of back-scattering from ionizedlayers with different photo-evaporation velocities.We also see this difference in the non-central portionsof the other Profiles with the exception of the S-N Southregion), where the velocity difference is 17.8 ± − for [N ii ] and 27.8 ± − for [O iii ]. This is the regionwhere V MIF , [O III] is about the same as the usual values forV low , [O III] . The velocity difference between V MIF , [O III] andV scat , [O III] is larger than normal (20 ± − in the NE-Region) This is expected if the PDR region has the usualvelocity (as indicated by the V MIF , [N II] velocities), but theV MIF , [O III] velocity is lower. However if the usual backscat-tering model applies, the velocity difference should be about47 km s − . This may indicate that the backscattering fromthe V MIF , [O III] component arises from emission farther fromthe LOS passing through the observed point. This is what would be expected from the V MIF , [O III] component origi-nated further from the PDR. This is also consistent withthe S(V scat , [O III] )/S(V MIF , [O III] ) ratio being 1.5 – 2 timeslarger than normal for quiescent regions. low Components
Although velocity components have previously been re-ported in the V low range, generally described as V blue , thisvelocity system is best defined in this study. The V low , [N II] average of 3.9 ± − and V low , [O III] average of 5.9 ± − are indistinguishable, although Fig. 4 and Fig. 5 sug-gest that there is a small difference of about the amount inTable 6, [O iii ] being higher by about 2 km s − .Addressing the important question of the presence of[O i ] in the V low system, we have averaged and analyzedthe Atlas spectra for the same three slits as described inSection 7.1.8 over a declination range of 45 (cid:48)(cid:48) north andsouth from θ Ori-C. V low components were seen in all thelines. The V
MIF (km s − ), V low (kms), and relative signalstrengths S(V low )/S(V MIF ) values are: [O iii ], 18.1, 3.2, 0.03;[N ii ], 22.7, 5.7, 0.08; [S ii ], 22.8, 4.1, 0.02; [O i ], 27.2, 8.0,0.04. The velocity determinations are all very close (aver-age 5.3 ± − ) and within the measurement spread ofthese weak components. With cautions about their uncer-tainties, the ratios of the V low components may be useful inmodels that compare the conditions in the V low layer withthat of the MIF.Given the agreement of the [N ii ] and [O iii ] velocitiesand those in [O i ] (8 km s − ) and [S ii ] (-2 km s − ) with thoseof the ionized absorption lines seen in the Trapezium stars(He i ± − , Na i − , Ca ii − ,P iii ± − , S iii ± − ) it is likely that all ofthese components arise from a common layer that is strat-ified in ionization. [O ii ] emission should arise in the sameionization zone (He o +H + ) as [N ii ] and the Wilson et al.(1959) average of the blue components of 11.1 ± − ,strengthens the interpretation of [N ii ] and [O ii ] as part ofthe V low system. As noted in Section 7.1.2, the three out-lying blue [N ii ] areas have velocities of 9 ± − , 11 ± − , and 9 ± − , which argues that the lowest ion-ization portion of the V low layer extends beyond the HuygensRegion.The best fit slope in the [N ii ] data (Fig. 4) is 0.8 whilethat for [O iii ] (Fig. 5) is 0.6. This argues that there is a goodbut not excellent correlation of V low , [N II] and V MIF , [N II] anda poorer correlation of V low , [O III] and V MIF , [O III] . Theseare consistent with the general trends that result from theV MIF , [O III] components arising at a greater distance andthicker layer than [N ii ]. Certainly the V low , [N II] layer knowsabout the velocity of the V MIF in each LOS. This presentsa dynamics problem since one is in the foreground and theother the background of θ Ori-C.Because of the fact that an [O i ] component is now rec-ognized, this layer is probably ionization bounded, unlikethe assumption in Abel et al. (2016) that it is mass-bounded.This presents a problem in explaining the radiation field thatreaches the two primary layers of the Veil. This problem willbe addressed in a future publication. Certainly the layer liesbetween θ Ori-C and the Veil and the high relative velocityof the V low layer towards the Veil indicates that a collision
MNRAS000
MNRAS000 , 1–29 (2018) C. R. O’Dell
Table 6.
