Excitation and ionization of gases by fission fragments.

MISSING IMAGE

Material Information

Title:
Excitation and ionization of gases by fission fragments.
Physical Description:
xv, 308 leaves. : ill. ; 28 cm.
Language:
English
Creator:
Walters, Roy Alan, 1941-
Publication Date:

Subjects

Subjects / Keywords:
Nuclear excitation   ( lcsh )
Ionization of gases   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis--University of Florida.
Bibliography:
Bibliography: leaves 302-307.
Statement of Responsibility:
By Roy Alan Walters.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000580518
notis - ADA8623
oclc - 14039068
sobekcm - AA00004931_00001
System ID:
AA00004931:00001

Full Text















EXCITATION AND IONIZATION OF GASES

BY FISSION FRAGMENTS














By

ROY ALAN WALTERS


A DI-:.ERi'AT:ON PRESENTED TO THP GRADUATE COUNCIL OF
THE 'UN-VERSI't'l OF FIORIDA IN PARTIAL
FULPILM[INT 01 TIhL : ;Cr- 1 F'.-L NTS FOR THE DEGREE OF
DOCTOR OF PHBLOSOP, I




INI'VL'PSITY OF FLORIDA
197 9





























To my father,

Harry Walters,

who was, in my eyes, the greatest of
engineers and the greatest of men.
















ACKNOWLEDGMENTS


The author would like to express his deepest appreciation and

gratitude to Dr. Richard T. Schneider, the chairman of his supervisory

committee, for his guidance and support in this research, and for the

faith and friendship he showed toward the author throughout this

academic endeavor. Sincere thanks are extended to the other members

of the supervisory committee, Drs. Hugh D. Campbell, Kwan Chen, George

R. Dalton, and William H. Ellis.

The author also wishes to thank Dr. Edward Carroll for his

ideas and encouragement on the LMFBR detector studies.

Special note should be made of the valuable contributions of

Mr. Ernest Whitman, who aided the author with the design and construc-

tion of the vacuum chamber, and of Mr. Richard Paternoster, who

developed the computer analysis and plotting programs. Thanks are

also extended to Mr. George Wheeler for his valuable assistance in the

construction of equipment and its operation.

The author acknowledges the technical assistance provided by

Mr. Henry Gogun and the other members of the UFTR crew.

Special thanks are extended to the author's wife and son for

their help in the production of this paper and for their patience and

support through the years this endeavor has taken to complete.

The author will be forever indebted to his father, the late

Harry Walters, for his unfailing faith and encouragement, and most











especially, for his aid in the actual production of this manuscript.

Many of the diagrams that are a part of this dissertation are products

of his highly skilled hands.

Appreciation is expressed to Mrs. Edna Larrick for typing the

final draft of this manuscript.

This research was supported by NSA Grant NGL-10-005-089.

















TABLE OF CONTENTS


Page


ACKNOWLEDGMENTS . . .

LIST OF TABLES . . .

LIST OF FIGURES . . .

ABSTRACT . . . .

CHAPTER
1 INTRODUCTION . . .

1.1 The Nuclear-Pumped Laser . .
1.2 Previous Studies of Fission
Fragment-Produced Plasmas . .
1.3 Glow Discharge Irradiation and Fission Fragment
Interaction Experiments . .

2 THEORETICAL CONSIDERATIONS . .


2.1 Energy Deposition by Fission Fragments
and Other Particles . .
2.2 Energy Deposition by the Reaction He3(n,p)T
in a Glow Discharge . .
2.3 Description of a Fission Fragment-Produced Plasma

3 PLASMA RESEARCH APPARATUS . .

3.1 Introduction . . .
3.2 Primary System . . .


3.2.1 Plasma Region . .
3.2.2 Vacuum Chamber; Optical, Gas
and Electrical Feeds ..


3.3 Uranium Coatings . .


3.3.1 Coating Requirements .
3.3.2 Coating Thickness .
3.3.3 Review of Methods and Chemistry .
3.3.4 Chemical Procedures .
3.3.5 Mechanical Coating Procedure .


iii


viii


. ix

. xiii



. 1

. 1

S. 9

. 11

. 14


14

S 37
40

S 49

49
S 50


. 51

. 55










TABLE OF CONTENTS (Continued)


CHAPTER Page
3 (Continued)

3.4 Support Systems . . 65

3.4.1 Reactor: Neutron Flux and Gamma Dose .. 65
3.4.2 Electrical Systems . ... 67
3.4.3 Gas Filling Systems . ... 68
3.4.4 Shielding . . 70
3.4.5 Safety . .... 71

3.5 Data Acquisition Systems . .. 75

4 ANALYSIS OF "IN-REACTOR" PLASMAS: FISSION FRAGMENT
GENERATED AND GLOW DISCHARGE . .. 77

4.1 Experimental Procedure and Data Analysis:
Fission Fragment Interactions . .. 78
4.2 Helium . . .. 86

4.2.1 Helium Kinetics and Spectral Analysis 88
4.2.2 Line Intensity and Excited State Density 92
4.2.3 Boltzmann Plot Analysis . ... .102
4.2.4 Field Amplification of Line Intensities 110

4.3 Argon Excitation by Fission Fragments ... 121

4.3.1 Argon Kinetics and Spectral Analysis .... .122
4.3.2 Line Intensity and Excited State Density 131
4.3.3 Boltzmann Plot Analysis .. .. 135
4.3.4 Field Amplification . ... 140

4.4 CF Fission Fragment Interactions .. .153
4.5 Glow Discharge Irradiations . .. 160

4.5.1 Experimental Procedures . .. 161
4.5.2 The Glow Discharge . .. 161
4.5.3 General Reactor Mixed Radiation Effects
on the Glow Discharge ... .. 164
4.5.4 Volume Deposition . ... 171
4.5.5 Cathode Deposition . .... .174

5 APPLICATIONS OF FISSION FRAGMENT-PRODUCED PLASMAS 178

5.1 The Nuclear-Pumped Laser. . 179
5.2 A Neutron Detector for the Liquid Metal
Fast Breeder Reactor .. . 187
5.2.1 Diagnostic and Power Supply Systems 190
5.2.2 Neutron Detector Experimental Results ..... 192
5.2.3 Data Projections and Realistic Chamber Design 198












TABLE OF CONTENTS (Continued)


CONCLUSIONS . . .


CHAPTER

6

APPENDIX

I


II

III


IV

V


BIBLIOGRAPHY . . .

BIOGRAPHICAL SKETCH . . .


Page

201


. 205

. 223


. 231

. 277

S. 295


302

308


PRESSURE DEPENDENCE OF RELATIVE LINE INTENSITIES
AND EXCITED STATE DENSITIES HELIUM .

BOLTZMANN TEMPERATURE PLOTS HELIUM .

PRESSURE DEPENDENCE OF RELATIVE LINE INTENSITIES
AND EXCITED STATE DENSITIES ARGON .

BOLTZMANN TEMPERATURE PLOTS ARGON .

FIELD AMPLIFICATION OF ARGON SPECTRAL EMISSION .
















LIST OF TABLES


Table Page

4-1 HELIUM DATA .............. ... 116


4-2 IDENTIFIED CASCADES OF ARGON II . ... 134


4-3 ARGON DATA ............... .* .......... 145


4-4 CF FISSION FRAGMENT IRRADIATIONS:
BAND PEAKS OBSERVED AT 760 TORR . .. 160


viii

















LIST OF FIGURES


Geometry . . .

Range of Fission Fragments in Helium and Argon .

Deposition of Energy into Helium by Fission Fragments,
11 n
= 3.8 10 . .
cm -sec


2-4 Deposition of Energy into Argon by Fission Fragments,

= 3.8 X 10 2
2 . o
cm -sec

2-5 Energy Deposition by Fission Fragments in Argon


and Helium A Fixed Cavity


Reactor Mounting . .

Chamber Experimental Sections .

Chamber Detail Reactor Region .

University of Florida Training Reactor .

K as a Function of Density .

The Kinetic Energy of Fission Fragments as
a Function of Mass Number [38] ... .

Chemical Procedures Coating Solution .

Mechanical Procedures for Coating .

Distribution of Thermal Neutron Flux
along Experimental Chamber .

General Gas Discharge I-V Characteristic

Chamber Mounting and Shielding Cave .


Helium Spectrum . . .

Helium Spectrum . . .


Figure

2-1

2-2

2-3


Page

21

29



33


. . 36


. 52

. 53

. 54

. 56

. 58


3-1

3-2

3-3

3-4

3-5

3-6



3-7

3-8

3-9



3-10

3-11

4-1

'4-2


. 66

. 69

. 72


I I I 0 .











LIST OF FIGURES (Continued)


Figure

4-3 Helium Spectrum . . .

4-4 Helium Spectrum . . .

4-5 Helium Spectrum . . .

4-6 Helium Excited States and Transitions .

4-7 Optically Viewed Energy Deposition in Helium
and Argon as a Function of Pressure .

4-8 Helium Intensity and Energy Deposition as
a Function of Pressure . .

4-9 Relative Excited State Population Density as
a Function of Pressure . .

4-10 Line Intensity versus Pressure Helium .

4-11 Boltzmann Plot Helium Glow Discharge .

4-12 Boltzmann Plot Fission Fragment Excited Helium .

4-13 Pressure Dependence of the Boltzmann Plot
Correlation Coefficients Helium . .

4-14 Boltzmann Temperature as a Function of Pressure .

4-15 Current-Voltage Characteristics of
Fission Fragment Excited Helium . .

4-16 Line Amplification and Current as a Function of
Applied Voltage 30.5 cm Long Helium Cavity .

4-17 Electron Energy Distribution Fission Fragment
Excited Helium . . .

4-18 Argon Spectrum . . .

4-19 Argon Spectrum . . .

4-20 Argon Spectrum . . .

4-21 Argon Spectrum . . .

4-22 Argon Spectrum . . .

4-23 Argon Spectrum . . .


Page

. 81

. 82

. 83

. 87


. 95



. 97



. 98

. 99

. 104

. 105


. 107

. 108


. 112


. 114


. 120

S. 124

S. 125

S. 126

S. 127

S 128

129











LIST OF FIGURES (Continued)


Figure Page

4-24 Intensity and Calculated Deposition (Viewed)
as a Function of Pressure . ... 133

4-25 Boltzmann Temperature Plot 600 torr Argon I .. 137

4-26 Boltzmann Plot Argon II, 600 torr . ... 138

4-27 Argon Temperature and Correlation Coefficient
versus Pressure . .... .. 139

4-28 Amplification and Current versus Applied Voltage -
150 torr Argon . . .. .... 141
E
4-29 Amplification Coefficient versus -, Argon ... 143

4-30 Electron Density and Temperature versus Pressure Argon 148

4-31 Ion Pair Generation Rate as a Function
of Pressure Argon . . .. 149

4-32 Electron Energy Distribution Argon . .. 151

4-33 Electron Energy Distribution Argon . .. 152
o 0
4-34 Glow Discharge through 5.5 torr CF4 2000 A 5087 A 155

4-35 Band Peak Intensity of 100 torr CF4 as a Function of Time 156

4-36 Spectrum of Fission Fragment Excited CF4,
o o 4
760 torr 2000 A 5087 A . .. 158

4-37 Spectrum of Fission Fragment Excited CF4,
o 0 4
760 torr 2646 A 3280 A . ... 159
3
4-38 Glow Discharge I-V Characteristics for He -C02-N2,
8:1:1, 10 torr . . .. 165

4-39 I-V Characteristics for Hollow Cathode Glow Discharges 167

4-40 I-V Characteristics, 3.3 torr Glow Discharge,
Flat Cathode . . ... 168

4-41 Voltage Decay, Glow Discharge, for Reactor Shutdown -
Constant I, 20 ma . .... 169

4-42 Voltage Decay, Glow Discharge, for Reactor Shutdown -
Hollow Cathode . . ... .. 170












LIST OF FIGURES (Continued)


Figure

5-1

5-2

5-3

5-4

5-5



5-6

5-7

5-8


Page

. 181

. 184

. 186

. 191



. 194

. 195

. 196

. 197


Nuclear Pumped Laser . .

