Devaraj Kiruthika 201112 PHD

THE CENTIMETER- AND
MILLIMETER-WAVELENGTH AMMONIA
ABSORPTION SPECTRA UNDER JOVIAN
CONDITIONS
A Dissertation
Presented to
The Academic Faculty
by
Kiruthika Devaraj
In Partial Fulfillment
of the Requirements for the Degree
Doctor of Philosophy in
Electrical and Computer Engineering
School of Electrical and Computer Engineering

Georgia Institute of Technology
December 2011
THE CENTIMETER- AND
MILLIMETER-WAVELENGTH AMMONIA
ABSORPTION SPECTRA UNDER JOVIAN
CONDITIONS
Approved by:
Professor Paul G. Steffes, Advisor Professor Oliver Brand

School of Electrical and Computer School of Electrical and Computer
Engineering Engineering
Georgia Institute of Technology Georgia Institute of Technology
Professor Waymond R. Scott Professor Carol Paty

School of Electrical and Computer School of Earth and Atmospheric
Engineering Science
Georgia Institute of Technology Georgia Institute of Technology
Professor Gregory D. Durgin Date Approved: September 19, 2011

School of Electrical and Computer
Engineering
Georgia Institute of Technology
To my parents
iii
ACKNOWLEDGEMENTS
I express my deepest gratitude to my advisor Professor Paul Steffes for his support
and guidance throughout the course of this research. I am grateful for his tremen-
dous encouragement and constant faith in my abilities as a researcher. I also thank
my committee members, Professor Oliver Brand, Professor Gregory Durgin, Profes-
sor Carol Paty, and Professor Waymond Scott for spending the time and effort in
reviewing this work.
I acknowledge the NASA Planetary Atmospheres Program and the NASA Juno
Mission Team for their financial support of this work. This work was supported by
NASA Contract NNM06AA75C from the Marshall Space Flight Center supporting
the Juno Mission Science Team, under Subcontract 699054X from the Southwest
Research Institute and by the NASA Planetary Atmospheres Program under Grants
NNG06GF34G and NNX11AD66G.
I spent two summers at the National Radio Astronomy Observatory (NRAO) in
Soccoro, NM as a graduate summer student and I thank NRAO for providing me with
an opportunity to work on planetary radio astronomy. I thank Dr. Bryan Butler for
his guidance on planetary radio observations, data reduction, and analysis. I wish to
express my gratitude to Dr. Brigette Hesman for her guidance on radiative transfer
models of planetary atmospheres and her constant support and encouragement ever
since.
The success of the laboratory experiments was made possible as a result of many
contributions over the years. I wish to thank Dr. Thomas Hanley for automating the
microwave measurement subsystem and data processing and Dr. Bryan Karpowicz for
building the high-pressure system, both of which were utilized in this research work.
iv
I thank them and Danny Duong for their support and many conversations. Thanks
to the all the staff in the school of ECE for keeping the department operational.
My graduate student experience would not have be complete had it not been for
all the wonderful friends I made over the years I was at Georgia Tech. I thank Manali
Tare for being my friend, and confidant, and for introducing me to masala chai, which
I consumed religiously every morning for a good part of four years. I also thank my
friends for making my stay at Georgia Tech nothing short of wonderful: Nandita
Yeshala, Smita Vemulapalli, Vibhore Kumar, Sonali Tare, Karan Singhal, Nagesh
and Deepthi Adiga, Swanand Gupte, Shreekrisha, Manoj Agarwal, Yogish Gopala,
and Sai Kumar Thumuluru.
The constant support and encouragement from my family sustained me through-
out these years. My parents Shanthamani Soundirarajan and Devaraj Ranganathan
have always supported me and I cannot thank them enough for their unconditional
love and unshakable belief in me. My brother Karthik Vignesh Devaraj, without
whom growing up would have been boring, ensured my life was filled with fun and
happiness, and the occasional squabbles. I spent a good part of my childhood with
my grandparents and I cannot express my gratitude for their unconditional love. I
thank my wonderful husband Lakshmi Narasimhan Chakrapani, without whom this
dissertation would not have been possible. His passion for research and curiosity to
understand the animate and the inanimate world continuously fueled and strength-
ened my research interests. I thank him for his support, love, and endless patience.
v
TABLE OF CONTENTS
DEDICATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii
ACKNOWLEDGEMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . iv
LIST OF TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
LIST OF FIGURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi
LIST OF SYMBOLS OR ABBREVIATIONS . . . . . . . . . . . . . . xxi
SUMMARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxvi
I INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Science Objectives and Applications . . . . . . . . . . . . . . . . . . 7
1.3 Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
II MICROWAVE SPECTROSCOPY AND AMMONIA SPECTRA 11

2.1 Absorption Spectra of Ammonia . . . . . . . . . . . . . . . . . . . . 12
2.2 Linewidths . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.3 Lineshapes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.4 Line Intensity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
III LABORATORY MEASUREMENTS OF AMMONIA . . . . . . . 22

3.1 Electromagnetic Propagation in a Homogeneous, Isotropic Medium . 22
3.2 Measurement Theory . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.3 Millimeter-Wavelength Measurement System . . . . . . . . . . . . . 28
3.3.1 Planetary Atmospheric Simulator . . . . . . . . . . . . . . . 28
3.3.2 Millimeter-Wavelength Subsystem . . . . . . . . . . . . . . . 31
3.3.3 Data Handling Subsystem . . . . . . . . . . . . . . . . . . . 40
3.4 High-Pressure Measurement System . . . . . . . . . . . . . . . . . . 40
3.4.1 High-Pressure Planetary Atmospheric Simulator . . . . . . . 40
3.4.2 Centimeter-Wavelength Subsystem . . . . . . . . . . . . . . . 43
vi
3.4.3 Data Acquisition System . . . . . . . . . . . . . . . . . . . . 49
3.5 Measurement Procedure . . . . . . . . . . . . . . . . . . . . . . . . . 51
3.6 Data Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
3.7 Measurement Uncertainties . . . . . . . . . . . . . . . . . . . . . . . 57
IV COMPRESSIBILITY OF FLUIDS . . . . . . . . . . . . . . . . . . . 67
4.1 Hydrogen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
4.2 Helium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
4.3 Ammonia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
4.4 Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
V MEASUREMENTS AND MODEL . . . . . . . . . . . . . . . . . . 86

5.1 Millimeter-Wavelength Ammonia Opacity Measurements . . . . . . . 86
5.2 High-Pressure Centimeter-Wavelength Ammonia Opacity Measure-
ments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
5.3 Data Fitting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
5.4 New Consistent Ammonia Absorption Formalism . . . . . . . . . . . 95
5.4.1 Line Parameters . . . . . . . . . . . . . . . . . . . . . . . . . 97
5.4.2 Ammonia Opacity Formalism . . . . . . . . . . . . . . . . . . 98
5.5 Model Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
5.5.1 Ultra-High-Pressure Extrapolation . . . . . . . . . . . . . . . 135
5.6 The Influence of Water Vapor on the Ammonia Absorption Spectrum 140
5.6.1 Ammonia/Water Vapor Opacity Measurements . . . . . . . . 141
5.6.2 Preliminary Ammonia/Water Vapor Opacity Model . . . . . 143
VI MILLIMETER-WAVELENGTH OBSERVATIONS OF JUPITER147

6.1 IRAM Observations . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
6.2 Radiative Transfer Model . . . . . . . . . . . . . . . . . . . . . . . . 152
6.3 Search for the 140 GHz Line . . . . . . . . . . . . . . . . . . . . . . 154
VII CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160

7.1 Contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
vii
7.1.1 Millimeter-wavelength measurement system . . . . . . . . . . 161
7.1.2 Laboratory measurements and model . . . . . . . . . . . . . 161
7.1.3 Search for the 140 GHz line . . . . . . . . . . . . . . . . . . . 162
7.2 Recommendations for Future Work . . . . . . . . . . . . . . . . . . 162
7.2.1 Millimeter-wavelength laboratory work . . . . . . . . . . . . 162
7.2.2 Centimeter-wavelength laboratory work . . . . . . . . . . . . 163
7.3 List of Publications . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
APPENDIX A — MILLIMETER-WAVELENGTH COMPONENTS168
APPENDIX B — VLA OBSERVATIONS OF THE 2009 JUPITER

IMPACT EVENT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
APPENDIX C — VLA OBSERVATIONS OF VENUS AT X BAND184
BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
VITA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
viii
LIST OF TABLES
3.1 High-pressure planetary atmospheric simulator instruments. . . . . . 46

4.1 Coefficients of the normal hydrogen ideal gas heat capacity equation. 72
4.2 Coefficients and parameters of the ideal part of the reduced Helmholtz
free energy equation for normal hydrogen. . . . . . . . . . . . . . . . 73
4.3 Parameters and coefficients of the residual part of the reduced Helmholtz
free energy term for normal hydrogen. . . . . . . . . . . . . . . . . . . 75
4.4 Coefficients of the mBWR equation of state for helium. . . . . . . . . 76
4.5 Coefficients of the ideal part of the reduced Helmholtz free energy for
ammonia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
free energy for ammonia. . . . . . . . . . . . . . . . . . . . . . . . . . 79
4.7 Coefficients and parameters of the ideal part of the reduced Helmholtz
free energy for water. . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
free energy for water. . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
5.1 Listing of all experimental conditions for the 2–4 mm-wavelength am-
monia opacity measurements conducted using the Fabry–Perot res-
onator as part of this work. . . . . . . . . . . . . . . . . . . . . . . . 87
5.2 Listing of all experiment sequences of the 5–20 cm-wavelength ammo-
nia opacity measurements conducted using the high-pressure system
as part of this work. . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
5.3 The breakdown in the fTPC space of the measurement database con-
sisting of the 75–150 GHz FPR measurements (this work) and the
1.5–27 GHz cavity resonator measurements (Hanley et al., 2009) used
in the first stage of optimization. . . . . . . . . . . . . . . . . . . . . 93
5.4 The breakdown in the fTPC space of the measurement database con-
sisting of the 75–150 GHz FPR measurements (this work), 1.5–27 GHz
cavity resonator measurements (Hanley et al., 2009), and the 1.5–6
GHz high-pressure measurements (this work) used in the second stage
of optimization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
5.5 Values of the low-pressure inversion model constants used for comput-
ing the H2 /He-broadened NH3 absorptivity when P <= 15 bar. . . . 101
5.6 Values of the high-pressure inversion model constants used for com-
puting the H2 /He-broadened NH3 absorptivity when P > 15 bar. . . 101
ix
5.7 Values of the model constants of the new model used for computing
the H2 /He-broadened NH3 absorptivity from the rotational transitions. 102
5.8 Values of the model constants of the new model used for computing
the H2 /He-broadened NH3 absorptivity from the ν2 roto-vibrational
transitions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
5.9 The percentage of the NH3 /He/H2 measurement data points within 2σ
uncertainty of the new model in comparison with the existing models. 106
5.10 Listing of all experiment sequences of the 5–20 cm-wavelength opacity
measurements of the NH3 /H2 O mixture. . . . . . . . . . . . . . . . . 142
5.11 Values of the H2 O model constants used for computing the H2 O-broadened
NH3 absorptivity from the inversion transitions. . . . . . . . . . . . . 144
A.1 AMC-10-RFH00 parameters. . . . . . . . . . . . . . . . . . . . . . . . 170
A.2 Quinstar QMH 922WHP/387 parameters in “external mixer” mode. . 170
A.3 Spacek amplifier and tripler parameters. . . . . . . . . . . . . . . . . 172
B.1 Summary of the VLA observations of Jupiter along with the duration
of observation and the geometry of the planet visible from Earth. . . 178
C.1 Ephemeris information and the calibrators for Venus observations. . . 184
x
LIST OF FIGURES
2.1 Ammonia molecule. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

2.2 The 140.14 GHz ν2 = 1 inversion transition occurs when the ammonia
molecule transitions between the lower inversion level of the ground-
state J = 2, K = 1 rotational transition and the upper inversion level
of the ground-state J = 1, K = 1 rotational transition. . . . . . . . . 15
2.3 Line positions and intensities of the NH3 and NH3 − ν2 transitions from
0.3–1300 GHz (Yu et al., 2010a,b,c). . . . . . . . . . . . . . . . . . . 17
3.1 A picture of the glass pressure vessel enclosing the Fabry–Perot res-
onator used for simulating the jovian atmospheric conditions. . . . . . 29
3.2 Block diagram of the spherical mirror Fabry–Perot resonator placed in
a near-confocal configuration. . . . . . . . . . . . . . . . . . . . . . . 32
3.3 Quality factor of the resonances of the Fabry–Perot resonator measured
at vacuum and room temperature (T=297 K). . . . . . . . . . . . . . 34
3.4 Effective path length of the resonances of the Fabry–Perot resonator
measured at vacuum and room temperature (T=297 K). . . . . . . . 35
3.5 Block diagram of the W band measurement system for studying am-
monia gas properties under simulated upper tropospheric jovian con-
ditions. Solid lines represent the electrical connections and the arrows
show the direction of signal propagation. Valves controlling the flow
of gases are shown by the small crossed circles. . . . . . . . . . . . . 36
3.6 Block diagram of the F band measurement system for studying am-
monia gas properties under simulated upper tropospheric jovian con-
ditions. Solid lines represent the electrical connections and the arrows
show the direction of signal propagation. Valves controlling the flow
of gases are shown by the small crossed circles. . . . . . . . . . . . . . 38
3.7 The Georgia Tech high-pressure system used for studying the centimeter-
wavelength properties of ammonia under simulated jovian conditions (Kar-
powicz and Steffes, 2011). The valves shown with a blue dot are high
temperature valves. . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.8 Assembly of the high-pressure system. . . . . . . . . . . . . . . . . . 45
3.9 Pictures of the pressure vessel in and after assembly. . . . . . . . . . 46
3.10 Cylindrical cavity resonator. . . . . . . . . . . . . . . . . . . . . . . . 47
3.11 The centimeter-wavelength subsystem and the data-acquisition com-
ponents of the high-pressure system (Karpowicz and Steffes, 2011). . 50
xi
3.12 Percentage contribution of the different measurement uncertainties to
the total uncertainty of the W band system at room temperature
(T=297 K). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
3.13 Percentage contribution of the different measurement uncertainties to
the total uncertainty of the F band system at T=218 K. . . . . . . . 66
3.14 Measured system sensitivity at room temperature (T=297 K) in the
2–4 millimeter-wavelength range. . . . . . . . . . . . . . . . . . . . . 66
4.1 The compressibility of various pure fluids at 500 K as a function of
pressure. For H2 and He, the molecular repulsive forces dominate un-
der these conditions and hence the measured pressure is greater than
the ideal pressure. For NH3 and H2 O, the molecular attractive forces
dominate under these conditions and hence the measured pressure is
less than the ideal pressure. . . . . . . . . . . . . . . . . . . . . . . . 69
4.2 The compressibility (Z) of pure normal hydrogen. . . . . . . . . . . . 74
4.3 The compressibility (Z) of pure helium. . . . . . . . . . . . . . . . . . 77
4.4 The compressibility (Z) of pure ammonia. . . . . . . . . . . . . . . . 80
4.5 The compressibility (Z) of pure water. . . . . . . . . . . . . . . . . . 85
5.1 Dry jovian adiabatic temperature-pressure (TP) profile along with the
TP space measurement points used in the model development and/or
evaluation of the new model performance. Red crosses are the 1.5–
27 GHz cavity resonator TP space points measured by Hanley et al.
(2009), blue triangles are the 22–40 GHz FPR TP space points mea-
sured by Hanley et al. (2009), black circles are the 75–150 GHz FPR
TP space points measured as part of this work, and black asterisks are
the 1.5–6 GHz high-pressure TP space points measured as part of this
work. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
5.2 Opacity data measured by Hanley et al. (2009) using the cavity res-
onators for a mixture of NH3 = 0.95%, He = 13.47%, H2 = 85.58% at a
pressure of 1.009 bar and temperature of 216.4 K compared to various
models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
onators for a mixture of NH3 = 0.77%, He = 13.5%, H2 = 85.73% at a
models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
onators for a mixture of NH3 = 4%, He = 13.06%, H2 = 82.94% at a
models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
xii
onators for a mixture of NH3 = 4%, He = 13.06%, H2 = 82.94% at a
models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
5.6 Opacity data measured by Hanley et al. (2009) using the Fabry-Perot
resonator for a mixture of NH3 = 4%, He = 13.06%, H2 = 82.94% at
a pressure of 2 bar and temperature of 295.3 K compared to various
models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
5.7 Opacity data measured using the 3–4 mm-wavelength system for a
mixture of NH3 = 4%, He = 13.06%, H2 = 82.94% at a pressure of
0.976 bar and temperature of 296.3 K compared to various models. . 109
mixture of NH3 = 6.04%, He = 12.78%, H2 = 81.18% at a pressure of
xiii
5.19 Opacity data measured using the 3–4 mm-wavelength system for pure
ammonia gas at a pressure of 0.255 bar and temperature of 295.1 K
compared to various models. . . . . . . . . . . . . . . . . . . . . . . 115
1.854 bar and temperature of 296 K compared to various models. . . 118
mixture of NH3 = 10.89%, He = 12.12%, H2 = 76.99% at a pressure
of 1.089 bar and temperature of 220.7 K compared to various models. 120
xiv
5.32 Opacity data measured using the high-pressure centimeter-wavelength
system for a mixture of NH3 = 100%, He = 0%, H2 = 0% at a pressure
of 0.09 bar and temperature of 376 K compared to various models. . 122
system for a mixture of NH3 = 0.71%, He = 99.29%, H2 = 0% at a
models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
system for a mixture of NH3 = 0.42%, He = 58.51%, H2 = 41.07% at a
pressure of 21.411 bar and temperature of 376 K compared to various
models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
system for a mixture of NH3 = 0.11%, He = 15.77%, H2 = 84.12%
at a pressure of 79.454 bar and temperature of 376.1 K compared to
various models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
xv
system for a mixture of NH3 = 0.59%, He = 0%, H2 = 99.41% at a
models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
various models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
various models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
xvi
various models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
various models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
various models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
various models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
5.58 Normalized nadir-viewing weighting functions in the 0.5–25 GHz range
computed by Karpowicz (2010) using a radiative transfer model for a
mean jovian atmosphere without cloud contribution. . . . . . . . . . . 136
5.59 Opacity data by Morris and Parsons (1970) for a mixture of NH3 =
0.9%, He = 99.1%, H2 = 0% at a frequency of 9.58 GHz and temper-
ature of 295 K compared to this work, Hanley (2008) with rotational
lines and Berge and Gulkis (1976) . . . . . . . . . . . . . . . . . . . 138
xvii
5.60 Opacity data measured by Morris and Parsons (1970) for a mixture
of NH3 = 0.44%, He = 0%, H2 = 99.56% at a frequency of 9.58 GHz
and temperature of 295 K compared to this work, Hanley (2008) with
rotational lines, and Berge and Gulkis (1976) . . . . . . . . . . . . . 139
system for a mixture of NH3 = 8.71% and H2 O = 91.29% at a pressure
of 1.041 bar and temperature of 452 K compared to Karpowicz and
Steffes (2011) H2 O model, the NH3 model described in Section 5.4.2
and the preliminary NH3 + H2 O model that includes the interaction
between NH3 and H2 O. . . . . . . . . . . . . . . . . . . . . . . . . . 145
system for a mixture of NH3 = 16.44% and H2 O = 83.56% at a pressure
of 0.18 bar and temperature of 373.9 K compared to Karpowicz and
Steffes (2011) H2 O model, the NH3 model described in Section 5.4.2
and the preliminary NH3 + H2 O model that includes the interaction
between NH3 and H2 O. . . . . . . . . . . . . . . . . . . . . . . . . . 146
6.1 Antenna temperature spectra obtained with the H polarization receiver
on September 17, 2010 for Jupiter and the Moon. The abscissa on top
is frequency in MHz and the ordinate is antenna temperature in K. . 149
6.2 Antenna temperature spectra obtained with the V polarization receiver
6.3 Antenna temperature spectra obtained with the H polarization receiver
6.4 Antenna temperature spectra obtained with the V polarization receiver
6.5 Ratio of the 140.14 GHz line to continuum spectrum (H polarization)
obtained on September 17, 2010. The abscissa on top is frequency in
MHz. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
6.6 Ratio of the 140.14 GHz line to continuum spectrum (H polarization)
obtained on September 18, 2010. The abscissa on top is frequency in
MHz. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
xviii
6.7 The temperature-pressure (TP) profile of Jupiter is shown as a black
line. The TP profile above the 1-bar level represents the Voyager radio
occultation results summarized by Lindal (1992), and the TP profile
below the 1-bar level represents the results of a wet-adiabatic extrapo-
lation using the thermochemical model. The deep ammonia abundance
is fixed at 800 ppm and the ammonia abundance profiles corresponding
to the saturation (Sat), hot spot (HS), north equatorial belt (NEB),
equatorial zone (EZ), and Galileo Probe (GP) models are also shown. 156
6.8 The TP profile of Jupiter (black line) and the normalized weighting
functions at 140.1 GHz for various ammonia abundance profiles. . . 157
6.9 Modeled disk-averaged brightness temperature of Jupiter for various
ammonia profiles. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
6.10 Disk-averaged brightness temperature measurements of Jupiter. . . . 159
A.1 Power output of the AMC for Pin =+10 dBm. . . . . . . . . . . . . . 169
A.2 Components of the W band subsystem. . . . . . . . . . . . . . . . . . 169
A.3 F band components: Spacek amplifier and frequency tripler. . . . . . 171
A.4 Power output of the Spacek amplifier and the tripler. . . . . . . . . . 172
A.5 F band harmonic mixer and diplexer. . . . . . . . . . . . . . . . . . . 173
A.6 JCA/Miteq amplifiers and bias tee. . . . . . . . . . . . . . . . . . . . 173
A.7 Output component parameters. . . . . . . . . . . . . . . . . . . . . . 174
B.1 Hubble image of Jupiter and its impact spot. Image credit: NASA,
ESA, H. Hammel (Space Science Institute, Boulder, CO), and the
Jupiter impact team. . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
B.2 Jupiter observation geometry on August 1, 2009. The red dots show
the progression of the coordinates of the impact site as the location
rotates across the disk during the observation. . . . . . . . . . . . . . 179
B.3 Maps of Jupiter at 3.5 cm made on July 26, 2009. . . . . . . . . . . . 181
B.4 Maps of Jupiter showing the thermal emission at 1.3 cm. . . . . . . . 182
B.5 Rotationally deprojected map of Jupiter at 3.5 cm averaged over ob-
servations from July 22–27, 2009. A black oval highlights the impact
site. No signature of the impact on the thermal emission at 3.5 cm, at
a depth of ∼ 1 bar, and with effective resolution of ∼ 6800 km, was
detected. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
C.1 Disk map of Venus as seen from Earth (left) and radar surface map of
Venus (right) on April 30, 1996. Image source: US Naval Observatory. 185
xix
C.2 Disk map of Venus as seen from Earth (left) and radar surface map of
Venus (right) on July 07, 2009. Image source: US Naval Observatory. 185
C.3 X band map of Venus taken with the VLA on July 07, 2009. . . . . . 187
C.4 Residual X band map of Venus for the July 07, 2009 observation. . . 188
C.5 Residual X band map of Venus for the April 30, 1996 observation. . . 188
C.6 The TP profile of Venus (left) and the weighting functions at the disk
center (right). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
xx
LIST OF SYMBOLS OR ABBREVIATIONS
α Multiple uses: Absorption, reduced Helmholtz free energy.
AIPS Astronomical Image Processing Software.
ALMA Atacama Large Millimeter Array.
AMC Active multiplier chain.
AU Astronomical unit. 1 AU = 149597870.691 km.
arcsec 1 arcsecond is 1/3600 of a degree.
a Helmholtz free energy.
A Absorption at the line center.
As Asymmetry of resonance.
β Phase constant.
BW Bandwidth.
B Confidence coefficient of a two-tailed distribution.
CH4 Methane.
CW Continuous wave.
C band Electromagnetic frequencies from 4 to 8 GHz.
c Speed of light.
cp Heat capacity.
∆ν Linewidth (half width at half-maximum).
D Multiple uses: Distance of separation between the mirrors in

a Fabry–Perot resonator, scale factor in the Ben-Reuven line-
shape.
dB Decibel.
dBm Decibel with respect to 1 mW.
DC Direct current.
DW Data weight.
E Electric field.
xxi
EMIR Eight MIxer Receiver frontend on IRAM 30-m telescope.
EPL Effective path length.
ESA European Space Agency.
ε Permittivity of the medium.
E Energy.
ε0 Real part of permittivity of the medium.
ε00 Imaginary part of permittivity of the medium.
FPR Fabry–Perot resonator.
FSR Free spectral range.
FS Full scale.
FWHM Full width at half maximum.
F band Electromagnetic frequencies from 90 to 140 GHz.
f Frequency.
F Lineshape function.
f0 Resonant frequency.
GEISA Gestion et Etude des Informations Spectroscopiques Atmosphériques:

Management and Study of Spectroscopic Information.
GE General Electric.
GPIB General Purpose Interface Bus.
G band Electromagnetic frequencies from 140 to 220 GHz.
H Magnetic field.
H2 Hydrogen.
He Helium.
h Planck constant.
h0 Ideal gas enthalpy.
ht Height of the cylindrical cavity resonator.
HP Hewlett-Packard.
xxii
HPBW Multiple uses: Half-power bandwidth, half-power beamwidth.
HWHM Half width at half maximum.
IF Intermediate frequency.
IRAM Institut de Radioastronomie Millimétrique.
I Line intensity.
JPL Jet Propulsion Lab.
Jy Jansky. Unit of spectral flux density. 1 Jy = 10−26 W/(m2 .Hz).
J Total angular momentum vector.
K band Electromagnetic frequencies from 18 to 26.5 GHz.
Ka band Electromagnetic frequencies from 26.5 to 40 GHz.
Ku band Electromagnetic frequencies from 12 to 18 GHz.
K Projection of the angular momentum vector onto the molecular

