a dissertation entitled lighting the dark molecular gas and a bok globule by aditya g. togi
Post on 11-Sep-2021
3 Views
Preview:
TRANSCRIPT
A Dissertation
entitled
Lighting the dark molecular gas and a Bok globule
by
Aditya G. Togi
Submitted to the Graduate Faculty as partial fulfillment of the requirements for the
Doctor of Philosophy Degree in Physics
Dr. John-David T. Smith, Committee Chair
Dr. Adolf Witt, Committee Member
Dr. Lee Armus, Committee Member
Dr. Rupali Chandar, Committee Member
Dr. Sanjay Khare, Committee Member
Dr. Patricia R. Komuniecki, DeanCollege of Graduate Studies
The University of Toledo
May 2016
Copyright 2016, Aditya G. Togi
This document is copyrighted material. Under copyright law, no parts of thisdocument may be reproduced without the expressed permission of the author.
An Abstract of
Lighting the dark molecular gas and a Bok globule
by
Aditya G. Togi
Submitted to the Graduate Faculty as partial fulfillment of the requirements for theDoctor of Philosophy Degree in Physics
The University of ToledoMay 2016
Stars are the building blocks of galaxies. The gas present in galaxies is the pri-
mary fuel for star formation. Galaxy evolution depends on the amount of gas present
in the interstellar medium (ISM). Stars are born mainly from molecular gas in the
GMCs. Robust knowledge of the molecular hydrogen (H2) gas distribution is neces-
sary to understand star formation in galaxies. Since H2 is not readily observable in
the cold interstellar medium (ISM), the molecular gas content has traditionally been
inferred using indirect tracers like carbon-monoxide (CO), dust emission, gamma ray
interactions, and star formation efficiency. Physical processes resulting in enhance-
ment and reduction of these indirect tracers can result in misleading estimates of
molecular gas masses. My dissertation work is based on devising a new temperature
power law distribution model for H2, a direct tracer, to calculate the total molecular
gas mass in galaxies. The model parameters are estimated using mid infrared (MIR)
H2 rotational line fluxes obtained from IRS-Spitzer (Infrared Spectrograph- Spitzer)
instrument and the model is extrapolated to a suitable lower temperature to recover
the total molecular gas mass. The power law model is able to recover the dark molec-
ular gas, undetected by CO, in galaxies at metallicity as low as one-tenth of our Milky
Way value. I have applied the power law model in U/LIRGs and shocks of Stephan’s
Quintet to understand molecular gas properties, where shocks play an important role
in exciting H2. Comparing the molecular gas content derived through our power law
iii
model can be useful in studying the application of our model in mergers. The pa-
rameters derived by our model is useful in understanding variation in molecular gas
properties in shock regions of Stephan’s Quintet.
Low mass stars are formed in small isolated dense cores known as Bok globules.
Multiple star formation events are seen in a Bok globule. In my thesis I also studied
a Bok globule, B207, and determined the physical properties and future evolutionary
stage of the cloud. My thesis spans studying ISM properties in galaxies from kpc to
sub-pc scales. Using the power law model in the coming era of James Webb Space
Telescope (JWST) with the high sensitivity MIR Instrument (MIRI) spectrograph we
will be able to understand the properties of molecular gas at low and high redshifts.
iv
This work is dedicated to my determination to pursue astronomy from past 25 years
Acknowledgments
There are number of people that deserve special thanks as they have played a spe-
cial role in successful completion of this thesis. Foremost, I would like to thank my
advisor, Dr. J. D. Smith for offering me to join his group and continue in pursuing
my research on galaxy ISM. His knowledge, insight, patience, encouragement, and
desire to help have been valuable stimulation for my scientific development. I want to
express my gratitude to Dr. Adolf Witt who has imparted the scientific wisdom and
amazed me with his scientific working style. I am very much thankful for their advise
on writing papers and my scientific afternoon talks with Dr. Witt on astronomy to
politics to philosophy will long be cherished. I want to express my thanks to Dr.
Lee Armus, who has been an inexhaustible source of scientific and technical wisdom
and is a wonderful person to work with. I am thankful to him for all the support he
provided during my IPAC Visiting Graduate Student Fellowship program at Caltech.
I am thankful to my other committee members Dr. Rupali Chandar and Dr. Sanjay
Khare for accepting to be in my Ph.D. committee.
I am grateful to to Dr. Rick Irving for all the help and took out time in resolving
my computer issues. Finally, I would like to thank all graduate students at the Uni-
versity of Toledo. Whether talking fundamental physics, helping with code, or editing
my writing, their help was fundamental to both my work and especially tolerating
my insanity throughout.
vi
I have always cherished the company of my friends Amogh, Hemant, Nikhil, Prab-
hakar, Rajesh, Santosh Kumar, Shaiju, Sushilkumar, Venkatesh, Vidhi Mishra, Vish-
wanath, and an endless list of people who always heard and supported me during
times of distress. Finally, I would like to thank my family members for their constant
support and motivation throughout my academic journey.
Thank you all for having been with me in this wonderful journey.
vii
Contents
Abstract iii
Acknowledgments vi
Contents viii
List of Tables xiii
List of Figures xv
List of Abbreviations xviii
1 The Interstellar Medium (ISM) 1
1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Constituents of the ISM . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2.1 The ISM Gas . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 Evolution of ISM dust . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3.1 Dust composition and size distribution . . . . . . . . . . . . . 6
1.3.2 Dust processing in the ISM . . . . . . . . . . . . . . . . . . . 8
1.4 The cosmic life cycle . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.5 ISM molecular gas . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.5.1 Importance of H2 . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.5.2 Detection techniques of molecular gas . . . . . . . . . . . . . . 11
1.5.3 Dark Molecular Gas . . . . . . . . . . . . . . . . . . . . . . . 13
viii
1.6 Thesis outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2 Lighting the dark molecular gas: H2 as a direct tracer 17
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.2 Sample . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.2.1 MIR H2 rotational line fluxes . . . . . . . . . . . . . . . . . . 21
2.2.2 Cold molecular gas mass from CO line intensities . . . . . . . 21
2.3 Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.4 Method & Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . 28
2.4.1 Warm H2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
2.4.1.1 Upper temperature, Tu . . . . . . . . . . . . . . . . 29
2.4.1.2 Power law index, n . . . . . . . . . . . . . . . . . . . 29
2.4.2 Total H2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.4.2.1 Model Sensitivity Temperature, Ts . . . . . . . . . . 32
2.4.2.2 Model extrapolated lower temperature, Tℓ . . . . . . 38
2.4.2.3 Mass distribution function . . . . . . . . . . . . . . . 40
2.5 Results, Discussions, & Applications . . . . . . . . . . . . . . . . . . 41
2.5.1 What are the typical molecular gas temperatures in galaxies? 41
2.5.2 Estimating total MH2 . . . . . . . . . . . . . . . . . . . . . . . 43
2.5.3 Model derived molecular gas mass in ULIRGs, LIRGs and radio
galaxies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
2.5.4 Effect of dust temperature on the warm H2 fraction . . . . . . 47
2.5.5 Molecular gas in low metallicity galaxies . . . . . . . . . . . . 51
2.5.5.1 Metallicity estimation . . . . . . . . . . . . . . . . . 52
2.5.5.2 Cold molecular gas from CO line emission . . . . . . 52
2.5.5.3 Molecular gas from dust emission . . . . . . . . . . . 53
2.5.5.4 Molecular gas mass using our model . . . . . . . . . 53
ix
2.5.6 Future prospects . . . . . . . . . . . . . . . . . . . . . . . . . 57
2.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
3 Molecular gas properties in ISM of U/LIRGs of GOALS 75
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
3.2 Sample . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
3.2.0.1 Ancillary data CO(J=1–0) . . . . . . . . . . . . . . . 77
3.2.1 Data reduction and Spectral fitting . . . . . . . . . . . . . . . 78
3.3 Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
3.3.1 Disk template . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
3.3.2 Scaling and adding the template spectrum . . . . . . . . . . . 80
3.3.3 PAHFIT- to recover H2 line flux . . . . . . . . . . . . . . . . . 82
3.3.4 H2 gas mass from power law model . . . . . . . . . . . . . . . 82
3.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
3.4.1 LIRG’s disk template . . . . . . . . . . . . . . . . . . . . . . . 84
3.4.2 Power law index . . . . . . . . . . . . . . . . . . . . . . . . . . 85
3.4.3 H2 mass- extrapolating power law model to 49 K . . . . . . . 86
3.5 Discussions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
3.5.1 Low power law index in U/LIRGs . . . . . . . . . . . . . . . . 87
3.5.2 Relation between LIR, PAHs and power law index . . . . . . . 88
3.5.3 Low αCO or high temperature . . . . . . . . . . . . . . . . . . 92
3.5.4 Tℓ from gas-to-dust mass ratio GDR . . . . . . . . . . . . . . 92
3.5.5 Variation in the L850
MISM. . . . . . . . . . . . . . . . . . . . . . 95
3.5.6 Caveats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
3.6 Summary & Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . 98
4 Molecular gas properties in shocks of Stephan’s Quintet 106
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
x
4.2 H2 in shocks of Stephan’s Quintet . . . . . . . . . . . . . . . . . . . . 108
4.3 Observations, Data reduction and Analysis . . . . . . . . . . . . . . . 109
4.3.1 MIR H2 emission and spectra . . . . . . . . . . . . . . . . . . 110
4.4 Excitation diagram Fits . . . . . . . . . . . . . . . . . . . . . . . . . 111
4.4.1 Discrete temperature fit . . . . . . . . . . . . . . . . . . . . . 111
4.4.2 Power law model fit . . . . . . . . . . . . . . . . . . . . . . . . 129
4.5 Result . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
4.5.1 Two temperature fits . . . . . . . . . . . . . . . . . . . . . . . 130
4.5.2 H2 line ratios . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
4.5.3 Power law index . . . . . . . . . . . . . . . . . . . . . . . . . . 133
4.5.4 Total and Warm H2 gas mass . . . . . . . . . . . . . . . . . . 137
4.5.4.1 Total H2 gas mass . . . . . . . . . . . . . . . . . . . 137
4.5.4.2 Warm H2 fraction . . . . . . . . . . . . . . . . . . . 140
4.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
5 Physical properties of the cometary globule: the case of B207 143
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
5.2 Observations & Archival Data . . . . . . . . . . . . . . . . . . . . . . 144
5.2.1 Discovery Channel Telescope (DCT) observations . . . . . . . 145
5.2.2 Archival WISE and Herschel data . . . . . . . . . . . . . . . . 147
5.3 Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
5.3.1 Surface brightness profiles . . . . . . . . . . . . . . . . . . . . 148
5.3.2 Dust characteristics . . . . . . . . . . . . . . . . . . . . . . . . 152
5.3.2.1 Scattered light intensities . . . . . . . . . . . . . . . 152
5.3.2.2 Coreshine effect in NIR band . . . . . . . . . . . . . 155
5.3.3 Temperature Analysis using SED fitting . . . . . . . . . . . . 157
5.3.4 Mass Distribution . . . . . . . . . . . . . . . . . . . . . . . . . 160
xi
5.3.4.1 Star count technique . . . . . . . . . . . . . . . . . . 160
5.3.4.2 Reddening of background stars . . . . . . . . . . . . 162
5.3.4.3 Far infrared (FIR) emission . . . . . . . . . . . . . . 163
5.3.5 Temperature and density-Abel Inversion method . . . . . . . . 164
5.3.6 Core energetics . . . . . . . . . . . . . . . . . . . . . . . . . . 169
5.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
5.4.1 Surface brightness in core and rim . . . . . . . . . . . . . . . . 173
5.4.2 Grain growth and age determination . . . . . . . . . . . . . . 173
5.4.3 Core physical properties . . . . . . . . . . . . . . . . . . . . . 174
5.4.4 Core collapse . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
5.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
6 Summary and Future Work 178
6.1 What did we learn? . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178
6.1.1 H2 power law model . . . . . . . . . . . . . . . . . . . . . . . 178
6.1.2 Molecular gas properties in GOALS galaxies . . . . . . . . . . 179
6.1.3 Stephan’s Quintet work . . . . . . . . . . . . . . . . . . . . . 180
6.1.4 Study of Bok globule, B207 . . . . . . . . . . . . . . . . . . . 181
6.2 Future work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
6.2.1 Study of molecular gas at high z . . . . . . . . . . . . . . . . . 181
6.2.2 Study on Bok globules . . . . . . . . . . . . . . . . . . . . . . 183
References 188
xii
List of Tables
1.1 Description of the phase components of the ISMa. . . . . . . . . . . . . . 5
2.1 Observed properties of our sample galaxies . . . . . . . . . . . . . . . . 60
2.1 Observed properties of our sample galaxies . . . . . . . . . . . . . . . . 61
2.1 Observed properties of our sample galaxies . . . . . . . . . . . . . . . . 62
2.2 Observed molecular hydrogen rotational line flux . . . . . . . . . . . . . 63
2.2 Observed molecular hydrogen rotational line flux . . . . . . . . . . . . . 64
2.2 Observed molecular hydrogen rotational line flux . . . . . . . . . . . . . 65
2.2 Observed molecular hydrogen rotational line flux . . . . . . . . . . . . . 66
2.2 Observed molecular hydrogen rotational line flux . . . . . . . . . . . . . 68
2.3 Observed molecular hydrogen lines . . . . . . . . . . . . . . . . . . . . . 69
2.4 Model derived parameters for SINGS galaxies . . . . . . . . . . . . . . . 70
2.4 Model derived parameters for SINGS galaxies . . . . . . . . . . . . . . . 71
2.5 Model derived parameters for radio, U/LIRGs galaxies . . . . . . . . . . 72
2.5 Model derived parameters for radio, U/LIRGs galaxies . . . . . . . . . . 73
2.6 Observed molecular hydrogen line fluxes and mass in low metallicity dwarfs 74
3.1 H2 line fluxes for GOALS . . . . . . . . . . . . . . . . . . . . . . . . . . 100
3.1 H2 line fluxes for GOALS . . . . . . . . . . . . . . . . . . . . . . . . . . 101
3.1 H2 line fluxes for GOALS . . . . . . . . . . . . . . . . . . . . . . . . . . 103
3.2 H2 line fluxes for GOALS . . . . . . . . . . . . . . . . . . . . . . . . . . 104
3.2 H2 line fluxes for GOALS . . . . . . . . . . . . . . . . . . . . . . . . . . 105
xiii
4.1 MIR H2 line flux for Quintet . . . . . . . . . . . . . . . . . . . . . . . . 115
4.1 MIR H2 line flux for Quintet . . . . . . . . . . . . . . . . . . . . . . . . 116
4.1 MIR H2 line flux for Quintet . . . . . . . . . . . . . . . . . . . . . . . . 117
4.1 MIR H2 line flux for Quintet . . . . . . . . . . . . . . . . . . . . . . . . 118
4.1 MIR H2 line flux for Quintet . . . . . . . . . . . . . . . . . . . . . . . . 119
4.1 MIR H2 line flux for Quintet . . . . . . . . . . . . . . . . . . . . . . . . 120
4.1 MIR H2 line flux for Quintet . . . . . . . . . . . . . . . . . . . . . . . . 121
4.2 Power law model results for Quintet . . . . . . . . . . . . . . . . . . . . 122
4.2 Power law model results for Quintet . . . . . . . . . . . . . . . . . . . . 123
4.2 Power law model results for Quintet . . . . . . . . . . . . . . . . . . . . 124
4.2 Power law model results for Quintet . . . . . . . . . . . . . . . . . . . . 125
4.2 Power law model results for Quintet . . . . . . . . . . . . . . . . . . . . 126
4.2 Power law model results for Quintet . . . . . . . . . . . . . . . . . . . . 127
4.2 Power law model results for Quintet . . . . . . . . . . . . . . . . . . . . 128
4.2 Power law model results for Quintet . . . . . . . . . . . . . . . . . . . . 129
5.1 Photometric calibration for 1 count/s/pixel . . . . . . . . . . . . . . . . 150
5.2 Dust Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
5.3 Intensity in MJy/sr at FIR and sub-mm wavelengths . . . . . . . . . . . 158
5.4 B207 core/rim temperature . . . . . . . . . . . . . . . . . . . . . . . . . 166
xiv
List of Figures
1-1 ISM phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1-2 The cosmic lifecycle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1-3 H2 energy levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1-4 Effect of metallicity on CO and H2 cloud . . . . . . . . . . . . . . . . . . 14
2-1 Power law model fit for the H2 excitation diagrams . . . . . . . . . . . . 30
2-2 Histogram for the power law index, n . . . . . . . . . . . . . . . . . . . . 31
2-3 ∆χ2 value in the model parameter space Tℓ and n . . . . . . . . . . . . . 33
2-4 χ2 distribution for NGC 5033 (blue) and NGC 3627 (red) . . . . . . . . . 35
2-5 Histogram for the sensitivity temperature, Ts . . . . . . . . . . . . . . . 36
2-6 Average ∆χ2 contours for SINGS sample . . . . . . . . . . . . . . . . . . 37
2-7 Histogram for the lower cut off temperature, Tℓ . . . . . . . . . . . . . . 39
2-8 Cold and warm molecular gas mass distribution in NGC 5033 . . . . . . 40
2-9 Model and CO derived molecular gas mass corelation . . . . . . . . . . . 44
2-10 Histogram for the lower cut off temperature (ULIRGs and 3C galaxies), Tℓ 48
2-11 Warm gas mass in relation to FIR dust color . . . . . . . . . . . . . . . . 50
2-12 Model fit for the H2 excitation diagram for a low metallicity dwarf, Haro 11 54
2-13 Model predicted molecular gas mass in relation to metallicity . . . . . . . 56
3-1 IRS-Spitzer and CO beam size differences . . . . . . . . . . . . . . . . . 79
3-2 LIRG’s disk template . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
3-3 Power law model fit to a GOALS galaxy . . . . . . . . . . . . . . . . . . 83
xv
3-4 Power law indices for GOALS sample . . . . . . . . . . . . . . . . . . . . 85
3-5 Molecular gas estimation from power law model in GOALS galaxies . . . 87
3-6 Power law indices for GOALS galaxies . . . . . . . . . . . . . . . . . . . 89
3-7 Relation of power law index with 6.2 µm PAH . . . . . . . . . . . . . . . 91
3-8 Power law model derived H2 mass comparison with the dust emission method 94
3-9 Power law model derived H2 mass comparison with the L850 . . . . . . . 96
4-1 Stephan’s Quintet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
4-2 Spectral extraction grid- SQ . . . . . . . . . . . . . . . . . . . . . . . . . 112
4-3 Discrete two temperature fit for the H2 excitation diagram . . . . . . . . 113
4-4 Power law model fit for the H2 excitation diagram . . . . . . . . . . . . . 131
4-5 Two temperature fits in shocks of SQ . . . . . . . . . . . . . . . . . . . . 132
4-6 H2 line ratios in shocks of SQ . . . . . . . . . . . . . . . . . . . . . . . . 134
4-7 H2 ratio contours in shocks of SQ . . . . . . . . . . . . . . . . . . . . . . 135
4-8 Variation of power law index in shocks of SQ . . . . . . . . . . . . . . . . 136
4-9 χ2 variation in different regions of SQ . . . . . . . . . . . . . . . . . . . . 139
4-10 Warm gas mass fractions in shocks of SQ . . . . . . . . . . . . . . . . . . 142
5-1 A 3-colour stacked image of B207 consisting of V, R, and I bands . . . . 146
5-2 East-West horizontal cut across B207 for SB profile . . . . . . . . . . . . 149
5-3 Optical surface brightness profile for B207 . . . . . . . . . . . . . . . . . 151
5-4 3.4 micron image of B207 . . . . . . . . . . . . . . . . . . . . . . . . . . 156
5-5 SED fit for core and rim of B207 . . . . . . . . . . . . . . . . . . . . . . 159
5-6 Gas column density profile for B207 . . . . . . . . . . . . . . . . . . . . . 162
5-7 Temperature and density profile for B207 . . . . . . . . . . . . . . . . . . 167
5-8 Energy ratio profiles for B207 . . . . . . . . . . . . . . . . . . . . . . . . 172
6-1 Sensitivity of JWST-MIRI . . . . . . . . . . . . . . . . . . . . . . . . . . 183
6-2 Spectroscopic sensitivities of different instruments . . . . . . . . . . . . . 184
xvi
6-3 Histogram of the globule’s projected dense core-star distance . . . . . . . 186
xvii
List of Abbreviations
AGB . . . . . . . . . . . . . . . . . . . . . . Asymptotic Giant Branch
AGN . . . . . . . . . . . . . . . . . . . . . . Active Galactic Nucleus
CNM . . . . . . . . . . . . . . . . . . . . . Cold Neutral medium
DGR . . . . . . . . . . . . . . . . . . . . . . Dust to Gas Ratio
FIR . . . . . . . . . . . . . . . . . . . . . . . Far InfraRed
FWHM . . . . . . . . . . . . . . . . . . . Full Width at Half Maximum
GDR . . . . . . . . . . . . . . . . . . . . . . Gas to Dust ratio
GMC . . . . . . . . . . . . . . . . . . . . . Giant Molecular Cloud
GOALS . . . . . . . . . . . . . . . . . . . Great Observatories All Sky LIRGs Survey
HIM . . . . . . . . . . . . . . . . . . . . . . Hot Ionized medium
HST . . . . . . . . . . . . . . . . . . . . . . Hubble Space Telescope
IR . . . . . . . . . . . . . . . . . . . . . . . . InfraRed
IGM . . . . . . . . . . . . . . . . . . . . . . InterGalactic Medium
ISM . . . . . . . . . . . . . . . . . . . . . . . InterStellar Medium
ISRF . . . . . . . . . . . . . . . . . . . . . InterStellar Radiation Field
JWST . . . . . . . . . . . . . . . . . . . . James Webb Space Telescope
LIRGs . . . . . . . . . . . . . . . . . . . . Luminous Infrared Galaxies
MIRI . . . . . . . . . . . . . . . . . . . . . Mid-Infrared Instrument
NED . . . . . . . . . . . . . . . . . . . . . . NASA/IPAC Extragalactic Database
NIR . . . . . . . . . . . . . . . . . . . . . . . Near InfraRed
OPR . . . . . . . . . . . . . . . . . . . . . . Ortho-to-para ratio
PAHs . . . . . . . . . . . . . . . . . . . . . Polycyclic Aromatic Hydrocarbons
SED . . . . . . . . . . . . . . . . . . . . . . Spectral Energy Distribution
SFH . . . . . . . . . . . . . . . . . . . . . . Star Formation History
SFR . . . . . . . . . . . . . . . . . . . . . . Star Formation Rate
S/N . . . . . . . . . . . . . . . . . . . . . . . Signal to Noise ratio
SN . . . . . . . . . . . . . . . . . . . . . . . . Supernova
SQ . . . . . . . . . . . . . . . . . . . . . . . . Stephan’s Quintet
ULIRGs . . . . . . . . . . . . . . . . . . Ultra Luminous InfraRed Galaxies
UV . . . . . . . . . . . . . . . . . . . . . . . Ultraviolet
WIM . . . . . . . . . . . . . . . . . . . . . Warm Ionized medium
WNM . . . . . . . . . . . . . . . . . . . . . Warm Neutral medium
xviii
Chapter 1
The Interstellar Medium (ISM)
1.1 Introduction
The subject of this thesis is the most interesting component of galaxies, i.e. the
Interstellar Medium (ISM), which constitutes gas and dust. The ISM is an important
constituent of galaxies, for it is the ISM that is responsible for forming stars, the
building blocks of a galaxy. At the initial stage of galaxy evolution the baryonic mass
in galaxies was primarily in the ISM gas and as galaxies evolve, the ISM is gradually
converted to stars (Stahler & Palla, 2005).
The first generation of stars were primarily composed of basic elements, hydrogen
(H) and helium (He). As stars evolve, H and He are fused to form heavy elements
like C, N, and O and very massive stars may also form Fe. After the violent death of
stars these elements are ejected back in the ISM thus enriching the ISM with heavy
elements. The presence of heavy elements in Sun and Earth, implies that our solar
system is a descendant of the earlier generation of stars. The Ca in our teeth, Fe in
our blood, O which we breath were all synthesized in star cores and hence, it is apt
to say we are children of stars. Stars and ISM are well connected system, it is from
the ISM gas stars are born, and as a debt to repay these stars after dying enrich the
ISM back with more heavy and complex elements.
1
In the process of a galaxy’s evolution the gas may be ejected from the galaxy in
the form of galactic winds or stripped by some major/minor merging process or can
also cause infall of hot/cold gas in the disks (Veilleux et al. 2005). At present epoch
about 10% of the baryons is in the ISM of our galaxy - Milky Way.
H2 lacks dipole moment but with the advent of Infrared Space Observatory and
Spitzer telescope astronomers were able to measure quadrupole emission of H2 in
nearby and high redshift galaxies. It has been long assumed in the astronomy com-
munity that the MIR H2 emission is only able to trace the warm component of molec-
ular gas and hence, cannot be used to trace the total molecular gas content. My
thesis is aimed at using the MIR H2 rotational lines as a direct tracer to detect total
molecular gas in galaxies. Since our method employ direct usage of H2 molecules,
it is independent of other indirect tracers like CO, dust etc, which rely on intrinsic
assumptions of constant αCO, DGR and τdep, respectively. We are able to estimate
molecular gas mass even in low metallicity galaxies, as low as 10% of the Milky Way,
where CO fails to detect the total molecular gas mass.
1.2 Constituents of the ISM
The ISM governs the evolution of a galaxy and hence, can be determined by study-
ing physical properties of the ISM. The constituents of the ISM are:
Interstellar gas: Atoms, molecules, ions can be in the gas phase forming neutral,
molecular and ionized gas, respectively. The gas is the major component of the ISM.
Interstellar dust: Small solid particles majority in the size range 0.01–0.2 µm is
mixed homogeneously with the ISM gas however, in cold dense molecular cloud’s core
regions the dust grains can coagulate and can grow to micron size particles.
Cosmic rays: Charged particles with energies as high as 1021 eV have been detected
and form the cosmic rays component of ISM.
2
Electromagnetic radiation: Stars in galaxies are major sources of electromagnetic
radiation however cosmic microwave background photons from the relics of big bang
is almost distributed homogeneously in the space. It is the radiation from AGN,
young and old stars that heat the ISM gas and dust.
Interstellar magnetic field: The electric currents in the ISM is the source of ISM
magnetic field. The magnetic field guides the charged particles, cosmic rays, compo-
nent of the ISM. The magnetic field can play a pivotal role in cloud dynamics and
hence, evolution of galaxies.
The gravitational field: The ISM matter, stars, dust, stellar remnants, dark mat-
ter, all produce a gravitational field.
The dark matter particles: The dark matter particles dominate the mass in galax-
ies, when compared to the visible baryonic mass. These particles may interact non
gravitationally with baryons, magnetic field and other components of ISM.
1.2.1 The ISM Gas
The ISM gas can be in the molecular, atomic, or ionized form.
Molecular gas: The molecular gas in our galaxy- Milky Way is localized in GMCs.
The molecular gas phase has a small filling factor but accounts for a substantial
fraction of the Galactic ISM mass. Molecular clouds can be gravitationally or pressure
bound. On average the mass and size of a GMC is 4× 105 M⊙ and 40 pc, respectively.
The dense regions of GMCs are called cores, which have densities greater than 104
cm−3 and size of the order of 1 pc. Stars are formed in the gravitationally bound
cores of GMC.
Cold and Warm Neutral Medium (CNM and WNM) The neutral atomic gas
can be traced by the 21 cm spin transition line of atomic hydrogen. This neutral
atomic medium can be classified in two categories according to the temperature, cold
neutral medium (CNM) and warm neutral medium (WNM). The cold and warm
3
neutral mediums are about 100 K and 8000 K, respectively. The CNM occupies 1–
4% of the total volume and with an average density of 50 cm−3. The WNM is with
a much lower density about 0.5 cm−3 but occupies a total of 30–60% of the volume.
Warm Ionized Medium (WIM) As the name suggests the gas in this medium
is warm and ionized and so are at high temperatures. This phase is most clearly
associated with H II regions, where the gas is ionized by the photons from young
and hot stars and emits in the Hα recombination line. The density, temperature and
the filling phase of this phase is about 0.1 cm−3, 8000 K and 0.25, respectively.The
ionization source in the mid-plane disk of galaxies is young hot stars however, at high
galactic latitudes shock heating or suprathermal particles are responsible.
Hot Ionized Medium (HIM) This phase of the ISM is hot about 106 K and density
of 10−3 cm−3. The HIM occupies a large fraction of the volume of the halo but with a
small mass fraction of the ISM. The HIM is heated by strong shocks driven by violent
stellar winds from early type stars and supernova explosions. The hot gas detected at
high latitudes can be due to the superbubbles, flown out of the disk mid-plane of the
galaxies. The hot gas can be detected in absorption against bright background stars
in highly ionized species like C IV, S IV, N V, O III, O VI, etc or through continuum
and line emission in the far UV and X-ray wavelength range.
Table 1.1 and Figure 1-1 lists and shows the description of different phase compo-
nents of the ISM, respectively. In the Figure 1-1 ionized medium surrounds the cold
and warm neutral mediums, which further embeds the cold molecular cloud regions.
The dark dense core regions of molecular clouds becomes gravitationally unstable to
form stars.
4
Figure 1-1 The structure of ISM phase adopted from McKee and Ostriker(1977). The cold and warm neutral mediums are embedded inthe ionized medium of the ISM. The molecular clouds is furtherembedded in the cold neutral atomic medium. Stars, the basicbuilding blocks of galaxies are born in the cold core dense regionsof molecular clouds.
Table 1.1. Description of the phase components of the ISMa
Phase Density Temperature Volume Mass Hydrogen(cm−3) (K) % (M⊙)
Molecular cloud 102–106 10–100 1 109 H2
CNM 1–103 50–100 1–5 1.5×109 H IWNM 0.1–10 103–104 10–20 1.5×109 H IWIM 0.01 103–104 20–50 109 H IIHIM 10−4–10−2 105–107 30–70 108 H II
aThis table is a very schematic classification, and the separation between the ISM phases is debated.The numbers are orders of magnitudes
bThe masses are rough estimates of our galaxy- Milky Way
cThe last column lists the main state of the hydrogen atom in different phases
5
1.3 Evolution of ISM dust
The universe is dusty, and dust grains constitute the dominant form of solid matter
in the universe. Dust is observed in a wide variety of astrophysical environments,
ranging from stellar envelopes around cool red giants to SN ejecta, in diffuse halo
regions to cool dark dense core regions, from comets to interplanetary space to nearby
and high redshift galaxies and quasars. Recent studies with ALMA have traced dust
in galaxies at z = 7.5, occurring within only 500 Myr of the beginning of star formation
in the universe (Watson et al. 2015).
Dust grains are the main source of continuum opacity within galaxies, longward
of the Lyman limit of hydrogen, resulting in absorption and scattering of starlight at
UV/visible wavelengths and thermal emission in the IR and sub-mm range. Dust play
an important role in heating the ISM through photoelectric effect. The surface of dust
grains acts as a catalyst in the formation of H2 molecules in the galaxy ISM. Dust
particles, which contributes just about 1% of the ISM mass, play a crucial role in the
galaxy evolution and hence can be termed as the spice in the curry of galaxy. “What
is it about a casserole that makes it taste so wonderful? It’s not the potatoes, the meat,
the vegetables- no it’s the spice. In the universe, dust is but a minor constituent, but
it gives it all the flavor. Dust truly is the spice of the universe——”- Adolf Witt (ca.
2010)
1.3.1 Dust composition and size distribution
The composition of dust is characterized by spectral features observed in the
absorption and emission. The extinction, Aλ, ratio of the flux I(λ) after crossing the
dust cloud to the incident I0(λ) follows
I(λ) = I0(λ)exp(−τλ), (1.1)
6
where τλ = 0.921 Aλ is the optical depth. The extinction is obtained by comparing
the reddened spectrum of stars to the intrinsic spectrum from a library of spectral
models.
The extinction curve can be used to constrain size distribution of dust particles.
Mathis et al. (1977) demonstrated that the average ISM extinction could be produced
containing graphite and silicate grains with a power law size distribution (called as
MRN distribution) with an index of -3.5
dn ∝ a−3.5da, (1.2)
where dn is the number of grains between the size range a and a+da. The distribu-
tion is truncated between sizes in the range, 0.005-0.25 µm. With this distribution
the large dust grains are fewer in number but however dominate the mass fraction.
However, NIR scattering in the dark dense molecular cores (coreshine phenomenon) is
observed, suggesting the presence of micron size dust grains possibly formed through
coagulation process (Pagani et al. 2010).
ISM dust grain models have been improved in the past three decades to fit the
observational constraints. Elemental abundance of heavy elements, UV visible and
IR absorption and scattering properties, IR emission, polarization properties help to
constrain the composition of ISM dust. The models include silicate and carbonaceous
grains with variable size distributions to reproduce the observed ISM dust properties.
After the discovery of the 2175 A bump immediate studies claimed small graphite
particles to be responsible for this phenomena. Other forms of carbonaceous grains
(PAHs, amorphous carbons, and organic mantles on silicate grains) are being consid-
ered. The absorption feature in the MIR spectra at 9.7 and 18 µm is due to interstellar
silicates due to Si-O stretching and O-Si-O bending modes, respectively.
7
1.3.2 Dust processing in the ISM
In the following section we summarize the mechanism causing structural changes,
disrupt or even completely destroy dust grains. Apart from photo-processing all the
processes are collisional processes, interactions with the ions, molecules, and other
dust grain particles.
Sputtering: This process causes atoms to be removed from the surface of dust
grains resulting in erosion and returns to the gas phase. The erosion is due to grain
collisions with high velocity atoms and ions in energetic environments (Jones et al.
1994). The sputtering can be classified as either non-thermal and thermal, depending
on the relative motion between grain-gas and thermal motion of the gas, respectively.
Thermal sputtering occurs at high temperatures above 105 K dominated by collisions
with H, He atoms and ions. In grain collisions with hydrogen, inertial sputtering
happens at velocities higher than 30-40 km/s and at lower velocities dominated by
collisions with heavy particles. The velocities depends on the size of dust grains and
hence the effect of sputtering is size dependent.
Vaporization: Grain-grain collisions with impact velocities above 20 km/s result in
partial or complete vaporization of dust grains, releasing the elements present in the
dust to the gas phase.
Accretion: The adsorption and sticking onto grain surfaces of atoms, ions, radicals
and molecules increase the grain radius and the mass. For Mg, Si, and Fe, accretion
can occur at temperatures as high as 1000 K, while temperatures below 100 K are
required for species leading to formation of icy mantles. In dense cores we find CO, an
effective molecular gas tracer, freezing onto the grains and hence there is a reduction
of CO in the ISM gas phase.
Shattering: This occurs in grain-grain collisions with velocities of few km/s and
cause fragmentation of dust grains resulting in redistribution of mass towards smaller
8
grains.
Coagulation: Grains colliding with each other below a critical value cause con-
glomeration with a redistribution of mass towards larger grains. The critical velocity
depends on the grain composition and radius, varying from ∼ 1m/s for micron sized
grains to 1 km/s for 100 A sized grains, and is enhanced by the icy mantles, which
cause increase in the surface area and interaction time during the collision.
1.4 The cosmic life cycle
Stars are formed in the dark cold dense gravitationally unstable core regions of
molecular cloud. When the temperature and density are sufficiently high to fuse
hydrogen to helium nucleus, a new star is born. Nuclear reactions form elements
upto C, O, and N. The radiation pressure supports the gravitational collapse and
make them stable during the long (107 - 1010 yr) lifespan of the star. After leaving
the main sequence phase the star enters the AGB phase, when the outer layers are
loosely bound to the core and the thermal instabilities bring the heavy elements like
C and O to be mixed and bought to the stellar surface. In the end the outer envelope
is stripped away enriching the ISM with heavy elements and the core further cools
down to white dwarf. However for a heavy mass star (M⋆ ≥ 8M⊙) they are able
to fuse C and O into heavy elements till Fe and the core finally collapse and later
reduced to a neutron star, pulsar, or a black hole.
Ionizing photons from hot stars, O and B type, create H II regions and ionize
the surrounding medium and even after their death including the low mass stars
ionize ISM with heavy elements with simple to complex molecules, CO, PAHs, and
other acetylenic chain derivatives. After injection into the ISM the stardust and
molecules can cycle many times between the intercloud and cloud phases. UV rays
in the diffuse and reactions with ions continuously break and form molecular bonds,
9
Figure 1-2 The cosmic lifecycle. Figure adapted from Jones (2004).
allowing stable species to survive. The molecular clouds consisting of dust and other
molecules may become gravitationally unstable and then collapse to form second
generation of stars. These stars are enriched with heavy elements, when compared
to the first generation of stars. Protoplanetary disks may form around low mass
stars and likely all interstellar grains are vaporized and recondensed as solar system
condensates. The circle is complete and the cycle starts again. The cosmic life cycle
is schematically summarized in the Figure 1-2.
10
1.5 ISM molecular gas
1.5.1 Importance of H2
Molecular gas is a primary fuel for star formation in galaxies. Stars are formed
from gravitationally unstable cores within molecular clouds. To understand the star
formation rate it is important to know the molecular gas surface density. The molec-
ular gas density is proportional to the star formation rate density (Bigiel et al. 2008,
Kennicutt 1998). In the Milky Way galaxy the atomic and molecular gas mass are
comparable however, in local mergers, ULIRGs, the atomic and molecular gas is fun-
neled at the center and in the ISM of these galaxies the molecular gas mass can dom-
inate over the atomic gas, subsequently increasing the star formation surface density
about an order or two magnitude higher than the normal star forming galaxies.
1.5.2 Detection techniques of molecular gas
H2 is a homonuclear molecule and hence, lacks a dipole moment. The lowest
energy transitions of H2 are its purely rotational quadrupole transitions in the mid
infrared spectrum. The upper energy level of this transition, S(0), occurring at λ =
28.22 µm is at an energy level of 510 K, much higher than the ISM temperature. Due
to these constraints, despite H2 is the most abundant molecule in the universe it is a
poor emitter and so invisible. Figure 1-3 shows rotational energy levels of the ground
vibrational state of H2.
The ISM molecular gas consists other molecules which can be used as an indirect
tracers to estimate molecular gas content. Helium being monoatomic, suffers from
similar observability problems in cold clouds. Carbon Monoxide (CO) has a weak
permanent dipole moment (0.11 D = 0.11 × 10−18 esu cm) and the ground rotational
transition is at an energy level of 5.5 K, which makes CO to be easily excited in cold
11
Figure 1-3 The upper energy level of the lowest rotational line transition ofH2 (S(0), occurring at 28.22 µm) of the ground vibrational state(ν = 0) is at a temperature equivalent of 510 K, approximatelyan order of magnitude higher than the average ISM gas temper-ature. With such high energy levels and no dipole moment H2 isclassically considered a poor emitter.
dense molecular clouds. At a wavelength of 2.6 mm, J = 1–0 transition, is in the
transparent atmospheric window. Hence, CO is widely used as a tracer to detect
molecular gas in galaxies.
Solomon et al. (1987) and Scoville et al. (1987) measured velocity line width, size,
virial mass, and CO luminosity for molecular clouds in the Milky Way and observed
a relationship between the cloud dynamical mass, measured by the virial theorem
with the CO luminosity. This led to the calibration of measuring molecular gas mass
using the CO luminosity measurements. To use CO as a tracer to detect molecular
gas in galaxies a conversion factor, αCO or XCO, is used to convert CO intensity into
the column density of molecular gas using
N(H2) = XCO × W (CO, J = 1 − 0), (1.3)
where the column density, N(H2), is in cm−2 and the integrated line intensity, W(CO),
12
is in K km s−1. If used in mass units the equation is
Mmol = αCOLCO, (1.4)
where Mmol is the molecular gas in solar masses and LCO is in K km s−1 pc2. Both
αCO or XCO are referred to as the “CO-H2 conversion factor”. The values of αCO,
XCO are 4.35 M⊙ (K km s−1 pc2)−1 and 2×1020 cm−2 (K km s−1)−1, respectively.
XCO accounts only the column density of H2 but when used αCO the correction for
He and other heavy elements are considered by multiplying with a factor of 1.36.
1.5.3 Dark Molecular Gas
For almost half a century, CO is traditionally being used as the tracer to estimate
molecular gas content in GMCs of Milky Way and in the ISM of extragalactic systems.
However, a substantial amount of H2 gas may lie outside the CO region, in the outer
regions of the molecular cloud where the gas phase carbon can be in the form of C or
C+. This H2 gas is termed as ”CO-dark” or ”CO-faint” gas and can be primarily due
to self shielding phenomenon of H2 or shielding by dust from UV photodissociation.
For a standard cloud of 106 M⊙ with an incident radiation field, G’0 = 10, the dark
gas mass fraction is 0.28 and 0.31 for constant thermal pressures of 105 and 106 K
cm−3, respectively (Wolfire et al. 2003). Theoretical studies by Wolfire et al. (2003)
indicates variation in dark gas fraction ranges from 0.25–0.33 for a range of G’0 =
3–30 and cloud mass 105–106 M⊙.
Recent results by Herschel Space Observatory’s GOT-C+ project estimated again
about 20–30% of dark gas mass fraction in the Milky Way (Pineda et al. 2013). C+
emission is majorly associated with spiral arms and traces the envelopes of clouds and
in the transition region between molecular and atomic gas. Most of the observed C+
is produced by dense PDRs, with smaller contributions from CO dark H2 gas, cold
13
Figure 1-4 Effect of metallicity on CO and H2 cloud in a uniform radiationfield. Blue shading indicates the molecular gas region. Increas-ingly darker shading areas where the carbon is in C+, C, and CO.The upper sequence show the effect of decreasing metallicity andDGR on the distribution of C+, CO, and H2. The lower sequenceshows the effect of changing clump size or column density at afixed metallicity. Lower metallicity or low DGR reduces the COemitting region and hence an increase in the conversion factor,αCO, is required (Figure from Bolatto et al. 2013)
atomic gas and ionized gas. The fraction of CO-dark H2 to total H2 increases with
the Galactocentric distance, ranging from 20–80% from 4–10 kpc. Most of the C+
emission emerges from the dense PDRs with modest far-ultraviolet fields in the range
1–30. Above mentioned theoretical and observational studies indicates possibility to
underestimate molecular gas content using CO as a tracer.
