DELAYED COINCIDENCE SPECTROSCOPY OF FISSION FRAGMENT

EXCITED GASES

By ■ ■
GEORGE ROBERT SHIPMAN

A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF

THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA
1976

. ■ ACKNOWLEDGMENTS

As with any research effort, this dissertation has not
been the result of one person's efforts. The patient and
fatherly concern of Professor R. T. Schneider in the con-
tinued support of this project is gratefully acknowledged.
The willingness of my chairperson, H. D. Campbell, to pro-
vide help and insights into the theoretical aspects of ■
fission fragment produced plasmas was balanced by the equal
and complementary efforts of E. E. Carroll who suggested the
path through two apparently insoluable experimental diffi-
culties. Mr. Ken Fawcett, a man with an uncanny ability to
revive aged and senile electronic equipment, provided for
the care and well-being of the detection system. Mr. Ernie
Whitman and Mr. Joe Mueller provided help and expertise in
numerous and diverse fields. I have benefited from discus-
sions with many of my fellow students but especially Drs.
Bob Davie, John Davis and Jim Fuller. I have also had the
benefit of four friends and helpers during the course of
this research. Dave Sterritt and George Strickland devised
and wrote the programs for the data reduction while Dudley
Carter helped in almost every phase of the work. I pass
this system and the remaining problems on to Tom Maguire
who has been of immense help in the later stages of this

work. I wish to thank my wife, Barbara Jo, for typing and

ii

and retyping the rough draft and for her nearly patient
encouragement during the long term of this effort.

Ill

TABLE OF CONTENTS

Page
LIST OF FIGURES

ACKNOWLEDGMENTS ^-^

VI

ABSTRACT viii

CHAPTER

1. INTRODUCTION , , 1

1.1 Background! 1

1.2 General Description 2

1 . 3 Previous Work . . ' 4

1.4 Organizational Plan 4

2. THEORETICAL CONSIDERATIONS ... 6

2.1 Introduction . 6

2.2 Fission Fragment — Gas Interactions. ... 7

2.3 Rate Equation for C^Ru 18

2.4 Relating Solutions to Observables .... 20

3. SYSTEM AND CALIBRATION 22

3.1 Introduction and Design Criteria 22

3.2 . System Description 23

3.3 Calibration ■ . . 29

3.4 Output 31

3.5 Scanning Mode 32

4. ANALYSIS AND RESULTS 35

4.1 Introduction 35

4.2 Programs 35

4.3 Analysis 37

4.4 Pulse Height Distribution 41

5. OTHER GASES AND GAS MIXTURES 4 7

5.1 General Comparison of Mixture Spectra . . 47

5.2 Neon/Nitrogen Mixtures 47

5.3 Helium/Nitrogen Mixtures 59

5.4 Argon/Nitrogen 66

5.5 Carbon Tetraf louride 71

5.6 Population Inversion Study. ........ 77

iv

Page

6. CONCLUSIONS AND FUTUIE WORK 81

6.1 Conclusions 81

6.2 Future Work 8 2

6.3 Scanning Mode Recommendations 84

APPENDICES

1. ENERGY DEPOSITION. 86

2. NE-111 AS A CALIBRATION SOURCE 89

3. GAMMA VERSUS TLP FOR START PULSE 9 2

4. GAS PURITY 94

5. FISSION FRAGMENT TRANSIT TIME 96

6. 252cf CHARACTERISTICS 99

BIBLIOGRAPHY 102

BIOGRAPHICAL SKETCH 104

V

LIST OF FIGURES

FIGURE

1. Typical I(t) Curve
2.

Page

3.

4.

5.

6.

7.

8.

9.
10.
11.
12.

13.
14.
15.
16.

17.
18.
19.
20.

Typical Decay Curve Showing Gamma Coincidence

Peak 9

Pressure Variation of Decay Time 12

Decay Time Versus Pressure for 337. Inm. ... 13

Inverse Decay Time Versus Pressure for 337. Inm 14

Metastable Transfer in He/N2 15

Double Decay Curve 16

Relevant Energy Levels of He, Ne , A, and N2 . 1^

System for Lifetime Measurement 24

Scan of Pure N2 ^^

Peak Intensity Versus Pressure for 337. Inm.

Integrated Intensity Divided by Decay Time
as a Function of Pressure .

Integrated Intensity Versus Pressure
Fission Fragment Energy Spectrum.

40

42
43
45

Fission Fragment Distribution from TLP Pulse Heights . ^^

Scans of Nitrogen Impurities Ne , A, He, and

Pure N-

Scan of Nitrogen Impurity in Pure Neon
Decay Time Versus Pressure for 585. 2nm
Decay Time Versus Pressure for 585. 2nm
Decay Time Versus Pressure for 585. 2nm

48
50
51
52
53

VI

Page

21. Decay Time Versus Pressure for 585. 2nm .... 54

22. Decay Time Versus Pressure for 337. Inm .... 55

23. Decay Time Versus Pressure for 337. Inm .... 56

24. Decay Time Versus Pressure for 337. Inm .... 57

25. Decay Time Versus Pressure for 337. Inm . ... 58

26. Decay Time Versus Pressure for 391. 4nm .... 60

27. Decay Time Versus Pressure for 391. 4nm .... 61

28. Decay Time Versus Pressure for 391. 4nm .... 62

29. Scan of Nitrogen Impurity in He 63

30. Scan of He with 1% N2 64

31. Relative Intensity Versus Pressure for

391. 4nm 65

32. Scan of He with 10% N2 67

33. Metastable Transfer of He/N2 68

34. Decay Time Versus Pressure for 391. 4nm .... 69

35. Scan of N2 Impurity in He, Excitation Period Only. 70

36 . Scan of Nitrogen Impurity in Argon 72

37. Scan of Argon with 10% N2 73

38., Fast and Slow Decay in A/N2 74

39. N2 Impurity in CF4 76

40

Population Densities on N2(B) and N2(C) Versus
Time 79

41. Measured Is.t and 2nd Positive Populations. . . 80

42. Using Gamma Versus TLP as Start Pulse 93

Vll

of Cf evaporated on a platinum disk. The ^^^Cf emitte

Abstract of Dissertation Presented to the

Graduate Council of tiie University of Florida

in Partial Fulfillment of the Requirements for the

Degree of Doctor of Philosophy

DELAYED COINCIDENCE SPECTROSCOPY
OF FISSION FRAGMENT EXCITED GASES

By

George Robert Shipman
December, 1976

Chairman: Hugh D. Campbell, Ph.D.

Major Department: Nuclear Engineering Sciences

A time resolved single photon counting system was used
to study the emission spectra of N2 under fission fragment
bombardment. The source of fission fragments was O.Olyg

:ed

two groups of fission fragments with energies of about 80
MeV and about 10 5 MeV at a rate of 6,000 per second.

The system used a time-to-amplitude converter (TAG)
which converted the time difference between two timing
signals into a voltage. The "start" pulse was supplied
by a plastic scintillator which detected the prompt fission
gamma rays. The "stop" pulse was generated when a single
photon struck the photomultiplier mounted on the spectro-
graph. The probability that a photon occurred was pro-
portional to the population of the excited state giving
rise to the observed spectral line. The number of photons
in each time interval was thus related to the excited
state density in that same time interval. After accumulat-
ing enough counts to give acceptable statistics, a curve

viii

was generated which plotted the number density of excited
states as a function of time-after-fission. The portion of
the curve which occurred after the fission fragment had
passed out of the field of view (the 'afterglow) was least
squares fitted to a decreasing exponential.

