Extinction Associated with the Ultracompact HII Region S88B

Table of Contents

Abstract
Introduction

Observations & Reduction

        University of Rochester  (UR) Observations & Data Reduction

        2 Micron All Sky Survey (2MASS) Observations & Data Reduction

Discussion

        2MASS Results, Completeness Limit & Reddening Law

        2MASS Extinction Map

        UR Mapping Extinction by Comparing Br$\gamma$ Emission and 5 GHz Radio Continuum Emission

        UR Inner 2' Stellar Environment

        UR Conclusions

Figures

We present high spatial resolution mapping of the inner arcminute of the compact HII region S88B (G61.5+0.1) in Br Gamma (2.166 um) line emission and compare with 5 GHz radio continuum emission (Garay et al. 1993) to estimate the line-of-sight extinction to the ionized gas at 2.166 um.  On a larger scale (1 deg x 0.5 deg), the stellar counts derived from 2 Micron All Sky Survey (2MASS) images are used to determine the extent of the molecular cloud surrounding S88B.  By varying the assumed line-of-sight extinction, we fit a spectral type to stars appearing in J (1.25 um),  H (1.65 um), K (2.23 um) and 3.3 um broad band images taken with the University of Rochester 3^rd Generation
Infrared Camera and compare with other extinction determinations.

The extinction associated with the star forming region S88B (also known as G61.5+0.1) varies dramatically over relatively small spatial scales. Spherical and ultracompact HII regions (Wood & Churchwell 1989, Garay et al.
1993), so deeply embedded in cloud material that they are only evident  at radio wavelengths, appear in close proximity to optical nebulosity delineated by Halpha emission.  Pipher et al. (1977) and Evans et al. (1981) detect infrared nebulosity and point-like objects to the west of the radio emission sources and along the eastern edge of the optical emission.

Felli & Harten (1981), mapping the region at 5 GHz at low spatial resolution, report a flux
density for the whole region of 6.1 Jy, which is consistent with ionization by a single O5.5 star.  This prediction agrees
with the results of Garay $\etal$ who deduce an O6 star responsible for powering the emission detected from S88B-1.

Phillips & Mampaso (1991) report a north-south elongated outflow detected in CO at low resolution.  The peak of the red (south) lobe coincides most closely with the peak of the 12.6 um emission (Pipher et al.), but emission peaks at many wavelengths are located within ~20" of each other.  The blue (north) lobe velocities range from 15.8 to -3.7 km s^-1 while the red lobe velocity range is 24.9 to 44.4 km s^-1.

We have adopted a distance of 2.5 kpc to S88B, based on the spectroscopic parallax to LS II +24 9, a
member of the nearby OB association Vul OB1 (Garmany & Stencel 1992, Watson (private communication)).

The near-infrared images of S88B were obtained at the Mount Lemmon Observing Facility 1.5 meter telescope on 20 May 1997 and 17 October 1998.  Observations were taken with the University of Rochester Third Generation 256 x 256 InSb array camera.  Individual frames were  linearized, background subtracted, and flat fielded.
 
 
 

• We chose 16 co-added images covering approximately 70' x 30' centered on the S88B region that had been processed with the 2MASS pipeline (Kirkpatrick et al.1997) to analyze further (see Figs. 4a & 4b).

• While the 2MASS pipeline generates a catalog of stellar positions and magnitudes, for  regions of high stellar density (e.g. the galactic plane) more  complicated fitting of stellar psfs  is required.

• Co-added images were further processed under the IRAF software reduction  package using the DAOPHOT package and using specialized routines to extract psfs, magnitudes and positions of all stars.
• Detected  stars in each band are compared with artifacts identified by the 2MASS pipeline, and any star close to an artifact is rejected.
• Stars detected in each band within 1\arcsec of detected star positions in the other two bands are kept in the final list.
• A photometric offset was determined from stars falling in overlap regions between the images, then applied to correct each image.
• The final list of stars utilized in further astronomical analysis contained only stars detected in each band at 5 sigma or greater.

Results

• Two of the 16 fields were selected as control fields (i.e. minimal dust extinction), and the mean colors of stars in those fields were determined to be:

• There is a trend for the brighter stars to have a redder [H - Ks] color (probably red giant background stars) which may result in a bias (< 1.5 mag) in the visual  extinction determinations since the majority of stars detected through dense regions of the cloud will be  intrinsically brighter and thus redder background stars.