Nuclear-Pumped Laser Output . .

Cavity Design . .

LMFBR Neutron Detector Signal Flow . .

Amplification of Argon Line Intensity versus
Reactor Power . . .

Ar I 6965 A Filtered Output Neutron Detector .

Total Spectrum Signal Neutron Detector .

P-P Voltage versus Reactor Power Neutron Detector












Abstract of Dissertation Presented to the
Graduate Council of the University of Florida in Partial
Fulfillment of the Requirements for the Degree of Doctor of Philosophy


EXCITATION AND IONIZATION OF GASES
BY FISSION FRAGMENTS

By

Roy Alan Walters

June, 1973



Chairman: Dr. Richard T. Schneider
Major Department: Nuclear Engineering Sciences


The excitation and ionization produced by fission fragments was

investigated to identify basic mechanisms that could be applied to

direct nuclear pumping or enhancement of gas lasers.

A cylindrical U235 foil and its axial electrodes were placed

in a vacuum chamber which was capable of transmission of fluorescence to
o 0
its exterior from 2000 A to 8000 A. The chamber was filled with argon,

helium or carbon tetrafluoride at various pressures and emersed in

a thermal neutron flux of 3.8 1011 n/cm -sec.

The spectrum obtained from this excitation was qualitatively

similar to a glow discharge for argon and helium except for the pres-

ence of excited ion species of He II and Ar III. The spectral output

from irradiation of 760 torr carbon tetrafluoride provided a band

system that is presently unidentified. Pressure dependent relative

intensity and excited state density data provide information on state

and species kinetics such as formation of molecular ions and their loss

mechanisms. Four possible population inversions were identified in


xiii











Ar II. Boltzmann analysis of the excited states supplied a temperature

for each species where the correlation coefficients of the fit lines

indicate that the plasmas are typical nonequilibrium cascade systems.
10 3
Electron densities around 10 e /cm and Maxwellian temperature

values for the collected electrons have been obtained from the recombina-

tion region I-V characteristics. Electron energy distributions formu-

lated from the data compare favorably with referenced calculations.

Ion pair generation rates were well within expected deviation compared

to calculations using a two-region energy deposition model.

In the ion chamber region of the I-V characteristics, line

emission increased as an exponential function of field strength.

A model for this amplification was developed for argon, utilizing an

amplification coefficient applicable to all pressures.

A neutron detector was developed for the liquid metal fast

breeder reactor by using optical transmission from the reactor core of

electric field modulated emission from fission fragment excitation.

Measurement of the modulated field effect eliminates the majority of

noise sources and gamma degradation signal loss associated with other

detectors.

Because of the excellent spectral output from Ar II, a nuclear

pumped argon ion optical cavity was constructed. Data from reactor

irradiation of the cavity indicate that it was lasing.

The effect of mixed radiation from the reactor on a glow dis-

charge was studied. For thermal neutrbn fluxes less than 3.8 X 1011

n/cm2-sec and gamma dose rates of 1.1 x 10 R/hr, irradiations of He3
4
He N and mixtures thereof show that there is no volume deposition












effect on the glow discharge. A cathode photoemission effect was found

that altered the balance of the discharge. Positive ion bombardment

from the He3(n,p)T reaction products produced a considerable electron

source that perturbed the cathode fall region. Enhancement of CO2 lasers

was shown to be a mechanism of preionization for low voltage glow dis-

charge initiation and subsequent maintenance with lower power input.

This lowers the temperature of the discharge and improves laser pumping

efficiency.















CHAPTER 1


INTRODUCTION




During the last decade, fission fragments and other energetic

heavy ions that are produced by neutron reactions have been examined

as possible sources of energy for generation of high temperature plas-

mas. These high energy-particles can be produced in great numbers in

a nuclear reactor where the reaction source (neutron flux) can range

up to 1017 n/cm2-sec in short pulses. The main thrust of research in

this area has been toward the demonstration of a nuclear-pumped laser.


1.1 The Nuclear-Pumped Laser

The term "Nuclear-Pumped Laser" refers to a laser that is

excited by products of nuclear reactions only and not by any electri-

cal or optical source.

The link between the reactor and laser is natural when one

considers that communications could be greatly aided with transmission

of data to earth by laser beam in advanced extraterrestial equipment

which will have to include a nuclear reactor for power generation.

Direct coupling of a laser to a nuclear reactor is necessary

in order to produce desired high input powers. One might envision

a reactor-laser system where an optical cavity and the reactor fuel
235
are combined. A set of U235 internally coated tubes could be grouped

together in a cylindrical shape where the optical cavity would be











formed by mirrors reflecting through the tubes. A neutron moderator

would surround the tubes and cooling could be accomplished by passing

the laser gas through the tubes and then through a heat exchanger.
235
The fissioning of the U235 would supply both neutrons for sustaining

the nuclear reaction and fission fragments for excitation of the gas.

Each fission event supplies two fission fragments, a heavy fragment

with an average energy of 67.5 MeV and a light fragment with an aver-

age energy of 98.7 MeV. These light and heavy fragments vary in weight

according to the familiar fission product distribution, but both frag-

ments have the very important characteristic of being emitted with

a charge of about 20 e. This results in a very large coulombic inter-

action rate and, thus, a very large deposition of energy in a small

path length, producing the excited states utilized in the laser.

Operating this reactor in a pulse mode will maximize the peak laser

power output.

Another nuclear-pumped laser scheme involves a gas core reactor

using UF6 as a fuel and.some fluoride molecular species derived from

UF6 dissociation reactions as the lasing species. A similar homogen-
235
eous concept involves insertion of U235 compounds into a liquid dye

laser in order to produce a critical mass and thus a reactor. This

system would derive its excitation from fission fragments rather than

from chemical or optical sources. One disadvantage of the liquid dye

laser concept is the apparent breakdown of long chain dye molecules by

radiation interaction. R. Schneider [ 1] has inserted Rotemin b dis-

solved in ethyl alcohol into a reactor and observed fluorescence.












When the dye was removed from the reactor, it was evident that it had

been completely destroyed, since there was no color, or fluorescence,

left in the liquid. Much experimental work has been done trying to

show the feasibility of a nuclear-pumped laser. To date, no proof has

been presented that confirms the operation of such a laser.

There are several alternatives when considering an energy

source for direct excitation of a laser. The most widely available

source is gamma radiation, but in order to effect efficient energy

transfer, high density solids must be used. A desire to optimize the

absorption led early workers [2,3] to concentrate on direct excitation

of solid state lasers. However, gamma radiation was found to suppress,

rather than aid, lasing action in all solids studied to date. This

"cutoff" phenomenon is generally attributed to radiation damage in the

solid.

An obvious solution to the problem of radiation damage is to

use a gas or liquid lasing medium. For gases, this implies the use of

a high energy heavy particle that has a large dE/dx. Such a particle

is exemplified by the fission fragment. Liquid lasers are still in

their infancy and only one published report is available on an attempt

at nuclear pumping [4]. Nuclear pumping of gas lasers has been studied

extensively starting with a comprehensive study by L. Herwig [5] in

1964. He recognized the radiation damage problem associated with

solids and decided to concentrate on gas laser nuclear pumping.

His calculations showed that He-Ne laser threshold requirements were

theoretically within the reach of some reactor and accelerator radiation













sources. Herwig also noted that a large diameter laser may be possible

due to the inherently low electron temperature expected in the radia-

tion produced plasma.

J. DeShong [6] carried out more detailed calculations and

showed that a high pressure, large diameter nuclear-pumped laser has

theoretical efficiency two orders of magnitude greater than small diam-

eter devices. In 1967 he undertook a series of experiments to verify

the feasibility of direct pumping. None of his devices showed proof of

lasing. Eerkins [7,8] decided to study noble gases as a possible lasing

medium because their high density allows a large dE/dx by the fission

fragment and because their ionization energy is low (E. = 15.68 eV for
1

argon). Because recombination and dissociative recombination is ex-

tremely efficient in a cold plasma, it was felt that the argon ion tran-

sitions would exhibit a population inversion greater than that seen in

conventional electrically pumped lasers. Eerkins's source of energy
17 2
was a TRIGA pulsed reactor with fluxes up to 101 n/cm -sec. The
235
reactor neutrons interacted with a U coating on the laser tube wall

and produced a fission fragment-generated plasma. Boron-coated walls
10 7
were also tried, using the reaction B (n,a)Li producing a plasma

utilizing the reaction product energy of 2.3 MeV. Although Eerkins

did not find any proof of lasing, he did generate a light output that

was quite intense. Guyot [9], concentrating on Ar and CO2 lasers, also

did not find lasing in his experiments.

Nuclear-pumped laser experiments previously mentioned concen-

trated on pumping with a highly energetic particle that was a result











10
of a neutron reaction with B or a fission event. These experiments

also concentrated the reacting material on tube walls while inserting

gas as a target for the particles. V. Andriakhin [10] attempted lasing

in an entirely different manner. He utilized the large neutron cross

section (5000b) of the He3 reaction [He (n,p)T + 760 keV] to provide
3
a source for excitation in a cavity filled with a gas mixture of He

and mercury vapor. His device generated light output of 10 mW, but no

proof of lasing was seen.

Several researchers (T. Ganley [11], F. Allario [32], H. Rhoads

[13]) noted the inability of others to generate a nuclear-pumped laser

and decided to look at the effects of several nuclear sources on an
3
operating laser. All of these studies involved the use of He as a

replacement for He4 in a He-CO 2-N2 laser gas mixture while operating

the laser in a nuclear reactor. It was felt that the high efficiency

of the CO2 laser would allow maximum conversion of the IHe(n,p)T reac-

tion energy, 760 keV, into laser output.

All three authors noted an increase in the efficiency of the

laser being irradiated in the reactor and, therefore, an increase in

laser output. Rhoads recognized this increased laser efficiency as an

increase of the glow discharge efficiency. He described this effect

as primarily due to the bombardment of the cathode surface by gammas

and He (n,p)T reaction products and thus an alteration in the cathode

fall equilibrium. The output of a CO2 laser depends, among other

things, on the electron temperature, Te, of the plasma. Bullis [14]

has shown that the electron temperature for optimum pumping of a CO2











laser is far below that resulting from normal glow discharge pumping.

The addition of the radiation source term to the cathode fall equilib-

rium allows the operation of a glow discharge at much lower currents

and field strengths, allowing lower Te and much more efficient pumping.

This irradiation effect also allows operation of glow discharges at

much higher pressures than previously attainable.

It would also be possible to increase the power output of

present operating laser systems by the addition of a radiation source

at the cathode.

The glow discharge irradiation work described in this disserta-

tion was done simultaneously with the nuclear-enhanced laser work de-

scribed above; the results shown in this paper basically agree with

those described above.

Upon termination of the nuclear-pumped laser experiments, it

became evident that the theoretical considerations used to calculate

the neutron flux needed for lasing action may be inaccurate. The

thermal flux level available in many reactors is well above the calcu-

lated lasing thresholds; therefore, some of the experimental devices

should have lased. Most of the theoretical work has been based on the

premise that one excited state is available for each 100 eV deposited

in the gas [5,6]. This relation results from the assumption that once

the energy is deposited in a gas by nuclear interactions, subsequent

distribution of excited states, ionization, etc., is identical to that

of electrically pumped lasers.












G. R. Russell [15,16] has noted at least for the case of

atomic argon direct nuclear excitation of gases is an entirely dif-

ferent kinetic process than that found in most CW electrically excited

lasers. Generally, in electrically pumped gas lasers population inver-

sions are formed by metastable state collisions with ground state atoms

(He-Ne) or other energy transfer systems initiated by the electron

swarm.