It was observed that the excited state population
decayed more rapidly as the pressure was increased. Ex-
trapolation to zero pressure allowed determination of the
radiative lifetime while the rate of change of decay time
with pressure was related to the collisional quenching
cross section. The resolving time of this system was less
than two nanoseconds.

" The radiative lifetime of N„ (C Flu) was found to be
42ns±1.5ns and the collisional transfer rate constant to
be 1.0x10 cm sec

Several other noble gas-nitrogen mixtures are described
in a more qualitative manner.

IX

CHAPTER I
INTRODUCTION

1 . 1 Background

This dissertation is divided conceptually into three
parts. First, there is the description of a new system
which was developed for the analysis of gases excited by
non-periodic, transient, randomly occurring events. Second,
this system was used to make a careful study of one parti-
cular excited level in one particular gas. Third, a more
qualitative description is given of some measurements which
were done on various gas mixtures exhibiting both resonant
and nonresonant behavior.

The work which led to this dissertation was initially
conceived as an alternative to an in-core reactor experi-
ment. The study of fission fragment excited gases had been
going on for some time in connection with both the gaseous

core reactor and the nuclear pumped laser program.' 2 3, 4

2 3 5
A typical experiment involved placing a U lined cylinder

v-zithin the core of the reactor, filling the cylinder with a
chosen gas or gas mixture, and observing the emitted photons
with a spectrographic system.' ^ There were, however, sever-
al limitations inherent in this type of experiment. Because
of the nature of the reactor environment, stringent safety
precautions had to be maintained. These precautions, while

1

2

necessary, severely limited access to the experimental
device and restricted data acquisition, system modification,
and system adjustment. Excitation using the spontaneous
fission fragment emitter ^^^Cf was seen as an attractive
alternative. This isotope provides a safe, small, self-
powered source of energetic fission fragments closely
resembling in all respects the fission fragments emitted by
235u (see Appendix 6). At first it was thought that the
extremely low flux, about nine orders of magnitude less
than in the reactor, would make the spectroscopic diagnos-
tics an insuperably difficult problem. The resolution of
this problem proved to be the adoption of single photon
counting techniques.
1.2 General Description

Preliminary intensity measurements indicated that only
about one photon was observed for each 100 fission fragments
which passed through the region of observation. In addition,
it was soon realized that the time that a fission fragment
took to cross the chamber (about four nanoseconds) and,
hence, the length of the excitation pulse was extremely
short as compared to the mean time between fission fragments
(about 0.3 milliseconds) thus allowing a unique assignment
of an observed photon to a particular fission fragment.
Since we could tell which fission fragment caused vv/hich
photon, a measurement of the time interval between the
passage of the fission fragment and the emission of the
photon allowed a determination of the lifetime of the atomic

3

or molecular level from which the phototi originated. We
had thus created a fluorescence decay system utilizing a
totally new type of excitation. Since the probability that
a photon will be observed at a certain point in time is
proportional to the population density at that time, a
history of the photon flux is a history of the excited
state population. After the time intervals between many
fission fragment-photon pairs had been' measured , such a
photon flux history was obtained. Determination of the
interval between excitation and photon emission required
accurate knowledge of when the fission fragment transited
the excitation region. In all previous fluorescence decay
systems, the excitation source had been triggerable, and
this trigger had provided the timing "start" pulse. Here,
however, the excitation was a randomly occurring, non-
periodic, transient event. Two independent methods were
developed for providing the timing "start" pulse, both of
which were found to give identical results.

What originally started as an alternative to a reactor
experiment evolved into a system of great versatility,
capable of providing information not only about fission
fragment excited gases, but also about the fluorescence
decay of atomic and molecular levels.

The extent to which the information obtained in this
experiment can be scaled to the reactor situation has not
yet been resolved. Several studies' ^ ^ have shown that
population density increases linearly with reactor power

over several orders of magnitude. Populations probably
remain linear until electron -electron interactions begin to
dominate the thermalization of the swarm, a density which
has not yet been reached in gases up to one atmosphere.
1 , 3 Previous Work

Previous work germane to this dissertation has appeared
in a diverse range of fields.- This work has been recently
reviewed'* and the interested reader is referred thereto.
If, from the vast body of literature available ' upon the
subject of the interactions of fast charged particles with
gases, we restrict our attention to those which are con-
cerned with the passage of fission fragments and the sub-
sequent emission of photons, we find that they are, with
one' exception, concerned with the steady state case of
excitation within a reactor. That exception is the work of
Axtmann^ and his students at Pri-nceton. They have also
studied the C^]\u level of nitrogen, but with quite a dif-
ferent experimental technique and obtained essentially
identical results.
1 . 4 Organizational Plan

The plan of this dissertation will be to first develop
in Chapter 2 the theory necessary to understand the optical
emission of fission fragment excited gases in a general way,
and then look at the rate equations for the C Ru level in
pure nitrogen. In Chapter 3 will be described the experi-
mental apparatus and its calibration. Chapter 4 will be an
analysis of the results for pure nitrogen, including values

5

for the radiative and collisional rate constants. Chapters
will describe in a semiquantitative way the measurements
which have been made on other gases and gas mixtures. After
a short concluding discussion, a series of appendices will
examine in greater detail some questions which relate to the
analysis rather than interpretation of the results.

CHAPTER 2
THEORETICAL CONSIDERATIONS

2 . 1 Introduction

The primary theoretical motivation for this disserta-
tion was the desire to understand in what respect plasmas
created by the passage of fast charged particles resembled
(and how they differed from) electrically produced plasmas.
During the early stages of the research, which eventually
led to a nuclear pumped laser, it was widely believed that
plasmas' created by ionizing radiation would differ in
fundamental respects from electrical plasmas. As more
insight was gained, however, it was realized that the bulk
of the excitation and ionization was cavised by the shower of
fast secondary electrons rather than by the fission fragment
itself.^ Thus the excitation process was seen to be elec-
tron impact excitation with the following difference: the
electron energy distribution was nonmaxwellian , with peaks
at high energies corresponding to the "injection" energy of
the secondary electrons.'' Since, in the nuclear plasma
case, the electrons are always slowing down from their
initial birth energy, while, in an electrical plasma, the
electrons are being constantly accelerated by the electric
field, we may expect some differences in the excited state
populations.

7

The basic operational problem was that of finding
relations between the obser\'able quantity N(t) , the number
of photons which were emitted at time "t" after fission,
and the properties of fission fragment excited gases in
general or atomic and molecular properties of particular
gases. We shall first describe .the interaction of the
fission fragment with the gas in a general way, and then,
these relations will be derived by starting with the
elementary rate equations for pure nitrogen. The parameters
most amenable to analysis were t, the 1/e decay time; ^^.-j^gg , '
the 1/e rise time; ZN-^, the sum of all photons observed
independent of time; and N^^gj,, the value of the peak in-
tensity (see Fig. 1) .
2 . 2 Fission Fragment — Gas Interactions

Californium was chosen as a fission fragment source
since it underwent spontaneous fission and, thus, elimi-
nated the need for a neutron supply as is the case with
uranium. When a spontaneous fission occur];ed, the first
particles which entered the gas were prompt gamma rays.
These ganimas were emitted within 10 seconds of fission,
but caused little excitation of the gas in the experimental
chamber. They could, however, cause a small ionization
pulse in the residual gas within the photomul tiplier . This
pulse was amplified by the dynodes and recorded by the
counting system. The gamma pulse occurred well before the
peak of the light pulse, and caused no increase in back-
ground during the decay period, (see Fig. 2). At the time of

3

t-

o

S-

E

u

o

c

4->

' —

c

o

(^

c

C\J to

oo

o

r r--

n

^

(U

>

u

U

4J

u

(D
Cj-,

(3LEDS jeauLL) SiMROO JO yjaNflfJ

(U

E

•1 —

■i-J

■

+->

o

c

o

3

•—

o
o

S-

T

S-

!-

o

C

b

OJ

4->

CM

(T>

^

-^

o

LD

O^

cu

o

CO

CT.

z

LO

LD

a>

;8[e3S jpauLL) SiNOOO JO yBaWON

10

the gamma burst, the fission fragments began to enter the
gas. As the fission fragments passed through the gas, they
struck atoms of the gas mixture causing excitation and
ionization. The resulting excited states and ion pairs
were not all produced in interactions with the fission
fragment. The bulk of the excitation and ionization was
produced by the secondary electrons ejected during the
ionization.