• We find a signal to noise ratio of 5 at Ks = 15, H = 15.5, and J = 16.5  (see Fig. 5).

• Turnover in the log cumulative luminosity plots is found  at Ks > 14.5 H > 15, and J > 15.5 (see Fig. 6).

• We thus conservatively adopt completeness limits of Ks = 14.5, H = 15, and J = 15.5.

Reddening Law

In order to deduce the color excess E(H-Ks) we adopt a near-infrared (NIR) dust extinction power law of the form:

where alpha = 1.8 (Martin 1990).
 

Using our control field colors to estimate typical intrinsic colors of S88 dust cloud background stars, we assume:

The extinction at Ks-band is:

and the extinction at V-band is:

A_V = 8.99 x A_Ks (Mathis 1990).

• To map the large-scale structure of the dust cloud we spatially convolve the individual stellar measurements of extinction with a circular filter 51" in size and subsample the image at the Nyquist frequency (see Fig. 7).

• In the region of highest extinction we detect an average of only 3 stars in our 51" aperture (an upper limit to the average number of foreground stars contaminating each pixel).

• In a region of low dust extinction we detect an average of greater than 40 stars in our 51" aperture. Thus foreground stars should comprise at most 10% of the stars contributing to our dust extinction  image. Since they will make up a progressively larger fraction of the stars in regions of higher dust obscuration, our map of dust extinction is a lower limit to the true local dust extinction, even at this spatial resolution.

• In our map of the estimated local visual extinction to the S88B region (see Fig. 8) we see a ~ 10' (7.2 pc) diameter central dense core region  overlapping and extending north of the S88B star formation region with  additional filamentary structures extending further north.

• Using a standard gas-to-dust ratio N(H + H₂) = 2 x10²¹ cm^-2 mag^-1 (Bohlin et al. 1978) we estimate that the mass of central 7.2 pc cloud core region to be ~ 5200 M_solar.

The mass of the dense core region compares with those of other high mass star formation regions:

S88B ~ 5200 M_solar

MonR2 ~ 4000 M_solar (Gonates et al. 1992)

K3-50A ~ 2600 M_solar (Howard  et al. 1997)

We detect Br Gamma line emission in the vicinity of the 2.2 um emission peak (Pipher et al. 1977, Evans et al. 1981).  The Br Gamma flux is (2.9 +/- 0.3) x 10^-15 W m^-2 in a 17" diameter aperture centered at RA₂₀₀₀ = 19^h 46^m 47.1^s and DEC₂₀₀₀ = 25 deg 12' 43".  Our value for the total Br Gamma flux is somewhat lower than that found in the literature; Evans et al. quote a corrected flux from Pipher et al. of (4.0 +/- 0.7) x 10^-15 W m^-2 in a similar aperture.  The circumstances necessitating this correction are unknown.  We are reasonably confident that our results are photometric and well calibrated.

Comparison of radio continuum emission with Br Gamma line emission provides an absolute measure of the line-of-sight extinction toward S88B.  Assuming the Brackett line emission to be optically thin, a 10% number abundance of He (all HeII), and A _V/A_{Br Gamma} = A_V/A_K= 9.29,  the extinction implied by the observed Br Gamma emission can be mapped.  The expected Br Gamma emission can be predicted from the 5 GHz radio continuum emission (graciously provided by G. Garay) by employing the following relation:

where S_nu is the radio flux density at frequency nu, and T_e is the electron temperature, which we take to be 7500 K (Herter et al.1981).  In comparing  the predicted Br Gamma flux with the observed flux, we can determine  A_BrGamma:

Using the relationship A_V= 9.29 x A_K from Mathis (1990), we can estimate the visual extinction.  It varies from ~ 30 to >60 mag from west to east.  In the regions devoid of Br Gamma emission, we place a lower limit on A_V of ~ 90 magnitudes.  Because of this high degree of visual extinction, we believe the optical nebulosity detected on the Palomar Observatory Sky  Survey E-plate (presumeably H alpha emission) is due to a small amount of ionized gas on the front side of the molecular cloud, and therefore, suffering from much less extinction.

From our J- (1.25 um), H- (1.65 um) and K-band (2.23 um) images, we selected the 32  point-like objects appearing in all images for stellar spectral type fitting, using two methods.  These objects were selected and their magnitudes measured using the IRAF/DAOPHOT package.   The uncertainty in our photometry is 10%, leading to an uncertainty of 0.1 mag in the J-, H-, and K-band magnitudes listed in Table 2.