The kinetics of the argon ion laser, where the population

inversion is produced by electron collisional excitation of the upper

state and electron recombination in the lower state, served as Russell's

example. On the average, it takes two collisions with electrons to

elevate the argon ground state to the ionized excited level required

for population inversion. Such a ladder-climbing system is not evident

in the low temperature plasma generated by nuclear sources. This low

temperature system is more analogous to a recombining plasma at low

thermal electron temperatures, where there is additional preferential

excitation due to high energy particles superimposed on the thermal

excitation. Therefore, it would not be expected that inversions formed

in electrically excited CW lasers would necessarily be found in nuclear-

excited lasers. It is also very probable that the additional excita-

tion due to high energy heavy particles will create new inversions not

previously observed in electrically excited lasers.

Russell supports this last premise by calculating population

inversions produced in argon by fission fragment interactions and not

available in conventionally pumpedlasers. His calculations included












formation of excited states due to recombination of thermal electrons

and the associated radiative and collisional decay of these states.

To complicate matters further, Miley [17] and others have indicated

that there is a difference between how fission fragments and alpha

particles lose their energy in a gas medium. This is because alpha

particles remain essentially charge invariant over their path length,

while fission fragments starting with an average charge of 20e lose

their charge by recombination while losing energy by coulombic, rather

than nuclear, elastic collisions. Recently several of the authors [ 1]

mentioned above have revealed that they now feel that once the fission

fragment deposits its energy into a gas, this excited gas may be

regarded as identical to the electrically excited gas.

From the above discussion it can be concluded that the exact

nature of a gas excited by fission fragments is not known. Also, there

is a distinct possibility of producing a nuclear-pumped laser using

some unusual (not normally available) population inversion; some gas

now known to exhibit population inversions may become the standard for

this type of laser. A less optimistic conclusion is that nuclear

pumping is not presently feasible.

No one has previously studied experimentally the excited

states of a plasma produced by an extremely large source of heavy,

highly charged particles such as fission fragments. Until detailed

analysis of heavy particle-produced plasmas is complete, the nuclear-

pumped laser experiments are without foundation.












1.2 Previous Studies of Fission
Fragment-Produced Plasmas

To study excited states of atomic and ionized species exper-

imentally, spectrographic analysis of the spontaneous photon emission

from these states is necessary.

Several researchers have studied experimentally alpha particle-

induced luminescence of gases with great success. S. Dondes et al. [18]

have been able to supply very good spectrographic plates (long time
210
exposures) of many gases exposed to a Po 5 MeV, alpha source.

Amplification of the gross light output was indicated upon placing a

350 V/cm field across the luminescing area. P. Thiess [19], using

a similar but more powerful source and photon counting techniques,

was able to obtain similar data. Other allied work has come from the

French gas counter research program [20,21,22,233 and two very early

studies [24,25]. These latter experimental works used sources of very

low intensity and the data were decidedly biased toward use in the

design of counting equipment. The source strengths used in all of

the alpha particle interaction work are weak enough, and the electron

density therefore small enough, that the gas should probably be described

as a scintillating light source rather than a plasma. As mentioned

previously, it is hard to draw comparisons between such low level

a source luminescence and a plasma generated by a large fission frag-

ment source, but these studies do give basic knowledge of the kinetics

of heavy particle interactions.

In contrast to excited state measurements, fairly extensive

measurements of electron densities in radiation-produced plasmas have












been reported. Jamerson et al. [26,27] worked with "in-reactor" fission

fragment-produced plasmas where, utilizing the I-V curves produced by

a field across the plasma, they were able to calculate values for

electron temperature and density. For a flux of 1012 n/cm2-sec a 600K

electron temperature was found which compares favorably with the micro-

wave cavity measurements of Bhattacharya [28]. Ellis et al. [29,30,31],
3
have studied recombination coefficients in plasmas produced by He reac-

tion products, alpha particles, and fission fragments, and have added

immensely to the knowledge of the kinetics of these plasmas.

The above works indicate that the general kinetics studies of

alpha and fission fragment-produced plasmas are no longer in their

infancy; therefore, one can draw on these data in explaining the source

terms for generating atomic excited states.

A very large library of cross sections for collision-produced

excited states is available in a book by E. W. Thomas [32]. These

cross sections have been generated by bombarding various gases by heavy

ions produced in accelerators. Unfortunately, none of the ions had

energies above 1.5 MeV and few had energies above 0.5 MeV. The appli-

cability of these cross section data to fission fragment interactions is

questionable, but certain techniques used to generate these data are

of great interest in this study.

Several researchers have attempted to study spectral emissions

from fission fragment plasmas. F. Morse et al. [33] studied the

luminescence of several gases under bombardment by fission fragments

in a nuclear reactor. They did see some line structure, but concluded












that it was too weak to study (96-hour exposures!). They then returned

to the use of alpha sources [18] where reactor associated problems were

not present. R. Axtmann [34] studied the luminescent intensity of
252
nitrogen bombarded by fission fragments from Cf2, but he just assumed

that the light was from the second positive system of molecular nitro-

gen emission and tried no spectroscopic analysis. The above studies

led Pagano [35] to attempt spectroscopic analysis of various gases
252
bombarded by Cf252 fission fragments, but his source was not strong

enough to allow recording of spectra.

During the nuclear-pumped laser studies of Eerkins [7,8],

several spectroscopic plates of their pulsed plasma were taken, but

at that time they were more interested in producing a laser than study-

ing the plasma that they assumed contained population inversions.

Thus, the spectroscopic work was limited to a few plates in which the

photon output was filtered by the mirrors of the laser system.

The above spectroscopic studies gave great encouragement to

this author in his studies of "in-core" fission fragment-produced

plasmas. After careful analysis of the techniques used by these

researchers, a set of experimental guidelines (see Chapter 3) were

generated in order to avoid the known problems associated with such

research.


1.3 Glow Discharge Irradiation and
Fission Fragment Interaction Experiments

The need for experimental data on fission fragment-produced

plasmas and on the radiation-produced changes in an already existing

plasma has been established. The dual purpose of this dissertation












is to fill some of this unknown area and to generate sound techniques

for the study of "in-reactor" particle produced plasmas. Fortunately,

as will be shown in the following chapters, one basic experimental

apparatus can be used for this dual purpose.

The first probe into this area involves the study of a known

plasma, the glow discharge, and what happens to its operation under

bombardment by gamma and reaction products of the He (n,p)T reaction.

Several cathode configurations and several gases are studied for their

response to the reactor sources. Conclusions are drawn as to the extent

of changes and how they occur.

The most difficult area, the investigation of fission fragment-

produced excitation, is presented utilizing the assumption that the

plasma is an interacting Maxwellian system. The analysis of the data

emphasizes the deviations from equilibrium of this system. This approach

was taken due to the availability of large numbers (45 Ar II spectral

lines) of easily measurable spectral lines emitted by atomic as well as
4
ion species of several gases. He Ar, and CF4 were studied for pos-

sible population inversions by using Boltzmann plot techniques and

analysis of deviations of excited state populations as a function of

pressure. Data are supplied on the effect of pressure and a DC field

on the intensity of line emission.

A unique neutron detector for the liquid metal fast breeder

reactor was suggested by Dr. Edward E. Carroll [36]. Drawing from the

above data, a detection system was developed and tested using electric

field amplified light variations as a basis for its operation. Also,

an argon ion laser was tested utilizing pumping by fission fragment







13




interactions only. Results from this device, although not proof of las-

ing, show promise for future investigators.

In summary, this chapter has outlined the great need for

experimental data on the nature of effects of mixed radiation sources

on a glow discharge and has shown why the study of fission fragment-

produced plasmas is necessary.
















CHAPTER 2


THEORETICAL CONSIDERATIONS




To assess adequately the effects observed in the experimental

procedure, certain theoretical considerations must be made. The two

basic considerations that will be reviewed in this chapter are, first,

the way a particle deposits its energy, and, second, how this energy

might be distributed in a plasma. Basic calculational techniques will

also be reviewed for use in analysis of the data described later in

this study.


2.1 Energy Deposition by Fission Fragments
and Other Particles

The passage of charged particles through matter has been

studied for at least half a century. It is not surprising that the

theoretical and experimental description of alpha-particle and proton

penetration phenomena is well advanced as these are the charged particles

with which most experiments were performed before the discovery of nu-

clear fission. Presently there is emphasis on the study of the inter-

action of ions with larger mass and charge than these elemental par-

ticles. It is common practice to label ions such as fission fragments

as "heavy ions" in order to distinguish them from light ions such as

protons and alpha particles. This distinction is strictly arbitrary

since most of the phenomena involved in energy deposition by these











particles are identical. Fission fragments distinguish themselves as

heavy ions because they are very massive and have, immediately after

formation, about twenty electrons stripped from their atoms. Thus, the

effective charge of these fragments is considerably higher than those

of the light ion group. This is an important distinction, since heavy

ions suffer coulombic interactions, as well as the nuclear elastic

scattering found in light ion interactions. A light ion is essentially

charge invariant over its path length, while the fission fragment is

charge variable over part of its path.

The life of a fission fragment or other heavy particle is

summarized by Northcliffe [37]. If an atom is given a velocity greatly

in excess of the orbital velocities of its electrons and allowed to

enter a material medium, these electrons will be stripped from the atom

and the bare nucleus will proceed through the medium, gradually losing

energy because of coulombic interactions with the electrons of the

medium. At this point, where the heavy particle velocity is high,

elastic or inelastic collisions with the nuclei of the medium will be

relatively rare and will add little to the energy loss process. At

first there is a small, but finite, probability that the ion will cap-

ture an electron in one of these collisions and a large probability

that the electron will be lost in the next collision; but as the ion

slows down and approaches velocities comparable with the orbital veloc-

ity of a captured electron, the capture probability increases and the

loss probability decreases. As the ion slows to velocities smaller

than the orbital velocity of the first captured electron, the capture

probability becomes very large and the loss probability approaches zero.












Meanwhile the probability of capturing a second electron grows and the

corresponding loss probability decreases, so that with increasing prob-

ability the second electron is retained. As the velocity decrease con-

tinues, a third electron is captured in the same gradual way, and then

a fourth, and so on. The major difference in the description of the

capture process for successive electrons is the change in velocity

scale necessary to match the progressive decrease of orbital velocity

of these electrons within the ion.

Eventually the ion reaches velocities smaller than the orbital

velocity of the least tightly bound electron and spends most of its

time as a neutral atom. By this time its kinetic energy is being

dissipated predominately by the energy transfer arising from elastic

collisions between the screened nuclear fields of the ion and atom,

and a diminishing amount of energy is being transferred to the atomic

electrons. The neutralized ion is said to be stopped when it either

reaches thermal velocities or combines chemically with the atoms of the

stopping material.

With respect to the medium into which the heavy particle or

fission fragment is penetrating, most of the ionization and excita-

tion is caused by secondary electrons (delta rays) produced during the

initial coulombic stripping and recombination interactions. This does

not hold true, though, for a fission fragment near the end of its track,

where it is essentially neutral.