Once the fission fragment had stopped, either by giving
up its energy or through hitting the wall, no more energy
was transferred to the gas, and relaxation proceeded (see
Appendix 5) . Though no more energy was entering the gas as a
whole, the populations of individual levels may still have
continued increasing. This continued increase may have
been due to dissociative recombination, cascading, reso-
nance trapping, or metastable transfer. In these cases it
was difficult to say exactly when the excitation of a given
level ended, and the afterglow commenced. Thus, while the
decay rate in the afterglow was determined by the radiative
and collisional lifetimes, if excitation processes continued
after the fission fragment had passed from view, then what
was being measured was the excitation or transfer rate
rather than the decay rate.

We have now followed the energy of the fission frag-
ment in a general way from the time the fragment entered
the gas until some of its energy reached a particular
atomic or molecular level. The energy thus stored in a

11

level may leave by a variety of routes. The energy
pathways for deexcitation depend upon pressure in a com-
plicated way, because collisional effects are a function
of pressure. At very low pressures where collisions may
be ignored, the major loss of energy from a level is due
to radiative decay. As the pressure increases, collisional
self-quenching will cause the lifetime of the level to
decrease (see Figs. 3 and 4). By extrapolating the lifetime
versus pressure curve back to zero pressure, the radiative
lifetime may be determined. Fig. 5. Thus, the rate at
which the lifetime changed with pressure may be used to
determine the collisional transfer coefficient for the
observed level. This discussion of deexcitation assumes
that no impurities are present. With impurities, the
kinetic behavior becomes a great deal more complicated.
Impurity quenching may cause the lifetime of a level to
decrease from what it was in the pure gas case, while the
effects of transfer from an impurity with a resonant
metastable level can cause a dramatic increase (see Fig. 6).
If two different deexcitation paths exist having quite
different time scales, the decay curve may show two decay
times, one for short times and one for long times (see Fig. 7).

With the exception of a small amount of work on CF^ ,
the bulk of this study was carried out on the gases
helium, neon, argon, and nitrogen. Partial energy level
diagrams for these gases are shown in Figure 8. This
figure illustrates the energetic resonance between the

12

800 torr

3 3 7 . 1 n m

Figure 3.

TIME (ns)
Pressure Variation of Decay Time

13

I

00

CM
O

I

o

tXJ

II

CVI

LD

c

ro
!-l

o

0)
S-i

in

0)
>

e

-rH

>H

ro
U

0)
Q

0

(su) 3WIi A\/33a

14

C3
O
CO

C5
O

UJ
D-

o
o

r-
ro
n

O

IH
0)

W
U)
0)

Oj

U5

W

S-i

>

0)

•H

tH

u
dJ
p

(U

u

OJ

>

o

in

o

CM

Q)

u

Ti

XJ^

■H

lu

s u 1. / L

15

CTl

1^

o
o

CM

c

•H

M
(1)

c

(0
4-)
W

4-)

OJ
5^

OJ

u

en

(9LB3S JBauLL) SiNnOD JO yjawpN

L6

o .—

C\J 4->

~\ O CO

O

a)

>

^-^

^-1

LO

3

T3

u

C

o

>i

u

en

OJ

U

in

0)

o

Q

c

ro

Q)

(~)

c

^

o

— -

,U

o

Q

0)

3
tn

■H

&4

C5

Ln

(9LPDS J5auL[) SiNnOD dO b^aWflN'

17

2
C

0)
2

O
in
0)

>

0)

u

GJ

c

u

c
>

0)
.—I
Q)

o'

-r

o

OvJ

Q. Q.

cn m

ro

n

:3

<u

.T

— n

CNJ CM

I I I I

1 — I — I — r

0)
■H

T1 — I — r

in

CM

I I I — r

o

CM

(A3) A9y3N3

2 + +

B Zu level of N2 and the metastable levels of helium and

also the argon metastable, C^du resonance. These reso-
nances lead to intense first negative and second positive
emission respectively. By observing mixtures of the three
inert gasses with nitrogen, we were thus able to examine
the cases of neutral excitation (A/N2) , ion creation and
excitation (He/N2) , and no interaction (Ne/N2) .
2 . 3 Rate Equation for C^IIu

In this chapter we will confine ourselves to the
relatively simple case of the C-^IIu level of N2 with no
foreign gas present. For this restricted case, there was
no cascading®, metastable transfer or resonance trapping,
and thus the rate equation is very simple, consisting of
two deexcitation terms, collisional and radiative decay,
and one excitation term, collisional excitation from the
ground state. The collisional excitation "turns off" at
tf, the time at which the fission fragment left the ob-
servable region or had lost so much energy that it could
no longer produce excited states. The rate equation thus
separated in a very natural way into two parts, one before
tf, and one after. ^

dN2 dE^

"dT ^ Q"dr - ^1^2 - 1^2PN2 0<t<tf, (1)

*
dN-, * *

dt - ^-

where, N2 is the population of the C-^lIu level, Q is the

number of C^Ilu states produced per unit energy deposited

in the gas, E^^ is the energy, deposited by the fission

19

fragment, P is the pressure, t^ is the time at which the
fission fragment collides with the wall or has slowed to
the point where it can no longer cause excitation and ioni-
zation. k-[^ and ^2 ^^^ the rate coefficients for the
deexcitation reactions shown below:

N2 (C^nu) -i N2 (B^ u) + hv

v-^ observed states.
N2 (C-^nu) + N2(X-'-Eg+) ->

k-]_ is the sum of all transition probabilities for the C-^Ilu
state,

^1 = ^UL = 7 sec-1,
where, Tj. is the radiative lifetime of tlie state, and k2 is
the two-body collisional quenching rate with units (sec"-'-
torr"^) .

After t^, the population decreases monotonically , and
Equation 2 may be solved to give

Solving the rate equations for 0£t<tf requires a functional
form for the energy deposition of the fission fragment.
While much has been written about the form of this func-
tion °, for nitrogen the energy deposition is adequately
described by the square law for electronic stopping

E = Egd-r/Re)^, (See Appendix 1) (4)

which, in terms of velocity is

V = Vgd-r/Rg) , . (5)

where E and v are the energy and velocity of the fission ■

fragment at the distance r from the source. Eg and vg are

20
the initial energy and velocity of an average fission
fragment. Rg is the extrapc^lated electronic range, Rg =

^^7-, where S is the range in mg/cm ; R is the gas constant;
PM

T the absolute temperature; P the gas pressure; and M the

molecular weight of the gas. Integrating Equation 5 over

the flight time and track length of the fission fragment

gives

tf = -^ and-—) . (6)

Vq Re

Solving Equations 3 and 6 together gives

?^ = ZEoe-^t ,
where Z = 2 Vq/Rq- 'Thus, for 0£t<tf

";(t) = (r^) [e-^^-e-^^l^^2^^h'. (7)

2. 4 Relating Solutions to Observables

■k

Since the intensity of I(t) equals k]^N2{t), determin-
ing the decay time of the intensity in the period tf£t<™

g

ives the excited state lifetime directly. Since Equations

_ * _ ZEoQk^
3 and 7 must be eq.ual at tf, I^ax ~ ''-1^2max ~ k +koP-Z

{g- (ki+k2P-Z) tf_-j^} _ As T goes to infinity, N2,^3j^ is just

the total number of excited states produced by the passage
of the fission fragment

NLax(^-") = EoQ(l-e'^^f).
Integrating k]^N2(t) from 0 to °° gives the observable inte-
grated intensity

* -Ztf

I = '(T '^iN2(t)dt - k^EgQi (1-e ).