axis.
k Wavenumber.
λ Wavelength.
LNA Low noise amplifier.
LO Local oscillator.
LSD Least significant digit.
L band Electromagnetic frequencies from 1 to 2 GHz.
MWR Microwave radiometer on board the Juno spacecraft.
mBWR Modified Benedict-Webb-Rubin equation of state.
µ Permeability of the medium.
M Molecular weight.
NASA National Aeronautics and Space Administration.
NH3 Ammonia.
NIST National Institute of Standards and Technology.
NPT National pipe thread.
xxiii
NRAO National Radio Astronomy Observatory.
ν Frequency (cm−1 ).
ν0 Center frequency of line transition.
ν2 Higher energy state.
N Refractivity.
n Number density of a gas.
NH Harmonic number.
nri Refractive index of a gas.
Nsamples Number of samples.
ω Radial frequency.
P Pressure.
Q band Electromagnetic frequencies from 33 to 50 GHz.
Q Quality factor.
RBW Resolution bandwidth.
ρ Density.
% Reduced density.
R Specific gas constant.
R0 Ideal or universal gas constant.
rd Radius of the cylindrical cavity resonator.
REFPROP Reference fluid thermodynamic and transport properties database.
RF Radio frequency.
RFI Radio frequency interference.
RTD Resistance temperature detector.
RTM Radiative transfer model.
SCPI Standard commands for programmable instruments.
SNR Signal to noise ratio.
S-parameter Elements of a scattering matrix.
xxiv
SiO2 Silicon dioxide.
σ Multiple uses: Standard deviation, measurement uncertainty.
S Insertion loss of resonator.
s0 Ideal gas entropy.
SN Standard deviation of measurements.
TCM Thermochemical model.
TEM Transverse electromagnetic mode.
TE Transverse electric mode.
TM Transverse magnetic mode.
TP Temperature-Pressure.
tan δ Loss tangent.
τ Reciprocal of the reduced temperature.
T Temperature.
t Transmissivity.
UHP Ultra-high-purity.
USB Universal serial bus.
UT Universal time.
VLA Very Large Array.
VSWR Voltage standing wave ratio.
V Volume.
W band Electromagnetic frequencies from 75 to 110 GHz.
X band Electromagnetic frequencies from 8 to 12 GHz.
ζ Line-to-line coupling element in the Ben-Reuven lineshape.
Z Compressibility.
xxv
SUMMARY
Accurate knowledge of the centimeter- and millimeter-wavelength absorptivity
of ammonia is necessary for the interpretation of the emission spectra of the jovian
planets. The objective of this research has been to advance the understanding of
the centimeter- and millimeter-wavelength opacity spectra of ammonia under jovian
conditions using a combination of laboratory measurements and theoretical formula-
tions. As part of this research, over 1000 laboratory measurements of the 2–4 mm-
wavelength properties of ammonia under simulated upper and middle tropospheric
conditions of the jovian planets, and approximately 1200 laboratory measurements
of the 5–20 cm-wavelength properties of ammonia under simulated deep tropospheric
conditions of the jovian planets have been conducted. Using these and pre-existing
measurements, a consistent mathematical formalism has been developed to reconcile
the centimeter- and millimeter-wavelength opacity spectra of ammonia. This formal-
ism can be used to estimate the opacity of ammonia in a hydrogen/helium atmosphere
in the centimeter-wavelength range at pressures up to 100 bar and temperatures in the
200 to 500 K range and in the millimeter-wavelength range at pressures up to 3 bar
and temperatures in the 200 to 300 K range. In addition, a preliminary investigation
of the influence of water vapor on the centimeter-wavelength ammonia absorptivity
spectra has been conducted.
This work addresses the areas of high-sensitivity centimeter- and millimeter-
wavelength laboratory measurements, and planetary science, and contributes to the
body of knowledge that provides clues into the origin of our solar system. The labo-
ratory measurements and the model developed as part of this doctoral research work
can be used for interpreting the emission spectra of jovian atmospheres obtained from
xxvi
ground-based and spacecraft-based observations. The results of the high-pressure am-
monia opacity measurements will also be used to support the interpretation of the
microwave radiometer (MWR) measurements on board the NASA Juno spacecraft at
Jupiter.
xxvii
CHAPTER I
INTRODUCTION
If we could understand how the jovian system formed and evolved, we could unlock
vital clues to the beginning and ultimate fate of the entire solar system.
— Morrison and Samz, Voyage to Jupiter, 1980
The jovian planets (Jupiter, Saturn, Uranus, and Neptune) are the most massive
planetary bodies in the solar system, and together they comprise 99.56% of the plan-
etary mass of the solar system. Because of strong gravity fields and relatively low
temperatures, the jovian planets, unlike the terrestial planets, have retained a large
portion of the composition of the primordial cloud from which our solar system was
formed. Hence, understanding their molecular composition and distribution could
reveal much information about the formation of our solar system and provide vital
clues to the formation of similar planetary systems.
For more than half a century, Jupiter has been one of the most extensively stud-
ied astronomical radio and microwave sources because of the fascinating diversity of
its emission characteristics. The discovery of radio emission from Jupiter was made
serendipitously at a frequency of 22.2 MHz by Burke and Franklin (1955). This com-
plex high-intensity emission was non-thermal in nature and was later found to be
synchrotron radiation originating from the relativistic electrons accelerating through
the planet’s magnetosphere. The microwave radiation from Jupiter is dominated by
thermal emission from its atmosphere at frequencies above 3 GHz and synchrotron
emission at frequencies below 3 GHz. The ammonia resonant structure at microwave
frequencies in Jupiter’s atmosphere was first measured by Law and Staelin (1968).
1
Since then, numerous other ground-based and spacecraft-based observations have
been made through the microwave region. Jupiter is also the only giant planet from
which in situ measurements of the atmospheric composition have been made (by the
Galileo entry probe; Niemann et al., 1996). Saturn was first detected at the radio-
wavelengths by Smith and Douglas (1957). Lindal et al. (1985) interpreted Voyager
radio occultation measurements at Saturn and attributed the absorption encountered
at 2.3 GHz and 8.4 GHz to ammonia. To comprehensively study Saturn, the Cassini
mission was conceptualized in the 1980s and launched in 1997. One of the objectives
of the Cassini mission is to study the atmospheric structure of Saturn using radio
occultation experiments conducted at 2.3 GHz, 8.4 GHz, and 32 GHz. Uranus and
Neptune, being the outermost planets in our solar system were the last to be scru-
tinized closely. Voyager 2 made its closest approach to Uranus on January 24, 1986
and Neptune on August 25, 1987, providing a wealth of information on the ice giants.
To interpret the observed emission spectra of the jovian atmospheres, the emis-
sion spectra are compared with appropriate jovian atmospheric models, and the at-
mospheric composition and distribution of various constituents are obtained. The
observable atmospheres of the jovian planets are dominated by hydrogen and he-
lium with small amounts of methane, water vapor, ammonia, hydrogen sulfide, and
other gases. A lack of laboratory measurements of the centimeter- and millimeter-
wavelength (microwave region) properties of various gases has been cited as a major
hindrance for modeling the atmospheres of the jovian planets (de Pater and Mitchell,
1993; de Pater et al., 2005).
The objective of this work has been to improve our understanding of the centimeter-
and millimeter-wavelength absorption spectra of ammonia under jovian conditions
using a combination of laboratory measurements and theoretical formulations (ab-
sorption formalism). The ammonia absorption formalism has been incorporated into
a radiative transfer model of the atmosphere of Jupiter to predict the emission from
2
the planet in the millimeter-wavelength region. Finally, observations of Jupiter have
been made with the Institut de Radioastronomie Millimétrique (IRAM) telescope
facility to search for the 140 GHz ν2 inversion transition of ammonia.
1.1 Motivation
The motivation for this doctoral research work has been to improve our understanding
of the absorption spectra of ammonia in the centimeter- and millimeter-wavelength
region under jovian conditions. An aggressive campaign of laboratory measurements
of the centimeter-wavelength opacity of ammonia in a hydrogen/helium atmosphere
under deep jovian conditions and the millimeter-wavelength opacity of ammonia under
upper tropospheric jovian conditions has been undertaken. Using these and pre-
existing laboratory measurements, a consistent ammonia opacity formalism that can
be used in the 1 mm–30 cm wavelength range in a hydrogen/helium environment
under jovian conditions has been developed. Additionally, laboratory measurements
and modeling of ammonia broadened by water vapor under jovian conditions have
been conducted for the first time to study the influence of water vapor on the ammonia
absorption spectra. Finally, the ammonia opacity model has been used to interpret
the emission spectra of Jupiter at 140 GHz.
Why conduct laboratory measurements?
Absorptivity data for planetary atmospheres obtained from ground-based and spacecraft-
based observations can be used to infer abundances of centimeter- and millimeter-
wavelength absorbing constituents in those atmospheres, as long as reliable informa-
tion regarding the absorbing properties of potential constituents is available. The-
oretical calculations of the absorption spectrum of gaseous molecules can be made
from classical theories using resonant line strengths and frequencies that exist in
spectral line catalogs such as the JPL catalog (Pickett et al., 1998) and the GEISA
catalog (Jacquinet-Husson et al., 2011). However, these theoretical calculations of
3
opacity are limited by the knowledge of the lineshapes and the broadening param-
eters of various molecules with the appropriate broadening agents under the tem-
perature and pressure conditions found in planetary atmospheres. For example, the
centimeter-wavelength laboratory measurements of the opacity of phosphine under
simulated jovian conditions reported by Hoffman et al. (2001) showed that the opac-
ity of gaseous phosphine is an order of magnitude higher than theory had predicted.
Even under terrestrial conditions, it has been known since the measurements near
atmospheric pressure by Becker and Autler (1946) that absorption of microwaves by
water vapor cannot be accounted for by standard lineshape theory. Models for atmo-
spheric water vapor absorption therefore include an empirical “continuum” compo-
nent (Rosenkranz, 1998). Hence, there is a need to conduct laboratory measurements
of the properties of gases under simulated planetary atmospheric conditions over a
range of temperatures and pressures that correspond to the altitudes probed by astro-
nomical observations, and over a range of frequencies that correspond to those used
in astronomical observations. These measurements can be used in the empirical de-
termination of the broadening parameters, temperature coefficients, and lineshapes of
various gases under simulated planetary conditions. These laboratory measurements
can also form the basis for developing accurate mathematical formalisms to model
the centimeter- and millimeter-wavelength properties of gases under those planetary
conditions.
Why use centimeter and millimeter waves?
Centimeter- and millimeter-wavelength astronomy are powerful tools for studying
the jovian atmospheres. Centimeter waves probe the middle and deep atmospheres
of jovian planets (pressures up to hundreds of bar), and hence help infer the interior
composition and dynamics of the planetary atmospheres. Millimeter waves probe
the upper and middle tropospheres of jovian planets (pressures less than a few bar)
4
providing unique insights into the atmospheric composition, chemistry, and dynamics
of those layers in the planetary atmospheres. Furthermore, the millimeter-wavelength
range is often more sensitive than any other spectral range for the detection of minor
species (trace gases).
Why study ammonia?
Gaseous ammonia contributes to strong absorption in the jovian planets in the
centimeter- and millimeter-wavelength range because of the presence of a series of
strong inversion transitions around 1.25 cm, several strong rotational transitions in
the sub-millimeter region, and a strong ν2 inversion transition at 2.15 mm. Hence,
knowledge of the opacity of gaseous ammonia directly impacts the accuracy of inter-
pretation of the observed emission spectra of the jovian atmospheres. Furthermore,
since ammonia is one of the predominant absorbers in the jovian planets, its opacity
must be known before the potential effects of other absorbing constituents can be
assessed. There has been tremendous interest in understanding the absorption prop-
erties of ammonia in the centimeter-wavelength region since they were first measured
in the laboratory by Cleeton and Williams (1934) (see, e.g., Hanley et al., 2009).
Recently, Hanley et al. (2009) made close to 2000 high-accuracy measurements of the
centimeter-wavelength properties of ammonia under simulated jovian atmospheric
conditions (pressures up to 12 bar and temperatures up to 450 K) and developed a
model to estimate the opacity of ammonia in the centimeter-wavelength range. The
millimeter-wavelength absorptivity of pure ammonia was first investigated more than
50 years ago by Nethercot et al. (1952) who measured the absorption of one atmo-
sphere of pure ammonia up to 260 GHz. A few 3.2 mm-wavelength measurements
of ammonia gas properties under simulated jovian conditions were made by Joiner
5
and Steffes (1991) and Mohammed and Steffes (2004). However, these millimeter-
wavelength measurements had large uncertainties because of the coarse instrumen-
tation used at that time and also did not account properly for the adsorption of
ammonia on the surface of the resonator and pressure vessel.
Several ammonia opacity models are currently used to estimate the centimeter-
and/or millimeter-wavelength opacity of gaseous ammonia under jovian conditions (Berge
and Gulkis, 1976; Spilker, 1990; Joiner and Steffes, 1991; Mohammed and Steffes,
2003, 2004; Hanley et al., 2009). The models that were derived based on centimeter-
wavelength measurements (Berge and Gulkis, 1976; Spilker, 1990; Hanley et al., 2009)
cannot be used to accurately estimate the absorptivity in the millimeter-wavelength
range because they do not account for the presence of all the absorption lines in
the millimeter and sub-millimeter region. The models that were derived based on a
limited number of millimeter-wavelength measurements (Joiner and Steffes, 1991; Mo-
hammed and Steffes, 2003, 2004) do not accurately represent the absorptivity of am-
monia over the entire millimeter-wavelength region because of the limited wavelength
ranges measured and the large uncertainties associated with those measurements.
Furthermore, none of the models account for the non-ideal nature of gases under high
pressures, and hence cannot be used to study the deep interiors of the jovian planets.
There were also difficulties in reconciling the centimeter- and millimeter-wavelength
opacity of ammonia (see, e.g., Mohammed and Steffes, 2004). Hence, there has been
a strong impetus to conduct a large number of highly accurate measurements of the
centimeter-wavelength properties of ammonia under simulated deep jovian conditions
and the millimeter-wavelength properties of ammonia under simulated upper tropo-
spheric jovian conditions, and to develop a model to estimate the opacity of ammonia
in a hydrogen/helium atmosphere over a wide range of pressures, temperatures, and
mixing ratios, that is consistent in both the centimeter- and millimeter-wavelength
range.
6
1.2 Science Objectives and Applications
The scientific objective of this research has been to advance the understanding of
the centimeter- and millimeter-wavelength absorption spectra of gaseous ammonia
under jovian conditions. Specifically, the research focused on conducting highly accu-
rate laboratory measurements of the 2–4 mm-wavelength opacity of ammonia under
simulated jovian conditions at pressures up to three bar and temperatures between
200 and 300 K, and the 5–20 cm-wavelength opacity of ammonia under simulated
deep jovian conditions at pressures up to 100 bar and temperatures up to 500 K.
These new measurements and close to 1500 centimeter-wavelength measurements of
the opacity of ammonia conducted by Hanley et al. (2009) were used to develop a
consistent ammonia opacity model to provide accurate interpretation of ground-based
and spacecraft-based observations of jovian planetary atmospheres in the 1 mm–30 cm
wavelength range. The results of the high-pressure centimeter-wavelength measure-
ments will also be used to support the interpretation of the microwave radiometer
(MWR) measurements on board the NASA New Frontiers Juno spacecraft that is
scheduled to arrive at Jupiter in 2016.
Millimeter-wavelength applications
Millimeter-wavelength astronomy is a powerful tool for studying the temperature
structure, composition, and dynamics of jovian planetary atmospheres. To date,
ground-based millimeter-wavelength observations have been used for disk-averaged
emission measurements of the jovian planets (Ulich, 1974; Griffin et al., 1986; Muhle-
man and Berge, 1991; Griffin and Orton, 1993; Kramer et al., 2008), interferometric
mapping of Saturn (Dowling et al., 1987; van der Tak et al., 1999; Dunn et al.,
2005), and interferometric observations of limb darkening of Jupiter (Valdes et al.,
1982). At millimeter-wavelengths, the planets Uranus and Neptune, with small ap-
parent diameters and large flux densities, are frequently used as primary calibrators
7
of astronomical sources and telescope parameters (Ulich, 1981; Kramer et al., 2008).
Future millimeter-wavelength observations with the Atacama Large Millimeter Array
(ALMA) will represent a major step forward in the study of planetary atmospheres
because of its high angular resolution, fast imaging capabilities, and wide instanta-
neous bandwidth. Combined millimeter-wavelength observations using the ALMA in
conjunction with space telescopes such as the Planck space telescope (Lamarre et al.,
2003) will greatly enhance studies of planetary atmospheres. The ammonia opacity
model developed as part of this work can be used for interpreting the millimeter-
wavelength emission spectra of jovian atmospheres observed by such facilities.
Centimeter-wavelength applications
The ammonia opacity model developed as part of this work will be used for inter-
preting the MWR measurements made by the Juno spacecraft at Jupiter, in addi-
tion to its application to the interpretation of past and future centimeter-wavelength
ground-based and spacecraft-based observations, radio occultation experiments, and
entry-probe observations of the jovian planets. The NASA mission Juno is a robotic
spacecraft scheduled to arrive at Jupiter in July 2016. Using a spinning, solar-powered
spacecraft, Juno will make maps of the gravity, magnetic fields, and atmospheric com-
position of Jupiter from a unique polar orbit. During its one-year mission, Juno will
complete 33 eleven-day-long orbits and will sample Jupiter’s full range of latitudes
and longitudes. The primary goal of the Juno MWR is to probe the deep atmosphere
of Jupiter at radio wavelengths ranging from 1.3–50 cm using six separate radiometers
to measure the planet’s thermal emissions and determine the atmospheric composition
beneath the cloud layers, down to hundreds of bar of pressure. These thermal emission
measurements with the aid of a radiative transfer model can help infer planet-wide
concentration and distribution of water vapor and ammonia, provided the radiative
transfer model has accurate absorption models for ammonia and water vapor that are
8
valid at the pressures and temperatures probed by the six MWR channels.
1.3 Organization
This dissertation addresses five main areas: theoretical discussion of microwave spec-
troscopy, laboratory measurements of the centimeter- and millimeter-wavelength ab-
sorption spectra of ammonia under simulated jovian conditions, discussion of the com-
pressibility of fluids under deep jovian conditions, development of a consistent model
to represent the absorption spectra of ammonia under jovian conditions, and an ap-
plication of this model for the interpretation of ground-based millimeter-wavelength
observations of the emission spectra of Jupiter. Each of these topics is discussed
within its own chapter.
Chapter 2 provides a brief description of microwave spectroscopy and presents
the fundamental theory of the absorption spectra of ammonia. Chapter 3 provides
a discussion of approaches for measuring absorption and refraction of gases in the
centimeter- and millimeter-wavelength region and a description of the millimeter-
wavelength system and the high-pressure centimeter-wavelength system used for per-
forming measurements of the properties of gases under simulated jovian conditions.
This chapter also provides a discussion of the laboratory measurement procedure, and
methods for processing the raw data and the uncertainties involved.
Chapter 4 introduces the concept of non-ideal gases (real gases) and provides
a thermodynamic framework for handling pure fluids under high-pressure conditions
that are characteristic of the deep jovian atmospheres. Chapter 5 provides a summary
of the centimeter- and millimeter-wavelength laboratory measurements, and the con-
sistent ammonia absorption formalism developed based on laboratory measurements
and theoretical models. The performance of the new model is compared with the pre-
vious models in the 1 mm–30 cm wavelength range. Additionally, a brief discussion
9
of the preliminary investigation (laboratory measurements and model) of the influ-
ence of water vapor on the ammonia absorption spectra in the centimeter-wavelength
range is provided.
Chapter 6 shows an application of the new ammonia opacity model for the inter-
pretation of ground-based observations of the 2.1 mm-wavelength emission spectra of
Jupiter. Chapter 7 concludes this dissertation with a summary of main contribution
and explores the future directions of research.
10
CHAPTER II
MICROWAVE SPECTROSCOPY AND AMMONIA
SPECTRA
Electromagnetic radiation incident on a gaseous molecule produces absorption or
emission of energy at a particular wavelength. Radiation incident on a molecule is
absorbed when the molecule transitions from a lower energy state to a higher energy
state, emission is the reverse process. The frequency associated with this energy
change is given by
∆E
f= , (2.1)
h
where ∆E is the change in energy between the upper and lower states, h = 6.624 ×
10−34 Jsec (Planck’s constant), and f is the frequency of absorption or emission. An
isolated molecule’s internal energies consist primarily of electronic, vibrational, and
rotational, typically associated with absorption or emission in the visible, infrared,
and microwave regions, respectively. Electronic transitions occur when electrons in
a molecule are excited from one energy level to a higher energy level. The energy
associated with the electronic transition is very large and correspond to the visible
region of the electromagnetic spectrum. Vibrational transitions occur when the atoms
in a molecule are in periodic motion while the molecule as a whole has a constant
rotational motion. A linear molecule with N atoms has 3N-5 normal modes of vibra-
tion since the rotation about the molecular axis cannot be observed, and a non-linear
molecule with N atoms has 3N-6 normal modes of vibration. The energy associated
with vibrational motion is considerably larger than that associated with rotational
motion, and hence the frequency associated with these transitions is in the infrared re-
gion of the spectrum. The energy studied in the microwave region results mostly from
11
rotational motion with the exception of molecular inversion, which is a form of vibra-
tional motion. Rotational transitions occur when a molecule that possesses either a
magnetic or electric dipole moment rotates about its center of mass. Polar molecules
have a structure that naturally possesses a permanent dipole moment creating an
asymmetry in the charge distribution which occurs as a result of the covalent bond-
ing. Polar molecules interact with electromagnetic radiation and are active absorbers
in the microwave region. A large portion of the centimeter- and millimeter-wavelength
absorption in the jovian planets stems from the presence of polar molecules such as
ammonia, hydrogen sulfide, water vapor, and phosphine. Non-polar molecules do
not possess permanent dipole moment and are not active absorbers in the microwave
region, but they do exhibit pressure-induced absorption. This absorption occurs be-
cause of the formation of transient collisionally-induced dipoles. Since the bulk of
jovian atmospheres consist of hydrogen, helium, and methane, it is important to
consider the collisionally-induced dipoles of H2 -H2 , H2 -He, and H2 -CH4 .
2.1 Absorption Spectra of Ammonia
Ammonia is a symmetric top molecule with a trigonal pyramidal shape and a bond
angle of 107.8◦ (Figure 2.1). The central nitrogen atom has five outer electrons and
each of the hydrogen atom has an electron, resulting in a total of eight electrons or
four electron pairs, which are arranged tetrahedrally. Three of these electron pairs are
used as nitrogen-hydrogen bond pairs, which leaves one lone pair of electrons. The
lone pair of electrons repel more strongly than bond pairs, therefore the bond angle is
107.8◦ and not 109.5◦ as expected for a regular tetrahedral arrangement. This shape
gives the ammonia molecule an electric dipole moment and makes it polar.
Ammonia has been a prime candidate to test theoretical and experimental mi-
crowave spectroscopy techniques because of the presence of a large number of easily
observable lines (Townes and Schawlow, 1955). Ammonia also provides a rich and
12
Figure 2.1: Ammonia molecule.
intense spectra involving hindered motion arising from quantum mechanical tunnel-
ing effect. The central nitrogen atom cannot be allowed in the plane of the hydrogen
atoms because of the large potential-energy hump at this position. However, the ni-
trogen atom can “tunnel” through the plane of the hydrogen atoms and vibrate from
one side to the other. These hindered motions are in principal vibrational transitions,
but are called inversion transitions. Although the vibrational transitions absorb or
emit infrared frequencies, the inversion transitions occur in the microwave region
because they are slowed down by the hindering potential.
The potential energy curve between the nitrogen atom and the plane of the hy-
drogen atoms shows two minima that correspond to the equilibrium position of the
nitrogen atom on either side of the plane of the hydrogen atoms (see, e.g., Townes
and Schawlow, 1955). The nitrogen atom may vibrate rapidly with respect to the
plane of the hydrogen atoms in one of the potential minima resulting in vibrational
transitions in the infrared region, and also may penetrate the potential barrier re-
sulting in inversion transitions in the microwave region and begin vibrating in the
other potential minima. The height of the potential hill above the minimum is ∼
2077 cm−1 and the ground-state inversion transitions occur in the microwave region
(∼ 23.8 GHz). Comparison of the ammonia potential hill with other symmetric hy-
drides such as phosphine and arsine can be made. The height of the potential hill
13
above the minimum for phosphine is ∼ 6085 cm−1 resulting in ground-state inversion
transitions in the radio wavelength range (∼ 0.14 MHz) and for arsine is ∼ 11220
cm−1 resulting in a ground-state inversion frequency of ∼ 1/2 cycle/year. Thus, the
ground-state inversion transitions of ammonia are rapid and provide a large number
of easily observable lines, whereas arsine takes two years to go through a cycle of
inversion and does not provide any observable transitions.
The interaction between the rotational and vibrational motion of ammonia results
in a series of closely spaced lines in the rotational spectrum, each corresponding to dif-
ferent vibrational states. The interaction between inversion and rotational transitions
of ammonia results in a series of lines in the inversion spectrum, each corresponding
to a different rotational state.
Due to the symmetric nature of the ammonia molecule, the quantum numbers J
and K are used to describe the rotation of the molecule. J represents the total angular
momentum vector of the ammonia molecule (rotational quantum number) and K is
the projection of J onto the molecular axis. Since K represents a component of the
rotational quantum number J, it can never be larger than J, and due to the symmetry
of the coordinate system negative values are not used. It is not possible to have to
have a molecule with zero angular momentum because of the uncertainty principle,
and hence J cannot be equal to zero. However, K can be equal to zero in pure
rotational states, but this does not correspond to inversion. Hence, while considering
inversion transitions, the quantum numbers are counted up from one, with J ≥ K
and the inversion transitions are represented as (J, K). Inversion splitting of the
rotational spectrum corresponds to different spins of the nitrogen nucleus that are
recognized through molecular inversion, giving rise to a third quantum number S
that can have a value of either 1 or 0. Inversion transition (1,1) occurs at 23.67
GHz, but the strongest inversion transition is the (3,3) transition that occurs at
23.87 GHz (Poynter and Kakar, 1975). A more complete description of the ammonia
14
S=1
J=2
K=1
S=0
S=1
J=1
K=1
S=0
≈ ≈
νQ
νR
S=1 J=1
S=0 K=1
Figure 2.2: The 140.14 GHz ν2 = 1 inversion transition occurs when the ammonia
molecule transitions between the lower inversion level of the ground-state J = 2, K
= 1 rotational transition and the upper inversion level of the ground-state J = 1, K
= 1 rotational transition.
spectrum and hindered motion can be found in Townes and Schawlow (1955).
In addition to the ground-state transitions (ν2 = 0), ammonia can have transitions
that occur at higher-states (ν2 = 1, 2, 3, and so on). The higher state transitions (also
called ν2 transitions) that influence the centimeter- and millimeter-wavelength opacity
of ammonia are similar to the ground state transitions, except that the transition
occurs from a higher-energy state. For example, the strongest millimeter-wavelength
ν2 = 1 inversion transition that occurs at 140.14 GHz is a transition between the lower
inversion level of the ground-state J = 2, K = 1 rotational transition and the upper
inversion level of the ground-state J = 1, K = 1 rotational transition (Figure 2.2 ).
15
Microwave and sub-millimeter spectral line catalogs such as the JPL catalog (Pick-
ett et al., 1998) provide a database of the line strengths and frequencies of transitions
of various molecules in the centimeter, millimeter and sub-millimeter region. The
JPL spectral line catalogs for ground-state NH3 transitions and higher-energy state
NH3 − ν2 transitions were recently updated (Yu et al., 2010a,b,c), and the line tran-
sitions in the 0.3–1300 GHz frequency range and their intensities are shown in Fig-
ure 2.3. The NH3 catalog has a total of 1716 ground-state transitions that include 415
inversion transitions in the 0.3–220 GHz frequency range and 1301 rotational tran-
sitions in the 0.3–20 THz frequency range. The higher-state NH3 − ν2 catalog has a
total of 4198 inversion-rotational transitions in the 0.02–46 THz frequency range. To
effectively model the centimeter- and millimeter-wavelength opacity of ammonia, it
is critical to consider the contribution from all the line transitions that influence its
opacity in that spectral region.
16
Inversion lines
−20
10 ν inversion−rotational lines
2
Rotational lines
Intensity (cm−1/molecule/cm2)
−25
10
−30
10
−35
10
0 1 2 3
10 10 10 10
Frequency (GHz)
Figure 2.3: Line positions and intensities of the NH3 and NH3 − ν2 transitions from
0.3–1300 GHz (Yu et al., 2010a,b,c).
17
2.2 Linewidths
Transitions of an isolated, undisturbed, and stationary molecule occur at very specific
energies and frequencies since these transitions are quantized. However, various types
of disturbances change the energy levels, giving a width to the spectral lines and
varying the center frequency, thereby broadening the line. The full width at half
maximum (FWHM) is defined as the width that causes transition in at least half of
the molecules of a particular species. Spectral linewidth, nominally called the half
width at half maximum (HWHM), is half of the FWHM.
The sources of line-broadening include natural line broadening, Doppler broaden-
ing, saturation broadening, molecular collisions against walls, and pressure broaden-
ing. Zero-point ambient electromagnetic energy, present in free space, disturbs the
molecules and results in natural line broadening or radiation broadening (Townes
and Schawlow, 1955). This width is negligible when compared to the other types
of broadening in the microwave region. Molecular motion relative to the direction
of propagation of the electromagnetic radiation results in a frequency shift due to
the Doppler effect. The linewidth that results from Doppler broadening is given
as (Townes and Schawlow, 1955)
r
−7 T
∆νDoppler = 3.581 × 10 ν , (2.2)
M
where ν is the center frequency of the line, T is the temperature of the molecules
in a gas, and M is the molecular weight. For the ammonia ν2 =1 transition at 140.14
GHz, the Doppler linewidth at room temperature is ∼ 210 kHz. The Doppler broad-
ening is significant for very low density gas mixtures. However, under the jovian
temperature and pressure conditions studied in this work, the Doppler effect is neg-
ligible. Saturation broadening results when the intensity of the microwave radiation
is so large that absorbing molecules cannot get rid of the absorbed energy rapidly
18
enough and the Beer-Lambert’s law breaks down. While gases under low-pressures
show significant saturation effect (see, e.g., Townes and Schawlow, 1955), it is negligi-
ble for the pressure conditions studied in this work. Additionally, collisions with the
walls of a test chamber in a laboratory setting can cause broadening. As an example,
at room temperature, gaseous ammonia in a parallel plate waveguide cell with a spac-
ing of 4 mm between the plates has a wall collision linewidth of 12 kHz (Townes and
Schawlow, 1955). Although the line broadening due to wall collisions could be signif-
icant for small waveguide cells (Danos and Geschwind, 1953), this effect is negligible
for large test chambers such as those used in this work.
Under the pressure and temperature conditions used in this study, the greatest
source of line broadening arises from pressure broadening due to molecular collisions.
These collisions allow for the transfer of kinetic energy and interactions between the
molecules due to van der Waals force. Almost all collision theories assume bimolecular
collisions and most collision theories also assume that these collisions are instanta-
neous, i.e., the time between collisions is very large compared to the duration of a
collision. While these assumptions are adequate for low pressure conditions, they are
invalid under very high pressure conditions. Each molecular species has a broadening
cross-section (collision diameter) that depends on its size and dipole moment. These
broadening cross-sections are critical in characterizing the microwave absorption by
various molecular species. When a molecule collides with another molecule of the
same species, the resulting broadening is called self-broadening. If the collision is
between two molecules of different species, the resulting broadening is called foreign-
gas-broadening. The foreign-gas-broadening occurs even if the colliding molecules do
not have any microwave absorption lines themselves, as in the case of hydrogen and
helium.
19
2.3 Lineshapes
The lineshape of molecular collisions is used to describe the spectral dependence of
broadening. Lorentz (1906) was the first to model the pressure broadening of gases.
Debye (1929) described the absorption and refraction in polar molecules with a the-
ory that differed from that of Lorentz at the zero resonant frequency. Van Vleck
and Weisskopf (1945) combined the two theories to derive the Van Vleck–Weisskopf
lineshape given as
2
1 ν ∆νj ∆νj
FV V W (ν, ν(0,j) , ∆νj ) = + ,
π ν(0,j) (ν(0,j) − ν)2 + ∆νj2 (ν(0,j) + ν)2 + ∆νj2
(2.3)
where for the line j, ∆νj is the half width at half-maximum, ν(0,j) is the center
frequency of the line transition, and ν is the frequency of the incident electromagnetic
wave. Gross (1955) assumed a Maxwellian distribution of molecular velocities, instead
of the Boltzmann one used by Lorentz and Van Vleck and Weisskopf and derived the
Gross or the Kinetic lineshape given by
" #
1 ν 4νν(0,j) ∆νj
FG (ν, ν(0,j) ) = 2
. (2.4)
π ν(0,j) (ν(0,j) − ν 2 )2 + 4ν 2 ∆νj2
Although the Van Vleck–Weisskopf and Gross lineshapes converge at the line cen-
ter, the Gross lineshape has higher skirts away from the linecenter than the Van
Vleck–Weisskopf lineshape. Ben-Reuven (1966) derived a lineshape with two addi-
tional parameters, a line shift term (δ) proportional to the gas density, and a line-to-
line coupling element (ζ). The Ben–Reuven lineshape is given by
2
(γj − ζj )ν 2 + (γj + ζj )[(ν(0,j) + δj )2 + γj2 − ζj2 ]

2 ν
FBR (ν, ν(0,j) , γj , ζj , δj ) = ,
π ν(0,j) [ν 2 − (ν(0,j) + δj )2 − γj2 + ζj2 ]2 + 4ν 2 γj2
(2.5)
20
where for the line j, γj = ∆νj is the linewidth. The Gross lineshape is a special
case of the Ben–Reuven lineshape under the assumption that only sense-reversing
collisions take place, in which case γj = ζj and δ = 0 (Waters, 1976).
2.4 Line Intensity
The absorption from a collisionally broadened gas is a function of the absorption at
the line center, the linewidth, and the lineshape function. The absorption at each line
center is calculated using the line intensity information from the latest JPL catalog
as per Pickett et al. (1998)
nIj (T )
Aj = , (2.6)
π∆νj
where for the line j, n is the number density of the gas, I j (T ) is the intensity of
the line at temperature T , and ∆νj is the linewidth. The line intensity is a measure
of the amount of energy associated with any particular molecular transition. The
line intensity at the measurement temperature (T ) is calculated as per Pickett et al.
(1998), using the values of the line intensity at the reference temperature (T0 ), lower
state energy of the transition, and temperature dependence parameter. The tempera-
ture dependence parameter for a diatomic or linear molecule is 1 and for a symmetric
rotor is 3/2. The line intensities (at the reference temperature), lower state energies,
and transition frequencies used in this study are provided by the JPL spectral line
catalog (Pickett et al., 1998). A detailed description on calculating the absorption of
a collisionally broadened gas is provided in Section 5.4.
21
CHAPTER III
LABORATORY MEASUREMENTS OF AMMONIA
One of the outstanding problems in the millimeter-wavelength spectroscopy of planets
is the lack of adequate information on lineshapes and linewidths of gases at relevant
pressures and temperatures, and with appropriate broadening agents. On the other
hand, at centimeter-wavelengths, although there is adequate information on line-
shapes and linewidths of gases, there is a lack of understanding of the influence of
the far-wings of the rotational lines on the continuum absorption spectrum and a
very poor understanding of the effects of compressibility of gases under the high-
pressure conditions typical of the deep jovian atmospheres. These problems may
be addressed by developing accurate opacity models grounded in accurate labora-
tory measurements. This chapter describes the theory, laboratory apparatus, mea-
surement procedure, and uncertainties in the measurements of the centimeter- and
millimeter-wavelength absorptivity of gaseous ammonia under simulated jovian at-
mospheric conditions.
3.1 Electromagnetic Propagation in a Homogeneous, Isotropic

Medium
The phasor expressions for the electric and magnetic fields of an electromagnetic wave
propagating in a homogeneous, isotropic, and lossy medium in the +x direction are
E(x) = E0 e−jkx = E0 e−jαx e−jβx , (3.1)
H(x) = H0 e−jkx = H0 e−jαx e−jβx , (3.2)
22
respectively, where E 0 and H0 are the amplitudes of the electric and magnetic
fields, respectively, k is the wavenumber, and α is the attenuation constant, and β is
the phase constant. The wavenumber k is given as
√
k = ω µε, (3.3)
where ω is the radial frequency, µ is the permeability, and ε is the permittivity.
For most gases under jovian conditions, µ may be assumed to be entirely real and
equal to µ0 , the permeability of free space. However, for lossy gases, the permittivity
is complex and is given as
ε = ε0 − jε00 , (3.4)
so that the wavenumber of the electromagnetic wave equations becomes
p
k=ω µ(ε0 − jε00 ). (3.5)
The frequency dependence of the real and imaginary parts of the permittivity are
not independent and are related by the Kramer-Kronig relation as (see, e.g., Ramo
et al., 1994)
∞
ω 0 ε00 (ω 0 )dω 0
Z
0 2
ε (ω) = ε0 + , (3.6)
π 0 (ω 02 − ω 2 )
∞
[ε0 (ω 0 ) − ε0 ]dω 0
Z
00 2ω
ε (ω) = − , (3.7)
π 0 (ω 02 − ω 2 )
where ω and ω 0 are the radial frequencies. A fully accurate characterization of
ε0 from ε00 requires knowledge of the value of ε00 over an infinite frequency range,
23
and vice versa. Practical calculations of the unknown quantity using the Kramer-
Kronig relation use all available information about the known quantity, while using
appropriate interpolations to fill any gaps, and assumptions at the extreme frequencies
where information is not available.
The wavenumber can be separated into the attenuation and phase constants α
and β (see, e.g., Ramo et al., 1994)
s
ε00

jk = α + jβ = jω µε0 1 − j , (3.8)
ε0
where
v
u s 00 2

u 0
u µε  ε
α = ωt 1+ − 1, (3.9)
2 ε0
and
v
u s 00 2

u 0
u µε  ε
β = ωt 1+ + 1. (3.10)
2 ε0
The frequency dependence can be removed by taking the ratio of α and β as
vq
ε00 2
u
α u 1+
u
ε0
−1
= tq . (3.11)
β ε00 2

1+ ε 0 +1
ε00
The term ε0
is the loss tangent (tan δ) of the medium. The inverse of loss tangent
ε0
is the quality factor of the medium Q = ε00
. For a gaseous medium that is not very
lossy,
ε00 1
0
= tan δ = << 1, (3.12)
ε Qgas
24
where Qgas is the quality factor of the gaseous medium. Hence, for a low-loss gas,
the terms in Equation 3.11 can be expanded using Taylor series and approximated as
α ε00
≈ 0. (3.13)
β 2ε
α
This approximation estimates β
within 0.5% when the loss tangent is less than
0.01 (or absorptivity less than 104 dB/km) and the loss tangents of gaseous media
are almost always less than 0.01 (Spilker, 1990).
3.2 Measurement Theory
The reduction in the quality factor (Q) of a resonant mode of a resonator in the
presence of a low-loss gas is used to measure the absorption of the gas (see, e.g.,
Hanley and Steffes, 2007). The quality factor of a resonance is given as (Matthaei
et al., 1980)
2πf0 × Energy Stored

Q= , (3.14)
Average Power Loss
where f0 is the resonant frequency. The Q can be computed as the resonant
frequency divided by its half-power bandwidth (HPBW).
f0
Q= . (3.15)
HPBW
The Q of the lossy gas and its opacity are related by
ε00 π 1 π
α≈ = , (3.16)
ε0 λ Qgas λ
where ε0 and ε00 are the real and imaginary permittivity of the gas, λ is the
wavelength in km, and α is the absorptivity of the gas in Nepers/km (1 Neper =
8.686 dB). The quality factor of a resonator filled with the lossy gas is given by
25
1 1 1 1 1
= + + + , (3.17)
Qm
loaded Qgas Qr Qext1 Qext2
where Qm
loaded is the measured quality factor of the loaded resonator, Qgas is the
quality factor of the gas under test, Qr is the quality factor of the evacuated res-
onator, less coupling losses, and Qext1 and Qext2 are the external coupling losses in
the resonator. For symmetric resonators such as the ones used in this work, we can
assume Qext1 = Qext2 . The coupling losses can be calculated by measuring the trans-
missivity t= 10−S/10 , where S is the insertion loss of the resonator in decibels (dB)
at the frequency of a particular resonance, and using the relation (Matthaei et al.,
1980)
2
Qm

t= 2 , (3.18)
Qext
2Qm
Qext = √ . (3.19)
t
The value of Qr is related to the measured Q at vacuum (Qm

vac ) by
1 1 1 1
= + + . (3.20)
Qm
vac Qr Qext1 Qext2
Substituting Equation 3.19 in equations 3.17 and 3.20, we get
√ √
1 1 − tloaded 1 − tvac
= − , (3.21)
Qgas Qm
loaded Qm
vac
where tloaded and tvac are the transmissivities of the resonances in the loaded and
vacuum conditions, respectively. The addition of the test gas causes a shift in the
center frequency of the resonances corresponding to the refractive index of the test gas.
26
There is a change in the quality factor of the resonances when the center frequency
changes owing to the changes in coupling to the resonator. This effect is called
dielectric loading (DeBoer and Steffes, 1994) and can be removed by performing
additional measurement of the quality factor of the resonances with a lossless gas
present and shifting the center frequency of the resonances by the exact same amount
as with the lossy gas. These matched measurements are used in place of the vacuum
measurements in Equation 3.21, and converting the units from Nepers/km to dB/km
(1 Neper/km=8.686 dB/km), the absorptivity is given as
√ √
π 1 − tloaded 1 − tmatched
α = 8.686 − (dB/km). (3.22)
λ Qm
loaded Qm
matched
The dielectric loading of resonances gives information about the refractive index of
a gas. The refractive index (nri ) of gases, albeit very close to one, should be known to
a very high accuracy since small changes in the refractive index can significantly alter
the propagation of electromagnetic radiation through an atmosphere. Refractivity
(N ) of a gas mixture is defined as refractive index less one multiplied by 106
N = 106 (nri − 1). (3.23)
Refractivity is measured by the change of the center frequency of the resonances
compared to their vacuum values and is given as
fvac − fgas
N = 106 × , (3.24)
fgas
where fvac and fgas are the center frequencies of the resonances measured with the
system under vacuum and with the test gases, respectively (Tyler and Howard, 1969).
For the pure ammonia experiments, the refractivity of ammonia can be directly cal-
culated using the above equation. However, for ammonia mixture experiments, the
27
measured refractivity represents the total refractivity of the test mixture which is
the sum of the individual constituents’ refractivities weighted by their mole frac-
tion. Refractivity depends on pressure (P ) and temperature (T ), and the normalized
refractivity is calculated as
N R0 T
N0 = , (3.25)
P
where R0 is the universal gas constant.
3.3 Millimeter-Wavelength Measurement System
A new high-sensitivity millimeter-wavelength measurement system developed as part
of this work has been used to measure the 2–4 millimeter-wavelength properties of
gases under simulated planetary conditions. The measurement system consists of a
planetary atmospheric simulator, millimeter-wavelength subsystems: W band (3–4
mm) and F band (2–3 mm), and a data handling subsystem. For the study described
in this dissertation, the W band and F band systems have been used for the measure-
ments of the opacity of pure ammonia and ammonia/hydrogen/helium mixtures un-
der simulated jovian conditions. Additional applications of the millimeter-wavelength
system for the study of hydrogen sulfide under jovian conditions, and sulfur dioxide
and sulfuric acid vapor under Venus conditions are discussed by Devaraj and Steffes
(2011).
3.3.1 Planetary Atmospheric Simulator
The planetary atmospheric simulator controls and monitors the environment experi-
enced by the measurement system, including the pressure and temperature conditions
of the gas under test. The simulator consists of a pressure vessel, temperature cham-
ber, gas-handling subsystem, and various measurement gauges. The main component
of the atmospheric simulator is a pressure vessel capable of withstanding pressures
28
Figure 3.1: A picture of the glass pressure vessel enclosing the Fabry–Perot resonator
used for simulating the jovian atmospheric conditions.
up to three bar. This pressure vessel is a Pegasus

R
conical glass pipe (6x9 inches)
with aluminum endplates. The endplates are sealed with Viton

R
O-rings sandwiched
between the lip of the glass and the flat endplate which is bolted to an inner flange.
The pressure vessel encloses the Fabry–Perot resonator used for making measurements
and is shown in Figure 3.1.
The pressure vessel and a part of the millimeter-wavelength measurement sys-
tem are contained in a temperature-controlled chamber which is a Revco

R
ultra-low
temperature freezer capable of operating between 173 K and 218 K. The temperature
chamber has a stability of approximately ±0.5 K. The freezer does not operate warmer
than 218 K, and hence measurements cannot be conducted in the 218–290 K range.
Room temperature measurements are also conducted with the system placed inside
the temperature chamber (with the cooling system shut-off) since the chamber helps
to maintain thermal stability. When the system is operating at cold temperatures,
the thermal stability is increased by placing iron blocks inside the freezer so that the
temperature fluctuations inside the pressure vessel do not exceed ±0.5 ◦ C as verified
by in situ temperature measurements. The temperature inside the pressure vessel is
monitored by an Omega
R
Model HH21 Microprocessor Thermometer (display unit)
29
connected to a Type T (copper/copper-nickel) thermocouple probe threaded into a
1/4” NPT fitting on one side of the aluminum endplate of the resonator. This allows
the probe to be in direct contact with the test gases and provides a very accurate
measurement of the temperature inside the vessel. The thermometer display unit has
a resolution of 0.1 ◦ C and a 3σ accuracy of 0.3% of the displayed value plus an ad-
ditional ±0.6 ◦ C uncertainty at temperatures below 0 ◦ C. Above 0 ◦ C, the additional
uncertainty remains the same, but the accuracy improves to 0.1%. The thermocouple
itself has an accuracy of 1.0 ◦ C or 0.75% of the reading, whichever is greater.
Pressure measurements in the 2–3 bar range are made with an Omega
R
DPG7000-
30 Digital Test Gauge, and pressure measurements in the 0–2 bar range (vacuum to 2
bar) are made with an Omega