To calculate the space dependent deposition of energy in a

medium, it is usual to start with a stopping power relation. Using

appropriate geometry, one first calculates the available energy per











unit volume, and then using ion or excited state generating terms, the

kinetics of the system. The Bohr stopping equation for fission frag-

ments is [38]

4 4
dE 2 e 22 e
2 = NZ L, + 2NZ L (1)
dz eff 2 e 12 2 v(
my M2v

where
Sa /3 -1/3 1 -1
L =L L (x-1/3 + -x
ee e 4 "x



Lv = ( (Z23z/3 2/3 m(M1+M2) 2 -)2 22
RM 2 1 2 M1M2



N = atom density of the density stopping material

M1,Z1 = mass and nuclear charge of the moving fragment


MZ2 = mass and nuclear charge of the stopping material

e,m = electronic charge and mass

v = velocity of the moving fragment

Zeff= effective charge of the fission fragment example:
eff
1/3
Z = Z1/3 v/v
eff o
v = velocity of a Bohr-orbit electron (2.2 x 10 cm/sec)
o
V
o
x= 2Z -
eff v

L = term for electronic stopping power



e
Le = the electronic stopping power for particles of comparable
velocities (about 6.33 10- v).
velocities (about 6.33 x 10 v).












The first term in the right-hand side of equation (1) describes

the electronic stopping power derived from coulombic interactions.

The second term describes energy transfer by nuclear elastic inter-

actions. It is standard practice to ignore the second term since the

amount of energy deposited by nuclear interactions is small compared

to the total energy deposited. Thus, using the first term only of

equation (1), the range of a fission fragment can be determined.
1 2
Assuming for the particle that E = -M v and that the Thomas-Fermi
2 1
1/3
effective charge, Zef = Z v/v is valid, the fragment velocity
eff o

follows from equation (1) as

dv
d K(N,Z1,M1) (2)


where

F 2/9
4 -8 6Z
K(N ) 2N e 6.33 x 10 1/3 1 (3)
K(N,Z1,M1) = 2N -- 8M + 1 (3)
2 8M 1 1 1/3
my 1 2
o -


K is therefore a function of the mass and charge of the moving

fragment and the density of the medium, but is velocity and space inde-

pendent. Solving this equation shows that

v(x) = v. Kx (4)


Solving this velocity-distance relationship for x when

v(x) = 0 or v(x) x=R = O, where R equals the range of a particle

with initial velocity v., the result is

V.
R(v.) = (5)
1 K











1 2
Assuming the initial energy E. = -M Vi and substituting into

equation (4) produces the well-known square law energy deposition rela-

tionship of a fission fragment,

x 2
E(x) = E(l1 -) (6)
x R


This relationship is therefore equivalent to the Bohr stopping

power equation using the Thomas-Fermi approximation for Zeff with the

nuclear stopping term neglected.

Several authors'using the general equation,

x n
E = E(l ) (7)


have disagreed with the n=2 value derived above. Axtmann [34], using

the luminescence of nitrogen under fission fragment bombardment, found

n equal to 1.7. Long [39] used n= 1 for his calculations where the

n value was obtained from collated range-energy data for a variety of

stopping materials. Steele [40] used n= 1.5 to compute energy deposi-

tion by fission fragments in water.

To generalize the square law equation (6) for a point source

in an infinite homogeneous medium, an energy transfer function can be

stated as

2
G(x,p) = E.(1 -) (8)
1R

-I
where cos -1 is the angle between the x-axis and the direction of

particle movement.

To calculate the deposition in a gas by a fission fragment,

one must remember that a fission fragment source such as a coating of











U308 has a finite thickness. The foil is a dense medium and thus

absorbs a large amount of the fission energy available. Calculations

should therefore include this second medium unless the fuel is in a

gaseous form, such as UF6. It is very tempting to assume that only

perhaps one-half of the energy available will get through the foil

into the gas. This assumption would be hard to prove because of the

difficulty of relating a measurement of this energy in one particular

experiment to a calculation where either the thickness or the density

of the uranium compound is different.

To calculate energy deposition at any point z of Figure 2.1,

one first assumes that scattering of the fission fragment by the medium

is negligible and that only straight line paths need be considered.

The origin of the geometry is at the left face of the source slab which

has a thickness R1, the range of a fission fragment in the source medium.

When x is larger than R1, the substitution z= x-R1 is made so that

z=0 at the interface. The source is assumed isotropic in emission

with azimuthal symmetry about x. The angle between the path and the

x-axis is 8. The slowing down of the particle in both media is

described by equation (4). While the fragment is in medium I moving along

the abscissa, the residual velocity at x v1(x ;x), of a particle borne

at x with initial velocity vi is


v1(x';x) = vi KIx' -xI x x' < R1 (9)


while for medium II


v2(z;x) = C(x) K2z ,


(10)















MEDIUM I
(fuel)


MEDIUM II
(gas)


X =0 \ I

X=R1
---- R ---- Z=0















r


/ MEDIUM II


ORIGIN


Figure 2-1


Geometry











where C(x) is the residual velocity at z=0 or the interface of a

particle born at x in medium I. Assuming continuity at the interface,


v (x';x) R v (z;x) ;


then

C(x) = v. K (R x) (11)



and thus, the residual velocity in medium II is


v2(z;x) = v. K1(R1 x) K2z (12)


Since the thickness of medium I is equal to the range of a

fragment in medium I, equation (11) is valid for any 0 : x R1.

For a fission fragment not moving along the abscissa, but in the
-1
direction 9 = cos -,equation (12) becomes

(R -x)
(R1X) z
v (z;p,x) =v K K (13)
2 v 1 p 2 p


provided that

(R -x)
(RIX) z
K + K -< v for all p > 0
1 p 2 p o


or, in other words, provided that the particle arrives in medium II.

Since it was previously shown in equation (5) that

vK v
K 2 '
2R R2
1 2


one can state that











.r Rl-X z ] 14
v2(z;LP,x) = v 1 (14)



The Energy Transfer Function of equation (8) can now be stated as the

two-medium function

1 2
F(z;p,x) = 2 M1v2(z;p,x)


r (R -x) 2
= E 1 (15)



In order to eliminate the dependence on R2, a conversion factor "a" is

derived from the Bragg-Kleeman rule which converts the range of a

charged particle in one medium to its range in another medium.


R P (A /A)1/2 R1 (16)
R2 = p 2

Therefore,

R2 = aR1

and

a (A2/A1/2 (17)
P2

where

p = density,

and

A = atomic weight.
,t ,t
a" may also be derived from the theoretical stopping equation

or from experimental measurements. Equation (14) thus becomes



G(z;p,x) = E 1 (R1 (18)
o 1 p.R1 PCaR












From the general geometry of Figure 2-1, a total energy current

is derived for a point r in medium II due to the source S(r ,Eo, ) in

medium I.


J (r)= drJ
Ero


-4
r
o
-4
r

O




E
o

S(r ,E ,X?)dr dE dc
0 0 0 0


dJ dE S(r ,E ,fQ) G(r;r ,E ,0) ,
S0 o 0 0
E
0


(19)


= a spacial point in medium I,

= a spacial point in medium II,

= a solid angle characterizing the direc-

tional distribution of the source,

= initial energy of the fission fragment.


= the distribution of the fission fragment

source at r in dr at n in dQ and
o o

at E in dE (usually a constant).
o o


G(r;ro,Eo,) =
o0 0


-4
the energy at r, carried by a fragment

originated at ro, moving in direction 0

with initial energy Eo, assuming no

scattering.


If it is assumed that C(x) is the fission density,

C(x) = f (x)$(x), and that f(E ) is the normalized fission fragment

spectrum, then with isotropic emission, the differential source within

a thin layer dx at x, emitted in the solid angle width dpdc with

initial energy E in dE is
,cp)dx dE

S(x,E ,pcp)dx dE dpqp = -Cf (E )dx dE ddp (20)
o o 4r o o


where











Using the energy transfer function of equation (18) and equation (20)

in (19), one obtains the total energy current at z in medium II,

JE(z), resulting from a distributed source of fission fragments in

medium I
2n co R1 1

S(z) d dE f(E )E dx d pC(x)
E 4TJ o /a aR1-ax+2

aR1


X ( -(R -x)
R1 a 1 (21)
R aR1 )


Nguyen [38] discusses the limits of integration and how one

would analytically integrate this function. With a constant fission

density of C

SCE pR 1 1 22 21 33 3 22 ] ( energy
JE(z) 2 --bz+ b z + bz (bz)
cm -sec

(22)

where
1 1
b = -
1 2

R1 and R2 being the ranges of fission fragments in mediurmsI and II,

respectively, of a fragment having the initial energy Ep. The required

boundary condition JE(z)l =0 (or at bz = 1) is satisfied.
2

Equation (22) represents the total residual energy at point z.

The instantaneous energy loss per volume as a function of z is obtained

by taking the derivative of JE(z) with respect to z.











JE(Z) CE 1 3 2 2 energy (23)
d z [b z 2b2z On (bz) b(23)
dz 2 3
cm -sec


An almost identical empirical energy deposition relationship can be

derived, as previously noted, based upon the relationship


E(x) = E (1 1 5 n 3 (24)


Both equations give similar results for small z, but vary considerably

for z approaching the range R2 in medium II.

This energy deposition function ignores any nuclear elastic

collisions; but if one calculates an ion production source, a so-called

"ionization defect" takes into account this nuclear deposition, which

is less effective in ionization than coulombic interactions.

At this point researchers split to several different techniques

for generating source terms for a kinetic system. Most studies have

constructed an ion source term and used the standard w values for

fission fragment interaction with various gases. These values include

the "ionization defect" and are experimental in origin.

Using the square law point deposition form of equation (23),

one can derive, simply by dividing by w, the volumetric ion production

rate.
dJ (z) CE. R. r22 a
I dE(z CEp i 2 z22bz ion pairs
Sw* dz 2w. Li x J 13
1 cm -sec
(25)

where the distinction is made between the light and heavy groups of

fission fragments. Therefore,

I1 1 (26)
IT(Z) = IL(z) + 2 I(z)(26)










eV
The assumption that w V is a constant value over the
ion pair

entire path of the fragment is false, but if one includes the "ioniza-

tion defect" and views the target as a whole, such as a plasma system,

this approximation should be close to the actual generation rate.

P. Thiess [17] approached the problem in an unusual manner. Using

the semiempirical energy deposition approach, a suggested alternative

shown above, he avoids the use of w values by calculating excited states

and inoization directly. This approach requires knowledge of a complete

set of cross sections for generating the source terms for excited levels.

Thiess used modified Bethe-Born cross section data based upon proton

impact. Russell [16] used another approximation, the Gryzinski electron

interaction cross section, for his excited state calculations. Both

authors clearly state that the use of these cross sections may be

entirely invalid, but must be used because there are no experimentally

measured cross section data available for such interactions. One factor

that may make the Gryzinski electron interaction approximation more

applicable than the others is the fact that about two-thirds of the

ionization and excitation is distributed to the gas by secondary delta

rays or fast electrons, rather than by the primary fission fragment

particle.

The range of a fission fragment in a gas is a function of the

density of that gas and its molecular weight. Range relations are

strictly empirical and are derived from measured data independent of

straggling or other statistical phenomena.












Range as a function of pressure can be calculated using the

following equation [411

KE2/3
R(cm) =


where K = 1.4 for most gases (Figure 3-5).

Figure 2-2 shows a plot of range vs. pressure for both the

light group (Eo = 98.7 MeV) and heavy group (Eo = 67.5 MeV) fission

fragments in argon and helium.

Experimental procedures such as those used in this dissertation

are based on cylindrical geometry. The average chord length

(s w 4 volume/area) best represents the distance that a particle--if born

on the surface or in the volume of the cylinder--would travel in a

straight line before it would collide with the surface. For a cylinder

30.5 cm in length and 3.7 cm in diameter, the average chord length

is 3.4 cm. These lengths are identical to those found in the experi-

mental apparatus used here. It is interesting to note in Figure 2-2

that at all pressures below 1 atmosphere (760 torr), the range of

fission fragments in helium is greater than the average chord length.