21

Physically, N2ni-,x ^^ ji^^st the maximuin excited state
density that is obtained, ai'd I is that fraction of the
excited states which decay by radiation. Note that

2max

I/t = 19^.. (T^"")

CHAPTER 3
SYSTEM AND CALIBRATION

3 . 1 Introduction and Design Criteria

The design of the experimental apparatus was deter-
mined in large part by the nature of the isotope source.
The source emitted about 3000 fission fragments each
second into the gas, and each fission fragment produced
about 10-10 photons. Because of the very small solid
angle subtended by the monochromator slit, it was ex-
pected that, on the average, one photon would be counted
for each 1000 fission fragments. This extremely low
light level demanded that single photon counting techniques
be employed. Since thermally emitted electrons or dark
counts could not be distinguished from single photons,
the photomul tiplier was cooled to decrease random back-
ground caused by thermionic emission of electrons from the
photocathode . The key to the noise reduction and the
advantage of this system ove'r other single photon counting
spectroscopy systems was the use of two photomultipliers
in a delayed coincidence circuit which virtually eliminated
the possibility of counting a random noise signal generated
in only one of the tubes. The low flux level also required
the ability to maintain suitable gas purity for the ex-
tended times necessary to establish acceptable statistics

22

23
(see Appendix 4) .

The experimental apparatus (see Fig. 9) may be conveni-
ently divided into three sections:

The experimental chamber,

The vacuum and gas handling system, and

The detection system.
The operation and calibration of each component section will
first be described, followed by a discussion of system
characteristics.
3 . 2 System Description

The experimental chamber was a 3.8cm i.d. double-sided
Varian high vacuum flange. The side toward the total lumi-
nosity photomultiplier (TLP) had a pyrex window which was
optically coupled to the TLP. The side facing the monochro-
mator had a sapphaire window within 2mm of the entrance slit
of the monochromator . The flange had two ports, one which
connected with the gas handling and vacuum system, and the

other through which the calibration source or high voltage

2 5 2

electrode could be inserted. The Cf source was cemented

to the inside of the flange in such a way that a fission
fragment emitted normal to the surface traveled along the
long axis of the monochromator slit. In this way the light
input to the monochromator was maximized.

The vacuum system was entirely stainless steel with
copper gaskets. A combination of roughing, oil diffusion,

and ion pumps were used to attain an ultimate vacuum of

-9
5x10 mm Hg before filling v\'ith a test gas. Tne system was

24

TO GAS HANDLING AND
VACUUM SYSTEM

SLP

DELAY

MONOCHRO
MATOR

PM

PLASTIC
SCINT."
FOR GAMI'lAS

TIME TO
PULSE HEIGHT

CONVERTER (TAC)

MULTICHANNEL
ANALYZER (fICA)

DISCRIMINATOR

Ficjure 9. System for Lifetime MeasurcmonL

25

then filled with research grade gas and the pressure
measured to within 1 torr with a Wallace and Tiernan gauge,
Ultek high vacuum valves were used to completely isolate
the system during the period of counting, which, in some
cases, took as long as 72 hours.

The detection system may be described in terms of
its four component functions :

a. Determine when a fission fragment has passed
through and excited the gas (the total lumi-
nosity photomul tiplier , TLP ) .

b. Determine when an excited atom or molecule
has decayed by the emission of a photon

(the specific luminosity photomul tiplier ,
SLP) .

c. Measure the time interval between the exci-
tation and decay (the time to amplitude
converter, TAG) .

d. Digitize, accumulate and store measurements
for successive fission fragments (the multi-
channel analyzer, MCA) .

a. When the fission fragment passed through the gas, a
large burst of photons was produced. These photons were
of all wavelengths and were spread in time between the
fission and the final decay of the longest lived of the
atomic and molecular states which, in some cases, could
involve times as long as 80 microseconds. When these
photons struck the photocathode of the TLP, they produced

26

a current pulse which triggered a-built-in discriminator.
The discriminator, which used zero crossover timing,
generated a fast negative logic pulse when half the total
charge had been collected. To use this signal for timing,
it was necessary to know how it was related in time to the
fission event. This time relationship was determined in
the following manner. Within 10 seconds of fission, a
burst of prompt gamma rays occurred and was detected by a
fast plastic scintillator mounted on a photomul tiplier
placed underneath the experimental chamber. The time
interval between the gamma burst or fission and the peak
of the total luminosity pulse was thus determined (see
Appendix 2). In addition, the lifetime of an excited state
was measured using first the gamma detector and then the
TLP to start the timing (see Appendix 3). The measured life-
times in both cases were found to be identical. Thus, the
output of the TLP was related to the time of fission and,
hence, to the time of excitation by a fixed and known time
lag .

b. The photons from the gas passed through a sapphire
window and into the entrance slit of the monochromator ,
Because of the close proximity of the monochromator slit
to the source, no system of lenses would have increased the
Ij.ght input. The monochromator was a Model 218, 0.3 meter
focal length Czerney-Turner , manufactured by the McPherson
Instrument Corporation, with the grating blazed for 300nra.
Mounted directly behind the exit slit was an RCA 8575

27 • . -

photomultiplier contained wj thin a thermo-electrically
cooled housing. The system was calibrated for relative
wavelength sensitivity using an .Eppley Tungsten filament
lamp certified by NBS. No absolute calibration was at-
tempted. The output of the anode was dropped across a
50 ohm load resistor to give fast time response and the
resulting pulse sent to an Ortec Model 260 time pickoff
unit. The time pickoff unit consisted of a pulse trans-
former, a wide-band transitor amplifier, a tunnel diode
discriminator, and a line drive buffer. The unit acted
essentially as a fast, low level timing discriminator.
When an input pulse exceeded the discriminator setting, a
NIM standard fast negative logic pulse was generated.
c. The time-to-amplitude converter measured the time
interval between the leading edges of logic pulses fur-
nished to its start and stop inputs and generated an analog
output pulse that was proportional to the measured time
interval. The start pulse initiated the charging of a
highly linear capacitor, and the stop pulse terminated the
charging. The voltage across the capacitor was thus pro-
portional to the time interval between the tvv'o signals and
served as the output. Since the TAC was triggered by the
leading edge of the input pulse, it was necessary for good
time resolution that the pulses have very fast rise times
and that they exhibit a small dynamic range. These criteria
were statisfied by using NIM-standard fast negative logic
pulses. The start pulse was supplied by a zero crossover

