Citation
Papers from Across the Disciplines: The distribution of gas densities in the Milky Way

Material Information

Title:
Papers from Across the Disciplines: The distribution of gas densities in the Milky Way
Creator:
O'Dougherty, Stefan
Place of Publication:
Gainesville, Fla.
Publisher:
University of Florida
Publication Date:
Language:
English

Subjects

Genre:
serial ( sobekcm )

Notes

Abstract:
Observations of the atomic and molecular phases of the interstellar medium that coincide with the 20 degree by 6 degree region covered by the Census of High- and Medium-mass Protostars survey (CHaMP) are used to derive mass estimates and a probability distribution function of gas with density. The Southern Galactic Plane Survey (SGPS) and The Galactic All Sky Survey (GASS) were used for the HI, while the Nanten surveys of 12CO, 13CO, C18O and HCO+ were used as tracers for H2. Maximum and minimum mass estimates for each of the surveys were calculated due to the distance ambiguity that occurs within the solar radius. The derived probability distribution function is compared to simulations of a model galaxy. Maps of the distribution of the mass were generated as viewed from above the Milky Way for the 20 degree Galactic longitude range covered by CHaMP in each of the surveys for a better understanding of the distribution of GMCs and the structure of any spiral arms in the region.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.

Downloads

This item has the following downloads:


Full Text



The Distribution of Gas Densities in the Milky Way


Stefan O'Dougherty


College of Liberal Arts and Sciences, University of Florida


Observations of the atomic and molecular phases of the interstellar medium that coincide with the 20� by 6� region covered by the
Census of High- and Medium-mass Protostars survey (CHaMP) are used to derive mass estimates and a probability distribution
function of gas with density. The Southern Galactic Plane Survey (SGPS) and The Galactic All Sky Survey (GASS) were used for the
HI, while the Nanten surveys of 12CO, 13CO, C180 and HCO+ were used as tracers for H2. Maximum and minimum mass estimates for
each of the surveys were calculated due to the distance ambiguity that occurs within the solar radius. The derived probability
distribution function is compared to simulations of a model galaxy. Maps of the distribution of the mass were generated as viewed
from above the Milky Way for the 20� Galactic longitude range covered by CHaMP in each of the surveys for a better understanding
of the distribution of GMCs and the structure of any spiral arms in the region.


Introduction

From stars to the diffuse matter that lies between them,
the universe is composed primarily of hydrogen. Within
the confines of the local Milky Way, hydrogen exists in
three states: atomic (HI), molecular (H2), and ionized. The
HI accounts for roughly 75% of gas mass, while the
remainder is H2 and a small fraction of gas that is ionized.
The H2 gas exists as structures known as Giant Molecular
Clouds (GMCs), which are the sites of star formation. By
studying these clouds, insight into the complex processes
that are involved with star formation can be gained. For
example, radio surveys of the HI 21 cm line and H2
molecular tracers can be used to study the distribution of
the mass among various density domains within these
GMCs. This distribution is an important constraint when
creating theoretical models of GMC formation in spiral
galaxies like our own Milky Way.
A section of the Galaxy covering parts of Vela, Carina,
and Centaurus was chosen for the purposes of this mass
analysis since the area has been covered by some of the
latest and highest resolution surveys dedicated to map
GMCs on a large scale. The Census of High- and Medium-
mass Protostars (CHaMP; Barnes et al. 2011) has a
coverage of 20� by 6� over this southern region of the
Milky Way. CHaMP surveys the region using several
tracers of different densities in a more complete and
unbiased way than previous surveys to further enhance
knowledge of the physical processes involved with star
formation.
This survey can be complemented by investigating the
amount of mass for the atomic and molecular gas in the
region. Analysis of the HI and H2 molecular tracers that
cover different densities can lead to the computation of a
probability distribution function of gas with density (gas
density PDF) that gives the mass fractions for each
molecular tracer as a function of density. Theoretical
models of the evolution of the Milky Way's interstellar