Averaged Properties of the V low components outside the Orion-S Crossing Area*Region < V MIF , [N II] > < V low , [N II] > < V MIF , [N II] -V low , [N II] < V MIF , [O III] > < V low , [O III] > V MIF , [O III] -V low , [O III] > NE-Region 23.7 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ◦ (Outside) 19.1 ± ± ± ◦ (Central) 24.2 ± ± ± ± ± ± ± ± − . Parentheses indicate the number of samples. Only samples with both V MIF and V low values were used.** Weighted by number of samples and without Minus90 ◦ (Outside). of the two is imminent, about 30,000 – 60,000 yrs (Abel etal. 2016).It is likely that the highly blue shifted H i absorptionlines seen by van der Werf et al. (2013) at the end of the fea-tures driving HH 202 (their Sample F at -3.0 – +7.3 km s − )are the products of shocks causing high compression densi-ties that have led to de-ionization in those areas of the V low layer.The fact that S(V low , [O III] ) and S(V MIF , [O III] ) becomeabout the same in the E-W Profile West Region whileS(V low , [N II] ) remains weak compared with S(V MIF , [N II] ) ar-gues that some form of selective shadowing of the V MIF , [O III] region is occurring, as discussed in Section 7.3. blue Components
The V blue components are difficult to isolate and are therarest of the components in the quiescent regions. Whenpresent, their inherent weakness makes determination oftheir velocities uncertain. Since the shocks from young staroutflows that we see in the Huygens Region are selectivelyblue shifted, it is possible that the components do not forma system. In this case the V blue components would be at-tributed to low velocity blue shifted shocks. new , [O III] Components
The V new , [O III] components have no counterparts in [N ii ].They are present in the high velocity resolution [O iii ] studyof Casta˜neda (1988) although not recognized there as a com-mon velocity system. Because of lying between V MIF , [O III] and V scat , [O III] they are difficult to identify except whenthey are strong relative to a clearly defined V scat , [O III] fea-ture.The V new , [O III] component appears in three areas andthe results are shown in Table 7. There we see thatthere are well defined groupings of the velocity differences. < V new , [O III] -V MIF , [O III] > is about 11 km s − , much lessthan the 21 km s − for < V new , [O III] -V low , [O III] > . TheS(V new , [O III] )/S(V low , [O III] ) ratio is usually similar, whichmakes it unlikely that V low , [O III] and V new , [O III] are related.The fact that the V new , [O III] component is characteristi-cally strong everywhere (except in S-N Profile North) yetis absent in [N ii ], that the velocity difference from V MIF isnearly constant, and V new , [O III] is too large to be the resultof back-scattering all argue that it must arise from a fully ionized layer. In Section 6.3.1 of Abel et al. (2016) thoseauthors discuss outflow away from the observer from theionized side from Component B of the Veil (the Veil compo-nent argued to be closer to θ Ori-C). They predict that this[O iii ] should occur at 28.1 km s − . This is indistinguishablythe same in all the sampled regions in Table 7 except forthe peculiar S-N Profile South Region, which is obviouslyvery different from other regions of the MIF. Abel et al.(2016) also predicted the outflowing [N ii ] emission to be at27.3 km s − . This lower velocity component would be lostin the much stronger higher velocity V MIF , [N II] component.How this outflowing emission from the Veil that we see in[O iii ] can be dynamically linked to the V MIF , [O III] layer isan open question. Nevertheless, the most likely explanationof the V new , [O III] component is that it is the outflow fromthe Veil predicted by Abel et al. (2016). The most dramatic changes in velocity occur in the Orion-S Crossing and those in the NE-SW and E-W Profiles areeasiest to understand. In these profiles V
MIF and V low aregreater than the adjacent regions by about 4 – 5 km s − in both [N ii ] and [O iii ]. This indicates that the underlyingPDR is moving away with respect to the nearby MIF.The changes of the signal levels in the NE-SW andE-W Profiles behave very differently than the velocities.S(V MIF , [N II] ) progresses across the Crossing as if it werea simple ionized layer. In contrast, the S(V low , [N II] ) val-ues increase significantly within the Crossing. S(V MIF , [O III] )drops dramatically in the core of the Crossing, while theS(V low , [O III] ) values increase to about the same signal levelas the V MIF , [O III] signals. The combined signals of theV MIF , [O III] and the V low , [O III] within the Crossing are aboutequal to an interpolation of nearby values. It is as if both theV MIF , [O III] and the V low , [O III] layers are receiving similaramounts of ionizing photons above 24.6 eV that are neces-sary to produce an He + +H + zone.In the S-N Profile there is a familiar increase inV MIF , [N II] at the crossing. However, the V MIF , [O III] tran-sitions from background values south of the Crossing to thehigher values to the north in the middle of the Crossing. Al-though the S(V MIF , [N II] ) varies smoothly across the Cross-ing, S(V low , [N II] ) temporarily increases there. Proceedingfrom north to south, we see that the S(V MIF , [O III] ) drops MNRAS , 1–29 (2018) ayers in the Central Orion Nebula Table 7.