Dust and gas are homogeneously mixed in the galaxy ISM and DGR is almost
constant in the galaxy ISM. Israel et al. (1997) using dust emission determined the
molecular gas mass in low metallicity galaxies, where CO is unable to trace the dark
molecular gas component. Assuming a constant DGR and estimating the dust and
atomic gas mass from IR and 21 cm line observation, respectively, can be used in
14
determining molecular gas mass using the equation
Mdust × GDR = M(HI) + M(H2) (1.5)
Molecular gas is the fuel for the star formation in galaxies. Assuming a constant
molecular gas depletion time (τdep) to form stars and estimating the star formation
rate (SFR) using observation tracers, Hα and IR continuum, the molecular gas content
can be estimated using,
Mmol = SFR × τdep (1.6)
(Schruba et al. (2012)).
Estimating molecular gas content using dust or SFR assume a constant DGR
or τdep, respectively, and are subjected to their own internal biases. Recent studies
have indicated at lower metallicities, 12+log[O/H] ≤ 8, the dependence of DGR is
steeper, implying relatively less dust (more gas) at lower abundances (Remy-Ruyer
et al. 2014). Physical processes in the galaxy ISM can alter these constants and the
molecular gas mass estimated from these indirect tracers can be under/over estimated.
A model employing the direct tracer of molecular gas, H2, free from such
internal bias is required for estimating molecular gas mass in galaxies, and
this holds the motivation for the thesis.
1.6 Thesis outline
The thesis outline is as follows
Chapter 2: I discuss here our method in using MIR H2 rotational lines to detect
total molecular gas mass in galaxies and apply our model in low metallicity galaxies
to estimate the molecular gas content.
Chapter 3: In this chapter I apply our power law model, discussed in the previ-
15
ous chapter, in local mergers, ultra/luminous infrared galaxies (U/LIRGs) of GOALS
sample. The ISM in mergers of GOALS are subjected to excess shock excitations,
high turbulence, increasing the dust and gas temperatures, when compared to normal
star forming galaxies. Using the power law model we study the ISM properties in
GOALS and compare the results with normal star forming galaxies.
Chapter 4: Stephan’s Quintet is a compact galaxy group undergoing violent col-
lisions. Radio observations in the early 1970’s revealed a mysterious filament of
emission in the IGM of these galaxies. The Spitzer telescope discovered powerful
MIR H2 rotational lines in the IGM of Stephan’s Quintet, powered by shocks. I use
our method in detecting molecular gas content and in understanding the ISM physics
of Stephan’s Quintet.
Chapter 5: Star formation in galaxies majorly occurs in GMCs, however there are
small isolated dark nebulae containing dense dust and molecular gas, known as Bok
globules, forming low mass stars. The cause of cometary structure of Bok glob-
ules is still a mystery and hence, a subject of intense research. Here, I studied the
properties of an isolated small molecular cloud, B207, harboring a Class I protostar,
IRAS04016+2610, in the Taurus-Auriga molecular cloud region. Using optical to
sub-mm observational data we study physical properties and the evolutionary fate of
B207.
Chapter 6: Here we state our future plans and how in the coming era of James
Webb Space Telescope (JWST) and other major coming up and existing telescopes,
with much better sensitive instruments, will revolutionize the study of ISM physics.
The Mid Infrared Instrument (MIRI) spectrograph, onboard JWST, with sensitivities
100× better than the previous IRS-Spitzer is capable of detecting H2 lines in high
redshift galaxies. Using these fluxes of H2 rotational lines in our power law model it
will be possible to understand molecular ISM properties in galaxies across different
redshifts and hence, star formation history that shaped our universe.
16
Chapter 2
Lighting the dark molecular gas:
H2 as a direct tracer
2.1 Introduction
By a factor of more than 104, H2 is the most abundant molecule in the universe,
found in diverse environments ranging from planet atmospheres to quasars hosts
(Hanel et al. 1979, 1981, Walter et al. 2003, Genzel et al. 2014). It was the
first neutral molecule to form in the Universe and hence dominated the cooling of
pristine gas at early times (Lepp et al. 2002). Stars form principally from molecular
clouds, and most physical prescriptions for stellar formation rate (SFR) are therefore
directly linked to the surface density of H2 gas (Kennicutt 1998, Bigiel et al. 2008).
To understand the star formation processes and how it varies and has evolved over
the star-forming history of the Universe, it is necessary to accurately measure the
mass and distribution of H2.
Despite its high abundance, H2 can be difficult to study directly. It possess no
permanent dipole moment, which makes it a weak rotational emitter. In addition,
the upper energy level for the lowest permitted rotational (quadrupole) transition is
E/k = 510 K above ground (Dabrowski 1984; van Dishoeck & Black 1986). Hence,
17
the H2 gas that comprises the bulk of the molecular interstellar medium (ISM) is
commonly believed to be too cold to be visible. And yet, despite these observational
disadvantages, the advent of the Infrared Space Observatory (ISO) and in particular
Spitzer’s Infrared Spectrograph (IRS, Houck et al. 2004) revealed pure rotational
H2 emission from a rich variety of extragalactic sources, including normal star form-
ing galaxies (Roussel et al. 2007), ultraluminous and luminous infrared galaxies
(U/LIRGs Lutz et al. 2003, Pereira-Santaella 2010; Veilleux et al. 2009; Stierwalt et
al. 2014), galaxy mergers (Appleton et al. 2006), radio-loud AGN (Ocana Flaquer
et al. 2010), UV-selected galaxies (O’Dowd et al. 2009), quasar hosts (Evans et al.
2001), cooling-flow cluster systems (Egami et al. 2006), sources with extreme shock-
dominated energetics (Ogle et al. 2010), and even star-forming sources at redshifts
& 3 (Walter et al. 2003; Solomon $ Vanden Bout 2005).
In the absence of direct measurements of H2, the lower rotational line transitions
of CO (the next most abundant molecule) are generally used as a molecular gas
tracer. To convert the measured integrated intensities of 12CO to H2 column density
a conversion factor, αCO, is needed, typically calibrated against virial mass estimates
in presumed self-gravitating clouds (Solomon et al. 1987; Scoville et al. 1987; Strong
& Mattox 1996; Abdo et al. 2010). However recent evidence indicates that the value
of αCO varies substantially, both within galaxies and at different epochs (Genzel et
al. 2012; Sandstrom et al. 2013).
In molecular clouds in our Galaxy, the observed γ-ray flux arising from the cosmic
ray interactions with H2 can be used to recover H2 gas mass accurately (Bhat et
al. 1985; Bloeman et al. 1984,1986), but this method requires information on the
cosmic ray distribution not available in other galaxies. Another method to assess H2
content is to use dust as a tracer, with a presumed or recovered dust-to-gas (DGR)
ratio (e.g. Sandstrom et al. 2013; Remy Ruyer et al. 2014). Among other biases,
both the CO and dust-based methods lose reliability at low metallicity. Due to
18
declining dust opacity and less effective self-shielding than H2, the CO abundance
drops rapidly with decreasing metallicity (Wolfire et al. 2010; Bolatto et al. 2013).
Variations in the radiation field can also lead to selective CO destruction (Hudson
1971; Bolatto et al. 2013). At the lowest metallicities, DGR itself begins to scale
non-linearly with the metal content of the gas (Herrera-Camus et al. 2012; Fisher
et al. 2014; Remy-Ruyer et al. 2014). Another method of inferring molecular gas
content is to couple measured star formation rates with an assumed constant H2
depletion timescale (τdep), equivalent to a constant star formation efficiency (SFE) in
galaxies (Schruba et al. 2012). None of these methods provide direct tracers of the
H2 reservoirs, and the relevant indirect tracers all rely on physical assumptions (e.g.
constant or measurable values of αCO, DGR, or τdep) which are unlikely to be valid
in all environments.
Here we challenge the long-stated assumption that H2 rotational emission is a poor
tracer of the total molecular content in galaxies. Our understanding of the structure
of the molecular material in galaxies is undergoing significant revision. While CO
traces the coldest component of the molecular ISM, half or more (and sometimes
much more) of the molecular gas in galaxies is in a warmer, CO-dark state, where H2
persists (Field et al. 1966; Tielens 2005; Draine 2011; Wolfire et al. 2010; Pineda et
al. 2013; Velusamy & Langer 2014). What this means is that the common wisdom
that all molecular gas is found at temperatures T = 10 − 20K is incorrect. With
molecular material existing in quantity at excitation temperatures 50 - 100K, it be-
comes possible to use the rotational emission spectrum of the Universe’s dominant
molecule to directly assess a substantial portion of the total molecular content in
galaxies.
We introduce here a simple continuous temperature model which enables direct
use of the H2 rotational emission from galaxies to recover their total molecular gas
content. While several assumptions are required to make such a model possible, the
19
resulting biases are expected to be completely distinct from those of the indirect gas
tracers. The paper is laid out as follows. We describe the archival sample in § 2. A
rotational temperature distribution model is presented in § 3, and in § 4 we describe
the method and its calibration procedure. Section 5 presents results, discussions,
applications and future prospects of our model. We summarize our conclusions in § 6.
2.2 Sample
We have selected a large sample of galaxies spanning a wide range of different
physical properties with reliable H2 rotational emission detections. Our sample in-
cludes 14 Low Ionization Nuclear Emission Regions (LINERs), 18 galaxies with nuclei
powered by star formation, 6 Seyfert galaxies, 5 dwarf galaxies, 11 radio galaxies, 19
Ultra Luminous Infrared galaxies (ULIRGs), 9 Luminous Infrared galaxies (LIRGs),
and 1 Quasi Stellar Object (QSO) host galaxy. The galaxies were chosen to have
at least three well-detected H2 pure-rotational lines with Spitzer/IRS (Houck et al.
2004), including either S(0) or S(1), occurring at 28.22 and 17.04 µm, respectively.
We also required an available measurements of CO line emission (either J = 1–0 or
2–1) for an alternative estimate of the total molecular gas mass. See Table 2.1 for a
complete sample list with relevant physical parameters and the references from which
each galaxy was drawn.
The flux ratios S70
S160are calculated using 70 and 160µm intensities obtained from the
Herschel-PACS instrument (Poglitsch et al. 2010). We convolved PACS 70 µm maps
to the lower resolution of PACS 160 µm to calculate flux ratios for mapped regions,
for which we have the H2 line flux. For ULIRGs and other unresolved galaxies, S70
S160
represents the global flux ratio. The PACS70/PACS160 ratios are listed in Table 2.1
for each galaxy.
20
2.2.1 MIR H2 rotational line fluxes
All MIR H2 rotational line fluxes were obtained from Spitzer-IRS observations at
low (R ∼ 60 – 120) and high (R ∼ 600) resolution between 5 – 38 µm. The Spitzer
Infrared Nearby Galaxies Survey (SINGS, Kennicutt et al. 2003) is a diverse sample
of nearby galaxies spanning a wide range of properties. The SINGS sample consists of
LINERs, Seyferts, dwarfs, and galaxies dominated by star formation in their nuclei.
We adopted the H2 rotational line fluxes for four lowest rotational lines (S(0) to S(3))
in SINGS galaxies from Roussel et al. 2007. The targets were observed in spectral
mapping mode and the details of the observing strategy is described in Kennicutt et
al. 2003 and Smith et al. 2004.
H2 rotational line fluxes for 10 3C radio galaxies are from Ogle et al. 2010. This
sample includes from S(0) to S(7). For the radio galaxy 3C 236, the fluxes of H2
rotational lines are obtained from Guillard et al. 2012.
The H2 rotational line fluxes for QSO PG1440+356 were taken from the Spitzer
Quasar and ULIRG Evolution Study (QUEST) sample of Veilleux et al. 2009, and
those for the 19 ULIRGs were obtained from Higdon et al. 2006. The fluxes of H2
rotational lines for the NGC 6240 and other LIRGs in the sample are from Armus et
al. 2006 and Pereira-Santaella et al. 2010, respectively. H2 rotational lines are also
observed in the molecular outflow region of NGC 1266 (Alatalo et al. 2011). Table
2.2 lists the flux of MIR H2 rotational lines for all our selected galaxies.
2.2.2 Cold molecular gas mass from CO line intensities
To test and calibrate a model which measures total molecular mass from the ro-
tational lines of H2, an estimate of the “true” cold H2 gas mass in a subset of galaxies
is required. For this purpose, we utilize the cold H2 gas masses for the SINGS sam-
ple compiled by Roussel et al. 2007, which are derived from aperture-matched CO
21
velocity integrated intensities. The observed line intensities of the 2.6 mm 12CO(1–0)
line within the measured CO beam size was chosen to match the aperture of the
Spitzer-IRS spectroscopic observations. Roussel et al. 2007 assumed a CO intensity
to molecular gas conversion factor (αCO) of 5.0 M⊙(K km s−1 pc2)−1 (equivalent to
XCO = 2.3× 1020 cm−2(K km s−1)−1). They applied aperture corrections by project-
ing the IRS beam and CO beam on the 8 µm Spitzer-IRAC map. The underlying
assumption is that the spatial distribution of the PAH bands and CO(1–0) line emis-
sions are similar at large spatial scales. PAH emission emerges from photo dissociation
regions, the outer regions of molecular clouds, where UV photons are able to excite
these molecules and hence the distribution are mostly co-related.
For the radio galaxies, molecular gas masses were compiled by Ogle et al. 2010, and
the molecular gas mass was estimated from the CO flux density corrected and scaled
to the standard Galactic CO conversion factor. For the radio galaxy 3C 236 Ocana,
Flaquer et al. (2010) estimated an upper limit to the cold H2 gas mass. The cold
H2 gas masses for the LIRGs in our sample were obtained from Sanders et al. 1991.
They used αCO = 6.5 M⊙(K km s−1 pc2)−1 (equivalent to XCO = 3.0 × 1020 cm−2(K
km s−1)−1). We derived aperture corrections to the CO intensities by projecting IRS
and CO beam on the 8 µm IRAC map, since the CO beam size (55 arcsec aperture)
is much larger than the IRS mapped region (13.4 × 13.4arcsec2). Required aperture
corrections are typically less than a factor of 2. For the LIRGs NGC 7591, NGC
7130, and NGC 3256, the value of molecular gas mass was obtained from Lavezzi &
Dickey (1998); Curran et al. (2000); Sakamoto et al. (2006), respectively. The global
molecular gas masses for ULIRGs are from Rigopoulou et al. (1996); Sanders et al.
(1991); Mirabel et al. (1989); Evans et al. (2002). For the quasar QSO PG1440+356,
Evans et al. (2001) derived the molecular gas mass.
Different studies have utilized substantially different αCO values to calculate molec-
ular gas masses. In the Milky Way disk, a modern value of αCO,Gal = 4.35±1.3M⊙(K
22
km s−1 pc2)−1 (equivalent to XCO = 2.0× 1020 cm−2(K km s−1)−1) has been found to
be consistent with a wide variety of constraints (Bolatto et al. 2013). In this work,
we scale all molecular gas mass estimates to correspond to αCO = αCO,Gal. Since we
require only the H2 gas masses in our analysis, the resulting molecular mass values
were further reduced by a factor of 1.36 to remove the mass contribution of Helium
and other heavy elements. Recent dust-based studies of a subset of the SINGS sample
have revealed variations in αCO both between and within these galaxies Sandstrom
et al. (2013).
Table 2.2 lists the CO-based H2 gas mass estimated for each galaxy in the sam-
ple, aperture matched to the region of Spitzer/IRS coverage, along with H2 masses
calculated using the conversion factor for central regions derived by Sandstrom et al.
(2013) in parentheses, where available.
2.3 Model
The primary methods of measuring H2 mass all rely on a set of assumptions —
that the αCO factor is known and constant, that the dust-to-gas ratio is fixed or tied
directly to metallicity, or that the star formation depletion time is unchanged from
environment to environment. All of these methods also rely on: the CO molecule,
which is 10,000× less abundant than H2, dust grains, with their complex formation
and destruction histories and varying illumination conditions, or newly formed stars,
which are presumed to be associated with molecular material with a consistent conver-
sion efficiency. All of these methods suffer biases based on the physical assumptions
made.
We propose a direct means of assessing total molecular gas mass using the MIR
quadrapole rotational line emission of H2. Interstellar or direct ultraviolet radiation
fields, shocks, and other mechanical heating processes are all potential sources of H2
23
excitation. The method we propose requires only the assumption that H2 rotational
temperatures are, when averaged over the diversity of emitting environments and
processes in galaxies, widely distributed, and smoothly varying. This is directly
analogous to modeling the radiation field intensity which heats dust grain populations
in galaxies with a smooth distribution (e.g. Draine et al. 2007; Galliano et al. 2003).
We model the H2 temperature distribution as a smooth, truncated power law with
fixed cut-off temperatures. Indeed, the H2 molecule is radiatively cooled by a cooling
function which can be well approximated as a power law with slope 3.8 (Hollenbach
& McKee 1979; Draine et al. 1993). A similar analysis by Burton (1987) derived a
cooling function with a power law index, n = 4.66, for H2 molecules in the temperature
range 10–2000 K.
The temperature equivalents of H2 rotational levels are high, starting at T = 510
K (see Table 2.3). The Boltzmann distribution of energy levels, however, leads to
substantial excitation even at more modest peak temperatures. The lowest upper
energy levels at J = 2–4 are well populated at excitation temperatures of 50 to 150
K, whereas the higher-J levels are excited by temperatures from a few hundred to
several hundred K. Typically, a small number of discrete temperature components
are fitted to low-J rotational line fluxes, recovering warm H2 temperatures in this
range (Spitzer et al. 1973; Spitzer & Cochran 1973; Spitzer et al. 1974; Spitzer $
Zweibel 1974; Savage et al. 1977; Valentijn & van der Werf 1999; Rachford et al.
2002; Browning et al. 2003; Snow & McCall 2006; Ingalls et al. 2011).
Other studies have also adopted continuous temperature distributions for H2. Za-
kamska et al. (2010) in her study of H2 emission in ULIRGs used a power law model
dM ∝ T−ndT to calculate the expected H2 rotational line ratios, where dM is the
mass of H2 gas with excitation temperatures between T and T + dT . In her sample,
power law indices 2.5 < n < 5.0 are required to reproduce the observed range of H2
line ratios. Similarly Pereira-Santaella et al. (2014) adopted a power-law distribution
24
in H2 to model H2 rotational emission in six local infrared bright Seyfert galaxies,
with power law indexes ranging from 4–5. In the shocked environment of supernova
remnant IC 443, Neufeld et al. (2008) found a power law temperature distribution for
H2 column density with power law index varying over the range 3–6 with an average
value of 4.5.
The present practice for measuring warm (T & 150K) H2 gas mass is to use
two or three distinct temperature components to model the H2 rotational line fluxes.
Yet even on the scales of individual molecular clouds, H2 is present at a wide range
of temperatures, calling into question whether discrete temperature components are
physical. When used purely for assessing warm gas mass (T & 100K), the power
law method provides a more robust, unique and reproducible measure than discrete
temperature fits, which are sensitive to the arbitrary choice of starting line pair. A
continuous distribution also makes possible extrapolation to a suitable lower temper-
ature to recover the total cold molecular gas mass (see § 2.4).
We fit the flux of H2 rotational lines using a continuous power law temperature
distribution by assuming
dN = mT−ndT, (2.1)
where dN is the column density of H2 gas between excitation temperature T and
T + dT , n is the power law index, and m is a constant. Integrating this distribution
to recover the total column density, the scaling co-efficient m is found to be
m =Ntot(n − 1)
T 1−nℓ − T 1−n
u
, (2.2)
where Tℓ and Tu are the lower and upper temperatures of the distribution, respectively,
and Nobs is the column density of H2 in the observed line of sight and temperature
is presumed equal to the rotational temperature. For a continuous distribution of
molecules with respect to temperature, the column density of molecules at upper
25
energy level j responsible for the transition line S(j) is
Nj =
∫ Tu
Tℓ
gj
Z(T )× e
−Ej
kT mT−ndT, (2.3)
where gj is the degeneracy value for the corresponding energy level Ej . The degener-
acy value for even and odd values of j corresponding to para and ortho H2 is
gj = 2j + 1, (2.4)
and
gj = 3(2j + 1), (2.5)
respectively. The factor of 3 for odd values of j is for ortho-hydrogen, which have
parallel proton and electron spins, forming a triplet state. Z(T ) is the partition
function at temperature T . The model assumes local thermodynamic equilibrium
(LTE) between ortho and para H2 and the abundance ratio of ortho to para H2 at
any temperature, T , in LTE is given by Burton et al. (1992). The partition functions
for para and ortho H2 are
Z(T ) = Zp(T ) =∞
∑
j=even
(2j + 1)e−Ej
kT (2.6)
and
Z(T ) = Zo(T ) =∞
∑
j=odd
3(2j + 1)e−Ej
kT , (2.7)
respectively.
Table 2.3 list the wavelength of each H2 rotational line transition, the upper en-
ergy level of the corresponding transition in temperature units, and their radiative
rate coefficients A (Huber 1979; Black & Dalgarno 1976).
26
Assuming H2 emission to be optically thin, the flux observed in a given transition
j is
Fj =hνANj+2Ω
4π, (2.8)
where Ω is the solid angle of the observation. Substituting the value of Nj from
equation 2.3 into Eq. 4.3 and then using equation 2.2, we obtain the total column
density
Ntot =4πFjλ(T 1−n
ℓ − T 1−nu )
AhcΩ(n − 1)∫ Tu
Tℓ
gj+2
Z(T )e
−Ej+2kT T−ndT
(2.9)
From Eq. 4.3, we conclude that the column density is proportional to the flux F
and the corresponding transition wavelength λ, and inversely proportional to the
spontaneous emission probability A,
Nj+2 ∝Fjλ
A(2.10)
To perform the fit, we develop an excitation diagram from the ratio of column den-
sities,
Nj+2
N3=
FjλjA1
F1λ1Aj, (2.11)
where N3 is the column density of the upper energy level of the transition S(1). We
select N3, corresponding to the S(1) transition, since it is the brightest and most
frequently detected line. Using equation 2.3, we find this column density ratio from
the continuous temperature model to be
Nj+2
N3=
∫ Tu
Tℓ
gj+2
Z(T )× e
−Ej+2kT T−ndT
∫ Tu
Tℓ
g3
Z(T )× e
−E1kT T−ndT
. (2.12)
We determined the parameters of our model by comparing the observed and the
modeled column density ratios from equation 4.5 and 2.12, by varying the upper and
27
lower temperature Tℓ and power law index n (see § 2.4). Knowing these parameters,
we substitute their values in equation 2.9, to obtain the total column density. The
total number of molecules is calculated using
ntot = NtotΩd2, (2.13)
where ntot is the total number of hydrogen molecules and d is the distance to the
object. We then calculate the total H2 gas mass,
MH2,tot = ntot × mH2 , (2.14)
where mH2 is the mass of a hydrogen molecule.
2.4 Method & Calibration
Applying the model developed in § 2.3 permits direct assessment of the warm H2
column. Extrapolating the model to lower temperatures to recover the full molecular
content requires a calibration procedure using a trusted independent estimate of H2
content. Here we describe both applications.
2.4.1 Warm H2
The power law temperature distribution of H2 molecules in Eq. 2.1 has three pri-
mary parameters: upper temperature Tu, power law index n, and lower temperature
Tℓ. An excitation diagram formed from the H2 line flux ratios is the distribution of
normalized level populations and constrains our model parameters. Excitation dia-
grams relate the column density of the upper level (Nu) of a particular transition,
normalized by its statistical weight, gu, as a function of its energy level Eu. In the
following sections we describe how the model can be used to estimate the warm H2
28
gas mass. In practice, except in a few cases, the lower temperature cutoff Tℓ is not di-
rectly constrained (see § 2.4.2.1). The model fit results themselves are given in Table
2.4.
2.4.1.1 Upper temperature, Tu
The bond dissociation energy for H2 is 4.5 eV, corresponding to E/k ∼ 5 ×
104 K, hence H2 is not typically bound above this temperature. H2 present in photo-
dissociation regions (PDRs) in the outer layers of the molecular cloud can reach
temperatures of a few 100 K (Hollenbach et al. 1997, Hollenbach et al. 1999). For
the SINGS galaxies, Roussel et al. (2007) concluded that the mass of H2 greater
than 100 K contributes 1–30% of the total H2 gas mass. In the MOlecular Hydrogen
Emission Galaxies (MOHEGs) of Ogle et al. (2010), M(H2) > 1500 K contributes
only 0.01% of the total H2. M(H2) > 300 K contributes less than 1% in ULIRGs
(Higdon et al. 2006).
In our model, when Tu is allowed to vary above 1000K, we found negligible impact
on the recovered total gas mass or the quality of the fit to the excitation diagram.
We therefore fixed the upper temperature of the power law model distribution at
Tu = 2000 K. H2’s ro-vibrational excitation spectrum (including the 1-0 S(1) line at
λ = 2.12µn) would provide sensitivity to even hotter gas, but the mass contribution
of this very hot gas in the context of our model is insubstantial.
2.4.1.2 Power law index, n
Keeping a fixed value of Tu = 2000K, the other two model parameters, Tℓ and
n, are varied using a Levenberg-Marquardt optimization (Markwardt et. al. 2009)
to match the observed line flux ratios (suitably converted to column density ratios
using Eq. 4.5). Figure 2-1 is an example excitation diagram for galaxy NGC5033.
The model fit converges to Tℓ = 51K, n = 4.65, with a fixed value of Tu = 2000K.
29
Figure 2-1 Excitation Diagram for NGC 5033 and 3c293. The Nu/gu ratiosare normalized with respect to the S(1) transition. The solid redline indicates our model fit to the observed normalized columns,denoted by black points. The error bars on the black points arecomparable to the symbol sizes. The resulting model parametersTu, and n are indicated. The two solid lines demonstrate thetwo discrete temperature fit adopted by Roussel et al. (2007).The blue diamonds show model-predicted ratio values for the(unobserved) S(4)–S(7) H2 rotational lines. The below panel for3c293 shows our model fit to all the MIR rotational lines S(0)–S(7) detected with the IRS-Spitzer
Model predicted ratio values for the unobserved lines S(4)–S(7) are indicated, as are
the two discrete temperature fits chosen by Roussel et al. (2007).
The power law index n in our sample ranges from 3.79–6.39, with an average
value n = 4.84 ± 0.61. Figure 2-2 is a frequency distribution of power law index
required to fit the MIR H2 rotational line fluxes in the SINGS galaxies. The power
law index range derived in our model is comparable to the indices required to fit the
H2 rotational line fluxes for ULIRGs (2.5 < n < 5.0) and Seyfert galaxies (3.4–4.9)
(Zakamska et al. 2010, Pereira-Santaella et al. 2014).
30
Figure 2-2 The frequency distribution of the power law index, n, required tofit the MIR H2 rotational lines. The average value n = 4.84±0.61for the SINGS galaxy sample.
Galaxies with a steep power law index have low warm gas mass fractions, since a
larger quantity of H2 in these systems are at lower temperatures. Molecular gas heated
by shocks and other turbulent energetic phenomena have higher temperatures and
hence, higher warm gas mass fractions than gas in photo-dissociation regions (PDRs)
(Appleton et al. 2015 in prep). The power law index therefore gives information on
the relative importance of gas heating by shocks, photoelectric heating, UV pumping,
etc.
2.4.2 Total H2
With a continuous power law model well reproducing the rotational emission lines
from warm H2, it is natural to consider whether the total molecular gas reservoir could
be probed by suitable extrapolation of our model to lower temperatures. Typically,
31
the entire reservoir of H2 cannot be probed through rotational emission, since the
model loses sensitivity at temperatures far below the first rotational energy state.
Recovering the total molecular content from rotational H2 emission therefore typically
requires an additional free parameter — an extrapolated model lower temperature,
Tℓ — which must be calibrated against known molecular mass estimates. We first
explore the models sensitivity to low temperatures, and then describe the calibration
procedure we adopt.
2.4.2.1 Model Sensitivity Temperature, Ts
The upper energy level of the lowest rotational transition of H2 is 510K — sub-
stantiallly higher than typical kinetic temperatures in molecular regions. Yet, even
at temperatures well below this value, molecules in the high temperature tail of the
energy distribution can yield detectable levels of rotational emission. Below some
limiting temperature, however, increasing the number of cold H2 molecules will in-
crease the implied molecular gas mass, but result in no measureable changes to the
H2 rotational line fluxes.
We define the sensitivity temperature, Ts, as that temperature below which the H2
reservoir is too cold for changes in the amount of gas to lead to measurable changes in
the excitation diagram. To evaluate this in the context of our model, we calculated the
difference between the model-derived and observed column density ratios, Nu/gu
(Nu/gu)S(1).
Note that with the adopted normalization to the S(1) line, a unique excitation curve
exists for any given pair n, Tℓ (∀Tℓ ≤ Ts). We varied these model parameters and
evaluated the quality of the model fit using
χ2 =
m∑
i=1
(
Ri,mod − Ri,obs
σRi,obs
)2
, (2.15)
where R = ln(
Nu/gu
(Nu/gu)S(1)
)
, Ri,mod and Ri,obs are, respectively, the modeled and ob-
32
Figure 2-3 The ∆χ2 value in the model parameter space Tℓ and n forNGC 5033. At temperatures lower than 88 K (the maximumTℓ value for 1σ contour plot), the model yields similar flux ra-tios. Molecules with temperature less than 88 K have negligiblecontribution to H2 rotational line flux in the galaxy NGC 5033.
served flux ratios for the ith transition, with uncertainty σRi,obs, and the summation is
over all independent line flux ratios. We map the χ2 space in Tℓ and with σ values,
following (Avni 1976), via ∆χ2 = χ2 − χ2min (where χ2
min is the minimum χ2 value).
Figure 2-3 shows example ∆χ2 contours for the galaxy NGC 5033. The value of Ts
is determined at the maximum value of lower cutoff temperature Tℓ along the 1σ
contour. For the galaxy NGC 5033, Ts ∼ 88K.
The near-horizontal orientation of the contours indicates little correlation between
n and Tℓ. Since for most of the sample, the contours do not close as you go towards
lower temperatures Tℓ < Ts, any chosen lower cutoff of the power law distribution of
temperatures below Ts remains consistent with the data.
Generally, differences between the model and observed flux ratios decrease as Tℓ
33
decreases, and do not significantly change below ∼80 K. However, in some LINER
and Seyferts systems (e.g. NGC 2798, NGC 3627, and NGC 4579), the contours close,
and χ2 has a defined minimum. Figure 2-4 illustrates this phenomenon in the galaxy
NGC 3627, evaluating χ2 along the ridge line of best-fitting indices n for each Tℓ, and
contrasting this galaxy with the more typical case of an indefinite minimum. Evident
in the fits to NGC 3627 is a single best-fitting lower cutoff temperature of Tℓ ∼ 120K.
In such cases, a continuation of the power-law distribution to temperatures below this
best-fitting value is counter-indicated, as it degrades the fit. This indicates that in
these cases, the bulk of H2 is directly detected via rotational emission in a warm gas
component excess.
To identify galaxies with warm molecular gas excess, we imposed the constraint
∆χ2 = χ220K − χ2
min > 2.3, where χ220K is the χ2 value at 20 K. Figures 2-5 and 2-
6 show the distribution of sensitivity temperature Ts and the corresponding average
confidence contour plot for all SINGS galaxies (excluding in total 8 galaxies, identified
as having a warm gas excess (NGC 2798, NGC 3627, and NGC 4579) and few other
galaxies (e.g. NGC 1266, NGC 1291, NGC 1316, NGC 4125, and NGC 5195) with
low signal-to-noise (S/N) ratio in their S(0) line flux). The excluded galaxies are
all LINERs or Seyferts with low S(0)/S(1) ratios (in the range 0.04–0.2, vs. median
0.33 in the SINGS sample, Roussel et al. (2007)). They exhibit evidence for shocks,
unusually high dust temperatures, and merger morphologies — all processes that
could result in substantial quantities of warm molecular gas (Alatalo et. al. 2011,
Skibba et. al. 2011, Pellegrini et. al. 2013, Horellou et. al. 2001, Wilson et. al.
2013, Kohno et.al. 2002, Mentuch et. al. 2012).
The average value of Ts is 81K. This implies that, with the quality of H2 rotational
spectroscopy provided by the Spitzer/IRS instrument, H2 gas down to rotational
temperatures of ∼80K can be reliably detected.
34
Figure 2-4 The χ2 distribution for a normal galaxy NGC 5033 (blue) andan excess warm molecular gas galaxy NGC 3627 (red). The min-imum χ2 occurs at a cutoff temperature of 120 K for NGC 3627however, for the galaxy NGC 5033 the χ2 curve decreasesthroughout as the lower temperature is decreased. For tempera-tures less than 50 K the model show no change in the flux ratiossince temperatures are too cold to occupy rotational energy levelsand hence, a constant χ2 curve.
35
Figure 2-5 The distribution of sensitivity temperatures Ts (the temperatureabove which H2 can be directly traced via rotational emission) forthe SINGS sample, omitting warm excess sources. On average,the sensitivity temperature is 81K.
36
Figure 2-6 The average ∆χ2 contours of the SINGS sample, excluding warmexcess galaxies. The average value is Ts = 81K with an av-erage power law index of n = 4.8. Using these parameters,MH2(> 81K)/MH2,total = 15% of the H2 mass is probed directlyvia rotational emission (§ 2.4.2.3).
37
2.4.2.2 Model extrapolated lower temperature, Tℓ
Given our model sensitivity to H2 columns down to rotational temperatures ∼80K,
we evaluate the potential for extrapolating the power law distribution to lower tem-
peratures to attempt to recover the the total molecular gas mass.
To calibrate an extrapolated lower temperature cutoff – Tℓ — we require a set of
galaxies with well-established molecular gas masses obtained from LCO. We began
constructing a training sample using SINGS galaxies, omitting galaxies with a central
warm excess (see § 2.4.2.1). The ULIRGs and radio galaxies are avoided due to
ambiguity in their αCO values. Recent studies have also shown αCO in low metallicity
galaxies is higher than the Galactic value (Bolatto et al. 2013, Schruba et al. 2012,
Leroy et. al. 2011) and hence we restricted our sample to galaxies with an oxygen
abundance value of 12 + log[O/H] ≥ 8.4. We used the average of their characteristic
metallicities derived from a theoretical calibration (KK04 values) and an empirical
calibration (PT05 values).
To perform this calibration, a given model fitted to the rotational excitation dia-
gram is extrapolated to the particular temperature where the model mass equals the
cold molecular gas mass, estimated from CO emission (LCO), using αCO,Gal (column
10 of Table 2.2).
Figure 2-7 shows the frequency distribution of extrapolated model lower temper-
ature, Tℓ, required to fit the molecular gas mass in the SINGS sample (with and
without warm excess sources). The full sample is described by an average extrapola-
tion temperature of T ⋆ℓ = 49± 9K. Excluding galaxies with warm gas excess narrows
the distribution, but does not significantly change the average value of Tℓ. Table 2.4
lists the value of Tℓ, with the value in parentheses calculated when the conversion
factor derived by Sandstrom et. al. (2013) through dust emission is assumed.
38
Figure 2-7 The frequency distribution of model lower temperature, Tℓ, re-quired to fit the molecular gas mass using the Galactic αCO,Gal.On average the value for Tℓ is 49 K for our SINGS training sam-ple. The below histogram (in blue) is plotted excluding warmgalaxies with high Ts. However, no change in the average valueof Tℓ is found including or excluding galaxies with warm gas ex-cess.
39
Figure 2-8 The distribution dM/dT vs molecular temperature for galaxyNGC 5033. The darker and lighter shades in the plot show twodifferent regions, below and above the sensitivity temperature,Ts = 88 K. The fraction of H2 gas mass below Ts = 88 K for thegalaxy NGC 5033 is about 86%.
2.4.2.3 Mass distribution function
Since dM/dT is a strong negative power law, the number of molecules decreases
rapidly with increasing temperature — most of the H2 gas mass resides at the lowest
temperatures. We can evaluate the fraction of the total molecular mass our model is
sensitive to by comparing M(T = Tℓ → Ts) and M(T = Ts → Tu)
Figure 2-8 shows an example mass-temperature distribution dM/dT for NGC 5033
as a function of molecular temperature. The total molecular gas mass of 1.9 ×108
M⊙ is distributed with an index n = 4.65 in the temperature range Tℓ = 51K to
Tu = 2000K. The fraction of cold H2 gas mass (< TS = 88K for NGC 5033) is
∼ 86%.
Using the average power-law index and sensitivity temperature from the SINGS
40
training sample (see Fig. 2-6) we find MH2(> Ts = 81K)/MH2,total ∼ 15%. Adopting
instead the 2σ contour, the value of Ts becomes 97K and the ratio MH2(> Ts =
97K)/MH2,total is ∼ 8%. This demonstrates that, despite the many shortcoming of
H2 as a rotational emitter, we can trace a substantial fraction of the H2 in galaxies
directly.
Our model using the power law continuous distribution fits the H2 MIR rotational
lines and estimate the warm molecular gas content. The discrete temperature fit
using two or three temperature components is a less robust technique to estimate
the warm molecular gas mass in galaxies. Furthermore, for the first time using the
extrapolation method we could estimate the total molecular gas content in galaxies,
independent of any indirect molecular gas tracers.
2.5 Results, Discussions, & Applications
2.5.1 What are the typical molecular gas temperatures in
galaxies?
UV pumping sets the level populations of upper level states, J > 3, in particular
at low column densities of H2 . As a result, these states may result in rotational tem-
peratures well in excess of their kinetic temperatures. For our purposes whatever
combination of collisional (including shocks) and UV pumping determines the level
populations and hence average H2 excitation, we assume a single power law distribu-
tion of rotational temperature describes the ensemble of molecules. Since the mass in
the power law model is dominated by gas with the lowest rotational temperatures and
hence lowest excitations, it is instructive to compare the derived lower temperature
cutoffs (§4.2.2) with various models and measurements of temperature in molecular
gas.
41
The average lower extrapolation cutoff, T ⋆ℓ = 49K, is comparatively higher than
what is typically assumed for molecular regions (∼10–30K). And yet, this high lower
cutoff is essential in describing the portion of molecular material following a power-law
distribution of temperatures. For example, if we forced the extrapolation temperature
to a lower value of 20K while retaining the average power law index n = 4.8 (§ 4.1.2),
the estimated molecular gas mass would be on average ∼ 30 times higher than the
mass measured using LCO.
To calculate the total molecular gas mass, we extrapolate using a single power
law index. The possibility of a broken or non-power-law temperature distribution at
temperatures lower than our sensitivity temperature cannot be excluded (and indeed
in warm gas excess sources, will be required to explain ongoing star formation).
The ratio of warm diffuse to cold dense molecular gas in galaxies is a key step
in understanding the H2 temperature distribution. Assuming an incident radiation
field G0 = 10 for a cloud with mass 105–106 M⊙, Wolfire et al. (2010) estimated
that molecules in the outer diffuse region (AV < 1) obtain temperatures in the range
50–80K, with stronger incident radiation fields leading to even higher temperatures.
Due to the differences in self shielding from dissociating radiation, molecular clouds
have an outer layer of varying thickness which contains molecular hydrogen, but little
or no CO. This gas cannot be traced through CO transitions and is hence called dark
molecular gas (Wolfire et al. (2010)). This dark gas can account for a significant
fraction (24–55%) of the total molecular gas mass in galaxies (Pineda et. al. 2013,
Wolfire et. al. 2010, Smith et. al.2014, Planck Collaboration 2011).
CO-dark and diffuse molecular gas is heated to higher temperatures than dense
CO-emitting gas in molecular cloud cores. Indeed, individual molecular cloud simula-
tions find average mass-weighted H2 temperatures of ∼45K for typical cloud masses
and radiation (Glover et al. 2012a, Glover, priv comm.). Galaxy-scale hydrodynamic
simulations with a full molecular chemistry network Smith et al. 92014) also show
42
that H2 gas in the CO-emitting regions is markedly biased to the low temperature
end of the full temperature distribution (Glover, priv comm.). Also, in diffuse and
translucent molecular clouds in the Galaxy, Ingalls et al. (2011) demonstrated that
additional sources of heating are required to explain the observed H2 and atomic
cooling-line power. Our extrapolated power law model traces both warm and cold
molecular gas mass in galaxies. The common assumption that the bulk of molecular
gas is found at rotational temperatures of ∼10–30 K may be true only in the CO
emitting core regions of the molecular cloud. The average mass-weighted molecular
gas temperature can be much higher when the full range of emitting environments are
considered. It is therefore not surprising to find typical molecular gas temperatures
of ∼ 50K in galaxies, similar to our model derived values.
2.5.2 Estimating total MH2
By fitting a continuous temperature distribution to the MIR H2 rotational lines
and calibrating an extrapolating temperature, this model can be used to calculate the
total molecular mass in galaxies directly from H2 rotational emission. The method is
independent of any indirect tracer like CO, DGR, or assumptions about star formation
depletion timescales. Any physical processes which bias these tracers will not impact
results derived from fitting H2. In this section we test the model’s capability to
estimate molecular gas mass in different types of galaxies.
Adopting a fixed model lower extrapolation temperature T⋆l = 49 K the total
H2 gas mass is calculated by extrapolating the fitted model. The total H2 gas mass
derived by our model is shown in Figure 2-9. It compares very well with LCO along
with αCO,Gal based estimates of gas mass. The scatter in our model is about 0.31
dex (factor of 2) for the SINGS sample and increases to 0.34 dex (factor of 2.2) for
the complete sample, including U/LIRGs and radio galaxies. Some of this scatter no
doubt arises from ignorance of the true αCO values in these systems.
