PROPLYDS AROUND A B1 STAR: 42 ORIONIS IN NGC 1977

The star formation environment is likely to affect the evolution of protostellar and protoplanetary disks. Thermally driven winds, heated by ultraviolet radiation from massive stars, can shorten the lifetime of disks around neighboring low mass stars with potentially important implications for giant and icy planet formation. Observational support for disk destruction in strongly irradiated environments is provided by the reduction of disk fraction in the vicinity of O stars in clusters such as NGC 6611 (Guarcello et al. 2007, 2009, 2010), Pismis24 (Fang et al. 2012), NGC 2244 (Balog et al. 2007), the Arches Cluster (Stolte et al. 2010), and Cygnus OB2 (Guarcello et al. 2016), although other studies (e.g., Roccatagliata et al. 2011 in IC 1795 and Richert et al. 2015 in NGC 6611) have not found such a decline. Likewise Mann et al. (2014) (see also Mann & Williams 2010) have shown that the dust component of disks tends to be less massive in the immediate vicinity of the dominant O star in the Orion Nebula Cluster (ONC), θ¹Ori C (though Mann et al. 2015 found no such trend in another cluster containing O stars, NGC 2024). Finally, it should be noted that although a decline of disk fraction in crowded areas could in principle be instead attributed to dynamical interactions (e.g., Pfalzner 2004; Protegies Zwart 2016) it can be shown that for a normal IMF, photoevaporation becomes an important disk destruction mechanism at substantially lower densities than dynamical encounters (Scally & Clarke 2001).

However, the most dramatic evidence of disk destruction is provided by the large number of proplyds, cometary objects imaged in Hα in the vicinity of θ¹Ori C in the ONC (O'Dell et al. 1993; Bally et al. 2000).⁵ The sizes (few hundred au) and morphologies of proplyds are well explained by a model in which FUV radiation from θ¹Ori C drives a neutral disk wind; the bright cometary feature then results from the interaction of ionizing radiation (also from θ¹Ori C) with this neutral wind (Johnstone et al. 1998). Radio free–free measurements of the ONC proplyds imply mass loss rates of ∼10⁻⁷ M_⊙ yr⁻¹ (Churchwell et al. 1987), which would result in extremely short disk lifetimes. It is therefore unsurprising that objects experiencing such extreme photoevaporation are observed rather rarely, with relatively small samples being identified in Carina (Smith et al. 2003), Pismis 24 (Fang et al. 2012), NGC3603 (Brandner et al. 2000), and CygOB2 (Wright et al. 2012); in the latter two environments the "giant" proplyds (on a scale of 10⁴–10⁵ au) are not necessarily derived from disk photoevaporation (Sahai et al. 2012a, 2012b, though see also Guarcello et al. 2014).

To date proplyds have only been detected around stars of spectral type O, where the high mass loss rates both measured and predicted imply that this should be a short-lived evolutionary stage. Investigating the formation potential of proplyds around O stars, Störzer & Hollenbach (1999) also argued that proplyds should only be detectable in the very close vicinity of such objects, where the FUV field exceeds 5 × 10⁴ G₀ since their calculations implied that at lower G₀, neutral wind driving would be too weak to push the ionization front (IF) away from the disk, given the strong ionizing flux produced by O stars (here G₀ is the local FUV interstellar field , 1.6 × 10⁻³ erg cm⁻² s⁻¹). As noted by Störzer & Hollenbach (1999), this lower limit on FUV field strength required for proplyd production is however sensitive to stellar spectral type since this controls the relative strength of the FUV and ionizing radiation fields.

More recent studies have emphasized that significant winds can be driven at considerably lower G₀ values (Adams et al. 2004; Facchini et al. 2016). The radius of a proplyd produced by interaction between this neutral wind and the B star's ionizing luminosity could then be used to measure the mass loss rates at lower G₀, a quantity that is of great importance in assessing the significance of photoevaporation in a wide range of star-forming environments. Although the lower ionizing flux of the B star would produce structures of lower surface brightness than in the O star case, it would also imply spatially larger proplyds for a given wind rate, rendering them potentially more detectable than in the O star case. Moreover, the fact that lower wind mass loss rates are expected implies that such structures would be longer-lived than their counterparts around O stars and hence more abundant on those grounds.