Averaged Properties of the V new , [O III] components outside the Orion-S Crossing Area*Region S(V new , [O III] )/S(V MIF , [O III] ) < V low , [O III] > < V MIF , [O III] > < V new , [O III] > V new , [O III] -V MIF V new , [O III] -V low NE-Region (18) 0.17 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ◦ (4) 0.14 ± ± ± ± ± ± − . Parentheses indicate the number of samples. abruptly at the Crossing, never to fully recover, probablybecause of increasing distance from θ Ori-C.V low , [N II] appears a few times in the S-N Profile NorthRegion, then becomes common in the Crossing and the re-gion to its south. Again V low , [O III] appears a few times in theNorth region, then does not reappear within the Crossing orto its south. The strength of the low velocity [O iii ] in andsouth of the Orion-S Crossing is what cause its identificationas the V MIF , [O III] component, even though its velocities arecharacteristic of the V low , [O III] component in the rest of theSpectra. The most straight-forward interpretation is that thehigher energy photons necessary to create an He + +H + zonedo not reach the S-N Profile South Region, probably becauseit is shadowed by the Orion-S Cloud. But, the lower energyphotons necessary to create the He o +H + zone do reach thesurface of the PDR. Since the [O iii ] emitting zone is furtherfrom the PDR, the contrasting behavior of [N ii ] and [O iii ]can be explained by the Orion-S Crossing being a cloud highenough to block the high energy photons that usually createthe [O iii ] emitting layer and the lower energy photons passunder the cloud to continue the MIF [N ii ] emission.The V new , [O III] velocity system, whose properties arepresented in Table 7 appears eight times in the Orion-SCrossing six of which in the core region ( – ) haveunusually large S(V new , [O III] )/S(V MIF ) values 0.53 ± .48 al-though their velocities (18.6 ± − ) are lower than inother regions in Table 7 except for the S-N Profile South Re-gion (18.5 ± − ). The S(V new , [O III] )/S(V MIF ) ratiosare very different than the other regions, with this ratio be-ing greater than 1.0 for two ( and ). These two Spectraare where the S-N Profile crosses the Dark Arc feature. In Section 6.1 we found evidence that the Red Fan Cloudis a low ionization discrete cloud moving into the OMC atabout 8 ± − . In the middle of the RFC the V low , [N II] value changes abruptly indicating that the RFC physicallyinterferes with the layer producing V low , [N II] . As first notedin the discussion of the Orion-S Crossing we again see thatthe S(V low , [O III] ) and S(V MIF , [O III] ) components are nearlythe same in the RFC and that S(V low , [N II] ) is higher withinthe Red Fan Cloud than in the adjacent background regions.These many similarities of the Orion-S Crossing and theRFC (velocities and velocity changes, S(V low ) and S(V MIF )changes) indicate that the two regions share several physicalcharacteristics.However, it is not the case in either region that we areseeing the ionized layer on a foreground cloud. This would require that the cloud is located far enough from the MIBthat only the far side of the cloud is photo-ionized by θ Ori-C. In such a model one would expect the evaporative flow tobe towards the OMC, which could explain the high value ofV
MIF , [N II] , but in this model the V MIF , [O III] velocity shouldbe more positive than the V MIF , [N II] velocity and it is not. In Section 6.2 we saw that there were no statistically relevantdifferences in the velocities of the different [N ii ] and [O iii ]components. Changes in the signal ratios behaved in thefashion expected, except that greater changes in the signalratios for the V low components suggest that the V low emit-ting layer is more strongly affected by the SW Cloud thanthe V MIF emitting layer. This produces a dilemma becausethis would be the only evidence that the V low emitting layeris further than the V
MIF emitting layer. This is in contrastwith the Dark Bay area presented in Section 6.3
In this intensive presentation of observational data and pat-terns, many questions of interpretation remain open. V low . There seems to be a correlation of the V low andV
MIF components. Given that this component produces ab-sorption lines in the Trapezium stars and is distant from theMIF, what gives rise to this correlation?