43
Figure 2-9 The model extrapolated molecular gas mass, assuming a lowercutoff temperature of 49K vs the total molecular gas mass ob-tained from LCO measurements using αCO,Gal. Different symbolsrepresent different galactic systems as indicated. The solid lineis one to one correspondence while the dashed and dotted linesare twice and 3× the value respectively.
44
Galaxies with warm gas excess and a high sensitivity temperature Ts (§ 2.4.2.1),
require similar values of extrapolated Tℓ as the training sample. The warm molecular
gas traced by MIR rotational lines may be completely isolated from the cooler gas
in these galaxies. Extrapolating to a lower temperature of 49 K yields molecular gas
masses which agree with CO-derived values, assuming αCO = αCO,Gal. A possibility of
enhanced CO excitation due to high molecular gas temperature with a corresponding
low αCO in such warm galaxies cannot be ruled out (see also § 2.5.3).
Taken together, a correlation between H2-derived and CO-based mass estimates
is found, spanning seven orders of magnitude in mass scale and across a wide range
of galaxy types. The mass calculated with a continuous temperature distribution
model, extrapolated to a fixed cutoff temperature of 49K, can provide an independent
measurement of total molecular gas mass in galaxies, good to within a factor of 2.2
(=0.34 dex). This dispersion is comparable to uncertainties in the αCO conversion
factor itself (Bolatto et al. (2013)) as well as methods using DGR (Sandstrom et al.
(2013)).
2.5.3 Model derived molecular gas mass in ULIRGs, LIRGs
and radio galaxies
Many local ULIRGs are recent ongoing galaxy mergers. In the merging process a
large amount of gas in the spiral disk is driven to the central nuclear region, increasing
the gas temperature. The increase in temperature and turbulence increases the CO
linewidth, resulting in a high value of LCO for a given molecular gas mass. αCO,Gal
therefore gives an overestimate of H2 gas masses in these galaxies (Downes et al. 1993,
Bryant et al. 1999). Moreover, αCO,Gal can yield molecular gas masses greater than
the observed dynamical mass (Solomon et al. 1997). To avoid this, a lower value of
conversion factor is suggested for ULIRGs and other merger systems — αCO = 0.8
45
M⊙(K km s−1 pc2)−1, 5.5× lower than the standard Galactic value (Downes et al.
(1998)). However, by considering the high-J CO ladder, some studies have suggested
that even for ULIRGs αCO,Gal values are possible (Papadopoulos et al. 2012). Some
H2 emission may lie outside photo-dissociation and star forming regions in ULIRGs,
powered by shocks (Zakamska et al. 2010). This H2 may reside in CO-dark gas, so
that applying a low αCO value in ULIRGs may yield an underestimate H2 mass.
In radio galaxies molecular gas can be predominantly heated by shocks through
powerful jets. The molecular gas clouds may be affected by turbulence, and not
gravitationally bound, defying the use of standard αCO,Gal. Ogle et al. (2014) using
DGR in NGC 4258, a low luminosity AGN (LLAGN) harboring a jet along the disk,
derived gas mass of about 108 M⊙, an order of magnitude lower compared to the
standard method of using αCO,Gal. The molecular gas mass could be overestimated
when used αCO,Gal in radio galaxies, which harbor long collimated powerful jets.
In applying a power law model to the sample of ULIRGs, LIRGs and radio galaxies,
the nominal model extrapolation temperature T ⋆ℓ = 49K is used to calculate the total
H2 gas masses, listed in column 3 of Table 2.5. The Tℓ in column 4 of Table 2.5 is the
required extrapolated temperature to match the cold molecular gas mass measured
from the CO line intensity.
In radio galaxies 3c424, 3c433, cen A, and 3c236 the estimated H2 gas mass using
the power law model after extrapolation to T ⋆ℓ = 49 K, is higher when compared to
the CO luminosity derived values using αCO,Gal. However, when accounted for the
intrinsic variation in αCO,Gal and T ⋆ℓ the H2 gas masses are in agreement with each
other except in 3c424, where the difference in masses is more than 10×.
Using αCO = 0.8 M⊙(K km s−1 pc2)−1, a factor of 5.5× lower than the standard
Galactic value, we derive a modified extrapolation temperature T ′ℓ , which is required
to match the lowered gas mass. Since the model mass rises rapidly to lower temper-
atures, T ′ℓ > Tℓ. The T ′
ℓ values are listed in column 7 of Table 2.5. On average for
46
ULIRGs, T ′ℓ = 80 ± 13K.
Figure 2-10 shows the temperature distribution of Tℓ (adopting αCO,Gal) and T ′ℓ
(adopting αCO,Gal/5.5) in ULIRGs. The lower temperature cutoff in ULIRGs and ra-
dio galaxies (except 3c424) is very similar to the normal galaxy training sample when
αCO,Gal is adopted, but much higher with the reduced molecular mass of αCO,ULIRG.
It is of interest that when αCO,Gal is used with the nominal calibration cutoff
temperature T ⋆ℓ , the ULIRG sample in Fig. 2-9 does not exhibit any particular bias.
Either αCO and Tℓ are similar to their normal Galactic values in these systems, or
they have reduced αCO and a higher H2 temperature floor. This may indicate that the
same physical processes that lead to reduced αCO in highly active system, including
increased ISM pressure and radiation density, globally elevate the gas temperature.
Since ULIRGs could indeed have uniformly elevated molecular gas temperatures,
this degeneracy between αCO decrease and Tℓ increase leads to a systematic uncer-
tainty in the total H2 gas mass identical in form to that obtained directly from mass
estimates based on LCO. A suggested prescription which side-steps this ambiguity, in
applying this model to systems with non-Galactic αCO is to calculate the total gas
mass using the the nominal T ⋆ℓ = 49K, and scale it by αCO/αCO,Gal, for the preferred
αCO. As can be seen in Fig. 2-9, which adopts a uniform αCO,Gal, the molecular mass
in ULIRGs is well recovered by this procedure.
2.5.4 Effect of dust temperature on the warm H2 fraction
Since typical sensitivity temperatures are Ts ∼ 80K (see § 2.4.2.1), we can directly
calculate the warm H2 mass above ∼ 100K without extrapoloation, and compare it
to the total mass as recovered by the calibrated extrapolation of § 2.4.2.2. Given an
estimate for the power law index, n, and lower cut off temperature, Tℓ, of the power
law distribution, we can calculate the fraction of molecular gas mass at temperatures
47
Figure 2-10 The distribution of model extrapolated lower temperatures inULIRG and radio galaxies when the H2 gas mass is evaluatedusing the Galactic conversion factor αCO,Gal (above, red), andwhen adopting αCO,Gal/5.5, as generally accepted for ULIRGs(below, blue). The mean lower temperature cutoff is 50K and80,K when αCO,Gal and αCO = αCO,Gal/5.5 are used, respec-tively. The Tℓ distribution for the SINGS normal galaxy sampleis shown above, for comparison.
48
above 100K as
M(> 100 K)
Mtotal=
∫ Tu
100KT−ndT
∫ Tu
TℓT−ndT
, (2.16)
where Mtotal is the total molecular gas mass estimated from the CO line intensity.
Assuming 100 K ≪ Tu and Tℓ ≪ Tu we find
M(> 100 K)
Mtotal≈
(
100 K
Tℓ
)1−n
. (2.17)
Table 2.4 lists the calculated mass fraction M(>100K)Mtotal
for each galaxy. Columns 6
and 8 of Table 2.5 are calculated using αCO,Gal and 15.5
×αCO,Gal as generally assumed
for ULIRGs, respectively. At lower αCO, the warm gas mass fraction is higher.
Figure 2-11 shows the fraction of warm molecular gas mass, M(>100K)Mtotal
, as a function
of far infrared (FIR) dust color temperature, νfν70
νfν160. The warm gas fraction obtained
using αCO,Gal ranges from 2–30%, and exhibits little correlation among different galaxy
types with dust color temperature. The ULIRGs and normal star forming galaxies
show similar warm gas fractions, though ULIRGs have warmer dust color tempera-
tures, and LINERs and Seyferts have somewhat higher warm gas mass fractions than
normal star forming galaxies on average.
In contrast, using the available dust-derived central αCO estimates of Sandstrom
et. al. (2013) for normal galaxies and a reduced value 15.5
× αCO,Gal for ULIRGs
and QSO’s, however, leads to a strong correlation, with warmer dust implying an
increasing warm H2 fraction. The average warm gas fraction for ULIRGs and QSO is
then ∼45%, significantly above that of normal galaxies, and on the same increasing
trend with dust temperature. Depending on which prescription for total gas mass is
correct, this could indicate a dependence of αCO on temperature.
49
Figure 2-11 The fraction of warm H2 gas mass (T > 100K) versus the dust
color temperature, νfν,70
νfν,160, obtained from PACS. At left a con-
sistent Galactic αCO is adopted, and little trend is seen. Atright, a reduced αCO is applied to ULIRGs and QSO’s, and thedust-derived central αCO values of Sandstrom et al. (2013) areused. Galaxies with warmer dust color temperatures have highwarm molecular gas mass fraction.
50
2.5.5 Molecular gas in low metallicity galaxies
As traced by their CO emission, many low metallicity dwarfs appear to have van-
ishingly low molecular gas content, but retain high star formation rates. That is, they
are strong outliers on the Schmidt-Kennicutt relation (Galametz et al. 2009, Schruba
et al. 2012). This discrepancy implies either that in dwarf galaxies star formation
efficiencies are higher compared to normal spirals, or they host large molecular gas
reservoirs than is traced by the CO emission (Schruba et al. 2012, Glover et al.
2012b). It is possible that a significant fraction of H2 exists outside the CO region,
where the carbon is in C+ (ionized) or C0 (neutral) states. Since H2 can self-shield
from UV photons in regions where CO is photodissociated (Wolfire et. al. 2010), at
low metallicity, not only is the CO abundance reduced, but as dust opacity is reduced
and ionized regions become hotter and more porous, αCO,Gal can severely underes-
timate the molecular gas mass. Considerable effort has been invested in detecting
and interpreting CO emission at metallicity 50 times lower than the solar metallicity,
∼ 7.0, to assess the molecular gas content (Leroy et al. 2011, Schruba et al. 2012,
Cormier et al. 2014, Remy et al. 2014). Dust emission can be used to estimate the
molecular gas mass in ISM, assuming a constant DGR however, it is essential to know
the change in DGR with metallicity. At very low metallicities, ≤ 8, DGR appears to
scale non-linearly with metallicity (Herrera et al. 2012, Remy et al. 2014).
Our direct detection of H2 gas mass through H2 rotational lines is independent of
any indirect tracers, which are affected by changing metallicity and local radiation
effects. Applying our model in low metallicity galaxies should yield an estimate of
molecular gas mass without the same inherent biases introduced by these dust and
CO abundance variations. In this section we estimate the H2 gas masses through
our power law model in a low metallicity galaxy sample selected to have detected
H2 rotational emission, faint CO detection, and (where available) estimates of dust
51
mass. We then compare these H2 -based estimates to models and other methods
which attempt to control for the biases introduced at low metallicity.
The low metallicity galaxies were selected on availability of MIR H2 rotational
lines to have atleast three rotational lines including S(0) or S(1) line fluxes along
with CO derived molecular mass estimates.
2.5.5.1 Metallicity estimation
To study the variation of H2 gas mass from CO derived measurements over the
metallicity range, it is essential to estimate the metallicity of galaxies. The metal-
licities were determined applying the direct Te method. CGCG007-025 is the lowest
metallicity galaxy in our dwarf sample with the value of = 7.77 (Izotov et al. 2007).
Guseva et al. (2012) estimated the value of in the two H II regions, Haro 11B and
Haro 11C as 8.1 and 8.33, respectively hence, we adopt the average value for Haro
11. The metallicity value, , for NGC 6822 is 8.2 (Israel et al. 1997). No litera-
ture value exist for the oxygen gas phase abundance for the specific region of Hubble
V in NGC 6822, mapped by the IRS-Spitzer. Peimbert et al. (2005) estimated =
8.42±0.06 for Hubble V, which is inconsistent with the previous value of 8.2.
For selected SINGS galaxies, the metallicity values in the circumnuclear regions,
which are approximately the size of our H2 line flux extracted regions, are estimated
by averaging the theoretical (KK04) and an empirical metallicity calibration (PTO5)
as recommended by (Moustakas et al. 2010).
2.5.5.2 Cold molecular gas from CO line emission
Although the CO abundance drops super-lineraly with decreasing metallicity, it
is detected in the low metallicity sample, and as the most common molecular tracer,
can be compared directly to our H2- based method. We adopted the literature val-
ues for 12CO(1–0) line intensities for CGCG007-025 and N66, while for Haro 11,
52
UM311 and Hubble V region 12CO(3–2) line intensities were scaled using the rela-
tionICO(3−2)
ICO(1−0)= 0.60 (in temperature units) due to unavailability of 12CO(1–0) line
intensities. Calculating LCO, (area integrated luminosity) the molecular gas masses
are estimated using αCO,Gal and were further scaled using the 8 µm map, to account
for the difference in the extracted IRS spectrum and the CO beam regions for each
galaxy. The molecular gas masses are listed in Table 2.6.
2.5.5.3 Molecular gas from dust emission
An alternative method for estimating molecular gas mass makes use of dust emis-
sion together with assumption of dust opacity and grain size distribution to calculate
a total mass, scaling dust mass to the total gas mass using a presumed or modeled
dust-to-gas ratio, and removing the measured atomic mass from the region.
Leroy et al. (2007) estimated H2 gas surface density using dust emission from
FIR map in N66 region of SMC. They derived αCO to be about 27 times the αCO,Gal.
For Haro 11, UM311 and Hubble V using metallicity-DGR relation from Sandstrom
et al. (2013) and with the known dust mass, we estimated the total gas mass and
after subtracting the atomic gas content subsequently calculated the molecular gas
mass. However, we find a negative value for the molecular gas for UM311, suggesting
the ISM mass is mainly dominated by the atomic gas. The H2 gas masses calculated
from the dust emission are given in Table 2.6.
2.5.5.4 Molecular gas mass using our model
The H2 line flux extraction was performed for the similar region, where the CO
emission was measured by Rubio et al. (1996). The cubes were prepared using
CUBISM (Smith et al. 2007b) (Jameson, K. et al. in prep) and the H2 line fluxes
were estimated using the PAHFIT, a MIR spectral decomposition tool (Smith et
al. 2007a). The flux of H2 rotational lines for Haro 11 are from Cormier et al.
53
Figure 2-12 Excitation diagram for low metallicity galaxy Haro11. TheNu/gu ratios are normalized with respect to S(1) transition.The dashed red line indicate the model fit to the observed ra-tios. The blue dashed line predicts the value of S(0) flux ratio.The model estimated Tℓ, and n are mentioned.
(2014), while for UM 311, and CGCG 0077-025. Hunt et al. (2010) measured the H2
rotational line fluxes. For Hubble V region, Roussel et al. (2007) derived the flux of
H2 rotational lines.
Figure 2-12 is an excitation diagram for low metallcity galaxy Haro 11, with the
power law model fit. Our model prediction for the unobserved S(0) line is included,
and is consistent with the estimated upper limit. The H2 total gas mass for each low
metallicity galaxy is measured using the power law model with lower temperature
extrapolation to T⋆ℓ = 49 K. Table 5 lists the value of metallicity with their distance
and the measured value of H2 rotational line fluxes with the power law index and the
H2 gas mass derived using CO, dust, and our model.
Figure 2-13 compiles H2 gas masses derived using the various indirect tracers,
54
together with the results from our H2-only model. All values are shown relative to
the H2 mass inferred using CO luminosities with αCO = αCO,Gal, and are plotted as
a function of metallicity. For the SINGS sample, a variation of about 2–3 times the
LCO-derived H2 gas mass is found, likely a consequence of the intrinsic variation in
αCO at high metallicity & 8.4. At intermediate metallicity, the molecular gas content
from CO line emission for the Hubble V region in NGC 6822 compares well with our
model-derived gas mass, which suggests a similar Galactic conversion factor (αCO =
αCO,Gal), consistent with the results of Rijcke et al. (2006). At lower metallicity,
however, our model disagrees strongly with naive CO-based estimates, yielding up to
∼ 100× the molecular gas mass inferred from CO emission.
The molecular masses we recover at low metallicity are in good agreement with
other measures which attempt to account for the impact of reduced metal abundance.
These include dust-derived measurements, where available (when the strong metal-
licity dependence of the dust-to-gass ratio is accounted for). They also agree well
with the prescription for the power-law like αCO variations with metallicity recov-
ered from inverting star formation densities among all non-starburst galaxies in the
HERACLES sample (Schruba et al. 2012). The theoretical model of varying αCO by
Wolfire et al. (2010) also agrees well, assuming an H2 column density of 1022 cm−2
in a molecular cloud, scaled accordingly to the solar metallicity value of 8.66 with
log(G0/n) = -0.3 adopted from the luminosity ratio LOI
LFIRfrom Malhotra et al. (2001).
Recent work by Shi et al. (2015) derived the conversion factor in Sextans A, with a
metallicity of 7% of the Milky Way, to be about 700× the galactic conversion factor.
This value is in good agreement with our analysis on the relation between conversion
factor with metallicity (Figure 2-13). The power law model reliably recovers the total
molecular gas masses at metallicity as low as 10% of the Milky Way, where CO and
other indirect tracers suffer strong and non linear biases.
55
Figure 2-13 The ratio of molecular gas masses estimated using differentmethods to the “naive mass” obtained using LCO (and αCO
= αCO,Gal), as a function of gas-phase metallicity. The blackpoints show the molecular gas masses derived from the H2 rota-tional model. The blue line shows a fit to this ratio derived frominverting the star formation law for HERACLES non-starburstgalaxies Schruba et al. (2012). The molecular gas masses tracedby the dust emission are denoted by the red points in the plot,and are shifted slightly in their metallicity values for clarity.The black solid line is the predicted mass ratio from the theo-retical model of Wolfire et al. (2010), assuming N(H2) = 1022
cm−2 and log(G0/n) = -0.3. A steep increase in the ratio ofmodel derived H2 gas mass to the LCO derived measurements isobserved at metallicity values ≤ 8.4.
56
2.5.6 Future prospects
At metallicities . 0.25 Z⊙, the direct power-law method recovers total molecular
gas content as reliably as other tracers that account for or avoid the impact of reduced
gas-phase metal content. At even lower metallicities, the CO abundance plummets,
with essentially all of the molecular gas in a CO-dark phase. In the first epoch of the
star formation in the universe, the extremely low abundance of heavy elements leaves
H2 as a principal coolant (Lepp et al. 2002).
The Mid Infrared Instrument (MIRI) onboard JWST is sensitive enough to detect
the S(1), S(2), S(3) and higher rotational lines of H2 in luminous galaxies (U/LIRGs)
till redshift 0.6, 1.3, 1.9, and higher, respectively. Assuming averageLH2S(1)
LIR= 10−4
(Bonato et al. 2015), and LIR = 3×1011 L⊙ (typical for LIRGs), for S(1) at z =
0.5 to have signal to noise S/N = 5 will require 30 minutes of integration time with
JWST-MIRI. It will be possible to measure pure H2 rotational lines at high red-
shifts of z ≈ 6–7, almost reaching the reionization era of universe, with the SPace
Infrared telescope for Cosmology and Astrophysics (SPICA) and the Cryogenic Aper-
ture Large Infrared Submillimeter Telescope Observatory (CALISTO), planned for
the 2020 decade (Roelfsema et al. 2012, Bradford et al. 2015).
The above mentioned future projects for the next decade provides an opportunity
to observe H2 rotational lines at high redshifts. The power law model can be an useful
tool in estimating molecular gas mass and study its variation and consequences at
different redshifts.
2.6 Summary
We present a new power law temperature distribution model capable of reliably
estimating the total molecular gas mass in galaxies purely from as few as three mid-
infrared pure-rotational H2 emission lines. Our model is independent of the biases
57
affecting indirect tracers like CO, dust emission, or star formation prescriptions. It
can hence be used even in environments where reliability of those indirect tracers is
poor, such as at low metallicity. We calibrate the model on a sample of local star-
forming galaxies with well-quantified CO-based molecular gas masses, and apply the
model to local ULIRGs and low metallicity systems. Our key results are:
1. A model based on a continuous power law distribution of rotational tempera-
tures well reproduces the H2 excitation from a wide range of galaxy types, and
can directly recover the warm H2 mass (T > 100K) more reproducibly than
arbitrary discrete temperature fits.
2. The power law index obtained for all SINGS galaxies ranges from 3.79 - 6.4,
with an average value of 4.84.
3. With typical Spitzer detection sensitivities, the model can directly recover the
H2 gas mass down to a limiting sensitivity temperature of Ts = 81K (when the
S(0) line is available), accounting for ∼ 15% of the total molecular gas mass.
4. By calibrating the model using a subset of the SINGS sample with well deter-
mined CO-based molecular masses, we find that extrapolating the H2 tempera-
ture distribution to a single temperature of Tℓ = 49K recovers the total H2 gas
mass within a factor of 2.2 (0.34 dex).
5. When αCO,Gal is used, the fraction of warm molecular gas mass (M> 100 K)
in this training sample ranges from 0.02 to 0.30. If a reduced αCO is adopted
for the ULIRGs, this fraction increases with increasing dust color temperature
νfν(70)/νfν(160).
6. In ULIRGs, the total molecular gas mass obtained by extrapolating the model
to T ⋆ℓ = 49K is consistent with the molecular gas mass derived using αCO,Gal.
58
Alternatively, if a reduced αCO is adopted, the model extrapolation tempera-
ture required rises to 80 K. Either the warm molecular gas fraction and lower
temperature cutoff in ULIRGs is higher than in normal star forming galaxies,
or αCO is closer to the Galactic value than has been presumed.
7. At low metallicity (12 + log[O/H] . 8.4), where indirect tracers of molecular
mass suffer increasingly strong and non-linear biases, and the mass of the molec-
ular reservoirs exceed their CO-derived estimates by 100× or more, the direct
power law model recovers the total molecular gas masses reliably, as assessed
by other methods which correct for metallicity.
8. With upcoming facilities including JWST, SPICA, and CALISTO, detection
of the relevant H2 rotational lines in galaxies at intermediate to high redshifts
becomes possible, opening a new window on the fueling history of star formation
in the Universe.
59
Table 2.1. Observed properties of our sample galaxies
Galaxy D Type LIRS70
S160Ref
Name (Mpc) 1010 L⊙
(1) (2) (3) (4) (5) (6)
N337 19.3 SF 0.568 0.807 S07N1097 14.2 SF 2.343 0.932 S07N1266 31.0 LIN 1.333 1.420 S07N1291 10.4 LIN 0.039 1.382 S07N1316 21.0 LIN 0.260 1.368 S07N1482 22.6 SF 2.812 1.169 S07N1566 20.4 SY 0.444 0.772 S07N2798 25.8 SF 2.316 0.407 S07N2976 3.6 DW 0.007 0.564 S07N3049 19.2 SF 0.307 1.107 S07N3184 11.4 SF 0.026 0.604 S07N3190 19.3 LIN 0.290 0.577 S07N3198 14.1 SF 0.147 0.813 S07N3265 19.6 SF 0.237 1.057 S07Mrk33 22.9 DW · · · 1.525 · · ·N3351 9.3 SF 0.268 0.924 S07N3521 11.2 LIN 0.266 0.572 S07N3627 6.4 SY 0.221 1.016 S07N3938 14.3 SF 0.105 0.436 S07N4125 23.9 LIN 0.048 0.975 S07N4254 14.4 SF 0.533 0.602 S07N4321 14.3 SF 0.513 0.671 S07N4450 16.5 LIN 0.092 1.116 S07N4536 14.5 SF 0.905 1.172 S07N4559 7.00 SF 0.083 0.498 S07N4569 16.8 LIN 0.426 0.777 S07N4579 16.4 SY 0.161 0.857 S07N4625 7.6 DW 0.023 0.518 S07N4631 9.3 SF 0.306 0.922 S07N4725 20.5 SY 0.034 0.564 S07N4736 4.7 LIN 0.115 1.708 S07
60
Table 2.1 (cont’d)
Galaxy D Type LIRS70
S160
Ref
Name (Mpc) 1010 L⊙
(1) (2) (3) (4) (5) (6)
N4826 5.3 LIN 0.116 0.782 S07N5033 14.8 SY 0.577 0.572 S07N5055 7.9 LIN 0.159 0.557 S07N5194 7.6 SY 0.308 0.686 S07N5195 7.6 LIN 0.154 1.652 S07N5713 21.4 SF 2.998 1.001 S07N5866 15.3 LIN 0.215 0.981 S07
N6822A 0.5 DW — 1.151 –N6946 6.8 SF 0.387 0.958 S07N7331 14.5 LIN 0.249 0.564 S07N7552 21.0 SF 5.052 1.244 S07N7793 3.9 DW 0.009 0.648 S073c31 66.7 3C — — —3c218 240 3C — — —
3c272.1 19.1 3C — — —3c293 195 3C — — —3c310 233 3C — — —3c326n 395 3C — — —3c424 568 3C — — —3c433 445 3C — —- —3c436 1016 3C —- — —CenA 11.0 3C — — —3c236 449 3C — — —
Arp220 77.6 ULI 145 1.678 H06IRAS 00188-0856 596 ULI 256 — H06IRAS 03521+0028 717 ULI 365 — H06IRAS 05189-2524 186 ULI 143 2.292 H06IRAS 06035-7102 356 ULI 166 — H06IRAS 06206-6315 418 ULI 169 — H06IRAS 07598+6508 700 ULI 337 — H06IRAS 08572+3915 258 ULI 137 3.124 H06
61
Table 2.1 (cont’d)
Galaxy D Type LIRS70
S160Ref
Name (Mpc) 1010 L⊙
(1) (2) (3) (4) (5) (6)
IRAS 10565+2448 188 ULI 109 1.362 H06F12112+0305 324 ULI 212 1.576 H06
IRAS 13451+1232 563 ULI 197 — H06F14348-1447 332 ULI 224 1.550 H06
IRAS 17208-0014 188 ULI 250 1.744 H06IRAS 19254-7245 273 ULI 121 1.468 H06IRAS 20087-0308 483 ULI 280 — H06IRAS 23365+3604 286 ULI 147 1.557 H06
Mrk 273 164 ULI 142 1.899 H06UGC5101 174 ULI 100 1.053 H06Mrk 463E 221 ULI 60 — H06
PG1440+356 347 QSO 42 1.355 E01N6240 101 LIR 69 1.415 H06N3110 72.5 LIR 16 0.772 P10N3256 40.1 LIR 40 1.577 P10N3690 44.9 LIR 63 1.663 P10N5135 58.8 LIR 16 1.096 P10N6701 56.2 LIR 10 0.883 P10N7130 69.9 LIR 25 1.192 P10N7591 70.9 LIR 10 0.821 P10N7771 62.1 LIR 25 0.773 P10
aRef: S07 - Smith et al. 2007; K09 - Kennicutt et al. 2009; NED - NASA/IPAC ExtragalacticDatabase; H06 - Higdon et al. 2006; P10 - Pereira Santaella et al. 2010; E01 - Evans et al.2001
bEntries in the last column are references for the the infrared luminosity (8-1000 ), LIR. ForSINGS galaxies LIR is calculated using the relation LIR = 0.94 × LTIR (D. Dale priv. comm.)and the corresponding LTIR values are from [?].
cDistance measurements are from the KINGFISH webpage for SINGS galaxies and for othersthrough NED IPAC Extragalactic Database
dThe S70
S160is the FIR ratio calculated from the PACS 70 and 160 fluxes
eThe 3c radio galaxies were selected from Ogle et al. (2010).
62
Table 2.2. Observed molecular hydrogen rotational line flux
Galaxy S(0)(err) S(1)(err) S(2)(err) S(3)(err) S(4)(err) S(5)(err) S(6)(err) S(7)(err) M(H2)Name (106 M⊙)(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
N03371,1 01.16(0.38) 02.10(0.49) 01.54(0.99) 00.97(0.52) · · · · · · · · · · · · 56 (228)N10971,1 21.31(4.26) 72.61(4.33) 29.36(2.94) 42.30(2.34) · · · · · · · · · · · · 1493N126625,1 01.64(0.85) 14.85(0.66) 12.18(0.71) 18.98(1.16) 10.27(1.74) 24.19(1.99) · · · 19.22(2.86) 1253N12911,1 00.37(0.18) 02.98(0.48) 01.39(0.82) 03.16(1.00) · · · · · · · · · · · · 9N13161,1 00.15(0.08) 03.60(0.61) 01.84(0.59) 08.40(0.99) · · · · · · · · · · · · 124N14821,1 10.68(3.51) 42.40(5.34) 18.40(2.28) 20.75(2.69) · · · · · · · · · · · · 1607N156626,1 02.40(0.22) 12.95(0.71) 05.53(0.48) 09.16(1.76) · · · · · · · · · · · · 389N27981,1 04.32(2.53) 20.78(2.21) 09.07(0.92) 10.52(1.21) · · · · · · · · · · · · 273N29761S,1 00.89(0.24) 01.57(0.30) 00.49(0.35) 00.52(0.42) · · · · · · · · · · · · 1.08 (0.25)N30491,1 00.64(0.22) 2.37(0.38) 01.17(0.65) 01.11(0.52) · · · · · · · · · · · · 148N31841S,1 00.98(0.23) 02.26(0.29) 00.63(0.35) · · · · · · · · · · · · · · · 26 (10.6)N31901,1 01.81(0.29) 07.53(0.80) 02.09(0.64) 07.16(1.26) · · · · · · · · · · · · 110N31981,1 01.32(0.49) 03.21(0.54) 01.02(0.35) 01.87(0.85) · · · · · · · · · · · · 31N32651,1 00.93(0.32) 03.06(0.35) 01.14(0.79) 01.53(0.74) · · · · · · · · · · · · 94Mrk331,1 01.39(0.56) 02.93(0.37) 01.15(0.53) 04.57(0.86) · · · · · · · · · · · · 81N33511S,1 06.28(1.05) 21.84(2.06) 08.83(1.01) 16.98(1.99) · · · · · · · · · · · · 197 (32)N35211,1 01.78(0.36) 03.12(0.69) 01.23(0.40) 02.05(0.88) · · · · · · · · · · · · 20N36271S,1 03.12(0.39) 31.86(1.22) 14.17(0.52) 20.91(1.26) · · · · · · · · · · · · 208 (27)N39381S,1 00.80(0.12) 01.34(0.33) 00.42(0.32) · · · · · · · · · · · · · · · 37 (30)
63
Table 2.2 (cont’d)
Galaxy S(0)(err) S(1)(err) S(2)(err) S(3)(err) S(4)(err) S(5)(err) S(6)(err) S(7)(err) M(H2)Name (106 M⊙)(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
N41251,1 00.26(0.16) 01.76(0.47) 01.26(0.64) 01.80(0.88) · · · · · · · · · · · · 26N42541S,1 01.76(0.13) 08.73(0.95) 04.66(0.92) 03.25(0.88) · · · · · · · · · · · · 342 (700)N43211S,1 07.82(1.17) 26.75(2.28) 13.21(2.46) 16.45(1.62) · · · · · · · · · · · · 1386 (200)N44501,1 01.03(0.16) 09.14(0.52) 03.37(0.95) 08.90(1.11) · · · · · · · · · · · · 63N45361S,1 10.87(3.23) 41.16(4.05) 17.56(2.33) 21.68(2.07) · · · · · · · · · · · · 810 (415)N45591,1 01.36(0.21) 01.79(0.35) 00.40(0.15) 02.34(1.37) · · · · · · · · · · · · 18N45691,1 04.16(0.69) 31.61(0.87) 15.24(0.52) 30.94(1.87) · · · · · · · · · · · · 760N45791,1 00.89(0.23) 16.41(0.52) 10.58(1.24) 25.63(1.00) · · · · · · · · · · · · 205N46251S,1 00.80(0.16) 00.96(0.19) 00.54(0.43) · · · · · · · · · · · · · · · 8 (21)N46311,1 05.82(0.64) 12.36(1.61) 06.04(0.50) 04.68(0.91) · · · · · · · · · · · · 73.4N47251S,1 01.17(0.17) 3.71(0.51) 1.99(0.81) 3.80(0.90) · · · · · · · · · · · · <170 (<28)N47361S,1 03.32(0.68) 25.08(1.64) 10.01(1.04) 22.17(1.36) · · · · · · · · · · · · 21 (1.4)N48261,1 07.13(0.65) 34.53(1.58) 15.09(1.09) 21.06(1.03) · · · · · · · · · · · · 65N50331,1 03.66(0.35) 18.20(1.04) 06.35(0.31) 12.69(1.91) · · · · · · · · · · · · 190N50551S,1 04.40(0.24) 15.80(0.91) 05.20(0.67) 08.02(1.06) · · · · · · · · · · · · 87 (20)N51941,1 01.77(0.28) 13.44(0.91) 07.57(0.47) 18.73(1.77) · · · · · · · · · · · · 32N51951,1 05.42(2.09) 31.04(1.23) 12.77(0.65) 27.50(1.72) · · · · · · · · · · · · 80N57131S,1 02.74(0.36) 15.16(1.43) 05.09(0.79) 08.56(0.70) · · · · · · · · · · · · 329 (238)N58661,1 01.55(0.15) 09.00(0.55) 03.91(0.70) 03.72(0.65) · · · · · · · · · · · · 81
64
Table 2.2 (cont’d)
Galaxy S(0)(err) S(1)(err) S(2)(err) S(3)(err) S(4)(err) S(5)(err) S(6)(err) S(7)(err) M(H2)Name (106 M⊙)(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
N6822A1,1 00.81(0.43) 02.11(0.43) 01.60(0.41) 02.59(0.38) · · · · · · · · · · · · 0.03N69461S,1 22.77(6.19) 63.69(3.35) 27.16(1.71) 28.63(2.93) · · · · · · · · · · · · 346 (32)N73311,1 01.64(0.16) 04.96(0.50) 01.38(0.19) 02.34(0.83) · · · · · · · · · · · · 120N75521,1 25.78(10.69) 56.53(4.26) 24.59(2.25) 31.59(1.48) · · · · · · · · · · · · 1950N77931,1 00.70(0.17) 01.50(0.32) 01.17(0.46) 00.53(0.37) · · · · · · · · · · · · 33c0312,8 00.64(0.11) 01.36(0.58) 00.59(0.15) 0.60(0.17) 00.76(0.27) 00.80(0.18) · · · · · · 6953c2182,9 00.55(0.11) 00.41(0.22) 00.23(0.09) 00.55(0.08) 00.42(0.15) 00.23(0.05) · · · 0.16(0.05) 18953c272.12,10 00.42(0.14) 00.55(0.29) 00.27(0.15) 01.40(0.20) · · · 01.00(0.40) 01.60(0.40) · · · 1.83c2932,11 01.60(0.20) 05.30(0.40) 01.94(0.07) 03.25(0.07) 01.20(0.10) 02.80(0.20) 00.31(0.11) 1.30(0.10) 144503c3102 00.12(0.06) 00.36(0.07) 00.08(0.04) 00.48(0.04) 00.19(0.06) · · · · · · 00.29(0.09) · · ·3c326n2,12 00.30(0.06) 00.69(0.06) 00.41(0.04) 01.26(0.05) 00.31(0.06) 00.46(0.22) 00.25(0.09) 00.29(0.09) 11833c4242,13 00.45(0.07) 01.12(0.06) 00.21(0.03) 00.43(0.04) 00.15(0.04) 00.17(0.07) · · · 00.28(0.05) <35603c4332,14 02.00(0.40) 01.40(0.20) 00.58(0.17) 01.20(0.20) · · · 01.60(0.50) · · · 01.20(0.30) <54953c4362 00.48(0.16) 00.46(0.07) 00.31(0.12) 00.28(0.04) · · · 00.22(0.10) 00.41(0.12) 00.21(0.10) · · ·CenA2,15 10.00(6.00) 57.00(9.00) 27.00(4.00) 09.30(3.70) · · · 12.00(3.00) · · · 27.00(4.00) <2103c2363,10 00.20(0.03) 00.95(0.07) 00.32(0.05) 00.64(0.05) · · · · · · · · · · · · <6220Arp2204,16 < 970 18.62(1.68) 09.80(1.30) 07.30(0.20) · · · · · · · · · 12.90(3.80) 22500IRAS 00188-08564,16 < 19 00.43(0.05) 00.20(0.04) 00.38(0.12) · · · · · · · · · < 32 21300IRAS 03521+00284,16 < 13 00.82(0.13) 00.27(0.09) 00.37(0.03) · · · · · · · · · < 3.0 29400
65
Table 2.2 (cont’d)
Galaxy S(0)(err) S(1)(err) S(2)(err) S(3)(err) S(4)(err) S(5)(err) S(6)(err) S(7)(err) M(H2)Name (106 M⊙)(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
IRAS 05189-25244,16 < 282 03.40(0.70) 01.50(0.20) 03.60(1.30) · · · · · · · · · < 341 3200IRAS 06035-71024,17 < 61 04.17(0.04) 02.33(0.58) 03.4(1.00) · · · · · · · · · < 116 19000IRAS 06206-63154,17 < 29 01.29(0.10) 00.50(0.05) 00.59(0.19) · · · · · · · · · < 18 41400IRAS 07598+65084,16 < 30 01.38(0.70) 00.37(0.06) 00.67(0.02) · · · · · · · · · < 225 36000IRAS 08572+39154,16 < 154 01.19(0.04) 00.51(0.10) 00.46(0.13) · · · · · · · · · < 430 4200IRAS 10565+24484,16 < 109 06.57(0.05) 02.56(0.09) 04.03(0.02) · · · · · · · · · < 54 17900F12112+03054,18 1.4(0.4) 4.12(0.44) 1.73(0.31) 2.37(0.03) · · · · · · · · · 4.10(0.8) 19200IRAS 13451+12324,18 < 45 02.83(0.65) 01.29(0.05) 02.19(0.26) · · · · · · · · · < 41 <46900F14348-14474,18 < 6.1 4.47(0.15) 1.95(0.10) 2.4(0.21) · · · · · · · · · < 31 20500IRAS 17208-00144,16 < 239 08.81(0.09) 04.97(0.85) 05.7(1.1) · · · · · · · · · < 85 22400IRAS 19254-72454,17 < 85 08.81(0.57) 03.82(0.95) 03.82(0.04) · · · · · · · · · < 4.00 21000IRAS 20087-03084,16 < 21 02.29(0.14) 00.84(0.33) 01.3(0.40) · · · · · · · · · < 22 51100IRAS 23365+36044,16 < 91 03.9(0.31) 02.01(0.95) 03.1(0.5) · · · · · · · · · < 30 27000Mrk 2734,16 < 262.5 10.24(0.09) 05.6(0.70) 10.4(0.9) · · · · · · · · · < 150 15870UGC51014,19 < 100 04.96(0.47) 02.7(0.5) 02.8(0.30) · · · · · · < 195 7500Mrk463E4,18 < 79 02.79(0.37) 01.31(0.08) 02.8(0.60) · · · · · · < 385 2180PG1440+3565,20 0.60(0.14) 1.14(0.14) 0.51(0.06) 0.64(0.11) · · · · · · · · · · · · 6400N62406,21 8.60(1.20) 50.40(1.50) 39.8(0.50) 70.90(1.90) 36.4(10.8) 95.2(16.6) · · · 33.7(12.5) 14100N31107,21 1.60(0.20) 10.00(0.90) 4.00(1.00) · · · · · · · · · · · · · · · 5750
66
67
Table 2.2 (cont’d)
Galaxy S(0)(err) S(1)(err) S(2)(err) S(3)(err) S(4)(err) S(5)(err) S(6)(err) S(7)(err) M(H2)Name (106 M⊙)(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
N32567,22 9.00(3.00) 61.0(5.00) 34.0(3.00) · · · · · · · · · · · · · · · 1400N36907,21 3.80(0.60) 22.0(1.00) 11.0(4.00) · · · · · · · · · · · · · · · 3700N51357,21 1.30(0.50) 17.00(1.00) 9.30(1.40) · · · · · · · · · · · · · · · 5500N67017,21 1.90(0.90) 10.00(1.00) 4.60(0.80) · · · · · · · · · · · · · · · 1200N71307,23 1.40(0.30) 09.90(0.70) 4.20(1.70) · · · · · · · · · · · · · · · 4130N75917,24 1.00(0.30) 05.50(0.60) 3.00(1.00) · · · · · · · · · · · · · · · 3270N77717,21 3.50(1.80) 16.00(1.00) 8.70(1.20) · · · · · · · · · · · · · · · 3280
aThe cold H2 gas mass is calculated using αCO,Gal = 3.2 M⊙(K km s−1 pc2)−1 (not considering Helium and other heavy element masscontribution)
bThe mass value in the parentheses is calculated using CO weighted mean αCO, suggested by Sandstrom et al. (Table 4) 2013 using aconstant dust to gas ratio (DGR) marked in the reference as “S”
cThe two numbers as the notemark in the first column of the table correspond to references. The first number is the reference from whichwe adopt the MIR rotational line flux and the second number is for the cold gas mass derived from the CO emission.Ref: (1) Roussel et al. 2007; (1S) Sandstrom et al.2013 + Roussel et al. 2007; (2) Ogle et al. 2010; (3) Guillard et al. 2012; (4) Hiqdon et al.2006; (5) QUEST sample by S. Veilleux (2009); (6) Armus et al. 2006; (7) M. Pereira-Santaella. 2010; (8) Okuda et al. 2005; (9) Salome &Combes 2003; (10) Ocana Flaquer et al. 2010; (11) Evans et al. 1999; (12) Nesvadba et al. 2010; (13) Saripalli & Mack 2007; (14) Evans et al.2005; (15) Israel et al. 1990; (16) Solomon et al. 1997; (17) Mirabel et al. 1989; (18) Evans et al. 2002; (19) Rigopoulou 1996; (20) Evans etal. 2001; (21) Sanders et al. 1991; (22) Kazushi Sakamoto et al. (2006); (23) Curran et al. (2000); (24) Lavezzi & Dickey 1998; (25) Young etal. (2011); (26) Harnett et al.(1991)
68
Table 2.3. Observed molecular hydrogen lines
Transition Short notation Rest λ Eu
kA
ν = 0 (µm) (K) (10−11s−1)(1) (2) (3) (4) (5)
J = 2 −→ 0 S(0) 28.219 0510 2.95J = 3 −→ 1 S(1) 17.035 1015 47.6J = 4 −→ 2 S(2) 12.279 1681 275.0J = 5 −→ 3 S(3) 9.665 2503 980.0J = 6 −→ 4 S(4) 8.025 3473 2640.0J = 7 −→ 5 S(5) 6.910 4585 5880.0J = 8 −→ 6 S(6) 6.109 5828 11400.0J = 9 −→ 7 S(7) 5.511 7196 20000.0
aThe rotational upper level energies were computed from the molecularconstants given by Huber & Herzberg 1979 and transition probabilities arefrom Black & Dalgarno 1976
69
Table 2.4. Model derived parameters for SINGS galaxies
Galaxy Tℓ n M(T>100K)Mtotal
Name (K) (in %)(1) (2) (3) (4)
N0337 59 (43) 5.47 9.46 (2.3)N1097 47 5.00 4.88N1266 26 3.80 2.30N1291 52 4.27 11.79N1316 30 3.79 3.48N1482 51 5.00 6.76N1566 42 3.89 8.15N2798 71 4.96 25.76N2976 66 (89) 5.87 13.22 (56.6)N3049 42 5.02 3.06N3184 57 (69) 5.57 7.67 (18.3)N3190 55 4.68 11.08N3198 60 5.20 11.70N3265 53 5.15 7.17Mrk33 43 4.28 6.28N3351 43 (70) 4.75 4.22 (26.2)N3521 63 5.42 12.97N3627 43 (77) 4.53 5.08 (39.7)N3938 57 (55) 6.00 6.02 (5.03)N4125 50 4.16 11.18N4254 35 (29) 4.68 2.10 (1.05)N4321 38 (62) 4.96 2.18 (15.1)N4450 53 4.28 12.46N4536 49 (58) 5.03 5.64 (11.1)N4559 51 5.90 3.69N4569 37 4.30 3.76N4579 37 3.92 5.48N4625 65 (54) 6.39 9.81 (3.61)N4631 51 5.25 5.72N4725 > 51 4.79 <7.8
70
Table 2.4 (cont’d)
Galaxy Tℓ n M(T>100K)Mtotal
Name (K) (in %)(1) (2) (3) (4)
N4736 52 (112) 4.54 9.88 (149)N4826 50 4.85 6.93N5033 51 4.65 8.56N5055 50 (72) 5.05 6.04 (26.4)N5194 39 3.94 6.28N5195 50 4.51 8.78N5713 56 (61) 4.91 10.36 (14.5)N5866 55 4.70 10.95
N6822A 39 4.20 4.91N6946 45 (82) 4.96 4.23 (45.6)N7331 49 5.21 4.96N7552 51 5.02 6.67N7793 48 5.25 4.42
aThe value in the parentheses is calculated assum-ing central αCO from the dust emission, evaluated bySandstrom et al. 2013
71
Table 2.5. Model derived parameters for radio, U/LIRGs galaxies
Galaxy Ref Model mass Tℓ n[M(T>100K)]
Mtotal
T ′ℓ
[M(T>100K)]Mtotal
Name 106M⊙ (K) in %, for Tℓ (K) in % for T ′ℓ
(1) (2) (3) (4) (5) (6) (7) (8)
3c031 1 360 41 4.69 3.73 65 20.403c218 1 1040 41 4.47 4.53 67 24.92
3c272.1 1 1.87 50 3.41 18.8 101 102.43c293 1 13210 48 4.77 6.28 75 33.813c310 1 515 · · · 4.11 · · · · · · · · ·3c326n 1 2604 64 4.05 25.0 112 126.23c424 1 40820 >88 5.19 >58.5 >132 >320.03c433 1 12850 >62 4.51 >18.7 >101 >103.63c436 1 18820 · · · 4.40 · · · · · · · · ·CenA 1 379 >58 4.64 >13.8 >93 >76.793c236 2 14083 >61 4.86 >14.8 >95 >82.04
Arp220 3 10000 41 5.07 2.65 62 14.13IRAS 00188-0856 3 5472 33 4.38 2.36 55 13.13IRAS 03521+0028 3 53970 57 5.35 8.67 84 48.02IRAS 05189-2524 3 3600 52 4.27 11.8 87 62.94IRAS 06035-7102 3 19240 50 4.39 9.54 82 51.55IRAS 06206-6315 3 11760 36 4.68 2.33 56 11.80IRAS 07598+6508 3 76700 60 5.27 11.3 88 58.93IRAS 08572+3915 3 6586 53 5.02 7.80 85 51.64IRAS 10565+2448 3 19746 51 5.04 6.60 78 35.82IRAS 12112+0305 3 38640 59 5.07 11.7 90 63.80IRAS 13451+1232 3 43890 >49 4.57 >7.83 >78 >41.27IRAS 14348-1447 3 37130 59 4.74 13.9 92 72.38IRAS 17208-0014 3 14900 45 4.59 5.69 71 29.40IRAS 19254-7245 3 89760 69 5.42 19.4 101 106.48IRAS 20087-0308 3 45390 49 5.03 5.64 74 28.89IRAS 23365+3604 3 18230 45 4.72 5.13 71 27.30
Mrk 273 3 9730 43 4.37 5.82 71 31.55UGC5101 3 11655 56 4.96 10.1 85 53.17Mrk 463E 3 4320 61 4.24 19.7 102 107.1
72
Table 2.5 (cont’d)
Galaxy Ref Model mass Tℓ n [M(T>100K)]Mtotal
T ′ℓ
[M(T>100K)]Mtotal
Name 106M⊙ (K) in %, for Tℓ (K) in % for T ′ℓ
(1) (2) (3) (4) (5) (6) (7) (8)
PG1440+356 4 11735 58 5.04 11.1 88 59.58N6240 5 7240 40 3.73 8.20 65 30.95N3110 5 1520 33 4.22 2.82 51 11.33N3256 5 1360 50 3.71 15.3
NGC3690/IC694 5 1310 36 4.04 4.48 56 17.21N5135 5 820 25 3.70 25.2 41 9.17N6701 5 800 44 4.13 7.65 68 30.36N7130 5 1320 35 4.18 3.55 54 14.13N7591 5 754 32 4.18 2.67 49 10.17N7771 5 1000 33 3.82 4.39 54 17.25
Aa. The molecular gas mass in column 3 is calculated extrapolating the power law model to T⋆ℓ
= 49K.b. The temperature in column 4 is the model extrapolated temperature required to fit the totalmolecular gas mass, calculated using the αCO,Gal.c. The warm gas mass fraction listed in column 6 is derived using the cold molecular gas mass byassuming αCO,Gal.d. The temperature in column 7 is the model extrapolated temperature required to fit the totalmolecular gas mass assuming a low αCO of 0.8 M⊙(K km s−1 pc2)−1.e. The warm gas mass fraction listed in column 8 is derived using the cold molecular gas mass byassuming a low αCO
73
Table 2.6. Observed molecular hydrogen line fluxes and mass in low metal-licity dwarfs
Galaxy 12+log[O/H] S(0) S(1) S(2) S(3) S(4) S(5) D M(CO) M(dust) M(model) n(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13)
CG007-025 7.77 · · · 5.45(0.43) 1.67(0.43) · · · 0.72(0.21) · · · 24.5 41.2 · · · 3935 5.24N66 8.10 475(181) 1025(116) 1283(175) 433(207) 586(400) 482(314) 0.06 0.016 0.435 0.451 3.59
Haro 11 8.20 <2.49 1.68(0.55) 1.01(0.13) 1.21(0.18) · · · · · · 92 1670 4500d 6700 4.52UM311 8.31 · · · 1.49(0.69) · · · · · · 0.73(0.35) 0.66(0.27) 24 95 400 4.51
HubbleV 8.42 0.81(0.43) 2.11(0.43) 1.60(0.41) 2.59(0.38) · · · · · · 0.5 0.27 0.28 0.16 4.20
aAll rotational lines, masses, and ditance are in units 10−17 W m−2, M⊙, and Mpc respectively
bRef for metallicity CGCG007-025:(Izotov et al. 2007); N66:(Dufour et al. 1982,Dufour et al. 1984); Haro11:(Guseva et al. 2012);UM311:(Izotov et al. 1998)
cFor CGCG007-025 we also used the S(7) line flux of H2, (0.46 ± 0.13) × 10−17Wm−2, from (Hunt et al. 2010)
dRef for CO derived H2 mass CGCG007-025:(Hunt et al. 2015); N66:(Rubio et al. 1996); Haro11 and UM311:(Cormier et al. 2014); HubbleV:(Israel et al. 2003)
eRef for dust derived H2 mass N66:(Leroy et al. 2007); Haro11:(Cormier et al. 2014); HubbleV:(Israel et al. 1996)
fThe molecular gas mass for Haro11 is in the range (4.5–11.5)×108 M⊙, due to uncertainty in H I mass. In the table the minimum value ismentioned
74
Chapter 3
Molecular gas properties in ISM of
U/LIRGs of GOALS
3.1 Introduction
The Infrared Astronomical Satellite (IRAS) discovered a large population of lu-
minous infrared galaxies (LIRGs), whose bolometric luminosities are dominated by
the infrared part of their spectral energy distribution (SED; LIR ≥ 1011 L⊙, Soifer et
al. (1987)) Galaxies with LIR ≥ 1012 L⊙, known as Ultra Luminous Infrared Galaxies
(ULIRGs) are generally major mergers of molecular gas rich galaxies. The star for-
mation rate in U/LIRGs are above a magnitude or two higher and also exhibit high
FIR color compared to normal star forming galaxies. The merging morphology, high
LIR, high star formation rate and high dust temperature makes U/LIRGs to stand
out in their interstellar medium (ISM) properties from normal star forming galaxies.