R
DPG7000-VAC/15 Digital Test Gauge. These gauges
have a resolution of 1 mbar and 0.05% full scale total error band accuracy. The gauges
only display differential pressure between the closed system and the ambient labora-
tory environment, and hence the gauge readouts can vary as the ambient pressure
varies even if the absolute pressure in the system remains constant. To correct for
this effect, digital barometers (Davis

R
Weather Monitor II 7440 and Young
R
61202L)
are used to measure the ambient pressure inside the laboratory.
A Welch DuoSeal
R
vacuum pump Model 1376B-01 is used to evacuate the gases
from the system from 1 bar down to vacuum. Gases at higher pressures are vented
through an exhaust valve. A glass cylinder with a volume of approximately 7.2 L is
connected to the venting line along with an Analytical Technology, Inc. PortaSens
II portable gas leak detector so as to detect any trace amounts of gas through a
series of interchangeable electrochemical cartridge sensors. Gases are fed into the
pressure vessel through a series of regulators and valves. All components in the gas
handling system are connected with seamless 3/8” outer diameter stainless tubing
and Swagelok
R
fittings.
30
3.3.2 Millimeter-Wavelength Subsystem
Currently, the 2–4 mm-wavelength measurements are made with two different subsys-
tems, namely the W band and F band systems. At the heart of both the measurement
systems is a spherical mirror Fabry–Perot resonator (FPR) in a near confocal config-
uration enclosed in the pressure vessel.
Fabry–Perot resonator
Resonant microwave cylindrical cavities have been used to observe molecular reso-
nances in gases and their absorption coefficients for over 60 years (Bleaney and Pen-
rose, 1947; Weidner, 1947; Gordy, 1948). In the millimeter-wavelength region, FPRs
provide low loss, high coupling efficiency, and high precision in measurements (Cul-
shaw, 1960, 1961, 1962; Zimmerer, 1963). FPRs have been successfully employed for
the measurement of atmospheric gas losses in the millimeter-wavelength range for
more than 40 years (Valkenburg and Derr, 1966). Absorptivity of gases under simu-
lated planetary conditions have also been measured using an FPR in the past (Joiner
and Steffes, 1991; Fahd and Steffes, 1992).
The FPR used in the millimeter-wavelength system consists of two concave gold-
plated mirrors whose surface is polished to µm tolerance (Figure 3.2). Electromag-
netic energy is coupled to and from the resonator (which acts as a band-pass filter)
through irises located in the center of each of the mirrors via WR-8 waveguides which
pass through the endplates to the exterior of the pressure vessel. The end of each
waveguide section is pressure-sealed by a circular piece of mica window held in place
by a low temperature O-ring and vacuum grease. The resonator is symmetrical and
the input/output ports are interchangeable. One of the ports is connected to the
signal source through a waveguide section and the other end is connected to a high-
resolution spectrum analyzer through a harmonic mixer and diplexer.
The spherical mirror FPR is used in a near-confocal configuration which has a
31
T-type thermocouple ~23 cm
WR-8
waveguide
D = 15.2 cm
Tuner
Concave Gas Inlet

mirror
Aluminum endplates
Figure 3.2: Block diagram of the spherical mirror Fabry–Perot resonator placed in
a near-confocal configuration.
number of advantages such as high quality factor and very good tolerance to the
alignment of mirrors (Herriott et al., 1964). The radius of curvature of the spherical
mirrors is approximately 30 cm and the distance of separation (D) between the mirrors
is adjustable thereby making it possible to measure resonances (transmission peaks)
at desired frequencies. For the measurements that were made with this resonator,
a mirror spacing of approximately 15.2 cm provided minimum diffraction losses and
optimal free spectral range (FSR). FSR is the frequency interval between adjacent
axial mode resonances and is given by
c
FSR = , (3.26)
2D
where c is the speed of light. For the current resonator set-up, FSR ∼ 1GHz.
There are three kinds of losses in an FPR: resistive losses (on the surface of the
mirrors), coupling losses (due to the energy coupling in/out of the resonator through
the irises), and diffraction losses (around the sides of the mirrors) (Culshaw, 1960,
1962). When a spherical mirror interferometer is illuminated by an off-axis ray of
light, the repeated reflections cause several ray paths and these ray paths give rise
to additional resonances (Herriott et al., 1964). These additional off-axis resonances
32
adversely affect the performance of the interferometer. Furthermore, diffraction loss
is caused by energy spilling over the sides of the mirrors because of poor mirror align-
ment. To suppress the off-axis resonances and to reduce diffraction losses, the mirrors
need to be very precisely aligned. This was achieved using a beam from a helium-neon
laser, which was directed to the input waveguide of the resonator. The other mirror
was adjusted so that the reflected beam focused precisely on the output iris. The
coupling losses are a function of frequency due to the standing waves between the
input/output irises and the signal source/detector. The coupling losses were reduced
by ensuring a good waveguide to mirror joint. The resistive losses were minimized
by using highly reflective gold-plated mirror surfaces with minimal surface irregulari-
ties. However, the mirror resistive losses increase at higher frequencies because of the
shallow skin depth of the gold-plated mirrors. The coupling losses also increase at
higher frequencies. Neglecting diffraction losses, the quality factor of the resonances
are dependent on the mirror resistive and coupling losses. For the resonator used
in this work, the mirror resistive and coupling losses increase faster as a function
of frequency compared to the theoretical increase in Q as a function of frequency
and become dominant at higher frequencies. Hence, the quality factor of the system
decreases as a function of frequency (Figure 3.3).
In order to achieve a high system sensitivity (which corresponds to a high Q), all
losses in an FPR should be minimized. After the alignments were made, the quality
factor of the resonator at vacuum and T=297 K in the 2–4 mm-wavelength range
was between 45,000 and 190,000 (Figure 3.3). The effective path length (EPL) of the
electromagnetic energy is given as (Valkenburg and Derr, 1966)
Qλ
EPL = . (3.27)
2π
For example, consider a resonance at 77.08 GHz (3.89 mm). The observed Q
of this particular resonance was 185,800. Hence, the effective path length is about
33
200
W−band system
180 F−band system
160
140
Q (x1000)
120
100
80
60
40
70 80 90 100 110 120 130 140 150
Frequency (GHz)
Figure 3.3: Quality factor of the resonances of the Fabry–Perot resonator measured
at vacuum and room temperature (T=297 K).
115.01 m. Figure 3.4 shows a plot of the effective path length of the resonances as
a function of frequency as measured at room temperature. The performance of the
FPR improves slightly at lower temperatures.
W band system
The W band measurement system is used to measure the 3–4 mm-wavelength
properties of ammonia, and is shown in Figure 3.5.
A swept signal generator (HP 83650B) is used to generate signals in the 12.5–18.3
GHz range which is fed to a times six active multiplier chain (AMC) via flexible,
low-loss, high frequency 2.9 mm male–2.9 mm male coaxial cable assembly. All coax
connections are tightened with an 8 in-lb (0.90 N-m) torque wrench to ensure reliable
connections. The operating temperature of the AMC is 0◦ C–45◦ C. Hence, the AMC
is heat sunk to keep the case temperature below +45◦ C when it is used at room
temperature. When the AMC is used at cold temperatures, the heat sink is removed
34
120
W−band system
F−band system
100
Effective Path Length (m)

80
60
40
20
0
70 80 90 100 110 120 130 140 150
Frequency (GHz)
Figure 3.4: Effective path length of the resonances of the Fabry–Perot resonator
measured at vacuum and room temperature (T=297 K).
and thermal insulation is provided so that the heat generated by the AMC could
sustain its own operation even if the external temperature is -100◦ C. The AMC is
hermetically sealed to prevent ice deposition at cold temperatures. The output from
the AMC is fed to one of the ports of the FPR via a solid WR-8 waveguide section.
The radio frequency (RF) signal from the output port of the FPR is fed to a
QuinStar Technology QMH series harmonic mixer via solid WR-8/WR-10 waveguide
sections so as to enable down-conversion of 3–4 mm-wavelength signal. The QMH
series harmonic mixer has a common SMA port for the local oscillator (LO) and
intermediate frequency (IF) signals, and hence an external diplexer (model MD1A) is
used to combine/separate the LO and IF signals. The typical conversion loss of this
mixer is approximately 40 dB. Both the mixer and the diplexer are hermetically sealed
when used in the freezer. The harmonic mixer locked to the 18th harmonic is used
in the “external mixer” mode with a spectrum analyzer (HP 8564E). The detector
within the spectrum analyzer operates in a positive peak mode, which displays the
35
2-3 bar 0-2 bar
Pressure Pressure
Gauge Gauge
P V
H2-
NH3 He Ar
Mix
Vacuum Pump
Temperature Chamber
Power
X6
Supply
Active Multiplier
Chain
Millitech
AMC-10-RFH00
HP 83650B
Fabry-Perot
Swept Signal
Resonator
Generator
HP8564E
High
Resolution
Spectrum
QuinStar MD1A Analyzer
Diplexer
Thermocouple Harmonic Mixer
QuinStar
922WHP/387
Computer
Interface
Figure 3.5: Block diagram of the W band measurement system for studying ammo-
nia gas properties under simulated upper tropospheric jovian conditions. Solid lines
represent the electrical connections and the arrows show the direction of signal prop-
agation. Valves controlling the flow of gases are shown by the small crossed circles.
36
maximum power level received during the integration time of each point on each
individual sweep. This mode is used primarily because it maximizes the data return
to the computer. The normal mode detects both the high and low signal (noise
floor) intensities at each frequency point, but when transferring to the computer,
the spectrum analyzer is limited to 601 points in both the frequency and amplitude
axes. In the normal mode, the peak level data becomes interspersed with the noise
floor data, which would result in only half the data transferred being of practical
use for these measurements and consequently would halve the frequency resolution.
The mixer can also be used as a down-converter with an external LO, but a different
diplexer is used (model MD2A).
As shown in Figure 3.5, the AMC and the harmonic mixer are placed as close to
the resonator input/output ports as possible to reduce the signal loss and increase
the signal to noise ratio (SNR) of the resonances. The signal power generated by
the signal source, the loss through the FPR, and the conversion loss of the mixer are
dependent on the frequency. Hence, the signal to noise ratio of the received signal is
also dependent on the frequency. For example, the swept signal generator is used to
generate a 10 dBm signal at 12.5 GHz. This is fed to the AMC, which produces the
RF signal at 75 GHz with a signal power of 5 dBm. At 75 GHz, the loss through the
FPR is approximately 20 dB and the conversion loss of the mixer is approximately 40
dB. Hence, the signal power that reaches the spectrum analyzer is approximately -55
dBm at 75 GHz. Additional information on the various components of the W band
subsystem is provided in Appendix A.
F band system
The F band measurement system is used to measure the 2–3 mm-wavelength prop-
erties of ammonia and is shown in Figure 3.6.
The signal generated by a swept signal generator (HP 83650B) in the 33–50 GHz
37
2-3 bar 0-2 bar
Pressure Pressure
Gauge Gauge
P V
H2-
NH3 He Ar
Mix
Vacuum Pump
Temperature Chamber
Spacek Labs Pacific Millimeter
SG4413-15-16W D3WO
Frequency
AMP Tripler
(100-150 GHz)
HP83712B
HP 83650B
Synthesizer
Swept Signal (12-20 GHz)
Generator
(37-50 GHz) Fabry-Perot Resonator
HP8564E
JCA Technologies
JCA 1920-612 High
Resolution
AMP Spectrum
Analyzer
Diplexer Bias Tee
Thermocouple Harmonic Mixer
Pacific Millimeter
DM
V
Pacific Millimeter
MD2A IF LNA
(1 GHz)
MITEQ
AMF-3F-012017
Computer
Interface
Figure 3.6: Block diagram of the F band measurement system for studying am-
monia gas properties under simulated upper tropospheric jovian conditions. Solid
lines represent the electrical connections and the arrows show the direction of signal
propagation. Valves controlling the flow of gases are shown by the small crossed
circles.
38
range is amplified (Spacek model SG 4413-15-16W) and fed to a frequency tripler.
To facilitate the operation of the amplifier inside the freezer, the heat sink mounted
on the amplifier is removed and adequate thermal insulation is provided. Both the
amplifier and the tripler are hermetically sealed while operating at cold temperatures.
The output from the tripler is fed to one of the ports of the FPR via a solid WR-8
waveguide section.
The RF signal from the output port of the FPR is fed to a harmonic mixer (Pacific
Millimeter Products model DM) via a solid WR-8 waveguide section to enable down-
conversion of 2–3 mm-wavelength signal. The harmonic mixer can operate with an
LO frequency as high as 18 GHz and has a common SMA port for LO and IF signals.
An external diplexer (model MD2A) is used to combine/separate LO and IF signals.
For a particular RF and IF frequency, LO frequency can be computed using
fRF − fIF
fLO = , (3.28)
NH
where NH is the lowest integer such that fLO < 18 GHz. The harmonic mixer
has one diode, and requires a DC return path for the diode current. This DC return
path follows the IF path in the diplexer, and the device attached to the IF port of
the diplexer must have provision for a bias-tee. Mixer conversion loss is dependent
upon the frequency, the harmonic number, applied LO power, and the diode current.
For the 10th harmonic (LO frequency=11–17 GHz for RF=110–170 GHz) conversion
loss is approximately 40 dB. Sensitivity of the receiver system will depend on the
receiver bandwidth, but the mixer conversion loss sets a minimum noise contribution
for the receiver system of 40 dB. If the internal spectrum analyzer local oscillator
(3–6 GHz LO frequency) is used, an even higher conversion loss associated with the
low LO frequency and high harmonic mixing number results. Hence, a separate CW
signal generator (HP 83712B) along with an amplifier (JCA Technologies model JCA
1920-612) are used as the LO, and harmonics less than the tenth order are used.
39
The IF signal is then enhanced using a low noise amplifier (MITEQ model AMF-3F-
012017) and displayed on the spectrum analyzer. The IF is chosen such that there
is minimal radio frequency interference (RFI). RFI is further mitigated by wrapping
aluminum foil tapes around the microwave components and connectors at the IF
end. Additional information on the various components of the F band subsystem is
provided in Appendix A.
3.3.3 Data Handling Subsystem
The data acquisition system consists of a computer connected to the spectrum ana-
lyzer (HP 8564E), swept signal generator (HP 83650B), and CW signal generator (HP
83712B) via a general purpose interface bus (National Instruments GPIB Controller
for Hi-Speed USB). The instruments are controlled via Matlab

R
and the Standard
Commands for Programmable Instruments (SCPI). The software used is similar to
that used by (Hanley and Steffes, 2007) with modifications to account for the suite
of instruments used in this measurement system.
3.4 High-Pressure Measurement System
A high-pressure measurement system was used to conduct measurements of the 5–20
cm-wavelength absorptivity of ammonia under deep jovian conditions at pressures up
to 100 bar and temperatures up to 500 K. The high-pressure measurement system
consists of a planetary atmospheric simulator, centimeter-wavelength subsystem, and
a data handling system.
3.4.1 High-Pressure Planetary Atmospheric Simulator
The planetary atmospheric simulator was designed by Karpowicz and Steffes (2011).
The simulator consists of a high-pressure vessel, temperature chamber, water-reservoir,
mass flow meter, temperature and pressure gauges, vacuum pump, various gas bottles,
gas handling valves, and pipes. The main component of the planetary atmospheric
40
simulator is a high-pressure vessel that operates at pressures up to 100 bar and tem-
peratures up to 520 K custom built by Hays Fabrication and Welding. The custom
pressure vessel was designed with two 1/2” NPT ports, one 1/4” NPT port, and
two CF-1.33 flanges with 3.38 cm diameter for microwave feedthroughs. One of the
1/2” NPT ports is used as a gas inlet while the other 1/2” NPT port is used for a
temperature sensor. The 1/4” NPT port serves as the gas outlet. A composite glass
fiber/nitrile butadiene rubber KLINGERsil

R
C-4430 O-ring is used to seal the pres-
sure vessel along with twenty 6 cm diameter nuts torqued to 1760 Nm. The pressure
vessel is constructed out of a 30.48 cm section of schedule 100 pipe that is 35.56 cm
in diameter (outer dimension). An elliptical head is welded to the bottom of the
pressure vessel giving it a maximum interior height of approximately 46 cm. The top
of the pressure vessel is an ANSI class 900 flange that is 10.16 cm thick, and the top
plate is 9.2 cm thick. The pressure vessel has a volume of approximately 32.3 liters
and weighs approximately 545 kgs.
The pressure vessel is contained in a temperature chamber, which is a Grieve

R
industrial oven (model AB-650) that is rated to a temperature of 615 K and weighs
approximately 740 kgs. The combined weight of the pressure vessel and the oven far
exceeded the load capacity of the laboratory floor. It was determined that an outdoor
concrete pad on which a decommissioned crane once stood would be ideal for such a
load. Hence, all the system components, except the network analyzer and the data
acquisition system, are placed on the outdoor concrete pad. A metallic EZEE shed
was built around the system components for protection.
A water reservoir made of a 304 stainless steel pipe that is 46 cm long and 3.8 cm in
diameter is connected the pressure-vessel and placed inside the oven to produce water
vapor. Various gas bottles (hydrogen, helium, ammonia, pre-mixed hydrogen/helium,
and argon) required for simulating the jovian atmosphere and for dielectric matching
measurements are placed on a six-pack gas cylinder rack adjacent to the EZEE shed.
41
The gases and mixtures are ultra-high-purity UHP300 grade and supplied by Airgas.
Gas delivery pressure regulators, Matheson

R
model 3030-580 were used for argon
and helium, Matheson

R
model 3030-350 was used for hydrogen, and Airgas
R
model
Y11E444B660 was used for ammonia. A Teledyne-Hastings HFM-I-401 industrial
flow meter capable of operating up to 103 bar pressure is used for mass flow sensing.
When the gases are added to the pressure vessel, the flow totalizer function of the
flow meter is used to estimate the system volume since the volume changes by a
small but measurable amount with temperature. All components in the gas handling
subsystem are connected with seamless 3/8” outer diameter stainless steel tubing and
Swagelok
R
fittings. Gases are vented down to ambient pressure though an exhaust
valve and a Welch DuoSeal

R
vacuum pump model 1376B-01 is used to evacuate the
pressure vessel from 1 bar down to vacuum.
Pressure measurements are made with a GE Druck DPI 10430A Digital Test
Gauge in the 0 to 2 bar range, GE Druck DPI 104300A Digital Test Gauge in the 2
to 20 bar range, and an Omega

R
PX1009L0-1.5KAV pressure transducer for pressures
up to 100 bar. The GE Druck DPI 10430A gauge has a resolution of 0.1 mbar and the
GE Druck DPI 104300A gauge has a resolution of 1 mbar. Both the pressure gauges
have a 0.05% full scale total error band accuracy and display the absolute pressure
within the pressure vessel. The pressure gauges are powered externally with a 9 V
power supply and have an RS-232 serial interface for data acquisition. The Omega
R
pressure transducer has an accuracy of ±0.25% full scale output and measures the
absolute pressure. The voltage from the pressure transducer is read via a HP 34401A
multimeter.
The temperature of the gases in the pressure vessel is monitored with an Omega
R
resistance temperature detector (RTD) (PR-11-2-100-1/8-9-E). The RTD has a 0.01◦ C
resolution and the accuracy is IEC/DIN Class A (±0.15◦ C at 0◦ C and ±0.55◦ C at
200◦ C). The RTD is connected to the 1/2” NPT port of the pressure vessel via a 1/8”
42
Swagelok
R
tube to 1/2” NPT adapter (SS-200-1-8BT). For experiments conducted
below 450 K, a high temperature thermometer/hydrometer assembly (E+E Elek-
tronik EE33-MFTI9205HA07D05/AB6-T52) inserted into the 1/2” NPT port of the
pressure vessel was used to measure the temperature. The temperature resolution of
the thermometer/hygrometer is 0.01◦ C and the accuracy is between ±0.25◦ C at 0◦ C
and ±0.5◦ C at 180◦ C. The temperature of the pressure vessel is maintained within
±0.1◦ C of the measurement temperature throughout an experiment sequence. The
temperature inside and outside the oven are monitored by various T type thermocou-
ple probes. A block diagram of the EZEE-shed along with all the system components
are shown in Figure 3.7. The valves shown in blue in Figure 3.7 are high tempera-
ture Swagelok
R
SS-1RS6-PK valves rated to 315◦ C at a maximum pressure of 215
bar, and the valves shown in black are Swagelok

R
SS-1RS6 valves rated to 93◦ C at a
maximum pressure of 295 bar. Table 3.1 lists the instruments used in the planetary
atmospheric simulator of the high-pressure system along with their operating condi-
tions and 3σ precision. Figure 3.8 shows the high-pressure system in assembly along
with the EZEE shed and the gas cylinder rack and Figure 3.9 shows the high-pressure
vessel in and after assembly.
3.4.2 Centimeter-Wavelength Subsystem
The centimeter-wavelength subsystem has greatly benefited by continuous improve-
ment over the past twenty years (see, e.g., DeBoer and Steffes, 1996; Hanley and
Steffes, 2007). At the heart of the subsystem is a type 304 stainless steel cylindri-
cal cavity resonator placed inside the high-pressure vessel. The inside of the cavity
resonator is plated with gold to improve the quality factor of the resonances, and to
prevent reaction with corrosive gases. The interior dimension of the cavity resonator
is approximately 13.1 cm in diameter and 25.5 cm in height, and is ideal for measure-
ments in the 5–20 cm wavelength range. The resonator consists of two closed-loop
43
EZEE-SHED
Analog High Pressure Gauge
Teledyne-Hastings
Flow Meter
0-20 0-2
bar bar
Ar H2 H2
High
Pressure
Transducer
103 bar max
Hays
H2/He He H2
Pressure
Vessel
NH3
Vacuum
Pump
Exhaust Water
Reservoir
Figure 3.7: The Georgia Tech high-pressure system used for studying the centimeter-
wavelength properties of ammonia under simulated jovian conditions (Karpowicz and
Steffes, 2011). The valves shown with a blue dot are high temperature valves.
44
Figure 3.8: Assembly of the high-pressure system.
45
Table 3.1: High-pressure planetary atmospheric simulator instruments.
Instrument Operating Measurement Pa- 3σ Precision
Range (◦ C) rameter
GE Druck DPI 10430A -10 to 50 Pressure (0–2 bar ab- ±0.05% FS
solute)
GE Druck DPI 104300A -10 to 50 Pressure (0–20 bar ab- ±0.05% FS
solute)
Omega
R
PX1009L0-1.5KAV -54 to 343 Pressure (0–103 bar ±0.25% FS
absolute)
HP
R
34401A multimeter 23±5 Voltage (pressure ±(0.0035% of read-
transducer output) ing + 0.0005% of
range)
Omega
R
PR-11-2-100-1/8-9-E -50 to 450 Temperature (50 to ±(0.15 + 0.002 of
450◦ C) reading)◦ C
Omega
R
DP41B 0 to 50 Temperature (-200 to ±0.2◦ C
900◦ C)
Type T thermocouple -200 to 400 Temperature (-200 to greater of ±1.0◦ C
400◦ C) or 0.75% of reading
E+E Elektronik EE33- -40 to 180 Temperature (-40 to ±0.25–0.5◦ C
MFTI9205HA07D05/AB6-T52 180◦ C)
Teledyne-Hastings HFM-I-401 -20 to 70 Flowrate 0–10 slm, op- ±(0.2% FS + 0.5%
erates up to 103 bar reading)
(a) Microwave cavity resonator and SiO2 ca- (b) Assembled pressure vessel
bles inside pressure vessel. and water reservoir inside the
oven.
Figure 3.9: Pictures of the pressure vessel in and after assembly.
46
(a) Gold-plated resonator with the top plate removed. (b) Resonator with di-
electric spacers.
Figure 3.10: Cylindrical cavity resonator.
antenna probes mounted on the top plate and oriented to maximize the Q of the
TEM(0,M,L) mode (Hanley et al., 2009). The resonator has two horizontal slits on its
circular side near the top plate to suppress unwanted degenerate TM resonant modes
and also to allow gases to enter the interior. Additionally, the top and the bottom
plates of the resonator are isolated by approximately 1.75 mm from the cylinder by a
series of Teflon
R
washers around the connecting screws. This isolation eliminates the
degenerate TM(1,M,L) mode and greatly improves the sensitivity of the system, but
increases the microwave energy leaking out of the resonator. Hence, the resonator
is wrapped with a stainless steel mesh screen to damp interference of the microwave
energy reflected off the pressure vessel walls back into the resonator. The two ports of
the resonator are essentially symmetric and Qs of the resonances range from 30,000
to 80,000. The resonator with the top-plate removed and the assembled resonator
with dielectric spacers are shown in Figure 3.10.
The center frequencies of the transverse electric (TE) and transverse magnetic
(TM) mode resonances in a cylindrical cavity resonator are calculated as
47
s 2 2
c pN M L×π
fT E(N,M,L) = √ + , (3.29)
2π µr εr rd ht
s 2 2
c qN M L×π
fT M (N,M,L) = √ + , (3.30)
2π µr εr rd ht
where N , M , and L refer to the zeros in the standing wave patterns in the cir-
cumferential, radial, and axial dimensions, respectively, µr and εr are the relative
permeability and permittivity of the material contained in the resonator, c is the
speed of light in cm/s, rd and ht are radius and height of the cylindrical resonator
in cm, qN M is the M th zero of the N th order Bessel function, and pN M is the M th
zero of the first derivative of the N th order Bessel function of the first kind (Pozar,
1998). In this work, only TE mode resonances are used because of their high quality
factors. The TM modes have been intentionally suppressed to minimize interference
with the neighboring TE mode resonances. One dozen high-Q, low-asymmetry reso-
nances were selected and used as “standard resonances” for all the measurements. A
detailed description of the selection criteria for the resonances was provided by Hanley
et al. (2009).
The resonator is connected to the high-pressure Ceramtec

R
microwave feedthroughs
(model 16545-01-CF) on the top-plate of the pressure-vessel using Times Microwave

R
SiO2 microwave cables, to couple microwave energy in and out of the pressure ves-
sel. The SiO2 cables are rated to 875 K and the SMA Ceramtec feedthroughs are
rated to 103 bar and 625 K. Exterior to the pressure vessel, two 1 m long sections
of high-temperature Times Microwave M17/86-00001 (RG-225) cables fitted with
type-N connectors are connected to the Ceramtec feedthroughs via high-temperature
type-N to SMA adapters. These cable assemblies were assembled in the laboratory
with high-temperature solder rated to 500 K. One end of these cables with type-N
bulkheads are positioned outside of the oven via holes that were drilled on the side
48
wall of the oven. Two sections of approximately 24.4 m length Andrews
R
CNT 600
microwave cables, rated to 350 K are connected to the type-N bulkheads on the RG-
225 cable assemblies outside the oven and connect to an Agilent

R
E5071C vector
network analyzer that is situated inside the laboratory. The network analyzer op-
erates from 300 kHz to 8.5 GHz with a high-stability timebase (Option 1E5). The
long CNT 600 microwave cables are required to enable the placement of the network
analyzer within the laboratory environment to ensure its temperature stability. The
S-parameters measured by the network analyzer are read into the data acquisition
system via GPIB.
3.4.3 Data Acquisition System
The data acquisition subsystem consists of a laptop computer connected to the net-
work analyzer and multimeter via a GPIB connected to a National Instruments NI-
488.2 interface card, and a data logging computer connected to the various pressure
and temperature sensors via extended USB buses and USB to RS-232 adapters. The
extended USB buses allow the data logging computer to remain inside the labora-
tory. A schematic of the centimeter-wavelength subsystem and the data-acquisition
components of the high-pressure measurement system are shown in Figure 3.11. The
network analyzer is controlled via Matlab

R
and the Standard Commands for Pro-
grammable Instruments (SCPI) and the suite of pressure and temperature sensors
are controlled via Python

R
. The Matlab software used is similar to that used by
Hanley and Steffes (2007) and a detailed description of the data acquisition system
is provided by Karpowicz (2010).
49
EZEE Shed
TeledyneHastings Audio
2, 20 bar Oven Flowmeter via
gauges RS232 RS232
Control USB extended
Gauges
connected to
Laptop
RS232 via Inside Van
USB High Leer
Pressure
Transducer
USB/ RS232 Young
Extender Barometer
SiO2 Cable Data Logging Computer

RS232
RS232
RTD
Times Microwave
RG225/M17/86-
Resonator
00001
DP41B RTD Audio

GPIB
Shielded Pair
Andrews C600
Microwave Cables Network
80 ft x 2 Voltmeter Analyzer
Figure 3.11: The centimeter-wavelength subsystem and the data-acquisition compo-

nents of the high-pressure system (Karpowicz and Steffes, 2011).
50
3.5 Measurement Procedure
One of the most important prerequisites for performing the measurements of gas
properties is a leak-proof system. Leak-proofing the system not only reduces the
uncertainties in the experiments, but also ensures that no toxic or flammable gases
leak from the system and pose a health-hazard. The pressure vessel is assembled and
sealed at the coldest operating temperature, which ensures that it remains leak-proof
at warmer temperatures since warming expands the materials and only improves the
seal. The pressure vessel, gas handling pipes, valves, and gas regulators are positive
pressure-tested with a liquid leak detection compound at the highest permissible
pressure. Further tightening of the pressure vessel is achieved after drawing a vacuum
since there is additional compression force on the surface of the pressure vessel due
to normal atmospheric pressure. Once a good seal is established, the pressure vessel
is filled with argon (three bar for the glass pressure vessel and 95 bar for the high-
pressure vessel) and the pressure is monitored for several days to ensure that there is
a good pressure seal.
Once a good pressure seal is established in the millimeter-wavelength system,
the measurement process begins by characterizing the resonances of the FPR at the
measurement temperature. Whenever the temperature of the FPR changes, the metal
stand-offs that hold the mirrors in place and the mirrors themselves expand/contract,
and hence the distance between the mirrors and their alignment change slightly. Any
change in mirror alignment affects the off-axis resonances which move in frequency
with respect to the axial mode resonances. As a result, some axial mode resonances
become contaminated with the off-axis resonances, resulting in compound asymmetric
resonances. Asymmetry (As ) is defined as
(fhigh − fcenter ) − (fcenter − flow )

As = 100 × %, (3.31)
(fhigh − flow )
51
(DeBoer and Steffes, 1994), where fhigh , fcenter , flow , represent the higher fre-
quency half power point, center frequency, and lower frequency half power point of
the resonances, respectively. Since asymmetric resonances might broaden dispropor-
tionately when lossy gases are added, only those resonances with As ≤ 5% are used
for the measurements. About a dozen high-Q, low-asymmetry resonances that are
roughly uniformly spaced in frequency are selected to be used as the “standard reso-
nances” for all the measurement sequences at a particular measurement temperature.
For the centimeter-wavelength system, one dozen high-Q, low-asymmetry “standard
resonances” were chosen based on the selection criteria provided by Hanley et al.
(2009) and used for all the measurements.
Owing to the difficulties in frequently changing the measurement temperature,
and because changes in pressure and mixing ratios can be achieved quickly, the gas
properties are measured at a particular temperature for different sets of mixing ra-
tios and gas pressures. Once the desired set of measurements are conducted at a
particular temperature, the temperature of the chamber is changed and a new set of
measurements are made at the new temperature.
When gas molecules are added to the pressure vessel, they have a tendency to ad-
here (physically or chemically adsorb) to the surface of the pressure vessel. Physisorp-
tion is the process of gaseous molecules of one substance, the adsorbate, adhering to
the surface of a liquid or solid substance, the substrate, due to weak attractive van
der Waals forces between the adsorbate and the substrate. This does not result in
any sharing of electrons and is not to be confused with chemisorption that requires
the formation of a chemical bond between the substrate and the adsorbate. The for-
mation of physically adsorbed layer may be likened to the condensation of a vapor to
form a liquid and most likely occurs when the gas is near its condensation point, but
can occur under warmer, lower pressure conditions (Young and Crowell, 1962). If p is
the equilibrium pressure of the adsorbed film and p0 is the saturation vapor pressure
52
of the gas at the experimental temperature, then below p/p0 = 0.01 no significant
adsorption takes place (Young and Crowell, 1962). While all molecules can encounter
physisorption, polar molecules are more prone to it, especially when the substrates
are conductors. An analogue is a dipole interacting electromagnetically with its image
in the plane of the conductor. Adsorbed molecules have limited degrees of freedom,
and hence do not interact with the incident electromagnetic energy to the extent they
would in the gaseous phase. Adsorbed molecules, however, do change the electrical
conductivity of the surfaces, to which they adsorb (Young and Crowell, 1962). In
this measurement system, the effect of adsorbed ammonia on the conductivity and
resulting Qs of the resonator was found to be negligible. Desorption is the process by
which the adsorbed molecules are returned to the gas phase. The most effective ways
for molecules to desorb are by the reduction of the partial pressure of the adsorbate
or by the increase of the temperature of the substrate to provide enough energy for
the adsorbate to break the weak attractive van der Waals forces that hold them in
place.
A particular example of ammonia adsorption/desorption is discussed here, but
it should be noted that most polar molecules behave in a similar fashion. A study
by Rodrigues and Moraes Jr. (2002) found that at all temperatures the physisorp-
tion capacity of ammonia increased with increasing ammonia concentration, but an
increase in temperature at any concentration led to a decrease in the adsorption ca-
pacity of ammonia. The study also concluded that the effect of temperature more
significantly affected desorption than ammonia concentration. Non-polar molecules
such as hydrogen and helium on the other hand do not adhere to the substrate at
our experimental conditions. It is difficult to quantify or predict adsorption due to
its dependence on the fine molecular structure of the substrate. Hence, precautions
must be taken to account for adsorption, whenever gas mixtures are added to the
system where one component adsorbs more than another. Earlier studies of ammonia
53
opacity made under simulated jovian conditions did not properly account for adsorp-
tion (Mohammed and Steffes, 2004; Spilker, 1990). When manufactured mixtures of
hydrogen, helium and ammonia are used (Mohammed and Steffes, 2004), some ammo-
nia in the mixture adsorbs to the surface of the measurement system thereby reducing
its concentration in the gas phase, and there is no effective method to determine the
exact concentration of ammonia left in the measurement system. When a series of
progressive dilutions are employed to arrive at mixtures with a small ammonia con-
centration (Spilker, 1990), there is a shift in the adsorption/desorption equilibrium
since the concentration of ammonia decreases after partial venting. This leads to
additional desorption and disproportionate increase of the concentration of that gas.
One way to compensate for adsorption is to presaturate the surface of the mea-
surement system with the adsorbate so that no more adsorption occurs since there
are only so many layers of adsorbate that can form on the substrate. This is achieved
by adding ammonia to the pressure vessel first and letting the system completely
saturate while simultaneously monitoring the Qs of the resonances. After ammonia
is added, there is a reduction in the Qs of the resonances because of the absorption of
electromagnetic energy by ammonia in the gas phase. Once ammonia starts adhering
to the surface of the resonator, its concentration is reduced, and hence the Qs increase
slightly. After ammonia completely saturates the surface of the pressure vessel, the
Qs stabilize, and at this stage adsorption equals desorption. After this stage, hydro-
gen and helium are added to the system. Measurements of gas properties are made
starting at the lowest pressure by adding ammonia and working up to higher pres-
sures by adding hydrogen and helium. Measurements are never made by venting the
gas mixture down to lower pressures because of the possibility of enhanced ammonia
desorption from the substrate due to the reduction in the partial pressure of ammonia
thereby leading to a disproportionate increase in the concentration of ammonia.
54
Each measurement cycle consists of first measuring the pre-selected standard res-
onances of the resonator at vacuum. Test gases (eg. ammonia, hydrogen and helium)
are added to the system and after the system stablizies, the frequency shifted reso-
nances are measured at different pressures. Because of the large volume and thermal
inertia, it takes much longer for the centimeter-wavelength system to stabilize when
compared to the millimeter-wavelength system. For the centimeter-wavelength sys-
tem, it takes about 8–10 hours for the system to stabilize after the test gases are
added and for the millimeter-wavelength system, it takes about 2–3 hours for system
stabilization. After the frequency shifted resonances are measured, the pressure vessel
is vented down to the ambient pressure and a vacuum is drawn for at least 12 hours
to remove the test gases. A second set of vacuum measurements of the resonances
is made. A non-absorbing gas such as argon or carbon dioxide is then added to the
system to shift the frequency of the resonances by the exact same amount as the test
gases and the dielectrically matched measurements are made. The dielectric match-
ing process is aided by a series of tones produced by the data acquisition computer
to help the experimenter precisely match the center frequency of the resonances. The
system is evacuated once again to remove the matching gas and a third set of vacuum
measurements is made. After this step, three sets of straight-through measurements
of signal levels are made (without the resonator present) under the same conditions
at each pressure/frequency point of the test gas, so as to correct the measurements
of transmissivities (loaded and matched) for losses in the cables or waveguides. For
the millimeter-wavelength system, these measurements involve disconnecting the W
band/F band signal generator and detector from the resonator and coupling 1/100th
of the RF signal power (via a WR-10 20 dB directional coupler) from the generator to
the detector and making multiple measurements at the standard resonant frequencies.
For the centimeter-wavelength system, these measurements involve disconnecting the
RG-225 cables from the Ceramtec feedthroughs and connecting the cables together
55
via a high-temperature female-to-female SMA adapter and making multiple mea-
surements at the standard resonant frequencies. Between the repeated signal level
measurements, all the waveguide/coax connections are disconnected and reconnected
to better statistically characterize the reproducibility of the electrical connections.
Signal levels are only measured at the end of the experiments so as to ensure that the
resonator stays at a constant temperature (± 0.5◦ C for the millimeter-wavelength sys-
tem and ±0.1◦ C for the centimeter-wavelength system) throughout the measurement
cycle. (Signal level measurements involve opening the temperature chamber to dis-
connect/reconnect the signal generator, detector, waveguides, and/or coax assembly,
and this might temporarily alter the temperature of the chamber.)
3.6 Data Processing
The data are processed after each measurement cycle is completed. Software written
in Matlab
R
is used to automatically load and process the raw data. The software used
is similar to that described by Hanley and Steffes (2007), but with modifications to
account for the topology of the new measurement systems. The automated software
loads the raw data files, runs smoothing algorithms, and calculates the absorptivity of
the test gas at the measured frequencies, pressures, temperatures, and mixing ratios.
The shape of any major variation in the straight-through signal level measure-
ments of the cables/waveguides/adapters over the measurement frequency range of
the resonances are deconvolved from the sweeps. Since this variation is < ±0.25 dB
for the centimeter-wavelength system, this effect is omitted. The cable/waveguide
losses at the center frequency of the resonances under each pressure, temperature,
and mixing ratio condition is calculated by averaging the values of the mean of each
set of sweeps of the signal level measurements. The loaded insertion loss (Sl ) and
matched insertion loss (Sm ) are obtained by subtracting the cable/waveguide losses
from the peak power measurements of the test gas and the dielectric matching gas,
56
respectively. The insertion losses are used to generate the quantities tloaded and tmatched
in Equation 3.22.
For the millimeter-wavelength system that employs the spectrum analyzer, the
measured bandwidth (BWmeasured ) of a resonance is a function of the resolution band-
width (RBW) and the filter response (shape) of the spectrum analyzer which is ap-
proximately Gaussian (1 MHz < RBW < 2 MHz) or synchronously tuned 4-pole filter
(300 Hz < RBW < 300 kHz). RBW is chosen to be ∼ 1/100th of the frequency span.
The actual bandwidth is calculated from the measured bandwidth by

q
2
BWactual = BWmeasured − RBW 2 . (3.32)
This correction is unnecessary for the centimeter-wavelength system that employs the
network analyzer, since it is programmed to display the true response of the device
under test.
3.7 Measurement Uncertainties
The uncertainties associated with the measurements include instrumental errors (σ n ),
errors in dielectric matching (σdiel ), transmissivity errors (σtrans ), errors due to asym-
metry in resonances (σasym ), and errors in measurement conditions (σcond ) due to
uncertainties in measurement temperature, pressure, and mixing ratio.
The instrumental errors and electrical noise arise because of the sensitivity of the
electrical devices and their ability to accurately measure the center frequency (fo ) and
bandwidth (BWmeasured ). Electrical noise arises from the noise of the internal elec-
tronics and the accuracies of the frequency references. Electrical noise is uncorrelated
and the best estimate of instrumental uncertainty is the mean of multiple measure-
2
ments. The variance of the best error estimate is given by the sample variance (SN )
weighted by the confidence coefficient (B) as