Here we present seven proplyds discovered around 42 Ori (c Ori, HD37018, B1V) in NGC 1977, an H ii region located at ∼30' north of ONC at ∼414 pc distance (Menten et al. 2007). There is no O star in the region, but NGC 1977 contains three young B stars and at least ∼170 young stellar objects (Peterson & Megeath 2008). 42 Ori has the earliest spectral type and is the major source for ionizing photons in NGC 1977. An irradiated disk near 42 Ori has been detected by Bally et al. (2012) in the Hubble Space Telescope (HST) image using Hα filter (F658N). They identified a bent protostellar jet HH1064 from Parengo 2042 (the Spindle) in NGC 1977 with numerous bow shock features. They argue that the arc feature in the Hα Spindle is centered on the star and the brightened side of its arc is facing toward 42 Ori, suggesting that it may be a proplyd. The seven proplyds that we describe here (Figure 1) were discovered in archival Spitzer and HST images; we will discuss the implication of finding proplyds around a B star and the wider implications for disk clearing in UV environments.

Zoom In Zoom Out Reset image size

We used the Spitzer Space Telescope/IRAC (3.6, 4.5, 5.8, 8.0 μm) archival data for the detection of a dusty proplyd, KCFF 1. All four IRAC bands show clear detection of the central source and its tail pointing radially away from the B1 star, 42 Ori (Figure 2). The mosaic images were obtained from the Spitzer Heritage Archive (SHA).⁶ The images were processed by the pipeline version S18.25.0, and the mosaic images were made using MOPEX version 18.5.4 and super mosaic pipeline version 2.0. The final mosaic image has resolution of 06. The median exposure times (seconds per pixel) of the images are 52 s for long exposures and 2 s for short exposures.

Zoom In Zoom Out Reset image size

We used a set of archival data of the HST/the Advanced Camera for Surveys (ACS) to identify proplyds KCFF 2–7 (Figure 3). The data were obtained from the MAST archive. The HST ACS/WFC images was observed on 2010 November 12 and 2011 November 14 (PI: Bally, proposal ID 12250, cycle 18) using Hα (F658N) filter with 2460 and 2510 s exposure times. The F658N narrow band filter transmits both Hα and and [N ii] lines. The observational details were discussed in Bally et al. (2012). The final mosaic images have pixel sizes of 005.⁷

Zoom In Zoom Out Reset image size

We present a total of seven new proplyds in NGC 1977 in Table 1 (also see Figure 1). The KCFF source ID number (1st column) is used when we discuss individual sources here and we also assign names for the proplyds based on their coordinates (second column of Table 1) similar to the designation given for proplyds in the ONC (O'Dell 1998). We use the last three digits in R.A. (J2000) (sss after 5^h35^m20^s) and the last five digits of decl. (J2000) (m:ssss) to assign names (Table 1, 2nd column) for the coordinate-based names of the seven proplyds. For example, proplyd KCFF 1 with coordinates 5:35:24.142, −4:50:09.21 is named as 414-50092.

Table 1.  Properties of Proplyds in Vicinity of 42 Ori

Source ID	Name	R.A. (J2000.0)	Decl. (J2000.0)	Teff	Distance to 42 Ori		rIFb		rdb
KCFF#		(hh mm ss.sss)	(° ' '')	(K)	('')	(pc)a	('')	(au)c	('')	(au)c
1	414-50092	5 35 24.142	−4 50 09.21	3243d	17.59	0.036	...	...	...	...
2	551-51201	5 35 25.505	−4 51 20.11	3630e	70.72	0.138	1.35	540	...	...
3	881-50220	5 35 28.812	−4 50 22.04	...	84.65	0.166	0.40	160	...	...
4	808-50020	5 35 28.076	−4 50 02.06	...	75.45	0.148	0.46	184	0.17	68
5	338-51180	5 35 23.381	−4 51 18.06	...	59.42	0.116	0.49	196	0.12	48
6	252-52365	5 35 22.522	−4 52 36.56	3328d	138.14	0.268	0.73	292	...	...
7	313-48277	5 35 23.134	−4 48 27.69	3847e	111.05	0.215	0.56	224	...	...

Notes.

^aProjected distances from 42 Ori to proplyds. We use the distance of NGC 1977 to be 400pc in this work. ^bUncertainty of measuring sizes of ionization front and central disk size ranges about 0025–005. ^cCalculations using d ∼ 400 pc to NGC 1977. ^dDa Rio et al. (2016). ^eM. Fang et al. (2016, in preparation)

Download table as:  ASCIITypeset image

The closest proplyd, KCFF 1 (414-50092, Figure 2), is discovered in the Spitzer images at a distance of ∼7000 au from the B1 star with its dusty tail evaporating away from 42 Ori. Sizes and distances from 42 Ori are calculated assuming that the distance to NGC 1977 is 400 pc (throughout the paper). This source is detected in all four IRAC bands and in MIPS 24 μm, but not in the HST image because the HST survey area did not cover KCFF-1. The central source is an M3.5–M4 star, and its dusty tail is about 20'' (∼8000 au) long.