28

discriminator built into the base of the TLP, an Ortec
Model 264 photomultiplier timing discriminator and pre-
amplifier. This unit also provided a linear or analog
output v-;hich was used in conjuction with an Ortec Model 410
linear amplifier to set the discriminator level above the
alpha particle peaks so that the system responded only to
the fission fragments. Since both the TLP and the SLP were
responding to photons produced at about the same time, the
pulses from the SLP were delayed by a nanosecond delay unit
to allow time for the zero crossover discriminator in the
TLP to determine the time of fission. The arrival of the
start pulse opened a time window whose width could be set
from 50 to 80,000 nanoseconds. During this time window,
a photon hitting the photocathode could stop the TAG,
thereby producing an output pulse. If no photon was re-
ceived before the end of the time window, the TAG automati-
cally reset and was then ready to receive another start
pulse. The fraction of start pulses which were followed
by a stop pulse was thus related to the number of photons
present in the experimental chamber. The detailed behavior
of the number of true stops as a function of time accurately
reflected the time decay of the photon density and, thus,
the population of the excited level.

d. The output of the TAG was sent to a TMG Gorporation
multichannel analyser (MGA) . Using a Model 210 pulse
height logic unit, the height of each output pulse from the
TAG was measured and sent to a channel whose address cor-

29

responded to the measured pulse height and, thus, to the
measured time-after-fission. Counts were accumulated for
a period of time determined by the arrival rate of photons
at the SLP. The fewer photons per second being counted,
the longer the measurement had to be made in order to ac-
cumulate adequate statistics. The period of counting could
be varied from five minutes up to 72 hours, the upper limit
being set by the necessity for maintaining adequate gas
purity (see AppendLx 4). After measurement was completed, the
contents of the MCA memory was read out to an X-Y plotter
for qualitative analysis and the digital values punched on
paper tape for subsequent computer analysis.
3 . 3 Calibration

To determine the minimum decay time that this system
could measure, a small chip of an "ultra fast" plastic
scintillator (NE-lll ) was suspended in the middle of the
excitation region (see Appendix 2). The measured decay time
of 1.67±.06ns showed no broadening compared with the
published value of 1.7 ns . Since no scintillator could be
found with a shorter decay time, the minimum that this
system could measure is still unknov.-n. The TAC- was
capable of resolving signals lOps apart. The system was
ultimately limited by the photomultipliers which have rise
times on the order of Ins.

The minimum resolution in a real experiment is limited
generally by the statistical accuracy of the number of
counts which can be measured in the time over which condi-

30

tions may be expected to rei^ain unchanged. To determine
the variation due to counting statistics and errors associ-
ated with filling the system to a chosen pressure, decay
measurements at two different pressures were repeated ten
times each. The standard deviation of what should be
identical measurements was 0.138ns at 800 torr and 0.324ns
at 100 torr, or about 4% in each case.

Identification and elimination of the gamma background
proved to be a major experimental difficulty. In our pre-
vious experiments in the reactor, signal noise caused by
gamma rays striking the photomul tiplier had required
extensive shielding, a massive low-pass filter, and extreme-
ly slow scanning speeds. The original reason for consider-
ing a coincidence system was the freedom it seemed to offer
from gamma background. Since the gammas are discrete
particles, and thus capable of striking only one photomul-
tiplier at a time, requiring coincidence between the outputs
of the photomultipliers should have eliminated the possibil-
ity of responding to gamma noise. Examination of early
decay curves, however, showed .a peak occurring shortly after
fission which was finally traced to a coincidence between
the fission fragment signal from the TLP and a gamma noise
pulse in the SLP . Because the prompt gammas, unlike decay
gammas, do not occur at random but are emitted within 10
seconds of fission, coincidence does occur between the
fission fragment and the gammas (see Fig. 2) . These gamma
pulses were thought to be due to ionization of the residual

-31

gases within the tube and to scintillation of the quartz
window. The background level was found to be greatly in-
creased by opening the system and exposing the photomulti-
pliers to light, even with no voltage applied. After such
an exposure, about six hours was required before the back-
ground level approached steady state. Extensive tests
were made to determine how . the signal to noise ratio
changed with photomultiplier voltage. The signal to noise
ratio was found to increase monotonically as the high
voltage was reduced. If the voltage was reduced too
much, however, the fission fragment could not be distin-
guished from the alpha particles. The optimum situation
was when the voltage of the two photomultipliers was the
same and equal, to 2200 volts.
3 . 4 Output

After a lifetime measurement had been completed, the
data was in the form of integer values between zero and
2 located in the memory of the MCA. Each channel cor-
responded to a time interval determined by the full scale
setting of the TAC . For example, system calibration showed
that when the full scale setting of the TAC was 50ns and
64 channels of the MCA were being addressed, 10ns covered
a range of 12 channels. Thus, each channel corresponded
to 0.83±.04ns. The channel by channel accumulated count
was read out along with the corresponding channel number
either in analog form onto an X-Y plotter, in decimal form
on adding machine tape, or punched in binary coded hexi-
decimal on paper tape. In practice, the data was always

32

displayed in analog X-Y forn^ for quick analysis. If the

data appeared to be of interest, it was punched on paper

tape and the paper tape later read into the IBM 1800

computer and stored in a file on a magnetic disk. This

data was then available for immediate analysis and future

reanalysis.

3 . 5 Scanning Mode

The system was also capable of being used as a very
sensitive scanning spectroscopic system (see Fig. 10). To
scan, the system was set to respond to all photons of the
correct wavelength regardless of the time of occurrence
and then to step the spectrograph and a counter simultaneous-
ly so that the number of photons at each wavelength was
counted. In this mode of operation, the time window of
the TAC' was set wide enough to include the entire decay .
curve of the excited level. Any photon of the correct
wavelength that was associated with a fission fragment
thus gave rise to an output pulse. Using the internal
single channel analyzer (SCA) , this output pulse was con-
verted into a NIM standard 5-volt slow logic pulse and
supplied to the input of a multiscaler logic unit in the
MCA. The multiscaler unit addressed each channel of the
MCA sequentially and stepped when it received a trigger
pulse. This trigger pulse was supplied by a recycling
timer which would supply trigger pulses with any chosen
interval. This trigger pulse was also supplied to a
stepping motor connected- to the wavelength drive of the

33

337.1 ^2

U

Dl

357.6 N2

WAVELENGTH
Figure ,10. Scan of Pure N2

34

r

spectrograph and served to step the spectrograph and multi-
sealer simultaneously. Because of the wide slits on the
spectrograph, it was found to be sufficient to sample the
intensity every 0.2nm corresponding to 46 steps of the
stepping motor for each trigger pulse. The length of time
counted at each wavelength varied with the intensity of the
light. When scanning a gas, such as aii argon-nitrogen mixture
which is very bright, 10 seconds per channel proved suffi-
cient while for some poor emitters, several hours at each
channel were necessary.

An interesting feature of this mode of operation was
the ability to make a scan at any point in time. That is,
by setting the TAG to respond only to photons which occurred
between, say, 10 and 18ns after fission, we had a spectro-
graph with an 8ns shutter which opened 10ns after fission
(see Fig. 35). A spectral scan at several times after fission
could provide a complete time history of the intensity at
all wavelengths.