Stars form from molecular clouds, hence the knowledge of molecular gas content is
essential to understand star formation in galaxies.
H2, the most abundant molecule in the universe, is difficult to observe because
of the lack of dipole moment and high rotational energy levels. The lower rotational
transitions of carbon monoxide (CO) are used as a tracer to determine molecular gas
75
masses in galaxies. To convert the measured integrated intensities of CO luminosity to
molecular gas mass a conversion factor, αCO, is used, calibrated against the virial mass
estimates in presumed self gravitating molecular clouds of the Milky Way (Solomon
et al. 1987, Scoville et al. 1987). Using the standard Milky Way αCO in ULIRGs
the molecular gas mass is overestimated and is about 5 times the dynamical mass of
the galaxy (Solomon et al. 1997). To avoid this a low conversion factor for ULIRGs,
of about 5×, is used as compared to the galactic value. The conversion factor is
proportional to√
n/T , where n and T are the number density and temperature of
molecular clouds (Rachford et al. 2002). The observed high FIR color (high dust
temperatures) in ULIRGs plausibly suggest elevated gas temperatures in U/LIRGs,
which can reduce the conversion factor, αCO. The FIR color of LIRGs on average
are in between the values of a normal star forming galaxy and ULIRGs, which may
suggest intermediate values of αCO.
We used the power law model to estimate molecular gas content and study the
molecular ISM properties in normal star forming galaxies of the Spitzer Infrared
Nearby Galaxy Survey (SINGS). The merging morphology, high dust temperature of
galaxies in GOALS suggests indication of shocks. In this chapter we would like to
test our power law model in galaxies, where shocks play a significant contribution in
exciting H2. The main motivation of this chapter is:
1. to test the power law model in estimating molecular gas mass
2. how molecular mass estimated from the power law model compare with the dust
based and sub-mm luminosity, L850 measurements
3. how does the power law indices compare with the SED infrared slope
4. is there any change in the power law indices for AGN and star formation nucleus
dominated galaxies.
The chapter is organized in the following manner. First we describe our sample
76
selection from GOALS in §,2. In §,3 we describe our analysis followed by our results
and discussion in §,4 and §,5 , respectively. Finally we summarize our results in §,6.
3.2 Sample
The IRAS Revised Bright Galaxy Sample (IRAS-RBGS; Sanders et al. (2003)) is a
complete flux limited sample of 629 galaxies, at Galactic latitudes |b| > 5 with S60µm
> 5.24 Jy. The Great Observatories All-Sky LIRG Survey (GOALS; Armus et al.
(2009)) includes a complete luminosity-limited sample of 180 LIRGs and 22 ULIRGs
from the IRAS-RBGS. The galaxies in the GOALS sample range over distances 15–
400 Mpc, with the median redshift z = 0.0215 (∼ 95.2 Mpc). The spectra of these
galaxies were obtained in the staring mode using the IRS Short-Low (SL: 5.5–14.5
µm) and Long-Low (LL; 14–38 µm) instruments. At the median galaxy distance the
nuclear spectrum covers the central 1.8 and 5.2 kpc in the SL and LL instruments.
The on-source integration time range from 30–610 s, depending on the brightness of
nuclei.
3.2.0.1 Ancillary data CO(J=1–0)
We would like to compare our power law derived molecular gas with the CO
derived mass values. For this research work we selected 49 galaxies from the GOALS
sample, for which the CO luminosity values were available in the literature. The CO
(J = 1–0) line intensities for our selected galaxies are from Sanders et al. (1991). We
estimated the molecular gas masses using the galactic conversion factor αCO = 4.35
M⊙ (K km s−1 pc2)−1 (equivalent to XCO = 2 × 1020 (K km s−1 cm2)−1) (Bolatto
et al. 2013). The molecular gas masses were further reduced by a factor of 1.36 to
remove the heavy element contribution in the ISM molecular gas mass.
77
3.2.1 Data reduction and Spectral fitting
The IRS-Spitzer spectroscopic nuclear data were reduced using the S17 and S18.7
IRS pipeline from the Spitzer Science Center. The pipeline software removes the bad
pixels, subtracts the background, corrects for non-linearity and performs wavelength
and flux calibration. The background is subtracted using a dedicated background
when available. The standard extraction aperture and the point source calibration
modes were used to extract the one-dimensional spectra using the Spitzer IRS Custom
Extraction software (SPICE). The data reduction is described in detail and all the
reduced GOALS spectra are presented in Stierwalt et al. (2013). For systems observed
in the mapping mode, spectra were extracted using the CUBISM (Smith et al. 2007b).
3.3 Analysis
The main aim of this research work is to calculate the total H2 gas mass and test
whether MIR H2 rotational lines can be used as a tracer for estimating molecular gas
content in U/LIRGs. We would like to compare the estimated molecular gas mass
with the CO derived masses. However, only the central nucleus region is observed in
the staring mode with the IRS spectrograph in most of the GOALS galaxies. The CO
beamsize is about 1 arcmin wide covering sometimes the entire or may be significant
regions of galaxies. There is a discrepancy between the observed regions of galaxies
by the IRS-Spitzer and CO beam. Figure 3-1 shows the IRS-Spitzer (blue rectangle)
and CO beam (blue circle) on one of our sample galaxy NGC 5135. The IRS-Spitzer
observes the central nuclear sub-region of the galaxy when compared to the large CO
beam. Several galaxies are unresolved in the IRS-Spitzer beam and hence the IRS
and CO beam observe complete galaxy region.
78
Figure 3-1 The observed IRS-Spitzer and CO beam regions on NGC 5135.The blue circular region is about 1 arcmin wide in diameter.The IRS-Spitzer slit and CO beam observe different regions ofthe galaxy.
79
3.3.1 Disk template
To compare our power law derived H2 gas masses with the CO derived masses
we require H2 line fluxes from similar observed regions. As discussed in the previous
section, the IRS-Spitzer and the CO beam coverages are different, hence a correction is
required to estimate the H2 line fluxes and compare for the similar region as observed
by the CO beam. A method to correct this mismatch between the IRS and CO beam
was to derive a disk spectrum template and appropriately scale and add it to the IRS
nucleus spectrum. The derived final spectrum is an equivalent MIR spectrum of the
CO beam observed region. The H2 rotational line fluxes is then measured from the
final spectrum to evaluate the molecular gas mass in galaxies.
To prepare the disk template we used galaxies, which were observed in the mapping
mode. We selected a total of 10 galaxies for this, after excluding systems which have
widely separated nucleus and a steep MIR slope. We defined the boundary of the
nucleus region at 1 kpc radius from the merging center. From the spectrum of the
entire region (nucleus and disk), we subtracted the nucleus spectrum to estimate the
disk spectrum for each of our selected 10 galaxies. We calculated the median value
of the flux from the disk spectra at each wavelength and prepared the disk template.
The median LIRG’s disk template is shown in Figure 3-2 (in blue). For comparison we
have plotted the starburst nucleus (in red) and star forming SINGS galaxy template
(in black) from Brandl et al. (2006) and Smith et al. (2007), respectively.
3.3.2 Scaling and adding the template spectrum
We scaled the LIRG’s disk template appropriately, and added to the IRS-Spitzer
nuclear spectrum of each galaxy to construct an equivalent spectrum for the observed
wide CO beam region. For this we used the IRAC- 8 µm and MIPS- 24 µm fluxes of
the same CO beam observed region. We scaled the SL and LL (short low and long
80
Figure 3-2 The LIRG’s disk template spectrum (blue) is prepared after sub-tracting the nucleus from the total spectrum obtained from themapped galaxies. For comparison we also plot the normal starforming galaxy (SINGS, (Smith et al. (2007)), in black) and thenucleus starburst template (Brandl et al. (2006)), in red). TheLIRG’s disk template is similar to the spectrum of normal starforming galaxies.
81
low) template and added to the IRS nuclear spectrum to match the IRAC - 8 µm and
MIPS - 24 µm fluxes. Ideally with this procedure the short low and the long low MIR
spectrum need to be continuous but a slight mismatch between the two was observed
in the spectrum. To get a continuous spectrum the new version of the SL and LL
templates were scaled accordingly to match the two parts of the spectrum, finally
deriving a continuous MIR spectrum. We evaluated the percentage differences, which
was less than 20%, between the IRAC - 8 µm and MIPS - 24 µm fluxes with the final
fluxes at 8 µm and 24 µm MIR spectrum estimated from the scaled and added IRS
spectrum.
3.3.3 PAHFIT- to recover H2 line flux
By scaling and adding the template we build the continuous MIR spectrum for
the CO beam observed region for each galaxy in the sample. We used the PAHFIT
tool developed by Smith et al. (2007) to estimate H2 line fluxes for our region of
interest. PAHFIT is a spectral decomposition tool to separate IRS low-resolution
spectra into broad Polycyclic Aromatic Hydrocarbon (PAH) features, unresolved line
emission, grain continuum, and the full line flux of any blended features. Several H2
rotational lines are blended with PAH features and atomic ionized lines in the IRS
low-resolution spectra. The line FWHM, line equivalent width, and uncertainty in
the fluxes were determined within PAHFIT. The PAHFIT tool uses Drude profiles to
fit PAH features of different width in the IRS observed wavelength range.
3.3.4 H2 gas mass from power law model
After recovering H2 line fluxes from PAHFIT, we fit the MIR rotational line fluxes
using the temperature power law distribution. Using the power law model we derived
the molecular gas properties in our selected GOALS galaxies.
82
Figure 3-3 Power law model fit (red solid) to the observed H2 line ratios(black points) for a sample galaxy NGC23 in GOALS
We assumed the column density of H2 molecules are distributed by a power law
function with respect to temperature, dN ∝ T−n dT, where dN is the number of
molecules in the temperature range T—T+dT (the model is described in detail in
chapter 2 of the thesis). Keeping the upper temperature, Tu, fixed at 2000 K and
varying Tℓ and n, we fit the H2 excitation diagram for each galaxy selected from the
GOALS sample. Figure 3-3 is a model fit (red solid) to the observed H2 line ratios
(black points) in the excitation diagram of NGC 23, with n = 4.32 in the temperature
range 50–2000 K. With a continuous power law model well reproducing the rotational
emission lines from warm H2, we extrapolated our model to lower temperatures to
estimate the total molecular gas reservoir. We calculated our molecular gas masses by
extrapolating the model to 49 K, the average value of extrapolated lower temperature,
determined by Togi & Smith, (2016, submitted) in their normal star forming SINGS
galaxies to recover similar amount of molecular gas mass derived using LCO.
83
3.4 Results
3.4.1 LIRG’s disk template
We prepared the LIRG’s disk template (in blue) from the mapped GOALS galaxies
and compared it with the star burst nucleus template (in red, Brandl et al. (2006)
and normal star forming galaxy template from SINGS (in black, Smith et al. (2007)),
all normalized at 24 µm.
The LIRG’s disk template is similar to the normal star forming galaxy template
but deviates significantly from the nuclear starburst template (Figure 3-2). The MIR
slope of the LIRG’s disk template is lower than the starburst nucleus. The high MIR
slope is an indication of warm dust. The flux ratio, F30
F15, a MIR slope indicator, can
be used to characterize a starburst or an AGN dominated nucleus (Laurent et al.
(2000)). The ratio F30
F15is greater than 5.9 and less than 4.8 in starburst nuclei and
AGN dominated environments, respectively (Brandl et al. (2006)). The MIR slope
in nuclei region of GOALS galaxies are in the range 2 < F30
F15< 35.4 (Stierwalt et al.
(2013)). A majority of the LIRGs (63%) have MIR slope in the range 4 < F30
F15< 10.
When compared to the less luminous LIRGs, ULIRGs show a much steeper slope
on average F30
F15= 12.5 compared to 7.11, for LIRGs. Stierwalt et al. (2013) also
discovered the most obscured nuclei have steeper MIR slopes since most of the warm
dust emission is hidden behind a large amount of cooler dust.
The energy source in the nucleus is primarily through the compact star forming
regions or by an AGN making the environment warmer than compared to the disk
region. The PAH features in the starburst nuclei regions are weak compared to the
disks of the galaxies which suggests the possibility of PAHs getting destroyed by the
harsh radiation from the nucleus.
84
Figure 3-4 The model derived power law indices for our selected GOALSgalaxies. The average power law index is 4.41±0.49
3.4.2 Power law index
H2 molecules in a galaxy’s ISM can be heated through various excitation mecha-
nism including UV radiation from stars, shocks from AGN and supernova. Cooling
of H2 molecules result in various different rotational lines, which can be fit using a
power law temperature distribution. Using a temperature power law distribution we
fit the MIR rotational line fluxes in our sample galaxies of GOALS.
The power law index in our sample range from 2.83–5.53, with a mean value of
4.41±0.49. Fig 3-4 is a histogram of derived power law indices observed for GOALS
galaxies. The power law indices observed in GOALS galaxies are found to be lower
than the average power law index of 4.8, for normal SINGS star forming galaxies, an
indication of warm gas in the galaxy ISM of GOALS sample.
Neufeld et al. (2008) in a shocked environment of supernova remnant IC 443
found a power law temperature distribution for H2 column density in the range 3–6,
85
with an average value of 4.50. (Zakamska et al. (2010)) in her observation of ULIRGs
used a power law model with indices in the range 2.5–5.0 to fit the observed H2 line
ratios. Similarly Pereira-Santaella et al. (2014) in six local infrared bright seyferts
required a power law index in the range 4–5 to reproduce the observed H2 Spectral
Line Energy Distribution (SLED). Our derived power law indices are consistent with
the theoretical predicted and other observed values in the literature (Neufeld & Yuan
(2008), Hollenbach & McKee (1979)).
3.4.3 H2 mass- extrapolating power law model to 49 K
The present method of using two or three temperature components to fit the MIR
H2 rotational line fluxes is a simple but contrary to the distribution of molecules
with continuous temperature in the ISM. We employed a temperature power law
distribution of H2 molecules to fit the rotational line fluxes. A continuous power
law distribution model can be extrapolated to lower temperatures to recover the
total molecular gas mass. For normal star forming galaxies in the SINGS sample for
calibration we estimated the average extrapolated temperature of 49±10 K to match
the amount of molecular gas mass estimated from CO emission, when used a galactic
conversion factor, within a factor of ≈ 2. Similarly extrapolating our model to 49 K
we estimated the molecular gas mass for GOALS galaxies.
Figure 3-5 show the deviation of the model derived molecular gas mass with the
CO derived luminosity values. The solid line is the one-to-one correspondence line
with a factor of two deviation, shown by dotted lines. Galaxies are color coded in 4
categories according to their LIR,
1011 ≤ log(
LIR
L⊙
)
≤ 1011.33,
1011.33 ≤ log(
LIR
L⊙
)
≤ 1011.66,
1011.66 ≤ log(
LIR
L⊙
)
≤ 1011.99, and
log(
LIR
L⊙
)
≥ 1011.99,
86
Figure 3-5 The model derived molecular gas mass extrapolated to 49 K cor-relates with the CO derived molecular gas with assuming Galac-tic conversion factor. The dotted lines are a factor of two devia-tion from one to one correspondent solid lines. Generally galaxieswith LIR
L⊙≥ 11.66 show deviation from the solid line. This may
be due to a low αCO or high molecular gas temperatures.
in black, blue, green and red respectively. Generally galaxies with log(
LIR
L⊙
)
≥ 1011.66
show deviation from the solid line, hinting for a requirement of high αCO or a high
extrapolation temperature than 49 K (Figure 3-5).
3.5 Discussions
3.5.1 Low power law index in U/LIRGs
The variation in H2 rotational line ratios can cause change in the power law index,
n, required to fit MIR H2 ratios. Galaxies with warm molecular ISM result in a large
number of molecules at high temperatures causing a less steep temperature power law
87
distribution or a low power law index. Hence, a low power law index is an indication
of warm molecular ISM. The average power law index in GOALS is calculated to be
4.41±0.49, lower than the normal star forming SINGS galaxies, an indication of a
warmer molecular ISM in GOALS than compared to normal star forming galaxies.
The turbulence and shocks due to merging phenomenon in addition to radiation
from newly forming stars at a rate of '100 M⊙/yr can be a plausible explanation for
heating of molecular gas in GOALS. The FIR color, a proxy for dust temperature,
is higher in U/LIRGs than the normal star forming galaxies, which may possibly
suggest a warm molecular ISM (Sanders & Mirabel (1996), U et al. (2012)). The
infrared luminosity in U/LIRGs is due to high star formation rate. The high infrared
luminosity can be due to intense radiation from young forming stars capable of heating
the ISM gas. Figure 3-6 show the relation between power law index, n, and the
infrared luminosity, LIR. The blue points in the plot are the galaxies which require
a template addition to construct the MIR spectrum for the observed CO region.
The red points are unresolved and require no template addition. The power law
index decreases with the increase in infrared luminosity. As discussed previously low
power law index is an indication of warm molecular ISM, galaxies with high infrared
luminosity also indicate a warm molecular ISM.
3.5.2 Relation between LIR, PAHs and power law index
We study the relation of the power law index, n, with the infrared luminosity
LIR and 6.2 PAH flux. ULIRGs emit most of their energy in the infrared spectral
region, implying that they are heavily dust obscured. To power their huge infrared
luminosities these galaxies must undergo an intense star formation rate of ∼ 100
M⊙/yr. The dust in these galaxies absorbs the radiation from young forming stars
and re-radiates at long wavelengths causing high infrared luminosities. Another pos-
sible energy source can be the accretion of large quantities of gas onto a central
88
Figure 3-6 The power law index decrease with the increase in the infraredluminosity for LIR ≥ 1012 L⊙. A decrease in the power law indexsuggests warm molecular ISM in ULIRGs.
89
supermassive blackhole (SMBH) in an AGN (Sanders et al. (1988)). In most cases a
combination of star formation and AGN can contribute to the emission from U/LIRGs
and the relative contribution of the two processes has been extensively studied (Lutz
et al. 1998, Genzel et al. 1998, Laurent et al. 2000, Armus et al. 2007, Spoon et
al. 2007, Veilleux et al. 2009, da Cunha et al. 2010). ULIRGs are in an advanced
merging stage than compared to LIRGs, with two or more spirals that consists of a
flat rotating disk containing stars, gas and dust and a central concentration of stars
in the bulge. The merging phenomenon may result in shocks and turbulence heating
H2 molecules and a high warm gas mass fraction, causing the power law index to
reduce.
The MIR spectrum includes different PAH bands along with other atomic, molec-
ular, and ionized lines. Different ionized and neutral PAH bands can be observed
in the MIR band. PAH molecules can be destroyed in harsh environments of AGN
(Smith et al. (2007)). The mean equivalent widths (EQWs) of 6.2 µm PAH feature
in the LIRGs of GOALS sample is 0.55 µm lower than compared to ULIRGs with 0.3
µm (Stierwalt et al. 2013). Hence, this indicates a possibility of change in the 6.2 µm
PAH feature with the LIR. This also suggests a variation in the 6.2 µm PAH feature
with the power law index, n. Figure 3-7 demonstrate the relation between power
law index, n, and the equivalent width of the 6.2m PAH band. The blue points in
the plot are the galaxies which require a template addition, while the red points are
unresolved and require no template addition. A low power law index also have low 6.2
PAH EQWs, implying weak PAH emission in the warm molecular ISM environments.
This may suggest PAH molecules getting destroyed or small PDRs than compared to
normal galaxies.
90
Figure 3-7 Relation between 6.2 µm PAH equivalent width as a functionof power law index. Galaxies with low power law index haveelevated warm molecular gas mass fraction. In such warm envi-ronments PAH molecules can be destroyed resulting in low equiv-alent widths
91
3.5.3 Low αCO or high temperature
The 49 K is the average extrapolation temperature required to recover the total
molecular gas mass for normal star forming SINGS galaxies when using a galactic
conversion factor. However, galaxies of GOALS are mergers with high infrared lumi-
nosity and far infrared color, with a plausibility of high gas temperatures. High gas
temperatures and turbulence driven by merging or high star formation rate result an
increase in CO linewidth, causing overestimation of molecular gas mass when used
a Galactic conversion factor. The median CO FWHM linewidth is broad (average
FWHM ∼800 km/s) especially in sub millimeter galaxies and are often observed with
double-peaked line profiles, indicating merging or a disk (Greve et al. 2005). The
merging phenomenon and turbulence may result an increase in the linewidth. The
normal galactic conversion factor yields M(H2) = 2.5 Mdyn, a clear overestimate in
molecular gas mass (Solomon et al. (1997)).
In Figure 3-5 galaxies are generally on the solid line, implying an average of 49
K model extrapolated temperature is able to recover the molecular gas mass derived
from CO emission. However, galaxies with high infrared luminosity, LIR
L⊙≥ 11.66
show deviation from the solid line. The deviation in these galaxies can be minimized
using a high conversion factor or high molecular gas temperature (> 49 K) or both.
Galaxies above the solid line may be dominated by the effect of high gas temperatures
and may require model extrapolation to temperatures above 49 K to estimate the true
molecular gas mass. Galaxies to the left of the solid line are overestimated in their
molecular gas masses estimated form their CO emission and may require a low αCO.
3.5.4 Tℓ from gas-to-dust mass ratio GDR
In the previous subsection we discuss the deviation of CO derived molecular gas
mass, M(H2,CO), from the H2 model derived mass, M(> 49 K). We aim to test
92
our model predicted mass with dust derived molecular gas mass estimates, another
molecular gas tracer. The dust and gas in galaxies are well mixed in the ISM and
the gas-to-dust mass ratio is usually consistent to be constant. Assuming a constant
gas-to-dust ratio and estimating the dust mass from FIR luminosity the total gas
mass in the galaxy ISM can be estimated. Subtracting the neutral atomic from the
total gas mass, molecular gas mass can be calculated.
We estimated dust mass through 850 µm emission using the relation
Mdust
M⊙
=D2 × S850
B(ν, Tdust) × κ850,dust
, (3.1)
Mdust
M⊙
=1.25 × 105 × D2 × S850
Tdust × κ850,dust
, (3.2)
where D, S850, Tdust, κ850,dust are the distance in Mpc, 850 µm flux in Jy, dust
temperature, and dust mass opacity in m2/kg, respectively. The dust temperature
is estimated using the ratio, S60
S100, 60-to-100 micron flux ratio (Lisenfeld et al. 2000).
We assumed a value of 0.066 m2/kg for the dust mass opacity.
Two facts stand out from the analysis of Figure 3-8. The Galactic conversion
factor underestimates molecular gas mass (abscissa values less than 1.0) and/or model
extrapolated lower temperature of 49 K overestimate the molecular gas mass (ordinate
values greater than 100). However, assuming a high conversion factor, the molecular
gas mass will be over-estimated by CO emission, higher than the dynamical mass
(Solomon et al. (1997)), ruling out this possibility in ULIRGs. The temperature
value of 49 K was the model extrapolated temperature to recover the total molecular
gas mass, estimated for the normal star forming SINGS galaxies. However galaxies
in GOALS are at different merging stages with high FIR color, all pointing to higher
gas temperatures.
Few galaxies with infrared luminosities 1011.66 > LIR/L⊙ > 1011.33 and M(>49K)Mdust
93
Figure 3-8 Galaxies with low M(H2,CO)M(>49K)
, lower than 1.0 correspondingly have
high M(>49K)Mdust
, higher than 100. The solid line show the ratio
value of M(>49K)Mdust
= 100. This may be due to warm temperature
in U/LIRGs resulting in Tℓ to be higher than 49 K.
94
< 100 also show M(H2,CO)M(>49K)
> 1.0. The model molecular gas masses in these galaxies
deviate by a factor of 2-3× from the CO derived masses. Galaxies in the lower infrared
luminosity range, 1011 < LIR/L⊙ < 1011.33, with similar physical ISM conditions of
normal star forming SINGS galaxies are with the ratio M(H2,CO)M(>49K)
≈ 1.0, implying Tℓ
= 49 K is the suitable model extrapolated temperature to recover the molecular gas
mass in them. However, the ratio of model molecular gas to the dust mass, M(>49K)Mdust
, in
a few galaxies are as high as about 1000, implying model extrapolated temperature
to 49 K is about a magnitude higher than the standard value of ≈100. Generally
galaxies with high infrared luminosities, LIR/L⊙ > 1011.66, require Tℓ greater than 49
K, implying warm ISM gas temperatures in these galaxies.
3.5.5 Variation in the L850
MISM
Scoville et al. (2014) empirically calibrated the ratio of 850 µm luminosity to ISM
mass, α850 = L850
MISM, and found to be constant in normal star forming, starburst and
U/LIRGs galaxies. Scoville et al. (2014) had three samples for the calibration: 12
local star forming and star-burst galaxies, galactic observations from Planck and a
sample of 28 SMGs at z<3 with CO(1–0) measurements and estimated α850 of 1.01,
0.79, and 1.01 × 1020 ergs/s/Hz/M⊙, respectively. The ISM masses were calculated
by assuming the atomic gas mass to be half of the molecular gas mass, hence MISM =
1.5 × M(H2). We calculate the ratio of L850 to our model extrapolated mass estimate
to understand the molecular gas properties in U/LIRGs.
Figure 3-9 show the relation of the ratio L850
MISM= α850 with M(H2,CO)
M(H2,T>49K). The
ISM mass in the left panel was calculated using 1.5×M(H2, T > 49K). In the left
panel galaxies with ratio M(H2,CO)M(H2,T>49K)
less than 1.0 (using galactic αCO) have low α850
(than ≈1020 ergs/s/Hz/M⊙, Scoville et al. 2014, 2015), an overestimation of model
molecular gas mass. This implies the model need to be extrapolated to temperatures
higher than 49 K, indicating warm molecular ISM than normal star forming galaxies.
95
Figure 3-9 When used the galactic αCO the ratio, L850
MISMis ≈ 1020
ergs/s/Hz/M⊙, shown on the right plot. The MISM mass is cal-culated using the molecular gas mass, MISM = 1.5×M(H2). Themolecular gas masses used on the left and right plot are fromLCO and power law model, respectively. On the left plot the ra-tio, L850
MISM, increases with the molecular mass ratio derived using
CO luminosity and the model. Galaxies in GOALS have warmISM gas hence, the extrapolated temperature need to be higherthan 49 K. Reducing the power law model derived molecular gasmass will increase the ratio of L850
MISMand M(H2,CO)
M(>49K).
96
In the right panel of Figure 3-9 the ISM mass was estimated using molecular gas
mass derived through CO emission, M(ISM) = 1.5×M(H2, CO). The α850 is constant
and is ( 0.67±0.17)×1020 ergs/s/Hz/M⊙, as estimated by Scoville et al. 2014, 2015.
This suggests the model extrapolation temperature to 49 K possibly overestimates
molecular gas mass and hence, a warmer ISM than compared to normal star forming
galaxies.
3.5.6 Caveats
The conclusion of warm ISM in GOALS sample in the previous two subsections are
subjected to many caveats due to assumptions made. The dust masses were estimated
using a modified blackbody single temperature from the FIR ratio, S60
S100, which can
lead to a factor of more than 2× difference in galaxy’s dust mass (Remy-Ruyer et al.
2015). Moreover a constant dust mass opacity values for all galaxies can add further
error in dust mass estimation. The aperture correction, discussed in section 3.3, to
match CO apertures with the IRS beam to estimate molecular gas masses can be
another possible source of error in the molecular gas to dust mass ratio. Future work
on estimating dust mass using entire IR SED can lead to true estimate of molecular
gas masses.
In the next section for estimating α850 the error arise from L850 and the molecular
gas mass estimation. As discussed in chapter 2, the power law model predicted
molecular gas masses have a scatter of ≈0.34 dex, a factor of 2.20. Also in the
analysis the ISM atomic gas is assumed to be half the molecular gas mass for all
galaxies. A future work on measuring atomic gas, dust, and molecular gas mass for
the same spatial region in galaxies can prove useful in studying and understanding
variation in DGR in GOALS galaxies. This work could also lead to test any change
in DGR for a merger and a normal star forming galaxy.
97
3.6 Summary & Conclusions
We selected ten mapped LIRGs from GOALS to prepare a disk template spectrum.
We used the scaled disk template in addition to the IRS-Spitzer nucleus spectrum
from the photometry flux value constraint from IRAC-Spitzer 8 µm and MIPS-Spitzer
24 µm to get an equivalent spectrum for the galaxy region observed by the large CO
beam to derive H2 line fluxes. Using H2 rotational lines we derived the power law
index for warm molecular gas to estimate the temperature distribution of molecules
and understand molecular gas properties in these mergers.
Our key points of this work are:
1. The disk spectrum of LIRGs are similar to normal galaxies and the effects of
merging processes occur only at the nuclei region of the merger systems.
2. The power law indices (average 4.41±0.49) for our selected GOALS galaxies are
generally found lower than the normal star forming galaxies (average 4.8±0.61).
3. The power law indices decrease with the infrared luminosity, LIR. High infrared
luminosity sources may have high warm gas temperatures due to high star
formation rate and turbulence resulting in lower power law indices.
4. We observe an indirect relation between the power law index and 6.2 µm PAH
equivalent width. Merger systems with low 6.2 µm PAH width are warmer
sources with high AGN fraction.
5. The model predicted molecular gas masses for high infrared luminosity galaxies(
LIR
L⊙
)
> 1011.66, with extrapolation temperature to 49 K show a slight over-
estimate of molecular gas masses than compared to CO derived luminosities
using a Galactic conversion factor when compared to scatter of 2.2× in model
predicted molecular gas masses
98
6. Galaxies with an overestimation of molecular gas masses through our model
when extrapolated to 49 K show high ratio of M(H2>49K)Mdust
, implying possible
overestimation of model molecular gas masses than compared to normal stan-
dard assumption of molecular gas to dust mass ratio.
7. Galaxies with an overestimation of molecular gas masses through our model
when extrapolated to 49 K show low ratio of L850
MISMthan the method when used
a galactic conversion factor to calculate the ISM mass. This suggests warm
molecular ISM in GOALS than compared to SINGS galaxies.