medium (Tasker & Tan 2009; Tasker 2011) with their own
gas density PDF can then be used as a comparison with the
observational data to evaluate theories of formation and
properties of the GMCs.
The CHaMP region is between Galactic longitudes 280�
to 3000 and Galactic latitudes -40 to 20 and covers the
frequency ranges of 85 to 93 GHz and 107 to 115 GHz.
CHaMP uses the Australia Telescope National Facility's
22 meter diameter Mopra antenna and the MOPS digital
filterbank to acquire high resolution maps, with a beam
size of 40", on numerous spectral lines in the region. The
CHaMP survey has identified 209 clumps (Barnes et al.
2011) based on the Nanten surveys conducted in the same
area (Yonekura et al. 2005; Fukui et al. 2008). The Nanten
survey was conducted using a recursive mapping technique
that covered the Hz molecular tracers of 12CO, 13CO, C180
and HCO+. Each tracer probes denser components of
molecular gas, thus the 12CO was used as a finder chart for
mapping out the 13CO, then the latter was used as a basis
for mapping out the C180. This technique allows for
unbiased sampling of the entire region, since the majority
of all active star formation is traceable from C180
emission. For this research, the mass properties of the
entire region from the Nanten survey will be combined
with the results from the detailed core study. In contrast to
CHaMP, Nanten has a lower angular resolution of 3.3' and
a velocity resolution of 0.1 km s-'; however, the cubes used
in this analysis had lower velocity resolutions of 0.5 and 1
km s-1.
In order to calculate the mass of HI in the region two
surveys were selected: the Southern Galactic Plane Survey
(SGPS; McClure-Griffiths et al. 2005) and the Galactic All
Sky Survey (GASS; McClure-Griffiths et al. 2009). The
SGPS has a higher angular resolution of 2.2'; however, it
only covers Ibl < 1.50. To complete the full coverage of the
CHaMP region the missing data is taken from the GASS
survey at an angular resolution of 16' for the latitudes not


University of Florida I Journal of Undergraduate Research I Volume 12, Issue 2 I Spring 2011





STEFAN O'DOUGHERTY


covered by the SGPS. The velocity resolution of both is
equivalent at 0.82 km s-'.

Techniques

Data from the SGPS, GASS, and Nanten surveys were
used to calculated masses for the HI and the tracers of H2.
An initial assumption of a flat rotation curve is made in
determining the distances of the signal that adopts the
values Ro = 8.4 kpc and 00 = 254 km s-' (Reid et al. 2009).
One significant problem that affects the calculation of the
mass is the distance ambiguity, brought about by our


location within the disk of the Milky Way. For any VLSR
that corresponds to R < R0, a distance ambiguity arises. It
is impossible to determine whether the mass being
calculated is located at the near or the far distance based on
the VLSR alone. In an attempt to remedy this, for one
calculation it is assumed the distance is near every time R
< Ro, and for the second calculation the far distance is
applied. As Table 1 shows, an upper and lower limit then
exists for each tracer, the average of which is used for the
gas density PDF.


Table 1: Mass Totals for Each Survey in Units of 106 M�.

Interior Near Interior Far Interior Average Outer Total

HI Total 10.53 34.57 22.58 230.42 253.00

HI SGPS 5.40 18.48 11.94 110.75 122.69

HI GASS 5.05 17.61 11.36 106.43 117.79
bl < 1.5�
12CO Nanten 2.95 6.16 4.56 7.85 12.41

13CO Nanten 1.27 2.62 1.94 2.52 4.46

C18O Nanten 0.56 1.13 0.84 0.46 1.30

HCO Nanten 0.122 0.129 0.126 0.145 0.271

Note. Results are divided into interior and outer totals to illustrate the effect of the distance ambiguity. Column 5 lists the total mass values using
the interior average and the outer mass.