CHAPTER 4 ■ •

ANALYSIS AND RESULTS

4 . 1 Introduction

The basic question addressed in this chapter is that
of relating the raw data to the fundamental properties of
the fission fragment excited gas. }io\<r the system charac-
teristics affected the raw data was discussed in Chapter 3.
Here we shall assume that the value of N(t) has been cor-
rected for instrumental influence and try to determine how
these values relate to the radiative and collisional decay
coefficients. Vie shall first discuss how the computer
programs processed the raw data and then proceed to a
detailed discussion of the four output parameters t, N(t),
SUMN^, Nj^gj^, and their interpretations (see Fig. 1) . Finally,
some consideration will be given to general problems re-
lating to any measurement carried out with this system.
4 . 2 Programs

After an experiment had been performed and a count
completed, the data was stored in the memory of the MCA.
This data was immediately punched in hexadecimal form onto
paper tape and then read into an IBM 1800 computer and
placed in a file on a magnetic disk. All the data and the
current working programs were kept on this disk and, to-
gether with the lab book, provided a complete record of the

35

36
experiment. Once the data v.'as on the disk, there were two
programs which could be called upon for analysis: SUMNj_
which printed the sum total of counts between any chosen
limits thereby giving a number proportional to the number
of excited states which decayed by radiation in the time
interval specified, and SLOGP.

SLOGP first plotted logN^, the log of the number of
counts, versus the channel number i. Since the decay curve
was expected to be a simple exponential decay, the SLOGP
plot was visually examined to determine if there was a
straight line region and, if so, what channels comprised it.
In addition, this plot allowed a good determination of just
where background occurred. Once the limits of the decay
region had been determined, the program performed a least
squares fit of the experimental data to an equation of the
form Yj_ = Ae'^-i/^+B in the region chosen from the log plot
and specified by the operator. The operator supplied an
initial value of background chosen to be just below the
value obtained from the log plot. An iterative calculation
was then performed for the number of iterations and with
the iteration interval specified by the operator. For each
value of the background, a least squares fit was made and
the value of the error function calculated. After the
iteration was complete, the value of background for which
the error function was a minimum was chosen, and both the
raw data and the chosen fit were plotted along with the
numerical values for A, t, B, and the error function. If

37

the fit appeared to be flawed, new values of starting chan-
nel, stopping channel and background could be chosen and a
new curve superimposed upon the old for comparison.
4 . 3 Analysis

Once the data had been analyzed by the programs, there
were four numbers available:

1. T, the time it took for the excited state
population to decrease by a factor of 1/e.

2. N(t) , proportional to the number of excited
states decaying at time t.

3. SUMN- , proportional to the number of ex-
cited states which decayed between the time
A and the time B.

4. Nj^^j^, proportional to the greatest emission
rate.

The decay time determined by the least squares fit of
the raw data was the sum of tlie radiative and collisional decay
times. Since the radiative decay time is an atomic con-
stant, while the collisional decay time changes with
pressure, it is possible to determine the two effects
separately. As the pressure decreased, collisions became
less frequent, and the collisional decay rate decreased
until at zero pressure, the decay of the excited states was
entirely determined by the radiative decay.

The decay time x is equal to 1/ (k-,^4-k2P) , 'so that by
plotting 1/t versus pressure, the data should fit a straight
line with an interce'pt ki and a slope k2- Deviation from a

38
straight line, indicat-ing nonlinear behavior, v;as ob-
served at the highest pressure studied (800 torr) . This
deviation indicated that three-body collisions were be-
coming important and any studies carried out at this
pressure or greater should include a quadratic pressure
term in the rate equation. Figure 5 shows the results for
the C IIu level of pure nitrogen with the determined values
of ki and k2 shown. This curve represents the analysis of
82 decay time measurements.

There were three sources of difficulty which may cause
errors in the interpretation of the lifetime curves. These
three problems all took the form of additional excitation
terms in the rate equation other than direct excitation,
some of which persisted sufficiently late in time that
there is no true "decay" period.

Resonance trapping will cause the population of a
level connected to the ground state by an optically allowed
transition to be maintained at an artificially high level
compared with the optically thin case. The two most
satisfying solutions to this problem are either to solve
the equation of radiative transfer simultaneously with the
rate equations or to choose a level for which there is no
radiation trapping.

Cascading causes the population of an intermediate
level to continue to be augmented by the. decay of higher
lying levels for some time after the end of the excitation
pulse. This problem may be solved by either solving the

39
coupled sets of rate equations for all the levels in the
gas or by choosing a level vhich has no significant cascade
contributions.

Collisional transfer between excited levels in a
decaying gas causes an irreversible flow of excitation
energy toward the lowest excited state within a closely
spaced group of levels. The collisional transfer rate, and
hence, the collisional quenching coefficient, is explicitly
solved for in this dissertation.

It should be noted that one source of uncertainty
usually present in plasma spectroscopic diagnostics was
avoided here. No assumptions about equilibrium or depar-
tures therefrom were necessary. The measurement of the
decay time was independent of the relative populations of
the various levels subject to the three preceding cautions.
In fact, one could, hardly imagine a more nonequilibrium
gas either spatially or temporally than that being present-
ly considered.

" The curves of intensity versus time yielded the
temporal variation of the population. It may bo noLiHl that
this variation was in fact the solution of the rate equations
for the particular level observed. The rise time of the
population was nearly constant for all levels in all gases
at all pressures, and this surprising result is discussed
at more length in Appendix 5.

Plotting ^^SLX versus pressure (Fig. 11) we find that
the maximum excited state density occurred at about 350
torr but did not exhibit a particularly sharp maximum.

4 0

o

o

CO

C\J

X

o
o

IX)

s-
o

q:

a:

o c^

o

CD
CD
C\J

CO

UD "d-

CNJ

o

o o

o

e[9j)

AlISN3iNI >IV3d

rn
ro

!-l
O
U-l

QJ

V-l

in
in
(U

en

0)
>

-U

■H

-p

c

M

<u

(1)

■H

41

From Section 3 we see that SUMN/t plotted against pressure
should show identical behavior which, in fact, it does (see
Fig. 12) .

As shown in Section 3, the sum of the counts in all
channels can be related to the ratio of the radiative decay
constant and the collisional decay constant. Alternatively,
if k]^ and k2 are known from the decay curve, SUMN can be
related to the total number of excited states created by
the fission fragment passing through the gas, QEq ( l-e~^^f ) .
Figure 13 shows the variation of the integrated intensity
with pressure. The curve indicates a rather narrow peak
occurring at about 175 torr. Comparing Figures 11 and 13,
we may observe that a pulsed laser might operate best at
a different pressure than would a CW laser. Since, in the
pulsed case, we are interested in the greatest instantane-
ous population, we would choose the peak in Figure 11,
while for the steady state case, we are concerned with the
average population density and would follow Curve 13.
4.4 Pulse H'eight Distribution

One tremendous advantage to using the decay curve for
determining the lifetime or transition probability of an
excited state was that it didn't depend upon the number of
excited states created, but rather upon hov^7 those that were
created changed with time. What was being measured then
was not intensity, but the rate of change of intensity
which was much easier to do. The number of excited states
created by the passage of a fission fragment was propor-

12

S-l
3

m

u

O
O

CO

4-1
O

c
o

-H
-P

o
c

d

ID on

O
CD

•

5-
S-

o

a:

ZD

oo

LlJ

O Q.

o

o

CO

hJ

o

CD

O

o

siNnoo

(0

(U
6

•H

E-<

>

U
01
Q

><

Ti

■rH
>

Q

>i
-U
■H
Ul
C
0)
-U

c

T3

0)
+J

(0

t7>
Q)

-U

c

fN)

5-1

4 3

CXI

E

•2L

c

1 —

LU

.

CC

r-^

ZD

oo

Q.

n

o

0)

CO

QJ
Oi

w
d
en

o

>

o

yo

^— ^

>

S-

+j

i-

■H

o

W

-M

C

4J

LU

C

CH

1— 1

rD

CO

^

oo

(1)

UJ

JJ

cc

rtl

o

D.