99
Table 3.1. H2 line fluxes for GOALS
Galaxy S(0)(err) S(1)(err) S(2)(err) S(3)(err) S(4)(err) S(5)(err) S(6)(err) S(7)(err)(1) (2) (3) (4) (5) (6) (7) (8) (9)
NGC 23 39.7 (4.06) 239 (5.47) 118 (4.86) 209 (19.6) 79 (11.6) 106 (22.8) · · · 23.4 (15.3)MCG-03-04-014 29.5 (7.28) 104 (7.48) 62.1 (2.11) 103 (3.15) · · · 98.7 (5.54) · · · · · ·
NGC 695 28.4 (10.4) 154 (12.0) 91.3 (4.29) 176 (7.31) 31.5 (10.1) 106 (16.7) · · · 16.1 (14.7)NGC 828 42.9 (9.71) 360 (13.7) 160 (10.8) 288 (25.7) 109 (35.2) 202 (37.4) · · · 8.47 (39.2)IC 0214 22.2 (5.5) 91.4 (7.11) 51.6 (2.95) 81 (4.93) 8.38 (6.41) 72.7 (10.7) · · · 7.51 (7.06)
NGC 877-S 23.1 (2.73) 91.7 (5.02) 24.3 (2.68) 124 (38.3) 18.8 (7.07) 48.4 (22.0) · · · 3.61 (10.8)NGC 958 46.2 (77.6) 214 (142) 74.6 (163) 146 (156) 50.9 (705) 70.0 (307) 19.3 (475) 20.1 (74.4)NGC 992 40.7 (11.7) 221 (15.1) 124 (13.7) 193 (30.8) 56.3 (48.6) 145 (48.7) · · · 2.78 (43.2)
UGC 02238 13.4 (7.11) 199 (9.72) 102 (7.92) 193 (23.4) 57 (27.6) 108 (26.8) · · · · · ·NGC 1365 281 (86.1) 1910 (57.7) 1090 (48.7) 18 (51) 773 (224) 1190 (275) 119 (152) 203 (104)NGC 1614 224 (7.98) 559 (11.1) 308 (7.01) 388 (17.2) 214 (19.2) 504 (48.7) · · · · · ·
IRAS05081+7936 7.92 (6.36) 94.5 (7.11) 55.4 (6.36) 84.2 (13.1) 81.6 (15.7) 75.5 (19.7) · · · 8 (17.9)IRAS05189-2524 13.3 (4.64) · · · · · · 148 (3.7) 110 (5.36) 37.4 (7.17) · · · 22.6 (6.33)
NGC 2146 1030 (24.6) 2420 (23.3) 1410 (35.2) 2970 (83.5) 1040 (86.8) 1180 (192) · · · 326 (145)IRAS 07251-0248 12.9 (4.42) 66.4 (9.93) · · · 82.3 (20.7) 200 (8.52) · · · · · · 224 (14.2)
NGC 2623 86.4 (14.7) 178 (10.4) 60.6 (6.16) 141 (16.3) 111 (17.2) 33 (19.3) · · · 33.7 (16.4)IRAS08572+3915 284 (4.62) 533 (18.6) 62.9 (8.92) · · · 2120 (20.8) · · · 1110 (41.2) 1160 (28.5)
UGC 4881-E 19.0 (4.04) 73.4 (5.62) 24.4 (1.84) 57.8 (44.5) 20.8 (49.7) 27.1 (5.58) · · · 5.68 (4.41)UGC 4881-W 21.7 (4.47) 65.4 (5.03) 21 (1.8) 38.9 (3.86) · · · 23.9 (4.72) · · · 1.89 (4.32)
UGC 5101 21.8 (6.96) 105 (5.29) 50.5 (5.95) 77.3 (15.7) 129 (15.5) 56.9 (15.4) · · · 281 (16.8)
100
Table 3.1 (cont’d)
Galaxy S(0)(err) S(1)(err) S(2)(err) S(3)(err) S(4)(err) S(5)(err) S(6)(err) S(7)(err)(1) (2) (3) (4) (5) (6) (7) (8) (9)
NGC 3110 · · · 305 (12) 149 (4.48) 268 (7.9) 83.7 (14.2) 162 (27) · · · 22.5 (15.1)NGC 3221 54.3 (6.4) 234 (8.53) 105 (5.59) 201 (25.9) 73.6 (14.4) 139 (31.7) · · · 18.5 (58.2)NGC 3256 655 (76.7) 1530 (52.9) 844 (23.3) 1200 (33.1) 442 (93.2) 1030 (176) · · · 93.4 (109)
IRAS 10565+2448 48.9 (6.63) 114 (5.15) 52.2 (5.27) 87.7 (11.5) 76.5 (12.6) 80.6 (15.2) · · · 26.9 (13.1)IC 2810 18.6 (6.01) 63.6 (5.55) 26 (1.57) 45.9 (2.8) 11.4 (3.98) 27.9 (4.27) · · · 5.03 (4.15)
NGC 3690E 1520 (19.7) 1900 (9.24) 582 (4.23) 2030 (8.12) 1140 (10.5) 242 (7.51) · · · 541 (8.14)IRAS 12112+0305 3.73 (4.91) 91.6 (6.33) 20.6 (7.64) 67.7 (14.8) 52.9 (14.2) 20.6 (23.1) · · · 46.2 (16.8)
NGC 4194 324 (428) 447 (508) 220 (625) 379 (559) · · · 228 (1550) · · · 31.7 (213)VV250a-E 89.4 (11.2) 110 (8.19) 58.5 (5.11) 94.3 (11.5) · · · 59.1 (14.6) · · · · · ·VV250a-W 11.4 (3.5) 23.3 (5.05) 12.4 (5.75) 32.7 (14.4) · · · 27.7 (19.4) · · · · · ·UGC 08387 83.5 (11.7) 365 (12.4) 97.8 (2.2) 335 (6.44) 25.8 (5.99) 156 (12) · · · 19 (5.52)NGC 5135 75 (11.2) 350 (9.37) 165 (10.6) 282 (29.7) 81 (30.6) 345 (38.9) · · · 27.9 (24.7)UGC 08696 9.29 (14.4) 306 (8.17) 30.4 (5.56) 403 (23.3) 258 (18.6) 96.9 (17.6) · · · 280 (21.4)NGC 5653 54 (5.31) 311 (6.98) 123 (10.8) 267 (28.5) 71.3 (22.2) 132 (50.4) · · · 28.6 (33.3)
IRAS 14348-1447 3.69 (6.9) 140 (8.25) 26.6 (2.71) 141 (7.68) 86.4 (4.74) 14.6 (6.43) · · · 83.7 (5.04)CGCG049-057 28 (2.95) 124 (3.85) 42 (1.89) 173 (5.39) 72.7 (5.24) 30 (7.22) · · · 49.2 (6.29)
NGC 5936 25 (11.2) 228 (12) 118 (11.4) 175 (34.1) 64.5 (31.2) 150 (41.6) · · · · · ·Arp 220 · · · 1530 (34.2) · · · 1100 (134) 1060 (30.5) 180 (28) · · · 889 (38.8)
NGC 6090 93.3 (2.54) 203 (3) 68.2 (6.45) 126 (9.5) 41.6 (18.9) 84.9 (35.3) · · · 18.4 (28.6)NGC 6240 162 (19.3) 1250 (16.4) 406 (6.5) 1670 (22.3) 625 (21.2) 1220 (17.9) 105 (17.8) 542 (17.1)
101
102
Table 3.1 (cont’d)
Galaxy S(0)(err) S(1)(err) S(2)(err) S(3)(err) S(4)(err) S(5)(err) S(6)(err) S(7)(err)(1) (2) (3) (4) (5) (6) (7) (8) (9)
NGC 6701 54.6 (9.17) 275 (12.6) 122 (10.2) 335 (77) 82.1 (30.1) 65.1 (30.9) 11.6 (29.4)IRAS 19297-0406 30 (7.07) 106 (5.58) 17 (6.76) 83.6 (18.8) 53.6 (15.2) 14.2 (26.8) · · · 38.5 (17.6)CGCCG448-020-E 57 (5.06) 199 (3.79) 22.9 (2.92) 140 (5.35) 31.8 (7.25) 47.3 (8.36) · · · 16.7 (6.5)
NGC 7130 · · · 266 (9.67) 140 (11) 162 (13.3) 125 (38.1) 118 (44.5) · · · 51.3 (33)IC 5179 48.3 (4.63) 464 (7.76) 279 (4.8) 454 (10.4) 163 (17.7) 298 (6.8) · · · · · ·
IRAS 22491-1808 18.4 (5.46) 50.3 (7.32) · · · 65.8 (4.81) 33.8 (4.05) 13.5 (5.42) · · · 14.6 (5.11)NGC 7469 90.2 (35.2) 393 (25.8) 169 (7.02) 200 (8.04) 33.8 (19.7) 137 (20.5) · · · 6.47 (16.5)NGC 7674 91.8 (6.66) 80.7 (10.3) · · · 13.9 (5.24) 85.5 (9.91) 27.9 (10) 103 (15.4) · · ·Mrk 331 39.7 (10.3) 260 (7.60) 103 (6.07) 186 (10.6) 92.4 (16.5) 175 (19.1) · · · · · ·
103
Table 3.2. H2 line fluxes for GOALS
Galaxy D n log (M(>49 K)) log (M(H2,CO) S850S60
S100
(Mpc) mJy(1) (2) (3) (4) (5) (6) (7)
NGC23 58.5 4.31 9.46 9.68 144 (25) 0.60 (0.10)MCG-03-04-014 79.6 4.27 9.85 10.44 · · · 0.62 (0.11)
NGC695 133 4.29 9.97 10.34 136 0.60 (0.10)NGC828 71.4 4.52 9.93 10.14 · · · 0.48 (0.07)IC0214 123 4.4 9.74 10.24 · · · 0.64 (0.11)
NGC877-S 50.6 4.83 9.22 9.63 0.39 (0.08)NGC958 76.5 · · · · · · 10.14 262 0.34 (0.07)NGC992 54.1 4.35 9.38 9.63 146 0.64 (0.08)
UGC02238 88.3 4.36 9.77 10.14 104 0.50 (0.09)NGC1365 21.2 4.41 9.54 10.27 · · · 0.53 (0.07)NGC1614 65.5 4.6 10.1 · · · 219 0.99 (0.06)
IRAS05081+7936 229 4.05 10.1 10.69 · · · 0.58 (0.09)IRAS05189-2524 181 · · · · · · 10.37 · · · 1.20 (0.06)
NGC2146 11.9 4.61 9.26 9.6 · · · 0.71 (0.06)IRAS07251-0248 387 2.83 9.57 10.63 · · · 1.00 (0.13)
NGC2623 80.5 4.51 9.72 9.77 91 0.83 (0.06)IRAS08572+3915 254 3.1 10.3 9.78 17 (7) 1.64 (0.13)
UGC4881-E 169 4.77 10.1 10.46 65 (13) 0.58 (0.08)UGC4881-W 169 5 10.2 · · · · · · —
UGC5101 168 3.31 9.37 10.32 · · · 0.57 (0.09)NGC3110 75.2 · · · · · · 10.3 · · · 0.56 (0.11)NGC3221 61.4 4.45 9.57 9.66 · · · 0.39 (0.08)NGC3256 42.8 5 10.1 10.18 188(28) 0.77 (0.08)
IRAS10565+2448 188 4.46 10.2 10.34 61(13) 0.80 (0.07)IC2810 149 4.76 9.97 10.26 106(18) 0.58 (0.13)
NGC3690E 42.2 · · · · · · · · · · · · 0.97 (0.08)IRAS12112+0305 324 4.34 10.5 10.62 49 (10) 0.85 (0.1)
NGC4194 36.8 · · · · · · 9.28 · · · 0.83 (0.13)VV250a-E 132 4.9 10.2 · · · · · · 0.92 (—–)VV250a-W 132 4.83 9.46 · · · · · · —UGC08387 109 4.8 10.5 9.87 113 (15) 0.61 (0.11)
104
Table 3.2 (cont’d)
Galaxy D n log (M(>49 K)) log (M(H2,CO) S850S60
S100
(Mpc) mJy(1) (2) (3) (4) (5) (6) (7)
NGC5135 60.9 4.34 9.68 10.06 · · · 0.59 (0.01)UGC08696 162 3.96 10.2 10.24 · · · 1.02 (0.06)NGC5653 51.8 4.53 9.6 9.63 205 (32) 0.53 (0.08)
IRAS14348-1447 366 4.16 10.7 10.78 24 (7.2) 0.91 (0.13)CGCG049-057 56.4 4.22 9.08 9.47 200 (27) 0.71 (0.07)
NGC5936 57.5 4.36 9.45 9.63 152 (28) 0.53 (0.09)Arp220 77.2 4.04 10.3 10.28 456 (47) 0.92 (0.05)
NGC6090 123 5.53 10.7 10.15 91 (15) 0.75 (0.06)NGC6240 101 4.25 10.6 10.29 150 (45) 0.82 (0.06)NGC6701 53.4 4.28 9.42 9.61 · · · 0.49 (0.05)
IRAS19297-0406 372 4.57 10.9 10.51 · · · 0.91 (0.13)CGCCG448-020-E 148 5.01 10.6 10.3 · · · 1.26 (0.09)
NGC7130 63.6 · · · · · · 10.16 · · · 0.64 (0.07)IC5179 43.7 4.23 9.44 9.93 · · · 0.52 (0.07)
IRAS22491-1808 334 4.26 10.3 10.43 19 (5.7) 1.22 (0.11)NGC7469 62.7 4.99 10.1 9.96 192 (27) 0.74 (0.09)NGC7674 116 4.11 9.46 10.06 108 (20) 0.69 (0.11)Mrk331 72 4.62 9.86 10.11 132 (25) 0.86 (0.09)
aTo calculate the molecular gas from CO luminosity, L’CO values are from Sanders et al.(1991)
bThe 850 µm luminosities, L850, are from Dunne et al. (2000)
cThe S60
S100are from IRAS catalog (NED Extragalactic Database)
dThe molecular gas masses are in units of M⊙
105
Chapter 4
Molecular gas properties in shocks
of Stephan’s Quintet
4.1 Introduction
Stephan’s Quintet (hereafter SQ) in the constellation of Pegasus is a visual group-
ing of five galaxies: NGC 7317, NGC 7318a, NGC 7318b, NGC 7319 and NGC 7320,
of which the first four (NGC 7320 is a foreground galaxy) form the first discovered
compact galaxy group by Edouard Stephan in 1877 at Marseille Observatory. These
galaxies are of interest because of their violent collisions. SQ is also one of the nearby
compact groups of galaxies and hence, studied extensively at all the wavelengths of
electromagnetic spectrum from X-rays (Bahcall et al. 1984, Sulentic et al. 1995,
Trinchieri et al. 2003, O Sullivan et al. 2009), UV (Xu et al. 2005), visible (Moles et
al. 1998, Gallagher et al. 2001, Fedotov et al. 2011), IR (Sulentic et al. 2001; Natale
et al. 2010, Cluver et al. 2010; Guillard et al. 2010; Suzuki et al. 2011; Bitsakis
et al. 2014), to radio waves (Allen & Hartsuiker 1972; van der Hulst & Rots 1981;
Xu et al. 2003; Nikiel-Wroczynski et al. 2013). Previous studies have detected CO
(1–0) emission from brightest regions in NGC 7319 and in the two extragalactic star
forming regions (Gao & Xu 2000, Smith & Struck 2001, Lisenfeld et al. 2002, Petitpas
106
Figure 4-1 Stephan’s Quintet in the constellation of Pegasus is a visualgrouping of five galaxies: NGC 7317, NGC 7318a, NGC 7318b,NGC 7319, and NGC 7320. NGC 7320 is a foreground galaxy.Image credit: amazing-space.stsci.edu
107
& Taylor 2005). The galaxy NGC 7318b is an intruder galaxy, between NGC 7318a
and NGC 7319, and the IGM between NGC 7318b and NGC 7319 is shock heated to
high temperatures and glows in the X-ray wavelengths. Shocks excite H2 molecules
efficiently and hence, IRS-Spitzer MIR spectroscopy revealed bright, broad, pure H2
rotational line emission in these regions (Appleton et al. 2006).
In this work we prepared an extensive grid of IRS spectra pertaining to the exci-
tation of H2. We derive H2 rotational line fluxes for each region in the grid and using
the power law model estimated the physical properties of IGM in the shocks of SQ.
4.2 H2 in shocks of Stephan’s Quintet
The gas in the SQ group halo mainly constitutes of neutral atomic HI clouds and
possibly H2 clouds embedded in hot intercloud plasma. The absence of dust features
or lines of ionized gas suggests the presence of warm molecular gas with little or no
star formation at all. Shocks created due to intruding galaxy and a tidal arm at a
relative speed of about 1000 km/s excite H2 molecules, which result in large flux of
H2 rotational line fluxes. Its necessary to understand in shock regions
1. why is H2 present in postshock gas?
2. how can we account H2 excitation?
3. why H2 is the dominant coolant?
H2 formation results from the density structure of the preshock gas. The collision
velocity is the shock velocity in low density regions ( 0.01 cm−3) which produces
dust free X-ray emitting plasma. The pressure of the plasma drives shocks with slow
speed (5–20 km/s) into dense gas in clouds. Gas with temperatures less than 106 K
preserved dust and has a cooling time less than the collision age ( 5106 yr). Due to
high postshock pressure the warm neutral medium is unstable the gas cools to cold
neutral medium temperatures, condenses and form molecules. Guillard et al. (2009)
108
have shown for a wide range of postshock pressures, preshock cloud sizes and densities
the H2 formation timescale is lower than the collision age and the turbulent mixing
timescale of the warm gas with the plasma.
Observations have shown that the non thermal kinetic energy of H2 is higher
than the thermal energy of the plasma. H2 emission is powered by the dissipation of
this kinetic energy. Guillard et al. 2009 have reproduced H2 line emission fluxes by
appropriate combination of continuous (C-type) and jump (J-type) shock velocities
with suitable preshock densities.
The thermal, Rayleigh-Taylor and Kelvin-Helmholtz instabilities produces frag-
mentary postshock clouds where atomic and molecular clouds are mixed on small
spatial scales and embedded in the hot plasma. The efficient transfer of the kinetic
energy associated with the bulk displacement of clouds to turbulent motions of low
velocity within molecular gas, is essential for H2 to be a dominant coolant of the
postshock gas.
There is no star formation in the shock regions of SQ, where H2 is detected suggests
that the molecular fragments are short lived, not enough to collapse and become
gravitationally unstable.
4.3 Observations, Data reduction and Analysis
The spectroscopy observations of SQ were made using a low resolution mode (R
= 60–127) of the IRS-Spitzer instrument. The LL module spectrally mapped an
area of ∼2.8 × 3.2 arcmin2 in 21 steps of 8 arcsec (0.75 × slit width), and was
designed to cover whole radio/X-ray emitting filament. Similarly the SL module
(covering a slightly smaller area than LL) performed two partially overlapping scans
perpendicular to the LL slit, in 2 × 23 steps of 2.8 arcsec (0.75 × slit width) for each
sub-module. All the individual frames were median combined, calibrated, corrected,
109
and assembled into 2D data cubes using the CUBISM (Smith et al. 2007) software
after processing through the Spitzer Science Center S17 pipelines. This resulted in
cubes for the SL and LL mapped onto the sky with a pixel scale of 1.85 × 1.85 arcsec2
and 5.1 × 5.1 arcsec2, respectively. Inorder to ensure that the extracted spectra of
the H2 line fluxes are at a common spatial resolution, set by the S(0) line occurring
at the maximum rest wavelength at 28.22 µm, we convolved at each wavelength, the
individual layers of the SL2 (5.2–7.7 µm), SL1 (7.4–14.5 µm), LL2 (14.0–21.3 µm)
cubes to the scale of 7 arcsec (FWHM of the 28 µm line) using a Gaussian kernel.
It was not necessary to smooth the LL1 (19.5–38 µm) cube because the only line of
interest in this paper is 28.2 µm line, and so the cube was left in the native form.
This procedure has two effects that our spectral extractions were not affected by the
resolution difference over the 5–28 µm wavelength range, and the procedure improved
S/N ratio of the shorter wavelength data significantly.
4.3.1 MIR H2 emission and spectra
Previous MIR spectral observations have shown strong rotational H2 emission
lines, especially in the main shocked filament (Appleton et al. 2006). We aim to
study spatial trends in the temperature and density of gas over the whole region that
was mapped by both IRS-LL and IRS-SL instrument. Spectra were extracted from
the SL2, SL1, LL2, and LL1 cubes and then stitched together to prepare the full MIR
spectra. In general, no scaling was required in stitching the spectra together. A small
subset of SL2 spectra required a small positive correction due to an over subtraction
of the background reference spectrum.
The IRS spectral extractions were performed in CUBISM on the SL and LL cubes
using the four corners of 3 × 3 native SL pixels to define the extraction areas over
a continuous grid. This resulted in boxes of area 5.5 × 5.5 arcsec2. which is about
half the width of the LL slit. Figure 4-2 shows the extraction grid of 212 spectra
110
superimposed on a B-band HST image of the Quintet. We only sample regions that
are well covered in both the SL and LL cubes, and so we have been conservative
about the edges of the mapped area to ensure good coverage at all wavelengths
4.4 Excitation diagram Fits
H2 being a symmetric molecule has no allowed dipole transitions, and only weak
quadrupole transitions are allowed. Excitation diagrams are a convenient way of
exploring molecular distribution of level populations in galaxies. We compare two
approaches to fit the H2 excitation diagram derived from the 2D IRS mapping of the
SQ. First, we employ the traditional one, two or three discrete temperature fits and
the second, by assuming a single power law distribution of temperatures. Both the
methods assume that the gas is in thermal equilibrium. We will discuss how each
method provides insight into the system leading to self consistent conclusions.
4.4.1 Discrete temperature fit
Assuming the MIR lines of H2 are optically thin, the column density of the upper
energy level (Nu) of each pure rotational transition is measured observationally from
the spectral line flux, F, according to
Nu =4πF
hνAω, (4.1)
where h, ν, A and ω are Planck’s constant, frequency, Einstein’s coefficient of the
transition and solid angle of the observed region, respectively (same equation as 2.8).
In a local thermodynamic equilibrium (LTE) the upper energy level column density is
related to both T and Ntot, the excitation temperature and the total column density,
111
Figure 4-2 Spectral extraction grid used to investigate the excitation of thegas over the main H2 filament, and includes areas common toboth the long-low and short-low modules. The numbered ex-traction boxes are 5.55 × 5.5 arcsec2 (2.5 × 2.5 kpc2, assumingD = 94 Mpc). The background image is using the F665 N HSTimage. The overall extent of the warm H2 emission is about 50× 35 kpc2.
112
Figure 4-3 A two temperature fit, T1 = 219 K and T2 = 722 K, of theexcitation diagram of H2 for the square region 105 in the centerof the SQ’s filament. The values for each of the seven H2 linesare shown with 1-σ uncertainties.
respectively, by
Nu
gu= Ntot
e−Eu/kT
Z(T )(4.2)
The H2 excitation is usually presented as a plot of the ln[Nu/gu] versus Eu/K, where
the slope of the line is proportional to 1/T.
We assumed the ortho and para H2 for the lower pure rotational H2 transitions are
in equilibrium. For H2 densities ≥ 103 cm−3, most of the lower rotational transition
lines are thermalized and temperature derived from the fits to the ortho and para H2
transitions should yield consistent temperatures (Roussel et al. 2007). Burton et al.
113
1992 showed that in collisional equilibrium, the ortho-to-para (OPR), increases from
about unity at T = 75 K, to a constant value of 3 at T ≥ 300 K temperatures. Since
our temperatures from the rotational H2 line transitions are in the range 120 > T >
1500 K, the equilibrium values are in the range 1.8–3.0. After normalizing for this
factor, significant deviations from LTE would appear on an excitation diagram. No
evidence for such deviations from equilibrium are observed in our detections.
An additional complication is that a single temperature fit is usually not able
to recover all the H2 line fluxes. A two temperature fit usually fits the excitation
diagram better. In the literature this is a common approach because it provides a
first approximation to the temperatures, column densities, and the total warm H2
masses. However a unique discrete temperature fit may not exist for an excitation
diagram and hence a need for more robust technique is required.
Single temperature solutions were found, where there were fewer than four tran-
sitions (22% of the data) and two temperature solutions were required for the other
transitions. The process for fitting a two temperature LTE model is iterative. We
created an initial excitation diagram with the line fluxes (derived from Table 1) for
each of the 212 extracted square regions. Initially we assumed an OPR of 3 to obtain
a first approximation to the temperature minimizing the chi-squared deviations for
a two temperature model. Once the chi-squared values are minimized, we assumed
those temperatures as a first guess and adjust the OPR appropriate for LTE from
Burton et al. (1992) and re-run and this process is repeated for several iterations
until a best fit value for the T and column density are determined. Figure 4-3 show
an example of such a fit for an extracted square region numbered 105 and in Table
4.2 we list the two temperature values for each 212 extracted regions.
[*]-1in
114
Table 4.1. MIR H2 line flux for Quintet
Reg S(0)(err) S(1)(err) S(2)(err) S(3)(err) S(4)(err) S(5)(err) S(6)(err)(1) (2) (3) (4) (5) (6) (7) (8)
1 — — — — — — —2 — 2.48(0.575) 3.99(0.706) — — 7.89(0.443) —3 — 4.74(0.596) 7.14(0.991) — — — —4 1.14(0.425) 13.2(0.612) 10.1(0.839) 30.5(1.25) — — —5 — 12.1(0.57) 6.61(1.6) 21.3(1.25) — — —6 — 9.12(0.633) — — — 2.92(0.441) —7 — 10.6(0.638) 8.47(4.0) 17.7(1.25) — 21.8(3.0) —8 — 11.1(0.494) — 21.4(1.25) — 8.61(0.546) —9 — 9.19(0.52) 6.63(0.811) 17.5(1.33) — — —10 2.21(0.479) 8.11(0.444) — 14.8(1.38) — 5.23(0.433) —11 — 2.42(0.416) 2.42(0.703) 2.67(1.11) — — —12 — 8.76(0.551) 5.23(0.767) 9.12(1.24) — — —13 2.29(0.424) 22.9(0.602) 8.03(0.774) 18.8(1.25) — — —14 2.62(0.426) 23.4(0.562) 15.1(0.774) 33.7(1.25) 7.38(1.63) 26.2(1.4) 16.7(1.3)15 1.69(0.507) 16.7(0.562) 11.0(0.767) 16.7(1.24) 4.86(1.66) 16.6(1.4) 16.6(1.3)16 1.35(0.429) 15.2(0.612) 7.53(0.839) 9.99(1.24) — — —17 2.95(0.481) 18.1(0.594) 11.5(0.818) 23.4(1.3) 5.41(1.7) 14.0(1.3) —18 1.12(0.567) 12.0(0.533) 7.53(0.839) 11.0(1.25) — — —19 — 7.06(0.531) 5.62(0.832) — — — —20 1.85(0.472) 7.23(0.546) 4.54(0.847) 5.75(1.11) — 8.76(1.4) 15.6(1.3)21 — 3.88(0.523) 5.96(0.767) — — — —22 — 10.6(0.551) 6.27(0.767) 5.14(1.14) — — —23 1.97(0.557) 19.0(0.572) 11.7(0.774) 15.3(1.87) — — —24 2.11(0.426) 23.7(0.569) 14.5(0.774) 21.1(1.79) — 15.2(1.3) —25 2.41(0.426) 22.1(0.558) 13.0(0.832) 19.3(1.87) — 15.6(1.3) —26 2.27(0.507) 15.4(0.585) 11.2(0.818) 13.7(1.15) — 11.3(1.3) —27 — 11.7(0.538) — 14.5(1.77) — 13.7(1.3) —28 — 10.7(0.57) — 6.42(1.14) — 6.56(1.3) —29 — 7.6(0.543) 6.46(0.767) — — — —30 — 6.04(0.551) 7.53(0.839) — — — —31 1.14(0.424) 2.37(0.553) 2.94(0.7) — — — —
115
Table 4.1 (cont’d)
Reg S(0)(err) S(1)(err) S(2)(err) S(3)(err) S(4)(err) S(5)(err) S(6)(err)(1) (2) (3) (4) (5) (6) (7) (8)
32 1.01(0.5) 4.57(0.512) 3.34(0.774) — — — —33 1.9(0.479) 9.99(0.557) 5.05(0.941) 7.1(1.14) — — —34 0.897(0.516) 20.7(0.569) 11.7(0.811) 20.1(1.25) 4.57(1.68) 12.8(0.601) —35 2.75(0.43) 19.5(0.567) 13.2(0.839) 25.3(1.71) — 14.7(0.9) —36 3.04(0.479) 13.7(0.553) 9.19(0.839) 13.7(2.03) — 3.73(1.38) —37 1.66(0.8) 10.8(0.528) 8.54(0.999) 7.89(1.15) — — —38 — 9.99(0.625) 8.32(0.818) 6.05(1.15) — — —39 — 8.54(0.679) 4.48(0.861) 5.39(1.15) — — —40 1.17(0.426) 7.59(0.5) 5.69(0.89) 5.18(1.15) — — —41 — 2.95(0.479) 3.74(0.818) 7.11(1.26) — — —42 1.49(0.42) 6.2(0.57) 6.63(0.79) 8.8(1.26) — — —43 2.33(0.553) 16.4(0.562) 8.32(0.839) 10.5(1.28) — — —44 1.6(0.478) 21.3(0.557) 12.2(0.839) 19.5(1.25) — 8.32(2.5) —45 1.47(0.425) 21.1(0.554) 13.2(0.839) 29.7(1.3) 6.14(1.65) 14.0(1.5) —46 3.09(0.426) 17.7(0.541) 7.82(0.825) 16.9(1.25) — 10.6(1.8) —47 2.32(0.548) 10.6(0.584) 6.94(0.839) 11.2(1.28) — — —48 1.69(0.527) 10.3(0.596) 6.69(0.818) 10.6(1.28) — — 3.9(0.679)49 2.06(0.425) 11.4(0.623) 4.5(0.731) 10.3(1.28) — 3.29(0.437) —50 — 9.84(0.654) — 8.25(1.28) — — —51 — 4.23(0.612) — — — — —52 2.35(0.47) 9.19(0.566) 6.97(0.839) 5.54(1.25) — 13.3(1.5) —53 2.16(0.433) 16.8(0.575) 13.5(0.839) 15.9(1.31) 3.35(1.9) 8.11(0.446) —54 3.49(0.48) 25.0(0.572) 21.2(0.847) 32.4(1.3) 5.83(1.86) 18.4(0.492) —55 4.03(0.428) 34.8(0.575) 21.6(0.839) 41.6(1.25) 10.9(2.04) 20.4(0.648) 6.66(0.581)56 4.05(0.426) 27.1(0.557) 17.4(0.839) 29.1(1.25) 6.02(—9.99e+21) 13.1(0.453) 8.25(0.714)57 2.55(0.424) 18.2(0.558) 9.19(0.832) 10.1(1.26) 4.08(1.64) — —58 2.32(0.422) 13.8(0.52) 6.77(0.825) 7.67(1.26) — 13.2(1.5) —59 2.79(0.423) 11.9(0.613) 6.59(0.825) 5.48(1.15) — 5.08(0.624) —60 1.71(0.47) 10.6(0.469) 5.33(0.832) 3.49(1.15) — 11.1(3.0) 3.9(0.711)61 — 2.34(0.484) 4.58(0.745) — — — —62 1.69(0.426) 8.39(0.509) 4.26(0.731) 4.23(1.29) — — —
116
Table 4.1 (cont’d)
Reg S(0)(err) S(1)(err) S(2)(err) S(3)(err) S(4)(err) S(5)(err) S(6)(err)(1) (2) (3) (4) (5) (6) (7) (8)
63 2.97(0.43) 16.1(0.623) 8.54(0.839) 9.91(1.32) — 2.86(0.443) —64 4.36(0.431) 25.2(0.639) 13.5(0.839) 18.4(1.32) — 7.82(2.0) —65 3.93(0.504) 28.9(0.57) 17.9(0.839) 31.1(1.25) 7.67(1.87) — —66 3.68(0.424) 33.2(0.568) 19.0(0.839) 35.2(1.7) 9.84(1.87) — —67 3.56(0.428) 25.7(0.555) 12.9(0.839) 18.7(1.74) 5.59(1.65) — —68 2.77(0.428) 17.0(0.6) 7.11(0.839) 10.3(1.66) — — —69 1.99(0.523) 12.6(0.6) 5.69(0.796) 7.18(1.15) — — —70 2.05(0.554) 10.9(0.575) 3.97(0.818) 6.79(1.15) — — —71 — 2.18(0.575) — 3.13(1.12) — — —72 1.22(0.429) 6.07(0.613) — 3.44(1.13) — — —73 2.97(0.538) 12.2(0.627) 3.91(0.738) 5.88(1.33) — — —74 3.29(0.504) 16.0(0.619) 7.53(0.839) 11.4(1.7) — — —75 3.67(0.523) 22.8(0.585) 7.53(0.839) 14.1(1.69) — — —76 3.73(0.427) 34.7(0.575) 15.3(0.825) 29.7(1.25) 8.39(1.71) 7.6(0.774) —77 4.94(0.428) 36.1(0.565) 23.3(0.825) 46.3(1.3) 9.26(1.88) 24.9(0.533) 7.2(1.5)78 4.15(0.433) 26.4(0.609) 12.2(0.839) 28.0(1.28) 4.83(1.9) 14.8(1.0) —79 2.74(0.426) 16.9(0.575) 4.56(0.825) 11.2(1.27) — — —80 2.32(0.429) 13.8(0.578) 4.49(0.847) 10.9(1.3) — — —81 1.98(0.43) 13.2(0.578) 3.95(0.745) 8.54(1.25) 8.32(1.9) — —82 1.56(0.425) 13.8(0.546) 3.63(0.731) 12.3(1.25) — — —83 2.0(0.423) 14.1(0.479) — 11.8(1.25) — — —84 — — — — — — —85 0.637(0.48) 3.43(0.628) — 7.6(1.3) — — —86 2.08(0.549) 9.7(0.621) 2.97(0.839) 7.45(1.3) — — —87 3.26(0.427) 16.0(0.604) 4.72(0.847) 11.7(1.3) — 2.86(0.8) —88 3.39(0.429) 20.3(0.575) 6.86(0.825) 16.3(1.3) 4.74(2.11) — —89 4.07(0.524) 31.7(0.576) 13.2(0.818) 31.1(1.3) 6.25(1.69) 9.55(0.9) —90 4.52(0.519) 41.5(0.559) 20.6(0.825) 52.0(1.3) 13.5(2.06) 25.9(1.3) —91 5.31(0.429) 36.1(0.567) 19.0(0.839) 48.1(1.29) 10.9(2.45) 29.7(1.2) —92 3.93(0.426) 25.3(0.574) 9.34(0.839) 22.0(1.29) 8.32(2.02) 11.1(1.2) —93 3.31(0.429) 18.8(0.622) 5.46(0.825) 15.0(1.3) 5.94(1.98) 2.72(0.8) —
117
Table 4.1 (cont’d)
Reg S(0)(err) S(1)(err) S(2)(err) S(3)(err) S(4)(err) S(5)(err) S(6)(err)(1) (2) (3) (4) (5) (6) (7) (8)
94 3.63(0.562) 17.9(0.564) 7.31(0.825) 15.2(1.25) 9.7(2.0) — —95 4.13(0.479) 18.5(0.587) 7.31(0.818) 17.8(1.26) — — —96 2.58(0.431) 18.8(0.585) 7.6(0.811) 14.4(1.3) — — —97 2.04(0.427) — — 6.26(—9.99e+21) — — —98 — 2.33(0.504) — 6.8(—9.99e+21) — — —99 0.941(0.43) 6.81(0.621) — 4.14(1.3) — — —100 2.16(0.43) 12.0(0.634) 1.92(0.738) 7.45(1.3) — — —101 2.71(0.432) 16.7(0.622) 4.77(0.825) 12.8(1.3) — — —102 3.89(0.429) 29.4(0.576) 10.4(0.825) 26.8(1.3) 4.82(1.9) 12.5(1.2) —103 3.98(0.543) 43.9(0.576) 19.5(0.839) 51.6(1.29) 11.6(1.9) 24.7(1.3) 5.64(1.3)104 5.43(0.43) 43.3(0.578) 22.5(0.825) 59.4(1.3) 17.3(2.49) 36.3(1.4) 7.31(1.3)105 4.43(0.426) 33.3(0.561) 15.8(0.839) 35.4(1.28) 8.76(1.79) 23.9(1.4) 4.86(1.2)106 4.78(0.559) 27.1(0.612) 7.38(0.839) 20.6(1.3) 5.49(2.04) 7.19(0.9) —107 4.43(0.567) 23.4(0.621) 6.83(0.818) 18.5(1.27) 8.97(2.08) — —108 4.99(0.478) 26.1(0.561) 8.47(0.825) 22.2(1.3) 11.5(2.0) 6.85(1.2) —109 5.21(0.554) 33.1(0.672) 11.7(0.818) 25.2(1.29) 8.47(2.18) 11.9(1.2) —110 2.04(0.548) 3.37(0.685) — — — — —111 — 3.44(0.688) — 7.45(1.27) — — —112 — 3.76(0.753) — 6.78(1.3) — — —113 1.5(0.53) 7.96(0.481) — 10.6(1.27) — — —114 2.39(0.429) 12.0(0.625) 2.36(0.839) 12.3(1.27) — — —115 3.37(0.432) 24.7(0.624) 7.67(0.847) 24.6(1.32) 4.52(1.85) 11.9(1.3) 3.9(0.667)116 4.9(0.546) 43.1(0.623) 19.3(0.847) 58.6(1.29) 14.7(2.82) 33.4(1.4) 6.21(1.4)117 4.86(0.425) 45.4(0.572) 22.4(0.825) 64.3(1.28) 18.4(1.89) 35.4(1.4) 7.06(1.4)118 4.72(0.428) 37.1(0.567) 15.1(0.839) 44.5(1.29) 13.7(2.83) 26.2(1.2) —119 4.8(0.558) 32.9(0.612) 10.5(0.839) 31.3(1.3) 10.2(2.01) 17.1(1.4) —120 5.41(0.43) 35.2(0.581) 13.2(0.839) 34.6(1.25) 9.41(2.26) 14.4(1.2) —121 5.91(0.568) 36.1(0.581) 15.1(0.839) 33.7(1.3) 8.61(1.85) 15.8(1.4) 6.12(1.4)122 4.81(0.428) 36.9(0.539) 14.7(0.839) 30.3(1.3) 9.91(1.83) 16.6(1.2) —123 2.16(0.474) 2.55(0.443) — — — — —124 1.21(0.476) 3.26(0.8) — — — — —
118
Table 4.1 (cont’d)
Reg S(0)(err) S(1)(err) S(2)(err) S(3)(err) S(4)(err) S(5)(err) S(6)(err)(1) (2) (3) (4) (5) (6) (7) (8)
125 0.223(0.479) 3.22(0.57) — 5.67(1.3) — — —126 — 5.99(0.475) — 5.87(1.29) — — —127 2.2(0.477) 8.11(0.631) — 6.9(1.3) — — —128 2.45(0.473) 17.7(0.563) 4.65(0.767) 14.3(1.32) 6.76(1.72) — —129 4.65(0.475) 35.1(0.625) 12.3(0.825) 33.1(1.3) 8.32(1.87) 17.4(1.3) 5.42(1.3)130 4.8(0.53) 41.0(0.572) 18.0(0.825) 47.3(1.27) 13.0(1.88) 22.9(1.3) 7.11(1.3)131 4.96(0.429) 36.5(0.562) 14.6(0.839) 38.1(1.25) 10.8(1.84) 19.0(1.4) 5.32(1.4)132 5.55(0.43) 37.5(0.611) 13.8(0.839) 37.0(1.25) 7.96(1.82) 18.7(1.4) —133 5.25(0.43) 35.8(0.624) 15.8(0.818) 37.6(1.25) 12.5(1.78) 14.6(1.3) 9.05(1.3)134 6.22(0.539) 35.9(0.615) 16.6(0.818) 35.8(1.25) 15.1(1.81) 14.6(1.3) 12.6(1.3)135 4.44(0.427) 35.6(0.478) 9.84(0.818) 20.3(1.66) 8.25(1.87) 11.9(1.4) —136 1.32(0.478) 3.11(0.562) — — — — —137 1.4(0.476) 4.04(0.59) — — — — —138 1.31(0.566) 6.58(0.484) — — — — —139 2.27(0.477) 9.84(0.628) 5.23(0.818) 6.08(1.25) — — —140 2.57(0.564) 19.0(0.619) 5.48(0.926) 14.9(1.28) — — —141 3.82(0.43) 31.0(0.588) 11.0(0.818) 30.5(1.3) — 11.9(1.3) —142 6.39(0.568) 43.9(0.578) 19.4(0.839) 47.4(1.3) 9.34(1.74) 25.3(0.589) 6.3(0.58)143 6.56(0.432) 40.4(0.567) 18.2(0.825) 47.1(1.3) 11.6(2.08) 24.5(0.524) 8.25(0.583)144 4.99(0.431) 33.3(0.611) 8.47(0.839) 24.6(1.25) 8.54(1.7) 9.19(1.3) —145 5.12(0.548) 31.5(0.612) 6.69(0.948) 19.6(1.25) 10.9(1.65) — —146 4.56(0.563) 19.8(0.601) 5.06(0.811) 12.7(1.25) 6.74(1.67) 14.3(1.3) —147 2.0(0.477) 4.07(0.56) 3.92(0.854) — — — —148 — 3.7(0.556) — — — — —149 — 4.83(0.476) 3.64(0.839) — — — —150 1.6(0.428) 6.69(0.478) 5.65(0.832) 3.62(1.26) — — —151 2.3(0.568) 12.9(0.612) 5.58(0.839) 8.47(1.25) — — —152 3.64(0.43) 29.8(0.58) 9.41(0.847) 22.1(1.46) — — —153 6.3(0.583) 46.5(0.579) 22.9(0.839) 47.0(1.5) 11.5(1.74) 26.4(1.4) —154 6.91(0.432) 45.4(0.57) 20.0(0.818) 41.7(1.63) 9.7(2.0) 29.2(1.4) 8.18(1.4)155 5.39(0.429) 33.0(0.609) 11.0(0.839) 18.8(1.75) 4.88(1.7) — —
119
Table 4.1 (cont’d)
Reg S(0)(err) S(1)(err) S(2)(err) S(3)(err) S(4)(err) S(5)(err) S(6)(err)(1) (2) (3) (4) (5) (6) (7) (8)
156 4.13(0.43) 19.9(0.611) 6.45(0.839) 9.91(1.15) 7.11(1.61) — —157 3.55(0.564) 14.5(0.615) 5.51(0.818) 7.09(1.27) 3.55(1.61) — —158 1.44(0.428) 2.71(0.581) 1.8(0.861) 3.42(1.3) — — —159 — 2.75(0.559) — — — — —160 — 4.83(0.567) — — — — —161 — 5.7(0.622) 2.76(0.774) — — — —162 2.5(0.429) 10.4(0.477) 2.58(0.738) 7.6(1.24) — — —163 3.13(0.426) 20.3(0.646) 5.91(0.796) 9.48(1.24) — — —164 5.54(0.476) 38.0(0.581) 14.3(0.839) 28.4(1.24) 7.38(2.03) 16.5(1.3) —165 5.7(0.428) 42.6(0.577) 18.2(0.839) 35.7(1.27) 12.4(2.04) 24.2(1.4) —166 5.5(0.431) 29.7(0.61) 13.8(0.839) 25.3(1.3) 4.82(1.73) 17.2(1.4) —167 4.0(0.43) 19.0(0.572) 7.67(0.839) 11.2(1.16) 4.54(1.98) 10.3(0.533) —168 3.32(0.429) 15.0(0.622) 5.46(0.767) 8.76(1.66) 4.21(1.66) — —169 1.61(0.426) 5.83(0.475) 1.65(0.724) — — — —170 2.55(0.478) 5.04(0.623) 4.12(0.731) — — — —171 1.79(0.429) 5.38(0.556) 3.07(0.825) 4.96(1.26) — — —172 1.32(0.428) 6.18(0.582) 2.36(0.745) 4.02(1.25) — — —173 2.71(0.43) 11.0(0.478) 1.63(0.745) 5.93(1.46) — — —174 3.72(0.43) 18.5(0.588) 4.1(0.753) 9.34(1.3) — 4.75(1.3) —175 6.4(0.567) 28.2(0.577) 10.4(0.818) 18.1(1.25) 3.43(1.79) 9.84(1.3) —176 7.22(0.565) 32.3(0.577) 15.2(0.818) 25.8(1.28) 15.1(1.82) 16.8(1.4) —177 6.04(0.432) 29.5(0.571) 11.6(0.948) 24.5(1.3) 6.06(2.08) 14.3(1.4) —178 4.93(0.43) 23.2(0.572) 7.02(1.01) 16.0(1.41) 6.33(1.95) 16.6(1.4) —179 4.44(0.429) 20.7(0.61) 6.2(1.16) 6.15(1.34) — 15.1(1.4) —180 2.4(0.472) 5.98(0.473) 4.02(0.839) — — — —181 2.63(0.425) 6.71(0.602) 4.66(0.714) — — — —182 3.06(0.426) 11.4(0.571) 6.06(0.955) 7.6(1.29) — 6.41(1.4) —183 4.23(0.43) 16.8(0.557) 6.36(0.825) 10.6(1.25) — 6.3(1.0) —184 4.76(0.429) 18.6(0.622) 5.73(0.847) 7.74(1.29) — — —185 7.38(0.569) 21.1(0.579) 5.68(0.825) 9.26(1.31) — 7.67(1.0) —186 6.14(0.429) 24.8(0.583) 8.11(0.825) 13.5(1.25) 4.25(1.74) 9.84(1.0) —
120
Table 4.1 (cont’d)
Reg S(0)(err) S(1)(err) S(2)(err) S(3)(err) S(4)(err) S(5)(err) S(6)(err)(1) (2) (3) (4) (5) (6) (7) (8)