The initial mass calculation starts with the determination
of the column density for each survey. In general form the
column density along each line of sight is calculated via


Na =Ya T dv


and NH2 = Xa Na


where a is specific to the particular tracer, Ya is the scaling
factor from the observed intensity to column density of that
tracer, while Xa scales directly to the H column density
(whether atomic or molecular), and T is the brightness
temperature measured. Assuming optically thin conditions,
we use YH = 1.8 x 1018 cm-2 K-1 km-1 s for the HI (Spitzer
1978) which is calculated independent of temperature. For
the '2CO we adopt X,2c = 1.8 x 1020 cm-2 K'-1 km-' s
(Dame et al. 2001). In order to convert the '3CO we use
NH2 = 1.0 x 104 N,1 ., and N,:.., = 70 N13co with Y13c =
8.76 x 1014 cm-2 K'-1 km-1 s. The C'80 uses N,-..,= 560
Nc18o with Yc18o = 1.16 x 1015 cm-2 K-1 km'1 s.
The HCO (J= 1-0) column density is calculated via


N(HCO) = 3h O(Ti exp(E,,/k T -c dV (3)
87i3 L2DJu (1-exp(hv/kTex))

where Q is the partition function calculated at a Tex of 10
K, and Ju , Eu are calculated for the upper level, J=1. The
optical depth T is then calculated at each point using


t = -ln( - Tmb /(Tex- Tbg))


where Tmb is the main beam temperature from the data
cube, Tex is 10 K and Tbg is 2.73 K.
Along each line of sight the column densities are
converted into a mass surface density and summed over the
entire region for a final mass result with the general
equation


Mtotal=Z (mHgipANa)


where Na is the column density, mH is the mass of
hydrogen, .ip = 1.4 or 2.8 is the mean mass per particle for
atomic or molecular gas respectively and accounts for the


University of Florida I Journal of Undergraduate Research I Volume 12, Issue 2 I Spring 2011





THE DISTRIBUTION OF GAS DENSITIES IN THE MILKY WAY


helium abundance in the gas and A is equal to the pixel's
projected area times the distance squared.
Besides having the Nanten HCO+ data, the CHaMP
HCO+ data is available for the same region (Barnes et al.
2011, hereafter Paper I). As a comparison between the two


data sets, several additional methods of calculating the
mass for the CHaMP sources and the whole region were
conducted. The mass of HCO+ in Table 1 is from the
Nanten survey and has been calculated as described above.
The following comparisons are summarized in Table 2.


Table 2: Comparison of Nanten and HCO+ Mass Calculated through Various Methods.

Source & Method Mass (106 Mo)
CHaMP Sources 2D Gaussian Method 0.219
CHaMP Sources Summed Intensity Method 0.263
Entire Mopra Cubes Method 0.268
Entire Nanten Cubes Method 0.541
Nanten Limited to CHaMP Region Method 0.336
Note. Refer to Section II for more information on the specific differences between each method.


Paper I has defined 301 clumps, from 121 Nanten
clumps, and their respective masses through a two
dimensional Gaussian fit model. Their locations, masses,
and other properties are all listed in Paper I. As an initial
comparison to this Gaussian fit model, a script was written
to calculate the mass of each of these clumps that summed
up the entire signal, as opposed to fitting a Gaussian, in a
defined ellipse. Paper I has quoted half power ellipse sizes
and in order to pick up the majority of the signal, an ellipse
a factor of two times the quoted size was taken and all the
mass within summed. To have the same basis for
comparison, the same velocity range, based on a 3cy
velocity range from the peak VLSR of the source, is applied
as a constraint. The majority of clumps have associated
sources with previously known distances that help to
resolve the distance ambiguity, therefore the quoted
distances from Paper I were applied instead of calculating a
kinematic distance to each clump.
Another method of comparison of the Gaussian mass
method included summing up all emission in the HCO+
Paper I dataset, not simply limiting it to the defined source
ellipses. Velocity and distance constraints unique to each
region were applied. In an effort to minimize noise caused
by edge pixels, noisier regions and noisier strips within
regions, only signal with intensity 3cy above the mean noise
as calculated over a line free region for each position was
counted.
The HCO+ of Nanten covers an area larger than that of
Mopra, but at a poorer resolution. In order to compare the
difference in the surveys, the calculation of the mass of the
Nanten data was split during calculation in longitude space
to match corresponding CHaMP regions. The velocities
from the CHaMP sources that overlapped within these
longitude slices were separated and had their mass
calculated using the same distances as the defined in Paper
I. This allowed for a total mass of just the overlapping
regions to be calculated, in addition to limiting the amount
of distance ambiguity in the Nanten HCO+ mass total.