The situation for argon is different since it is ten times as dense for

equal pressures; therefore, the average chord length is equal to the

range of the light fragment at 360 torr and equal to the range of the

heavy fragment at 280 torr. A quick conclusion could be that for most

of the experimental data that one would observe, only a small fraction

of the fission fragment energy would be deposited in the gas. This is

not necessarily true, because the energy deposition from the foil is































ht FragmEnts

I Heavy Fra t HELIU
-- \ \- -- -__









Siht Fragmnts

Heavy Fra t meit ARGON
i ARGO


10


PRESSURE (torr)








Figure 2-2 Range of Fission Fragments
in Helium and Argon


in0











skewed towards the foil surface due to the finite thickness of the

source and the fact that the great majority of the fission fragments

do not leave the source surface with the typical 67.5 MeV or 98.7 MeV

average energies that they are born with.

A much better view of the energy deposition can be gained by

calculating the deposition profile at each pressure. A calculation

of the energy deposition utilizes the square law deposition function

and geometry used for equation (23). First, several assumptions must

be made in order to equate the slab geometry calculation to the cylin-

drical geometry that is presented in most experimental situations.

1. The slab and cylindrical two-region energy current functions

are essentially identical.

This is a good approximation since the range of a fission frag-

ment in the U308 source foil used in these calculations is only
-4
7.5 x 10-4 cm; therefore, the great majority of the energetic fission

fragments that have a considerable range are emitted perpendicularly

from the surface.

2. Little energy is emitted to the gas when a fission fragment

collides with a surface.

This assumption is not adequate for exact analysis but should

be valid for the accuracy required here.

3. Energy deposition by other sources is a very small fraction

of the fission fragment deposition.

This assumption has been proved experimentally to be valid

by Leffert [263, where he has shown that other sources, such as gamma












radiation, deposit less than 1 per cent of the total energy to a volume

in normal reactor situations.

In order to proceed further, the fission rate must be calcu-

lated as follows:

C = R = Nt O = E (27)


Nt = number density of target nuclei

a = fission cross section

= average neutron flux along the foil

S= macroscopic cross section.


For these conditions,

Thickness = 6.2 mg/cm2 (the range of a fission

fragment in U30 )

11 n
Average thermal flux = 3.8 x 10 2
cm -sec

Fission cross section = 505 barns 93% enriched uranium,

2 11 missions
the generating function per cm surface area is 8.06 x 10 fiss
cm -sec


From equation (23)


dJE(z) dJEi (Z)
dz X dz
i=1,2


Cf EpiRi b3z2 -2b2z n (bz)- b energy (28)
L= 2 cm -sec
i=1,2












where

C = fission rate

E = most probable energy at birth (each group)

1 1
b
R2i aRli

i = fission fragment group.


This generating formula is calculated by splitting the depend-

ence on light and heavy particles, then adding the results, giving the

energy deposition profile shown in Figure 2-3 for helium and Figure 2-4

for argon.

The gas pressure, or atom number density, is the most important

factor in the deposition of energy in a fixed cavity. For pressures

below 760 torr in helium, the energy deposition across a 3.4 cm average

chord length cavity is approximately uniform. But, in argon, only

below 75 torr is the energy deposition somewhat uniform across the

cavity. Since this calculation takes into account only coulombic inter-

actions and ignores the nuclear elastic and inelastic scattering of the

particles when they reach the neutral status, the energy deposition

curves fall off extremely fast. If the nuclear scattering terms were

included, the range would be extended slightly, but only a small addi-

tion would be made to the deposition of energy at the end point of the

fission fragment path. The effect on the total deposition would also

be small [37].

One of the unknowns, as previously described, involves how the

energy is utilized, what excited states or ions are produced, and what


















109











10


U






DJ
-7

LU
I


DISTANCE FROM U308 FOIL, Z (cm)


Figure 2-3


Deposition of Energy into Helium
by Fission Fragments, D = 3.8x1011 n-
cmn-sec


760 trr








1 00 torr


50 torr



























---- -- -- --- ---- ~~ : :


100


102













-70 or

S0 t r








150 tor"




75 tor























-- -


-1 0 1
10 100 101

DISTANCE FROM U308 FOIL, Z (cm)





Figure 2-4 Deposition of Energy into Argon
by Fission Fragments, D = 3.8x10 cm2sec


u

m
E
u


>-


08
LL 10

LU

I-
L-
C,
'a







107












photon emissions are coming from the "plasma." These items are

a function of cross section for the various species that are present

in the gas. Measurements of the photon emission of the gas are based

on total emission from the optical cavity; thus, this photon output

can be compared to the total energy deposited into the gas in this

cavity. A description of the total energy input can be obtained by

integrating the energy deposition function over all source areas and

over the average chord length. From equation (23),


dJ (z)
E = dA (Ez )dz (29)
JA f dz
f z

z
SA CE R
E = 1 b3z2 -2b2z m (bz) -b] dz (30)

o


where

z = average chord length, or

z = R2, if R2 < z

1
= R2 = range of fission fragments in gas and is

a function of pressure.

Integrating,

A CE R 3 3 z
f p z 22 n b b1 MeV
E= --bz (2bz--)-bz (31)
2 L3 2 sec
o


Again, as in equation (28), the calculation is split for each

group of fission fragments. Figure 2-5 shows the solution of equa-

tion (31), where E is calculated as a function of the gas pressure for


































1011


1010


PRESSURE (torr)


Figure 2-5


Energy Deposition by Fission
Fragments in Argon and Helium
A Fixed Cavity -











the representative cylinder with an average chord length of 3.4 cm.

The deposition in the cavity filled with helium is almost a linear

function of pressure. In argon, the effect of the range being less

than the cavity dimensions is evident by the leveling off of the curve

above 200 torr. Depending on the recombination and diffusion of elec-

trons at pressures above 200 torr, the fission fragment-produced

excitation may generate a torroidal luminescent output in the cylin-

drical cavity. This would alter the uniformity of the photon output

into the fixed solid angle view of the diagnostic equipment and may

provide erroneous data, especially if some of the surface region were

optically shielded from the detector system. It is estimated that such

a shading effect exists in the experimental equipment associated with

this work. Further review of this problem can be found in Chapter 4.


2.2 Energy Deposition by the Reaction He (n,p)T
in a Glow Discharge

High energy products of the reaction He (n,p)T are of interest

here because of the use of He3 in the glow discharge experiments to be

described later.

The proton and triton share the reaction energy of 760 keV with

the heavier triton taking 190 keV and the proton 570 keV. The initiat-

ing neutron energy is in the range of less than a few eV; therefore,

little momentum is transferred and the reaction particles travel

randomly in opposite directions. Thus, the interactions with the gas

are independent of one another.












Fortunately, the linear stopping power formulation for protons

and heavy-heavy protons (tritons) has been established as satisfactory

for calculating energy loss phenomena. Interesting calculations for

this reaction include total energy deposited in the cylindrical glow

discharge and total ionization produced by this deposition. These

calculations are done assuming that the gas is He3 at a pressure of
dE
15 torr for an 8:1:1 mixture of He-CO -N is 2.6 times that of
dx 2 2

helium [13]; therefore, the calculation of total energy deposited

should be multiplied by this amount for experiments involving CO2 gas

mixtures. Reference 8 gives the stopping power of helium as

dE
--P = 105 eV for 570 keV protons,
dx cm-torr


dET eV
E = 270 for 190 keV tritons.
dx cm-torr


For a cylindrical cavity 3.7 cm in diameter and 12.7 cm long

(identical to the dimensions of the experimental apparatus) the average

chord length is 2.87 cm.

The total energy deposition is calculated using the following

equations:

1. Reaction rate
-3
R = N to cm-3 (32)

3
where N = number density of He atoms

S= thermal neutron cross section = 5400 b

= thermal neutron flux 3.8 x 101 n/cm -sec.












2. Total energy deposition


E E (dE s PRV V (33)
dx/ +dx)J sec
p T

where s = average chord = 2.87 cm

P = pressure = 15 torr
3
V = volume = 136 cm .


3. Total energy available

keV
E = (760 keV) RV ke. (34)
sec


Upon application of the data to these equations, it is found
13 eV
that only 1.75 x 10 1 is being deposited into the cavity. This is
sec
14 eV
only one-tenth of the total available energy, 1.12 x 10 generated

in the cylindrical volume.
-6
This energy input is equal to only 2.8 x 10- watts, so the

total energy both available and deposited at this neutron flux level

is but a small fraction of the energy deposited by electrical excitation.

In fact, measurements that will be detailed in Chapter 4 for glow dis-

charges show that the minimum electrical power input needed to generate

a glow discharge in such a cavity is 0.5 watt. The one conclusion that

can be drawn from the calculations is that volume ionization or excita-
3
tion by He reaction products probably does not account for any signif-

icant changes in the operation of typical low power glow discharges,

especially for neutron fluxes below 1012 n/cm 2-sec.

Butler and Buckingham [42] state that for high energy ions

whose velocity is much greater than the thermal ion or electron












velocity, the loss rate of energy to the electrons is larger than to
2 3
ions by the factor (m./m )(p /zi p.). For He this ratio is approx-
1 e e 1 1

imately 20. This could account for some volume enhancement of energy,

especially in the case of fission fragment deposition. But, pe and

Pi are extremely small in both cases and most of the energy transfer

is to neutral particles. This effect then, is not significant in the

tenuous plasmas described here. Since it has been established [11,13]

that the nuclear reactor does affect the operation of a glow discharge

and thus laser operation, the changes occurring must be a function of

either changes in glow discharge structure or irradiation of the elec-

trodes. Data describing these effects are presented in Chapter 4.


2.3 Description of a Fission Fragment-
Produced Plasma

At present much effort is being expended in the area of char-

acterization of the tenuous "plasma" produced by fission fragment

sources [ 1 ]. The assembly of a set of kinetic rate equations is the

ideal approach to the characterization of this gas. But, because of

this method's detailed description of the number density of all species

and their important excited states, all reaction cross sections must

be known. Considering the number of species of a gas (atoms, ions, and

molecular combinations) and the excited states possibly present in

these species, this becomes an arduous task.

In most experimental processes, a small amount of impurity

gases are always present. These impurities enter into the kinetics of

the system and complicate the rate equation approach even more.












An example of this is the presence of a small amount of nitrogen.

Even in amounts of less than one part per million, spectroscopic anal-

ysis of an alpha particle or fission fragment excited gas show the

presence of the first negative system of N2 with very intense band

peaks. This indicates the presence of an additional ion generating

term of significant magnitude to alter the population of many species.

N is formed in several ways. The two most important transfer reac-
N2

tions are

+
Hemett + N He + N + e + AE (35)
metastable 2 2

and

He ++ N2 2He + N + AE. (36)
2 2 2



Thenormally considered te predominant reaction speified in reaction (35) ison of
normally considered the predominant reaction for the formulation of

the N ions. Thus the population of N2 is predominately a function
2 2

of the population of the metastable He(23s)'state and the recombina-

tion rate of N. N + will then increase as a function of increased
2 2

helium gas pressure, since the collision rate, as well as the meta-

stable population, also increases as a function of pressure. The reac-

tion described in equation (36) also produces the N2 ion, but at a

rate about five times slower than the Penning type ionization rate [43].

This is still significant, but the effect is diminished even more

because the population density of the molecular ion is far less than

the metastable state density.

The molecular helium ions are formed in many ways. The follow-

ing reactions generate the majority of the ions.












He + 2He He + He (37)
2

He(23s) + He(23s) He (38)


More information on formation and decay of these molecules is given

in References 43, 44, 45 and 46.

If the excited states are neglected and only number densities

of ion, atomic, and molecular species are included in the rate equa-

tions, the set of equations is reasonable and easily solved with the

inclusion of only a few unknown reaction cross sections. Examples of

this technique, which include the effects of wall losses from the

excited gas, are given in References 26 and 47.

In order to study fission fragment-produced excitation without

using rate equations, it is advantageous to assume some model. Such

models, although probably invalid for exact representations, should

use an equilibrium distribution of excited states based on Maxwell-

Boltzmann statistics or some combination of equilibrium distribution,

plus a calculation of individual excited states by approximate cross

section.