2
2 SN
σN =B× , (3.33)
Nsamples
57
where Nsamples is the number of independent measurements of the sample. For
the millimeter-wavelength system that employs the spectrum analyzer, five sets of
independent measurements of each resonances are taken, and for the centimeter-
wavelength system that employs the network analyzer, 30 sets of independent mea-
surements of each resonances are taken. A confidence coefficient corresponding to
the 95% confidence interval (approximately 2σ) is used. The confidence coefficient
is equal to the critical value of the Student t test (Student, 1908) for a two-tailed
significance of 0.05 and the degrees of freedom (Nsamples − 1). This means that the
true mean will fall above the confidence interval 2.5% of the time and below the con-
fidence interval 2.5% of the time. The critical value of B for 95% confidence for five
samples is 2.776 and for 30 samples is 2.045. The center frequency sample standard
deviation of the measurements is very small and its effect on the uncertainty in Q is
negligible. Hence, SN refers only to sample standard deviation of the bandwidth of
the measurements.
The HP 8564E spectrum analyzer is used for measuring the resonances in the
millimeter-wavelength system and the manufacturer specified instrumental uncertain-
ties are the 3σ values (Hewlett-Packard, 1997). The 3σ standard deviation for the
center frequency and bandwidth measurements, respectively, are estimated by
σo ≤ ± (fo × fref acc + 0.05 × SPAN + 0.15 × RBW + 10) (Hz), (3.34)
σBW ≤ ±(BWmeasured × fref acc ) + 4 × NH + 2 × LSD (Hz), (3.35)
where fref acc is given as
fref acc = (aging × time since calibration) + initial achievable accuracy
+ temperature stability, (3.36)
and fo , SPAN, RBW, NH , and LSD are the center frequency, frequency span, res-
olution bandwidth, harmonic number, and least significant digit of the bandwidth
58
measurement, respectively. LSD is calculated as LSD = 10x for the smallest posi-
tive integer value of x such that SPAN < 10x+4 . These formulae are for SPAN > 2
MHz×NH , which is a wide-span scenario. For SPAN ≤2 MHz×NH , SPAN multipli-
cation factor of 0.05 is replaced with 0.01. For the spectrum analyzer used in the
experiments, fref acc reduces to
fref acc = (10−7 × years since calibrated) + 3.2 × 10−8 . (3.37)
There is a 3σ uncertainty associated with the RBW (±25%). However, this
changes the value of BWactual (Equation 3.32) by less than 0.01%, and hence can
be omitted for all practical purposes in the computation of the measurement uncer-
tainties.
For measurements made with the Agilent E5071C-ENA Vector Network Analyzer,
the relative uncertainty in measured frequency is calculated as
σo = fmeasured × 5 × 10−8 + 5 × 10−7 × years since calibrated

(Hz), (3.38)
with fmeasured given in Hz. The absolute uncertainty in measured frequency is as low as
1 ppm of the frequency reading (Hanley et al., 2009) and hence can be ignored. Since
no specifications were available for calculating uncertainties in measured bandwidth
for the network analyzer, the formula provided by Hanley et al. (2009) was used
√
2 × BWmeasured × 5 × 10−8 + 5 × 10−7 × years since calibrated

σBW = (Hz),
(3.39)
where BWmeasured is given in Hz.
The formulae for calculating σo and σBW for the spectrum analyzer and the net-
work analyzer represent the 3σ uncertainties and are scaled by 2/3 to obtain the
2σ uncertainties. The instrumental uncertainties are also based on the specifications
provided by the manufacturer and the instruments are typically more accurate than
this. However, these formulae represent the worst case condition for instrumental
uncertainties and are used in our calculations.
59
The worst case scenario (greatest error) is considered in the computation of un-
certainties of the instrumental error and electrical noise and is computed as given
by DeBoer and Steffes (1994)
σψ2 = hz2l i + hz2m i − 2hzl zm i, (3.40)
where
Υ2i σo2

2σo σBW
hz2i i = 2 2 2
+ σBW + σN i + , i= l,m, (3.41)
foi Q2i Qi
2
Υl Υm σo 2 σo σBW σo σBW
hzl zm i = − × + σBW + + , (3.42)
fol fom Ql Qm Ql Qu
foi
Qi = , i = l, m, (3.43)
fBW i
√
Υi = 1 − t, i = l, m, (3.44)
where l, m denote loaded and dielectrically matched cases, respectively, and fol,om
and fBW l,BW m represent center frequency, and bandwidth of loaded and dielectrically
matched cases, respectively. The 2σ uncertainty of the measured gas absorption due
to instrumental errors and electrical noise is given by
8.686π
σn = σψ (dB/km), (3.45)
λ
where λ is the wavelength in km.
Errors due to dielectric matching, transmissivity, and asymmetry are computed
in a fashion similar to that used by Hanley et al. (2009), but with modifications
to account for the topologies of the new measurement systems. Errors in dielectric
matching are due to the minor mis-alignments in the center frequency between the
loaded and the matched measurements. This is the most trivial error because of
the highly accurate software-controlled matching of the center frequency between the
60
loaded and matched measurements. Though the gas used for dielectric matching
is lossless, the Q of the resonator can change slightly from that of vacuum values.
The magnitude of this effect is calculated by comparing the Q of the three vacuum
measurements to that of the dielectric matched measurements.

dQ Qvac,i − Qmatched
= for i = 1, 2, and 3. (3.46)
df i fvac,i − fmatched
The maximum of the three values is used to calculate a dQ value

dQ
dQ = × |floaded − fmatched | , (3.47)
df max
where floaded and fmatched are the center frequencies of the resonances under loaded
and matched conditions, respectively. The error in absorptivity due to imperfect
dielectric matching is then computed by propagating ±dQ through Equation 3.22
8.686π
σdiel =
λ √ √
1 − tloaded 1 − tmatched
× − m
Qmloaded Q matched + dQ
√ √
− − m (dB/km). (3.48)
Qmloaded Q matched − dQ
Transmissivity errors are due to the uncertainties in the measurement amplitude.
This is caused by the centimeter- and millimeter-wavelength instruments (signal gen-
erators, spectrum analyzer, and network analyzer), cables, adapters, and waveguides
used in the system. Absolute magnitude uncertainties for the signal sources can be
safely disregarded since the loaded, matched, and signal level measurements are all
made with the same signal power. The greatest uncertainty in the signal level mea-
surements results from disconnecting and reconnecting the waveguides and cables.
Since the waveguides and cables have to be disconnected at least once from the sys-
tem to characterize their loss, this procedure is repeated two more times to ensure a
better statistical bound on the uncertainty parameters. The standard deviation (SN )
of these measurements is calculated and is weighted by the confidence coefficient for
61
a two-tailed distribution for three samples with a 95% confidence interval (Student,
1908) and is divided by the square root of the number of measurements to yield
4.303
σmsl = √ SN . (3.49)
3
For the millimeter-wavelength system, the signal level measurements involve sam-
pling the RF power with a WR-10 20 dB directional coupler to feed to the harmonic
mixer for down-conversion and detection. This procedure ensures that RF power
input to the harmonic mixer does not exceed its maximum allowed input power of
-10 dBm. However, the WR-10 directional coupler does not uniformly sample the
input signal throughout the entire frequency range. Hence, an additional 1.5 dB
uncertainty is included in the calculation of the uncertainty due to insertion loss
(σins loss ). The signal generator has a temperature stability of 1 dB/10 ◦ C, but an
internal temperature equilibrium is reached after two hours of operation at a stable
ambient temperature (Hewlett-Packard, 1997). Since the measurement environment
outside the temperature chamber is maintained at a constant temperature throughout
the measurement cycle, this uncertainty can be disregarded. The total uncertainty in
insertion loss for the millimeter-wavelength system is calculated by
σins loss = σmsl + 1.5 (dB). (3.50)
For the centimeter-wavelength system, the signal level measurements involve dis-
connecting the RG-225 cables from the resonator and connecting them in a thru
configuration so that the cable losses can be measured at each resonant frequency.
Cable losses are measured three times by disconnecting and reconnecting the cables
to statistically characterize their loss and the measured signal level uncertainty σmsl
is calculated using Equation 3.49. While σmsl takes into account the variations in the
cables that arise when they are disconnected and reconnected, it does not account
for the variabilities in the SiO2 cables internal to the pressure vessel that cannot
be removed. To account for the variability in these SiO2 cables, a value 0.25 dB is
62
assumed based on worst case models of cable performance (Karpowicz and Steffes,
2011). An additional uncertainty of 0.5 dB is included to account for the variations
in the cable losses due to ambient temperature variations in the approximately 8.5
m long outdoor section of each of the 24.4 m Andrews

R
CNT 600 microwave cable.
The total uncertainty in insertion loss for the centimeter-wavelength system is given
as
q
2
σins loss = σmsl + 0.252 + 0.52 (dB). (3.51)
The error in insertion loss is used to compute the transmissivity error
1
10 −(Si −σins loss ) − 10 −(Si +σins loss ) ,

σt,i = (3.52)
2
i = l, m,
where l, m are the loaded and matched cases, respectively, and S is the insertion
loss of the resonator. This is used to compute the 2σ uncertainties in opacity and is
expressed as
8.686π
σtrans =
√2λ √ √ √
tl + σt,l − tl − σt,l tm − σt,m − tm + σt,m
× − (dB/km). (3.53)
Qm loaded Qm matched
Errors from asymmetry are due to the asymmetric nature of the measured res-
onances which are caused by the overlapping of the off-axis/TM mode resonances
on the axial/TE mode resonances. The asymmetric nature of the resonances are
more prominent at the shorter wavelength and colder temperatures. Errors due to
asymmetry result from the disproportionate asymmetric broadening of the loaded
measurements compared to the matched measurements. Equivalent full bandwidths
based on assuming symmetry of the high and low sides of the resonances are calculated
as
BWhigh = 2 × (fhigh − fcenter ), (3.54)
63
BWlow = 2 × (fcenter − flow ), (3.55)
where BWhigh , BWlow , fhigh , fcenter , and flow are the high bandwidth, low bandwidth,
higher frequency half power point, center frequency, and lower frequency half power
point, respectively. The difference between the opacities calculated using BWhigh and
BWlow (treated as 2σ errors) is defined as σasym and calculated as

8.686π
σasym =
λ √ √ !
1− t 1 − tmatched
loaded
× − m

Qm loaded,high Q matched,high
√ √ !
− − (dB/km), (3.56)
Qm Qm

loaded,low matched,low
where Qm m
loaded,high/low and Qmatched,high/low are the measured Qs evaluated using the high
and low bandwidths for loaded and matched cases, respectively.
The measured uncertainty in temperature, pressure, and concentration contribute
to the total uncertainties due to the measurement conditions (σcond ). Although σcond
does not directly affect the measurements, it still needs to be accounted for while
creating accurate models for opacity based on experimental data, and is computed as
q
2
σcond = σtemp + σp2 + σc2 , (3.57)
with σtemp , σp , and σc representing the 2σ uncertainties in opacity corresponding
to the uncertainties in temperature, pressure, and concentration, respectively. Each
of these uncertainties is calculated by halving the difference between the maximum
modeled opacity and the minimum modeled opacity, with each measurement uncer-
tainty. Since σcond dependence on the opacity and refractivity are not known while
making the measurements, this uncertainty is maintained separately from σtot . Thus,
the total 95% confidence for the measurement uncertainty is expressed in dB/km as
per Hanley et al. (2009)

p
σtot = σn 2 + σdiel 2 + σtrans 2 + σasym 2 . (3.58)
64
100
σn
90 σdiel
80 σtrans
σasym
70
Percentage Contribution
60
50
40
30
20
10
0
75 80 85 90 95 100 105 110 115
Frequency (GHz)
Figure 3.12: Percentage contribution of the different measurement uncertainties to

the total uncertainty of the W band system at room temperature (T=297 K).
The centimeter-wavelength system exhibits a maximum 2σ sensitivity in the opac-
ity ranging from 0.01 dB/km at 1.5 GHz to 0.1 dB/km at 6 GHz (Hanley et al., 2009).
For the W band and F band systems, the percentage contribution of different uncer-
tainties to the total 2σ uncertainty of a typical measurement are shown in Figures 3.12
and 3.13. The dominant factor in the total uncertainty in most cases for the W band
system is σn and that for the F band system is σtrans . For the cold temperature
measurements, σasym starts to dominate at the shorter wavelengths. The aggregate
sensitivities of the millimeter-wavelength system are shown in Figure 3.14.
65
100
σn
90 σdiel
80 σtrans
σasym
70
Percentage Contribution
60
50
40
30
20
10
0
100 105 110 115 120 125 130 135 140 145
Frequency (GHz)
Figure 3.13: Percentage contribution of the different measurement uncertainties to

the total uncertainty of the F band system at T=218 K.
1
10
Sensitivity (dB/km)
0
10
W−Band
F−Band
−1
10
80 90 100 110 120 130 140 150
Frequency (GHz)
Figure 3.14: Measured system sensitivity at room temperature (T=297 K) in the

2–4 millimeter-wavelength range.
66
CHAPTER IV
COMPRESSIBILITY OF FLUIDS
The thermodynamic and transport behavior of pure fluids and their mixtures are
controlled by the nature of the molecules and the intermolecular forces that exist
between them. The simplest theory, the theory for ideal gases, assumes that point
mass molecules with negligible volume are in constant, random motion, and elastically
collide with each other. The duration of collision is assumed to be negligible when
compared to the time between collisions and the intermolecular forces are assumed
to be negligible when compared to the kinetic energy generated by the molecular
collisions. The theory for ideal gases assumes that pressure arises because of molecular
collisions and not because of static repulsions between molecules. The common form
of the ideal gas law is
P V = nRT, (4.1)
where P is the absolute pressure, V is the volume occupied by the gas, n is the
molar density, R is the specific gas constant, and T is the absolute temperature. The
ideal gas equation may also be written as
P = ρRT, (4.2)
where ρ is the density of the gas. Gases are hardly ideal under the conditions
found in the deep planetary atmospheres. Real gases, as opposed to ideal gases, ex-
hibit compressibility effects. For real gases, the intermolecular forces are very large
and the force fields exhibited by the molecules are larger than the size of individual
67
molecules. Hence, there is indeed an action at a distance, albeit very weak. Real
gases also undergo inelastic collisions (kinetic energy is not conserved). An approxi-
mate formulation for the behavior of real gases was first proposed by van der Waals
(1873). Although the van der Waals equation of state provides a simple intuitive
way to account for real gas behavior, it is rarely accurate. Over the years, numerous
expressions have been developed to account for the behavior of real gases. To correct
for non-ideality, the simplest equation of state uses a correction factor, known as the
compressibility factor (Z):
P V = nZRT. (4.3)
The Z factor can be considered as being the ratio of the volume occupied by the
real gas to the volume occupied under the same temperature and pressure conditions
if the gas were ideal. The factor Z is a function of temperature, pressure, and gas
composition, and is often determined experimentally. If Z = 1, then the gas is
considered ideal. If Z <1, then the molecular attractive forces dominate and hence
the measured pressure (real pressure) appears to be less than the ideal pressure.
If Z >1, the molecular repulsive forces dominate and hence the measured pressure
appears to be greater than the ideal pressure.
To accurately study the thermophysical and thermochemical properties of the
jovian atmospheres, highly precise equations of state of various molecular species are
required. Pressure-explicit equations of state, such as the modified Benedict-Webb-
Ruben (mBWR) equation of state, can be used for directly deriving pressure for a
given gas density and temperature (Span, 2000). The Helmholtz free energy form
is also frequently used to accurately represent the fundamental equations of state of
pure substances. The thermodynamic fluid properties used in this work were obtained
from the REFPROP database of the National Institute of Standards and Technology
(NIST) (Lemmon et al., 2007) and subroutines were written in Matlab to access the
68
1.01
0.99
Compressibility
0.98
0.97
0.96 Hydrogen
Helium
Ammonia
Water Vapor
0.95 −2 −1 0 1
10 10 10 10
Ideal Pressure (bar)
Figure 4.1: The compressibility of various pure fluids at 500 K as a function of pres-
sure. For H2 and He, the molecular repulsive forces dominate under these conditions
and hence the measured pressure is greater than the ideal pressure. For NH3 and
H2 O, the molecular attractive forces dominate under these conditions and hence the
measured pressure is less than the ideal pressure.
database. An example of the compressibilities of hydrogen, helium, ammonia, and
water, in the 0.01–10 bar pressure range at 500 K, computed using the state of the
art equations of state for these fluids (described later in this chapter) is shown in
Figure 4.1.
In this work, the latest equations of state for pure fluids have been employed
to compute their thermodynamic properties at pressures up to 100 bar. Under the
ultra-high-pressure conditions characteristic of the very deep interiors of the jovian
planets (pressures much greater than 100 bar), in addition to the breakdown of the
ideal gas law for pure substances, Dalton’s law of partial pressures also breaks down.
Hence, in the ultra-high-pressure realm (while modeling the very deep interiors of the
jovian atmospheres), the non-ideal nature of fluid mixtures should also be included,
in addition to accounting for the non-ideality of pure fluids (Span, 2000).
69
4.1 Hydrogen
Hydrogen is the most abundant molecular species in the universe and constitutes
more than 80% of the atmospheric composition of the jovian planets (Irwin, 2003).
Depending on the relative orientation of the nuclear spin of the individual atoms in
the hydrogen molecule, diatomic hydrogen can exist in either the lower-energy sin-
glet “para” state (the nuclear spins are antiparallel) or higher-energy triplet “ortho”
state (the nuclear spins are parallel). The equilibrium ratio of orthohydrogen and
parahydrogen depends on the temperature. At very low temperatures, there is in-
sufficient thermal energy to populate higher energy states and hence hydrogen exists
exclusively in the “para” form (e.g., at 19 K, a sample of gaseous hydrogen is 99.75%
parahydrogen, Leachman et al., 2009). As temperature is increased, the higher energy
states are populated and the equilibrium shifts towards the “ortho” form. At 80 K,
the equilibrium concentration is approximately 50% parahydrogen and 50% orthohy-
drogen. At room temperature, an equilibrium distribution of 75% orthohydrogen and
25% parahydrogen is reached. This 3:1 distribution, commonly referred as “normal
hydrogen”, is also maintained at temperatures above ambient room temperature since
each of the four possible energy states will remain equally populated. Under the deep
jovian conditions, hydrogen exists in the normal form and this section contains infor-
mation about the currently accepted formulations for the thermodymanic properties
of normal hydrogen available in REFPROP (Lemmon et al., 2007).
The equation of state for normal hydrogen has been formulated using the Helmholtz
energy as the fundamental property with density and temperature as independent
variables. The Helmholtz free energy (a) is expressed as
a(T, ρ)
= α(τ, %), (4.4)
RT
where α is the reduced Helmholtz free energy, τ is the reciprocal of the reduced
70
temperature, and % is the reduced density, given as
Tc
τ= , (4.5)
T
ρ
%= , (4.6)
ρc
where the subscript c denotes a critical-point property. The critical values for
normal hydrogen are (Leachman et al., 2009)
Tc = 33.145 (K), (4.7)
Pc = 1.2964 (MPa), (4.8)
ρc = 31.26 (kg/m3 ). (4.9)
The reduced Helmholtz free energy is composed of the ideal gas component (α0 )
and the residual component (αr ) which corresponds to the intermolecular forces:
α(τ, %) = α0 (τ, %) + αr (τ, %). (4.10)
All thermodynamic properties can be calculated as the derivative of the Helmholtz
free energy and are given in Span (2000). For example, pressure can be calculated
from the Helmholtz free energy as

r
∂α
P = ρRT 1 + % , (4.11)
∂% τ
and compressibility is obtained as
∂αr

P
Z= =1+% . (4.12)
ρRT ∂% τ
The ideal gas Helmholtz free energy is expressed as
h0 s0
α0 = −1− , (4.13)
RT R
71
Table 4.1: Coefficients of the normal hydrogen ideal gas heat capacity equation.
k uk vk
1 1.616 531
2 -0.4117 751
3 -0.792 1989
4 0.758 2484
5 1.217 6859
where h0 is the ideal gas enthalpy and s0 is the ideal gas entropy. The ideal gas
enthalpy is given as
Z T
0
h = h00 + c0p dT , (4.14)
T0
and the ideal gas entropy is given as

T c0p
Z
0 ρT
s = s00 + dT − R log , (4.15)
T0 T ρ0 T0
where cp 0 is the ideal gas heat capacity, and s00 and h00 are the ideal gas entropy
and enthalpy at a reference density (ρ0 ) and temperature (T0 ). The ideal gas heat
capacity (c0p ) equation is given as

Nl v 2
X k exp(vk /T )
c0p = 2.5R + R uk , (4.16)
k=1
T [exp(vk /T ) − 1]2
and the coefficients uk and vk are listed in Table 4.2
Combining Equations 4.13-4.15, the reduced Helmholtz equation for normal hy-
drogen is given as
h00 τ s00 τ c0p τ c0p

Z Z
0 %τ0 τ 1
α = − − 1 + log − dτ + dτ , (4.17)
RTc R %0 τ R τ0 τ2 R τ0 τ
where %0 = ρ0 /ρc and τ0 = Tc /T0 .
A computationally convenient parametrized form of the ideal part of the reduced
Helmholtz free energy is given as (Leachman et al., 2009)

Nl
X
0
α = log % + 1.5 log(τ ) + a1 + a2 τ + ak log[1 − exp(bk τ )]. (4.18)
k=3
72
Table 4.2: Coefficients and parameters of the ideal part of the reduced Helmholtz
free energy equation for normal hydrogen.
k ak bk
1 -1.4579856475
2 1.888076782
3 1.616 -16.0205159149
4 -0.4117 -22.6580178006
5 -0.792 -60.0090511389
6 0.758 -74.9434303817
7 1.217 -206.9392065168
For normal hydrogen, Nl = 7 and the coefficients ak and bk are listed in Table 4.2.
The residual contribution to the reduced Helmholtz free energy is given as

l
X m
X
r di ti
α (τ, %) = Ni % τ + Ni %di τ ti exp(−%pi )
i=1 i=l+1
Xn
+ Ni %di τ ti exp[−ϕi (% − ϑi )2 − κi (τ − ςi )2 ], (4.19)
i=m+1
where the first summation is a polynomial comprising seven terms (l = 7), with
exponents di and ti on the reduced density and temperature, respectively. The sec-
ond summation consists of an exponential density component comprising two terms
(m = 9) to aid in the calculation of liquid and critical-region properties. The third
summation comprises of five modified Gaussian bell-shaped terms (n = 14) to accu-
rately model the critical region. The values of the parameters and coefficients are
given by Leachman et al. (2009) and listed in Table 4.3. The compressibility of pure
normal hydrogen computed using the reduced Helmholtz free energy as a function of
temperature and pressure is shown in Figure 4.2.
73
1.06
1.05
Compressibility
1.04
1.03
1.02
1.01
1
2
10
600
0 550
10
500
450
−2
10 400
Ideal Pressure (bar) Temperature (K)
Figure 4.2: The compressibility (Z) of pure normal hydrogen.
74
Table 4.3: Parameters and coefficients of the residual part of the reduced Helmholtz
free energy term for normal hydrogen.
i Ni ti di pi ϕi κi ςi ϑi
Polynomial
1 -6.93643 0.6844 1
2 0.01 1 4
3 2.1101 0.989 1
4 4.52059 0.489 1
5 0.732564 0.803 2
6 -1.34086 1.1444 2
7 0.130985 1.409 3
Exponential
8 -0.777414 1.754 1 1
9 0.351944 1.311 3 1
Gaussian
10 -0.0211716 4.187 2 1.685 0.171 0.7164 1.506
11 0.0226312 5.646 1 0.489 0.2245 1.3444 0.156
12 0.032187 0.791 3 0.103 0.1304 1.4517 1.736
13 -0.0231752 7.249 1 2.506 0.2785 0.7204 0.67
14 0.0557346 2.986 1 1.607 0.3967 1.5445 1.662
4.2 Helium
For pure helium, a modified Benedict-Webb-Rubin (mBWR) pressure-explicit equa-
tion of state developed by McCarty and Arp (1990) was used. The mBWR equation
of state is given as
P = ρR0 T + ρ2 (N1 T + N2 T 1/2 + N3 + N4 /T + N5 /T 2 ) + ρ3 (N6 T + N7 + N8 /T + N9 /T 2 )
+ ρ4 (N10 T + N11 + N12 /T ) + ρ5 (N13 ) + ρ6 (N14 /T + N15 /T 2 ) + ρ7 (N16 /T )
+ ρ8 (N17 /T + N18 /T 2 ) + ρ9 (N19 /T 2 ) + ρ3 (N20 /T 2 + N21 /T 3 ) exp(−Ψρ2 )
+ ρ5 (N22 /T 2 + N23 /T 4 ) exp(−Ψρ2 ) + ρ7 (N24 /T 2 + N25 /T 3 ) exp(−Ψρ2 )
+ ρ9 (N26 /T 2 + N27 /T 4 ) exp(−Ψρ2 ) + ρ11 (N28 /T 2 + N29 /T 3 ) exp(−Ψρ2 )
+ ρ13 (N30 /T 2 + N31 /T 3 + N32 /T 4 ) exp(−Ψρ2 ), (4.20)
75
Table 4.4: Coefficients of the mBWR equation of state for helium.
N1 = 0.4558980227431 × 10−4 N17 = 0.6446881346448 × 10−12
N2 = −0.1260692007853 × 10−2 N18 = 0.3298960057071 × 10−10
N3 = −0.7139657549318 × 10−2 N19 = −0.3555585738784 × 10−12
−2
N4 = 0.9728903861441 × 10 N20 = −0.6885401367690 × 10−2
N5 = −0.1589302471562 × 10−1 N21 = 0.9166109232806 × 10−2
−5
N6 = 0.1454229259623 × 10 N22 = −0.6544314242937 × 10−5
N7 = −0.4708238429298 × 10−4 N23 = −0.3315398880031 × 10−4
N8 = 0.1132915223587 × 10−2 N24 = −0.2067693644676 × 10−7
−2
N9 = 0.2410763742104 × 10 N25 = 0.3850153114958 × 10−7
N10 = −0.5093547838381 × 10−8 N26 = −0.1399040626999 × 10−10
N11 = 0.2699726927900 × 10−5 N27 = −0.1888462892389 × 10−11
−4
N12 = −0.3954146691114 × 10 N28 = −0.4595138561035 × 10−14
N13 = 0.1551961438127 × 10−8 N29 = 0.6872567403738 × 10−14
−7
N14 = 0.1050712335785 × 10 N30 = −0.6097223119177 × 10−18
N15 = −0.5501158366750 × 10−7 N31 = −0.7636186157005 × 10−17
N16 = −0.1037673478521 × 10−9 N32 = 0.3848665703556 × 10−17
R = 0.008314310
where Ψ = 1/ρ2c . The critical values for helium are (McCarty and Arp, 1990)
Tc = 5.1953 (K), (4.21)
Pc = 0.22746 (MPa), (4.22)
ρc = 69.3 (kg/m3 ) = 17.399 (mol/L). (4.23)
The coefficients used in the mBWR equation are provided in Table 4.4. The pressure
P is in MPa, density ρ is in mol/L, and temperature T in K. The compressibility of
pure helium computed using the mBWR equation as a function of temperature and
pressure is shown in Figure 4.3.
76
1.035
1.03
1.025
Compressibility
1.02
1.015
1.01
1.005
1
2
10
600
0 550
10
500
450
−2
10 400
Figure 4.3: The compressibility (Z) of pure helium.
77
4.3 Ammonia
For modeling the equation of state for pure ammonia, the formulations given by Tillner-
Roth et al. (1993) was used. The critical quantities for ammonia are taken as (Tillner-
Roth et al., 1993)
Tc = 404.45 (K), (4.24)
Pc = 11.333 (MPa), (4.25)
ρc = 225 (kg/m3 ). (4.26)
The ideal part of the reduced Helmholtz free energy is expressed as (Tillner-Roth
et al., 1993)
α0 (τ, %) = ln(%) + a01 + a02 τ − ln(τ ) + a03 τ 1/3 + a04 τ −3/2 + a05 τ −7/4 , (4.27)
where the coefficients are listed in Table 4.5.
The residual part of the reduced Helmholtz free energy for ammonia is expressed
as (Tillner-Roth et al., 1993)

5
X 10
X
r ti di
α = ai τ % + exp(−%) ai τ ti %di
i=1 i=6
17
X 21
X
2 ti di 3
+ exp(−% ) ai τ % + exp(−% ) ai τ ti %di , (4.28)
i=11 i=18
Table 4.5: Coefficients of the ideal part of the reduced Helmholtz free energy for
ammonia.
k ak
1 -15.81502
2 4.255726
3 11.47434
4 -1.296211
5 0.5706757
78
Table 4.6: Parameters and coefficients of the residual part of the reduced Helmholtz
free energy for ammonia.
i ai ti di
Polynomial
1 0.04554431 -0.5 2
2 0.7238548 0.5 1
3 0.01229470 1 4
4 -1.858814 1.5 1
5 0.2141882×10−10 3 15
Exponential
6 -0.0143002 0 3
7 0.3441324 3 3
8 -0.2873571 4 1
−4
9 0.2352589×10 4 8
10 -0.03497111 5 2
11 0.02397852 3 1
12 0.001831117 5 8
13 -0.04085375 6 1
14 0.2379275 8 2
15 -0.03548972 8 3
16 -0.1823729 10 2
17 0.02281556 10 4
18 -0.006663444 5 3
19 -0.008847486 7.5 1
20 0.002272635 15 2
21 -0.0005588655 30 4
where the first summation is a polynomial comprising five terms, with exponents
di and ti on the reduced density and temperature, respectively. The remaining sum-
mation terms consists of exponential density components to aid in the calculation of
liquid and critical-region properties. The values of the parameters and coefficients
are given by Tillner-Roth et al. (1993) and listed in Table 4.6. The compressibility
of pure ammonia computed using the reduced Helmholtz free energy equation as a
function of temperature and pressure is shown in Figure 4.4.
79
1.1
0.9
Compressibility
0.8
0.7
0.6
0.5
0.4
2
10
600
0 550
10
500
450
−2
10 400
Figure 4.4: The compressibility (Z) of pure ammonia.
80
4.4 Water
The latest formulation for the equation of state for water employs the reduced Helmholtz
free energy equation. The critical quantities for water are taken as (Wagner and Pruß,
2002)
Tc = 647.096 (K), (4.29)
Pc = 22.064 (MPa), (4.30)
ρc = 322 (kg/m3 ). (4.31)
The ideal part of the reduced Helmholtz free energy for water is given as (Wagner
and Pruß, 2002)