187 8.25(0.43) 33.7(0.58) 12.7(0.825) 20.8(1.3) 4.02(1.76) 15.5(1.3) 5.04(1.5)188 7.53(0.431) 35.5(0.583) 10.5(0.825) 20.6(1.25) 5.71(1.78) — —189 7.96(0.555) 29.2(0.563) 8.68(0.818) 19.8(1.25) 6.29(2.13) 14.9(1.3) —190 5.84(0.568) 24.0(0.61) 8.9(0.839) 11.9(1.77) — 13.8(1.3) —191 1.77(0.425) 5.2(0.612) 5.85(0.774) 3.1(1.47) — — —192 2.94(0.43) 8.9(0.567) 7.74(0.839) 7.11(1.25) — — —193 3.85(0.554) 16.4(0.546) 10.1(0.825) 8.97(1.25) — — —194 6.15(0.477) 22.1(0.558) 14.2(0.839) 17.8(1.25) — 14.2(1.3) —195 6.48(0.543) 26.3(0.622) 13.2(0.839) 17.9(1.25) 5.75(1.81) 20.8(1.3) —196 8.11(0.428) 28.4(0.659) 11.7(0.847) 17.4(1.25) — 16.5(1.3) —197 9.99(0.43) 37.3(0.626) 13.8(0.847) 21.0(1.32) 3.32(1.72) 7.14(1.3) 4.96(1.3)198 12.4(0.44) 47.0(0.627) 18.7(0.839) 30.4(1.32) 5.74(1.74) 16.3(1.3) —199 9.34(0.564) 41.2(0.619) 17.7(0.847) 31.6(1.3) 5.02(1.74) 15.6(1.3) —200 6.94(0.562) 31.3(0.572) 14.3(0.839) 23.9(1.25) — 12.3(1.3) —201 5.3(0.579) 24.8(0.56) 12.0(0.803) 14.4(1.69) — — —202 1.14(0.425) 4.92(0.561) 6.59(0.774) 5.74(1.25) — — —203 3.0(0.541) 10.7(0.624) 4.31(0.796) 12.1(1.25) — — —204 4.68(0.558) 16.0(0.572) 5.3(0.839) 10.2(1.25) — — —205 5.63(0.475) 23.4(0.561) 9.63(0.825) 19.0(1.25) — — —206 6.72(0.561) 29.4(0.554) 12.9(0.825) 22.1(1.26) 14.7(1.87) — —207 9.63(0.427) 45.6(0.606) 16.8(0.847) 29.8(1.26) 19.3(1.87) 10.7(1.3) —208 13.0(0.564) 46.0(0.615) 16.1(0.839) 32.9(1.32) 20.8(2.18) 16.3(1.3) —209 12.2(0.578) 49.3(0.619) 17.9(0.839) 37.9(1.3) 15.7(2.09) 22.6(1.3) —210 11.9(0.432) 49.1(0.638) 18.5(0.839) 35.2(1.32) — — —211 7.0(0.567) 35.0(0.61) 13.8(0.825) 19.5(1.25) — — —212 4.33(0.43) 18.9(0.61) 9.55(0.817) 13.3(1.25) — — —
aAll values in 10−19 W m−2
121
Table 4.2. Power law model results for Quintet
Reg T1 (T2) M(> T1) M(> T2) n Tℓ M(> Tℓ)[M(T>120K)]
Mtotal
(K) (K) (106M⊙) (104M⊙) (K) (106M⊙) (%)(1) (2) (3) (4) (5) (6) (7) (8) 9
1 —– —– —– —– —– —– —– —–2 —– —– —– —– —– —– —– —–3 100 582.5 6.528 6.4 —– —– —– —–4 133.4 484.4 2.496 17.41 3.44 50 10.8 11.85 176.1 554.5 1.728 7.232 3.84 80 4.71 31.76 —– —– —– —– —– —– —– —–7 —– —– —– —– 3.78 80 3.9 32.48 187.7 637.5 1.216 4.416 4.19 80 6.08 27.49 125.1 466.6 3.776 12.10 3.65 80 2.97 34.110 138.7 602 5.76 3.776 4.37 50 25.6 5.2211 —– —– —– —– —– —– —– —–12 142.7 432 2.56 10.56 4.24 80 5.01 26.913 195.6 505.3 2.56 8.96 4.62 110 5.8 7314 215.7 906 2.368 3.264 3.92 50 38.7 7.7915 275.8 1500 0.896 0.064 3.79 50 23.2 8.6616 170.3 381 1.28 22.4 4.58 125 2.38 10017 179.2 619.6 2.56 6.85 4.16 50 42.5 6.2918 100 375.8 5.12 25.66 4.30 80 7.24 26.319 116.4 1500 0.64 0.064 —– —– —– —–20 207.9 816.9 5.312 6.46 3.21 50 4.15 14.421 100 612.1 6.4 5.12 —- —– —– —–22 113.5 370.1 4.48 22.02 4.44 80 7.29 24.823 100 370.9 9.6 41.79 4.06 120 2.66 10024 346.4 1500 0.64 0.256 4.34 135 2.59 14825 333.9 1460.5 0.64 0.32 4.34 120 3.59 10026 340.2 1500 0.64 0.192 4.31 50 44.3 5.5427 194.7 847 1.6 0.128 4.13 80 6.03 28.228 195.2 875.7 1.28 0.64 4.74 80 9.52 21.929 100 475.2 8.32 11.14 —– —– —– —–30 100 556.1 6.4 8.064 —– —– —– —–
122
Table 4.2 (cont’d)
Reg T1 (T2) M(> T1) M(> T2) n Tℓ M(> Tℓ)[M(T>120K)]
Mtotal
(K) (K) (106M⊙) (104M⊙) (K) (106M⊙) (%)(1) (2) (3) (4) (5) (6) (7) (8) 9
31 100 654.5 8.32 2.368 —– —– —– —–32 137.2 615.3 2.56 2.944 —– —– —– —–33 147.9 441.6 3.84 8.128 4.64 50 45.2 4.1334 348.2 1500 0.64 0.256 4.51 80 15.1 24.135 176.6 606.8 3.2 9.152 4.26 50 52.5 5.7736 118.2 439.1 11.52 18.50 4.29 50 38.6 5.6137 100 377 7.68 23.17 4.23 80 6.11 2738 100 388.6 5.12 21.12 3.96 80 4.4 30.139 284.7 1500 0.64 0.192 4.67 80 7.17 22.640 100 397.3 7.04 14.4 4.21 50 19.1 6.0241 100 532.7 8.32 3.712 3.11 80 0.54 42.442 100 439.5 8.96 10.88 3.72 50 7.74 9.2243 117.5 355.4 7.68 35.264 4.57 100 5.7 52.144 349 1500 0.64 0.128 4.39 135 2.37 10045 362.7 947.1 0.64 0.896 4.15 135 2.14 10046 173.6 635.6 3.84 4.288 4.5 50 66.3 4.6747 100 394 15.36 19.84 4.24 50 27.7 5.8848 315.8 1254.6 0.64 0.384 4.06 50 21.1 6.8449 163.6 558.5 3.2 3.904 4.94 95 6.27 39.350 198.3 1472.4 1.28 0.448 —– —– —– —–51 —– —– —– —– —– —– —– —–52 141.5 670.6 5.12 4.352 4.03 50 17.9 7.0653 358.6 1500 0.64 0.128 4.55 135 1.98 10054 375.1 1500 0.64 0.32 3.9 50 40.4 7.9155 357.3 1246.5 0.64 0.576 4.43 120 5.88 10056 347.1 1500 0.64 0.256 4.51 100 9.04 52.757 111.5 347.2 9.6 42.432 4.55 100 6.25 52.358 242.7 1035.3 0.64 0.64 4.35 50 42.3 5.3259 147.4 559.5 5.76 5.504 4.78 50 64.5 3.6660 265.5 1469.2 0.64 0.192 4.17 50 25.2 6.24
123
Table 4.2 (cont’d)
Reg T1 (T2) M(> T1) M(> T2) n Tℓ M(> Tℓ)[M(T>120K)]
Mtotal
(K) (K) (106M⊙) (104M⊙) (K) (106M⊙) (%)(1) (2) (3) (4) (5) (6) (7) (8) 9
61 100 664.7 6.4 3.2 —– —– —– —–62 145.7 481 3.84 6.016 4.56 50 34.1 4.4363 151.5 470.3 5.76 11.456 4.98 95 8.78 39.564 147.2 435.7 8.32 23.872 4.55 75 23.9 18.965 100 383.7 23.04 58.56 4.22 50 73.3 5.9966 139.8 408.1 7.04 50.048 4.18 100 8.82 5667 132.6 378.7 8.32 43.584 4.41 80 17.1 25.168 149.3 405.8 5.12 17.536 4.6 80 13.4 23.369 152.7 406.1 3.2 12.928 4.73 85 8.86 27.670 165.8 545.1 3.2 3.136 4.85 85 8.42 26.571 123.3 892.6 4.48 0.64 —– —– —– —–72 179.9 1500 1.28 0.256 4.93 80 6.3 20.373 151.6 525 5.76 3.776 5.15 80 19.3 14.674 130.9 394.7 8.96 21.568 4.56 50 65.3 4.4275 161.8 440.6 5.76 13.632 4.77 100 8.95 50.376 212.2 533.3 2.56 12.032 4.72 120 6.72 10077 277.2 715.3 1.28 6.272 4.37 100 11 54.178 182 638.8 4.48 6.656 4.5 80 19.1 24.279 173.5 606.9 3.84 2.944 4.81 90 10.1 33.580 160.3 545 3.84 4.288 4.67 85 9.24 28.281 204.2 1500 1.92 0.512 4.53 87 7.28 32.482 354.7 188.9 0.64 0.896 4.63 100 4.97 51.683 184.6 863 2.56 1.216 4.61 100 5.01 51.884 —– —– —– —– —– —– —– —–85 187 1031.3 0.64 0.64 —– —– —– —–86 186.1 863.7 1.92 0.832 4.81 68 16.6 11.887 157.8 536.5 5.76 4.736 4.92 80 16.5 20.488 170.8 633.3 4.48 3.712 4.65 80 16.7 22.889 184.3 551.3 4.48 11.392 4.65 110 8.17 72.890 242.8 669.6 1.92 9.216 4.32 110 8.9 74.9
124
Table 4.2 (cont’d)
Reg T1 (T2) M(> T1) M(> T2) n Tℓ M(> Tℓ)[M(T>120K)]
Mtotal
(K) (K) (106M⊙) (104M⊙) (K) (106M⊙) (%)(1) (2) (3) (4) (5) (6) (7) (8) 9
91 183.6 672.6 5.12 9.856 4.24 50 94.5 5.8792 179.4 636.8 5.12 5.184 4.71 95 11.9 41.693 162.5 526.2 5.76 6.272 4.79 85 13.9 27.194 164.7 625.3 5.12 4.352 4.49 50 66.1 4.7295 142.6 458.9 9.6 13.184 4.55 50 74.2 4.4796 160.6 421.9 3.84 16.512 4.6 100 6.67 51.897 —– —– —– —– —– —– —– —–98 —– —– —– —– —– —– —– —–99 205.6 1500 0.64 0.256 4.81 100 2.75 49.9100 167.7 1092.8 3.2 0.512 4.88 85 9.49 26.2101 171.5 593.5 3.84 3.52 4.71 85 13.4 27.9102 189 633.5 4.48 6.016 4.64 100 10.7 51.5103 249.6 684.4 1.92 8.128 4.95 198 2.16 100104 216.1 701 3.84 10.048 4.2 100 11.7 55.8105 218.5 721.5 2.56 5.568 4.4 100 10.3 53.8106 168.4 590.9 7.68 5.696 4.91 80 27.7 20.5107 168.2 673.7 6.4 3.648 4.68 50 111 4108 162.3 590.4 8.32 6.656 4.67 50 123 4.01109 185 618.5 5.76 6.464 4.78 100 13.1 50.2110 121.6 856.1 7.68 0.96 —– —– —– —–111 137.8 897.2 3.2 0.768 —– —– —– —–112 143.6 932.4 2.56 0.704 —– —– —– —–113 166.9 968.1 2.56 0.896 4.19 50 19.6 6.1114 164.2 985.4 3.84 1.024 4.52 50 46 4.6115 182.5 725.8 4.48 3.648 4.57 100 8.57 52.2116 214.3 695.8 3.84 9.728 4.25 100 12.1 55.3117 221.4 687 3.2 11.072 4.2 110 9.04 75.7118 193.6 691.3 5.12 7.808 4.4 100 11.7 53.8119 182.5 675.9 5.76 5.76 4.59 90 16.8 35.6120 175.1 605.5 7.04 9.152 4.59 80 27.5 23.4
125
Table 4.2 (cont’d)
Reg T1 (T2) M(> T1) M(> T2) n Tℓ M(> Tℓ)[M(T>120K)]
Mtotal
(K) (K) (106M⊙) (104M⊙) (K) (106M⊙) (%)(1) (2) (3) (4) (5) (6) (7) (8) 9
121 190.4 653 5.12 7.296 4.59 75 35.6 18.5122 204.8 658.2 3.84 6.336 4.66 105 11.3 61.4123 111.2 841.5 10.88 1.024 —– —– —– —–124 135.3 873.6 3.2 0.896 —– —– —– —–125 180 480 0 7.424 —– —– —– —–126 182.9 1489.8 1.28 0.32 —– —– —– —–127 177.9 814.9 1.92 0.832 4.78 50 43.9 3.66128 189.7 861.6 3.2 1.472 4.61 100 6.31 51.8129 195.2 689.7 4.48 5.696 4.66 107 10.4 65.7130 224.5 683.4 3.2 8.064 4.42 105 10.9 63.4131 185.7 655.2 5.76 7.936 4.49 100 12 52.9132 180.5 642.6 7.04 8.064 4.55 80 28.4 23.7133 199.4 703.7 4.48 6.336 4.45 90 16.6 37134 220.7 824.2 3.2 3.968 4.45 50 120 5.04135 193.5 670.9 5.12 4.16 4.86 105 12.3 59.7136 130.1 863.7 3.84 0.96 —– —– —– —–137 139.5 894.1 3.2 0.832 —– —– —– —–138 178.1 1371.1 1.28 0.384 —– —– —– —–139 130.8 416.5 6.4 11.456 4.63 50 44.1 4.15140 183.5 609.5 3.2 3.648 4.72 95 8.77 41.9141 188.5 605.4 4.48 7.744 4.58 105 9.09 62142 213.4 692.9 3.84 7.936 4.55 100 15 52.4143 185 692.4 7.04 8.192 4.5 81 27.5 25.8144 176 615.9 7.04 5.888 4.81 100 13.4 49.9145 175.5 990.9 7.04 1.664 4.8 90 18.8 33.5146 155.5 832.4 8.32 1.664 4.61 50 85.9 4.25147 100.8 566.6 14.08 4.288 4.61 50 17.7 4.24148 —– —– —– —– —– —– —– —–149 158.2 451.1 —– 8.448 —– —– —– —–150 100 445.6 10.88 10.432 4.05 50 13.4 6.92
126
Table 4.2 (cont’d)
Reg T1 (T2) M(> T1) M(> T2) n Tℓ M(> Tℓ)[M(T>120K)]
Mtotal
(K) (K) (106M⊙) (104M⊙) (K) (106M⊙) (%)(1) (2) (3) (4) (5) (6) (7) (8) 9
151 151.8 429.2 4.48 10.56 4.72 80 11.3 22.2152 180.1 486.5 5.12 12.352 4.68 105 9.26 61.2153 245.9 688 2.56 7.808 4.47 100 15.1 53.1154 192.8 696.5 6.4 8.448 4.49 85 26.3 30.1155 156.7 409.2 9.6 26.176 4.73 85 23.2 27.6156 182.4 873.5 2.56 1.216 4.82 50 115 3.52157 139.8 410.8 8.32 12.288 4.92 50 93.9 3.24158 128.2 1024.3 4.48 0.512 4.85 50 16.1 3.44159 —– —– —– —– —– —– —– —–160 —– —– —– —– —– —– —– —–161 154.3 714.1 2.56 1.664 —– —– —– —–162 174.5 913.4 1.92 1.088 4.86 50 62.8 3.4163 172.7 470.2 4.48 7.104 4.92 100 8.74 48.9164 197.6 660.8 5.12 5.952 4.73 100 14.6 50.6165 215.5 708.9 3.84 6.016 4.57 100 14.8 52.1166 171.6 634.9 7.04 7.168 4.59 71 36.3 14.9167 169.7 706.6 5.12 2.688 4.69 50 91.8 3.95168 148.5 470.2 7.04 8.128 4.81 65 31.8 9.48169 165.2 1125.8 1.28 0.448 4.99 50 41.5 3.04170 105.3 549.1 15.36 4.928 4.79 50 27.9 3.61171 135.1 583.7 5.12 2.368 4.68 50 25.6 3.99172 169.8 877.7 1.92 0.704 4.78 50 33.5 3.66173 158 1500 5.12 0.32 5.21 77 17.1 15.1174 164.8 644.9 6.4 2.176 5.03 85 16.3 24.9175 158.7 579.4 10.24 6.848 4.89 50 176 3.32176 159.9 610.6 10.88 9.28 4.57 50 133 4.39177 162.5 634.9 9.6 6.208 4.68 50 140 4178 164.7 811.9 7.68 2.112 4.6 50 98.9 4.3179 168.4 953.2 6.4 1.024 4.57 50 85.5 4.38180 119.6 563.5 9.6 4.416 4.81 50 33.7 3.57
127
Table 4.2 (cont’d)
Reg T1 (T2) M(> T1) M(> T2) n Tℓ M(> Tℓ)[M(T>120K)]
Mtotal
(K) (K) (106M⊙) (104M⊙) (K) (106M⊙) (%)(1) (2) (3) (4) (5) (6) (7) (8) 9
181 114.2 523.4 12.16 6.272 4.67 50 31.7 4.02182 146.5 596.4 6.4 3.776 4.71 50 56.3 3.9183 152.6 602 8.32 3.84 4.91 54 78 4.52184 144.2 426.5 10.88 10.688 5.05 50 144 2.88185 142.7 663.5 17.28 2.688 5.15 50 183 2.65186 156 649.9 10.88 3.84 4.93 50 163 3.21187 155.7 650.9 14.72 6.144 4.82 50 193 3.52188 150.7 434.5 15.36 20.224 4.91 50 229 3.26189 152.2 699.2 15.36 3.904 4.84 50 171 3.48190 154.5 657.9 10.88 5.12 4.71 50 119 3.89191 100 533.9 12.16 6.848 3.91 50 8.49 7.86192 100 409.4 19.2 16 4.37 50 27.8 5.25193 100 357.5 23.68 37.056 4.47 50 59.2 4.79194 140.1 567.9 14.08 10.944 4.52 50 85.1 4.59195 158.1 681.4 10.24 5.952 4.5 50 99.1 4.65196 148.8 657.5 16.64 5.696 4.79 50 155 3.64197 149.5 571.8 20.48 9.344 5.02 56 174 4.7198 148.1 554.4 25.6 14.656 4.9 50 300 3.28199 155.5 559.8 16 13.44 4.76 50 218 3.72200 157.2 556.3 11.52 10.816 4.72 50 158 3.83201 100 342.8 33.28 60.224 4.56 50 101 4.42202 100 466.1 5.76 8.832 3.44 50 4.03 11.8203 140.5 539 7.04 4.8 4.47 50 38.7 4.78204 139.5 459 11.52 8.384 5 50 115 3.02205 132 403.2 15.36 26.624 4.69 50 113 3.95206 165.8 597.1 7.68 7.552 4.56 50 119 4.45207 225.7 1069.4 3.84 1.408 4.77 50 243 3.7208 147.6 605.2 27.52 10.688 4.83 50 266 3.51209 155.2 634 21.76 10.112 4.76 50 260 3.73210 131.5 386.6 33.28 60.352 4.8 50 273 3.6
128
Table 4.2 (cont’d)
Reg T1 (T2) M(> T1) M(> T2) n Tℓ M(> Tℓ)[M(T>120K)]
Mtotal
(K) (K) (106M⊙) (104M⊙) (K) (106M⊙) (%)(1) (2) (3) (4) (5) (6) (7) (8) 9
211 124 344.1 21.12 69.12 4.85 80 34 21212 100 352.8 28.16 42.88 4.64 50 85.4 4.13
4.4.2 Power law model fit
Considering the diverse range of heating environments in a galaxy’s ISM, the
traditional method of fitting two or three discrete temperature molecule components
to the excitation diagram may not be realistic. Assuming a continuous power law
temperature distribution for H2, we could fit the excitation diagram and calculate the
total H2 gas mass in the ISM (Chapter 2). Theoretical studies have demonstrated
that cooling of H2 molecules in shocks follows a power law temperature distribution
(Hollenbach & Mc. Kee 1979, Burton 1987. Following this in shocked environments
of supernova IC 443 (Neufeld et al. 2008) used a power law temperature distribution
to fit the mid infrared (MIR) H2 rotational line fluxes. Since shocks are the dominant
excitation mechanism for H2 we used our power law model to study molecular gas
properties in SQ (Guillard et al. 2009, Cluver et al. 2010, Guillard et al. 2012).
Assuming MIR H2 emission to be optically thin, the flux observed in a rotational
line S(J) is
FJ =hνJAJNJ+2Ω
4π, (4.3)
where FJ , νJ , AJ , NJ+2, Ω are the flux, frequency, spontaneous emission probability,
column density of the upper energy level, solid angle of the observation, respectively,
for the H2 rotational line S(J). The column density is proportional to the flux F , the
corresponding transition wavelength λ, and is inversely proportional to the sponta-
129
neous emission probability A,
NJ+2 ∝FJλJ
AJ, (4.4)
hence the observed ratio of column densities are,
NJ+2
N3=
Nu
NS(1)
=FJλJA1
F1λ1AJ, (4.5)
where N3 is the column density of H2 molecules at energy level J = 3, the upper
energy level for the S(1) transition.
We assumed the column density of H2 molecules are distributed by a power law
function with respect to temperature, dN ∝ T−n dT, where dN is the number of
molecules in the temperature range T—T+dT. The model consists of three parame-
ters, upper and lower temperature cut off with power law index, denoted by Tu, Tℓ,
and n, respectively. Keeping the upper temperature, Tu, fixed at 2000 K and varying
Tℓ and n we fit the H2 excitation diagram for each square region in the grid of our
mapped area of SQ. Figure 4-4 is a model fit (red solid) to the observed H2 line ratios
(black points) in the excitation diagram of the square region 105, with n = 4.4 in the
temperature range 100–2000 K .
4.5 Result
4.5.1 Two temperature fits
Figure 4-5 shows the distributions of T1 and T2 temperatures across the SQ. The
lowest temperature map shows cooler gas in the northern part of the filament with
a quite broad distributions, and peaks near the center and to the south of the main
filament reaching over 200 K. The situation is similar to the T2 component map,
where there are several peaks in the hottest component, at the ends of the filament.
and in the center. Note that the warmer T2 component at 650 K extends towards
130
Figure 4-4 Power law model fit (red solid) to the observed H2 line ratios(black points) in an excitation diagram for the square regionnumbered- 105 of our mapped area.
131
Figure 4-5 Map of the temperature distribution of the warm molecular gasderived from two-temperature fit to the H2 excitation diagramsat each sampled position. (a) lowest temperature (T1) map im-age, with contours in K, (b) same map superimposed on the HSTF665N image of the Quintet, (c) The higher temperature compo-nent (T2) superimposed on the F665N HST image, and (d) T2map as an image with contours.
the direction of the Seyfert galaxy NGC 7319. The hottest regions in both the T1
and T2 maps is the south of the filament. The coolest T1 component lies near the
extragalactic star formation site called SQ-A by Xu et al. 2005. In this region the
H2 may be heated by a combination of UV radiation and shocks (Cluver et al. 2010,
Appleton et al. 2013). One consequence of the large warm gas in the center of
the shock is that the measured mass in warm molecules is biased towards the lower
temperature regions.
4.5.2 H2 line ratios
The molecular gas in SQ can be substantially heated by shocks. H2 molecules
can be excited to their high rotational energy levels, resulting in significant amount
132
of MIR H2 emission. An analysis of H2 line ratios can provide insights for shock
energetics. In this section we study the variation of line ratios over the shock regions
of SQ.
The H2 MIR rotational line ratios S(0)/S(1), S(3)/S(1), and S(5)/S(1) for each
square region through shocks are color plotted, shown in Figure 4-6. Two facts clearly
emerge from this analysis. First, shock regions have low S(0)/S(1) ratios but high
S(3)/S(1) and S(5)/S(1) and second, the line ratios follow S(0)/S(1) < S(5)/S(1) <
S(3)/S(1). The contour levels of S(3)/S(1) and S(5)/S(1), shown in the Figure 4-7,
map the shock areas with two central regions, where these ratios are elevated. The
two center regions of contour correspond to central shock regions between galaxies,
NGC 7318b–NGC 7319 and NGC 7318b–NGC 7320. The white squares in the ratio
maps correspond to regions where S(0), S(3), and S(5) lines are undetected.
H2 molecules are excited to high energy levels (J > 2) in shocks resulting a decrease
in the number of molecules at energy level J = 2 hence, resulting in a low S(0) flux.
This explanation is consistent with the observed increase in the flux ratios of high
rotational lines, S(3)/S(1) and S(5)/S(1). However, shocks are unable to excite H2
to very high rotational levels (J ≥ 5), resulting in low S(5)/S(1) when compared to
S(3)/S(1). We conclude that shocks in SQ are efficient in exciting H2 molecules till J
≤ 5. The regions with undetected S(3) and S(5) lines (white squares) are less warm
unable to excite molecules to J = 5 and 7, respectively, resulting in low flux/non-
detection of these lines.
4.5.3 Power law index
In the last sub-section, we discussed variation in the H2 rotational line flux ratios
caused by shocks. This variation in line ratios result in a change in the power law
index of our model. An analysis of power law index can provide insights for physical
conditions of molecular gas.
133
Figure 4-6 H2 line ratios, S(0)/S(1), S(3)/S(1) and S(5)/S(1) are estimatedin shock regions of SQ, which are marked by the blue line and thecenter shock region is assumed to be approximately at the regionnumbered-91, shown in Figure–4-2. H2 line ratios, S(3)/S(1) andS(5)/S(1), are elevated in shocks of Stephan’s Quintet. However,S(0)/S(1) is constant over the shock filament of SQ, implyingthat molecules are excited to high energy levels (J > 2).
134
Figure 4-7 A 2D color plot for the H2 line ratios, S(0)/S(1), S(3)/S(1) andS(5)/S(1) in our mapped regions of SQ. Smooth contours areoverlaid on the color plot (left) and the Hα HST image fromWide Field Camera 3 (WFC3) (right). The position of contoursclearly mark the shock regions of SQ.
135
Figure 4-8 A 2D color plot for the power law index in the mapped regions ofSQ. Smooth contours are overlaid on the color plot and the HαHST image from WFC3. The position of contours clearly markthe shock regions of SQ. The shock regions have a high powerlaw index (considering the negative sign), implying the presenceof warm molecular gas.
Figure 4-8 is a 2D color plot for power law indices in the shock region of SQ.
The contour level map the main shock region of the filament very well, creating the
impression of a bow-shock shaped which follows faint Hα features shown on the HST
image. The power law indices estimated from our model are higher, about 4.3–4.4,
at the center of the main shock and lower still to the south (4.2-4.3), considering the
negative sign.
Shocks in SQ excite H2 molecules to high rotational energy levels, resulting in
large flux in high-J rotational lines, which in turn flattens the excitation diagrams. A
large number of H2 molecules are at high temperatures, leading to a less steep, less
negative in warmest part of the shocks. The values of the power law index are almost
all lower than the mean found for normal galaxies (4.84±0.61, chapter 2). Indeed the
only region near SQ-A, where the value of n lies between 4.8< n <5.1.
136
4.5.4 Total and Warm H2 gas mass
4.5.4.1 Total H2 gas mass
The upper energy level of the lowest rotational transition of H2 is at 510 K, higher
than the typical ISM kinetic temperatures. As a result low temperature H2 molecules
in the ISM are too cold to populate high energy levels (J = 2 and higher), and majority
of them remain in the ground state. We extrapolate our model to lower temperatures
to estimate the total molecular gas reservoir. The MIR H2 rotational lines are not
sensitive to molecules less than the temperature Ts, defined in chapter 2, and as a
consequence our model show no change at lower temperatures, T < Ts.
In our analysis three different cases emerge in determining Ts
Case1: As we decrease the lower temperature, Tℓ, the difference between the
observed and the model fit ratios (R) continuously decrease and below a certain
temperature (Ts) the deviation becomes insignificant, minor change in the model fit
occurs. The parameter, R, is defined as,
R =
m∑
i=1
(
fi,mod − fi,obs
σfi,obs
)2
, (4.6)
where f = ln(
Nu/gu
(Nu/gu)S(1)
)
, fi,mod and fi,obs are the modeled and observed flux ratios
for the ith line with uncertainty σfi,obs, and the summation is over all the independent
line flux ratios. The black solid curve in Figure 4-9, demonstrates this case. In such a
scenario we extrapolate our power law model to 50 K to calculate the total molecular
gas mass. The calculated molecular gas mass found by extrapolating our model to 50
K, correlates with the molecular gas mass derived using the Galactic-CO conversion
factor, αCO,Gal
Case2a: As we decrease the lower temperature, Tℓ, the value of R decreases but
below a certain temperature it starts increasing and at low temperatures show no
137
significant change. The red dotted curve in the Figure 4-9, demonstrates this case in
the square region- 76. The value of R decreases untill the lower temperature of about
120 K but increase in the range 50–120 K and remains almost constant below 50 K.
In such case we adopt the corresponding temperature for Tℓ, here 120 K, where the
deviation is found to be minimum. The presence of excess warm molecular gas due
to shocks could be the reason for high Tℓ and the complete molecular gas in these
regions are traced by the MIR H2 rotational lines. Case2b: The blue-dashed curve
is a similar case but the minimum deviation occurs at very high Tℓ. In our analysis
of SQ this was observed in the square region 103, the central shocked region of SQ,
where the model yield the best fit value of Tℓ = 198 K. The molecular gas in this
central shocked region of SQ is heated to a temperature of ∼ 200 K.
Case3a: In some regions the S(0) line is undetected but with detections of high
J lines of H2. The S(1) and higher J lines of H2, correspond to energy levels, J ≥
3. High J levels requires high temperatures for excitation, which result in high Ts.
In such cases we extrapolate our power law model to 80 K to calculate the total
molecular gas mass. Several Ultra Luminous InfraRed Galaxies (ULIRGs), which
are local mergers, with warm dust color temperatures have non detections of S(0)
but good S/N (signal-to-noise) detection for high-J rotational lines (Higdon et al.
2006; Stierwalt et al. 2014). In these galaxies we need to extrapolate our power
law model till 80 K to recover the total molecular gas mass to be consistent with
their dynamical mass estimates (Togi & Smith submitted). Case3b: The final case
is when no rotational H2 lines are detected. Two interpretations can be made, no
molecular gas in these regions or presence of very cold molecular gas in the ground
rotational state. The latter case is highly unlikely in case of SQ, where shocks are
observed leaving with a possibility of no significant amount of molecular gas content
in these regions.
Laying the criteria as discussed, we extrapolated our model to the appropriate
138
Figure 4-9 The difference between the model derived and the observed H2
line ratios (R) as a function of lower temperature to determineTℓ. The black solid, red dotted, and blue dashed curves representdifferent cases. For the black curve cases we adopt the value ofTℓ = 50 K. The red and blue curve occur in warm regions, wherethe temperature of all H2 molecules are high enough to be tracedcompletely by the MIR rotational lines. In such cases we adoptthe value of Tℓ at the temperature where we have the minimumdeviation. For instance for the red and blue curve we adopt Tℓ
= 120 and 198 K, respectively, given by our power law model.
139
Tℓ and estimated the total molecular gas mass for each square region in our mapped
area of SQ. The estimated total molecular gas mass in our mapped regions is about
8.3 × 109 M⊙. This molecular gas mass is consistent with the CO luminosity derived
gas mass of 5 × 109 M⊙ in the bridge and the ridge regions of SQ, using the Galactic
conversion factor αCO,Gal (Guillard et al. 2012). However, the difference in the
mapped areas and the uncertainty in the conversion factor for shock regions could
lead to variation in molecular gas masses.
4.5.4.2 Warm H2 fraction
Knowing the temperature distribution from our power law model and the total
molecular gas mass for each square region, we investigated the warm gas fraction in
shocks of SQ.
The H2 gas greater than 120 K, show their signatures in the MIR H2 rotational
lines. We selected the molecular gas temperature above 120 K to define our warm
gas. We calculated the molecular gas mass fraction above 120 K using,
M(> 150K)
Mtotal=
∫ Tu
150T−ndT
∫ Tu
TℓT−ndT
. (4.7)
The power law index, n, is estimated from the model and the lower temperature is
decided as discussed previously in the last sub-section. Calculating the warm gas
fraction for each square region in our mapped area of SQ we plotted a 2D color plot,
as shown in Figure 4-10.
The warm gas fraction is high in shock regions of SQ. On average the warm gas
fraction is about 30% in shocks. The central shock region of SQ (square region- 103),
the warm gas fraction is higher than 1.0, implying temperature of molecular gas is
higher than 150 K (here we defined our warm gas temperatures higher than 150 K).
The warm gas fraction is about 60% in the southern shock regions, between NGC
140
7318b–NGC 7320. The northern parts of our mapped regions are cold, with warm
gas fraction less than 10%. The molecular gas in the shocks of SQ are warmer when
compared to normal star forming galaxies, where the warm gas fractions is ∼ 10%
(Roussel et al. 2007)
4.6 Summary
1. The discrete two temperatures fits shows high temperatures in the center main
shock filaments and also at the observed southern region.
2. The line ratios in shocks of SQ are observed with S(0)/S(1) < S(5)/S(1) <
S(3)/S(1).
3. The line ratio S(3)/S(1) is higher than S(5)/S(1), implying shocks are able to
excite molecules till J = 5.
4. The power law index is higher, implying high warm molecular gas fractions in
shock regions of SQ.
5. The total molecular gas mass in our mapped regions of SQ is 8.3 × 109 M⊙, in
agreement with CO based estimates.
6. The warm molecular gas mass fraction is ∼ 30% in shock regions. In the
central shock regions the molecular gas temperature is about 200 K. The warm gas
mass fraction in the southern shock regions are found to be as high as ∼ 60%.
141
Figure 4-10 A 2D color plot for the warm gas mass fraction in the mappedregions of SQ. The top right figure show X-ray data from theChandra Observatory in blue superposed on optical data in yel-low (APOD, Credit: X-ray: G. Trinchieri et al.). The X-raysare produced by gas heated to ≈ 106 K by shocks on cosmicscales. Smooth contours are overlaid on the color plot and theHα HST image from WFC3. The position of contours clearlymark the shock regions of SQ, where the warm gas fraction ishigher. The warm gas fraction is about 30% in shocks, with anincreased value of about 60% in the southern regions. In thecentral shocked regions the molecular gas is heated to temper-atures of about 200 K.
142
Chapter 5
Physical properties of the
cometary globule: the case of B207
5.1 Introduction
Stars form inside molecular clouds, which have sizes ranging from a few tens of
parsec down to small, sub-parsec cores. Isolated Bok globules (Bok & Reilly, 1947)
are the simplest starless and star-forming molecular clouds (Clemens & Barvainis,
1988, Yun & Clemens 1990, Yun & Clemens 1992, Launhardt et al. 2010). Nearby
Bok globules are some of the best places to study in detail the physical and chemical
processes occurring during star formation (Clemens & Barvainis, 1988, Bourke et al.
1995, Launhardt et al. 1997, Henning & Launhardt 1998, Stutz et al. 2008, Ward et
al. 2007). Bok globules, being small in size and forming only one or at most a few
stars, offer an opportunity to study star forming processes in clouds unaffected by
the presence of other stars. Also, the large number of Bok globules within a distance
of 500 pc provides excellent spatial resolution to study star formation processes in
molecular clouds.
Sometimes more than one protostars at different evolutionary stages are seen embed-
ded in a single Bok globule (Launhardt et al. 2010). This led Launhardt et al. 2010
143
to suggest that multiple star formation in isolated globules is non-coeval, whereby a
Class I protostar, a much younger Class 0 protostar, and a gravitationally unstable
sub-mm core may all be co-existing in the same isolated, low-mass molecular cloud
without requiring a causal interrelationship among them. Studying such globules will
help understand time evolution and conditions required for the successive star forma-
tion process in molecular clouds.
In this paper, we study the cometary globule Barnard-207 (B207), located at a dis-
tance of 140 pc in the Taurus - Auriga molecular cloud region (Loinard et al. 2005,
Kenyon et al. 1994). B207 (LDN 1489) is a high Galactic latitude Bok globule, lo-
cated at equatorial coordinates (J2000.0) RA = 61.216 Dec = 26.318 (ℓ = 168.08,
b = -19.15). A 1.6 M⊙ protostellar source, IRAS 04016+2610, is currently forming
at the West side of B207’s core (Yen et al. 2014). This protostar is located at a
projected distance of about 10,000 AU from the center of the present-day dense core
of B207. Benson et al. 1998, using high spatial and spectral line observations of
N2H+ and C3H2, show an asymmetric profile for B207 (LDN 1489). They conclude
that the line shapes for B207 are consistent with models of cloud cores undergoing
gravitational collapse. Therefore, the globule B207 may be in a second star formation
phase. Hence, studying physical properties of B207 will help in our understanding of
the core collapse phenomenon and the star forming process occurring in such objects.
The goal of this paper is to perform an in-depth analysis of B207 and to determine
the physical properties and evolutionary state of LDN 1489.
5.2 Observations & Archival Data
The objective of the observations is to measure both the extinction of the back-
ground stars through different parts of B207 as well as the surface brightness of the
scattered light, and to study the general properties of the cloud in different optical
144
and infrared bands.
5.2.1 Discovery Channel Telescope (DCT) observations
B207 was imaged in several optical bands, using the Large Monolithic Imager
(LMI; Massey et al. (2013)) of the 4.3-metre Discovery Channel Telescope (DCT;
Degroff et al. (2014)), operated by Lowell Observatory, on 2013 February 5-6 and
2013 December 5. LMI exposures were taken using its e2v CCD231, having 6144 ×
6160 pixels, with 2 × 2 binning. Imaging was performed with the Johnson UBV and
Kron-Cousins R and I filters, at effective wavelengths 366, 428, 537, 633, and 806
nm, respectively (these calculated wavelengths are the arithmetic means of the two
wavelengths at which the transmission is half of the maximum1 in each filter). The
LMI field of view was 12.3×12.3 arcmin2, large enough to cover most of the dense
parts of the globule. The plate scale was 0.24 arcsec/pixel for 2 × 2 binning.
The globule was observed with total exposure times of 180, 240, 105, and 75
minutes in U, B, V, and R bands, respectively, consisting of sub-exposures of 15 -
20 minutes. In the I band, two exposures were taken, each of 10 minutes duration.
The observable characteristics of the globule are represented in 5-1. It is a 3-colour
stacked image in which blue, green and red colours represent the V, R, and I bands,
respectively.
The basic data reductions such as flat fielding, bias correction, and sky subtrac-
tion were performed using the Image Reduction and Analysis Facility (IRAF). The
effects of cosmic rays in the images were removed using the L.A. Cosmic package (van
Dokkum. 2001) within IRAF.