As a supplement to calculating the mass of the HI and
H2, information on the distances and the corresponding
longitudinal pixels were saved for each mass run, which
allowed for overhead maps of the gas distributions to be
plotted. They have been split into 0.5 x 0.5 kpc bins along
each longitude point. The corresponding longitude and
distance were then plotted onto a polar coordinate system
with a contour overlay of the mass density in terms of M�
pc-2 to illustrate the structure of the GMCs and the arms in
which they are found.

Results

The initial mass results are shown in Table 1 in units of
106 Mo. Listed for the HI is the amount of mass in the
SGPS survey, the amount in the higher latitudes of the
GASS survey, and the total amount of HI across the full
range. Each tracer has a calculation for the amount of mass
within the solar radius using the near distance, the far
distance, and an average of the two. Column 4 gives the
total mass outside the solar radius, for which there is only
one solution for the kinematic distance. Column 5 gives the
total mass based on the average of the near & far distances
and the total outside mass.
Comparing the average total of the SGPS and the GASS
with Ibl < 1.5� yields a difference of -5 x 106 M�, which
corresponds to a -4% difference between the mass totals.
This a reasonably small and acceptable difference between
the two surveys considering the difference in resolution
(2.2' vs 16').
For the gas density PDF, only mass that falls within a
distance of 8 kpc is taken into account. As illustrated in the
overhead maps in Figure 3, the majority of the mass across
all tracers falls within 8 kpc. For the molecular tracers,
there is no mass detected by the surveys above their
sensitivity levels past 10 kpc. Furthermore, for the weaker


University of Florida I Journal of Undergraduate Research I Volume 12, Issue 2 I Spring 2011





STEFAN O'DOUGHERTY


emission of CO8 and HCO+, the noise tends to dominate
over any actual emission in the outer regions, leading to
slightly negative results past 8 kpc. Limiting the radius to 8
kpc thus helps limit the weaker emission and is justifiable
as a reasonable cutoff point for comparisons between the
HI and the H2.
With the above limit imposed, the final gas density PDF
is calculated and can be seen in Figure 1. The mass


1





x
-o


cn
0













0.001
0.001


fractions are plotted over a range of densities thought to be
associated with their respective tracer. The Mopra data
does have a calculated weighted mean density; however,
the other ranges are our best estimates. We also include
error bars for each tracer, taking into account the
ambiguities associated with the distance, a flat rotation
curve, column density calculations, and conversion factor
uncertainties.


Probability Distribution Function of Gas with Density



HI..



_' . . . .... .... .1 i\ ".






'HCO
I
HHCO CO


HCO +

HCO'


1 10 100 1000 104 11
nH (cm-3)


Figure 1: Gas Density PDF for the HI & H2 in the CHaMP region. Plotted black squares represent the fraction per dex for each tracer at the center of
their estimated density range. Horizontal bars indicate the range of densities each covers, and the vertical bars span the uncertainty range we have
calculated for each. The dotted line represents a model gas density PDF with no FUV heating feedback from a simulation from Tasker & Tan (2009).
The dashed line is a model run by Tasker (2011) that includes the FUV heating.