8
X
α0 = ln(%) + n01 + n02 τ + n03 ln(τ ) + n0i ln(1 − exp(−γi0 τ )). (4.32)
i=4
The coefficients and parameters of the ideal part of the reduced Helmholtz free
energy for water are listed in Table 4.7.
The residual contribution to the reduced Helmholtz free energy is given as

l
X m
X
αr (τ, %) = Ni %di τ ti + Ni %di τ ti exp(−%pi )
i=1 i=l+1
n p
X X
di ti 2 2
+ Ni % τ exp[−ϕi (% − ϑi ) − κi (τ − ςi ) ] + Ni Υbi %ψ, (4.33)
i=m+1 i=n+1
Table 4.7: Coefficients and parameters of the ideal part of the reduced Helmholtz
free energy for water.
i n0i γi0
1 -8.32044648201 -
2 6.6832105268 -
3 3.00632 -
4 0.012436 1.28728967
5 0.97315 3.53734222
6 1.27950 7.74073708
7 0.96956 9.24437796
8 0.24873 27.5075105
81
where the first summation is a polynomial comprising seven terms (l = 7), with
exponents di and ti on the reduced density and temperature, respectively. The second
summation consists of an exponential density component comprising 45 terms (m =
51), the third summation comprises of three modified Gaussian bell-shaped terms
(n = 54), and the fourth summation comprises of two critical terms (p = 56). The
critical term components are given as
Υ = Θ2 + Bi [(% − 1)2 ]ai , (4.34)
Θ = (1 − τ ) + Ai [(% − 1)2 ]1/2βi , (4.35)
ψ = exp(−Ci [% − 1]2 − Di [τ − 1]2 ). (4.36)
The values of the parameters and coefficients are given by Wagner and Pruß (2002)
and listed in Table 4.8. The compressibility of pure water computed using the reduced
Helmholtz energy formulation as a function of temperature and pressure is shown in
Figure 4.5.
Table 4.8: Parameters and coefficients of the residual
part of the reduced Helmholtz free energy for water.
i ci di ti Ni
Polynomial
1 - 1 -0.5 0.12533547935523×10−1
2 - 1 0.875 0.78957634722828×101
3 - 1 1 -0.87803203303561×101
4 - 2 0.5 0.31802509345418
5 - 2 0.75 -0.26145533859358
6 - 3 0.375 -0.78199751687981×10−2
7 - 4 1 0.88089493102134×10−2
Exponential
8 1 1 4 -0.66856572307965
82
i ci di ti Ni
9 1 1 6 0.20433810950965
10 1 1 12 -0.66212605039687×10−4
11 1 2 1 -0.19232721156002
12 1 2 5 -0.25709043003438
13 1 3 4 0.16074868486251
14 1 4 2 -0.40092828925807×10−1
15 1 4 13 0.39343422603254×10−6
16 1 5 9 -0.75941377088144×10−5
17 1 7 3 0.56250979351888×10−3
18 1 9 4 -0.15608652257135×10−4
19 1 10 11 0.11537996422951×10−8
20 1 11 4 0.36582165144204×10−6
21 1 13 13 -0.13251180074668×10−11
22 1 15 1 -0.62639586912454×10−9
23 2 1 7 -0.10793600908932
24 2 2 1 0.17611491008752×10−1
25 2 2 9 0.22132295167546
26 2 2 10 -0.40247669763528
27 2 3 10 0.58083399985759
28 2 4 3 0.49969146990806×10−2
29 2 4 7 -0.31358700712549×10−1
30 2 4 10 -0.74315929710341
31 2 5 10 0.47807329915480
32 2 6 6 0.20527940895948×10−1
33 2 6 10 -0.13636435110343
34 2 7 10 0.14180634400617×10−1
35 2 9 1 0.83326504880713×10−2
36 2 9 2 -0.29052336009585×10−1
37 2 9 3 0.38615085574206×10−1
38 2 9 4 -0.20393486513704×10−1
39 2 9 8 -0.16554050063734×10−2
40 2 10 6 0.19955571979541×10−2
41 2 10 9 0.15870308324157×10−3
42 2 12 8 -0.16388568342530×10−4
43 3 3 16 0.43613615723811×10−1
44 3 4 22 0.34994005463765×10−1
45 3 4 23 -0.76788197844621×10−1
83
i ci di ti Ni
46 3 5 23 0.224462773320×10−1
47 4 14 10 -0.62689710414685×10−4
48 6 3 50 -0.55711118565645×10−9
49 6 6 44 -0.19905718354408
50 6 6 46 0.31777497330738
51 6 6 50 -0.11841182425981
Gaussian
i ci di ti Ni ϕi κi ςi ϑi
52 - 3 0 -0.31306260323435×102 20 150 1.21 1
53 - 3 1 0.31546140237781×102 20 150 1.21 1
54 - 3 4 -0.25213154341695×104 20 250 1.25 1
Critical
i ci di ti Ni Ci Di Ai βi
55 3.5 0.85 0.2 -0.14874640856724 28 700 0.32 0.3
56 3.5 0.95 0.2 0.31806110878444 32 800 0.32 0.3
84
1
0.8
Compressibility
0.6
0.4
0.2
0
2
10
600
0 550
10
500
450
−2
10 400
Figure 4.5: The compressibility (Z) of pure water.
85
CHAPTER V
MEASUREMENTS AND MODEL
One of the primary objectives of this work has been to develop a consistent formal-
ism for the centimeter- and millimeter-wavelength opacity of ammonia under jovian
atmospheric conditions. The model is based on an extensive set of measurements of
the opacity of ammonia conducted as part of this work and previous measurements
conducted by Hanley et al. (2009). A dry jovian adiabatic temperature-pressure
(TP) profile overlayed with the TP distribution of the centimeter- and millimeter-
wavelength measurement points that were used in the model development and/or
assessment of the model performance are shown in Figure 5.1.
5.1 Millimeter-Wavelength Ammonia Opacity Measurements
Over 1000 high-accuracy measurements of the 2–4 mm-wavelength absorptive prop-
erties of pure ammonia and ammonia pressure-broadened by hydrogen and helium
have been conducted using the millimeter-wavelength measurement system. Cer-
tified ultra-high-purity (UHP grade) ammonia and premixed hydrogen/helium gas
cylinders from Airgas, Inc. were used for the experiments and certified UHP grade
argon and carbon dioxide gas cylinders were used for dielectric matching experi-
ments. In the premixed hydrogen/helium cylinder, the helium mixing ratio was
(13.6±0.272)% and the remainder was hydrogen. This is approximately the helium
mole fraction at Jupiter measured by von Zahn et al. (1998). A total of 718 data
points of the opacity of ammonia in a hydrogen/helium environment and 295 data
points of the opacity of pure ammonia were measured with each data point uniquely
representing a combination of pressure, temperature, mixing ratio, and frequency.
86
Table 5.1: Listing of all experimental conditions for the 2–4 mm-wavelength ammonia
opacity measurements conducted using the Fabry–Perot resonator as part of this work.
Experiment Dates Temperature Pressure NH3 Mole Frequency
(K) (bar) Fraction (GHz)
06/2007–10/2007 295 1–3 0.02–0.04 75–120
12/2007–02/2008 222 1–3 0.02 75–120
08/2008–09/2008 295 1–3 0.02–0.04 110–150
10/2008–11/2008 295 0.1–1 1 110–150
11/2008–12/2008 295 0.1–1 1 75–120
01/2009–02/2009 222 0.1–3 0.02–1 110–150
02/2009–03/2009 222 0.1–3 0.02–1 75–120
03/2009–04/2009 205 0.1–3 0.02–1 110–150
04/2009–05/2009 205 0.1–3 0.02–1 75–120
Pure ammonia opacity measurements were made to accurately characterize its self-
broadening parameters. Table 5.1 lists the measurements taken along with the ex-
periment dates. The processed data are provided online (N H3 mmwavelength.dat)
at http://users.ece.gatech.edu/~psteffes/palpapers/models.html.
87
−2
10
TP Profile of Jupiter
1.5−27 GHz measurements (Hanley et al., 2009)
22−40 GHz measurements (Hanley et al., 2009)
75−150 GHz measurements (This Work)
−1
1.5−6 GHz measurements (This Work)
10
Pressure (bar)
0
10
1
10
2
10
100 200 300 400 500 600 700
Temperature (K)
Figure 5.1: Dry jovian adiabatic temperature-pressure (TP) profile along with the
TP space measurement points used in the model development and/or evaluation of
the new model performance. Red crosses are the 1.5–27 GHz cavity resonator TP
space points measured by Hanley et al. (2009), blue triangles are the 22–40 GHz FPR
TP space points measured by Hanley et al. (2009), black circles are the 75–150 GHz
FPR TP space points measured as part of this work, and black asterisks are the 1.5–6
GHz high-pressure TP space points measured as part of this work.
88
5.2 High-Pressure Centimeter-Wavelength Ammonia Opac-
ity Measurements
This measurement process involved an extensive series of measurements of the opacity
of ammonia under simulated deep jovian conditions at pressures up to 100 bar and
temperatures up to 500 K. A total of 1176 measurements of the 5–20 cm-wavelength
absorptive properties of pure ammonia and ammonia broadened by hydrogen and/or
helium have been made using the high-pressure centimeter-wavelength measurement
system. Certified UHP grade ammonia, helium, hydrogen, and argon gas cylinders
from Airgas, Inc. were used for the experiments and dielectric matching. Each
measurement sequence first involved the addition of gaseous ammonia to an evac-
uated chamber up to the desired pressure, and pure ammonia measurements were
made to accurately characterize its self-broadening parameters. In this work, pure
ammonia measurements were made in the 0.046–0.133 bar pressure range. Gaseous
helium was then added to the system up to the desired pressure and ammonia/helium
mixture measurements were made to accurately characterize the helium-broadening
parameters. The ammonia/helium mixture measurements were made in the 7–20
bar pressure range. Gaseous hydrogen was then added to the pressure vessel, in
10–20 bar pressure increments, and the ammonia/helium/hydrogen mixture mea-
surements were made at pressures up to 100 bar. A few measurements of the ammo-
nia/hydrogen mixture opacity at pressures up to 100 bar were also made to accurately
characterize the hydrogen-broadening parameters. A total of 180 data points of the
opacity of pure ammonia, 156 data points of the opacity of ammonia/helium mix-
ture, 120 data points of the opacity of ammonia/hydrogen mixture, and 720 data
points of the opacity of ammonia/helium/hydrogen mixture were measured, with
each data point uniquely representing a combination of pressure, temperature, mixing
ratio, and frequency. Table 5.2 lists the measurements taken along with the exper-
iment dates. The pressure values provided in the table correspond to the measured
89
Table 5.2: Listing of all experiment sequences of the 5–20 cm-wavelength ammonia
opacity measurements conducted using the high-pressure system as part of this work.
Experiment Dates Temperature Maximum NH3 Pressure He Pressure
(K) Pressure (bar) (bar)
(bar)
09/2009 376.0 94.66 0.0899 12.53
09/2009 375.5 95.51 0.0798 13.63
10/2009 375.2 96.11 0.0580 14.0
10/2009 373.2 95.59 0.1096 11.92
11/2009 446.5 95.34 0.1 7.35
12/2009 446.8 93.54 0.0824 0
12/2009 446.6 94.01 0.0805 19.07
01/2010 322.6 92.66 0.0838 16.55
02/2010 332.9 91.73 0.0462 15.58
02/2010 401.0 96.17 0.0813 11.98
03/2010 452.1 97.6 0.133 12.42
09/2010 503.6 94.17 0.0850 16.64
10/2010 502.8 89.99 0.0662 9.40
11/2010 502.6 96.47 0.1046 0
11/2010 502.1 98.44 0.1248 10.75
maximum pressure readings of the DPI pressure gauges or the pressure transducer.
The processed data are provided online (N H3 cmwavelength highpressure.dat) at
http://users.ece.gatech.edu/~psteffes/palpapers/models.html.
5.3 Data Fitting
One of the primary goals of this work has been to develop a model, based on
highly accurate laboratory measurements, that can be used to estimate the opac-
ity of ammonia in a hydrogen/helium atmosphere under diverse conditions in the
frequency/temperature/ pressure/concentration (fTPC) space. For the data fitting
process, the FPR measurements of the 75–150 GHz opacity of ammonia and the high-
pressure cavity resonator measurements of the 1.5–6 GHz opacity of ammonia, made
as part of this work, along with the cavity resonator measurements of the 1.5–27 GHz
opacity of ammonia made by Hanley et al. (2009) were utilized so as to obtain a con-
sistent model in the fTPC space. The FPR measurements of the 22–40 GHz opacity
90
of ammonia made by Hanley et al. (2009) were not used in the model development
process, but instead were used to evaluate the model performance. The method used
for data fitting was a Levenberg–Marquardt optimization technique (Levenberg, 1944;
Marquardt, 1963) with a minimization function
√
DW × (αmeasured − αmodel )
χ= , (5.1)
σmeasured
where DW is the data weight assigned to each data point, αmeasured , αmodel , and
σmeasured are the measured opacity, modeled opacity of the model under optimiza-
tion, and measured uncertainty in opacity, respectively. The measured uncertainty in
opacity is calculated as
σmeasured = σtot + σcond , (5.2)
where σtot is the total measurement uncertainty and σcond is the uncertainty due to
measurement conditions. The sum of squared value of the χ function was minimized
multiple times using random input seed values until a convergent solution was found.
The data weight is given as (Hanley et al., 2009)
1 1 1 1
DW = + + + , (5.3)
nf nT nP nC
where nf , nT , nP , and nC represent the number of measurements conducted at
each frequency, temperature, pressure, and gas concentration range, in the four-
dimensional fTPC space. The approach used divides the data points into roughly
equally spaced bins that span the fTPC space and each data point is scaled with its
data weight so as to prevent the accuracy of the derived model from being skewed
toward the most often measured conditions. The data fitting process was done in two
stages in this work. In the first stage, the free parameters for the rotational and ν2
roto-vibrational transitions were estimated using a measurement database consisting
91
of the 75–150 GHz FPR measurements made as part of this work and the 1.5–27 GHz
cavity resonator measurements conducted by Hanley et al. (2009). Nominal values of
the free parameters for the inversion transitions were also obtained in this stage and
later revised during the second stage of data fitting. In the second stage, the 1.5–6
GHz high-pressure data points were added to the measurement database and the free
parameters for the inversion transitions were estimated.
In the first stage of data fitting, the 75–150 GHz FPR measurements and the
1.5–27 GHz cavity resonator measurements were used to create the fTPC space. The
breakdown of the fTPC space is listed in Table 5.3. The free parameters for pure
ammonia were estimated before those for hydrogen/helium. The pure ammonia mea-
surements were divided into two groups with one group constituting the data points
with f ≤100 GHz (group I) and the other group constituting the data points with
f > 100 GHz (group II). For the data points in group I, the inversion and rotational
transitions of ammonia contribute significantly to the measured opacity and the con-
tribution from the ν2 roto-vibrational transitions is negligible. The ν2 roto-vibrational
transitions contribute significantly to the measured opacity when f > 100 GHz. Data
fitting was achieved for group I by optimizing the ammonia free parameters for the
inversion and rotational transitions. After this step, group II data points were used
for optimization of the ammonia free parameters for the ν2 roto-vibrational transi-
tions. Subsequent optimization steps involved assigning the values obtained from the
previous optimization steps to the ammonia free parameters and optimizing only the
hydrogen and helium free parameters using the ammonia/hydrogen/helium mixture
measurements. The mixture measurements were subdivided into two groups: group
III comprising the data points with f ≤ 100 GHz and group IV comprising the data
points with f > 100 GHz. The hydrogen and helium free parameters for the inversion
and rotational transitions were optimized using group III data points and the hydro-
gen and helium free parameters for the ν2 roto-vibrational transitions were optimized
92
Table 5.3: The breakdown in the fTPC space of the measurement database consist-
ing of the 75–150 GHz FPR measurements (this work) and the 1.5–27 GHz cavity
resonator measurements (Hanley et al., 2009) used in the first stage of optimization.
Frequency nf Temperature nT Pressure nP Concentration nC
range (GHz) range (K) range range (%)
(bar)
f<6 874 T<210 291 P<0.5 459 C<0.5 380
6≤f<27 557 210≤T<230 650 0.5≤P<1 213 0.5≤C<1 578
75≤f<100 329 230≤T<300 1195 1≤P<2 533 1≤C<4 481
100≤f< 120 393 300≤T<380 112 2≤P<5 761 4≤C<10 557
120≤f<150 291 380≤T<450 196 5≤P≤12.5 478 10≤C≤100 448
using group IV data points. This procedure was repeated several times with different
input seed values for the free parameters, until a convergent solution was obtained.
The final values for the free parameters for the rotational and ν2 roto-vibrational tran-
sitions of ammonia were obtained after the first stage of optimization. The nominal
values for the free parameters for the inversion transitions that were obtained after
the first stage of optimization were revised during the second stage of optimization.
The nominal values that were obtained for the inversion transitions of ammonia
after the first stage of optimization, could in fact be used to estimate the centimeter-
wavelength ammonia opacity under jovian conditions at pressures up to 12 bar (De-
varaj et al., 2011). However, these nominal values cannot be used under the deep
jovian conditions because they were not optimized to perform under those conditions.
To obtain a consistent formalism that can operate under diverse pressure conditions,
two sets of free parameters for the inversion transitions of ammonia were estimated
in the second stage of optimization. One set of parameters is used for modeling am-
monia opacity at pressures less than 15 bar and a second set of parameters is used
for modeling the ammonia opacity at pressures greater than 15 bar and up to 100
bar. It was possible to obtain a consistent model that performs well in both the
low-pressure and the high-pressure regimes, by using two sets of parameters for the
inversion transitions and incorporating a pressure-dependent switch.
93
The fTPC space for the second stage of data fitting includes the 1.5–6 GHz
high-pressure measurements in addition to the 1.5–27 GHz measurements conducted
by Hanley et al. (2009), and the 75–150 GHz FPR measurements conducted as part
of this work. The breakdown in the fTPC space of the measurement database used
in the second stage of optimization is listed in Table 5.4. In this stage of data fitting,
only the free parameters for the inversion transitions were optimized. To best fit the
measured data, two sets of parameters for the inversion transitions were derived, and a
pressure-dependent switch was included in the model to assure consistent results over
the wide pressure range (a few bar to hundreds of bar) characteristic of the middle
and deep tropospheres of the jovian planets. This was, however, not required for the
rotational and the ν2 roto-vibrational transitions of ammonia since these transitions
occur at very high frequencies (the first rotational transition of ammonia occurs at 572
GHz and the first strong ν2 inversion transition of ammonia occurs at 140.14 GHz)
and the jovian atmospheric layers that contribute to the emission at these frequencies
have pressures less than a few bar (upper and middle tropospheres of the jovian plan-
ets). For example, the atmospheric layers of Jupiter that contribute to the emission
at 140 GHz have pressures between 0.5 and 3 bar, whereas at 1 GHz, the contribution
is from the atmospheric layers with pressures between 20 and 300 bar. Hence, it was
sufficient to derive two sets of parameters (low-pressure and high-pressure) for the
inversion transitions alone.
In the second stage of data fitting, the measurements were split into two groups
(group V and group VI), with group V consisting of the data points with P <= 15
bar and group VI consisting of data points with P > 15 bar. The pure ammonia
measurements were used to obtain the ammonia free parameters for the inversion
transitions. Since the pure ammonia measurements were made under low-pressure
conditions (P < 0.8 bar), only one set of parameters were estimated. The ammo-
nia/helium mixture measurements with P <= 15 bar (group V) were used to obtain
94
Table 5.4: The breakdown in the fTPC space of the measurement database consisting
of the 75–150 GHz FPR measurements (this work), 1.5–27 GHz cavity resonator
measurements (Hanley et al., 2009), and the 1.5–6 GHz high-pressure measurements
(this work) used in the second stage of optimization.
Frequency nf Temperature nT Pressure nP Concentration nC
range range (K) range range (%)
(GHz) (bar)
f<5 1340 T<230 941 P<1 852 C<0.1 208
5≤f<10 847 230≤T<310 1195 1≤P<3 914 0.1≤C<0.5 1000
10≤f<27 420 310≤T<385 592 3≤P<12 867 0.5≤C<1 710
75≤f<100 329 385≤T< 460 592 12≤P<40 363 1≤C<10 1074
100≤f<150 684 460≤T<510 300 40≤P≤100 624 10≤C≤100 628
the low-pressure helium-broadening free parameters for the inversion transitions and
the ammonia/helium mixture measurements with P > 15 bar (group VI) were used
to obtain the high-pressure helium-broadening free parameters for the inversion tran-
sitions. The ammonia/hydrogen mixture measurements with P <= 15 bar (group
V) were used to obtain the low-pressure hydrogen-broadening free parameters for the
inversion transitions and the ammonia/hydrogen mixture measurements with P >
15 bar (group VI) were used to obtain the high-pressure hydrogen-broadening free
parameters for the inversion transitions. After this step, the low-pressure and high-
pressure inversion transition parameters were further adjusted by optimizing over
the ammonia/helium/hydrogen mixture measurements. This procedure was repeated
several times with different input seed values for the free parameters, until convergent
solutions were obtained.
5.4 New Consistent Ammonia Absorption Formalism
The absorption from a collisionally broadened gas can be expressed as
X
α= Aj π∆νj Fj (ν, ν(0,j) , . . .) (cm−1 ), (5.4)
j
where for the line j, Aj is the absorption at the line center in cm−1 , ∆ν j is the
95
linewidth (half width at half max) in cm−1 , F j (ν, ν0,j , . . .) is the lineshape function
in cm, ν(0,j) is the frequency at the line center in cm−1 , and ν is the frequency of the
incident electromagnetic wave in cm−1 (see, e.g., Townes and Schawlow, 1955).
The absorption at each line center is calculated using the line intensity information
from the latest JPL catalog as per Pickett et al. (1998)
nIj (T )
Aj = (cm−1 ), (5.5)
π∆νj
where n is the number density of the gas in molecules/cm3 , I j (T ) is the intensity
of the line in cm−1 /(molecule/cm2 ) at temperature T , and with ∆νj expressed in
cm−1 . The number density n is calculated as
0.1 × Pideal
n= , (5.6)
kB T
where Pideal is the ideal partial pressure of the gas under consideration in bar, T
is the temperature in K, and kB = 1.38 × 10−23 J/K (Boltzmann’s constant). The
ideal partial pressure of a gas is computed from its density and temperature. In this
work, the density of a gas is computed from its equation of state and further details
are provided in Chapter 4. The ideal pressure (in bar) is computed as,
R0
Pideal = ρ T, (5.7)
M
where ρ is the density in g/m3 , R0 is the ideal gas constant for the gas, M is the
molecular mass of the gas in g/mol and T is the measured temperature in K. The

−5 m3 bar
value for R0 used in this work is equal to 8.314472 × 10 Kmol
for all the gases
except helium. For helium, the equation of state requires the use of the older value

−5 m3 bar
8.31431 × 10 Kmol
. The line intensity at the measurement temperature (T ) is
calculated as
96
η+1
T0

1
− T1 E(l,j) (hc/kB )
Ij (T ) ≈ Ij (T0 ) exp T0
, (5.8)
T
where Ij (T0 ) is the intensity of the line at the reference temperature T0 , E(l,j) is
the lower state energy of transition in cm−1 , c is the speed of light in cm/s, h is
the Planck’s constant in Jsec, and η is the temperature dependence and is ≈ 3/2 for
non-linear and symmetric-top molecules such as ammonia, and 1 for linear molecules.
The values of ν(0,j) , Ij (T0 ), and E(l,j) at the reference temperature are taken from the
latest JPL spectral line catalog (Pickett et al., 1998) (Version 5, September 2010).
The values of Ij (T0 ) given in the JPL catalog have units of log10 (nm2 MHz) and must
be taken as the exponent of 10 and then divided by 2.99793458 × 1018 to be converted
to cm−1 /(molecule/cm2 ). The linewidth for a gas mixture is calculated by
ξij
X T0
∆νj = o
∆ν(i,j) P(i,ideal) (cm−1 ), (5.9)
i
T
o
where for the gas i and line j, ∆ν(i,j) is the line broadening parameter in cm−1 /bar,
P(i,ideal) is the ideal partial pressure of the gas in bar (computed using Equation 5.7),
and ξij is the temperature dependence of the line broadening parameter. The tem-
perature dependence is calculated as (e.g., Townes and Schawlow, 1955)
∆νj ∝ T −(m+1)/2(m−1) = T −ξ , (5.10)
where 1 < m < ∞. For neutral gases, m = 3 is a lower limit and hence 0.5 < ξ
< 1.0.
5.4.1 Line Parameters
The JPL spectral line catalogs for NH3 and NH3 − ν2 have been recently updated (Yu
et al., 2010a,b,c). The NH3 catalog has a total of 1716 transitions that include
97
415 inversion transitions in the 0.3–220 GHz frequency range and 1301 rotational
transitions in the 0.3–20 THz frequency range. The NH3 − ν2 catalog has a total of
4198 roto-vibrational transitions in the 0.02–46 THz frequency range. A total of 5914
transitions are used in the new opacity formalism for ammonia.
For the development of the new ammonia opacity model described in this work,
different lineshapes were investigated in order to obtain the best fit for the mea-
surements. By using a modified Ben–Reuven lineshape (Ben-Reuven, 1966) for the
inversion transitions, and a modified Gross lineshape (Gross, 1955) for the rotational
and ν2 roto-vibrational transitions, it was possible to achieve a robust fit for the
measurements. The lineshapes rely on the knowledge of the linewidths of the gases
being studied. Laboratory measurements of the self- and foreign-gas-broadening pa-
rameters of various line transitions were used where those data were available. When
laboratory measurements of the broadening parameters were not available, those pa-
rameters were made free variables in the optimization process and the values closest to
the average of the available measured broadening parameters that best fit the opacity
measurements described in this work were used.
5.4.2 Ammonia Opacity Formalism
The new model for the hydrogen- and helium-broadened opacity of ammonia uses a
modified Ben–Reuven lineshape for the inversion transitions and a modified Gross
lineshape for the rotational and ν2 roto-vibrational transitions. The cumulative opac-
ity is calculated by
α = (αinv + αrot + αν2 ) ×434294.5 (dB/km), (5.11)
where αinv , αrot , and αν2 are the opacities from the inversion, rotational, and ν2
transitions, respectively, in cm−1 . The multiplication factor of 434294.5 converts the
total absorptivity from cm−1 to dB/km.
98
The opacity from the inversion transitions is calculated using a modified Ben–
Reuven lineshape in a fashion similar to that described by Hanley et al. (2009) but
with different sets of values for the model constants than those used by Hanley.
Two sets of model constants were obtained by optimizing the free parameters for the
inversion transitions. When P <= 15 bar, the low-pressure model constants are used
and when P > 15 bar, the high-pressure model constants are used. The opacity from
the inversion lines is given as
η+1 X 2
0.1Dinv PN H3 2 T0 ν

1 1 hc
− E
αinv = × Ij (T0 ) exp T0 T (l,j) kB
kB T π T j
ν(0,j)
(γj − ζj )ν 2 + (γj + ζj )[(ν(0,j) + δj )2 + γj2 − ζj2 ]

(cm−1 ),
[ν 2 − (ν(0,j) + δj )2 − γj2 + ζj2 ]2 + 4ν 2 γj2
(5.12)
where for the inversion line j, ν(0,j) , γj , ζj , and δj are the center frequency,
linewidth, coupling parameter, and shift parameter, respectively, in cm−1 , and Dinv
is a unitless scale factor. The frequency, linewidth, coupling, and shift parameters
are nominally expressed in GHz and should be converted to cm−1 before they are
used in the equation for opacity. (Note that 1 cm−1 ∼ 30 GHz.) The linewidth
and coupling parameter of the lines are calculated by summing the contribution from
different gases and are given by
ΓH2 ΓHe ΓN H3

300 300 295
γj = γH2 PH2 + γHe PHe + γN H3 γ(0,j) PN H3 (GHz),
T T T
(5.13)
ZH2 ZHe ZN H3

300 300 295
ζj = ζH2 PH2 + ζHe PHe + ζN H3 γ(0,j) PN H3 (GHz),
T T T
(5.14)
99
where for the inversion line j and i = H2 , He, and NH3 , γi and ζi are constant
scale terms, and Γi and Zi represent the constant temperature dependences of the
broadening of each of the gases, Pi are the ideal partial pressures in bar, and γ(0,j) are
the self-broadening linewidths of the inversion transitions of ammonia in MHz/Torr.
The units of γ(0,j) remain in MHz/Torr and the conversion to GHz/bar is incorporated
in the scale terms γN H3 and ζN H3 in Equations 5.13 and 5.14. The values for γ(0,j)
are from the calculations of Poynter and Kakar (1975) assuming a T0 of 295 K. For
the lines with center frequencies below 7.2 GHz and J > 16, where J represents the
total angular momentum vector for the ammonia molecule, γ0 is expressed as
K
γ0 (J, K) = 25.923 p (MHz/Torr), (5.15)
J(J + 1)
where K is the projection of J onto the molecular axis. The γ(0,j) of all the other
inversion lines which are not listed by Poynter and Kakar (1975) and whose J < 16
or center frequency > 7.2 GHz are assigned a constant value equal to the average of
the γ0 of the lines calculated by Poynter and Kakar. The pressure shift parameter is
calculated by
δj = d × γj (GHz), (5.16)
where d is an empirically derived constant. All the inversion transitions are as-
signed the same set of model constants (either low-pressure or high-pressure), even
though each line transition behaves differently (see, e.g., Hanley et al., 2009). The
equation for computing the opacity from the inversion transitions has 14 free parame-
ters. Two sets of free-parameters were empirically derived by data fitting (Tables 5.5
and 5.6) and the appropriate set of values should be used for calculating the ammonia
opacity for a given pressure.
100
Table 5.5: Values of the low-pressure inversion model constants used for computing
the H2 /He-broadened NH3 absorptivity when P <= 15 bar.
i=H2 i=He i=NH3
γi 1.7465 0.9779 0.7298
Γi 0.8202 1 1
ζi 1.2163 0.0291 0.5152
Zi 0.8873 0.8994 2/3
d -0.0627
Dinv 0.9862
Table 5.6: Values of the high-pressure inversion model constants used for computing
the H2 /He-broadened NH3 absorptivity when P > 15 bar.
i=H2 i=He i=NH3
γi 1.6361 0.4555 0.7298
Γi 0.8 0.5 1
ζi 1.1313 0.1 0.5152
Zi 0.6234 0.5 2/3
d 0.2
Dinv 1.3746
The opacity from the rotational transitions is calculated using a modified Gross
lineshape and is given as
η+1 X
0.1Drot PN H3 1 T0 ν

1
− T1 E(l,j) khc
αrot = × Ij (T0 ) exp T0 B
kB T π T j
ν(0,j)
 
4νν(0,j) ∆νj
(cm−1 ),
 
 2 
2
ν(0,j) − ν 2 + 4ν 2 ∆νj2
(5.17)
where for the rotational line j, ν(0,j) is the frequency of transition, ∆νj is the
linewidth parameter, and Drot is an empirically derived unitless scale factor. The
linewidth parameter is given as
101
Table 5.7: Values of the model constants of the new model used for computing the
H2 /He-broadened NH3 absorptivity from the rotational transitions.
i=H2 i=He i=NH3
ci 0.2984 0.75 3.1789
ξi 0.8730 2/3 1
Drot 2.4268
ξH2 ξ
300 300 He
∆νj = cH2 ∆ν(H2 ,j) PH2 + cHe ∆ν(He,j) PHe
T T
ξ
300 N H3
+ cN H3 ∆ν(N H3 ,j) PN H3 (GHz), (5.18)
T
where for i = H2 , He, and NH3 , ci and ξi are the empirically derived model
constants, Pi are the ideal partial pressures in bar, and ∆ν(i,j) are the broadening
parameters in GHz/bar. The units of ∆νj should be converted from GHz to cm−1
before it is used in the equation for computing the opacity from the rotational lines
(Equation 5.17). The self-broadened linewidth for the J = 1 ← 0 transition of the
ammonia molecule is assigned the value measured by Belov et al. (1983), and the
hydrogen- and helium-broadened linewidths for the J = 1 ← 0 transition are assigned
values measured by Bachet (1973). The self- and hydrogen-broadened linewidths
for each of the other rotational lines are assigned values measured by Brown and
Peterson (1994). The linewidths for the lines that were not measured by Brown
and Peterson were assigned values computed using their extrapolation formula. The
helium-broadened linewidths are assigned the values computed using the formula
given by Pine et al. (1993). The empirically derived model constants for the rotational
transitions are listed in Table 5.7.
The opacity from the ν2 roto-vibrational transitions is calculated using a modified
102
Gross lineshape and is given as
η+1 X
0.1Dν2 PN H3 1 T0 ν

1 1 hc
− E
αν2 = × Ij (T0 ) exp T0 T (l,j) kB
kB T π T j
ν(0,j)
 
4νν(0,j) ∆ν −1
 (cm ),
 
 2
2
ν(0,j) − ν 2 + 4ν 2 ∆ν 2
(5.19)
where, for the ν2 roto-vibrational line j, ν(0,j) is the frequency of transition, ∆ν is
the linewidth parameter, and Dν2 is an empirically derived unitless scale factor. The
linewidth parameter is given by
ξH2 ξHe ξN H3