The latter approach was used by Russell [16] in his calculations

of population inversions in argon. This model is presented here since

it is reasonably complete and takes into account most of the processes

for forming excited states in an individual manner, rather than by

empirical statistical distribution methods. It does ignore all

excited states other than those of atomic argon, and it would require

extensive modification to include analysis of ion excited states which

are experimentally available for study in fission fragment-generated











plasmas. Also, no provisions are made for inclusion of impurity species

in the equation set, but they could be added without great difficulty,

since most impurity interactions are loss terms for the primary gas.

This semiequilibrium model is similar in many ways to that used

by Leffert [47], except that in the latter case no attempt was made to

calculate excited states densities.

Using the theory of Bates, Kingston, and McWhirter [48],

the production terms for argon excitation have been reduced to five

principal processes:

1. Recombination of thermal electrons with atomic ions

2. Inelastic collisions between excited atoms and thermal

electrons

3. Radiative transitions

4. Direct excitation due to fission fragments and high

energy secondary electrons

5. Formation of excited states in the products of dissocia-
+ *
tive recombination of diatomic ions (He + e He + He ).

Combining the above processes, an infinite set of excited

state density functions are obtained.



n(p)= () [ne (K(p,c)+ K p,q)) + (pq) (pq)
qjp q

+ ne n "(qK(q,p) + "(qGqpA(q,p)
q:p q>P


+ n2 FK + B + S n + S (39)
e 1_ (c,p) (p) a (p) D











where


n(p) = excited state number density, state p

n = electron number density

n(q) = excited state number density, state q


K(p,c) = inelastic cross section for collision of an excited

state with a thermal electron producing ionization

(loss term)

K = inelastic cross section for collision of an excited
(p,q)
state with an electron producing excitation from

p to q (loss term)

A( = radiative transition rate for loss by a transition
(p,q)
from excited state p into q

G( = correction term to account for optical trapping in

resonant transitions, transition from p to q

K( = inelastic cross section for collision of a thermal

electron with the excited state q producing the

excited state p (gain)

A = radiative transition rate for gain by a transition

from excited state q into p

G, = correction term to account for optical trapping in

resonant transitions, transition from q to p

K(c,p) = cross section for 3-body recombination from continuum

to excited state p (gain)

P = radiative recombination cross section producing

excited state p (gain)












S = source term for formation of excited states in the

process of dissociative recombination (gain, but

usually ignored).

S$ na (p) is a production term for excited states (p) due

to direct excitation from fission fragments and the secondary electrons

they produce. 'This term is assumed constant throughout the volume in

question and is analogous to the production rate for atomic ions.

S n = partial generation function that includes neutron
a

flux ', atomic number density n and a geometry

term S

= ratio of total excitation rate to atomic ion produc-

tion rate = 0.53 [24]

i() = ratio of the excitation cross section for array p

normalized with the cross section for the first

excited state (i.e., for argon, the 4s array),

assuming



I *(p) = 1.0 (40)
n=2
n-_


The cross section ratios can be calculated using the theory of

Gryzinski [49] for electron-atom interaction or by any other method.

If relative excitation cross sections were known from experimental

measurements, they also could be inserted at this point.

The argon ion density can be.represented as follows:












dAr+ 2 + Ar+
d- = S n mn Ar D 1
dt a a 1 2
nA
a



n (n K(cp) + (p)] n(P)K(Pc (41)

p=l

where

S@n = direct ionization source term from fission fragments

and secondary electrons

2+ +
-mn Ar = loss term by generation of the molecular ion Ar
a 2

Ar+
-D = loss by diffusion from the volume of interest.
1 2
n
a


The next three terms are, as described previously, loss by

3-body recombination, radiative recombination,

The molecular ion density equilibrium is


+ +
dAr Ar
2 2 + + 2
n Ar D Ar2 n D2 -- (42)
dt a D 2 e 2 2 (42)
nA
a

where

2 +
mn Ar = 3-body collisional production term
a

-a Ar2 n = recombination loss term
D 2 e


Ar
-D -- = diffusion loss term.
2 2
nA
a

The electron balance is

n = Ar + Ar, (43)
e 2












d(- n kT ) m 8kT
= 4 e kn (n q + n q )(T -T )
dt m e r e ei a ea e a
a e


+ Qre- Qrad + Sn E (44)
rec rad a Se

The first term on the right-hand side is the elastic collision energy

loss assuming a Maxwellian distribution of electron energy about the

electron temperature T .
e

q = electron ion cross section
ei

q = electron atom cross section

Qre = gain in thermal electron energy from inelastic

electron atom collisions

Qrad = line and continuum radiation losses

Sna E = energy source term from fission fragment interactions.
a Se

The time dependence of all the rate and equilibrium equations

presented is much faster than any change in the fission fragment source

term; therefore, steady state solutions can be obtained by setting the

time rate of change of Ar, Ar n T and n to zero.
2 e e (p)

Equations (39), (41), (42), (43), and (44) are an infinite set;

in order to generate a solution an approximation is required. Since

there exists a level adjacent to the continuum that is in Saha equilib-

rium at the electron temperature, it can be assumed that all levels

above this are populated with a normal Maxwell-Boltzmann statistical

distribution of states. If the n(p) set of equations is truncated at

this p level, a closed set is obtained. The obvious problem is what

level to truncate the excited state population calculation. Russell












assumed equilibrium above the 6p level and solved a set of thirteen

simultaneous equations by computer methods. His solutions for Ar I

indicate the presence of inversions in the 5S-4p and 4d-5p transitions.

Although the capability of lasing these transitions has not been

determined, other transitions of Ar I have been made to lase.

In review, this chapter has shown how a fission fragment

deposits energy and how one could calculate this deposition in several

geometries. It has been shown that the He (n,p)T reaction products do

not deposit significant amounts of energy into the volume of a glow
3
discharge in 15 torr He A model of the reaction kinetics of a fis-

sion fragment-generated "plasma" has been presented in order to

describe the basic processes that occur in such an excited gas.















CHAPTER 3


PLASMA RESEARCH APPARATUS




3.1 Introduction

Since previous researchers [33,35] have been fraught with great

difficulties when attempting spectroscopic analysis of in-reactor

plasmas, it was imperative that certain basic design criteria, devel-

oped as a result of this previous work, be applied to the construc-

tion of this apparatus.

Most of the difficulties involved the transmission of the

spectrograph of the low level light produced by the fission fragment

plasma. The glow discharge experiments have adequate light output to

overcome these problems; therefore, design was optimized for the fis-

sion fragment interaction experiments. The experimental chamber and

peripheral equipment were developed using the following criteria:

1. Chamber size must be large enough in average path length

to take full advantage of the energy available from the

average fission fragment.

2. U235 fissionable coatings should be maximized in area and

thicker than the mean free path of a fission fragment in

the coating.

3. Solid angle light availability to the optical system from

the plasma should be optimized.












4. Luminescent components such as quartz (light pipe) should

be eliminated from the optical path where subjected to

large neutron fluxes.

5. The experimental chamber should be located at the highest

flux position available in the reactor and should contain

only low thermal neutron cross section material (except
235
for U235 in the fission fragment plasma experiment).

6. A high light power spectrograph and extremely sensitive

photomultiplier should be used instead of photographic

techniques.

7. The design should include the capability of using high

purity gases.

The above guidelines were followed, utilizing trade-offs where

necessary, and the resulting apparatus, as described in this chapter,

proved quite successful in providing excellent spectral data.

A detailed set of criteria for generating data of the quality needed

to determine cross sections is available in Reference 32.


3.2 Primary System

Several alternate methods of studying spectroscopically an

in-core plasmawere investigated and two basic solutions were identi-

fied: (1) a small spectrograph installed next to the plasma, internal

to the reactor, providing a large solid angle light gathering capabil-

ity, and (2) the optical transmission of light up a tube to a spectro-

graph.












The second method was chosen as most practical. It employed

a stainless steel vacuum chamber 9 feet in length and 7/8 inch in

diameter, with a standard Ultek cross on top (Figures 3-1,3-2), where

the light is transmitted at right angles to the vertically positioned

spectrograph. Various experiments were inserted into the tube where

they were held at the bottom by gravity.

Emphasis was placed on the use of standard fittings, flanges,

and sizes. This allowed the replacement and storage of activated

9-foot tube sections, thereby allowing experimental procedures to con-

tinue without waiting for an activated section to decay.


3.2.1 Plasma Region

The outer stainless steel vacuum jacket allows insertion of

both the glow discharge and fission fragment apparatus into the high

neutron flux region of the reactor. The glow discharge experiments

included study of three different types of cathodes, flat, brush, and

hollow. The insulating structures for the flat and brush cathodes

(Figure 3-3) consisted of an outer 2 feet by 40-mm O.D. Vycor tube

with the bottom beveled inward. The cathodes made continuity with the

outer shell by gravity contact on the bottom of the casing. The ring

anode was spaced from the cathode by a 35 mm by 5-inch Vycor tube.

Electrical connection to the anode was made by an insulated wire run-

ning up the inside of the tube.

The hollow cathode structure was similar except a Vycor insert

was built to support the small cathode.





















GAS FILL
TUBE -




ELECTRIC F


'LENS ALIGNMENT
MECHANISM

MIRROR AND MOUNT

-QUARTZ WINDOW


RING ANODE U3 VYCOR TUBE


COATING

DETAIL


WIELDING CAVE

,- SPECTROGRAPH


CAVE /
SPECTROGRAPH

TOP VIEW


Figure 3-1 Reactor Mounting































GLOW DISCHARGE EXPERIMENT


FISSION FRAGMENT EXPERIMENT


Figure 3-2 Chamber Experimental Sections




















L---- u3i s'0'-,-



I---
0



<
\ zF--






CO

m L C0LL.
0 0 0
00 C3 )
00Z0


C) I--




m-a
O "- 0 LL.>
3c = CD uj = 0
0 I- Z LU 4


--- w z'=I = / ou



--
C<
LA4

^C- CD 4)---







3 "0 3 H
too <
C)
03

S0U





























"-'Z >' 0 0 ', 0 o.
__ >> < C
-J
I- LU
oo Co


LUJ : La







U- L
7fU 0











C I L) cm LaJ <
D CLLLL ) u









< co < I- -
V)1-
^*^" *^ c :
V^ V -1

"- ~-- ^ -^ ^ ^l c
^ ^ __ ___ SL ^ Y
^ = = ^ F S
<.^ .V .. < ^ ^ ^ ^ H y T-\ | z

\ / LU
\/ 0 C
V 0 c
LU OC 00 3












Removal of these structures to allow insertion of the next

assembly was a simple matter of turning the chamber upside down, and

sliding out the apparatus.

The fission fragment interaction assembly (Figure 3-3) was

a 40 mm diameter, 3-foot-long Vycor tube with a 30.5 cm length inter-
235
nal coating of U3 0 and an indentation at the 30.5 cm distance to
3 8

support the anode. For comparison, a glow discharge was occasionally

started inside this assembly, using the vacuum chamber bottom as

a cathode.


3.2.2 Vacuum Chamber; Optical, Gas
and Electrical Feeds

All chamber parts were designed around "Ultek" 2-inch

flanged fittings in order to aid in replacement and modification.

The outer shell (Figure 3-1) consists of a 9-foot, 1-7/8 inch O.D.

by 0.065 inch thick No. 304 stainless steel tube with a heliarc-

welded plug on one end a standard 2-inch fitting heliarc-welded on

the other end.

Placed vertically on top of the long tube is the 2-inch

standard cross. Matched to the cross is electrical feed through on

one side for connection to the anode, gas feed on the top and a vinyl

sealed quartz window on the other side.