The molecular tracers in general have all their mass
located within 8 kpc, so a larger percentage of their mass
falls within the solar radius than for the atomic mass.
Therefore, for the H2 tracers the distance ambiguity
uncertainty ranges from 13% to 18%;f however, the HI
uncertainty is only 5%. To estimate the uncertainty for
assuming a flat rotation curve, the velocity used to
calculate a distance at a specific longitude was varied by
+/-10 km s-' to see the resulting change in mass. They all
vary between 2-5%, which was concluded to be a fairly
small and acceptable uncertainty for the assumption of a
flat rotation curve.
The uncertainty associated with calculation of the
column density is dominated by the conversion values
chosen. For 12CO the widely accepted Wco value (Dame et
al. 2001) is used, with a 10% uncertainty to account for any
variations in the uniformity of the tracer in the region. The
abundance ratios and X values for "3CO, C80, and HCO


are much more uncertain and can range greatly based on
various assumptions. This has been taken into account and
the uncertainties have been adjusted accordingly. Since the
HCO+ is the most uncertain, a factor of 3 error for the
abundance ratio (Barnes et al. 2011) is applied. The "3CO
X-factor has a wide degree of variance dependent on local
properties and the amount of star formation in the region
(Pineda et al. 2008). A wide range of X-factors have been
derived, and in comparison to this study the following are
applied: a factor 2 uncertainty for the "3CO and a factor 2.5
uncertainty for the C80 (Pineda et al. 2008, Glassgold et
al. 1985).
One other source for uncertainty in calculating the
column density is the error introduced by integration of the
brightness temperature over the velocity range. For the
Nanten data the sensitivity appears to be deep enough to
get a very low rms per channel. As a result the integration
errors were very close to zero. The HI shows the same


University of Florida I Journal of Undergraduate Research I Volume 12, Issue 2 1 Spring 2011





THE DISTRIBUTION OF GAS DENSITIES IN THE MILKY WAY


negligible errors. The combined errors are plotted as the
vertical axis bars for the observed mass fraction values in
Figure 1. The HI and the 12CO have the least amount of
error, mostly due to the minimization of uncertainty in
their calculations and conversions. The remaining
molecular tracer errors exhibit the level of uncertainty that
exists. The mass fractions are plotted against a simulated
galaxy's mass fractions (Tasker & Tan 2009; Tasker 2011)
as a rough comparison. The simulation was modeled to
create a galaxy similar to the Milky Way that can be used
to study GMC evolution and is plotted against the
observed gas density PDF in Figure 1. The dotted line
shows a PDF from the simulation of Tasker & Tan (2009)
that did not include FUV heating, while the dashed line
includes those effects. The model with no FUV heating
assumes a flat rotation curve and does not include star
formation and feedback, but instead it focuses on
gravitational instabilities and cloud collision effects.
With the inclusion of the heating, the model estimates an
overabundance of mass over the 12CO and "3CO density
range. The higher density observations covered by CO80
and HCO+ do drop off sharply and appear to match the
drop of the model near the Paper I sources' mass weighted
density average of 4.0x 103 cm-3.
The gas density PDF also includes points from the
comparison analysis between the Mopra and Nanten data.
As shown in Table 2, the total mass of the CHaMP sources
from Paper I as calculated through a two dimensional
Gaussian fit is 0.219 x 106 Mo and is plotted as the point
HCO+c in Figure 1. Calculating the summed intensity in an
ellipse twice the size of the half power ellipses defined
from the Paper I yields a total mass of 0.263 x 106 Mo. Yet
this value may be double counting mass if nearby ellipses
are overlapping. While the majority of sources are mostly