M

o

rri

>*

U3
O

O

CM

o

u.

m

(U

M

3

o

tP

o

•H

CM

\m

Nuns 3Aiivi3y

44
tional to the energy deposited which was, in turn, dependent
upon the initial or birth energy of the fission fragment.
Fission fragments at birth showed a distribution of energies
divided into two rather well defined groups. .The energy
distribution of the 252Qf source used in the experiment,
measured with a surface barrier detector, is shown in Figure
14. Since some fission fragments have a greater initial
energy than others, they will, on the average, create more
excited states in passing through the gas. Thus, the
intensity we observed after the passage of a fission frag-
ment was dependent to some extent upon the particular
particle we chose to observe. Note, however, the rate at
which the excited states died away did not depend on the
energy of the fission fragment which created them. Looking
at the output of the TLP , the amplitude of the total lumi-
nosity pulse was proportional to the number of photons
resulting from the passage of the fission fragment. At
constant pressure, this was proportional to the energy of
the incident particle.

The pulse height distribution in the gas is shown in
Figure 15.' This distribution was measured in CF^ , a gas
which gi.ves a large number of photons per unit energy
deposited and, hence, provided good scintillator resolution.

45

1000

o

100

ENERGY
Figure 14. Fission Fragment Energy Spectrum

46

PULSC IILIGIIT

Figure 15. Fission Fragment Distribution from TLP Pulse

Heights

CHAPTER 5
OTHER- GASES AND GAS MIXTURES

5 . 1 General Comparison of Mixture Spectra

Figure 16 shows scans of the fission fragment excited
nitrogen impurity spectrum of one atmosphere neon, argon,
and helium alongside that of pure nitrogen. Notice that
the argon and helium were scanned five times faster than
the neon and nitrogen, and hence, their peak heights or
intensities should be increased by a factor of five for
direct comparison. That the nitrogen impurity in argon
and helium was so much brighter than either pure nitrogen
or nitrogen in neon, was a consequence of the efficient,
transfer of energy from the metastable energy traps in
argon and helium in to nitrogen. The dominant emission
from helium was the 391. 4nm N2 band, while that from argon
was in the 337. Inm N2 band, both of v;hich results follow
immediately from a consideration of Figure 8. The dominant
emission from neon was also the 391. 4nm band, and this has
not been explained. The relative peak heights adjusted for
the different scan speeds of the most intense emission
from each gas goes as neon to argon to helium to nitrogen,

2.1 : 1.3 : 13.7 : 1 .

5 . 2 Neon/Nitrogen Mixtures

Besides the measurementof the decay time of the C Ru

47

48

en

tr>

0)
D

C

<

0)
2

U)
0)
•H
-U
•H
i-l

a

E

o

U-l

o

to
c
to
u

CO

en

(a[PDS jeauLL) AiISN3ifJI 3AIiV13y

4 9
state of -nitrogen, several other studies were carried out
to illustrate the versatility of the system for the study
of fission fragment excited gases. An extensive study was
made of neon and neon-nitrogen mixtures. Using this mix-
ture provided the opportunity of viewing, simultaneously,
three lines from three different species, atomic neon,
molecular nitrogen, and the nitrogen molecular ion. Figure
17 is a spectral scan of fission fragment excited neon
showing the nitrogen impurity emission. Looking first at
the 585. 2nm neon line in pure neon, we find a radiative
lifetime of 16.6±.7nm (see Fig. 18). As the nitrogen impurity
concentration increased from 1% to lOo to 50o (Figs. 19, 20,
21), there was^ no significant change in the lifetime of the
neon excited state. This might be expected since there
was a very poor energy match between the levels of neon
and those of nitrogen. The values for k2/ the collisional
transfer coefficient, appeared to show a systematic de-
crease with nitrogen concentration. If this trend is real,
it means that the nitrogen is less effective than the neon
in quenching the level that gives rise to the 585. 2nm line.
Looking next at the C-^llu level of nitrogen witli neon con-
centration values of 1% , lO"!, 90", and 99" (Figs. 22, 23, 24,
25) , we see that when the nitrogen was the dominant gas,
the lifetime varied strongly with pressure; whereas, when
the neon was dominant, there wa's very little change. This
again tells us that there was not much interaction between
the levels of the two gases.

50

O)

QJ

c

c

rt3

s-

_C

S-

O

o

-~^

-t-J

u

OJ

o

tn

UD

O

r~^

LT)

o

OJ

OJ
S-l

c:

•H
V-l

:3
a,
e

0)

en
o

4-1

M-l

o

c
u

(XI

(D

(3[\?os jeauL[) AiISN3iNI BAIiVHBy

51

S-

0

-M

■—

(/I

--

1
0

l/l

CO

1—

c

c:

X

UD

1X>

UD

O)

E

0

.

UD

2:

CO

•

CXI

'—

UJ

.

II

M

II

q;

LD

=1

00

1 —

CO

Q.

Ln

^^

-V

H

o
o
CO

6

a

r>j

•

0,

LO

0

00

<o

in

J-i
0

M-l

QJ
U

:3

0)

, — ^

s^

!-

a.

S-

0

U)

+->

3

^ ^

w

0

s^

0

UJ

0)

q;

>

^)

00

0)

tn

e

LlJ

•H

d;

H

Cl

0
0)
Q

o

CI
CO

0)

o

CO

to

(su) 3WIi AVD3a •

52

s-

s-

0

-M

•—

00

c

•—

--

^

o

cyi

0

I/)

o

c

1 —

c

' —

00

X

CO

E

0

•

r^

c

•

OvJ

. —

CNJ

CM

2:

LT)

II

II

II

CD

CO

- 00

^

tn

-bii

J^

H

I J*"-^!

0

0

1-0

c

in

CO

U

-— ,

0

S-

m

s-

0

Q)

+->

rn

0

UJ

U)

«^

QL

(U

:z)

M

oo

Oj

00

UJ

U5

en

3

Q_

>

I L

o
o

CNJ

>i

(ti
u

(U
Q

C7^

S-l

•rH

a.

cc

(su) 3WIi AVDBQ

53

s-

S-

o

4->

•—

tn

c

■—

1

VI

o

C/1

c

r —

c

1.0

X

CO

LD

p^

•

E

O

r-^

c

, —

r—

CNJ

II

II

II

un

CO

■ CM

s

LD

-^

J^

H

o

CO

(su) 3wii AVDBQ

54

^

CD
O
00

s-
s-
o
+->

c

CM OsJ

■^-> LD
OJ CO

— 1/1

I C

1/1 ^

C I

CO o

Ln I —

o X

• CO

1/1
c:
c\j

t —

o

CD

. — CM

S-

e

G

CM

in

CO

Ln

1-1
O

o

o
o

0)
S-l

00

0)

>

(1)
e

o
o

CO

(0
U

Q

CM

u

■rH

CO

us

(su) 3WIi AV33a

55

-•—I -I

u.