The photometric calibration was derived using a standard star, GD71, from the
list of Landolt A. U. (1992), located close to the globule’s location. Using the LMI-
1http : //www.lowell.edu/techSpecs/LMI/specs.html
145
Figure 5-1 A 3-colour stacked image of B207. The blue, green, and redcolours represent the V, R, and I bands, respectively. Theprotostar, IRAS 04016+2610, at R.A. = 4h4m43.071s , Dec =+261856.39 is forming at a distance of about 10000 AU west ofthe core, marked by the red arrow. A sharp boundary is seen to-wards the outer rim of the cloud to the East. A less dense regionof material, appearing as a hole is located towards the immediateWest of the protostar.
146
DCT manual, the photometric scale was fixed and the zero points were determined.
The magnitudes of other stars in the field were determined and found to be consistent,
within 10% uncertainty in the observed bands, when compared to entries in SIMBAD.
5.2.2 Archival WISE and Herschel data
We employed archival Wide-field Infrared Survey Explorer (WISE; Wright et al.
(2010)) W1 and W2 image data at 3.4 and 4.6 µm to perform star counts and to
investigate the presence of coreshine in the inner regions of the B207 globule. The
WISE band images were 10×10 arcmin2 wide. The pixel scale was 1.375 arcsec/pixel
in the W1 and W2 bands. We used the WISE point source catalog (PSC) to identify
stars in the image.
To determine the spectral energy distribution (SED) of the core and the rim of
the globule, we used archival Herschel Space Observatory’s Photoconducting Array
Camera and Spectrometer (PACS; Poglitsch et al. (2010)) - 160 µm image with
Spectral and Photometric Imaging Receiver (SPIRE; Griffin et al. (2010)) - 250,
350, and 500 µm maps that were about 1 × 0.9 wide, observed in the Herschel-
Guaranteed-Time Key Program (KPGT) program (PI Ph. Andre). The full width
half maximum (FWHM) for PACS 160, and SPIRE 250, 350, and 500 µm point
spread functions (PSFs) are 12.8, 17.6, 23.9, and 35.2 arcsec, respectively.
5.3 Analysis
The mass, temperature, density distribution, along with other dust and gas prop-
erties provide hints to the past, present and future evolutionary fate of a globule. In
this section, we provide a description of our methods for determining the physical
properties of B207.
147
B207 shows many interesting features, as seen in Figure 5-1. First, a protostellar
object (IRAS 04016+2610) is seen near the centre of the image, at a projected offset
of ∼ 10,000 AU from the dark core of B207. Second, a nearly transparent region
of gas and dust is seen ∼ 30, 000 AU towards the West of the newly forming star.
Third, a sharp edge is seen towards the eastern side of the cloud. Fourth, a small
dark core in B207 is seen as a dark region to the East of IRAS 04016+2610. The
bright rim surrounding the core is typical for high-latitude globules and is interpreted
as resulting from scattering (see section 3.2.1) of the galactic radiation field in the
optically thin outer portion of the globule (Fitzgerald et al. (1976)).
5.3.1 Surface brightness profiles
In contrast to many globules at lower Galactic latitudes, B207 is readily discernible
through its surface brightness at optical wavelengths (Figure 5-1). In its optically thin
outer portions of the globule this surface brightness is proportional to the line of sight
optical depth. Mapping the optical surface brightness of globules provides, therefore,
an excellent means of tracing the optical depth distribution in the outer parts of
a globule like B207. With additional information on the line-of-sight interstellar
radiation field (ISRF) and the intensity of the diffuse galactic light (DGL) in the sky
adjacent to the globule, it is also possible to estimate the globule’s dust albedo at
different wavelengths.
We obtained surface-brightness profiles of the globule B207 from the U, B, V,
R, and I band images. A horizontal East-West (E-W) cut, shown in Figure 5-2,
was made to construct the surface brightness profile. Individual surface brightness
values were averaged over a box of size 5×20 arcsec2, perpendicular to the E-W cut,
oriented along the north-south (N-S) direction, yielding sky subtracted intensities in
erg/cm2/s/A/sr. Table 5.1 lists the value of the conversion factor from count/s/pixel
to erg/cm2/s/A/sr in U, B, V, R, and I bands, calibrated by measuring the flux from
148
Figure 5-2 The position of East-West horizontal cut region across B207 ona R band image for surface brightness analysis.
a Hubble standard star GD71.
The surface brightness profiles for the five LMI bands are shown in Figure 5-3. At
the eastern rim, the surface brightness rises sharply, reaching a peak almost exactly
at the same spatial location for all the observed wavelengths. Given that the peak
surface brightness of any given wavelength occurs at the same (wavelength-dependent)
optical depth, this finding demonstrates that the line-of-sight optical depth increases
sharply near the portion of the brightness peaks. This suggests that the material is
compressed towards the Eastern side of the globule. At the core position, the surface
brightness increases with increasing wavelength. The intensity of the core region is
lower than the background sky in U and almost equal to the sky in B. The region
149
Table 5.1. Photometric calibration for 1 count/s/pixel
Band Effective wavelength (λ) IntensityName (nm) (10−7erg/s/cm2/A/sr)(1) (2) (3)
U 366 9.678B 428 1.625V 537 1.006R 633 0.683I 806 0.555
ATo calculate surface brightness, we used the conversion of 1 pixel2 =1.354×10−12 steradian
BThe wavelength listed for each band is the arithmetic mean of the twowavelengths at which transmission is half of the maximum. The transmis-sion curves for each band can be found atwww2.lowell.edu/rsch/LMI/specs.html
west of the protostar, appearing as a hole in the globule, displays an intensity value
greater than the clear background sky in B, V, and R bands, implying the presence
of some diffuse dust. The flux of the protostar increases from undetectable at U to
very bright at I, consistent with its SED peaking at ∼ 50 µm (Furlan et al. 2008).
The surface brightness distribution of B207 is characterized by a bright outer
rim surrounding a dark core. This pattern is the consequence of a strongly forward-
throwing phase function of the scattering dust, which controls the transfer of the
illuminating ISRF through the object (Witt et al. 1974, Witt et al. 1990). High-
latitude globules like B207 are detectable by their scattered-light surface brightness,
because the diffuse Galactic light (DGL) arising in the surrounding diffuse interstellar
medium has a lower surface brightness due to its much lower line-of-sight optical
depth, thus providing the contrast necessary for visibility. At low Galactic latitudes,
the surface brightness of the DGL and of the bright outer rim of globules are more
nearly the same, thus erasing this contrast. Those globules are then recognized simply
as dark core seen against a brighter background.
Given the extreme asymmetry of the scattering phase function, the globule’s sur-
150
Figure 5-3 The sky subtracted surface brightness (in ergs/cm2/s/A/sr) forB207 in U, B, V, R, and I bands of the DCT-LMI. The East -West direction is marked in the U-band plot starting from East(left) to West (right). The position of the eastern bright rim,core, protostar IRAS 04016+2610, and the transparent hole re-gion are marked as R, C, S, and D, respectively. The East-Westdistance is measured from the protostar, where the zero positioncorresponds to R.A. = 4h4m43.5s , Dec = +2618′50′′
151
face brightness is determined almost exclusively by the intensity of the ISRF seen by
the far side of the globule from directions within a few degrees around the line-of-sight
direction from the observer to the globule. The surface brightness at the side facing
the observer is determined by the transfer of scattered radiation through different
sections of the globule. The actual intensity is proportional to the product of two
probabilities: the probability of scattering, measured by a(1 − e−τ ), and the proba-
bility that once-scattered radiation escapes through the front surface of the globule,
measured by e−τabs . Here, a is the dust albedo, τ is the line-of-sight extinction optical
depth through the globule, and τabs is the typical absorption optical depth for the
average once-scattered photon. A suitable approximation is τabs ≈ 0.5(1− a)τ (Witt
et al. 2008). The bright outer rim of an externally illuminated globule occurs where
the product of these two probabilities reaches a maximum. For the high dust albedos
typical for globules (Witt et al. 1990), this maximum occurs at line-of-sight optical
depths near τ = 1.5, occurring at about ∼-270 arcsec from the protostar in B207, as
seen in Figure 5-3.
5.3.2 Dust characteristics
5.3.2.1 Scattered light intensities
The maximum surface brightness of B207 observed in the different LMI bands
may be used to constrain the albedo of the dust present in this globule. Our tool is
the relation between the ratio of maximum surface brightness to the intensity of the
illuminating radiation field and the dust albedo, as found from Monte Carlo radiative
transfer simulations by Witt et al. (1974). In order to employ this tool, we need
to correct the observed maximum surface brightnesses in the five LMI bands for the
background DGL intensity still present in the intensity of the sky adjacent to B207,
and we also need the value of the interstellar radiation field (ISRF) at the location
152
of B207. It is this particular intensity that is crucial for determining the surface
brightness of B207 because of the strongly forward-throwing phase function of the
dust.
The sky intensity adjacent to B207 consists of foreground components stemming
mainly from terrestrial airglow, scattering in the Earth’s atmosphere, and zodiacal
light, and a background component consisting of DGL, arising from scattering of
the ISRF by diffuse Galactic dust along the line of sight. The fact that the surface
brightness core of B207 in U has a substantially lower surface brightness than that
of the adjacent sky, suggests that most of the adjacent sky DGL arises at distances
larger than that of B207. This is also supported by the recent work, which places
most of the dust in the line of sight of B207 at a distance of about 250 pc and beyond
(Green et al. 2015). To find the absolute surface brightness of B207, we need to
add the intensity of the adjacent DGL to the intensities displayed in Figure 3. The
foreground components, contributing equally to both the sky and globule intensities,
have been eliminated by the original sky subtraction and are of no further concern.
We estimate the DGL intensities in the sky adjacent to B207 by first finding the
optical depth of the dust in the sky region. The all-sky map of Schlegel et al. (1998)
yields a value of E(B - V) ∼ 0.197, and with the re-normalization by Schlafly et
al. (2010) and Schlafly et al. (2011) yields a more probable value of E(B - V) ∼
0.169. With a value of RV = 3.1 for average diffuse Galactic dust, we find AV = 0.53
mag. Scaling this value with the extinction cross sections per H nucleon (Cext/H) as
a function of wavelength from Draine (2003)2 for ISM dust (RV = 3.1), we find the
optical depth values listed in column 3 of Table 2.
Noting that these optical depths are within the optically thin regime, we use the
radiative transfer approach based on single scattering (Equation 4, Witt et al. (2008))
to estimate the corresponding DGL intensities. For this, we use the ISRF intensities
2http : //www.astro.princeton.edu/ draine/dust/dustmix.html
153
Table 5.2. Dust Characteristics
Filter ISRF τ a τabs DGL SBmax(DGL+SBmax)
ISRFac
(1) (2) (3) (4) (5) (6) (7) (8) (9)
U 4.77 0.721 0.625 0.135 1.338 0.259 0.335 0.62±0.05B 5.55 0.624 0.648 0.110 1.497 0.839 0.421 0.70±0.05V 6.41 0.485 0.671 0.080 1.528 1.811 0.521 0.75±0.05R 6.80 0.402 0.676 0.065 1.425 2.361 0.557 0.78±0.03I 6.37 0.259 0.656 0.045 0.911 3.272 0.657 0.87±0.03
1The value of ISRF, DGL, and SBmax are in units of 10−8erg/s/cm2/A/sr. The ISRF fordifferent wavelengths are from Porter & Strong (2005)
2τ and a are the extinction optical depth and albedo of the dust in the sky adjacent to B207assuming RV = 3.1. The albedo values are from Draine (2003)
3τabs is the absorption optical depth of cloud evaluated using τabs = 0.5(1 − a)τ
4DGL is the diffuse galactic light adjacent to B207, derived using aIISRF (1 − e−τ )e−τabs
4SBmax is the peak intensity of the globule after subtracting the nearby background sky intensity
5The (DGL+SBmax)/ISRF is the ratio of the total maximum intensity scattered by the globuleto the interstellar radiation field
6The albedo of the cloud, ac, evaluated for the corresponding ratio (DGL+SBmax)/ISRF as-suming the forward scattering asymmetry parameter, 0.6 ≤ g ≤ 0.9 from Figure 6, Witt et al.(1974)
of Porter & Strong (2005) at the globule’s position, listed in column 2 of Table 2, and
the dust albedos from Draine (2003) for RV = 3.1 Milky Way dust, listed in column
4 of Table 2. The resulting DGL intensities are listed in column 6 of Table 2.
In order to arrive at the background-corrected absolute surface brightness val-
ues for the maximum surface brightness, we add this DGL component to the sky-
subtracted maximum surface brightness listed in column 7 of Table 2. The ratio of
the corrected absolute maximum surface brightness to the ISRF listed in column 8 is
the quantity to be compared directly with the model predictions in Figure 6 of Witt
et al. (1974). The resulting estimates for the dust albedo, ac, in B207 are shown in
the last column of Table 2 with their estimated uncertainties. The albedo values were
calculated under the assumption that the phase function asymmetry, g, is close to
0.9, consistent with the results found in other dark nebulae (Gordon (2004)). Henyey
154
& Greenstein (1941) introduced the phase function asymmetry parameter, g, in the
range −1 ≤ g ≤ +1, from back scattering to isotropic scattering (g = 0) to forward
scattering. A change in the assumption of phase function asymmetry to g ≈ 0.7
causes a small reduction in the estimated albedo values Gordon (2004).
Two facts stand out from this analysis. First, the globule albedo values are gen-
erally larger than those of the dust in the adjacent diffuse dust background. Second,
the albedo is increasing with wavelength, reaching its highest value of about 0.87
in the I band. These characteristics are typical for size distributions of grains that
include micron-size grains, which are significantly larger than the largest grains in
diffuse ISM dust, e.g. Kim & Martin (1996). The dust albedo values are comparable
with those found in other studies of globules and dark nebulae. Fitzgerald et al.
(1976) evaluated the dust albedo in B-band, a = 0.70 ± 0.08, for the Thumbprint
nebula. The Coalsack and Libra cloud have a dust albedo of ∼ 0.6 for g ∼ 0.8 in U,
B, and V bands (Mattila (1970)). However Witt et al. (1990) found the albedo value
decreasing from 0.80 at 469 nm to 0.58 at 856 nm for the Bok globule CB4, which
has a much lower central optical depth compared to B207.
5.3.2.2 Coreshine effect in NIR band
Coreshine is an effect in which starlight in the 3–5 µm wavelength range is effi-
ciently scattered by larger than normal ISM grains, e.g. µm size grains versus the
normal size range 0.01–0.25 µm. The dense cores of molecular clouds frequently ex-
hibit this behaviour (Pagani et al. 2010, Steinacker et al. 2010). The dense cores
are sufficiently optically thick to prevent UV photons from penetrating and exciting
polycyclic aromatic hydrocarbon molecules (PAHs), thus ruling out the possibility of
3.3 µm PAH emission contributing to the 3–5 µm coreshine. Pagani et al. (2010)
investigated a sample of 110 cores from which 95 cores were detectable in the Spitzer-
IRAC band at 3.6 µm. Numerous cores, about 50, clearly had signatures of coreshine.
155
Figure 5-4 A 10×10 arcmin2 WISE W1 image of B207 at 3.4 µm. The pro-tostar IRAS 04016+2610 is at the center of the image. The bluecross sign is the core position. B207 is exhibiting the coreshineeffect by showing a clear excess of 3.4 µm emission compared tothe background. The coreshine effect is due to the scattering bylarge grains inside the globule
The emission at 3.6 µm is spatially coincident with the densest regions of the cores.
The coreshine phenomenon can be explained by near-infrared (NIR) radiation
from the ISRF being scattered by micron-size grains. The dense cores of the globules
are cold and thus provide favorable physical environments for the grains to acquire
ice mantles, coagulate, and grow in size to exhibit the coreshine effect (Hirashita &
Li (2013), Andersen et al. (2014)). The intensity of the coreshine depends on the
incident radiation, the extinction of the background radiation, the grain properties,
and the core properties (Steinacker et al. 2014b).
We investigated the presence of coreshine in the core of B207 using the 3.4 µm im-
age of WISE. The coreshine effect as a surface brightness excess in the core compared
156
to the background sky is observed in the 3.4 µm WISE W1 image, shown in Figure
5-4. The observed core intensity excess above sky of B207 over an aperture of 10
arcsec radius was found to be 0.05 MJy/sr (conversion from DN to Jy units adopted
from WISE Preliminary Data Release). The surface brightness value of the B207
core at 3.4 µm is comparable to the values obtained in the Taurus-Perseus sample
of Lefevre et al. (2014). Theoretical modeling by Steinacker et al. (2014b) yields a
similar value of ∼ 0.06 MJy/sr for the surface brightness at 3.6 µm.
The core of B207 was found to be opaque, completely blocking the background
starlight even at such a long NIR wavelength. The observed coreshine effect further
supports our idea of the presence of large grains inside B207, as already inferred from
the dust albedo measurements.
5.3.3 Temperature Analysis using SED fitting
The gas temperature and turbulence are major parameters in determining condi-
tions for the core collapse, providing pressure support against the gravitational infall
of the material and hence, deciding the evolutionary fate of the globule (Dickman
& Clemens (1983), Nelson & Langer (1999)). The difference between core and rim
temperatures can help us estimate temperature gradients of gas and dust in the glob-
ule. The rim temperature depends on ambient ISM conditions in which the globule
is located.
We used the PACS 160 µm, SPIRE 250, 350, and 500 µm dust emission maps
from Herschel Space Observatory to model the line-of-sight average dust temperature
and column density toward the core and rim of B207. We convolved 160, 250, and
350 µm maps to the equivalent resolution of the 500 µm map using a low-resolution
kernel (Aniano et al. 2011). The core was defined as a circular region of 40 arcsec
diameter (> 35.2 arcsec, the PSF of SPIRE instrument at 500 µm) at the position
RA = 61.2023, Dec = 26.314. Table 5.3 lists the background subtracted intensities
157
Table 5.3 Intensity in MJy/sr at FIR and sub-mm wavelengths
Wavelength (µm) Core Rim160 99.5±10.0 16.4±4.0250 155.9±12.5 24.7±5.0350 148.2±12.2 17.2±4.1500 84.4±9.2 7.1±2.7
measured in the four Herschel (PACS 160 µm, SPIRE 250, 350, 500 µm) bands.
Spectral energy distributions (SEDs) were constructed and single temperature
modified black-body curves were used to fit the SEDs. We used a single dust tem-
perature black body model Sν = ΩB(ν, Td)(1 − e−τ(ν)), where Ω, B(ν, Td), and τ(ν)
are solid angle of the emitting element, Planck function at dust temperature Td, op-
tical depth at frequency ν, respectively. We used a dust opacity model, κν ∝ νβ
with β = 1.8 (Planck Collaboration 2011) to fit the SED using the MPFIT program
(Markwardt (2009)). Figure 5-5 shows single dust temperature SED model fits at the
core and rim.
The derived core and rim temperatures are 11.7±0.1 K and 13.3±1.4 K, respec-
tively. Our derived core temperature is in agreement with the 11.7 K value predicted
for B207 (LDN 1489) by Ford & Shirley (2011), who used a radiative transfer model
with the ISRF as the heating source. The core and rim temperatures are consis-
tent with literature studies for other similar Bok globules (CB244-Stutz et al. 2010,
DC000.4-19.5 Hardegree-Ullman et al. 2013, Launhardt et al. 2013, CB17- Schmalzl
et al. 2014). The presence of the nearby protostar, IRAS 04016+2610, could raise
the core temperature. However, Launhardt et al. (2013) showed that the presence of
protostars in a globule affected the temperature only to a distance of about 5000 AU.
In our case, IRAS 04016+2610 is at a projected distance of ∼10000 AU from the core
of B207 and, therefore, is not expected to have significant effects on the core temper-
ature. The rim temperature of the globule is affected by the external heating from
the ISRF and the outside physical environment in which the globule is embedded.
158
Figure 5-5 A modified single temperature blackbody spectrum for the dustemission at the core and rim assuming β = 1.8 to fit the ob-served intensity values. The model yields a temperature value of11.7±0.1 K and 13.3±1.4 K for core and rim, respectively.
159
5.3.4 Mass Distribution
Mass is one of the key physical parameters that decides a globule’s fate. The
evolutionary future of the core, whether it forms a star or remains starless, depends
on the likelihood of core collapse, which is directly linked to the core mass. Moreover,
the type of star formed inside the core depends on the dense core mass, with a star
formation efficiency of ǫdensecore = Mstar/Mdensecore ≈ 30% (Motte et al. 1998, Alves
et al. 2007, Andre et al. 2014). Different methods can be employed to estimate the
core mass. We derive the core mass of B207 by three different methods: star counts,
reddening of background stars, and sub-mm 500 µm emission.
5.3.4.1 Star count technique
The star count method, based on the number density of background stars (stars/arcmin2)
seen in each part of the cloud, is used to obtain the extinction, an estimate of the
column density (Dickman (1978)). The number density of stars is compared to that
on the adjacent, clear, background sky region. As the core is approached, the increase
in the column density of dust leads to higher extinction, causing the fainter stars to
disappear, below some detectability threshold. Assuming an appropriate gas to dust
ratio (GDR), the total column density of gas can be estimated. Summing over the
entire globule area allows us to calculate the globule’s mass.
We used a rectilinear grid of squares, each 2×2 arcmin2 on a 3.4 µm WISE W1
map of B207. In each square, we counted the number of stars. The average num-
ber of stars in the clear background sky towards the east of B207 was found to be
1.75 stars/arcmin2. Groszschedl (2012)3 found the slope for log(Nm) versus limiting
magnitude λ for WISE W1 in an Orion A control field of 36.7 degree2 to be ∼0.33,
where Nm is the number of stars brighter than magnitude λ per square degree in the
3http : //othes.univie.ac.at/20156/1/2012− 04 − 16 0408170.pdf
160
sky. The galactic latitude of the control field is identical to that of B207. Comparing
the number of stars in each square region on the globule with the clear background
sky, and using the slope of 0.33 from the plot of log(Nm) versus λ we calculate the
extinction at 3.4 µm (A3.4). Knowing the extinction value in each square region we
calculate the mass of the globule B207, using the relation
M = (γd)2µNH
A3.4
∑
i
Ai3.4, (5.1)
where
N(H)
A3.4=
AV
A3.4× N(H)
AV(5.2)
N(H)
A3.4= 17.63 × 1.87 × 1021 = 3.3 × 1022cm−2mag−1. (5.3)
Here, γ is the angular size in radians of each square in the grid, d = 140 pc is the
distance to the cloud, µ = 1.36 the mean molecular weight corrected for helium and
heavy element abundance, and i is the counting index of the square on the extinction
map (Cambresy (1999)). The value of N(H)AV
= 1.87×1021 cm−2 mag−1 is from Bohlin
et al. (1978). For the 3.4 µm map this value needs to be scaled accordingly. From,
Draine (2003) we calculated the ratio of AV /A3.4 = 17.63 for an extinction curve with
RV = 5.5, applicable for molecular clouds.
We find the molecular gas column density at the core to be N(H2) = (4.2 ±
1.0) × 1022 cm−2, while the corresponding mass out to 160 arcsec is (12.1± 3.0)M⊙,
calculated from the star count method (our core mass includes He and other heavy
element mass correction and 160 is the core radius estimated from the Bonnor-Ebert
modeling, Sec 3.5). The star count method implicitly requires that all stars counted
in an obscured region lie behind the cloud. This fact, along with the Poisson counting
errors in the small number of stars in each square grid and the error in the distance
to B207, contribute to the total error in the mass calculation of the core (Dickman
161
Figure 5-6 The variation of column density as a function of the distance fromthe core. The black points are the column density measurementsfrom the WISE band colour excess (W1 - W2) of backgroundstars in the field of view (FOV). The blue curve is a fit to the ex-tinction profile obtained from the reddening of background stars.The red diamonds are from the FIR intensity photometry fromthe SPIRE-500 µm map. The molecular column density of thecore is of the order 3.6×1022 cm−2, causing an extinction of AV
∼50 mag for RV = 5.5.
(1978) Appendix).
5.3.4.2 Reddening of background stars
The change in colour of background stars seen through the globule can be used
to estimate the column density of material. Assuming an appropriate value of the
intrinsic colour of background stars and the GDR, we can estimate the globule’s mass.
We selected a region of 300 arcsec radius centered on the core of B207 to estimate
the column density profile. We determined the W1-W2 colour of background stars.
162
Galaxies in the region were removed, since they are intrinsically redder compared to
stars (Wright et al. (2010)). We chose the WISE band intrinsic colour W1 - W2
= 0.05 mag for unreddened background stars (Wright et al. 2010, Kirkpatrick et
al. 2011). Assuming the value of RV = 5.5, the relation between molecular column
density and the colour excess in W1 - W2 from Draine (2003) is
N(H2)
E(W1 − W2)= 3.73 × 1022cm−2mag−1 (5.4)
We chose RV = 5.5 because the sizes of dust particles in the globule are greater
compared to normal ISM dust. The morphological structure of our cloud is not
spherically symmetric, as the western part of the globule is more diffuse than the
compressed eastern part. Hence, there is a considerable scatter seen in the extinction
values at a fixed radial distance from the core. We averaged the extinction values
estimated from stars in 9 radial bins of 30 arcsec width and fit the extinction profile.
An exponential fit to the column density profile was marginally better than a power
law fit, over the range 0–300 arcsec from the core center. Our exponential fit to the
column density is 6.69×1022 exp(-0.015r), where r is the radial offset from the core
in arcsec, as shown in Figure 6. The mass, measured out to 160 arcsec radius is
12.8 ± 0.5M⊙.
5.3.4.3 Far infrared (FIR) emission
We used the Herschel-SPIRE 500 µm map to estimate the column density profile,
as an independent check. We calculated B500(Td) (Planck function) for the core dust
temperature Td = 11.7 K (see Section 3.3). The core mass, Mcore, can be obtained
from
Mcore =I500 × D2
κ500 × B500× MH
Md× 1.36, (5.5)
163
where I500 and D are the intensity at 500µm and globule’s distance, respectively.
MH
Md= 110 is the hydrogen gas-to-dust mass ratio in the solar neighborhood (Sodroski
et al. 1997) and κ500 is the dust opacity at 500 µm. We used the value for κ500 = 2.9
cm2g−1 from Ossenkopf & Henning (1994) for the MRN dust distribution with thin
ice mantles.
We calculated the core mass of 11.7±2.9 M⊙ using 500 µm emission. A large error
in this mass estimation is from MH/Mdust, which is uncertain by a factor of about 10–
35% (Sodroski et al. (1997)). In Figure 5-6, the red diamonds are the column density
values derived from 500 µm intensities. The molecular hydrogen column density,
N(H2), can be obtained from
N(H2) =1
2× N(H) =
1
2× MH
Mdust× τ500
κ500 × mH(5.6)
The masses 12.1±3.0 M⊙, 12.8±0.5 M⊙, 11.7±2.9 M⊙, estimated out to the 160
arcsec radius, derived by the three different methods, star count technique, reddening
of stars, and FIR 500 µm emission, respectively, appear to be in good agreement
with each other. However, assuming a constant temperature and dust opacity over
the entire core region can introduce a minor error in our mass estimate values. The
molecular column density at the core centre derived from the 500 µm intensity map
is N(H2) = 3.6 × 1022cm−2, causing an extinction of AV ∼ 50 mag for RV = 5.5 (
Draine 2003). The high column density categorizes the globule B207 as an Infrared
Dark Cloud (IRDC) (Egan et al. 1998).
5.3.5 Temperature and density-Abel Inversion method
Recently, the radial variation in dust temperature and density has been studied
in a few starless cores. Roy et al. (2014), using the inverse Abel transform technique
on FIR Herschel maps of B68 and L1689B, concluded that the actual core dust tem-
164
perature is about 2 K lower compared to the temperature obtained by the line of
sight averaged SED fitting. Evaluating the core temperature using a SED fit gives
an average temperature along the line of sight and does not take into account the
temperature gradients within the sources. We adopted the inverse Abel measurement
technique to study the variation in temperature along the radial direction towards
the core of B207. We used Herschel-SPIRE FIR maps at 250 and 350 µm for the
analysis.
We considered a spherically symmetric globule with radial density profile, ρ(r),
surrounded with a uniform background and isotropic ISRF. With an assumption of
optically thin dust emission, the specific intensity, Iν(p), of the globule observed at
impact parameter p, is
Iν(p) = 2
∫ ∞
p
ρ(r)B(ν, Td(r))κνrdr
√
r2 − p2+ Iν,bg + Iν,N , (5.7)
where Iν,bg and Iν,N are the background and instrumental noise, respectively, B(ν, Td(r))
is the Planck function at the dust temperature value Td(r) for a given radius r from
the core centre, and κν is the frequency dependent dust opacity (Roy et al. 2014).
Using the inverse Abel transform (e.g. Bracewell 1986) we obtain the integrand
of the above equation at observed frequency ν,
ρ(r)B(ν, Td(r))κν =−1
π
∫ ∞
r
dIν
dp
dp√
p2 − r2. (5.8)
Evaluating the above equation at two different frequencies and using their ratio with a
predefined assumption about the dust opacity law κν ∝ νβ with β = 1.8, we calculated
the temperature value at distance, r, from the core centre.
We used convolved 250 and 350 µm Herschel-SPIRE intensity values to evaluate
the temperature profile of B207. The cloud was divided into 12 equal angular sector
regions along with equal radial steps for each angular sector region. A radial temper-
165
Table 5.4 B207 core/rim temperature
Region TSED(K) TAbel(K)Core 11.7±0.1 9.4±0.1Rim 13.4±1.4 13.3±1.0
ature profile for each angular sector was computed. Figure 5-7a represents the radial
temperature profile of B207, averaged over all 12 angular sector regions. The error in
the temperature values are computed using the standard deviation from 12 angular
sector regions. However, because of the presence of the protostar, IRAS04016+2610,
and the transparent hole region, angular sector regions towards the western side of
the core and at a distance greater than 10,000 AU were not considered. The protostar
can heat gas and dust in its vicinity causing deviations in the temperature profile of
the globule.
The temperature profile, when extrapolated to the core, reaches a minimum of
about 9.4±0.1 K. The temperature increases to 13.3±1.0 K in the outer regions of the
cloud at a distance of about ∼ 40000 AU, where it is heated by the ISRF. The core
temperature derived by the Abel inversion technique is 2.3 K less than that found
from the SED fitting technique. Table 5.4 lists the core and rim temperatures for
B207 derived from the SED fit and Abel inversion method.
Knowing the radial temperature distribution in B207 and equation (5.8), we calcu-
lated the radial density profile from the 250 µm map. We used the dust mass opacity
at 250 µm, κ250 = 10.6 cm2 g−1, from Ossenkopf & Henning (1994) for the MRN dust
distribution with thin ice mantles. Assuming a hydrogen gas-to-dust mass ratio of
110 in the analysis, we adopted a gas mass opacity at 250 µm, κ250,gas = κ250/110
= 0.1 cm2 g−1. The density distribution was obtained again for each angular sector
region. Figure 5-7b represents the density profile for B207 averaged over all angular
sector regions. The density distribution flattens out towards the core center and is
significantly different in slope from that in the outer regions of the cloud.
166
Figure 5-7 a. Dust temperature profile of B207 obtained by applying theAbel inversion method to circularly averaged intensity profilesobserved with Herschel at 250µm and 350µm. The error rep-resent the standard deviation of the mean in Td(r) values ob-tained from independent profile reconstruction along 12 angulardirections. b. Density profile of B207 obtained by applying theAbel inversion method to circularly averaged intensity profilesobserved with Herschel at 250 µm.
167
We analyzed the density profile using the Bonnor-Ebert (BE) model by solving
the modified Lane-Emden equation for an isothermal sphere (Bonnor 1956, Ebert
1955. Any given solution is characterized by three parameters: the temperature T ,
the central density ρc, and the outer radius of the cloud Rmax. Once an outer radius
is specified, the model is in an unstable equilibrium if ξmax > 6.5, where
ξmax =R√
4πGµmHρc
c2s
. (5.9)
cs, µ, and mH are the sound speed, mean atomic mass per particle in molecular
clouds, and mass of hydrogen atom, respectively.
We generated a set of BE models over a 3D grid in temperature, outer radius and
central density. Our temperature grid ranged from T = 4 to 40 K in steps of 2 K.
Our radial grid varied from θmax = 80 to 240 in steps of 5 arcsec. For the central
density grid, we varied the ξmax from 3.0 to 12.8 in steps of 0.2, since varying ξmax
is equivalent to varying the central density, nc (or ρc). We evaluated our density
profiles and calculated the best fit model by finding the minimum χ2 over the grid in
temperature, θmax, and ξmax
χ2 =1
N − 3
N∑
i=1
(ni − nBE,i)2
σ2i
, (5.10)
where ni, σi, and nBE,i are the density profile, observed uncertainties, and BE model
density profiles, respectively.
We found the best-fit BE model with parameters T = 24±1.2 K, θmax = 160±24
arcsec and ξmax = 7.0 ± 0.9, which corresponds to central density of nc = 3.4 × 105
cm−3. The value of ξmax is marginally greater than 6.5, indicating that the core
may be gravitationally unstable and subject to collapse. The calculated BE mass
parameter, MBE , is 2.2 M⊙, while the actual core mass calculated with the density
distribution, shown in Figure 7b, out to radius of 160 arcsec is 9.3 M⊙.
168
5.3.6 Core energetics
The observed evidence for a collapsing core for B207 is presented by Benson et
al. (1998) in their survey of dense cores in dark clouds. The spectra of C3H2 (21,2
- 10,1) and the F1,F = 0,1→1,2 component of the N2H+(1 - 0) transition in B207
show asymmetric profiles, consistent with the models for core collapse. Knowing the
number density and temperature distribution of the core we computed energy values
due to different processes occurring inside the core to determine whether the core is
stable or collapsing.
We calculated the gravitational potential energy, Egrav, from the density distribu-
tion
Egrav = −1
2× 4πGmHµH
∫ r1
0
M(≤ r)nH(r)rdr. (5.11)
Finding the density distribution profile, and assuming r1 = 160 arcsec = 0.108 pc =
22,400 AU, and adopting µH = 1.36, we calculated Egrav = -1.78 × 1036J. The factor
1/2 is to compensate for the double counting of point masses.
Next, to estimate the thermal energy we determined the temperature as a function
of radial distance from the core, shown in Figure 5-7,
T (r) = 0.012r + 9.42, (5.12)
where r is in arcsec. Goldsmith (2001) showed that at densities nH > 104.5 cm−3
dust and gas molecules are closely coupled, leading to identical temperatures, despite
significant depletion of coolant molecules like CO. We assumed the dust and gas
temperature in the core of B207 to be identical in our analysis since the densities are
greater than 105 cm−3. From Schmalzl et al. (2014), we find
Ethermal =3kB
24π
∫ r1
0
T (r)np(r)r2dr, (5.13)
169
where kB is the Boltzmann constant, and np=nHµH/µp is the volume density of
particles with µp = 2.32 being the mean atomic mass per particle in molecular clouds
(Przybilla et al. 2008). The calculated value of Ethermal = 1.03× 1036J.
Caselli et al. (2002), determined the ratio of rotational kinetic energy to the
gravitational potential energy for the core of B207 to be about 0.74× 10−3. Adopting
this value, the rotational K.E. of the core is Erot,K.E. = 0.001× 1036J.
The magnetic energy for B207 was not measured. No literature value exists for
the magnetic field strengths for this particular core. Thus we attempted to estimate
a magnetic field value for B207. Caselli et al. (2002) adopted a typical magnetic field
strength of 10 µG for their sample of globules based on the OH Zeeman measurements
in L1544 of Crutcher & Troland (2000). On the other hand, in a sample of 3 globules,
Wolf et al. (2003) found a magnetic strength of about 200 µG on average. In addition,
polarization measurements indicate an absence of a well ordered magnetic field in
B207 (D. Clemens; priv. commn).
We assumed a constant magnetic field strength throughout the core, to calculate
the magnetic energy
Emag =4πr3
1
3umag, (5.14)
where umag = B2/2µ0 is the magnetic energy density. We calculated Emag = (0.06–
6.0)× 1036 J for the value range of B = 10–100 µG.
We calculated the turbulence energy in the core of B207. The observed linewidth
from Benson et al. (1998), using C3H2, provides a value of 370 m/s. The turbulent
velocity dispersion, σ, is related to the line width ∆v by
σ2 = ∆v2 − ln(2)8kT
m, (5.15)
where m is the mass of C3H2 molecule Myers (1983). The calculated turbulent velocity
dispersion is σ = 354 m/s. The turbulence energy can then be calculated using the
170
relation
Eturb =3Mcoreσ
2
16 × ln2(5.16)
The calculated value of Eturb = 0.62× 1036J
The equivalent energy due to the external ISM pressure estimated from the BE
model at the outer radius, assuming T = 24 K and density of 1.91×104 cm−3 is
Eext =4
3πR3p0 = 1.03 × 1036J (5.17)
(Equation 41.39 Draine (2011)).
With these different energy values we find the net sum energy for B207,
Etot = Egrav + Ethermal + Erot,K.E. + Emag + Eturb + Eext (5.18)
Etot = (−1.78 + 1.03 + 0.001 + 0.06 + 0.62 − 1.03) × 1036J (5.19)
Etot = −1.72 × 1036J (5.20)
However, note the analysis is done assuming a low magnetic field strength of 10 µG.
A high magnetic field of 100 µG would lead to a positive value for Etot.
Figure 5-8 is a plot of energy ratio profiles for B207 as a function of distance from
the core centre. We do not consider the Erot since it is less than 1% of the thermal
energy and has negligible effect on the total energy. Neglecting the core rotational
energy, the criterion for a core to be bound is
Etherm + Eturb + Emag
|Egrav| + |Eext|≤ 1. (5.21)
The energy ratio in equation 5.21 reaches unity at a core distance of about 20,000
AU and decreases further out, neglecting the external ISM pressure, implying a grav-
itationally bound core and suggesting core collapse likely occurring in B207. When
171
Figure 5-8 Energy ratio profiles for B207. The total energy is dominated bythe negative gravitational energy demonstrating a gravitationallybound core in B207. The energy ratio reaches unity at a coredistance of ∼20,000 AU however, when the effects of externalISM pressure is included the ratio reaches unity at a much smallerdistance of ∼10,000 AU.
172
included the effects of external ISM pressure, the energy ratio reaches unity at a
distance of about 10,000 AU from the core centre.
5.4 Discussion
5.4.1 Surface brightness in core and rim
The surface brightness profiles in U, B, V, R, and I bands display a bright rim
surrounding a dark core with increasing core surface brightness at longer wavelengths.
The darkness of the core is due to the inefficiency of back scattering with a highly
forward directed scattering phase function. The increase in dust albedo, in addition
to an increase in the ISRF intensity at longer wavelengths, causes the core intensity
to increase in redder bands. Given the forward directed phase function and the high
photon escape probability in the optically thin rim, the surface brightness is greatly
enhanced. Hence, the rim is brighter than the core.
The adjacent sky is darker mainly because of lower optical depths, declining with
increasing wavelengths. In addition, the albedo of the dust grains in the nearby sky
region to B207 are lower compared to the dust in the globule, and this difference
increases with increasing wavelengths (Table 2). As a result, the contrast between
the rim and the nearby sky region increases with wavelength, causing the rim intensity
to increase significantly above the sky background, while at the shortest wavelength
in the U band the intensities of the rim and the sky become almost equal, making
the rim nearly undetectable.
5.4.2 Grain growth and age determination
The dust albedo ac of the cloud listed in Table 5.2 increases with wavelengths
from about 0.62 in the U band to around 0.87 in the I band. The increase of the
173
albedo with increasing wavelength suggests an increase in the grain size inside the
globule compared to the the dust in the diffuse ISM (Kim & Martin 1996, Draine
2003a), where the dust albedo is ∼0.65, as listed in column 4 of Table 5.2 for RV =
3.1. This suggests that the cloud possesses conditions favourable for grain growth.
The observed phenomenon of coreshine in the WISE W1 band also indicates the
presence of large grains in the core of B207. With a core density of 105 cm−3 the time
required for grains to coagulate and grow to sub-micron size is about 105–106 years
(Hirashita & Li 2013, Ysard et al. 2013, Steinacker et al. 2014a). The presence of
Class I protostar, IRAS 04016+2610, also indicates an age of the globule in the range
few×105 years (Yen et al. 2014).
5.4.3 Core physical properties
The temperatures at the projected core position and the rim of B207 are 11.7±0.1
and 13.4±1.4 K, respectively, from our SED fitting technique. However, the Abel
inversion technique yields a lower value of 9.4±0.1 K at the actual core of B207. The
number density at the core position is nH = 3.4×105 cm−3, decreasing to 400 cm−3 at
the outer regions of the cloud (40000 AU). The Bonnor-Ebert model fit to the density
profile yields a temperature of 24±1.2 K and ξmax = 7.0±0.9. The value of ξmax,
marginally greater than 6.5, suggests the core may be gravitationally unstable. The
calculated Bonnor-Ebert mass, MBE , is 2.2 M⊙, which is lower than 9.3 M⊙ estimated
using the density profile. The calculated mass being greater than MBE also suggest
a gravitationally unstable, collapsing core.
Lippok et al. (2013) showed an increase in CO depletion with the number density
in their sample of 6 cores. The depletions in their core sample range from 48% to 98%
and at an average density of nH = (1.1 ± 0.4) × 105cm−3, 50% of CO is frozen onto
dust grains. The dust temperatures in the Lippok et al. (2013) core sample range
from 9.8–12.9 K. The physical conditions at the core position of B207 with dust
174
temperature 9.4 K and nH = 3.4×105 cm−3 are thus favourable for CO molecules to
freeze on to the dust grains. Ford & Shirley (2011) modeled the B207’s core with
a canonical abundance of CO and found a significant depletion by a factor of 1000
compared to the ISM value. This indicates freezing of CO and other major molecules
like H2O on dust grains, increasing the size of the dust grains.