In the first method, we assumed all stellar objects were located at the distance of S88B (2.5
+/- 0.5 kpc) and varied the line-of-sight visual extinction until a fit was found to one of eleven Main Sequence stars (using tau_lambda ~ lambda^alpha, where alpha ~ 1.4, as seen in other star forming regions (Howard, Pipher & Forrest
1994)).  The stellar spectral types considered were:  O6-8, B0, B3, B5, B8, A0, A5, F0, F5, G0, and K0.  The visual extinction deduced for each star is listed in Table 2 and in plotted on Figure 2.  The small amount of visual extinction, when compared with the visual extinction deduced from Br Gamma line emission ratios, implies that if these objects are at the distance of S88B, they are on the near side of the molecular cloud.

Our second method of fitting spectral type was to measure the line-of-sight visual extinction to each object
from a [J-H] vs. [H-K] color-color plot and compare the dereddened colors with those of Main Sequence stars (Koornneef 1983).  Assuming absolute magnitudes for the fitted spectral types (Lang 1992), the distance to each object was calculated.  The measured visual extinction, the spectral type  fit and the calculated distances are
listed in Table 2.  Two-thirds of the objects appear to be foreground stars.  This result is not surprising as by limiting our sample to objects with J-, H-, and K-band magnitudes, we eliminate highly reddened objects, including background stars.  Three  stars (14, 24 and 35) are definitely at the distance of S88B.   Star 24, located near the center of the region, is a B-type star experiencing A_V ~ 11 mag. This is considerably less extinction than estimated by
the comparison of predicted and observed Br Gamma emission. The close spatial proximity of Stars 24 and 25 to the presumed YSO (Star 7) and their massive nature  could indicate recent star formation and they are a probably cluster.  Thus we believe that the proximity and the nebulosity in which Stars 24 and 25 and the YSO are embedded may be contaminating the point-spread-function subtraction, leading to erroneous values of line-of-sight extinction and distance.

The above methods of determining spectral fit and extinction can be compared for stars at the distance of S88B.  Both methods give comparable results for stars 14, 24, 25 and 35 when the associated uncertainties are considered (distance, magnitudes of visual extinction and discrete bins of spectral type in the first method).

In the region where there is sufficient Br Gamma emission to give adequate signal-to-noise, the
amounts of extinction derived from the above methods can be contrasted with those determined from comparison of predicted and observed Br Gamma line emission.  In almost all cases, the extinction determined from the Br Gamma emission exceeds the stellar fitting estimations.  As mentioned earlier, we acknowledge a bias in our stellar selection criteria toward foreground objects.  The ionized gas emission detected in the infrared must originate deeper within the
molecular cloud.

A final check of our estimates of visual extinction is comparison of our values with the visual extinction derived from the silicate absorption feature.  Pipher et al. (1977) quote tau_9.7 = 3.6.  Assuming Rieke & Lebofsky's (1985) ratio A_V/tau_9.7= 16.6 +/- 1.2, we find A_V = 60 +/- 8 mag, consistent with the A_V estimated from the Br Gamma line emission.

Figure 1: (Return to Table of Contents)

Observed Br Gamma (2.166 um) emission map.  The green contours represent A_V as estimated from the comparison of predicted and observed Br Gamma line emission.  The white contour refers to regions of 5 sigma detection of Br Gamma emission.

Figure 2:  (Return to Table of Contents)

Figure 3:  (Return to Table of Contents)

140" by 120" composite image of S88B.  Red indicates K-band (2.23 um), green H-band (1.65 um) and blue J-band (1.25 um).  From west to east, the central nebulosity rapidly changes color as an increase in extinction allows the K-band emission to dominate.

Figure 5:  (Return to Table of Contents)

Uncertainty estimate (mag) as a function of brightness (mag) for all stars in one of the co-add frames. Ks is plotted in red, H in green and J in blue.

Figure 6:  (Return to Table of Contents)

 Cumulative luminosity function of all stars used in the extinction estimate. The log of the sum of all stars brighter than a given magnitude is plotted as a function of that magnitude in all three bands.

Figure 7:  (Return to Table of Contents)

 Weight plane image used for the A_V map. The image is  scaled from 0 (white) to 102 (black) stars and represents the number of  stars measured at each point to determine the dust extinction.  The regions of overlap between the 2MASS coadd frames are clearly visible due to multiple measurements of the same star in these regions.

Figure 8:  (Return to Table of Contents)

(Return to Table of Contents)