The optical system consists of a 500-mm focal length fused

quartz plano-convex lens held and focused vertically by three align-

ment wires (Figure 3-1). A right angle front surface mirror is held

and aligned by a mount inserted into the center of the cross. The

vinyl seal quartz window allows light to escape from the chamber and
















B E
I- i

"B i


=1
= =


I












be focused 2 feet to the side on the entrance slit of the spectrograph.

Thus, the plasma is viewed through the ring anode at a solid angle
-4
of 1.18 10-4 steradians, and focused by the lens on the spectro-

graph entrance slit after passage through the exit window.
-6
The maximum vacuum attained by the system was 1 x 10- torr

and the maximum design pressure was 760 torr.


3.3 Uranium Coatings

3.3.1 Coating Requirements

One of the most important factors in the production of

a fission fragment interaction plasma is the fabrication of a suitable

uranium coating for the experimental chamber. Requirements for the

coating included the capability of deposition on the interior of a var-

iety of cylindrical surfaces, such as Vycor and stainless steel.

Interior dimensions from 10 mm to 44 mm and lengths up to 3 feet were

encountered.


3.3.2 Coating Thickness

In order to maximize the source strength and thus the

number of fission fragments available, coatings were made thicker than

the range of a fission fragment in the source material. The range of

a fission fragment passing through U30 the final product of the

coating procedure, is calculated [41] as follows, with K and E from

Figures 3-5 and 3-6:


R [2] = KE2/3 (45)
cm







58
















.6"" i':lt : + ^E ^::::::
.6





.4

"E: -- lo- ... .



.2




0 4
0 5 10 15 20
DENSITY (g/cm3)


K as a Function of Density


Figure 3-5
























s

X 100

CO

o 90


0
5 80
80
z

u 70

z
M.


- CS














U 2











I I I I I I I I

80 90 100 110 120 130 140 150


MASS NUMBER A














Figure 3-6 The Kinetic Energy of Fission Fragments as

a Function of Mass Number r38]











Density of U30 = 8.3 ,
cm


K for U0 = 0.294 mg
3 8 (cm )(MeV)2/3'


Maximum expected fission fragment kinetic energy = 97 MeV,

for mass = 95; therefore,

mg
R = 6.206 mg
2
cm

A coating thickness of 10 2 was considered adequate to
2
cm
account for some nonuniformity of application.


3.3.3 Review of Methods and Chemistry

There are several methods for producing coatings including

ion deposition, electroplating, powder distribution on a binder, and

coating and ignition.

For the purposes of the fission fragment interaction exper-

imental work the most flexible and inexpensive technique, coating and

ignition, was chosen.

Much detailed work was done in this area at Los Alamos Scien-

tific Laboratory around 1945 under the title of "Zapon Spreading

Techniques" [ 51].

The basic principle [52] of the Zapon spreading technique is

as follows: An alcoholic solution of the nitrate of the substance to

be deposited is mized with a dilute solution of Zapon lacquer in alco-

hol, acetone, or Zapon thinner. The resulting solution is spread or

painted on the foil backing, allowed to dry, and then ignited to remove

organic substances and to convert the nitrate to oxide.












Many variations of this technique are possible, and it is

applicable to a wide variety of substances. The procedures used in

this project generally follow those suggested by this work.

Many alternate methods of placing coatings on surfaces were

evaluated, but the only successful results were obtained when many

thin coatings were added in succession.

The requirements for a successful coat include a liquid that

can be deposited in thin layers, does not evaporate too quickly, and

includes a binder to transfer the fluid to a very viscous substance

upon evaporation of the solvents. The binder is necessary to prevent

an unequal distribution on the surface.

The Zapon spreading technique for uranium used ethyl alcohol
235
as a solvent for U 02[NO3 2. 6H20 and added Zapon lacquer as a

binder for the uranyl nitrate salt upon evaporation of the ethyl

alcoIol.

Zapon lacquer is not available, so a chemical substitution of

"Testors" butyrate dope containing methyl cellosolve and isobutyl

alcohol was used. Unfortunately, this lacquer forms a gel with ethyl

alcohol so n-butyl acetate was added to dissolve the lacquer.
235
A detailed description of the preparation of uranium foils is

given below.


3.3.4 Chemical Procedures

The procedure for producing the coating solution, as

detailed in Figure 3-7, is variable depending upon the form of metal

available. The following procedure starts with highly enriched

uranium (93%) metal.






































L

-o

C O
e'lJ 0

r o
3.


t0
CM 0






3 cn
-c
-0










































many

proc


1. Dissolve the metal in concentrated nitric acid (HNO3 ).

2. Dry the solution on a hot plate with the temperature below

2000F in order to avoid hot spots and thus production of

any UO2. The result is UO2[NO312 *6H20, a bright yellow

powder.
235
3. For each 4 grams of U2, add 20 ml ethyl alcohol, 20 ml

of butyl acetate, and 20 drops of butyrate dope.




4. Variations in the above mixtures are possible, but the pro-

portions given seem to work well in the coating process.


3.3.5 Mechanical Coating Procedure

The art of producing a U308 coating involves application of

thin coats in order to avoid flaking of material. The following

:edure is detailed in Figure 3-8.

1. a. Vycor base Using a clean tube, flush with 6NaOH

and then with distilled water.

b. Metal base Using a clean tube, flush with 6N HNO3

and then with distilled water.

2. Place the cylinder on a rotating horizontal mount.

3. While the cylinder is rotating, place a thin, even coat

of solution on the inside of the tube, using a coating

tool such as a camels hair brush or, for long tubes,

a tube swab mounted on a wooden dowel.

4. While the cylinder is rotating, dry the coating with air.


















































Q)

1:
0







-4J
0
C,
u










o


0







r-l
r74
bO (
C 0











5. Fire the coating using a Fischer burner or propane torch.

This will first oxidize the uranyl nitrate to brown-colored

UO2 and then at higher temperatures to black-colored U30 .

The butyrate dope will evaporate and leave a residue of

less than 1 per cent.
2
6. Repeat until coating thickness is at least 10 mg/cm2

Forty coats were adequate for the Vycor substrate.


The resultant coatings appear as a black ceramic material with

a surface area much greater than that of a flat substrate. This is

due to the porosity of the ceramic coat and has likely aided greatly

in the deposition of larger amounts of fission fragment energy into

the surrounding gas.
235
Since U is radioactive and most samples contain traces of

other isotopes, care was taken to properly handle and contain the

chemicals involved in the coating procedure.



3.4 Support Systems

Acquisition of various data describing the effects occurring

in a plasma is dependent on the following support systems.


3.4.1 Reactor: Neutron Flux and Gamma Dose

The experimental chamber was designed to be inserted into

the University of Florida Training Reactor. As shown in Figure 3-4,

the UFTR is an Argonaut-type reactor licensed to operate at 100 kW.

The thermal neutron flux available.in this facility at 100 kW is shown

in Figure 3-9 as a function of distance from the bottom of the





















































I I I I I I
5 10 15 20 25 30 35

DISTANCE FROM END OF CHAMBER (cm)


Figure 3-9


Distribution of Thermal Neutron Flux
along Experimental Chamber


U ,

cU




I
C\J










U--



1




0












experimental chamber. These measurements were taken, using standard

gold foil dosimetry. In order to measure the thermal neutron flux

actually available to the plasma region, all foil irradiations were

done inside the stainless steel experimental chamber. At 100 kW the
11 2
average flux over the length of the glow discharge is 4.7 X10 n/cm -

sec, and, over'the fission fragment plasma, 3.8 X 1011 n/cm2-sec.

Gamma dose in this area at 100 kW is estimated to be 1.1 x107 R/hr

at equilibrium.

Reactor power data relate directly to the flux and are based

upon two calibrated compensated ion chambers placed near the core.

The experimental apparatus was placed in the 1-7/8 inch center verti-

cal access port located between the two fuel regions. This is the

highest thermal flux region available in the reactor, but it is also

the highest reactivity worth region. This presented certain problems

which will be discussed in Section 3.4.5.


3.4.2 Electrical Systems

There are two systems involved in operation of the apparatus,

other than those power supplies and leads involved in diagnostic equip-

ment.

In order to supply a high voltage DC field to the chamber

when studying fission fragment interaction light amplification, a 10 mA

5 kV calibrated DC supply was used. For current measurements, the

power supply was floated with respect to ground on a 100 ohm resistor

across which was placed a Keithly DC microvoltmeter.












Electrical systems for operating the glow discharge were more

sophisticated because an arc in the chamber could destroy the contin-

uity of the experimental procedure by altering the cathode surface.

Therefore, a sensing circuit was designed to monitor the voltage across

the glow discharge. Upon sensing the large negative transient when

the glow discharge changes from an abnormal glow to an arc, Figure 3-10,

an extra resistive load is switched into the series load chain in order

to quench the arc and return the system to a glow discharge. Although

the quenching system operated effectively, it was rarely needed, as

will be explained in Chapter 4.

Glow discharge I-V data were obtained either by two voltmeters,

one connected to read glow discharge voltage and another across a 2 ohm

calibrated resistor to measure current, or by direct connections to an

x-y plotter.

The glow discharge electrical source consisted of three series-

connected ultra-stable Lambda power supplies with both variable volt-

age and variable load controls. This system allowed operation over

wide ranges of gas pressure and glow regions.

All of this equipment was installed in racks and operated

remotely on the floor of the reactor cell, since dose levels were too

high on the reactor top face to allow long term access by personnel.


3.4.3 Gas Filling Systems

Two different systems were used for filling the chamber.

One of the systems, designed for ultra pure gas filling and mass anal-

ysis of filled chambers f31], was located remotely from the reactor
































C- /,L
U w



-J I Z -- -- *-




0f 1 0I 0
--Z I LLJ UJ ,
0 I 5I C
O^ (- 1 i o
0 C.5 Ct: Z (A


2r














-S.


NOI II SNM

MOlo 3W


U


-w


VNOa)OD


30)IVHOSIa

IN2SNMOI


SNOI93d
IN3ano3 MOi


--t











building, and could only be used in the glow discharge experiments where

neutron activation of equipment was minimal. Its capabilities allowed

precision filling of gases with impurities of less than 1 part per

million, and bake-out of wall adhered impurities such as H20. The min-

imum chamber pressure that was reached during pumping and bakeout was
-6
1 X 10-6 torr, assuring the absence of substantial leaks in the system.

For the fission fragment interaction experiments, where the

U30 foil is highly activated, gas fill was done while the apparatus

was mounted in the reactor. The gas filling system used in this case

consisted of a remotely situated two-stage vacuum pump, needle valve

mixing equipment and pressure indicating systems. Minimum pump pres-
-3
sure on the system was 1 x 10 torr, thus allowing an impurity source

5
of air, with a 1 atm fill of pure gas, of less than 1 part per 105

It was expected that the main impurity would be N2 and that this would

show up in the spectrographic analysis.

Since the above facility was not capable of ultra pure filling,

it was decided that commercial grade gases would be adequate for these

early experiments. The analysis of the argon and helium that was used

did, in fact, show impurities, mostly N2, of less than 1 part per 105

All experimental work was done with the chamber sealed due to

safety regulation as will be described in Section 3.4.5.


3.4.4 Shielding

Two main purposes were served when the elaborate shielding

cave system was built.











First, the dose available from the center vertical port without

shielding would not allow reactor operation much above 1 kW, two orders

of magnitude below needed flux levels. Second, the ultra sensitive

photomultiplier used in diagnostics is sensitive to gamma and without

shielding the DC level and subsequent noise is intolerable.

The shielding system is diagramed in Figure 3-1 and shown in

Figure 3-11. The cave around the center vertical port consists of

lead filler rings surrounded by steel and lead blocks. Borated poly-

ethylene blocks were used to moderate and absorb either fast or thermal

neutron flux transmitted up the access port.