The S-N Profile South Region.
What occurs in thisregion causing the V
MIF , [O III] component to have velocitiesnormally associated with V low , [O III] , while V MIF , [N II] doesnot? Ionization in the V low
Layer.
Given that one seesionized states in the absorption lines of the Veil, how is thisreconciled with the low ionization [O i ] and [S ii ] features inthe intervening V low system? The regions with the most positive V
MIF , [N II] . These are the Orion-S Crossing and Red Fan Cloud features(the other high velocity region at S-N Profile is appar-ently linked to the High Ionization Arc). These velocities arenot caused by highly tilted regions because there is no lo-calized peak in S(V MIF , [N II] ). This leaves the uncomfortableanswer that the Orion-S Crossing and Red Fan Cloud arerapidly approaching the OMC. V new , [O III] . Is this component an outflow away fromComponent B of the foreground Veil? How is the layer pro-ducing high ionization emission dynamically linked to the[O iii ] emission coming from near the MIF?
Is the Orion-S Cloud a finger-tip or a cloud?
Doesthe selective shadowing of the MIF [O iii ] producing emission
MNRAS000
MNRAS000 , 1–29 (2018) C. R. O’Dell outside the Orion-S Crossing indicate that the Cloud is anisolated cloud and not the fingertip of a column of materialpointed towards θ Ori-C? • Deconvolution of spectra across the Huygens Regionof the Orion Nebula establish that there are several distinctvelocity systems, each having their own origin. • The faint red velocity component (V scat ) is confirmedas back-scattering of MIF emission by grains in the densebackground PDR that lies just beyond the MIB. • V MIF velocities vary at small scales due to tilted re-gions in the MIF, but at scales of > (cid:48)(cid:48) , the variations arecaused by local variations in motion of OMC gas near thePDR. • A low velocity (V low ) component is seen in many areasof the Huygens Region. The layer producing V low must bebeyond (away from the observer) with respect to the fore-ground Veil (because the Veil is very low ionization) butmust lie closer to the observer than the Trapezium stars(because the absorption line velocities seen in the stars co-incide with those in V low ). • In the Orion-S Crossing both V
MIF , [N II] andV MIF , [O III] move about 4 km s − away from the observerwith respect to the adjacent regions. In this region theS(V MIF , [O III] ) drops sharply while the S(V low , [O III] ) com-ponents increase enough that the sum of these two compo-nents is about what is expected from interpolating acrossadjacent regions. • In the region south of the Orion-S Crossing the bright-est [O iii ] component has velocities characteristic of the V low component and V low , [O III] is not seen. In this same regionthe V MIF , [N II] component remains at velocities typical ofthe MIF and a V low , [N II] component is seen. • In addition to the Orion-S Crossing, regions of similarsignals of the V
MIF , [O III] and V low , [O III] components arefound in an extended region to the west of the Crossing. • The above points indicate that the Orion-S Cloudis selectively shadowing the higher layers where theV
MIF , [O III] components are formed but that θ Ori-C lowerenergy photons that create the [N ii ] emitting layer pass be-low the Orion-S Cloud.This shadowing could also explainthe drop in S(V MIF , [O III] ) and the increase in S(V low , [O III] )in the Orion-S Crossing. • The Red Fan feature is identified as an isolated cloudmoving rapidly towards the OMC and has many of the fea-tures of the Orion-S Cloud Region. • A new velocity system (V new , [O III] ) is found in manyregions. The velocity of this [O iii ] component and the lack ofa detected [N ii ] component is interpreted as flow away fromthe observer of the Veil’s ionized layer that faces θ Ori-C.