Six other proplyds (KCFF 2−7) are identified in the HST ACS/WFC image (Figure 3). Bally et al. (2012) presented Parengo 2042, the Spindle, residing in a large proplyd. However, we do not discuss this source in this paper, since its tail is not clearly visible in the HST image alone. We measure the radius of the IF (r_IF) and disk radius (r_d) in a manner similar to Section 3.2 in Vincente & Alves (2005), where they measure the IF chord diameter. For the brightest proplyd, KCFF-2 (551-51201), we fit a circle to ∼30% higher than the background level, and for other fainter sources we fit a circle to ∼10% higher than the background using contour, radial profile, and cross-section analyses. The measurement uncertainties in size estimation are about half a pixel to a pixel (∼0025–005, ∼10–20 au); this is introduced when we fit a circle to the 10% or 30% above the average background level, and also because of slight departure from a circular shape of the proplyd's IF. The IF radius is estimated as the radius of the circle fitting the IF chord.

All proplyds are found within 0.3 pc distance from 42 Ori, with their IF pointing toward 42 Ori and their tails pointing radially away from it (see Figures 2 and 3). KCFF 2 has a bright thick IF and a bright central source with r_IF ∼ 135 (540 au) and a central source with radius ∼035 (140 au). We do not list this central source as a disk source, since the central source does not have a clear disk morphology. KCFF 3 (881-50220) has r_IF ∼ 04 (160 au), but the central source is not detected while the ionizing front is very bright.

KCFF 4 and 5 have resolved central sources. The central source of KCFF 4 (808-50020) is an illuminated disk or quasi-spherical material with r_d ∼ 017 (∼70 au), which is slightly asymmetric with the semimajor axis, located inside the proplyd with r_IF ∼ 046 (184 au). The central source of KCFF 5 (338-51180) is smaller than KCFF-4 with r_d ∼ 012 (48 au) and r_IF ∼ 049 (196 au). The size of r_IF for these two sources are similar.⁸

KCFF 3–5 are undetected in the Spitzer photometry catalog from Megeath et al. (2012), which would suggest that these sources are very low mass objects. The Spitzer catalog can detect very low mass objects, down to brown dwarfs mass objects, because their Spitzer IRAC band 1 (3.6 μm) data go as deep as ∼16 mag with 10σ detection in Orion. The non-detection of these sources in the Spitzer IRAC band 1 would equate to an upper mass limit of around 15 Jupiter masses according to the evolutionary models of Baraffe et al. (2015). Partial obscuration of the central object by the disk would, however, imply that the mass was considerably greater than this. We incline toward this explanation on the grounds that the disk masses and consequent disk lifetimes against photoevaporation would otherwise be very short. Note that by invoking obscuration we are requiring some quasi-spherical distribution in the vicinity of the star, or an inclined edge-on disk, which may or may not impact our interpretation of the ∼50 au scale structures within KCFF 4 and 5 as being disks.

KCFF 6 (252-52365) has a bright central object with low surface brightness IF with r_IF ∼ 292 au. The IF size appears to be larger than other proplyds, except KCFF 2. The central source harbors an object with T_eff of 3328 K (Da Rio et al. 2016). The IF is faint, but clearly shows a half-circular morphology as a proplyd, but we note that its tail is extremely faint (Figure 4). KCFF 7 (313-48277) has a similar r_IF size and a faint tail. The central source has the highest T_eff (3847 K) among the seven proplyds in Table 1.

Zoom In Zoom Out Reset image size

In order to estimate the expected size of proplyds in this environment (specifically the distance between the center of the proplyd source and the offset IF) we need to (i) estimate the ionizing flux from the neighboring B star, (ii) estimate the expected mass loss rate in the neutral wind from the proplyd, and (iii) impose a condition of ionization balance in the ionized flow close to the IF. This approach can be applied whatever the mechanism driving the neutral wind (Clarke & Owen 2015). Here we follow Johnstone et al. (1998) in assuming that this is driven by the FUV flux from the same neighboring massive star which also provides the ionizing photon source. We however differ from Johnstone et al. (1998) in that we use new calculations of photoevaporative mass loss in regions of relatively low FUV fields (Facchini et al. 2016), noting that the observed proplyds in NGC 1977 are exposed to an FUV field that is less than that irradiating the well-studied proplyds in the ONC.