A secondary lead cave was constructed around the photomulti-

plier tube and constituted the light tight mounting case for the photo-

multiplier as well as a container for the immediate electronics required

for its operation.

With this shielding, the DC shift from gamma sources was 7.5 mV

at 100 kW reactor power with a measurable spectrum line height as small

as 2.5 mV. After processing the signal, a noise component of about

1.1 mV overlaps the data.


3.4.5 Safety

Experimental procedures that involve insertion of absorbers

and unclad fuel into a high worth region of a reactor must also include

many safety considerations.

The center vertical access port, where the experimental work

was done, is located, as shown in Figure 3-4, between two fuel regions

in a graphite moderator. This area is considered a high reactivity
































































Figure 3-11 Chamber Mounting and Shielding Cave












worth region because it couples the two fuel regions together. Any

thermal neutron absorber inserted into this region affects the reactor

criticality situation much more than this same absorber placed, for

instance, in the thermal column, where the reflection of neutrons back

into fuel is minimal. The experimental apparatus is not an extremely

heavy absorber, but its negative reactivity effect on the core was

large enough so that in certain circumstances, where heavy absorber

gases were used and where the tube was inserted to a greater depth

than normal, the reactor could either not achieve criticality or could

not run at high power due to the negative temperature coefficient of

reactivity.

This problem could be overcome only by allowing removal of

enough reactor control blade (neutron absorber) so that the decreased

absorption would overcome the negative reactivity of the experiment.

Initial control blade removal levels were limited by the reac-

tor subcommittee of the University of Florida Radiation Control Com-

mittee to less than 0.6 per cent positive reactivity insertion. With

a positive reactivity insertion of greater than 0.6 per cent, the

reactor is in a prompt critical excursion or critical on prompt neutrons

only and thus in an uncontrollable highly. dangerous situation (pos-

sible destruction of the core). But, due to the large negative reac-

tivity presented by the experimental apparatus that would counter-

balance positive reactivity, the reactor subcommittee allowed removal

of the control blades to a position corresponding to greater than 0.6

per cent positive reactivity when the following requirements were met:











1. The apparatus is absolutely fixed and cannot move.

This requirement was satisfied by proper placement of the

massive weight of the shielding cave which could not be moved without

deliberate actions.

2. The gas fill is fixed in pressure, less than one atmo-

sphere, and will not be changed while the reactor is

critical.

This requirement was satisfied by sealing the chamber before

each run and by placing a relief valve at the fill valve to keep pres-

sures greater than one atmosphere from entering the chamber.

The fission fragment interaction assembly somewhat eased the

negative reactivity problem due to the coating's addition of about

4 grams of fuel to the high reactivity worth region.

After many hours of high power irradiation, the glow discharge

experiments had an activity which gave surface dose rates of about

400 mR/hr. This decreased to about 100 mR/hr in a few days and was

deemed not much of a problem, especially since there was no loose con-

tamination. The fission fragment interaction assembly contained a

large area of unclad fuel (coating); when removed from the reactor, the

dose rates at the tube surface were about 100 R/hr. Therefore, remote

handling of the apparatus by the reactor overhead crane and long term

storage in the water shield tank (see Figure 3-4) was necessary before

disassembly of the chamber. The internal components of the chamber,

especially the upper section, were heavily contaminated with mixed fis-

sion products and had to be handled properly during disassembly.











The vacuum pump exhaust which contained various radioactive
41 40
gases, usually Ar from activated Ar4, was connected to the

UFTR's air handling system for filtration, dillution, and disposal.

The problem of cooling the glow discharge was handled by the

same air handling system, which displaces about 12 CFM through the

center vertical port.



3.5 Data Acquisition Systems

I-V, reactor power, and spectroscopic data were taken during

experimental runs. Current and voltage measurement are an integral

part of the electrical supply system and were described in Section

3.4.2. Data on reactor power, and thus neutron flux, were obtained

from the reactor operating console, utilizing two compensated ion

chambers.

Spectroscopic data were obtained using a McPherson, 3/4 meter,

criss-cross, Czerny-Turner mount, scanning monochrometer with the
o
grating blazed for 3000 A and an EMI 9558 QB photomultiplier tube

mounted in the shield cave previously described. Since the lens does

not identically focus all wavelengths on the entrance slit, the system

was operated in a slightly defocused manner to allow for less precise

alignment of the optical system. The photomultiplier cathode was

operated at -1300 volts DC and its load was 1 megohm (plotter input).

A Moseley 7000A X-Y plotter was operated on X time base while the

spectrograph grating was driven at constant speed, thus giving the

spectral line structure as a function of wavelength. Noise problems











were minimized by the use of a low pass filter and by operating both

plotter and grating drive slowly enough to allow detailed plotting of

line shapes.

The system was grounded, using a 10-foot rod driven into earth

below the reactor. Braided copper strap was used for connections,

allowing no ground loops and using the stainless steel experimental

chamber as the main ground bus.

The following chapters will detail the use of the equipment

herein described and present results and analysis of the effects seen.
















CHAPTER 4


ANALYSIS OF "IN-REACTOR" PLASMAS: FISSION
FRAGMENT GENERATED AND GLOW DISCHARGE




When referring to the photon output from the "in-reactor"

plasma, it can be assumed that the plasma is optically thin or that

there is very little absorption of photon energy by the gas. In order

to analyze the optical emissions and determine some characteristics of

the excited gas, some model is useful. Since it is presupposed that

the excited gas is in thermal nonequilibrium, the most valid model, as

previously stated, is a coupled set of interaction rate equations.

Without a complete set of cross section data for generation and decay

of excited states, this type of model is not useful. These data are

not available in complete form; therefore, the rate equation and com-

bination type approaches [15] are by-passed in favor of a more easily

applied method, Boltzmann analysis.

The assumption that fission fragment excitation produces a gas

that is in thermal equilibrium, where the electron temperature and

distribution of excited states is defined by Fermi-Dirac or Boltzmann

statistics, is probably not valid. But, there are certain features of

this type of model that can prove useful in determining trends in the

excited state densities and in the search for a so-called "negative

temperature" or population inversion for pumping of a laser. The most










important feature of this model is the capability of describing the

plasma by an exponentially dependent factor analogous to the Boltzmann

temperature. Pressure-dependent plots of this factor and the excited

state densities from which it is derived are described in this chapter.

Also included is a review of all experimental fission fragment inter-

action data generated by this research and an analysis of these data to

determine the characteristics of the generated plasmas. A review of
3 4
the experimental results of the He and He4 glow discharge reactor irra-

diations is also included.


4.1 Experimental Procedure and Data Analysis:
Fission Fragment Interactions


The three gases bombarded with fission fragments--helium, argon

and carbon tetrafluoride--gave a luminous output that was easily mea-

surable both for atomic spectral line peaks and for molecular band peaks.

Data were recorded by the methods described in Chapter 3. A typical

spectral sweep is shown in Figures 4-1 to 4-5 for helium at 100 torr

pressure. A second sweep is required with a different recorder sensi-

tivity when the line or band intensity exceeds the plotter peak value

(Figure 4-1).

An interesting feature of these data is the significant back-

ground produced by gamma interaction with the photomultiplier. This

background supplies the error in the spectral line intensity measurements.

The large amount of gamma shielding shown in the previous chapter reduced

this background to a tolerable level so that the measurement error on

the low level lines was acceptable. Other sources of error, such as

deviations in reactor operating power and changes in the nature of





* i ..


1 ?

2700
2700 A + 33361


0A
A


Figure 4-1 Helium Spectrum


I,0e n~,


}4<0]t- ^'o I'" a^. l ofc 9^'2" jW























0 0
3850 A 4410 A


j. l** JS UWd( D w .* A'.., 30d P*+,/ r ,,;'/.


,,


























I-- I




p
I
in~ ZT

tci"


13


O O



3336 A 3967 A
1 ,


Pi7A*- ItS Cl,,- /12 e" I 141- IS-A o 0"' 1 l
"t ro V'M LI 110F.I


! a
i ,' "


.i r
- 9 A'6








3967 A 461


Figure 4-2 Helium Spectrum


+rr
*
n;

i

Y"PL.


~j


A- wy


3aU /-- U4 I


Hw.1~*A Awtj sk.U /^ ro*~.










f5r~


* r11


I I
i l 1 : 1 :


o o
46A 2 ,4 A

* 4615A A5247A


5 7 / s A '-*"
Awl$ 0& rw ~


934-r v~
Jksa .1&L -


- ~ ,~1 7


i ii

5247 A 5877 A


Figure 4-3 Helium Spectrum


'"'"''


Ma".4,l,(, l /,>? io,,


Wis- 9 51 '(17


"^ I T


'Ir- .


~i ri


I'
ii


--- i











0'irl %nW- Mel fwVT soo .i *ic<-v
eB *f' "


F-


0 0
5877 A +- 6506 A


.1~.


06
6506 A + 7136 A


Figure 4-4 Helium Spectrum


y.41 f^ 5


t ~ ~ of /1. aftof^ ^ '^
isoi ? wri, Uf'c, aod 'Y)l
NoIbc IMAO


I.*- '. I ,*











W. 6s, AP,^ Itto 4 AW6% .7. 9C "*. -_j /"
if go 400-V-
r~crCLIjA ,4~ V.-Y 71 o


[ I


0 0
7136 A 7762 A


774: ';k. 4070


- '* ."rv rtx, *-I'- ,,,,O WP .--
Cp 6A_ Jd


--I i ---i- 9


7762 0 0
7762 A + 8390 A


Figure 4-5 Helium Spectrum


I: 1i






,74" a


'F Qe
T gC
d ^`


7/16 1 17 )L










the gas fill, were deemed insignificant in most cases and are noted

where applicable. Errors associated with the band height measurement

of CF4 and its reaction products cannot be established since the molec-

ular dissociation and recombination of the various species present are

continuing kinetic processes reaching equilibrium in a long time

frame (about 3 hours).

At 100 KW reactor power, the error in the intensity is a con-

stant factor equal to 1.125 mV divided by the appropriate calibration

factor for the wavelength of interest. This error (AI) is shown on the

linear intensity versus pressure plots. For the logarithmic calculation

of relative excited state densities, the error can be expressed as AF,

the standard error, where

AF = AI (46)


and

F = log [ (47)


Therefore,

6F
= log e] (48)

and

AF = [log0e] (49)


This is the error indicated by the error bars on the

Boltzmann plots.

Calibration of the system was accomplished by using a tungsten

lamp calibrated by the National Bureau of Standards and a mockup of

the experimental system. All of the optical elements included in the










experimental apparatus were included in the calibration procedure.

To avoid both reflection and second-order errors in the continuum

measurements, stepwise filtration using calibrated dielectric cutoff

filters was applied to the light source. No reflection or second-order
O o
terms were observed over the calibration range, 3500A to 8500A. The

system response curve was folded into the spectral radiance curve and

the result normalized to the highest value. This resulted in a rela-

tive response function which provides calibrated relative intensity data

when divided into the measured intensity of a line. For wavelengths
0
below 3500A, the cutoff value for reasonable accuracy using a tungsten

lamp, theoretical response curves were used. The theoretical response

calibrated line intensities were not used in the Boltzmann plots, thus

introducing no systematic error in these calculations. These low

wavelength data are given as a function of pressure and population

density where their intensity values are not compared to other line

intensities. Second-order output was observed for the large bands of

the N+ first negative system in helium and the OH 3064. system in argon.
2

Since the system response below 3000A is small, other second-order lines

or bands were not observed from the ultraviolet component of the spectral

output.

For the fission fragment work, the spectrometer entrance and

exit slits were set at 100 microns, providing a resolution of about
o o
0.5A with a dispersion of 16A per millimeter. With this resolution,

accurate separation of some of the argon lines such as 4332.OA and
0
4331.2A was not possible; therefore, for these few cases, individual

line intensities were estimated using the relative intensity ratio