300 300 300
∆ν = ∆νH2 PH2 + ∆νHe PHe + ∆νN H3 PN H3 (GHz),
T T T
(5.20)
where for i = H2 , He, and NH3 , ξi are the empirically derived temperature coef-
ficients, Pi are the ideal partial pressures in bar, and ∆νi are the empirically derived
broadening parameters for the ν2 transitions in GHz/bar. The units of ∆ν should
be converted from GHz to cm−1 before it is used in the equation for computing the
opacity (Equation 5.19). The self-broadening parameter of the strongest ν2 transition
in the millimeter-wavelength region (ν0 = 140.14 GHz) was theoretically calculated
by Belli et al. (1997). However, the theoretically calculated value of 13.72 GHz/bar
was too large to fit the opacity measurements. The self-broadening parameter of the
466 GHz ν2 transition was measured by Belov et al. (1982). The self-, hydrogen-, and
helium-broadening parameters of most of the ν2 transitions have not been measured.
Hence, the broadening parameters were made free variables during model optimiza-
tion and empirically derived constant values are used for all lines. The empirically
derived model constants for the ν2 roto-vibrational transitions are listed in Table 5.8.
The values of the model constants listed in this paper were optimized to the latest
JPL spectral line catalogs (Pickett et al., 1998) (Version 5, September 2010). The
103
Table 5.8: Values of the model constants of the new model used for computing the
H2 /He-broadened NH3 absorptivity from the ν2 roto-vibrational transitions.
i=H2 i=He i=NH3
∆vi (GHz/bar) 1.4 0.68 9.5
ξi 0.73 0.5716 1
Dν2 1.1206
frequency, line intensity, and lower state energy of the various transitions of ammonia
given in the latest JPL catalogs along with the self- and hydrogen-/helium- broad-
ening parameters for these transitions are provided online (ammonia inversion.dat,
ammonia rotational.dat, and ammonia rotovibrational.dat) at http://users.ece.
gatech.edu/~psteffes/palpapers/models.html. Software for running the model
in Matlab
R
is also provided online (nh3devarajstef f esmodel.m). This model can be
used to estimate the opacity of ammonia under jovian conditions in the centimeter-
wavelength range at pressures up to 100 bar and temperatures in the 200 to 500 K
range and in the millimeter-wavelength range at pressures up to 3 bar and tempera-
tures in the 200 to 300 K range.
5.5 Model Performance
Assessing the performance of the model with respect to the measured data provides
insights into the goodness of fit of the model and the conditions of its effectiveness.
The model was compared with the 1431 data points of the 1.5–27 GHz opacity of
ammonia and the 250 data points of the 22–40 GHz opacity of ammonia measured
by Hanley et al. (2009), in addition to the 1013 data points of the 75–150 GHz opacity
of ammonia and the 1176 data points of the 1.5–6 GHz opacity of ammonia measured
as part of this work.
The new ammonia opacity model fits 66.84% of the 1.5–6 GHz high-pressure cav-
ity resonator measurements, 66.14% of the 75–150 GHz FPR measurements, 93.92%
of the 1.5–27 GHz cavity resonator measurements, and 90.4% of the 22–40 GHz
104
FPR measurements within 2σ uncertainty. Overall, the model fits 78.2% of the
3870 measurements in the 1.5–150 GHz range within 2σ uncertainty. Comparison
of the new model peformance with the models of Berge and Gulkis (1976), Spilker
(1990), Joiner and Steffes (1991), Mohammed and Steffes (2003), Mohammed and
Steffes (2004), Hanley et al. (2009), Hanley (2008) with rotational lines, and Devaraj
et al. (2011) is listed in Table 5.9. Plots comparing some of the measured data to the
different models are shown in Figures 5.2–5.57. The error bars shown in the plots are
the 2σ measurement uncertainties. In the implementation of the model by Joiner and
Steffes (1991), there were difficulties in matching the numerical values given by the
author. Hanley et al. (2009) describe similar difficulties in reproducing the original
numerical values given by Joiner and Steffes (1991). However, the numerical val-
ues estimated by Hanley et al. (2009) for the various models are consistent with the
values estimated in this work. The Hanley et al. (2009) model includes only the in-
version transitions of ammonia and was developed using their centimeter-wavelength
measurements conducted at pressures up to 12 bar. Hanley (2008) provided a modi-
fication to this model for use in the high-pressure regime by including the lowest 20
rotational transitions of ammonia in addition to the inversion transitions, and both
the models are shown in the figures for comparison.
105
Table 5.9: The percentage of the NH3 /He/H2 measurement data points within 2σ uncertainty of the new model in comparison
with the existing models.
NH3 opacity model Cavity (1.5– FPR (22–40 GHz) FPR (75–150 GHz) High Pressure Total
27 GHz) (1.5–6 GHz)
Berge and Gulkis (1976) 53.11 89.2 23.59 56.72 48.81
Spilker (1990) 70.23 48.4 8.69 25.6 39.15
Joiner and Steffes (1991) 82.88 84.4 19.84 60.03 59.53
Mohammed and Steffes 57.16 55.2 3.95 37.5 37.13
106
(2003)
Mohammed and Steffes 49.13 86.4 25.07 52.55 46.28
(2004)
Hanley et al. (2009) 96.09 85.2 10.46 61.73 62.5
Hanley (2008) w/ rot. lines 96.09 85.6 12.04 61.82 62.99
Devaraj et al. (2011) 95.11 94.8 66.44 35.46 69.46
This work 93.92 90.4 66.14 66.84 78.19
P=1.009 bars, T=216.4 K, NH =0.95%, He=13.47%, H =85.58%
3 2
2
10
Opacity (dB/km)
1
10
Measured Data
0 Berge and Gulkis (1976)
10 Spilker (1990)
Joiner and Steffes (1991)
Mohammed and Steffes (2003)
−1
Hanley et al. (2009)
10 Hanley w/ rot. lines (2008)
Devaraj et al. (2011)
This Work
5 10 15 20 25
Frequency (GHz)
Figure 5.2: Opacity data measured by Hanley et al. (2009) using the cavity resonators
for a mixture of NH3 = 0.95%, He = 13.47%, H2 = 85.58% at a pressure of 1.009 bar
and temperature of 216.4 K compared to various models.
P=5.782 bars, T=216.3 K, NH3=0.77%, He=13.5%, H2=85.73%
2
10
Opacity (dB/km)
Measured Data
1
Spilker (1990)
Hanley w/ rot. lines (2008)
0 Devaraj et al. (2011)
10
This Work
5 10 15 20 25
Frequency (GHz)
for a mixture of NH3 = 0.77%, He = 13.5%, H2 = 85.73% at a pressure of 5.782 bar
107
P=2.96 bars, T=293.6 K, NH3=4%, He=13.06%, H2=82.94%
3
10
2
Opacity (dB/km) 10
Measured Data
10 Spilker (1990)
0
This Work
2 4 6 8 10 12 14 16 18 20 22
Frequency (GHz)
for a mixture of NH3 = 4%, He = 13.06%, H2 = 82.94% at a pressure of 2.96 bar and
temperature of 293.6 K compared to various models.
P=5.927 bars, T=293.3 K, NH3=4%, He=13.06%, H2=82.94%
3
10
Opacity (dB/km)
2
10
Measured Data
Berge and Gulkis (1976)
Spilker (1990)
1 Mohammed and Steffes (2003)
10
This Work
2 4 6 8 10 12 14 16 18 20 22
Frequency (GHz)
for a mixture of NH3 = 4%, He = 13.06%, H2 = 82.94% at a pressure of 5.927 bar
108
P=2 bars, T=295.3 K, NH =4%, He=13.06%, H =82.94%
3 2
3
10
Opacity (dB/km)
Measured Data
Spilker (1990)
10
This Work
24 26 28 30 32 34 36 38
Frequency (GHz)
Figure 5.6: Opacity data measured by Hanley et al. (2009) using the Fabry-Perot
resonator for a mixture of NH3 = 4%, He = 13.06%, H2 = 82.94% at a pressure of 2
bar and temperature of 295.3 K compared to various models.
P=0.976 bars, T=296.3 K, NH3=4%, He=13.06%, H2=82.94%
1
10
Opacity (dB/km)
Measured Data
Spilker (1990)
10 Hanley et al. (2009)
This Work
80 85 90 95 100 105 110
Frequency (GHz)
Figure 5.7: Opacity data measured using the 3–4 mm-wavelength system for a mix-
ture of NH3 = 4%, He = 13.06%, H2 = 82.94% at a pressure of 0.976 bar and
109
P=1.999 bars, T=296.2 K, NH =4%, He=13.06%, H =82.94%
3 2
Opacity (dB/km)
1
10 Measured Data
Spilker (1990)
This Work
80 85 90 95 100 105 110
Frequency (GHz)
P=3.009 bars, T=296.3 K, NH3=4%, He=13.06%, H2=82.94%
2
10
Opacity (dB/km)
Measured Data
Spilker (1990)
1
This Work
80 85 90 95 100 105 110
Frequency (GHz)
110
P=1.041 bars, T=219.3 K, NH =2%, He=13.33%, H =84.67%
3 2
1
10
Opacity (dB/km)
Measured Data
Spilker (1990)
0
This Work
80 85 90 95 100 105 110
Frequency (GHz)
Figure 5.10: Opacity data measured using the 3–4 mm-wavelength system for a
mixture of NH3 = 2%, He = 13.33%, H2 = 84.67% at a pressure of 1.041 bar and
P=2.07 bars, T=219.2 K, NH3=2%, He=13.33%, H2=84.67%

Opacity (dB/km)
1 Measured Data
10
Spilker (1990)
This Work
80 85 90 95 100 105 110
Frequency (GHz)
111
P=3.085 bars, T=219.6 K, NH3=2%, He=13.33%, H2=84.67%
2
10
Opacity (dB/km)
Measured Data
Spilker (1990)
10
This Work
80 85 90 95 100 105 110
Frequency (GHz)
P=0.994 bars, T=221.5 K, NH3=6.04%, He=12.78%, H2=81.18%

Opacity (dB/km)
1 Measured Data
10
Spilker (1990)
This Work
80 85 90 95 100 105 110 115
Frequency (GHz)
mixture of NH3 = 6.04%, He = 12.78%, H2 = 81.18% at a pressure of 0.994 bar and
112
P=1.925 bars, T=221.1 K, NH3=3.12%, He=13.18%, H2=83.7%
2
10
Opacity (dB/km)
Measured Data
Spilker (1990)
1
10 Joiner and Steffes (1991)
This Work
80 85 90 95 100 105 110 115
Frequency (GHz)
P=2.766 bars, T=221.6 K, NH3=2.14%, He=13.31%, H2=84.55%
2
10
Opacity (dB/km)
Measured Data
Spilker (1990)
This Work
80 85 90 95 100 105 110 115
Frequency (GHz)
113
P=1.083 bars, T=207.7 K, NH =9.17%, He=12.35%, H =78.48%
3 2
2
10
Opacity (dB/km)
Measured Data
Spilker (1990)
10
This Work
80 85 90 95 100 105 110 115
Frequency (GHz)
P=1.943 bars, T=207.7 K, NH3=5.06%, He=12.91%, H2=82.03%
2
10
Opacity (dB/km)
Measured Data
Spilker (1990)
10
This Work
80 85 90 95 100 105 110 115
Frequency (GHz)
114
P=2.855 bars, T=207.7 K, NH =3.36%, He=13.14%, H =83.5%
3 2
2
Opacity (dB/km) 10
Measured Data
Spilker (1990)
1
10 This Work
80 85 90 95 100 105 110 115
Frequency (GHz)
P=0.255 bars, T=295.1 K, NH3=100%, He=0%, H2=0%
2
10
Opacity (dB/km)
Measured Data
Spilker (1990)
This Work
80 85 90 95 100 105 110 115
Frequency (GHz)
Figure 5.19: Opacity data measured using the 3–4 mm-wavelength system for pure
ammonia gas at a pressure of 0.255 bar and temperature of 295.1 K compared to
various models.
115
P=0.505 bars, T=296.7 K, NH =100%, He=0%, H =0%
3 2
Opacity (dB/km)
2 Measured Data
10
Spilker (1990)
This Work
75 80 85 90 95 100 105 110 115
Frequency (GHz)
various models.
P=0.755 bars, T=296.8 K, NH3=100%, He=0%, H2=0%
3
10
Opacity (dB/km)
Measured Data
Spilker (1990)
10
This Work
75 80 85 90 95 100 105 110 115
Frequency (GHz)
various models.
116
P=0.124 bars, T=295.9 K, NH =100%, He=0%, H =0%
3 2
Opacity (dB/km)
1
10
Measured Data
Spilker (1990)
This Work
105 110 115 120 125 130 135 140 145
Frequency (GHz)
various models.
P=0.262 bars, T=295.5 K, NH3=100%, He=0%, H2=0%
2
10
Opacity (dB/km)
Measured Data
Spilker (1990)
1
This Work
105 110 115 120 125 130 135 140 145
Frequency (GHz)
various models.
117
P=0.103 bars, T=220.9 K, NH3=100%, He=0%, H2=0%
2
10
Opacity (dB/km)
Measured Data
10
Spilker (1990)
This Work
105 110 115 120 125 130 135 140 145
Frequency (GHz)
various models.
P=1.854 bars, T=296 K, NH3=3.97%, He=13.06%, H2=82.97%

Opacity (dB/km)
1
10
Measured Data
Spilker (1990)
This Work
110 115 120 125 130 135 140 145
Frequency (GHz)
temperature of 296 K compared to various models.
118
P=2.998 bars, T=296.2 K, NH =3.97%, He=13.06%, H =82.97%
3 2
2
10
Opacity (dB/km)
Measured Data
Spilker (1990)
10
This Work
110 115 120 125 130 135 140 145
Frequency (GHz)
P=2.841 bars, T=220.9 K, NH3=2.21%, He=13.3%, H2=84.49%
2
10
Opacity (dB/km)
Measured Data
Spilker (1990)
1
This Work
105 110 115 120 125 130 135 140 145
Frequency (GHz)
119
P=1.089 bars, T=220.7 K, NH =10.89%, He=12.12%, H =76.99%
3 2
2
10
Opacity (dB/km)
Measured Data
Spilker (1990)
This Work
105 110 115 120 125 130 135 140 145
Frequency (GHz)
P=1.909 bars, T=220.7 K, NH3=6.2%, He=12.76%, H2=81.04%
2
10
Opacity (dB/km)
Measured Data
Spilker (1990)
1
This Work
105 110 115 120 125 130 135 140 145
Frequency (GHz)
120
P=2.826 bars, T=220.6 K, NH =4.07%, He=13.05%, H =82.88%
3 2
2
10
Opacity (dB/km)
Measured Data
Spilker (1990)
10 This Work
105 110 115 120 125 130 135 140 145
Frequency (GHz)
P=2.791 bars, T=208.1 K, NH3=3.41%, He=13.14%, H2=83.45%
2
10
Opacity (dB/km)
Measured Data
Spilker (1990)
1
10 This Work
110 115 120 125 130 135 140 145
Frequency (GHz)
121
P=0.09 bars, T=376 K, NH =100%, He=0%, H =0%
3 2
0
Opacity (dB/km) 10
Measured Data
Spilker (1990)
−1
This Work
2 3 4 5 6 7
Frequency (GHz)
Figure 5.32: Opacity data measured using the high-pressure centimeter-wavelength

system for a mixture of NH3 = 100%, He = 0%, H2 = 0% at a pressure of 0.09 bar
and temperature of 376 K compared to various models.
P=12.618 bars, T=375.9 K, NH3=0.71%, He=99.29%, H2=0%
1
10
Opacity (dB/km)
Measured Data
0
10 Spilker (1990)
This Work
2 3 4 5 6 7
Frequency (GHz)

system for a mixture of NH3 = 0.71%, He = 99.29%, H2 = 0% at a pressure of 12.618
122
P=21.411 bars, T=376 K, NH =0.42%, He=58.51%, H =41.07%
3 2
1
10
Opacity (dB/km)
Measured Data
Spilker (1990)
0
This Work
2 3 4 5 6 7
Frequency (GHz)

system for a mixture of NH3 = 0.42%, He = 58.51%, H2 = 41.07% at a pressure of
21.411 bar and temperature of 376 K compared to various models.
P=39.649 bars, T=375.9 K, NH3=0.23%, He=31.6%, H2=68.18%
1
10
Opacity (dB/km)
Measured Data
Spilker (1990)
0
This Work
2 3 4 5 6 7
Frequency (GHz)

39.649 bar and temperature of 375.9 K compared to various models.
123
P=61.608 bars, T=376 K, NH3=0.15%, He=20.34%, H2=79.52%
1
10
Opacity (dB/km)
Measured Data
Spilker (1990)
0
This Work
2 3 4 5 6 7
Frequency (GHz)

P=79.454 bars, T=376.1 K, NH3=0.11%, He=15.77%, H2=84.12%
1
10
Opacity (dB/km)
Measured Data
Spilker (1990)
0
This Work
1 2 3 4 5 6 7
Frequency (GHz)

124
P=94.665 bars, T=376 K, NH =0.09%, He=13.23%, H =86.67%
3 2
1
10
Opacity (dB/km)
Measured Data
Spilker (1990)
0
This Work
1 2 3 4 5 6 7
Frequency (GHz)

P=0.082 bars, T=446.8 K, NH3=100%, He=0%, H2=0%
0
10
Opacity (dB/km)
Measured Data
Spilker (1990)
−1
This Work
2 3 4 5 6 7
Frequency (GHz)

125
P=14.025 bars, T=446.9 K, NH =0.59%, He=0%, H =99.41%
3 2
1
10
Opacity (dB/km)
Measured Data
Spilker (1990)
0
This Work
2 3 4 5 6 7
Frequency (GHz)

system for a mixture of NH3 = 0.59%, He = 0%, H2 = 99.41% at a pressure of 14.025
P=29.862 bars, T=446.9 K, NH3=0.28%, He=0%, H2=99.72%
1
10
Opacity (dB/km)
Measured Data
Spilker (1990)
0
This Work
2 3 4 5 6 7
Frequency (GHz)

126
P=51.48 bars, T=446.9 K, NH =0.16%, He=0%, H =99.84%
3 2
1
10
Opacity (dB/km)
Measured Data
Spilker (1990)
10
This Work
2 3 4 5 6 7
Frequency (GHz)

P=69.01 bars, T=446.8 K, NH3=0.12%, He=0%, H2=99.88%
1
10
Opacity (dB/km)
Measured Data
Spilker (1990)
This Work
1 2 3 4 5 6 7
Frequency (GHz)

127
P=93.545 bars, T=446.9 K, NH3=0.09%, He=0%, H2=99.91%
1
10
Opacity (dB/km)
Measured Data
0 Spilker (1990)
10
This Work
1 2 3 4 5 6 7
Frequency (GHz)

P=0.046 bars, T=332.9 K, NH3=100%, He=0%, H2=0%
0
10
Opacity (dB/km)
Measured Data
−1
Spilker (1990)
This Work
2 3 4 5 6 7
Frequency (GHz)

128
P=15.628 bars, T=333.2 K, NH =0.3%, He=99.7%, H =0%
3 2
1
10
Opacity (dB/km)
0
10 Measured Data
Spilker (1990)
This Work
2 3 4 5 6 7
Frequency (GHz)

P=31.738 bars, T=333 K, NH3=0.15%, He=49.1%, H2=50.76%
1
10
Opacity (dB/km)
Measured Data
Spilker (1990)
0
This Work
2 3 4 5 6 7
Frequency (GHz)

129
P=49.567 bars, T=333.1 K, NH =0.09%, He=31.44%, H =68.47%
3 2
1
10
Opacity (dB/km)
Measured Data
Spilker (1990)
10
This Work
2 3 4 5 6 7
Frequency (GHz)

P=64.177 bars, T=333.1 K, NH3=0.07%, He=24.28%, H2=75.65%
1
10
Opacity (dB/km)
Measured Data
0 Spilker (1990)
10
This Work
2 3 4 5 6 7
Frequency (GHz)

130
P=91.728 bars, T=332.9 K, NH =0.05%, He=16.99%, H =82.96%
3 2
1
10
Opacity (dB/km)
Measured Data
0
Spilker (1990)
This Work
1 2 3 4 5 6 7
Frequency (GHz)

P=0.125 bars, T=502.4 K, NH3=100%, He=0%, H2=0%
0
10
Opacity (dB/km)
Measured Data
−1 Spilker (1990)
This Work
2 3 4 5 6 7
Frequency (GHz)

131
P=10.877 bars, T=502.6 K, NH =1.15%, He=98.85%, H =0%
3 2
1
10
Opacity (dB/km)
0
10 Measured Data
Spilker (1990)
This Work
2 3 4 5 6 7
Frequency (GHz)

P=19.882 bars, T=502.5 K, NH3=0.63%, He=54.08%, H2=45.29%
1
10
Opacity (dB/km)
Measured Data
Spilker (1990)
This Work
2 3 4 5 6 7
Frequency (GHz)

132
P=42.01 bars, T=502.3 K, NH =0.3%, He=25.59%, H =74.11%
3 2
1
10
Opacity (dB/km)
Measured Data
Spilker (1990)
0
This Work
2 3 4 5 6 7
Frequency (GHz)

P=58.305 bars, T=502.4 K, NH3=0.21%, He=18.44%, H2=81.34%
1
10
Opacity (dB/km)
Measured Data
Spilker (1990)
0
This Work
1 2 3 4 5 6 7
Frequency (GHz)

133
P=82.432 bars, T=502.1 K, NH3=0.15%, He=13.04%, H2=86.8%
1
10
Opacity (dB/km)
Measured Data
Spilker (1990)
0
This Work
1 2 3 4 5 6 7
Frequency (GHz)

P=98.439 bars, T=502.1 K, NH3=0.13%, He=10.92%, H2=88.95%
1
10
Opacity (dB/km)
Measured Data
Spilker (1990)
0
This Work
1 2 3 4 5 6 7
Frequency (GHz)

134
5.5.1 Ultra-High-Pressure Extrapolation
The new consistent ammonia opacity model was developed for use under jovian con-
ditions in the centimeter-wavelength range at pressures up to 100 bar and in the
millimeter-wavelength range at pressures up to 3 bar. The new model corrects for the
non-ideal pure fluid behavior under high-pressure conditions and the high-pressure
inversion model parameters were empirically derived by data fitting to a high-pressure
data set (P > 15 bar). While this model can be extrapolated with reasonable cer-
tainty to slightly higher pressures, some caution must be exercised when using the
model under ultra-high-pressure conditions (pressures much greater than 100 bar).
When modeling the microwave emission from a planetary atmosphere, weighting
functions (or contribution functions) indicate the altitudes (pressure layers) that con-
tribute most to the brightness temperature at a particular frequency. The weighting
functions probe less deep into the atmosphere at the limb of the planet than at nadir
(the path length through the atmosphere at the limb is longer than at nadir). For
example, the nadir-viewing weighting function for Jupiter show that at 1 GHz the
atmospheric layers from 20 to 300 bar contribute to the brightness temperature. At
even lower frequencies, the jovian atmospheric layers at hundreds of bar of pressure
contribute to the brightness temperature. The normalized nadir-viewing weighting
function in the 0.5–25 GHz range computed by Karpowicz (2010) using a radiative
transfer model for the atmosphere of Jupiter without contribution from cloud opacity
is shown in Figure 5.58.
Morris and Parsons (1970) made measurements of the microwave absorption of
ammonia in a predominantly hydrogen or helium atmosphere at 295 K and pressures
up to 700 bar. A tunable resonant cavity which was operated in the TE013 mode
was used for these measurements and the binary mixture absorption measurements
were conducted at 9.58 GHz. These measurements show a steep increase in opacity
with pressure up to around 80 bar, and then the opacity increases less steeply. Morris
135
Figure 5.58: Normalized nadir-viewing weighting functions in the 0.5–25 GHz range
Figure 5.4:
computed The normalized
by Karpowicz (2010)weighting function transfer
using a radiative at each model
frequency
for aasmean
a function
jovian of
pressure for a nadir viewing angle using
atmosphere without cloud contribution. the NH 3 opacity model of this work for the mean
condition of Figure 5.3.
136146
(1971) attempts to explain this behavior as a shift from resonant to non-resonant
Debye absorption along with greater frequency of collision of molecules. Berge and
Gulkis (1976) developed a model that was fit to the Morris and Parsons (1970) am-
monia/hydrogen data using a correction factor for hydrogen. However, this model
does not include a correction factor for helium, which behaves in a similar fashion
to hydrogen. Hanley (2008) attempted to fit to the Morris and Parsons (1970) data
by including the lowest 20 rotational transitions of ammonia. The Berge and Gulkis
(1976), Hanley (2008) with rotational lines, and the new model developed as part
of this work are shown along with the Morris and Parsons (1970) measured data in
Figures 5.59 and 5.60. Since the absolute uncertainties of the Morris and Parsons
(1970) ammonia measurements are not precisely known, the measurements were as-
sumed to be accurate within ±10%. There could be some additional uncertainties in
the Morris and Parsons (1970) data since they did not account for the adsorption of
ammonia on the walls of their system. Additionally, the measurements were made
only at one resonant frequency and that resonance might have been contaminated by
other resonances.
Although the new model performs very well in its consistency with the Morris
and Parsons (1970) data, some caution must be exercised when using the new model
at pressures much greater than 100 bar. The new model accounts for the non-ideal
behavior of pure fluids, but does not account for the non-ideal behavior of fluid mix-
tures. Additionally, only the inversion transitions of the new model were optimized
for use under high-pressure conditions (pressures up to 100 bar). The parameters for
the rotational transitions described in this work were not optimized to perform under
high-pressure conditions. Finally, the assumptions underlying most lineshape theo-
ries, such as binary and elastic collisions of molecules, might become invalid under
ultra-high-pressure conditions. All of the possible shortcomings of the new model at
pressures much greater than 100 bar indicate that some caution must be exercised
137
T=295 K, NH3=0.9%, He=99.1%, H2=0%
1100 Morris and Parsons (1970)

This Work
900
800
Opacity (dB/km)
700
600
500
400
300
200
100
100 200 300 400 500 600 700

Pressure (bar)
Figure 5.59: Opacity data by Morris and Parsons (1970) for a mixture of NH3 =
0.9%, He = 99.1%, H2 = 0% at a frequency of 9.58 GHz and temperature of 295 K
compared to this work, Hanley (2008) with rotational lines and Berge and Gulkis
(1976)
138
T=295 K, NH3=0.44%, He=0%, H2=99.56%
Morris and Parsons (1970)

400 This Work
350
Opacity (dB/km)
300
250
200
150
100
50
100 200 300 400 500 600 700
Pressure (bar)
Figure 5.60: Opacity data measured by Morris and Parsons (1970) for a mixture of
NH3 = 0.44%, He = 0%, H2 = 99.56% at a frequency of 9.58 GHz and temperature
of 295 K compared to this work, Hanley (2008) with rotational lines, and Berge and
Gulkis (1976)
139
when extrapolating the new model to ultra-high-pressure conditions characteristic of
the very deep atmospheres of jovian planets.
5.6 The Influence of Water Vapor on the Ammonia Ab-

sorption Spectrum
Water vapor is the third most abundant constituent deep in the atmosphere of Jupiter
after hydrogen and helium (see, e.g., Karpowicz, 2010). Current ammonia opacity
models for jovian conditions, including the new model that was described earlier in
this chapter, account for the self-broadening and foreign-gas-broadening due to hy-
drogen and helium. While the broadening effects of hydrogen and helium are well
characterized, it is critical to investigate any possible pressure-broadening effects of
water vapor on the ammonia absorption spectrum. Karpowicz and Steffes (2011)
made extensive measurements of the centimeter-wavelength opacity of water vapor
under jovian conditions and found that the self-broadening from water vapor domi-
nates its absorption spectrum. Water vapor, with its a large broadening cross-section
(collision diameter), has the potential to broaden the ammonia transitions. Hence,
it is important to study the influence of water vapor on the ammonia absorption
spectrum.
Preliminary experimental investigation of the pressure-broadening of ammonia by
water vapor in the centimeter-wavelength region has been conducted in the 375–450
K temperature range. Prior to this investigation, there was at least one laboratory
measurement study which indicated that water vapor can efficiently broaden the 572
GHz rotational transition of ammonia (Belov et al., 1983). Hence, understanding the
enhanced opacity of ammonia due to the presence of water vapor is crucial to the
determination of the abundance of water vapor in the jovian atmospheres.
140
5.6.1 Ammonia/Water Vapor Opacity Measurements
Laboratory measurements of the centimeter-wavelength properties of ammonia pressure-
broadened by water vapor have been conducted using the high-pressure system.
The measurement process involved four sequences of measurements of the 5–20 cm-
wavelength opacity of ammonia/water vapor mixture in the 375–450 K temperature
range. The measurements at 450 K were conducted by this author and the measure-
ments at 375 K were conducted by fellow graduate student Danny Duong. Certified
UHP grade ammonia and argon gas cylinders from Airgas, Inc. were used for the
experiments and dielectric matching. A water reservoir (similar to the one used by
Karpowicz and Steffes, 2011), filled with distilled water, ACS Reagent Grade with
ASTM D 1193 specifications for reagent water, type II (manufactured by Ricca Chem-
ical Company), was connected to the pressure vessel and placed inside the oven to
generate the required amount of water vapor for the experiments. The water vapor
adsorption is very small (see, e.g., Karpowicz, 2010) when compared to ammonia un-
der our experimental conditions. Hence, ammonia was added first to the system and
the walls of the pressure vessel and cavity resonator were presaturated with ammonia
before the pure ammonia measurements were made.
Each measurement sequence first involved vacuum measurements, followed by the
addition of gaseous ammonia to the evacuated chamber up to the desired pressure
and pure ammonia measurements were made. Water vapor was then added to the
system up to the desired pressure, by opening the oven door and quickly opening
and closing the valve on the water reservoir, and ammonia/water vapor mixture
measurements were made. Since there is a possibility of water vapor condensing in
the pipes exterior to the oven (which are at room temperature), the high-temperature
inlet and outlet valves that are inside the oven were closed before water vapor is
added into the pressure vessel. A coarse estimate of the water vapor pressure, while
adding water vapor to the system, was obtained from the pressure transducer readings
141
Table 5.10: Listing of all experiment sequences of the 5–20 cm-wavelength opacity
measurements of the NH3 /H2 O mixture.
Experiment Dates Temperature NH3 Pressure H2 O Pressure
(K) (bar) (bar)
12/2010 452.11 0.0902 0.945
12/2010 452.52 0.0462 1.0912
02/2011 373.85 0.0295 0.1499
03/2011 373.85 0.0473 0.2065
to ensure that its partial pressure was well below the saturation vapor pressure at
the measurement temperature. After the measurements of the water vapor/ammonia
mixture were conducted, the pressure was measured using a “buffer method”. A small
amount of helium was added to the pipe section connecting the DPI 10430A gauge
to the inlet valve to a pressure known to be slightly greater than that in the pressure
vessel. Then, while monitoring the DPI 10430A gauge, the oven was quickly opened
and the inlet valve to the pressure vessel was opened and the pressure quickly read
before closing the valve and shutting the oven door. The pressure reading stabilizes
very quickly and the possibility of any water vapor entering and condensing in the
pipes external to the oven is very low. Since the volume of the pipe section between
the DPI 10430A gauge and the inlet valve is very small relative to the volume of the
pressure vessel, which is approximately 32 liters, no additional corrections were needed
to account for the true pressure in the vessel prior to opening the inlet valve. After
this step, a vacuum was drawn in the pressure vessel and a second set of vacuum
measurements were conducted. After this step, dielectric matching measurements
were conducted followed by a third set of vacuum measurements. Finally, three sets
of straight-through measurements of signal levels were conducted. Table 5.10 lists
the measurements and the experiment dates.
142
5.6.2 Preliminary Ammonia/Water Vapor Opacity Model
Based on the initial four sequences of the ammonia+water vapor measurements, a
preliminary model of the opacity of ammonia pressure-broadened by water vapor
has been developed in a fashion similar to that explained in Section 5.4.2. The
preliminary model for the water vapor-broadened opacity of ammonia uses a modified
Ben-Reuven lineshape (Ben-Reuven, 1966) for the inversion transitions. The opacity
of ammonia broadened by water vapor in the centimeter-wavelength region is given by
Equation 5.12. The linewidth and the coupling parameter of the lines are calculated
by summing the contribution from ammonia and water vapor and are given by
ΓH2 O ΓN H3
300 295
γj = γH2 O PH2 O + γN H3 γ(0,j) PN H3 (GHz), (5.21)
T T
ZH2 O ZN H3
300 295
ζj = ζH2 O PH2 O + ζN H3 γ(0,j) PN H3 (GHz), (5.22)
T T
where for the inversion line j and i = H2 O and NH3 , γi and ζi are constant
scale terms, and Γi and Zi represent the constant temperature dependences of the
broadening of each of the gases, Pi are the ideal partial pressures in bar, and γ(0,j) are
the self-broadening linewidths of the inversion transitions of ammonia in MHz/Torr.
The pressure shift parameter is calculated by
δj = d × γj (GHz), (5.23)
where d is an empirically derived constant. The free parameters for water va-
por were empirically derived using the four sequences of measurements, while the
free parameters for ammonia were assigned the same values derived in Section 5.4.2.
Table 5.11 lists the values of the water vapor and ammonia model constants of the
preliminary NH3 /H2 O model.
143
Table 5.11: Values of the H2 O model constants used for computing the H2 O-
broadened NH3 absorptivity from the inversion transitions.
i=H2 O i=NH3
γi 8.4993 0.7298
Γi 1 1
ζi 1.364 0.5152
Zi 2/3 2/3
d -0.0627
Dinv 0.9862
The laboratory measurements and the preliminary model of the opacity of am-
monia in a water vapor atmosphere have given insights into the pressure broadening
effects of water vapor on ammonia opacity and can potentially improve retrievals
of the atmospheric abundance of water vapor at Jupiter from MWR measurements.
Plots comparing some of the measured data to the preliminary NH3 /H2 O model,
along with the models for pure ammonia opacity (this work) and pure water vapor
opacity (Karpowicz and Steffes, 2011) are shown in Figures 5.61–5.62. The error bar
shown in the plots are the 2σ measurement uncertainties. It can be seen that there is
a significant influence of water vapor on the centimeter-wavelength opacity of ammo-
nia. An aggressive campaign of laboratory measurements of the opacity of ammonia
pressure-broadened by water vapor/hydrogen/helium mixture under simulated deep
jovian conditions are currently being conducted by fellow graduate student Danny
Duong. An updated model of the opacity of ammonia pressure-broadened by water
vapor/hydrogen/helium mixture will be developed by a future graduate student after
the measurements are completed.
144
P=1.041 bars, T=452 K, NH3=8.71%, H2O=91.29%, He=0%, H2=0%
1
10
Opacity (dB/km)
0
10
−1
10 Measured Data
Karpowicz and Steffes, 2011 (H2O only)
This Work (NH3 only)
This Work (NH3+H2O)
2 3 4 5 6 7
Frequency (GHz)

system for a mixture of NH3 = 8.71% and H2 O = 91.29% at a pressure of 1.041 bar
and temperature of 452 K compared to Karpowicz and Steffes (2011) H2 O model,
the NH3 model described in Section 5.4.2 and the preliminary NH3 + H2 O model that
includes the interaction between NH3 and H2 O.
145
P=0.18 bars, T=373.9 K, NH3=16.44%, H2O=83.56%, He=0%, H2=0%
0
10
Opacity (dB/km)
−1
10
Measured Data
−2 Karpowicz and Steffes, 2011 (H2O only)
10
This Work (NH3 only)
This Work (NH3+H2O)
2 3 4 5 6 7
Frequency (GHz)

system for a mixture of NH3 = 16.44% and H2 O = 83.56% at a pressure of 0.18 bar
and temperature of 373.9 K compared to Karpowicz and Steffes (2011) H2 O model,
the NH3 model described in Section 5.4.2 and the preliminary NH3 + H2 O model that
includes the interaction between NH3 and H2 O.
146
CHAPTER VI
MILLIMETER-WAVELENGTH OBSERVATIONS OF
JUPITER
Radiative transfer simulations of the atmosphere of Jupiter performed early in this
work (Devaraj et al., 2011) predicted that the jovian brightness temperature at 140.14
GHz corresponding to the ν2 = 1 inversion transition of ammonia was at least 4 K
lower than the continuum brightness temperature at these wavelengths. Hence, in
an attempt to detect the line, millimeter-wavelenth observations of Jupiter were con-
ducted with the Institut de Radioastronomie Millimétrique (IRAM) 30 m telescope
facility as described below. The disk-averaged brightness temperature spectrum of
Jupiter obtained from these observations showed no features corresponding to the
absorption line. Recently, the ammonia spectral line catalogs were significantly im-
proved (Yu et al., 2010a,b,c). New radiative transfer simulations of the atmosphere of
Jupiter were performed using the ammonia absorption model presented in this work
(with the new spectral line catalogs for the ammonia ground-state and ν2 transitions)
and the simulations predict that the 140 GHz transition is only < 0.5 K darker than
the continuum. As discussed below, this is consistent with the latest observations of
Jupiter (non-detection of the line)1 .
6.1 IRAM Observations
The IRAM 30 m telescope located on Pico Veleta in the Spanish Sierra Nevada at an
elevation of 2850 m is one of the most sensitive millimeter-wave telescopes. The 30
m single dish parabolic antenna has a Cassegrain-Nasmyth configuration supported
1
Observations were carried out with the IRAM 30 m Telescope. IRAM is supported by
INSU/CNRS (France), MPG (Germany) and IGN (Spain)
147
on an alt-azimuth mount. Eight MIxer heterodyne Receivers (EMIR) that operate in
the 0.8–3.6 mm-wavelength range were recently installed on the telescope and used
for this study. Observations of Jupiter were conducted using two E1 band (120–170
GHz) horizontal (H) and vertical (V) polarization receivers tuned to 140.142 GHz on
two consecutive nights (September 17 and 18) in 2010. The half-power beamwidth
(HPBW) of a single dish telescope with a 70% illumination efficiency is given by