5.4.4 Core collapse
We determined the core mass using several approaches, a) star count technique,
b) reddening of background stars, c) 500 µm FIR emission, and d) Abel inversion
technique. All methods yielded consistent estimates of the core mass ∼12 M⊙. Caselli
et al. (2002), measuring the velocity dispersion in B207, deduced the virial mass of
the core (major and minor axes length of 1.6 and 1.1 arcmin, respectively) to be
1.47±0.06 M⊙. Our derived core mass is substantially higher than the virial mass,
indicating an unstable condition and likelihood of collapse.
The total net energy, Etot, is found to be negative in the core of B207 (Equation
5.20) and Etherm+Eturb+Emag
|Egrav|is less than unity at the scale of the core size radius,
assuming a weak magnetic field of 10 µG. This implies the core is gravitationally
bound and can be a collapsing core (Ward et al. 2007, Francesco et al. 2007, Andre
et al. 1993). Our analysis of a collapsing core is consistent with the spectroscopic
result of Benson et al. (1998). The ratio of Egrav to Eth is about 2 in a ∼12 M⊙
core of B207 with a central density of ∼ 3.4 × 105 cm−3, making the core unstable
and subject to collapse in the context of Bok globule sample studies of Lippok et al.
(2013) and Launhardt et al. (2013).
Apart from the IRAS 04016+2610 no other protostellar source is detected in B207
at FIR wavelengths as high as 500 µm. The temperatures determined from the SED fit
and the Abel inversion technique do not show evidence of central warming, indicating
an absence of a warm pre-stellar core. This leads us to conclude that the collapse of
175
B207 is in its very initial phase to form a low mass star in the future.
5.5 Conclusions
We present optical observations of B207 made with the Large Monolithic Imager
(LMI) instrument attached to the Discovery Channel Telescope (DCT) in U, B, V, R,
and I bands, together with MIR and FIR data from WISE and Herschel Observatory.
We have studied the physical conditions in the globule B207. The main conclusions
of our study are:
1. The dust albedo in the cloud increases with wavelength from 0.62 in the U-band
to 0.87 in the I-band, indicating grain sizes substantially in excess of those found in
the ISM.
2. The core-shine phenomenon observed in the 3.4 µm WISE W1 band suggests the
presence of micron-size dust particles and the globule’s age of ∼ 106 years, required
for the dust grains to coagulate and grow.
3. The radial density profile can be approximated with a BE model with central den-
sity 3.4 × 105 cm−3 and ξmax = 7.0±0.9, suggesting a gravitationally unstable core.
4. The line-of-sight averaged core and the rim temperatures are about 11.7±0.1 and
13.4±1.4 K, respectively, derived with the SED fitting technique. However, the Abel
inversion technique yields a lower temperature of 9.4±0.1 K at the position of the
core center.
5. The averaged molecular column density (from the three different methods: star
count technique, reddening of background stars, FIR 500µm emission), N(H2), at the
core is about (5.2 ± 1)×1022 cm−2, causing an extinction of about AV = 70 mag,
assuming RV = 5.5. This yields a core mass of ∼ 12 M⊙, out to radius 160 arcsec
with a central number density of nH = 3.4 × 105 cm−3.
6. The energy values relating different processes governing stability indicate a col-
176
lapsing core, which is well supported by observations from line transitions of N2H+
and C3H2 by Benson et al. (1998).
7. The simultaneous presence of a collapsing core and a Class I protostar at a pro-
jected distance of ∼10000 AU indicates non-coeval star formation in B207, while the
absence of a Class 0 source in the core of the globule, suggests that the collapse is in
its initial phase.
177
Chapter 6
Summary and Future Work
Molecular gas is an important constituent of the ISM and a primary fuel for star
formation. To understand star formation and hence galaxy evolution it is essential
to understand physical properties of molecular gas in galaxies. Despite H2 being the
most abundant molecule in the universe, due to lack of a dipole moment and high
rotational energy levels it has never been directly used to study the bulk molecular
ISM properties in galaxies. From almost half a century CO, with an abundance of
10−4 in number relative to H2, has been the primary tracer to estimate molecular
gas content in nearby and high redshift galaxies. This thesis work is based on
developing and discussing a model, using the rotational lines of H2, a direct
tracer, to estimate the total molecular gas content in galaxies
6.1 What did we learn?
6.1.1 H2 power law model
The weak quadrupole pure rotational lines, detected with the ISO and IRS-Spitzer
are presently employed to study warm molecular gas component in galaxies. H2
excitation diagrams are modeled using two or three discrete temperature components
however, this method is not robust and does not yield unique temperature fits to
178
the molecular gas emission. Physically H2 molecules in the ISM are in a continuous
temperature distribution. A need for the continuous distribution of H2 molecules
with temperature, which can also model the H2 excitation diagrams was required.
We employed a temperature power law distribution for H2 and the model parameters
were estimated using the rotational line fluxes obtained from the IRS-Spitzer. Our
model was able to recover the H2 excitation diagram with robust and repeatable
model parameters, unlike discrete temperature models. Moreover, extrapolating our
model to lower temperatures, we recover the total molecular gas mass in galaxies as
traced by CO and other indirect tracers. This resulted in a new detection method of
ISM molecular gas content in galaxies using the warm MIR H2 rotational lines.
We extended our method to low metallicity dwarfs, where CO fails to recover the
true molecular gas mass. Our method, which is independent of any indirect tracers
is able to recover the total molecular gas mass in low metallicity galaxies, where CO
is 50–100× under luminous and CO estimated ISM molecular gas mass are about an
order or two magnitude lower when used a Galactic conversion factor. We employed
the power law method in several low metallicity dwarfs (as low as about 10% of the
Milky Way abundance). The results yielded by the power law method was in good
agreement with other indirect tracers like dust and star formation rate.
6.1.2 Molecular gas properties in GOALS galaxies
We estimate the model parameters from the emission of MIR H2 lines and with
an extrapolation to lower temperature at 49 K, the molecular gas mass estimated is
comparable to the CO derived molecular masses, but subjected to the internal biases
of the CO-conversion factor and the error involved in the flux of H2 lines. Applying
the model in local mergers of GOALS, where turbulence and shocks result in physical
properties of the gas, which differ from normal star forming galaxies and may test
179
our model for any alterations.
Using our model with an extrapolation to 49 K temperature yielded molecular
gas masses consistently higher than the CO derived values. The model estimated
molecular gas mass is also higher when compared to standard gas-to-dust ratio. Two
possibilities arise from this result, either the dust derived molecular gas mass is lower
or the model estimated molecular gas masses are higher. A solution to this problem is
to increase the extrapolation temperature, higher than 49 K, which may subsequently
reduce the ISM gas mass, resulting in standard dust-to-gas ratio and also a higher
warm molecular mass fraction.
6.1.3 Stephan’s Quintet work
Due to collisions occurring in the galaxies of Stephan’s Quintet the molecular
gas is heated to high temperatures. Shocks excite the H2 molecules and the flux of
rotational lines of H2 are higher than normal star forming galaxies. We used the
power law model to understand the physical properties of molecular ISM/IGM gas
content. Since most of the H2 gas is warm and hot, the power law model was able to
estimate the molecular gas content in these shocks without any further extrapolation.
When compared to the single dish CO observations the molecular gas content was
in good agreement with our power law estimates. We estimated the warm gas mass
fraction in Stephan’s Quintet to be about,MH2
(>150K)
Mtotal= 0.3, higher than normal star
forming galaxies. However, in the central shock regions the temperature of warm gas
was estimated to be about 200 K and in the southern regions the warm gas mass
fraction was estimated to be as high as 0.6.
180
6.1.4 Study of Bok globule, B207
Stars are not only formed in core dense regions of GMCs, but also in small isolated
dense core regions of molecular clouds, known as Bok globules. We study the physical
properties of the Bok globule B207 located in the Taurus Auriga molecular region at
a distance of 140 pc, which harbor a Class I protostar, IRAS04016+2610. The dust
properties along with the temperature, volume and column number density profile
is estimated using optical to sub-mm wavelength observations. The different energy
profiles governing the stability of the dark dense core region and the Bonner Ebert
model fit show a gravitationally unstable and collapsing core indicating that B207 is
undergoing a second star formation phase.
6.2 Future work
6.2.1 Study of molecular gas at high z
The James Webb Space Telescope, successor to the Hubble and Spitzer space
telescopes, is planned for launch in late 2018. The Mid Infrared Instrument (MIRI)
onboard JWST has both a camera and a spectrograph covering the wavelength range
from 5 to 28 microns. The detectors of the instrument are sensitive enough to detect
the S(1), S(2), S(3) and higher rotational lines of H2 of nearby and low redshift objects
till z = 0.6, 1.3, 1.9, and higher respectively. The wavelength dependent sensitivity
limit of the MIRI instrument is shown in Figure 6-1. The sensitivities and spectral
resolution are about 100× and 5× better than the previous IRS-Spitzer instrument.
Assuming an averageLH2S(1)
LIR= 10−4 (Bonato et al. 2015), and LIR = 3×1011 L⊙
(typical for LIRGs), the luminosity of S(1) line is 1.2×1034 W, corresponding to flux
at z = 0.5 (dL = 2.9 Gpc) about 1.2×10−19 Wm−2. To observe this line and a signal
to noise ratio S/N = 5, will require 30 minutes of integration time, hence MIRI-
181
JWST is capable of detecting rotational lines of H2 at low and intermediate redshifts.
Using these H2 line fluxes in our power law model, we can characterize the physical
conditions of molecular gas and its role in the early universe.
The Cryogenic Aperture Large Infrared Submillimeter Telescope Observatory (CAL-
ISTO) planned for 2020 decade is a 5m space based observatory cooled to T∼4 K
emphasizing moderate resolution spectroscopy in the wavelength range 35–600 µm
band. The 5σ 1 hour line sensitivity of the telescope is less than 10−20 W m−2 (Brad-
ford et al. 2015). At such high line sensitivity, it is possible to measure pure H2
rotational lines at high redshifts of z ≈ 6–7, almost reaching the reionization era of
universe. Figure 6-2 shows spectroscopic sensitivities in the FIR and sub-mm wave-
lengths of different telescopes. The left panel shows the spectrum of a galaxy, with
luminosity L = 1012 L⊙ and the flux dilution occurring at different redshifts. With the
sensitivity of CALISTO, the MIR flux of 1012 L⊙ galaxy at z = 6.0 can be observed.
At this redshift the age of the universe is about 900 Myr, 7% of the universe’s present
age. Shocks produced in the early universe due to rapid cold gas accretion and AGN
feedback may enhance H2 line emission in many systems, when dust and metals are
in their first cycle of enrichment. Using our power law model with the H2 line flux
we can measure molecular gas properties at high redshifts.
The SPace Infrared telescope for Cosmology and Astrophysics (SPICA) is a joint
mission between European Space Agency (ESA) and Japan Aerospace Exploration
Agency (JAXA), optimized for MIR and FIR spectroscopy with a cryogenically cooled
telescope, targeted to launch in 2027-2028. SPICA will be equipped with two primary
instruments: the Spica FAR infrared Instrument (SAFARI; Roelfsema et al. 2012)
and the SPICA Mid-infrared Instrument (SMI). The limiting flux of point sources for
5σ-1hr is few×10−19 Wm−2 for SMI and SAFARI instruments. Different molecular,
atomic, ionized, PAH line/band intensities for star forming galaxies with 1012 L⊙ can
be observed with the SPICA telescope even at high redshifts. The S(1) line intensity
182
Figure 6-1 Sensitivity of MIRI spectrograph onboard JWST as a function ofwavelength, The sensitivity is about 10−20 Wm−2 at 10m, whichis about two orders of magnitude better and with a spectral reso-lution of about 5 times higher than the IRS-Spitzer spectrograph.
of H2 can be observed till redshift, z = 1, for galaxy with 1012 L⊙.
The above mentioned future projects for the next decade provides an opportunity
to observe H2 rotational lines at high redshifts. Our power law model can be useful
tool in estimating molecular gas mass and study its variation and conditions at dif-
ferent redshifts.
6.2.2 Study on Bok globules
The existence of the cometary shaped morphology of Bok globule has been sur-
prising to astronomers for many decades. Bok globules are sites of low mass star
formation. Sometimes more than one protostars at different evolutionary stages are
seen embedded in a single Bok globule. A Class I protostar, a much younger Class
0 protostar, and a gravitationally unstable sub-mm core all co-exist in the same iso-
183
Figure 6-2 Spectroscopic sensitivities in the FIR and sub-mm. In the leftplot the sensitivity is in Wm−2 for a single pointed observa-tion. Galaxy spectra assuming L = 1012 L⊙ at various redshiftsare overplotted using light curves, with continuum smoothed toR=500. The magenta dashed line is the sensitivity of a quantum-limited heterodyne receiver in a bandwidth of 10 km/s The rightpanel shows he speed for a blind spatial spectral survey reachinga depth of 10−19 Wm−2 over a square degree, including the num-ber of spatial beams and instantaneous bandwidth. (adaptedfrom Bradford et al. 2015)
184
lated, low-mass molecular cloud (Bok globule) (Launhardt et al. 2010). The presence
of Class I protostar, IRAS 04016+2610, along with a collapsing core qualify B207 as a
perfect case for studying sequential star formation. Many Bok globules are observed
with an embedded protostar along with a dark core in the globule sample studied
by (Launhardt et al. 2010), proving B207 is not an unique object. Figure 6-3 is a
histogram showing the projected distance between the dense core and the protostar
in a sample of 38 Bok globules. Further, these observations impose many queries on
Bok globules:
1) What is the distance between the dark core and the embedded protostar?
2) How many Bok globules show collapse features?
3) What are the physical conditions required to have core collapse?
4) Do protostars have any effects on core collapse?
5) What are the age differences between newly forming stars in a Bok globule?
With the new advanced facility of Atacama Large Millimeter/submillimeter Array
(ALMA), we can now look into the deep dense core region of Bok globules. The ab-
sence of any proto stellar source in the collapsing core of B207 at submm wavelengths
observed with Herschel Space Observatory definitely suggests B207 as a potential tar-
get for ALMA. We hope, in the coming decade, new ALMA observations will uncover
the mystery of sequential star formation in Bok globules.
We have yet to fully grasp the physical processes that control atomic to molecular
gas conversions. The fraction of molecular gas content in low density diffuse envi-
ronments and cold dense gravitationally bound molecular clouds in a galaxy is still
unknown. Future research on molecular gas fraction variation in diffuse environments
with galaxy types can provide clue in study of star formation and hence, galaxy evo-
lution. On a much small spatial scale, about a parsec size, low mass star formation
185
Figure 6-3 Distribution of globules according to the projected distance be-tween the core and the protostar in a sample of 38 globules. Afuture study on a much bigger sample with the available FIR Her-schel data will be helpful in determining this distribution moreexact.
186
occurring in isolated Bok globules may help us understand effects of star formation
processes on molecular clouds. In future along-with polarization maps, sensitive and
good resolution IR observations can reveal the mystery of cometary shape structures
of Bok globules. New avenues of research are opening with present observations of
ALMA and also in the coming era of JWST. With other large telescopes in plan def-
initely promise a new age of astronomy and have potential prospect to revolutionize
many areas of astrophysics.
+——————————————————————————————————+
187
References
Abdo, A. A., Ackermann, M., Ajello, M., et al. 2010, ApJ, 710, 133
Alatalo, K., et al. 2011, ApJ, 735, 88
Allen, R. J., & Hartsuiker, J. W. 1972, Nature, 239, 324
Alves, J., Lombardi, M. & Lada, C. J. 2007, A&A, 462, L17
Anathpindika, S. 2015, arxiv:1501.01401
Andre, P. et al. 2014, Protostars and Planets VI, 27
Andersen, M., Thi, W.-F., et al. 2014, A&A, 568, L3
Andre, P., Ward-Thompson, D. & Barsony, M. 1993, ApJ, 406, 122
Aniano, G. et al. 2011, PASP, 123, 1218
Appleton, P. N. et al. 2006, ApJ, 639, L51
Appleton, P. N., Xu, K. C., Reach, W., et al. 2006, ApJ, 639, L51
Appleton, P. N., Guillard, P., Boulanger, F., et al. 2013, ApJ, 777, 66
Armus, L. et al. 2006, ApJ, 640, 204
Armus, L., Charmandaris, V., Bernard-Salas, J., et al. 2007, ApJ, 656, 148
Avni, Y. 1976, ApJ, 210, 642
Bahcall, N. A., Harris, D. E., & Rood, H. J. 1984, ApJL, 284, L29
Barnard, E. E., Chicago: University of Chicago Press, 1927
188
Basu, S. & Mouschovias, T. C. 1994,432, 720
Basu, S. 2000, ApJL, 540, 103
Benson, P. J., Caselli, P. & Myers, P. C. 1998, ApJ, 506, 743
Bhat, C. L., Issa, M. R., Houston, B. P., Mayer, C. J., & Wolfendale, A. W. 1985,Nature, 314, 511
Bigiel, F., Leroy, A., Walter, F., et al. 2008, AJ, 136, 2846
Bitsakis, T., Charmandaris, V., Appleton, P. N., et al. 2014, A&A, 565, A25
Black, J. H., & Dalgarno, A. 1976, ApJ, 203, 132
Bloemen, J. B. G. M., Caraveo, P. A., Hermsen, W., et al. 1984, A& A, 139, 37
Bloemen, J. B. G. M., Strong, A. W., Mayer-Hasselwander, H. A., et al. 1986, A&A, 154, 25
Bohlin, R. C., Savage, B. D. & Drake, J. F. 1978, ApJ, 224, 132
Bok, B. J. & Reilly, E. F. 1947, ApJ, 105, 255
Bolatto, A. D., Wolfire, M., & Leroy, A. K. 2013, ARAA, 51, 207
Bonato, M., Negrello, M., Cai, Z.-Y., et al. 2015, MNRAS, 452, 356
Bonnell, I. A. & Clarke, C. J. 1999, MNRAS, 309, 461
Bonnell, I. A. & Bate, M. R. 2006, MNRAS, 370, 488
Bonnor, W. B. 1956, MNRAS, 116, 351
Bourke, T. L. et al. 1995, MNRAS, 276, 1067
Bracewell, R. N. 1986, McGraw-Hill Series in Electrical Engineering, Networks andSystems, New York: McGraw-Hill, 2nd rev.ed.
Bradford, C. M., Goldsmith, P. F., Bolatto, A., et al. 2015, arXiv:1505.05551
Brandl, B. R., Bernard-Salas, J., Spoon, H. W. W., et al. 2006, ApJ, 653, 1129
189
Browning, M. K., Tumlinson, J., & Shull, J. M. 2003, ApJ, 582, 810
Bryant, P. M., & Scoville, N. Z. 1999, AJ, 117, 2632
Burton, M. G. 1987, Ph.D. Thesis
Burton, M. G., Hollenbach, D. J., & Tielens, A.G.G.M. 1992, ApJ, 399, 563
Cambresy, L. 1999, A&A, 345, 965
Caselli, P., Benson, P. J., Myers, P. C. & Tafalla, M. 2002, ApJ, 572, 238
Clemens, D. P., Sarcia, D., et al. 2007, PASP, 119, 1385
Clemens, D. P., Pavel, M. D., Cashman, L. R., 2012, ApJS, 200, 21
Clemens, D. P. & Barvainis, R. 1988, ApJS, 68, 257
Cluver, M. E., Appleton, P. N., Boulanger, F., et al. 2010, ApJ, 710, 248
Cormier, D., Madden, S. C., Lebouteiller, V., et al. 2014, A& A, 564, A121
Crutcher, R. M. & Troland, T. H. 2000, ApJL, 537, L139
Crutcher, R. M. 2012, ARA&A, 50, 29
Curran, S. J., Aalto, S., & Booth, R. S. 2000, A& As, 141, 193
Dabrowski, I. 1984, Canadian Journal of Physics, 62, 1639
da Cunha, E., Charmandaris, V., Dıaz-Santos, T., et al. 2010, A&A, 523, A78
DeGroff, W. T. et al. 2014, Society of Photo-Optical Instrumentation Engineers(SPIE) Conference Series, 9145, 2
de Rijcke, S., Buyle, P., Cannon, J., et al. 2006, A& A, 454, L111
Di Francesco, J. et al. Protostars and Planets V, 2007, 17
Dickman, R. L. 1978, AJ, 83, 363
Dickman, R. L. & Clemens, D. P. 1983, ApJ, 271, 143
190
van Dishoeck, E. F., & Black, J. H. 1986, Interstellar Processes: Abstracts of Con-tributed Papers, 149
Downes, D., Solomon, P. M., & Radford, S. J. E. 1993, ApJL, 414, L13
Downes, D., & Solomon, P. M. 1998, ApJ, 507, 615
Draine, B. T., & McKee, C. F. 1993, ARAA, 31, 373
Draine, B. T. 2003, ARA&A, 41, 241
Draine, B. T. 2003, ApJ, 598, 1017
Draine, B. T., Dale, D. A., Bendo, G., et al. 2007, ApJ 663, 866
Draine, B. T. 2011, Physics of the Interstellar and Intergalactic Medium by BruceT. Draine. Princeton University Press, 2011. ISBN: 978-0-691-12214-4,
Dufour, R. J., Shields, G. A., & Talbot, R. J., Jr. 1982, ApJ, 252, 461
Dufour, R. J. 1984, Structure and Evolution of the Magellanic Clouds, 108, 353
Dunne, L., Eales, S., Edmunds, M., et al. 2000, MNRAS, 315, 115
Ebert, R. 1955, ZAP, 37, 217
Egami, E.; Rieke, G.H., Fadda,D. & Hines, D. C. 2006, ApJ, 652, L21
Egan, M. P. et al. 1998, ApJL, 494, L199
Evans, A. S., Sanders, D. B., Surace, J. A., & Mazzarella, J. M. 1999, ApJ, 511, 730
Evans, A. S., Frayer, D. T., Surace, J. A., & Sanders, D. B. 2001, AJ, 121, 1893
Evans, A. S., Mazzarella, J. M., Surace, J. A., & Sanders, D. B. 2002, ApJ, 580, 749
Evans, A. S., Mazzarella, J. M., Surace, J. A., et al. 2005, ApJS, 159, 197
Fedotov, K., Gallagher, S. C., Konstantopoulos, I. S., et al. 2011, AJ, 142, 42
Field, G. B., Somerville, W. B., & Dressler, K. 1966, ARAA, 4, 207
191
Fisher, D. B., Bolatto, A. D., Herrera-Camus, R., et al. 2014, Nature, 505, 186
Fitzgerald, M. P., Stephens, T. C. & Witt, A. N. 1976, ApJ, 208, 709
Ford, A. B. & Shirley, Y. L. 2011, ApJ, 728, 144
Furlan, E. et al. 2008, 2008, ApJS, 176, 184
Galametz, M., Madden, S., Galliano, F., et al. 2009, A& A, 508, 645
Gallagher, S. C., Charlton, J. C., Hunsberger, S. D., Zaritsky, D., & Whitmore, B. C.2001, AJ, 122, 163
Galliano, F., Madden, S. C., Jones, A. P., et al. 2003, A& A, 407, 159
Gao, Y., & Xu, C. 2000, ApJ, 542, L83
Genzel, R., Tacconi, L. J., Combes, F., et al. 2012, ApJ, 746, 69
Genzel, R., Tacconi, L. J., Lutz, D., et al. 2015, ApJ, 800, 20
Glover, S. C. O., & Clark, P. C. 2012a, MNRAS, 421, 9
Glover, S. C. O., & Clark, P. C. 2012b, MNRAS, 426, 377
Goldsmith, P. F. 2001, ApJ, 557, 736
Gordon, K. D. 2004, Astrophysics of Dust, 309, 77
Green, G. M. et al. 2015, arxiv 1507.01005
Greve, T. R., Bertoldi, F., Smail, I., et al. 2005, MNRAS, 359, 1165
Griffin, M. J. et al. 2010, A&A, 518, L3
Guillard, P., Boulanger, F., Pineau Des Forets, G., & Appleton, P. N. 2009, A&A,502, 515
Guillard, P., Boulanger, F., Cluver, M. E., et al. 2010, A&A, 518, A59
Guillard, P., Ogle, P. M., Emonts, B. H. C., et al. 2012, ApJ, 747, 95
192
Guillard, P., Boulanger, F., Pineau des Forets, G., et al. 2012, ApJ, 749, 158
Guseva, N. G., Izotov, Y. I., Fricke, K. J., & Henkel, C. 2012, A& A, 541, A115
Hanel, R., Conrath, B., Flasar, M., et al. 1979, Science, 206, 952
Hanel, R., Conrath, B., Flasar, F. M., et al. 1981, Science, 212, 192
Hardegree-Ullman, E. et al. 2013, ApJ, 763, 45
Harnett, J. I., Wielebinski, R., Bajaja, E., Reuter, H.-P., & Hummel, E. 1991, Pro-ceedings of the Astronomical Society of Australia, 9, 258
Hennebelle, P. & Fromang, S. 2008, A&A, 477, 9
Henning, T. & Launhardt, R. 1998, A&A, 338, 223
Henyey, L. G. & Greenstein, J. L. 1941, ApJ, 93, 70
Herrera-Camus, R., Fisher, D. B., Bolatto, A. D., et al. 2012, ApJ, 752, 112
Higdon, S. J. U., Armus, L., Higdon, J. L., Soifer, B. T., & Spoon, H. W. W. 2006,ApJ, 648, 323
Hirashita, H. & Li, Z.-Y. 2013, MNRAS, 434, L70
Hocuk, S., Cazaux, S., & Spaans, M. 2014, MNRAS, 438, L56
Houck, J. R., Roellig, T. L., van Cleve, J., et al. 2004, ApJS, 154, 18
Hollenbach, D., & McKee, C. F. 1979, ApJS, 41, 555
Hollenbach, D. J., & Tielens, A. G. G. M. 1997, ARAA, 35, 179
Hollenbach, D. J., & Tielens, A. G. G. M. 1999, Reviews of Modern Physics, 71, 173
Horellou, C., Black, J. H., van Gorkom, J. H., et al. 2001, A& A, 376, 837
Huber, K. P. & Herzberg, G. 1979, Constants of Diatomic Molecules (New York: VanNostrand)
Hudson, R. D. 1971, Reviews of Geophysics and Space Physics, 9, 305
193
Hunt, L. K., Thuan, T. X., Izotov, Y. I., & Sauvage, M. 2010, ApJ, 712, 164
Hunt, L. K., Garcıa-Burillo, S., Casasola, V., et al. 2015, A& A, 583, A114
Ingalls, J. G., Bania, T. M., Boulanger, F., et al. 2011, ApJ, 743, 174
Israel, F. P., Bontekoe, T. R., & Kester, D. J. M. 1996, A& A, 308, 723
Israel, F. P. 1997, A& A, 328, 471
Israel, F. P., van Dishoeck, E. F., Baas, F., et al. 1990, A& A, 227, 342
Israel, F. P., Baas, F., Rudy, R. J., Skillman, E. D., & Woodward, C. E. 2003, A&A, 397, 87
Izotov, Y. I., & Thuan, T. X. 1998, ApJ, 500, 188
Izotov, Y. I., Thuan, T. X., & Stasinska, G. 2007, ApJ, 662, 15
Jones, A. P., Tielens, A. G. G. M., Hollenbach, D. J., & McKee, C. F. 1994, ApJ,433, 797
Jones, A. P. 2004, Astrophysics of Dust, 309, 347
Kennicutt, R. C., Jr. 1998, ApJ, 498, 541
Kennicutt, R. C., Jr., Armus, L., Bendo, G., et al. 2003, PASP, 115, 928
Kenyon, S. J., Dobrzycka, D. & Hartmann, L. 1994, AJ, 108, 1872
Kim, S.-H. & Martin, P. G. 1996, ApJ, 462, 296
Kirkpatrick, J. D. et al. 2011, ApJS, 197, 19
Kohno, K., Tosaki, T., Matsushita, S., et al. 2002, PASJ, 54, 541
Landolt, A. U. 1992, AJ, 104, 340
Launhardt, R., Ward-Thompson, D. & Henning, T.,1997, MNRAS, 288, L45
Launhardt, R. et al. 2010, ApJS, 188, 139
194
Launhardt, R. et al. 2013, A&A, 551, A98
Laurent, O., Mirabel, I. F., Charmandaris, V., et al. 2000, A&A, 359, 887
Lavezzi, T. E., & Dickey, J. M. 1998, AJ, 115, 405
Lefevre, C. et al. 2014, A$A, 572, A20
Lepp, S., Stancil, P. C., & Dalgarno, A. 2002, Journal of Physics B Atomic MolecularPhysics, 35, 57
Leroy, A., Bolatto, A., Stanimirovic, S., et al. 2007, ApJ, 658, 1027
Leroy, A.K., et al. 2011, ApJ, 737, 12
Li, Z.-Y. & Shu, F. H. 1997, ApJ, 475, 237
Lippok, N. et al. 2013, A&A, 560, A41
Lisenfeld, U., Isaak, K. G., & Hills, R. 2000, MNRAS, 312, 433
Lisenfeld, U., Braine, J., Duc, P.-A., et al. 2002, A& A, 394, 823
Lizano, S. & Shu, F. H. 1989, 342, 834
Loinard, L., Mioduszewski, A. J. & Rodrıguez, L. F. et al. 2005, ApJL, 179
Lutz, D., Spoon, H. W. W., Rigopoulou, D., Moorwood, A. F. M., & Genzel, R. 1998,ApJL, 505, L103
Lutz, D.; Sturm, E.; Genzel, R.; Spoon, H. W. W.; Moorwood, A. F. M.; Netzer, H.& Sternberg, A. 2003, A& A, 409, 867
Malhotra, S., Kaufman, M. J., Hollenbach, D., et al. 2001, ApJ, 561, 766
Markwardt, C. B. 2009, Astronomical Data Analysis Software and Systems XVIII,411, 251
Massey, P. et al. 2013, American Astronomical Society Meeting Abstracts, 221, 345.02
Mathis, J. S., Rumpl, W., & Nordsieck, K. H. 1977, ApJ, 217, 425
195
Mattila, K. 1970, A&A, 9, 53
McKee, C. F., & Ostriker, J. P. 1977, ApJ, 218, 148
Mentuch Cooper, E., Wilson, C. D., Foyle, K., et al. 2012, ApJ, 755, 165
Mirabel, I. F.; Sanders, D. B. & Kazes, I. 1989, ApJl, 340, L9
Moles, M., Marquez, I., & Sulentic, J. W. 1998, A& A, 334, 473
Motte, F., Andre, P. & Neri, R. 1998, A&A, 336, 150
Mouschovias, T. C. & Spitzer, Jr. L., 1976, ApJ, 210, 326
Moustakas, J., Kennicutt, R. C., Jr., Tremonti, C. A., et al. 2010, ApJS, 190, 233
Myers, P. C. 1983, ApJ, 270, 105
Natale, G., Tuffs, R. J., Xu, C. K., et al. 2010, ApJ, 725, 955
Nelson, R. P. & Langer, W. D. 1999, ApJ, 524, 923
Nesvadba, N. P. H., Boulanger, F., Salome, P., et al. 2010, A& A, 521, A65
Neufeld, D. A., & Yuan, Y. 2008, ApJ, 678, 974
Ngoumou, J., Hubber, D., et al. 2014, arxiv:1410.5279
Nikiel-Wroczynski, B., Soida, M., Urbanik, M., Beck, R., & Bomans, D. J. 2013,MNRAS, 435, 149
Ocana Flaquer, B., Leon, S., Combes, F., & Lim, J. 2010, A& A, 518, A9
O’Dowd, M. J., Schiminovich, D., Johnson, B. D., et al. 2009, ApJ, 705, 885
Ogle, P., Boulanger, F., Guillard, P., et al. 2010, ApJ, 724, 1193
Ogle, P. M., Lanz, L., & Appleton, P. N. 2014, ApJl, 788, L33
Okuda, T., Kohno, K., Iguchi, S., & Nakanishi, K. 2005, ApJ, 620, 673
Ossenkopf, V. & Henning, T. 1994, A&A, 291, 943
196
O’Sullivan, E., Giacintucci, S., Vrtilek, J. M., Raychaudhury, S., & David, L. P. 2009,ApJ, 701, 1560
Padelis P. Papadopoulos, Paul van der Werf; Xilouris, E.; Isaak, K. G. & Yu Gao2012, ApJ, 751, 10
Pagani, L., Steinacker, J., Bacmann, A., Stutz, A., & Henning, T. 2010, Science, 329,1622
Peimbert, A., Peimbert, M., & Ruiz, M. T. 2005, ApJ, 634, 1056
Pellegrini, E. W., Smith, J. D., Wolfire, M. G., et al. 2013, ApJL, 779, LL19
Pereira-Santaella, M., Alonso-Herrero, A., Rieke, G. H., et al. 2010, ApJs, 188, 447
Pereira-Santaella, M., Spinoglio, L., van der Werf, P. P., & Piqueras Lopez, J. 2014,A& A, 566, A49
Petitpas, G. R., & Taylor, C. L. 2005, ApJ, 633, 138
Pilbratt, G. L. et al. 2010, A&A, 518, 1
Pineda, J. L., Langer, W. D., Velusamy, T., & Goldsmith, P. F. 2013, A& A, 554,A103
Planck Collaboration, Ade, P. A. R., Aghanim, N., et al. 2011, A& A, 536, AA19
Planck Collaboration 2011, A&A, 536, A19
Planck Collaboration 2015, arxiv 1502.04123
Poglitsch, A., Waelkens, C., Geis, N., et al. 2010, A& A, 518, L2
Porter, T. A. & Strong, A. W. 2005, International Cosmic Ray Conference, 4, 77
Przybilla, N., Nieva, M.-F., & Butler, K. 2008, ApJL, 688, L103
Rachford, B. L., Snow, T. P., Tumlinson, J., et al. 2002, ApJ, 577, 221
Remy-Ruyer, A., Madden, S. C., Galliano, F., et al. 2014, A& A, 563, A31
Remy-Ruyer, A., Madden, S. C., Galliano, F., et al. 2015, A & A, 582, A121
197
Rigopoulou, D.; Lawrence, A.; White, G. J., Rowan Robinson, M., & Church, S. E.1996, A& A, 305, 747
Roelfsema, P., Giard, M., Najarro, F., et al. 2012, Procspie, 8442, 84420R
Roussel, H., et al. 2007, ApJ, 669, 959
Roy, A. et al. 2014, A&A, 562, A138
Rubio, M., Lequeux, J., Boulanger, F., et al. 1996, A& As, 118, 263
Sakamoto, K., Ho, P. T. P., & Peck, A. B. 2006, ApJ, 644, 862
Salome, P., & Combes, F. 2003, A& A, 412, 657
Sanders, D. B., Soifer, B. T., Elias, J. H., Neugebauer, G., & Matthews, K. 1988,ApJL, 328, L35
Sanders, D. B., Scoville, N. Z., & Soifer, B. T. 1991, ApJ, 370, 158
Sanders, D. B., & Mirabel, I. F. 1996, ARAA, 34, 749
Sanders, D. B., Mazzarella, J. M., Kim, D.-C., Surace, J. A., & Soifer, B. T. 2003,AJ, 126, 1607
Sandstrom, K. M., Leroy, A. K., Walter, F., et al. 2013, ApJ, 777, 5
Saripalli, L., & Mack, K.-H. 2007, MNRAS, 376, 1385
Savage, B. D., Bohlin, R. C., Drake, J. F., & Budich, W. 1977, ApJ, 216, 291
Schlafly, E. F. et al. 2010, ApJ, 725, 1175
Schlafly, E. F. & Finkbeiner, D. P. 2012, ApJ, 737, 103
Schlegel, D. J., Finkbeiner, D. P. & Davis, M. 1998, ApJ, 500, 525
Schmalzl, M. et al. 2014, A&A, 569, A7
Schruba, A., Leroy, A. K., Walter, F., et al. 2012, AJ, 143, 138
198
Scoville, N. Z., Yun, M. S., Sanders, D. B., Clemens, D. P., & Waller, W. H. 1987,ApJS, 63, 821
Scoville, N., Aussel, H., Sheth, K., et al. 2014, ApJ, 783, 84
Scoville, N., Sheth, K. Aussel, H., et al. arxiv 1505.02159
Shi, Y., Wang, J., Zhang, Z.-Y., et al. 2015, ApJL, 804, L11
Shu, F. H. 1977, ApJ, 214, 488
Skibba, R. A., Engelbracht, C. W., Dale, D., et al. 2011, ApJ, 738, 89
Smith, J. D. T., Dale, D. A., Armus, L., et al. 2004, ApJS, 154, 199
Smith, J. D. T., Armus, L., Dale, D. A., et al. 2007a, PASP, 119, 1133
Smith, J. D. T., Draine, B. T., Dale, D. A., et al. 2007b, ApJ, 656, 770
Smith, R. J., Glover, S. C. O., Clark, P. C., Klessen, R. S., & Springel, V. 2014,MNRAS, 441, 1628
Smith, B. J., & Struck, C. 2001, AJ, 121, 710
Snow, T. P., & McCall, B. J. 2006, ARAA, 44, 367
Sodroski, T. J. et al. 1997, ApJ, 480, 173
Soifer, B. T., Neugebauer, G., & Houck, J. R. 1987, ARAA, 25, 187
Solomon, P. M., Rivolo, A. R., Barrett, J., & Yahil, A. 1987, ApJ, 319, 730
Solomon, P. M.; Downes, D.; Radford, S. J. E. & Barrett, J. W. 1997, 478, 144
Solomon, P. M., & Vanden Bout, P. A. 2005, ARAA, 43, 677
Spitzer, L., Jr., & Cochran, W. D. 1973, ApJL, 186, L23
Spitzer, L., Drake, J. F., Jenkins, E. B., et al. 1973, ApJL, 181, L116
Spitzer, L., Jr., & Zweibel, E. G. 1974, ApJl, 191, L127
199
Spitzer, L., Jr., Cochran, W. D., & Hirshfeld, A. 1974, ApJS, 28, 373
Spoon, H. W. W., Marshall, J. A., Houck, J. R., et al. 2007, ApJL, 654, L49
Sreenilayam, G. et al. 2014, AJ, 147, 53
Stahler, S. W., & Palla, F. 2005, The Formation of Stars, by Steven W. Stahler,Francesco Palla, pp. 865. ISBN 3-527-40559-3. Wiley-VCH , January 2005., 865
Steinacker, J., Pagani, L. et al. 2010, A&A, 511, A9
Steinacker, J., Ormel, C. W. 2014, A&A, 564, A96
Steinacker, J., Andersen, M. et al. 2014, A&A, 563, A106
Stierwalt, S., Armus, L., Surace, J. A., et al. 2013, ApJS, 206, 1
Stierwalt, S., Armus, L., Charmandaris, V., et al. 2014, ApJ, 790, 124
Strong, A. W., & Mattox, J. R. 1996, A& A, 308, L21
Stutz, A. M. et al. 2008, ApJ, 687, 389
Stutz, A. et al. 2010, A&A, 518, L87
Sulentic, J. W., Pietsch, W., & Arp, H. 1995, A& A, 298, 420
Sulentic, J. W., Rosado, M., Dultzin-Hacyan, D., et al. 2001, AJ, 122, 2993
Suzuki, T., Kaneda, H., Onaka, T., & Kitayama, T. 2011, ApJL, 731, L12
Terebey, S. & Shu, F. H. and Cassen, P. 1984, ApJ, 286, 529
Testi, L., Palla, F. & Natta, A. 1999, A&A, 342, 515
Testi, L., Palla, F. & Natta, A., 2001, Astronomical Society of the Pacific ConferenceSeries, 243, 377
Tielens, A. G. G. M. 2005, The Physics and Chemistry of the Interstellar Medium,by A. G. G. M. Tielens, pp. . ISBN 0521826349. Cambridge, UK: CambridgeUniversity Press, 2005.,
200
Trinchieri, G., Sulentic, J., Breitschwerdt, D., & Pietsch, W. 2003, A& A, 401, 173
U, V., Sanders, D. B., Mazzarella, J. M., et al. 2012, ApJS, 203, 9
Valentijn, E. A., & van der Werf, P. P. 1999, ApJL, 522, L29
van der Hulst, J. M., & Rots, A. H. 1981, AJ, 86, 1775
van Dokkum, P. G. 2001, PASP, 113, 1420
Veilleux, S., Cecil, G., & Bland-Hawthorn, J. 2005, ARAA, 43, 769
Veilleux, S., Rupke, D. S. N., Kim, D.-C., et al. 2009, ApJS, 182, 628
Velusamy, T., & Langer, W. D. 2014, A& A, 572, A45
Walter, F., Bertoldi, F., Carilli, C., et al. 2003, Nature, 424, 406
Ward-Thompson, D. et al. Protostars and Planets V. 2007, 33
Watson, D., Christensen, L., Knudsen, K. K., et al. 2015, Nature, 519, 327
Wilson, C. D., Cridland, A., Foyle, K., et al. 2013, ApJl, 776, LL30
Witt, A. N. & Stephens, T. C. 1974, AJ, 79, 948
Witt, A. N., Oliveri, M. V. & Schild, R. E. 1990, AJ, 99, 888
Witt, A. N., Mandel, S., et al. 2008, ApJ, 679, 497
Wolf, S., Launhardt, R. & Henning, T. 2003, ApJ, 592, 233
Wolfire, M. G., David Hollenbach., & Christopher, F. McKee. 2010, ApJ, 716, 1191
Wright, E. L. et al. 2010, AJ, 140, 1868
Xu, C. K., Lu, N., Condon, J. J., Dopita, M., & Tuffs, R. J. 2003, ApJ, 595, 665
Xu, C. K., Iglesias-Paramo, J., Burgarella, D., et al. 2005, ApJL, 619, L95
Yen, H.-W. et al. 2014, ApJ, 793, 1
201
Young, L. M., Bureau, M., Davis, T. A., et al. 2011, MNRAS, 414, 940
Ysard, N. et al. 2013, A&A, 559, A133
Yun, J. L. & Clemens, D. P. 1990, ApJL, 365, 73
Yun, J. L. & Clemens, D. P., 1992, ApJL, 385, 21
Zakamska, N. L. 2010, Nature, 465, 60
202
top related