ACKNOWLEDGEMENTS
The author is grateful to Will Henney of the Instituteof Radio Astronomy and Astrophysics for digital copiesof the spectral Atlas, including [O i ] without the sky linesubtracted, and for clarification of the properties of scat-tered emission lines. Gary Ferland of the University of Ken-tucky extracted and shared the information from the Cloudy model for the central nebula that allowed estimating thethickness of the [N ii ] and [O iii ] emitting layers and was aconstructive reviewer of the penultimate draft manuscript.In this study we have made extensive use of the SIM-BAD database, operated at CDS, Strasbourg, France andits mirror site at Harvard University and to NASA’s Astro-physics Data System Bibliographic Services. We have usedIRAF, which is distributed by the National Optical Astron-omy Observatories, which is operated by the Association ofUniversities for Research in Astronomy, Inc. under cooper-ative agreement with the National Science foundation.The observational data were obtained from observationswith the NASA/ESA Hubble Space Telescope, obtained atthe Space Telescope Science Institute, which is operated bythe Association of Universities for Research in Astronomy,Inc., under NASA Contract No. NAS 5-26555; the Kitt PeakNational Observatory and the Cerro Tololo InteramericanObservatory operated by the Association of Universities forResearch in Astronomy, Inc., under cooperative agreementwith the National Science Foundation; and the San PedroMartir Observatory operated by the Universidad NacionalAut´onoma de M´exico.Since this is likely to be the last lead-author paper onthe Orion Nebula by the author, it is appropriate to ac-knowledge the colleagues who have enabled this series (thefirst was (O’Dell & Hubbard 1965)) of studies. These beginwith Arthur D. Code, Rudolph L. Minkowski and DonaldE. Osterbrock, all then at the University of Wisconsin; IraS. Bowen, Guido M¨unch, and Olin C. Wilson of the (then)Mt. Wilson and Palomar Observatories; Manuel Peimbertthen a graduate student at the University of California atBerkeley, now at the Institute for Astronomy-Mexico City;my former graduate students at Rice University, Hector O.Casta˜neda, Zheng Wen, Michael R. Jones, Xihai Hu, andTakao Doi; and my long time collaborators Nicholas P. Abelof the University of Cincinnati, Gary J. Ferland of the Uni-versity of Kentucky, and William J. Henney of the Institutefor Radio Astronomy and Astrophysics-Morelia, Mexico. APPENDIX A: ACCURACY OF THEDECONVOLUTIONS
The emission line deconvolution task ‘splot’ takes assumedapproximate velocities of line components and produces bestfits. Our method of measurement was to first do a fit to asingle assumed line. From the region of a poor fit a secondline was added to the solution and if there was a third regionof poor fit a third line was added. When well separated highvelocity lines were seen, they were treated separately. Anillustration of our observed spectra is shown in Fig. A1. Itis from Sample -30 (cid:48)(cid:48) ,Line 17.This procedure was tested by creating a series of sim-ulated spectra, then deconvolution of them in the abovemanner. In order to best mimic the nebular spectra we as-sumed three components for the simulations. Since the sep-arations and typical FWHM are similar for both [N ii ] and[O iii ], it was not necessary to make a line-specific seriesof calculations. We adopted velocity shifts of ±
17 km s − and a FWHM of 16 km s − . The typical value for theS(V scat )/S(V MIF ) ratio of 0.06, was employed. This left onlythe more highly variable S(V low )/S(V
MIF ) ratio as the quan-
MNRAS , 1–29 (2018) ayers in the Central Orion Nebula
450 -25 250 50 75
Sample [N II] Spectrum
MIF low scat
Figure A1.
The [N ii ] spectrum of Sample -30 (cid:48)(cid:48) ,Line 17 is pre-sented for comparison with the following artificial spectrum, asdescribed in this section. The heavy line depicts the observed pro-file, the dashed lines the components given from deconvolution.For each component the velocity, the total signal relative to thestrongest (MIF) component, and the FWHM are given. The sumof the three fitted components falls within the width of the lineshowing the observed profile. Artificial Spectrum
250 V
MIF low -17 km s-1Ratio 0.10FWHM 16 km s-1V scat
17 km s-1Ratio 0.06FWHM 16 km s-1-50
Figure A2.