λ
HPBW = 1.22 × (radians), (6.1)
D
where λ is the observation wavelength and D is the diameter of the telescope. The
antenna HPBW at 140 GHz is approximately 17.5 arcsec and the spectral resolution
is 4 MHz. An IF bandwidth of 4 GHz was used for the observations. During the
observations, Jupiter was approximately 3.95 AU from Earth and had an angular
diameter of 49.8 arcsec. The astrometric right ascension was ∼ 23 h 58 m and decli-
nation was ∼ -01 deg 55 m. The Moon was chosen as the calibration source for these
observations since it is a constant radiator without any features at millimeter wave-
lengths, and the night side of the Moon is approximately at the same temperature as
Jupiter. It was also at a convenient elevation in the sky during the observation period.
The observations of the Moon were used for calibration, and disk-averaged antenna
temperature measurements of Jupiter were made. It was possible to remove part of
the standing waves inherent to the receiver by comparing the antenna temperature
spectra of Jupiter and the Moon.
Figures 6.1 and 6.2 show the observed antenna temperature spectra of Jupiter
and the Moon for the H and V polarizations on September 17, 2010, and Figures 6.3
and 6.4 show the observed antenna temperature spectra of Jupiter and the Moon
for the H and V polarizations on September 18, 2010. These antenna temperature
spectra have been corrected for the system temperature. It can be seen that the
spectra between the two days are quite reproducible. It can also be seen that on both
days H polarization spectra have much better SNR than V polarization spectra.
148
Figure 6.1: Antenna temperature spectra obtained with the H polarization receiver
on September 17, 2010 for Jupiter and the Moon. The abscissa on top is frequency
in MHz and the ordinate is antenna temperature in K.
Figure 6.2: Antenna temperature spectra obtained with the V polarization receiver
149
Figure 6.3: Antenna temperature spectra obtained with the H polarization receiver
Figure 6.4: Antenna temperature spectra obtained with the V polarization receiver
150
Figure 6.5: Ratio of the 140.14 GHz line to continuum spectrum (H polarization)
obtained on September 17, 2010. The abscissa on top is frequency in MHz.
The primary objective of the observations was to detect a potential dip in Jupiter’s
emission spectrum resulting from the 140 GHz line. Standing waves inherent to the
receivers (∼ 300 MHz) can be seen on both the Jupiter and Moon observations. After
removing most of the standing waves by dividing the Jupiter spectrum with that of
the Moon, the ratio of the line (4 MHz spectral resolution centered at 140.142 GHz) to
continuum spectrum (4 GHz) was obtained. The final spectra of the ratio of the line
and continuum for the two observations are shown in Figures 6.5 and 6.6. Only the H
polarization spectra are shown here since the V spectra were very noisy. These spectra
have a typical rms of ∼0.2–0.3% mainly due to the residual of the standing waves.
It can be seen from the figures that there is no dip in the spectrum corresponding to
the 140 GHz ν2 inversion transition of ammonia.
Ulich (1974) conducted very high-precision absolute brightness temperature mea-
surements of the new Moon at 2.1 mm (140 GHz) and reported a value of 145 ± 9
K. Since the night-side of the Moon is a constant radiator at millimeter-wavelengths,
it serves as an excellent calibration source at millimeter-wavelengths. The measured
151
Figure 6.6: Ratio of the 140.14 GHz line to continuum spectrum (H polarization)
obtained on September 18, 2010. The abscissa on top is frequency in MHz.
antenna temperature at 140 GHz (corrected for the system temperature) for the
night-side of the Moon is ∼ 125 K and for Jupiter is ∼ 133 K. Since the brightness
temperature of the new Moon is available to a very-high precision (145 ± 9 K), it is
possible to calibrate the corrected antenna temperature measurements of Jupiter with
that of the Moon. The computed disk-averaged brightness temperature of Jupiter at
140 GHz is ∼ 155 ± 15 K.
6.2 Radiative Transfer Model
A radiative transfer model (RTM) can be used to simulate the millimeter-wavelength
emission spectrum of Jupiter as observed from Earth. By incorporating the new am-
monia opacity formalism in the radiative transfer calculations of Jupiter, it is possible
to analyze the effects of the new formalism on the modeled brightness temperature
of the planet. The RTM used for this analysis was developed by Hesman et al.
(2007) and modified by this author. The RTM relies on a thermochemical model
(TCM) to define the distribution and abundances of the constituent elements and
152
the temperature-pressure (TP) profile of the atmosphere of Jupiter. The construc-
tion begins, in general, from some assumption of the deep abundance of atmospheric
constituents. The TP profile is then computed using a wet adiabatic extrapolation
in discrete layers of pressure, starting at the deepest layer and ending at the 1 bar
pressure level. For pressures less than 1 bar, the Voyager radio occultation results
summarized in tabular form by Lindal (1992) are used. The assumed mole fraction
of various constituents in the deep atmosphere of Jupiter for the RTM construction
are as follows: XHe =0.1346, XH2 O =0.0055, XH2 S =66.01 ppm, XNH3 = 800 ppm, and
rest H2 . These abundance values are based on recent studies of the composition of
Jupiter (e.g., Atreya et al., 2003). Since the abundance of PH3 in Jupiter is very
small (e.g., Kunde et al., 1982), and because there are no strong absorption lines of
PH3 in the 2–4 mm-wavelength range (e.g., Pickett et al., 1998), it is not included in
the RTM simulations.
Once the constituent abundances, temperature, and pressure in each layer has
been defined, the absorption of each layer is calculated. The collision induced ab-
sorption between H2 –H2 , H2 –He, and H2 –CH4 are calculated using a model given
by Orton et al. (2007). Absorption from NH3 is computed using the latest formal-
ism described in this paper. Absorption from H2 O is computed using a formalism
developed by Karpowicz (2010), and absorption from H2 S is computed using a for-
malism developed by DeBoer and Steffes (1994). The formalisms for H2 O and H2 S
are based on centimeter-wavelength laboratory measurements. The outgoing electro-
magnetic radiation exiting the atmosphere of Jupiter at the measurement frequency
is predicted by the black body thermal emission of each layer and the absorption of
its constituents.
153
6.3 Search for the 140 GHz Line
Based on the results of radiative transfer simulations of the atmosphere of Jupiter
performed early in this work (Devaraj et al., 2011), observations were conducted with
the IRAM 30 m telescope to search for a potential dip in Jupiter’s spectrum resulting
from the 140 GHz ammonia line transition. However, no feature was detected in
the observed spectrum of Jupiter. Since then, the ammonia line catalogs have been
extensively updated (Yu et al., 2010a,b,c). New radiative transfer simulations were
performed using the ammonia model described in this dissertation.
Based on the attenuation of the Galileo Probe’s radio signal in Jupiter’s atmo-
sphere, the deep atmospheric ammonia abundance was estimated as 800 ppm (Folkner
et al., 1998; Hanley et al., 2009). The low temperatures in the upper troposphere of
Jupiter cause ammonia to condense and form ammonia ice clouds, thereby depleting
gaseous ammonia from the atmosphere. Another mechanism by which a relatively
large amount of ammonia gets depleted is by combining with hydrogen sulfide result-
ing in the formation of ammonium hydrosulfide cloud (see, e.g., Sromovsky and Fry,
2010). A range of ammonia abundance profiles from a variety of sources were used in
these simulations so as to study their impact on the millimeter-wavelength emission
spectrum of Jupiter. If ammonia is assumed to follow a saturation profile (Sat), then
ammonia abundance should be constant in the troposphere until it reaches the satu-
ration layer; at that point the ammonia abundance profile follows the vapour pressure
curve. The saturation vapor pressure equation for ammonia used in this study is given
by Overstreet and Giauque (1937). The Galileo Probe (GP) ammonia profile includes
the abundance values inferred from thermal net flux measurements in the 0.5–3 bar
range (Sromovsky et al., 1998), and from the attenuation of radio signals during the
probe descent in the 4.3–15 bar range (Folkner et al., 1998; Hanley et al., 2009). The
probe results were from the edge of a 5 µm hotspot (Orton et al., 1998) and display
a substantial decrease in ammonia mixing ratio with altitude and never reach the
154
saturation level of ammonia. The Sat and GP profiles along with three other profiles,
the hot spot (HS), north equatorial belt (NEB), and equatorial zone (EZ) profiles
listed by Sromovsky and Fry (2010) were used for modeling the ammonia abundance
in Jupiter’s atmosphere. The temperature-pressure (TP) profile of Jupiter and the
ammonia abundance profiles used in the RTM simulations are shown in Figure 6.7.
Figure 6.8 shows the TP profile of Jupiter overlayed with the normalized weighting
functions (which indicate the atmospheric layers that contribute most to the bright-
ness temperature at a particular frequency) at 140.1 GHz for the different ammonia
abundance profiles.
The RTM simulations were performed for various ammonia abundance profiles
and the predicted disk-averaged brightness temperature of Jupiter in the 138–142
GHz range is shown in Figure 6.9. The model predicts that irrespective of the as-
sumed ammonia abundance profile, the brightness temperature of Jupiter at 140
GHz (corresponding to the ν2 transition of ammonia) is only < 0.5 K darker than the
continuum. With the current capabilities of millimeter-wavelength telescopes, this
shallow line cannot be detected. Future generation of high-sensitivity telescopes such
as the ALMA might help in detecting the line at Jupiter. Of particular interest is
the predicted brightness temperature of Jupiter for various ammonia abundance pro-
files. The disk-averaged brightness temperature varies by approximately 50 K for the
two extreme cases of ammonia abundance profiles used in the simulations (saturation
profile and hot spot profile which is highly subsaturated). Even among the other
profiles, there is a variation in the predicted brightness temperature of at least 5 K.
Hence, highly accurate millimeter-wavelength observations of Jupiter can potentially
be used to retrieve the global abundance profile of ammonia in the 0.5–5 bar region.
The brightness temperature of Jupiter was estimated to be ∼ 155 ± 15 K from
the latest observations using the “relative” calibration technique described earlier.
155
Ammonia Mole Fraction
−6 −5 −4 −3
10
−1 10 10 10
10
Sat
HS
NEB
EZ
GP
Pressure (bar)
0
10
1
10
100 150 200 250 300 350
Temperature (K)
Figure 6.7: The temperature-pressure (TP) profile of Jupiter is shown as a black line.
The TP profile above the 1-bar level represents the Voyager radio occultation results
summarized by Lindal (1992), and the TP profile below the 1-bar level represents
the results of a wet-adiabatic extrapolation using the thermochemical model. The
deep ammonia abundance is fixed at 800 ppm and the ammonia abundance profiles
corresponding to the saturation (Sat), hot spot (HS), north equatorial belt (NEB),
equatorial zone (EZ), and Galileo Probe (GP) models are also shown.
156
Normalized Weighting Function
−3 −2 −1 0
10
−1 10 10 10
10
Sat
HS
NEB
EZ
GP
Pressure (bar)
0
10
1
10
100 150 200 250 300 350
Temperature (K)
Figure 6.8: The TP profile of Jupiter (black line) and the normalized weighting
functions at 140.1 GHz for various ammonia abundance profiles.
157
230
Sat
HS
NEB
220 EZ
GP
210
Brightness Temperature (K)
200
190
180
170
138 138.5 139 139.5 140 140.5 141 141.5 142

Frequency (GHz)
Figure 6.9: Modeled disk-averaged brightness temperature of Jupiter for various

ammonia profiles.
158
220
200
Brightness Temperature (K)
180
160
140
120
Existing Observations
IRAM Observations
100 1 2
10 10
Frequency (GHz)
Figure 6.10: Disk-averaged brightness temperature measurements of Jupiter.
Although these observations had large uncertainties, the measurements are still con-
sistent with the existing brightness temperature measurements. The latest measured
Jupiter brightness temperature is graphically compared against those in the litera-
ture in Figure 6.10. The existing values are taken from the compilations by Berge
and Gulkis (1976) and Joiner and Steffes (1991) along with individual observations
from de Pater et al. (1982), Griffin et al. (1986), de Pater et al. (2001), Gibson et al.
(2005), and Weiland et al. (2011).
159
CHAPTER VII
CONCLUSIONS
The objective of this doctoral research has been to advance the understanding of
the centimeter- and millimeter-wavelength properties of gaseous ammonia under jo-
vian conditions. As part of this research, extensive laboratory measurements of the
2–4 mm-wavelength properties of ammonia under simulated upper and middle tro-
pospheric conditions of the jovian planets, and the 5–20 cm-wavelength properties
of ammonia under simulated deep tropospheric conditions of the jovian planets were
conducted. These and pre-existing laboratory measurements (Hanley et al., 2009)
were utilized to develop the most accurate and consistent model to date to represent
the opacity of ammonia pressure-broadened by hydrogen/helium in the centimeter-
wavelength range at pressures up to 100 bar and temperatures up to 500 K and in
the millimeter-wavelength range at pressures up to 3 bar and temperatures up to
300 K. Additionally, preliminary laboratory investigation of the 5–20 cm-wavelength
opacity of ammonia pressure-broadened by water vapor were made and an initial
model for the centimeter-wavelength opacity of ammonia broadened by water vapor
was developed.
7.1 Contributions
The main focus of this doctoral research work was to better understand the centimeter-
and millimeter-wavelength opacity spectra of ammonia under jovian conditions. Dur-
ing the course of this work, several unique contributions were made to the fields of
microwave spectroscopy and planetary science.
160
7.1.1 Millimeter-wavelength measurement system
A high-sensitivity millimeter-wavelength measurement system was developed as part
of this work, to accurately measure the propagation properties of gases under sim-
ulated planetary atmospheric conditions. The measurement system operates in the
2–4 mm-wavelength range and withstands up to 3 bar of pressure. This system was
employed in the measurements of the 2–4 mm-wavelength opacity of ammonia under
jovian conditions as part of this work. With minor modifications, this system can
be used to study the properties of various gases such as hydrogen sulfide under Nep-
tune atmospheric conditions, and sulfur dioxide and sulfuric acid vapor under Venus
atmospheric conditions (Devaraj and Steffes, 2011).
7.1.2 Laboratory measurements and model
As part of the laboratory measurement campaign, extensive measurements of the
centimeter- and millimeter-wavelength properties of ammonia under simulated jovian
conditions were made. The millimeter-wavelength measurements contributed to the
empirical estimation of the self-, hydrogen-, and helium-broadening parameters of the
140.14 GHz ν2 inversion transition of ammonia for the first time in a laboratory set-
ting. Efforts toward developing a unified formalism to estimate the centimeter- and
millimeter-wavelength opacity spectra of ammonia at the pressures, temperatures,
and mixing ratios characteristic of the outer planets have been on-going for more
than 30 years since the work by Berge and Gulkis (1976) (see, e.g., Mohammed and
Steffes, 2004). As part of this research, a consistent absorption formalism was devel-
oped to accurately characterize the opacity of ammonia in the centimeter-wavelength
range at pressures up to 100 bar and temperatures up to 500 K, and millimeter-
wavelength range at pressures up to 3 bar and temperatures up to 300 K. This model
represents the first successful attempt at reconciling the centimeter- and millimeter-
wavelength opacity spectra of ammonia under jovian conditions. This model can
161
be used for accurate interpretation and modeling of the broad emission spectrum of
jovian planets and for accurate retrievals of ammonia and other constituents in the
jovian planetary atmospheres from ground-based and spacecraft-based radio obser-
vations. The influence of water vapor on the ammonia absorption spectrum in the
centimeter-wavelength region was studied in the laboratory for the first time as part
of this research and this work will aid in the retrievals of the atmospheric abundance
of water vapor and ammonia at Jupiter from the Juno MWR measurements.
7.1.3 Search for the 140 GHz line
Observations of Jupiter were conducted at 140 GHz to search for the ν2 =1 inversion
transition of ammonia using the IRAM 30 m telescope facility and the ν2 transition
was not detected. Radiative transfer simulations using the new ammonia model show
that the 140 GHz line is < 0.5 K darker than the continuum, and this result is
consistent with the non-detection of the line from the IRAM observations. The disk-
averaged brightness temperature of Jupiter was estimated to be ∼ 155 ± 15 K based
on the observations.
7.2 Recommendations for Future Work

7.2.1 Millimeter-wavelength laboratory work
Ground based millimeter-wavelength astronomy is very powerful in probing planets
because of the low attenuation millimeter waves suffer in Earth’s atmosphere com-
pared to submillimeter waves. New millimeter arrays such as the ALMA are capable
of providing an unprecedented combination of sensitivity, angular resolution, spec-
tral resolution, and imaging fidelity. Hence, accurate data concerning the absorptive
properties of various atmospheric constituents are required to interpret those observa-
tions and derive planet-wide abundances of those constituents. The new millimeter-
wavelength measurements of ammonia have proven that the pre-existing models did
not accurately predict the opacity of ammonia in the millimeter-wavelength range
162
under relevant planetary conditions. Hence, there is a need to accurately study
the properties of other millimeter-wavelength absorbing gases under those planetary
conditions. The system described in this paper can be used for making such high-
precision measurements. Future work will involve modifying the measurement system
for studying the millimeter-wavelength properties of sulfur dioxide and sulfuric acid
vapor under simulated Venus atmospheric conditions (Devaraj and Steffes, 2011).
Measurements of gas properties at shorter wavelengths can yield much information
about the behavior of the rotational transitions of various molecules. The FPR that
was built for the measurements of ammonia opacity has the capability to operate in
the 1–4 mm-wavelength range. A G band (140–220 GHz) planar frequency doubler
along with the active multiplier chain (AMC) can be used to generate signals in the
1.5–2 mm-wavelength range and a series of harmonic mixers can be used for down-
conversion. Such a system can be used for studying the rotational transitions of gases
such as hydrogen sulfide under simulated planetary atmospheric conditions.
7.2.2 Centimeter-wavelength laboratory work
Based on preliminary results of the ammonia/water vapor experiments, it is clear
that there is a necessity to understand the enhancement of the opacity of ammonia
due to the presence of water vapor. Extensive measurements of the properties of
ammonia in a hydrogen/helium/water vapor atmosphere under jovian conditions are
currently being conducted by fellow graduate student Danny Duong. The initial
measurements made as part of this study and the current measurements that are
being conducted will be used to update the preliminary model described in Section 5.6
to more accurately characterize the centimeter-wavelength properties of ammonia
under jovian conditions. In addition, high-pressure measurements of other centimeter-
wavelength absorbing gases with relatively large abundance, such as hydrogen sulfide,
can provide unique insights into the behavior of those gases under the deep jovian
163
conditions and also help limit uncertainties in the Juno MWR retrievals at Jupiter.
An accurate ammonia opacity model for jovian atmospheres should include the
effects of broadening from all the major atmospheric constituents. For example, the
third most abundant constituent in the atmosphere of Uranus is methane (∼ 2.3%).
Although methane is non-polar and its inherent centimeter-wavelength opacity is
very small (Jenkins and Steffes, 1988), because of its large abundance and broad-
ening cross-section it could potentially broaden ammonia under Uranus conditions.
Hence, laboratory measurements of the centimeter-wavelength opacity of ammonia in
the presence of methane under the deep atmospheric conditions of Uranus can yield
information about the potential broadening nature of methane.
7.3 List of Publications
Journal Articles
K. Devaraj and P. Steffes (2011), “A new consistent model for the microwave
opacity of ammonia under deep jovian conditions,” Icarus, in preparation.
K. Devaraj and P. Steffes (2011), “The Georgia Tech millimeter-wavelength
measurement system and some applications to study of planetary atmospheres,”
Radio Sci., 46, RS2014, doi:10.1029/2010RS004433.
K. Devaraj, P. Steffes, and B. Karpowicz (2011), “Reconciling the centimeter-
and millimeter-wavelength ammonia absorption spectra under jovian condi-
tions: Extensive millimeter-wavelength measurements and a consistent model,”
Icarus, 212, pp. 224–235.
Conference Proceedings
K. Devaraj and P. Steffes, “Laboratory measurements and a consistent model
of the microwave properties of ammonia under jovian conditions,”. To be pre-
sented at the EPSC-DPS Joint Meeting 2011, Nantes, France, October 7, 2011.
164
K. Devaraj, D. Duong, and P. Steffes, “Preliminary results for the 5-20 cm
wavelength opacity of ammonia pressure broadened by water vapor under jovian
conditions,”. To be presented at the EPSC-DPS Joint Meeting 2011, Nantes,
France, October 5, 2011.
P. Steffes and K. Devaraj, “VLA observations of Venus at X band,” Work-
shop on New Results from Venus Express, Ground-Based Observations and Fu-
ture Missions. To be presented at the 2011 VEXAG International Workshop,
Chantilly, Virginia, August 31, 2011.
K. Devaraj and P. Steffes, “New laboratory measurements of the centimeter-
wavelength properties of ammonia under deep jovian atmospheric conditions,”
Bulletin of the American Astronomical Society, vol. 42, 2010, p.1040. Pre-
sented at the 42nd Annual Meeting of the Division for Planetary Sciences of
the American Astronomical Society, Pasadena, CA, October 7, 2010.
K. Devaraj, B. Butler, B. Hesman, P. Steffes, and R. Sault, “VLA observations
of the Jupiter impact,” Geophysical Research Abstracts, vol. 12, 7661. Presented
at the EGU General Assembly, Vienna, Austria, May 2010 (Invited Talk).
K. Devaraj and P. Steffes, “The 2–4 millimeter-wavelength opacity of ammo-
nia,” Geophysical Research Abstracts, vol. 12, 7675. Presented at the EGU
General Assembly, Vienna, Austria, May 2010.
K. Devaraj and P. Steffes, “Laboratory measurements of microwave properties
of ammonia under deep jovian atmospheric conditions,” Lunar and Planetary
Institute, Contribution No. 1553, pp. 1875 -1876, 2010. Presented at the 41st
Lunar and Planetary Science Conference, The Woodlands, TX, March 4, 2010.
K. Devaraj and P. Steffes, “The 2–4 millimeter-wavelength opacity of am-
monia: Extensive laboratory measurements and a new model,” Bulletin of the
165
American Astronomical Society, vol. 41, no. 3, 2009, p. 1049. Presented at
the 41st Annual Meeting of the Division for Planetary Sciences of the American
Astronomical Society, Fajardo, PR, October 7, 2009.
B. Butler, K. Devaraj, P. Steffes, and B. Hesman, “Observations of the Jupiter
impact with the VLA,” Bulletin of the American Astronomical Society, vol. 41,
no. 3, 2009, p. 1194. Presented at the 41st Annual Meeting of the Division for
Planetary Sciences of the American Astronomical Society, Fajardo, PR, October
7, 2009.
B. Hesman, M. Hofstadter, B. Butler, and K. Devaraj, ‘Microwave observa-
tions of Neptune,” Bulletin of the American Astronomical Society, vol. 41, no.
3, 2009, p. 196. Presented at the 41st Annual Meeting of the Division for Plan-
etary Sciences of the American Astronomical Society, Fajardo, PR, October 7,
2009.
K. Devaraj and P. Steffes, “A new laboratory system for measurement of the
millimeter-wave properties of gases and preliminary results for the continuum
opacity of ammonia,” Bulletin of the American Astronomical Society, vol. 40,
no. 3, 2008, p. 497. Presented at the 40th Annual Meeting of the Division for
Planetary Sciences of the American Astronomical Society, Ithaca, NY, October
14, 2008.
M. Hofstadter, B. Butler, M. Gurwell, B. Hesman, and K. Devaraj, ‘The tro-
pospheres of Uranus and Neptune as seen at microwave wavelengths,” Bulletin
of the American Astronomical Society, vol. 40, no. 3, 2008, p. 488. Presented at
the 40th Annual Meeting of the Division for Planetary Sciences of the American
Astronomical Society, Ithaca, NY, October 14, 2008.
K. Devaraj and P. Steffes, “Laboratory measurements of w-band continuum
166
opacity of ammonia using a fully confocal Fabry-Perot resonator,” International
Union of Radio Science Programs and Abstracts: 2008 National Radio Science
Meeting, pp. J1-4. Presented at the 2008 URSI National Radio Science Meeting,
Boulder, CO, January 3, 2008.
K. Devaraj and P. Steffes, ‘Preliminary results for the 2–4 millimeter wave-
length continuum opacity of ammonia based on new laboratory measurements
under simulated jovian conditions,” Bulletin of the American Astronomical So-
ciety, vol. 39, no. 3, 2007, p. 447. Presented at the 39th Annual Meeting of the
Division for Planetary Sciences of the American Astronomical Society, Orlando,
FL, October 9, 2007.
P. Steffes, T. Hanley, B. Karpowicz, and K. Devaraj, “Laboratory measure-
ments of the microwave and millimeter-wave properties of planetary atmo-
spheric constituents: The Georgia Tech system,” Workshop on Planetary At-
mospheres, pp. 117-118. LPI Contribution No. 1376, Lunar and Planetary In-
stitute, Houston. Presented at the 2007 Workshop on Planetary Atmospheres,
Greenbelt, MD, November 6, 2007.
167
APPENDIX A
MILLIMETER-WAVELENGTH COMPONENTS
The millimeter-wavelength subsystem consists of a Fabry–Perot resonator and the
W band and F band components (signal generators and receivers). This appendix
provides an overview of the characteristics of the W band and F band components.
Complete design and construction information of the millimeter-wavelength compo-
nents can be obtained from the manufacturer’s websites.
A.1 W band Components
The W band subsystem consists of the following components along with the Fabry–
Perot resonator.
. Signal Generator (HP 83650B)
. Active Multiplier Chain (Millitech AMC-10-RFH00)
. Harmonic Mixer (QMH 922WHP/387)
. Diplexer (MD1A)
. Spectrum Analyzer (HP 8564E)
The first stage of the two-stage active multiplier chain (AMC) consists of a
Ku band (12–18 GHz) amplifier followed by a frequency doubler, and the second
stage of the AMC consists of a Ka band (26.5–40 GHz) amplifier followed by a fre-
quency tripler. The output power of the AMC varies as a function of frequency and it
is shown in Figure A.1. The harmonic mixer can be used either in an “external mixer”
mode with a spectrum analyzer LO and an MD1A diplexer or as a down-converter
168
9
Power Output (dBm)

6
1
75 80 85 90 95 100 105 110
Frequency (GHz)
Figure A.1: Power output of the AMC for Pin =+10 dBm.
with an external LO (fLO up to 18 GHz) and MD2A diplexer. The AMC, mixer, and
diplexer are are shown in Figure A.2. Tables A.1 and A.2 list some of the parameters
of the AMC and mixer, respectively.
When the components are placed at cold temperatures, bias is continuously pro-
vided to the AMC, even if the system is not under operation, to ensure that the
heat dissipated by the AMC along with the thermal insulation layers maintain the
(a) Active multiplier chain. (b) Harmonic mixer (c) Diplexer

(QMH). (MD1A).
Figure A.2: Components of the W band subsystem.
169
Table A.1: AMC-10-RFH00 parameters.
Output frequency 75–110 GHz
Input frequency 12.50–18.33 GHz
Multiplication factor 6
Input power +10 dBm (nom)
Output power see Figure A.1
Maximum input power +13 dBm
Signal purity (max) -20 dBc
DC input 8–12 V @ 600 mA (typ)
Operating temperature 0◦ C to +45◦ C
Table A.2: Quinstar QMH 922WHP/387 parameters in “external mixer” mode.

Input frequency 75–110 GHz
Input power -10 dBm (max)
LO frequency (from spectrum analyzer) 3–6.1 GHz
LO power (from spectrum analyzer) 15 dBm
Harmonic number 18
Conversion loss 40 dB (nom)
Operating temperature -55◦ C to +125◦ C
AMC within its operating temperature at all times. It is critical to prevent condensa-
tion from occurring inside the components when they are used at cold temperatures.
Hence, all the components are hermetically sealed in an enclosure with some silica
gel to prevent condensation. The silica gel beads are periodically replaced to ensure
minimum water vapor content inside the enclosure.
A.2 F band Subsystem
The F band subsystem consists of the following components along with the Fabry–
Perot resonator.
. Signal Generator (HP 83650B)
. Q band Amplifier (Spacek SG4413-15-16W)
. Frequency Tripler (Pacific Millimeter Products D3WO)
. Harmonic Mixer (Pacific Millimeter Products DM)
170
(a) Spacek amp. (b) Tripler.
Figure A.3: F band components: Spacek amplifier and frequency tripler.
. Diplexer (Pacific Millimeter Products MD2A)
. LO Amplifier (JCA 1920-612)
. Local Oscillator (HP 83712B)
. Bias Tee
. Low Noise Amplifier (MITEQ AMF-3F-012017)
. Spectrum Analyzer (HP 8564E)
A swept signal generator (HP 83650B) is used to generate the input signal in
the 37 to 50 GHz range (Q band) that is amplified (Spacek amplifier) and fed to
a frequency tripler to generate the desired RF in the 2 to 3 mm-wavelength range
(Figure A.3). The output power of the Spacek amplifier and the tripler are shown in
Figure A.4 and some additional parameters are listed in Table A.3.
A harmonic mixer and the diplexer (model MD2A) are used for down-converting
the RF signals (Figure A.5). The conversion loss of the harmonic mixer when used
with the spectrum analyzer LO in the “external mixer” mode is > 70 dB. Hence, the
harmonic mixer is used in the “down-converter” mode with an MD2A diplexer and
an external LO. Increasing the LO frequency reduces the conversion loss, for fLO up
to 18 GHz. Hence, the lowest possible harmonic (7–10) is chosen for down-converting
171
22
21
20
Power Output (dBm)

19
18
17
16
15
36 38 40 42 44 46 48 50
Frequency (GHz)
(a) Power output of the Spacek amplifier

(P1dBoutput ).
2
Power Output (dBm)
−1
−2
110 120 130 140 150 160 170
Frequency (GHz)
(b) Power output of the tripler (Pin =+20 dBm).
Figure A.4: Power output of the Spacek amplifier and the tripler.
Table A.3: Spacek amplifier and tripler parameters.

Spacek amplifier Frequency 33–50 GHz
Gain 15 dB (min) / 19 dB (typ)
P-1dB 17.5 dBm (typ) / 15 dBm (min)
Psat 19 dBm (typ)
VSWR in/out 2:1 typ
Bias 375 mA @ +8 to +12 VDC
RF connectors K-female In / WR-22 Out
Operating temperature 0◦ C to +50◦ C
Frequency tripler Input frequency 33–56.7 GHz
Output frequency 100–170 GHz
Input power +15 dBm
172
(a) Harmonic (b) Diplexer.
mixer.
Figure A.5: F band harmonic mixer and diplexer.
(a) JCA amp. (b) LNA. (c) Bias tee.
Figure A.6: JCA/Miteq amplifiers and bias tee.
the RF signal while keeping the IF constant at 1 GHz. A JCA amplifier is used to
amplify the LO signals and a spectrum analyzer is used to view the IF via a MD2A
diplexer. Bias is provided to the harmonic mixer via a bias tee. An LNA amplifies
the signals prior to displaying in the spectrum analyzer. The JCA amplifier, bias tee,
and the LNA are shown in Figure A.6. The power output of the JCA amplifier and
the S-parameters of the bias tee are shown in Figure A.7.
RFI was reduced by wrapping aluminum foil tapes around the bias tee, LNA, and
all the microwave adapters. A 1 k ohm series resistor was added to the biastee and
a bias voltage of 8VDC was provided. Without this series resistance, the optimum
bias-voltage that maximizes the system sensitivity is 388 mVDC . The amplifiers and
the mixer must be biased before any millimeter wave energy can be provided to the
components.
173
18
17
16
Power Output (dBm)
15
14
13
12
11
10
10 11 12 13 14 15 16 17 18 19 20
Frequency (GHz)
(a) Power output of the JCA amp (Pin =-5 dBm).
(b) S-parameters of the bias tee.
Figure A.7: Output component parameters.
174
APPENDIX B
VLA OBSERVATIONS OF THE 2009 JUPITER IMPACT
EVENT
Jupiter was bombarded by an object of unknown origin on UT July 19, 2009 (Sánchez-
Lavega et al., 2010). The collision itself was not recorded, but Anthony Wesley (an
amateur astronomer) spotted an anomalous feature centered at ∼ 305◦ W longitude

◦
and ∼ 58 S planetographic latitude in System III (λIII ) coordinates within hours
after the impact. The Hubble image of Jupiter and its impact spot is shown in Fig-
ure B.1. Visible and infrared observations that followed confirmed that the feature
was exogenic and it spanned a total area of ∼ 4800 km (east-west) x 2500 km (north-
south) (de Pater et al., 2010; Sánchez-Lavega et al., 2010; Fletcher et al., 2010).
Since different wavelengths probe different regions in the atmosphere of Jupiter,
multi-wavelength observations of the impact, spanning radio, infrared, and visible
wavelengths, are best suited for studying the impact characteristics, such as the im-
pact dynamics, direction from which the impactor arrived, depth penetration of the
impactor, and the type of body involved.
175
Figure B.1: Hubble image of Jupiter and its impact spot. Image credit: NASA, ESA,
H. Hammel (Space Science Institute, Boulder, CO), and the Jupiter impact team.
B.1 VLA Observations of the Impact Event
The Jupiter impact event of 2009 was observed with the Very Large Array (VLA)
situated on the Plains of San Agustin near Socorro, New Mexico, operated by the
National Radio Astronomy Observatory (NRAO). The VLA consists of 27 antennas,
each with a diameter of 25 m. The antennas are arranged in a Y-shape with nine
antennas on each arm of the Y. There are four basic configurations for the VLA: A, B,
C, and D. The A-configuration is the largest (highest resolution) and D-configuration
is the most compact. In order to observe Jupiter using the VLA immediately after
the impact, three rapid response-target of opportunity proposals were written and
submitted to NRAO (VLA proposal IDs: VLA/09B-205, VLA/09B-206, VLA/09B-
208). The proposals were reviewed by a panel of scientists and later accepted by the
NRAO scheduling committee and a total of 38.5 hours of observation time was alloted
between July 22, 2009 and August 10, 2009.
176
The VLA observations of Jupiter were conducted when the array was in the C-
configuration, with a maximum physical antenna separation of 3.4 km. The observa-
tions were conducted at L, C, X, and K bands. During the observations, the L band
electronics were tuned to 1.3851 and 1.4649 GHz, C band to 4.86 GHz, X band to
8.44 GHz, and K band to 22.46 GHz. Continuum mode was used in all the obser-
vations, and the four Stokes parameters were measured in two 50 MHz passbands.
This provides an equivalent bandwidth of 100 MHz when the two passbands are av-
eraged together (except the two L band frequencies, where the IFs are not averaged
together), and the total intensity images (Stokes I) are formed (see, e.g., Butler et al.,
2001). During the observations, Jupiter was approximately 4 AU from Earth and
had an angular diameter of 48 arcsec. The astrometric right ascension was ∼ 21 h
45 m and the declination was ∼ -14 deg 36 m. Quasar 3C286 or 3C48 was used
as the primary calibrator (flux), while the secondary calibrator (phase) was selected
depending on which phase calibrator was approximately at the same elevation in the
sky during the time of observation.
A summary of the observations along with the geometry of the planet visible
from Earth and the duration of observations is listed in Table B.1. The observations
were split into 1-4 hour time slots such that the impact site was visible as viewed
from Earth at the time of the observation. The sidereal rotation rate of Jupiter is
approximately 9 h 55 m, so the coordinates of the impact location rotate across the
disk during the course of the observation. An example of the 1.3 cm observation made
on August 1, 2009 is shown in Figure B.2.
The total emission from Jupiter at radio wavelengths is due to the thermal com-
ponent from the planet as well as the synchrotron component from the relativistic
electrons spiraling in the magnetosphere of the planet (see, e.g., Berge and Gulkis,
1976). At long wavelengths, synchrotron emission dominates and at short wavelengths
thermal emission dominates. The cross-over is at approximately 10 cm (Berge and
177
Table B.1: Summary of the VLA observations of Jupiter along with the duration of
observation and the geometry of the planet visible from Earth.
Date Wavelength λIII longitude Duration
(cm) (hours)
July 22, 2009 3.5 197-288 1.5
July 23, 2009 3.5 257-293 1
July 26, 2009 3.5 258-7 3
July 27, 2009 3.5 261-327 2
July 29, 2009 6, 20 87-232 4
July 30, 2009 6, 20 153-292 4
August 1, 2009 6, 20 82-227 4
August 1, 2009 1.3 227-336 3
August 5, 2009 1.3 259-7 3
August 6, 2009 6, 20 104-249 4
August 6, 2009 1.3 249-358 3
August 10, 2009 1.3 268-340 2
August 10, 2009 6, 20 340-125 4
Gulkis, 1976). For the L and C band observations, the contribution to the total
emission is mostly from the synchrotron radiation, and for the X and K band ob-
servations, the major contribution is from Jupiter’s atmospheric thermal emission.
Different wavelengths have different thermal and synchrotron contributions and they
also probe to different depths in the atmosphere, with the longer wavelengths prob-
ing deeper into the atmosphere of the planet (see, e.g., Janssen et al., 2005). The
weighting functions indicate the altitudes (pressure layers) that contribute most to
the brightness temperature at a particular frequency. The normalized nadir-viewing
weighting function in the 0.5–25 GHz range computed by Karpowicz (2010) using a
radiative transfer model for the atmosphere of Jupiter is shown in Figure 5.58. The
3.5 cm wavelength probes the atmospheric layers between 0.8 bar and 3 bar and the
1.3 cm wavelength probes the atmospheric layers between 0.3 bar and 0.8 bar.
When an object bombards Jupiter, a plume of material from the object as well as
from Jupiter’s atmosphere is brought up from Jupiter’s troposphere into the strato-
sphere. Hence, there will be a change in the composition of the stratosphere as well
178
Figure B.2: Jupiter observation geometry on August 1, 2009. The red dots show the
progression of the coordinates of the impact site as the location rotates across the
disk during the observation.
as the troposphere at the impact site. Enhanced emission associated with ammo-
nia gas and an increase in temperature in the upper troposphere at the site of the
impact were observed in Jupiter following the 2009 impact event (de Pater et al.,
2010; Fletcher et al., 2010). The 1.3 and 3.5 cm wavelengths probe the upper and
middle troposphere of Jupiter. The VLA is capable of measuring changes due to the
impact because of its sensitivity and resolution. Unfortunately the array was not in
a high-resolution configuration at the time of the impact. The VLA resolution in
C-configuration at 3.5 cm is ∼ 2.3 arcsec (6800 km linear at Jupiter) and at 1.3 cm
is ∼ 0.8 arcsec (2300 km linear at Jupiter). Additionally, it was summer in New
Mexico; so observing at high frequencies was difficult because of the North American
Monsoon. In spite of these difficulties, observations were conducted at 1.3 and 3.5 cm
wavelengths to study the changes to the thermal emission from Jupiter, and at 6 and
20 cm wavelengths to study the changes to the synchrotron emission from Jupiter.
179
B.2 Data Reduction and Analysis
The calibration of the data was performed in the normal fashion for VLA data,
using the AIPS reduction package (http://www.cv.nrao.edu/aips/). A detailed
description of the initial calibration and self-calibration of the visibilities obtained for
the observations of Venus is provided by Butler et al. (2001) and a similar approach
was followed for the calibration of Jupiter observations in this study. After the self-
calibration process, the data product is a set of fully calibrated visibilities for Jupiter.
These visibilities were used to make maps of the received flux density, in units of Jy,
across the visible disk of Jupiter. Initial data reductions were performed for the
1.3 and 3.5 cm observations. The synchrotron component of the emission at 3.5 cm
is approximately 10% and at 1.3 cm is < 1%. Hence, at 3.5 cm, it is possible to
observe both thermal and synchrotron emissions. Some approximations about the
spatial extent of the synchrotron belts were made, and post processing was done
on the reduced 3.5 cm maps of Jupiter to seperate the thermal and synchrotron
components. Figure B.3 shows the maps of the thermal and synchrotron component
of Jupiter at 3.5 cm wavelength made on July 26, 2009. Figure B.4 shows the maps
of the thermal component of Jupiter at 1.3 cm wavelength made on August 01, 2009
and August 06, 2009. Work is in progress to reduce and analyze the 6 and 20 cm
observations and to study the changes to the synchrotron emission due to the impact
at these wavelengths.
Initial data reduction and analysis show no significant changes to the thermal
component because of the impact at both the observation wavelengths (Devaraj et al.,
2010). Using the method by Sault et al. (2004), the 3.5 cm images of Jupiter obtained
over four days between July 22–27, 2009 were rotationally deprojected into a planetary
cartographic system. The resulting images were averaged together and is shown in
Figure B.5 and the impact site is highlighted with a black oval. It can be seen from the
figure that no signature of the impact on the thermal emission at 3.5 cm, at a depth
180
(a) Thermal component.
(b) Synchrotron component.
Figure B.3: Maps of Jupiter at 3.5 cm made on July 26, 2009.