All proplyds here (KCFF 1–7) have structures with tails pointing radially away from 42 Ori. It is thus reasonable to associate the ionization source with this star; indeed the relatively close proximity of these sources to what is the earliest type star in the region strengthens this expectation. We estimate the stellar mass from the B1V spectral type to be around 10 M_⊙ (e.g., Lorenz et al. 2005; Lorenzo et al. 2016) and obtain ionizing photon outputs and FUV luminosities of 10⁴⁵ s⁻¹ and 2 × 10³⁷ erg s⁻¹ from Diaz-Miller et al. (1998) and Armitage (2000), respectively. We can then obtain a maximum FUV flux of ∼3000 G₀ in the vicinity of the proplyds. This maximum is obtained by neglecting dust extinction between the B star and the proplyd and by setting the distance between star and proplyd to be its separation on the sky, ∼0.2–0.3 pc. Störzer & Hollenbach (1999) showed that the creation of proplyds around O stars requires a minimum FUV flux of 5 × 10⁴ G₀. We here re-examine this issue using the lower ionizing fluxes of B stars and more recent models of neutral winds from protoplanetary disks in mild FUV environments by Facchini et al. (2016) (see also the pilot solutions of Adams et al. 2004). The mass loss rate in this regime is a sensitive function of the disk outer radius and also depends somewhat on the effect of grain growth in modifying the FUV opacity in the wind. As an example, we take the cases of moderate grain growth (maximum grain size of 3.5 μm) and disk radii of 40 and 50 au, for which the mass loss rates for an FUV field of 3000 G₀ are 10⁻⁹ and 10⁻⁸ M_⊙ yr⁻¹ respectively (see Figure 12 of Facchini et al. 2016). Combining Equations (6) and (10) of Johnstone et al. (1998) in order to remove their dependence on the explicit formula for FUV-driven mass loss rate as a function of system parameters, we obtain ${r}_{\mathrm{IF}}=1200\,\mathrm{au}{{\rm{\Phi }}}_{45}^{-1/3}{\dot{M}}_{-8}^{2/3}{d}_{\mathrm{pc}}^{2/3}$ where Φ₄₅ = Φ/10⁴⁵ s⁻¹(Φ being the stellar ionizing luminosity taking into account possible absorption by dust and the background nebula), ${\dot{M}}_{-8}=\dot{M}/{10}^{-8}{M}_{\odot }$ yr⁻¹, and d_pc is the distance to the ionizing source in parsecs. Adopting a distance between the proplyds and the B1 star of ∼0.2 pc and the mass loss rates given above, we obtain r_IF = 70 and 400 au for disk radii of 40 and 50 au respectively; r_IF values that are higher by a factor of a few are obtained in the case of dust growth to mm sizes (see the right hand panel of Figure 12 in Facchini et al. 2016).

We thus see that IFs with offset distances on the observed scale (∼200 au) are to be expected given the distances of the proplyds from the B1 star in NGC 1977.⁹

We have presented seven new proplyds around a B1 star, 42 Ori, in NGC 1977, about 30' north of the Orion Nebula (M42). This is, to our knowledge, the first time that proplyds (i.e., imaged structures showing clear evidence of external photoevaporation) have been detected in the neighborhood of stars of spectral type later than O. This discovery therefore opens up the possibility of testing theories of FUV photoevaporation in much weaker FUV background fields than has been possible hitherto (the estimated FUV background at the location of these new proplyds is ∼3000 G₀, more than an order of magnitude less than the estimated fields in the vicinity of the classical proplyds in the ONC).

Two proplyds (KCFF 3 and 4) contain bright interior structure on a scale of ∼50–70 au. We have used the recent models of Facchini et al. (2016) to estimate the mass loss rate from such disks in radiation fields of ∼3000 G₀ and find values in the range 10⁻⁹–10⁻⁸ M_⊙ yr⁻¹. Such rates are comparable with typical accretion rates in T Tauri stars, and therefore suggest that external photoevaporation will be a major player in the evolution of such disks. We have already noted that the (lack of) Spitzer detection in KCFF 4 and KCFF 5 might imply very low mass central objects in these cases; if so, the low expected mass of associated disks would in turn imply very short disk depletion timescales. Alternatively these masses may be substantially under-estimated if the source is partially obscured by an edge-on disk or quasi-spherical material.

Given estimates for the ionizing flux from 42 Ori that is incident on the proplyds, we use these mass loss estimates to predict the expected radii for offset IFs in these objects and obtain values of order 100 au (the predicted mass loss rates and the resulting proplyd radii are a sensitive function of disk radii, which can be estimated only in a few cases). These predicted proplyd sizes are in excellent agreement with the scales of structures seen in the proplyds of 42 Ori.

We thank the anonymous referee for the helpful comments. This work has been partially supported by the DISCSIM project, grant agreement 341137 funded by the European Research Council under ERC-2013-ADG. This work benefits from the EOS NExSS collaboration. The HST data presented in this paper were obtained from the Mikulski Archive for Space Telescopes (MAST) at https://archive.stsci.edu/hst/. STScI is operated by the Association of Universities for Research in Astronomy, Inc., under NASA contract NAS5-26555. Support for MAST for non-HST data is provided by the NASA Office of Space Science via grant NNX09AF08G and by other grants and contracts.