An artificial spectrum used in the determinationof accuracies of the results of using task ‘splot’ as described inSection A. It is model Std+Low 0.100 from Table A1 The fit ofthe sum of the individual components is within the breadth ofthe line tracing the artificial line profile. The sum of the threefitted components falls within the width of the line showing thecalculated profile tity changing within the set of models. The parameters ofthe models are given in the left column in Table A1 andthe results of the splot deconvolution are in the remainingcolumns. The model most closely agreeing with the repre-sentative spectrum (Fig. A1) is shown in Fig. A2.Examination of Table A1 shows that the derived char-acteristics of the V
MIF and V scat components are alwaysclose to the assumed values, even over a wide range ofS(V low )/S(V
MIF ) values. The derived values of V low show nosystematic trends with S(V low )/S(V
MIF ) although the devi-ation is -1.7 km s − when there is only a one percent con- tribution of S(V low ). The derived S(V low )/S(V MIF ) valuesagree with the assumptions down to a five percent contribu-tion but the derived S(V low )/S(V
MIF ) values are 1.4 timestoo large for the 2.5 and 1.0 percent models.The above comparison indicates that the velocity datafor V low is accurate down through a true contribution of 2.5percent and the relative signals are accurate down to fivepercent, but at assumed values of 2.5 percent and 1.0 per-cent, the derived values are about 40 percent too large. Thismeans that almost all of the data used in our discussions areclose to the real values, with the caveat that the lower S/Nindividual slit spectra must have more uncertainty.
MNRAS000
MNRAS000 , 1–29 (2018) C. R. O’Dell
Table A1.
Deconvolution of Artificial Spectra*Spectrum V
MIF
FWHM(MIF) V scat
S(V scat )/S(V
MIF ) FWHM(V scat ) V low
S(V low )/S(V
MIF ) FWHM(V low )Assumed Value 0.0 16.0 17.0 0.06 16.0 -17.0 — 16.0MIF 0.0 16.0 — — — — — —MIF+Scat 0.0 16.0 17.0 0.060 15.9 — — —Std**+Low 0.01 0.0 16.0 16.8 0.063 16.4 -15.3 0.014 18.3Std+Low 0.025 0.0 15.8 16.8 0.067 14.5 -17.1 0.036 17.7Std+Low 0.050 0.0 16.0 17.4 0.059 16.0 -17.2 0.049 16.0Std+Low 0.075 0.0 16.1 17.2 0.059 16.1 -17.4 0.072 15.9Std+Low 0.100 0.0 16.1 17.3 0.059 16.0 -17.1 0.099 16.0Std+Low 0.150 0.0 15.8 17.0 0.061 16.4 -17.1 0.145 16.0Std+Low 0.200 0.0 16.0 17.0 0.064 16.4 -17.0 0.201 16.0*All velocities are in km s − with respect to the assumed MIF component at 0.0 km s − . **Std means the MIF+Scat spectrum REFERENCES
Abel, N. P., Ferland, G. J., O’Dell, C. R., Troland, T. H., 2016,ApJ, 819, 136Arthur, S. J., Medina, S.-N. X., Henney, W. J., 2016, MNRAS,423, 2864Baldwin, J. A., Ferland, G. J., Martin, P. G., Corbin, M. R., Cota,S. A., Peterson, B. M., Sletteback, A., 1991, ApJ, 374, 580Balick, B., Gammon, R. H., Hjellming, R. M., 1974, PASP, 86,616Bally, J., O’Dell, C. R., McCaughrean, M. J., 2000, AJ, 119, 2119Casta˜neda, H. O., 1988, Ap&SS 67, 93Deharveng, L., 1973, A&A, 29, 341Doi, T., O’Dell, C. R., Hartigan, P., 2004, AJ, 127, 3456F´ur´esz, G., Hartmann, L W., Megeath, S.T., Szentgyorgyi, A. H.,Hamden, E. T., 2008, ApJ, 676, 1109Garc´ıa-D´ıaz, Ma. T., Henney, W. J., 2007, AJ, 133, 952Garc´ıa-D´ıaz , Ma.