181
(a) August 01, 2009.
(b) August 06, 2009.
Figure B.4: Maps of Jupiter showing the thermal emission at 1.3 cm.
182
Figure B.5: Rotationally deprojected map of Jupiter at 3.5 cm averaged over obser-
vations from July 22–27, 2009. A black oval highlights the impact site. No signature
of the impact on the thermal emission at 3.5 cm, at a depth of ∼ 1 bar, and with
effective resolution of ∼ 6800 km, was detected.
of ∼ 1 bar, and with effective resolution of ∼ 6800 km, was detected. Preliminary
analysis of the 1.3 cm thermal emission data (depth of ∼ 500 mbar and an effective
resolution of ∼ 2300 km) also indicates that no signature of the impact was detected.
183
APPENDIX C
VLA OBSERVATIONS OF VENUS AT X BAND
Previous observations of Venus were conducted at X band (3.6 cm) in 1996 and new
observations were conducted as part of this work in 2009 with the NRAO VLA. These
observations have now been calibrated, reduced, and analyzed in a consistent fashion.
Spatial variations were observed in the microwave emission from Venus originating
from the deep atmosphere.
C.1 VLA Observations
Observations of Venus were made at X band using the VLA on April 30, 1996 by But-
ler et al. (2001) and July 07, 2009 by this author. The VLA was in C-configuration
during both the observations, and at 3.6 cm, the angular resolution of the array is
∼ 2.3 arcsec. Continuum mode was used in all the observations. A summary of
the ephemeris information and the calibrators used for the observations is listed in
Table C.1.
Table C.1: Ephemeris information and the calibrators for Venus observations.
Observation Date April 30, 1996 July 07, 2009
Time range (UT) 0148-0224 0240–0450
Right ascension 05 h 21 min 04 h 03 min
Declination +27 deg 39 min +18 deg 05 min
Distance from Earth (AU) 0.486 0.947
Angular diameter (arcsec) 34.28 17.61
Primary calibrator 3C286 3C48
Secondary calibrator 0555+398 0344+159
Figure C.1 shows the disk of Venus as seen from Earth (left) and the radar map
of the surface of Venus on April 30, 1996 and Figure C.2 shows the disk and radar
maps of Venus on July 07, 2009.
184
Figure C.1: Disk map of Venus as seen from Earth (left) and radar surface map of
Venus (right) on April 30, 1996. Image source: US Naval Observatory.
Figure C.2: Disk map of Venus as seen from Earth (left) and radar surface map of
Venus (right) on July 07, 2009. Image source: US Naval Observatory.
185
C.2 Data Analysis
Calibration and reduction of the Venus observations were performed using the AIPS
data-reduction package in a fashion similar to the Jupiter observations explained in
Appendix B.2, and maps of the 3.6 cm emission from Venus were created. Figure C.3
shows the reduced X band map of Venus for the 2009 observation. The reduced maps
of Venus were referenced to a radially-averaged, limb-darkened disk and residual maps
were created. These residual maps are shown in Figures C.4 and C.5 for the two
observations.
Significant structures are visible on the residual maps. Of note are the bright/dark
regions around the limb of the planet, which correspond to the variations in the
abundance of microwave-opaque sulfuric acid vapor below the cloud base, caused
by the Hadley cell circulation. (The limb-viewing weighing functions correspond
to such altitudes.) However, the features in the center of the disk correspond to
variations in either temperature or microwave absorbing gases in the deepest part
of the troposphere. At X band, emission from the disk center is from the 0-15 km
altitude range (see, Figure C.6).
Sulfur dioxide (SO2 ) and gaseous sulfuric acid (H2 SO4 ) dominate the centimeter-
wavelength emission from Venus (see, e.g., Butler et al., 2001; Jenkins et al., 2002).
At altitudes below 35 km, gaseous H2 SO4 thermally dissociates and forms H2 O and
SO3 , both of which exhibit relatively small amounts of microwave absorption at the
abundance levels present in the Venus atmosphere. Thus, in the deep atmosphere of
Venus, only SO2 , and to a lesser extent OCS have the potential to affect the observed
microwave emission. The collisionally-induced absorption from CO2 has a significant
effect on the microwave emission from Venus, but it is not in any way localized
since CO2 is the dominant atmospheric constituent (∼ 96%). Hence, it is critical to
study the centimeter-wavelength properties of SO2 in the deep Venus atmospheric
conditions.
186
Figure C.3: X band map of Venus taken with the VLA on July 07, 2009.
187
Figure C.4: Residual X band map of Venus for the July 07, 2009 observation.
Figure C.5: Residual X band map of Venus for the April 30, 1996 observation.
188
189
Figure C.6: The TP profile of Venus (left) and the weighting functions at the disk center (right).
Laboratory measurements of the centimeter-wavelength opacity of SO2 under
Venus atmospheric conditions were conducted by Suleiman et al. (1996) at pressures
up to 4 bar. However, the weighting functions for Venus central disk show that the
contribution to the 3.6 cm emission is from altitudes with pressures in the 40–92 bar
region (Figure C.6). Hence, laboratory measurements of the microwave properties
of SO2 under simulated deep tropospheric conditions of Venus are currently being
planned with the high-pressure laboratory system described in this dissertation. The
results of the laboratory study will aid in interpreting the observed spatial variations
in the microwave emission from Venus originating from the deep atmosphere.
190
Bibliography
Atreya, S., Mahaffyy, P., Niemann, H., Wong, M., Owen, T., 2003. Composition and
origin of the atmosphere of Jupiter–an update, and implications for the extrasolar
giant planets. Planet. Space Sci. 51, 105–112.
Bachet, G., 1973. Etude des elargissements de la raie de rotation a 19 cm−1 de NH3
perturbee par des gaz etrangers comprimes. J. Quant. Spectrosc. Radiat. Transfer
13, 1305–1308.
Becker, G. E., Autler, S. H., 1946. Water vapor absorption of electromagnetic radia-
tion in the centimeter wave-length range. Phys. Rev. 70, 300–307.
Belli, S., Buffa, G., Tarrini, O., 1997. On the extension of the Rydberg-Ritz principle
to the collisional relaxation of molecular and atomic lines. Chem. Phys. Lett. 193,
277–284.
Belov, S. P., Kazakov, V. P., Krupnov, A. F., Markov, V. N., Mel’nikov, A. A.,
Skvortsov, V. A., Tret’yakov, M. Y., 1982. The study of microwave pressure
linewidths. J. Mol. Spectrosc. 94, 264–282.
Belov, S. P., Krupnov, A. F., Markov, V. N., Mel’nikov, A. A., Skvortsov, V. A.,
Tret’yakov, M. Y., 1983. Study of microwave pressure lineshifts: dynamic and
isotopic dependences. J. Mol. Spectrosc. 101, 258–270.
Ben-Reuven, A., 1966. Impact broadening of microwave spectra. Phys. Rev. 145, 7–22.
Berge, G. L., Gulkis, S., 1976. Earth-based radio observations of Jupiter: millimeter
to meter wavelengths. In: Gehrels, T. (Ed.), Jupiter: Studies of the Interior, At-
mosphere, Magnetosphere, and Satellites. Univ. of Arizona Press, Tuscon, AZ, pp.
621–692.
Bleaney, B., Penrose, R. P., 1947. The inversion spectrum of ammonia at centimeteric
wavelengths. Proc. Roy. Soc. A189, 358.
Brown, L. R., Peterson, D. B., 1994. An empirical expression for linewidths of am-
monia from far-infrared measurements. J. Mol. Spectrosc. 168, 593–606.
Burke, B. F., Franklin, K. L., 1955. Observations of a variable radio source associated
with the planet Jupiter. J. Geophys. Res. 60, 213–217.
Butler, B. J., Steffes, P. G., Suleiman, S. H., Kolodner, M. A., Jenkins, J. M., 2001.
Accurate and consistent microwave observations of Venus and their implications.
Icarus 154, 226–238.
Cleeton, C. E., Williams, N. H., 1934. Electromagnetic waves of 1.1 cm wave-length

and the absorption spectrum of ammonia. Phys. Rev. 45, 234–237.
191
Culshaw, W., 1960. High resolution millimeter wave Fabry-Perot interferometer. IRE
Trans. Microwave Theory Tech. 8, 182–189.
Culshaw, W., 1961. Resonators for millimeter and submillimeter wavelengths. IRE
Trans. Microwave Theory Tech. 9, 135–144.
Culshaw, W., 1962. Futher considerations on Fabry-Perot type resonators. IRE Trans.
Microwave Theory Tech. 10, 331–339.
Danos, M., Geschwind, S., 1953. Broadening of microwave absorption lines due to
wall collisions. Phys. Rev. 91, 1159–1162.
de Pater, I., DeBoer, D., Marley, M., Freedman, R., Young, R., 2005. Retrieval
of water in Jupiter’s deep atmosphere using microwave spectra of its brightness
temperature. Icarus 173, 425–438.
de Pater, I., Dunn, D., Romani, P., Zahnle, K., 2001. Reconciling Galileo Probe data
and ground-based radio observations of ammonia on Jupiter. Icarus 149, 66–78.
de Pater, I., Fletcher, L. N., Pérez-Hoyos, S., Hammel, H. B., Orton, G. S.,
Wong, M. H., Luszcz-Cook, S., Sánchez-Lavega, A., Boslough, M., 2010. A multi-
wavelength study of the 2009 impact on Jupiter: Comparison of high resolution
images from Gemini, Keck and HST. Icarus 210, 722–741.
de Pater, I., Kenderdine, S., Dickel, J. R., 1982. Comparison of the thermal and
nonthermal radiation characteristics of Jupiter at 6, 11, and 21 cm with model
calculations. Icarus 51, 25–38.
de Pater, I., Mitchell, D., 1993. Radio observations of the planets: the importance of
laboratory measurements. J. Geophys. Res. 98, 5471–5490.
DeBoer, D. R., Steffes, P. G., 1994. Laboratory measurements of the microwave prop-
erties of H2 S under simulated Jovian conditions with an application to Neptune.
Icarus 109, 352–366.
DeBoer, D. R., Steffes, P. G., 1996. Estimates of the tropospheric vertical structure
of Neptune based on microwave radiative transfer studies. Icarus 123, 323–335.
Debye, P. J. W., 1929. Polar Molecules. The Chemical Catalog Company Inc., New
York.
Devaraj, K., Butler, B., Hesman, B., Steffes, P., Sault, R., 2010. VLA observations
of the Jupiter impact. Geophys. Res. Abs. 12.
Devaraj, K., Steffes, P. G., 2011. The Georgia Tech millimeter-wavelength measure-
ment system and some applications to the study of planetary atmospheres. Radio
Sci., RS2014, doi:10.1029/2010RS004433.
192
Devaraj, K., Steffes, P. G., Karpowicz, B. M., 2011. Reconciling the centimeter-
and millimeter-wavelength ammonia absorption spectra under jovian conditions:
Extensive millimeter-wavelength measurements and a consistent model. Icarus 212,
224–235.
Dowling, T. E., Muhleman, D. O., Berge, G. L., 1987. Aperture synthesis observations
of Saturn and its rings at 2.7-mm wavelength. Icarus 70, 506–516.
Dunn, D. E., de Pater, I., Wright, M., Hogerheijde, M. R., Molnar, L. A., 2005.
High-quality BIMA-OVRO images of Saturn and its rings at 1.3 and 3 millimeters.
Astron. J. 129, 1109–1116.
Fahd, A. K., Steffes, P. G., 1992. Laboratory measurements of the microwave and
millimeter-wave opacity of gaseous sulfur-dioxide (SO2 ) under simulated conditions
for the Venus atmosphere. Icarus 97, 200–210.
Fletcher, L. N., Orton, G., de Pater, I., Edwards, M., Yanamandra-Fisher, P., Ham-
mel, H., Lisse, C., , Fisher, B., 2010. The aftermath of the July 2009 impact on
Jupiter: Ammonia, temperatures and particulates from Gemini thermal infrared
spectroscopy. Icarus 211, 568–586.
Folkner, W. M., Woo, R., Nandi, S., 1998. Ammonia abundance in Jupiter’s atmo-
sphere derived from the attenuation of the Galileo Probe’s radio signal. J. Geophys.
Res. 103, 22847–22856.
Gibson, J., Welch, W. J., de Pater, I., 2005. Accurate jovian radio flux density mea-
surements show ammonia to be subsaturated in the upper troposphere. Icarus 173,
439–446.
Gordy, W., 1948. Microwave spectroscopy. Rev. Mod. Phys. 20.
Griffin, M. J., Ade, P. A. R., Orton, G. S., Robson, E. I., Gear, W. K., Nolt, I. G.,
Radostitz, J. V., 1986. Submillimeter and millimeter observations of Jupiter. Icarus
65, 244–256.
Griffin, M. J., Orton, G. S., 1993. The near-millimeter brightness temperature spectra
of Uranus and Neptune. Icarus 105, 537–547.
Gross, E. P., 1955. Shape of collision-broadened spectral lines. Phys. Rev. 97, 395–403.
Hanley, T. R., 2008. The microwave opacity of ammonia and water vapor: Application
to remote sensing of the atmosphere of Jupiter. PhD dissertation, Georgia Institute
of Technology, Atlanta, GA.
Hanley, T. R., Steffes, P. G., 2007. A high-sensitivity laboratory system for mea-
suring the microwave properties of gases under simulated conditions for planetary
atmospheres. Radio Sci. 42 (RS6010).
193
Hanley, T. R., Steffes, P. G., Karpowicz, B. M., 2009. A new model of the hydrogen
and helium-broadened microwave opacity of ammonia based on extensive labora-
tory measurements. Icarus 202, 316–335.
Herriott, D. R., Kogelnik, H., Kompfner, R., 1964. Off axis paths in spherical mirror
interferometers. Appl. Opt. 3, 523–526.
Hesman, B. E., Davis, G. R., Matthews, H. E., Orton, G. S., 2007. The abundance
profile of CO in Neptune’s atmosphere. Icarus 186, 342–353.
Hewlett-Packard, 1997. HP 8560 E-Series Spectrum Analyzer Calibration Guide.

Santa Rosa, CA.
Hoffman, J. P., Steffes, P. G., DeBoer, D. R., 2001. Laboratory measurements of

the microwave opacity of phosphine: Opacity formalism and application to the
atmospheres of the outer planets. Icarus 152, 172–184.
Irwin, P., 2003. Giant planets of our solar system: Atmospheres, composition, and
structure. Springer-Praxis, Chichester, UK.
Jacquinet-Husson, N., Crepeau, L., Armante, R., Boutammine, C., Chédin, A., Scott,
N. A., Crevoisier, C., Capelle, V., Boone, C., Poulet-Crovisier, N., Barbe, A.,
Campargue, A., Benner, D. C., Benilan, Y., Bézard, B., Boudon, V., Brown, L. R.,
Coudert, L. H., Coustenis, A., Dana, V., Devi, V. M., Fally, S., Fayt, A., Flaud,
J. M., Goldman, A., Herman, M., Harris, G. J., Jacquemart, D., Jolly, A., Kleiner,
I., Kleinböhl, A., Kwabia-Tchana, F., Lavrentieva, N., Lacome, N., Xu, L.-H.,
Lyulin, O. M., Mandin, J. Y., Maki, A., Mikhailenko, S., Miller, C. E., Mishina,
T., Moazzen-Ahmadi, N., Müller, H. S. P., Nikitin, A., Orphal, J., Perevalov, V.,
Perrin, A., Petkie, D. T., Predoi-Cross, A., Rinsland, C. P., Remedios, J. J., Rotger,
M., Smith, M. A. H., Sung, K., Tashkun, S., Tennyson, J., Toth, R. A., Vandaele,
A. C., , Auwera, J. V., 2011. The 2009 edition of the GEISA spectroscopic database.
J. Quant. Spectrosc. Radiat. Transfer 112, 2395–2445.
Janssen, M. A., Hofstadter, M. D., Gulkis, S., Ingersoll, A. P., Allison, M., Bolton,
S. J., Levin, S. M., Kamp, L. W., 2005. Microwave remote sensing of Jupiter’s
atmosphere from an orbiting spacecraft. Icarus 173, 447–453.
Jenkins, J. M., Kolodner, M. A., Butler, B. J., Suleiman, S. H., G.Steffes, P., 2002.
Microwave remote sensing of the temperature and the distribution of sulfur com-
pounds in the Venus atmosphere. Icarus 158, 312–328.
Jenkins, J. M., Steffes, P. G., 1988. Constraints on the microwave opacity of gaseous
methane and water vapor in the jovian atmosphere. Icarus 76, 378–382.
Joiner, J., Steffes, P. G., 1991. Modeling of Jupiter’s millimeter wave emission utilizing
laboratory measurements of ammonia (NH3 ) opacity. J. Geophys. Res. 96, 17,463–
17,470.
194
Karpowicz, B. M., 2010. In search of water vapor on Jupiter: Laboratory measure-
ments of the microwave properties of water vapor and simulations of Jupiter’s
microwave emission in support of the Juno mission. PhD dissertation, Georgia In-
stitute of Technology, Atlanta, GA.
Karpowicz, B. M., Steffes, P. G., 2011. In search of water vapor on Jupiter: Labo-
ratory measurements of the microwave properties of water vapor under simulated
jovian conditions. Icarus 212, 210–223.
Kramer, C., Moreno, R., Greve, A., 2008. Long-term observations of Uranus and
Neptune at 90 GHz with IRAM 30 m telescope. Astron. Astrophys. 482, 359–363.
Kunde, V., Hanel, R., Maguire, W., Gautier, D., Baluteau, J. P., Marten,
A., 1982. The tropospheric gas composition of Jupiter’s north equatorial belt
(NH3 , PH3 , CH3 D, GeH4 , H2 O) and the Jovian D/H isotopic ratio. Astrophys. J.
263, 443–467.
Lamarre, J. M., Puget, J. L., Bouchet, F., Ade, P. A. R., Benoit, A., Bernard, J.,
Bock, J., Bernardis, P. D., Charra, J., Couchot, F., Delabrouille, J., Efstathiou,
G., Giard, M., Guyot, G., Lange, A., Maffei, B., Murphy, A., Pajot, F., Piat,
M., Ristorcelli, I., Santos, D., Sudiwala, R., Sygnet, J., Torre, J., Yurchenko, V.,
Yvon, D., 2003. The Planck high frequency instrument, a third generation CMB
experiment, and a full sky submillimeter survey. New Astron. Rev. 47, 1017–1024.
Law, S. E., Staelin, D. H., 1968. Measurements of Venus and Jupiter near 1-cm
wavelength. Astrophys. J. 154, 1077–1086.
Leachman, J. W., Jacobsen, R. T., Penoncello, S. G., Lemmon, E. W., 2009. Funda-
mental equations of state for parahydrogen, normal hydrogen, and orthohydrogen.
J. Phys. Chem. Ref. Data 38, 721–748.
Lemmon, E. W., Huber, M. L., McLinden, M. O., 2007. NIST Standard Refer-
ence Database 23: Reference Fluid Thermodynamic and Transport Properties-
REFPROP, Standard Reference Data Program Version 8.0. National Institute of
Standards and Technology, Gaithersburg.
Levenberg, K., 1944. A method for the solution of certain nonlinear problems in least
squares. Quart. Appl. Math. 2, 164–168.
Lindal, G. F., 1992. The atmosphere of Neptune - an analysis of radio occultation

data acquired with Voyager 2. Astron. J. 103, 967–982.
Lindal, G. F., Sweetnam, D. N., Eshleman, V. R., 1985. The atmosphere of Saturn
- an analysis of the voyager radio occultation measurements. Astron. J. 90, 1136–
1146.
Lorentz, H. A., 1906. The width of spectral lines. Proc. Neth. Acad. Arts Sci. 18,
134–150.
195
Marquardt, D., 1963. An algorithm for least-squares estimation of nonlinear param-
eters. SIAM J. Appl. Math. 11, 431–441.
Matthaei, G. L., Young, L., Jones, E., 1980. Microwave filters, impedance matching
networks and coupling structures. McGraw-Hill, New York.
McCarty, R. D., Arp, V. D., 1990. A new wide range equation of state for helium.
Advances in Cryogenic Engineering 35, 1465–1475.
Mohammed, P. N., Steffes, P. G., 2003. Laboratory measurements of the Ka-band

(7.5 to 9.2 mm) opacity of phosphine (PH3 ) and ammonia (NH3 ) under simulated
conditions for the Cassini-Saturn encounter. Icarus 166 (423–435).
Mohammed, P. N., Steffes, P. G., 2004. Laboratory measurements of the W band (3.2
mm) properties of phosphine (PH3 ) and ammonia (NH3 ) under simulated conditions
for the outer planets. J. Geophys. Res. 109 (E07S13).
Morris, E. C., 1971. Microwave absorption by gas mixtures at pressures upto several
hundred bars. II. Discussion of results. Aust. J. Phys. 24, 157–175.
Morris, E. C., Parsons, R. W., 1970. Microwave absorption by gas mixtures at pres-
sures upto several hundred bars. I. Experimental technique and results. Aust. J.
Phys. 23, 335–349.
Muhleman, D. O., Berge, G. L., 1991. Observations of Mars, Uranus, Neptune, Io,
Europa, Ganymede and Callisto at a wavelength of 2.66 mm. Icarus 92, 263–272.
Nethercot, A. H., Klein, J. A., Loubser, J. H. N., Townes, C. H., 1952. Spectroscopy
near the boundary between the microwave and infrared regions. Suppl. Nuovo Ci-
mento 9, 358–363.
Niemann, H. B., Atreya, S. K., Carignan, G. R., Donahue, T. M., Haberman, J. A.,
Harpold, D. N., Hartle, R. E., Hunten, D. M., Kasprzak, W. T., Mahaffy, P. R.,
Owen, T. C., Spencer, N. W., Way, S. H., 1996. The Galileo probe mass spectrom-
eter: Composition of Jupiter’s atmosphere. Science 272, 846–849.
Orton, G. S., Fisher, B. M., Baines, K. H., Stewart, S. T., Friedson, A. J., Ortiz,
J. L., Marinova, M., Ressler, M., Dayal, A., Hoffmann, W., Hora, J., Hinkley, S.,
Krishnan, V., Masanovic, M., Tesic, J., Tziolas, A., Parija, K. C., 1998. Character-
istics of the Galileo Probe entry site from Earth-based remote sensing observations.
J. Geophys. Res. 103, 22791–22814.
Orton, G. S., Gustafsson, M., Burgdorf, M., Meadows, V., 2007. Revised ab initio
models for H2 − H2 collision-induced absorption at low temperatures. Icarus 189,
544–549.
Overstreet, R., Giauque, W. F., 1937. Ammonia. The heat capacity and vapor pres-
sure of solid and liquid. Heat of caporization. The entropy values from thermal and
spectroscopic data. J. Am. Chem. Soc. 59, 254–259.
196
Pickett, H. M., Poynter, R. L., Cohen, E. A., Delitsky, M. L., Pearson, J. C., Müller,
H. S. P., 1998. Submillimeter, millimeter, and microwave spectral line catalog. J.
Quant. Spectrosc. Radiat. Transfer 60, 883–890.
Pine, A. S., Markov, V. N., Buffa, G., Tarrini, O., 1993. N2 , O2 , H2 , Ar and He
broadening in the ν1 band of NH3 . J. Quant. Spectrosc. Radiat. Transfer 50, 337–
348.
Poynter, R. L., Kakar, R. K., 1975. The microwave frequencies, line parameters, and
spectral constants for 14NH3 . Astrophys. J. Suppl. Ser. 29, 87–96.
Pozar, D. M., 1998. Microwave Engineering. 2nd ed., Wiley and Sons, New York.
Ramo, S., Whinnery, J. R., Duzer, T. V., 1994. Fields and Waves in Communications
Electronics. Wiley and Sons, New York.
Rodrigues, C. C., Moraes Jr., D., 2002. Control of the emission of ammonia through
adsorption in a fixed bed of activated carbon. Adsorpt. Sci. Technol. 20, 1013–1022.
Rosenkranz, P. W., 1998. Water vapor microwave continuum absorption: A compar-

ison of measurements and models. Radio Sci. 33, 919–928.
Sánchez-Lavega, A., Wesley, A., Orton, G., Hueso, R., Perez-Hoyos, S., Fletcher,
L. N., Yanamandra-Fisher, P., Legarreta, J., de Pater, I., Hammel, H., Simon-
Miller, A., Gomez-Forrellad, J. M., Ortiz, J. L., Garca-Melendo, E., Puetter, R. C.,
Chodas, P., 2010. The impact of a large object with Jupiter in July 2009. Astrophys.
J. 715, 155–159.
Sault, R. J., Engel, C., de Pater, I., 2004. Longitude-resolved imaging of Jupiter at
λ=2 cm. Icarus 168, 336–343.
Smith, H. J., Douglas, J. N., 1957. First results of a planetary radio astronomy
program of the Yale Observatory. Astron. J. 62, 247.
Span, R., 2000. Multiparameter Equations of State. Springer, New York.
Spilker, T. R., 1990. Laboratory measurements of the microwave absorptivity and

refractivity spectra of gas mixtures applicable to giant planet atmospheres. PhD
dissertation, Stanford University, CA.
Sromovsky, L. A., Collard, A., Fry, P., Orton, G., Lemmon, M., Tomasko, M., Freed-
man, R., 1998. Galileo Probe measurements of thermal and solar radiation fluxes
in the jovian atmosphere. J. Geophys. Res. 103, 22929–22978.
Sromovsky, L. A., Fry, P. M., 2010. The source of 3-µm absorption in Jupiter’s clouds:
Reanalysis of ISO observations using new NH3 absorption models. Icarus 210, 211–
229.
Student, 1908. The probable error of a mean. Biometrika 6, 1–25.
197
Suleiman, S. H., Kolodner, M. A., Steffes, P. G., 1996. Laboratory measurement of
the temperature dependence of gaseous sulfur dioxide (SO2 ) microwave absorption
with application to the Venus atmosphere. J. Geophys. Res. 101, 4623–4635.
Tillner-Roth, R., Harms-Watzenberg, F., Baehr, H. D., 1993. Eine neue fundamen-
talgleichung für ammoniak. DKV-Tagungsbericht 20, 167–181.
Townes, C. H., Schawlow, A. L., 1955. Microwave Spectroscopy. McGraw-Hill, New

York.
Tyler, G. L., Howard, H. T., 1969. Refractivity of carbon dioxide under simulated
Martian conditions. Radio Sci. 4, 899–904.
Ulich, B. L., 1974. Absolute brightness temperature measurements at 2.1-mm wave-

length. Icarus 21, 254–261.
Ulich, B. L., 1981. Millimeter-wavelength continuum calibration sources. Astron. J.

86, 1619–1626.
Valdes, F., Welch, W. J., Haber, D., 1982. The Jovian ammonia abundance from
interferometric observations of limb darkening at 3.4 mm. Icarus 49, 17–26.
Valkenburg, E. P., Derr, V. E., 1966. A high-Q Fabry-Perot interferometer for water
vapor absorption measurements in the 100 Gc/s to 300 Gc/s frequency range.
Proceedings of the IEEE 54, 493–498.
van der Tak, F., de Pater, I., Silva, A., Millan, R., 1999. Time variability in the radio
brightness distribution of Saturn. Icarus 142, 125–147.
van der Waals, 1873. On the continuity of the gaseous and liquid states. PhD disser-
tation, Universiteit Leiden, Leiden, Netherlands.
Van Vleck, J. H., Weisskopf, V. F., 1945. On the shape of collision-broadened lines.
Rev. Mod. Phys. 17, 227–236.
von Zahn, U., Hunten, D. M., Lehmacher, G., 1998. Helium in Jupiter’s atmosphere:
Results from the Galileo probe helium interferometer experiment. J. Geophys. Res.
103, 22815–22829.
Wagner, W., Pruß, A., 2002. The IAPWS formulation 1995 for the thermodynamic
properties of ordinary water substance for general and scientific use. J. Phys. Chem.
Ref. Data 31, 387–535.
Waters, J. W., 1976. Absorption and emission by atmospheric gases. In: Meeks, M. L.
(Ed.), Methods of Experimental Physics. Vol. 12B. New York Academic Press, New
York, NY.
Weidner, R. T., 1947. Microwave spectrum of iodine monochloride at 4.5 centimeters

wavelengths. Phys. Rev. 72, 1268.
198
Weiland, J. L., Odegard, N., Hill, R. S., Wollack, E., Hinshaw, G., Greason, M. R.,
Jarosik, N., Page, L., Bennett, C. L., Dunkley, J., Gold, B., Halpern, M., Kogut,
A., Komatsu, E., Larson, D., Limon, M., Meyer, S. S., Nolta, M. R., Smith, K. M.,
Spergel, D. N., Tucker, G. S., , Wright, E. L., 2011. Seven-year Wilkinson microwave
anisotropy probe (WMAP) observations: Planets and celestial calibration sources.
Astrophys. J. Suppl. Ser. 192, doi: 10.1088/0067–0049/192/2/19.
Young, D. M., Crowell, A. D., 1962. Physical Adsorption of Gases. Butterworths,

London.
Yu, S., Drouin, B. J., Pearson, J. C., 2010a. Species tag: 17002 Version 5.
URL http://spec.jpl.nasa.gov/ftp/pub/catalog/doc/d017002.pdf
Yu, S., Drouin, B. J., Pearson, J. C., 2010b. Species tag: 17004 Version 5.
URL http://spec.jpl.nasa.gov/ftp/pub/catalog/doc/d017004.pdf
Yu, S., Pearson, J. C., Drouin, B. J., Sung, K., Pirali, O., Vervloet, M., Martin-
Drumel, M.-A., Endres, C. P., Shiraishi, T., Kobayashi, K., Matsushima, F.,
2010c. Submillimeter-wave and far-infrared spectroscopy of high-J transitions of the
ground and ν2 = 1 states of ammonia. J. Chem. Phys. 133, doi:10.1063/1.3499911,
174317.
Zimmerer, R. W., 1963. Spherical mirror Fabry-Perot resonators. IRE Trans. Mi-
crowave Theory Tech. 11, 371–379.
199
VITA
Kiruthika Devaraj was born in Coimbatore, Tamil Nadu, India. After completing her
Bachelor’s degree in Electronics and Communication Engineering from PSG College
of Technology, Coimbatore, she enrolled at the Georgia Institute of Technology and
earned a Master’s degree in Electrical and Computer Engineering in May 2009, fol-
lowed by a doctorate in December, 2011. As a PhD candidate, Kiruthika Devaraj
conducted laboratory measurements of centimeter- and millimeter-wavelength opacity
of gases under simulated planetary atmospheric conditions.
200

Devaraj Kiruthika 201112 PHD

Uploaded by

Copyright:

Available Formats

Devaraj Kiruthika 201112 PHD

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Devaraj Kiruthika 201112 PHD

Uploaded by

Copyright:

Available Formats

THE CENTIMETER- AND

School of Electrical and Computer Engineering

Professor Paul G. Steffes, Advisor Professor Oliver Brand

Professor Waymond R. Scott Professor Carol Paty

Professor Gregory D. Durgin Date Approved: September 19, 2011

dous encouragement and constant faith in my abilities as a researcher. I also thank

my committee members, Professor Oliver Brand, Professor Gregory Durgin, Profes-

reviewing this work.

NNG06GF34G and NNX11AD66G.

I spent two summers at the National Radio Astronomy Observatory (NRAO) in

and Sai Kumar Thumuluru.

The constant support and encouragement from my family sustained me through-

out these years. My parents Shanthamani Soundirarajan and Devaraj Ranganathan

my grandparents and I cannot express my gratitude for their unconditional love. I

thank my wonderful husband Lakshmi Narasimhan Chakrapani, without whom this

LIST OF SYMBOLS OR ABBREVIATIONS . . . . . . . . . . . . . . xxi

II MICROWAVE SPECTROSCOPY AND AMMONIA SPECTRA 11

III LABORATORY MEASUREMENTS OF AMMONIA . . . . . . . 22

V MEASUREMENTS AND MODEL . . . . . . . . . . . . . . . . . . 86

VI MILLIMETER-WAVELENGTH OBSERVATIONS OF JUPITER147

VII CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160

APPENDIX A — MILLIMETER-WAVELENGTH COMPONENTS168

APPENDIX B — VLA OBSERVATIONS OF THE 2009 JUPITER

APPENDIX C — VLA OBSERVATIONS OF VENUS AT X BAND184

3.1 High-pressure planetary atmospheric simulator instruments. . . . . . 46

2.1 Ammonia molecule. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

α Multiple uses: Absorption, reduced Helmholtz free energy.

AIPS Astronomical Image Processing Software.

ALMA Atacama Large Millimeter Array.

AMC Active multiplier chain.

AU Astronomical unit. 1 AU = 149597870.691 km.

arcsec 1 arcsecond is 1/3600 of a degree.

a Helmholtz free energy.

A Absorption at the line center.

B Confidence coefficient of a two-tailed distribution.

C band Electromagnetic frequencies from 4 to 8 GHz.

∆ν Linewidth (half width at half-maximum).

D Multiple uses: Distance of separation between the mirrors in

dBm Decibel with respect to 1 mW.

EPL Effective path length.

ESA European Space Agency.

ε Permittivity of the medium.

ε0 Real part of permittivity of the medium.

ε00 Imaginary part of permittivity of the medium.

FPR Fabry–Perot resonator.

FSR Free spectral range.

FWHM Full width at half maximum.

F band Electromagnetic frequencies from 90 to 140 GHz.

GEISA Gestion et Etude des Informations Spectroscopiques Atmosphériques:

GPIB General Purpose Interface Bus.

G band Electromagnetic frequencies from 140 to 220 GHz.

h0 Ideal gas enthalpy.

ht Height of the cylindrical cavity resonator.

HWHM Half width at half maximum.

IRAM Institut de Radioastronomie Millimétrique.

JPL Jet Propulsion Lab.

Jy Jansky. Unit of spectral flux density. 1 Jy = 10−26 W/(m2 .Hz).

J Total angular momentum vector.