isolated, there are certain regions that have numerous
sources overlapping. A future fix to this could be to
develop a routine that determines overlapping pixels in
three dimensional space and then assigns the mass in that
pixel to the closer source. The expected result would lead
to a mass that would approach the two dimensional
Gaussian total mass. The HCO,+ point in Figure 1 shows
the total mass calculated from the Nanten cubes, 0.541 x
106 Mo. This used the method of taking overlapping
regions from the CHaMP HCO+ survey and setting them to
the equivalent distances of the sources in the corresponding
region in order to minimize distance ambiguities. A
separate point, HCO+b shows only the mass from the
Nanten survey that overlaps with the regions in longitude
space covered by Paper I. This value of the Nanten mass,
0.336 x 106 Mo, differs from the Paper I full region total
by -7 x 105 Mo. The beam size of Nanten compared to that
of Mopra would lead to some undersampling, but as Paper
I discusses, barely 2% of the CHaMP sources could be
considered point like. Therefore we conclude that the
majority of the extra mass seen is most likely from low
level emission in widespread low density clumps.
The final part of this research focuses on visualizing the
results. When calculating the mass totals for each survey,
the distance was divided up into 0.5 by 0.5 kpc boxes along
each longitude in order to create overhead maps of the
mass surface density. Figure 2 shows the latest artistic
representation of the Milky Way as viewed from above
(NASA/JPL Caltech/R. Hurt (SSC-Caltech)), with an
outline of the CHaMP region. On the right is the first of the
overhead maps, from the HI GASS survey. For simplicity,
all the maps assume the near distance ambiguity. The
Carina Arm is quite visible in the map as the regions that
have the greatest mass surface density trace out a distinct


HI GASS-Neor


0 5 10 15 20
X (kpc)


S2W(? 69W9
(M pc)


s2wo 11SoW 13


Figure 2: On the left, an artist's conception of the Milky Way as viewed from above. The red lines outline the CHaMP region in longitude, from 280�to
300�.
Source: NASA/JPL Caltech/R. Hurt (SSC-Caltech). On the right, a mass surface density map of the region as viewed from above based on the HI
GASS survey covering a distance out to 20 kpc and assuming near distances for the distance ambiguity.
University of Florida I Journal of Undergraduate Research I Volume 12, Issue 2 I Spring 2011





THE DISTRIBUTION OF GAS DENSITIES IN THE MILKY WAY


13CO


2 4 6 a 10
X (kpc)


HCO+


4 6 8 10
X (kpc)


0 2 4 6
M. pc-2


4 6 8 10
X (kpc)


8 10 12


Figure 3: Maps for HI and the H2 molecular tracers of the mass surface density over the CHaMP region (2800 to 3000) out to 10 kpc. The red line
represents the radius cutoff of 8 kpc imposed in creating the gas density PDF.


spiral arm pattern. Figure 3 displays the GASS HI and all
the Nanten H2 tracers, out to a distance of 10 kpc.

Conclusions

The atomic and molecular hydrogen was studied in a 200
by 60 region of the Milky Way in order to derive a gas
density PDF. This region coincides with the area covered
by the CHaMP survey and will allow for further insight
into the study of star formation and GMCs. The results
confirm the expectation that the majority of gas in the local
region is mostly atomic hydrogen. The mass estimates
required addressing the distance ambiguity brought about
by any sources that fell within the solar radius. Basing the
distance calculation only on the velocity of the source, two
mass calculations were run to get an upper and lower limit.
These were averaged and combined with the mass outside
the solar radius for the final mass totals.
In constructing the gas density PDF, uncertainties from
the ambiguity as well as uncertainties due to conversion
factors of the tracers to H2 and any Y & X factors used in
the column density calculations must be taken into account.
The HI and 12CO have the least amount of uncertainty
compared to the other tracers, which is key since they
account for a large percentage of the mass for the region. It
should be noted that the mass results for the "3CO are
slightly lower than expected and are in fact a lower limit
since they are based on the assumption of being optically
thin. Also, the HCO+ conversion factor to H2 is uncertain
and as a result the mass fraction varies by a factor of three.