-T., Henney, W. J., L´opez, J. A., Doi, T., 2008,Rev. Mex. Astron. Astrofis., 44, 181Goicoechea, J. R., et al. 2015, ApJ, 812, 75Goicoechea, J. R., Pety, J., et al. 2016, Nature, 537, 207Goudis, C. 1982, The Orion complex: A case study of interstellarmatter (Dordrecht, Netherlands, D. Reidel Publishing Co.),Astrophysics and Space Science Library,Volume 90G¨udel, M., Briggs, K. R., Montmerle, T., Audard, M., Rebull, L.,Skinner, S. L., 2008, Science, 319, 309Henney, W. J., 1994, Rev. Mex. Astron. Astrofis., 29, 192Henney, W. J., 1998, ApJ, 503, 760Henney, W. J., Arthur, S. J., Garc´ıa-D´ıaz, Ma.-T., 2005, ApJ,627, 813Hobbs, L. M.,1978, ApJS, 38, 129Kounkel, M., et al. 2016, ApJ, 834, 142Leroy, J. L., Le Borgne, J. F., 1987, A&A, 186, 322Menten, K. M. Reid, M. J., Forbrich, J., Brunthaler, A., 2007,A&A, 474, 515Mesa-Delgado, A., N´u˜nez-D´ıaz, M., Esteban, C., L´opez-Mart´ın,L., Garc´ıa-Rojas, J., 2011, MNRAS, 417, 420 (M-D),Muench, A., Getman, K., Hillenbrand, L, Preibisch, T., 2008, inHandbook of Star Forming Regions, Vol. 1:The Northern Sky,ASP Monograph Publications, Vol. 4, ed. B. Reipurth, p. 483O’Dell, C. R., 2001, ARAA, 39, 9O’Dell, C. R., 2009, PASP, 121, 428O’Dell, C. R., Ferland, G. J., Henney, W. J., Peimbert, M.,Garc´ıa-D´ıaz, Ma. T., 2015, AJ, 150, 108 (O15)O’Dell, C. R., Ferland, G. J., Peimbert, M., 2017, MNRAS, 464,4835 (O17a)O’Dell, C. R., Harris, J. A., 2010, AJ, 140, 985O’Dell, C. R., Henney, W. J., Abel, N. P., Ferland, G. J., Arthur,S. J., 2009, AJ, 137, 367O’Dell, C. R., Hubbard, W. B., 1965, ApJ, 142, 591 O’Dell, C. R., Kollatschny, W., Ferland, G. J., 2017, ApJ, 837,151O’Dell, C. R., Muench, A., Smith, N., Zapata, L., 2008, in Hand-book of Star Forming Regions, Vol. 1:The Northern Sky, ASPMonograph Publications, Vol. 4, ed. B. Reipurth, p. 544O’Dell, C. R., Valk, J. H., Wen, Z., Meyer, D. M., 1993a, ApJ,403, 478O’Dell, C. R., Walter, D. K., Dufour, R. J., 1992, ApJ, 399, L67O’Dell, C. R., Wen, Z., 1992, ApJ, 387, 229O’Dell, C. R., Wong, S. K., 1996, AJ, 111, 846O’Dell, C. R., Yusef-Zadeh, F., 2000, ApJ, 120, 382 (OY-Z)Ossenkopf, V., R¨ollig, M., Neufeld, D. A., Pilleri, P., Lis, D. C.,Fuente, A., van der Tak, F. F. S., Bergin, E., 2013, A&A, 550,57Osterbrock, D. E., Ferland, G. J., 2006, Astrophysics of GaseousNebulae and Active Galactic Nuclei (second edition), Univer-sity Science Books (Mill Valley)Pelligrini, E. W., Baldwin, J. A., Ferland, G. J., Shaw, G., Heath-cote, S., 2009, ApJ, 693, 285Shaw, G., Ferland, G. J., Henney, W. J., Stancil, P. C., Abel, N.P., Pellegrini, E. W., Baldwin, J. A., van Hoof, P. A. M., 2009,ApJ, 701, 677Sicilia-Aguilar, A., et al., 2005, AJ, 129, 363Tielens, A. G. G. M., Meixner, M. M., van der Werf, P. P., Breg-man, J., Tauber, J. A., Stutzki, J. Rank, D., 1993, Science,262, 86Troland, T. H., Goss, W. M., Brogan, C. L., Crutcher, R. M.,Roberts, D. A., 2016, ApJ, 825, 2van der Werf, P. P., Goss, W. M., 1989, A&A, 224, 229van der Werf, P. P., Goss, W. M., 1990, A&A, 364, 157van der Werf, P. P., Goss, W. M., O’Dell, C. R., 2013, ApJ, 762,101Walmsley, C. M., Natta, A., Oliva, E., Testi, L., 2000, A&A, 364,301Weilbacher, P. M., et al. 2015, A&A, 582, A114Wilson, O. C., M¨unch, G., Flather, E. M., Coffeen, M. F., 1959,ApJS, 4, 199Zuckerman, B.,1973, ApJ, 183, 163 MNRAS , 1–29 (2018) ayers in the Central Orion Nebula This paper has been typeset from a TEX/L A TEX file prepared bythe author.MNRAS000