The observed gas density PDF was also compared to
simulated gas density PDFs of a galaxy with and without
FUV heating feedback. There is a general agreement with
the FUV heating model, though the model has too much
mass in the density ranges covered by 12CO and 13CO.
However, this model has a steeper drop off at the C180 and
HCO+ densities that roughly corresponds with the observed
values.
Comparison of the CHaMP source mass total to the
complete total covered by the Mopra and Nanten cubes
leads to the conclusion that there is fair amount of low
density gas at fainter emissions that is contributing to the
gas mass.
Further work will be done to investigate and reduce the
uncertainty in the density ranges for each of the molecular
tracers in the gas density PDF. Future galaxy simulations
could lead to a closer correlation between observed and
modeled PDFs. Additional research can build from this
project to solve the distance ambiguities that arise. By
using the defined clumps from the CHaMP region, we can
study the HI data from the SGPS survey to possibly
eliminate the distance ambiguity by checking for HI self-
absorption features (Goldsmith & Li 2004). These
absorption features are the result of a cold HI cloud that is
in front of a warmer HI cloud (Anderson & Bania 2009).
Several studies have shown that molecular clouds generate
these HI self absorption features, and it is generally agreed
that evidence of these features at the same velocity and
velocity width as the molecular tracers means the cloud is
at the near distance (Anderson & Bania 2009).


University of Florida I Journal of Undergraduate Research I Volume 12, Issue 2 I Spring 2011


GASS


2 4


12CO


OB55N-- 1
2 4 6 8 1B
X (kpc)


X (kpc)


C180


rrr :sll





THE DISTRIBUTION OF GAS DENSITIES IN THE MILKY WAY

References McClure-Griffiths N. M., Pisano, D. J., Calabretta, M. R., Ford, H. A., Lockman,
F. J., Staveley-Smith, L., Kalberla, P., Bailin, J., Dedes, L., Janowiecki, S.,
Gibson, B. K., Murphy, T., Nakanishi, H., Newton-McGee, K. 2009, ApJS,
Anderson, L. D., Bania, T. M., 2009, ApJ, 690, 706 181, 398

Barnes, P. J., Yonekura, Y., Ryder, S.D., Hopkins, A. M., Miyamoto, Y., Reid, M. J., Menten, K.. M., Zheng, X. W., Brunthaler, A., Moscadelli, L., Xu,
Furukawa, N., Fukui, Y., 2010, MNRAS, 402, 73 Y., Zhang, B., Sato, M., Honma, M., Hirota, T., Hachisuka, K., Choi, Y. K.,
Moellenbrock, G. A., Bartkiewicz, A., 2009, ApJ, 700, 137
Barnes, P. J., Yonekura, Y., Fukui, Y., Miller, A. T., Miihlegger, M., Agars, L.
C., Miyamoto, Y., Furukawa, N., Papadopoulos, G., Jones, S. L., Hernandez, Rohlfs K. & Wilson T.L. 2006, Tools of Radio Astronomy, 4th ed. (Springer:
A. K., O'Dougherty, S. N. Tan, J. C., 2011, ApJ, submitted Berlin)

Dame, T. M., Hartmann, D., Thaddeus, P., 2001, ApJ, 547,792 Pineda, J. E., Caselli, P., Goodman, A. A., 2008, ApJ, 679, 481

Fukui, Y., Kawamura, A., Minamidani, T., Mizuno, Y., Kanai, Y., Mizuno, N., Spitzer, L. Jr., 1978, Physical Processes in the Interstellar Medium, (John Wiley
Onishi, T., Yonekura, Y., Mizuno, A., Ogawa, H., Rubio, M., 2008, ApJS, & Sons: New York)
178, 56
Tasker, E., Tan, J., 2009, ApJ, 700, 358
Goldsmith, P.F., Li, D., 2004, ApJ, 622, 938
Tasker, E. 2011, ApJ, 730, 11
McClure-Griffiths N. M., Dickey, J.M. Gaensler, B. M., Green, A. J., Haverkorn,
M., Strasser, S. 2005, ApJS, 158, 178 Yonekura, Y. Asayama, S., Kimura, K., Ogawa, H., Kanai, Y., Yamaguchi, N.,
Barnes, P.J., Fukui, Y., 2005, ApJ, 634, 476


University of Florida I Journal of Undergraduate Research I Volume 12, Issue 2 I Spring 2011
7