S-

o

+->

■—

(/)

c:

-—

—

LD

C3

00

o

l/l

O

c:

, —

c

■~

1^

CNJ

X

U3

n

E

O

t^

c

■

, —

ro

CO

, —

-2L

r^

II

II

II

QJ

m

C\J

■21

m

J^

J^

H

o
CO

I— • — I

I — •-

»— •-

I » I

o

o

F

UJ

C

ro

M
O

U-l

^ — ,

u

s-

3

s_

U5

o

U)

-1_>

(U

— ^

u

o

Cu

o

1 1 1

"*

q:

m

^)

3

I/)

m

CyO

S-l

UJ

ai

Qi

>

Cl_

e
>1

(0
O
0)
D

o

C2

,

CM

CM
CM

Q)

iX>

^

CM

o

CO

lO

^

CM

ro

ro

ro

CO

CM

OJ

CM

CM

(su) 3i.|ii AV33a

56

o
o
CO

o

CD
IX)

(/)

Ln

c

c:

r~~

<*

C\J

E

o

i-D

c

•

o

OO

,

II

II

II

r^

oo

, —

CM

S-

ro

-^

-:^

H

m
m

S-l

o

o 3

+-> U)

— in

8 QJ

en

Q-

en

0)
>

e

•rH

Eh

o

o

CO

O
0)
Q

S-l

:3
en

•H

Cm

CD

CD

CD
CVI

O

(su) 3i,jii AV33a

5 7

'^

on

o

1/1

c

r—

c

oo

X

LO

-

C\J

O-i

B

o

•

Ln

c

•

1 —

■*

F

•

II

II

II

r-^

oo

CM

s.

n

J^

J^

H

o
o
cx)

o

o

lO

e
c

t— 1

•

n
n

O

u-i

i_

CIJ

s-

i-i

o

D

■t->

cn

—

0)

CJ

UJ

i-l

Oi

Ul

CO

Wl

=5

UJ

cn

ai

!m

Q_

0)

e

. EH

>i

U

n

CJ

o

<XI

en

fa

o

LO

o

o

O
CM

(su) 3WIi A\/33a

58

o
o
CO

U-

s-

0

+->

■—

1/1

c

o

, —

^

o

1

1

, —

C/l

0

00

~~^

c

f —

c

•-

CSJ

X

10

E

0

•

l~^

c

•

m

^

CXI

, —

^

r^

II

II

II

(V

ro

, —

C\J

2:

m

■^

-iii

t-

J-i
O

OJ

in

Q)

u
Cm

in

!-l

>

(D
E
•H

EH

>^

U
QJ
Q

in

(U

•H

o

o

CD
CO

o

CSJ

(su) 3WIi A\/33a

59

When we turn to the analysis of the 391. 4nm and due
to the N2 molecular ion, which has been reported to have a
radiative lifetime of about 65ns, some very strange results
are observed (see Figs. 26, 27 ,-28). In pure neon, the first
negative band due to nitrogen impurities within the gas
had a lifetime of 90 , 000+ 3500ns . As the nitrogen concen-
trations increased to 10% and 50%, the lifetime dropped to
10,000±400 and finally, to 3800±150ns. Anomalous excitation
of N2 has been seen before in laser produced air plasmas,'^
argon plasma jets,^^ radio frequency plasma torches,''^ and
manganese seeded plasma jets,''* but has received no adequate
explanation. '
5 . 3 Helium/Nitrogen Mixtures

Next, let us consider helium-nitrogen mixtures. For
pure helium, the emission was dominated, just as in the case
of electrical excitation, by the first negative bands due to
the N2 molecular ion (see Fig. 29). As the nitrogen concentra-
tion increased to 1% (Fig. 30) , the dominant band became the
second positive system, but the first negative was still
visible. This behavior is to be compared with the case of
electrical excitation in glow discharge where the 391.4
emission dominated to much higher concentrations of N . The
persistence of the N^ emission may be due to the electric
field inhibiting recombination and, hence, keeping the
population of N^"*" elevated with respect to the fission
fragment excited case. The intensity variation of 391. 4nm
with pressure for several mix ratios is shown in Figure 31.

60

o
o

C30

S-

S-

o

-(->

--

-

' —

KTi

<J\

c

c

Ln

r^

Kn

+

1

1

c

(XI

o

o

o

■^

, —

1 —

o

X

X

o

, —

CO

«\

0)

H

.

o

z

c

, —

C\J

cr>

^

LU

.

M

II

II

a:

, —

:z>

CTs

- CO

s-

Q_

ro

:^

-^

p

CD
O

o

o
o

Qi

UJ

m
O

0)
J-(

tn

0)
s-i

CM

in

tn

u

0)

>

6

■H
E-i

>i

(0
O
QJ
Q

o
o

CM

fNl

0)

3

o

O

o

C3

o
o

o

CM

(su) 3WIi AVD3a

tj

61

CO

o

<D

1

S-

0

4J

1

1

1/1

c

uo

1X1

I/)

c

1

c

+

<d-

0

0

c^

1

, —

0

2:

0

X

0

, —

00

-

E

X

•

0

c

, —

iX)

1 —

•=d-

II

II

II

. —

CT>

r—

- C\J

s-

ro

.^

_^

H

o

o

IX)

i-

o
+->

o
o

CO
t>0

o

CD
CM

(Ti

m

!h
O
14-1

0)

>-l
cn

0)
S-l
Oi

D

u

>

s

•H
EH

>i
to
U

OJ
Q

CNl

0)
u

D

•rH

h4

o

O
LD
OvJ

o

O
CD
CSJ

o
o
in

o
o

C3

o
o
in

(su) 3WIi AV33a

62

o
o
CO

S-

o

^

U3

— h

o

O

to

■-- C\J

, —

, —

c

— -21

X

X

o

ix)

o

LD

E

.

CO

c

CM

CNJ

CO

CNJ "^

z

n

II

II

^^^ , —

OJ a^

oo

S-

z: ro

:^

-^

H

o

CD

V-l

O

U-l

0)
S-i

en

0)

a.
in

D
in

S-l

>

(1)

u

Q)
Q

00
CNJ

0)
U

tn

•H

(su) iwii AV33a

63

CO

es

C

•H

u

C

GJ
tr
O

■H

O

C

u

(M

0)

(aLPDS .;eauLL) AiISN3iNI 3AIiV13y

64

o
o

o

i3LP3S jeauL[) AiISN3ifJl 3AIiV13y

65

QJ

S-l

LD

3

m

CO

■H

'

tM

AiISN3iNI 3AIiV13y

66

At 10% nitrogen, all observable emission was due to the
neutral molecule (see Fig. ;-2).

Turning to the decay curve with 1/2% N2 in helium,
we see that the difference between 800 torr and 100 torr
was dramatic (see Fig. 33) . There was a lot more energy com-
ing out of the gas as light, and it came out for a lot
longer time even though the energy deposited by the
fission fragment was less for a gas that was mostly helium
than for the heavier nitrogen. Figure 34 shows the change
with pressure of the decay time for various mix ratios.
As the amount of N2 decreased, the lifetimes of the N2
level was increasingly determined by the helium metastable
lifetime. As the amount of nitrogen increased ,. recombina-
tion took place much more rapidly, causing an additional
depletion of the level.

Figure 35 shows a scan where the system had been
adjusted to respond only to photons which V7ere emitted
during the first four nanoseconds after fission, while
the fission fragment was still within the field of view,
and excitation was still taking place. There appears to
be no difference between this curve and Figure 29, which
is from photons emitted at any time after fission.
5 , 4 Argon/Nitrogen

Because of the intense emission observed in the past
from fission fragment excited mixtures of argon and nitro-
gen, ^ ^ ^^ and also, because of the interest in using argon
as a cover gas in the gaseous core reactor, preliminary

67

o
■J-J

o

ro

es

CM

2

O

x:

•H
0)

as

o

c
.-a
u
en

rsi

Cn

■H

(91535 JBOUL[) A i I S N 3 i f J I 3 A I i \/ 1 3 ^

en
I —

in CO

o

o

Ln

a:

u

0)

C
oi

in
(U

r^-i

rn

d)

U

LT)

:3

CVI

en

■H

u,

(a[PDS ^eauL [) SlNnOD

69

CD
CD

CO

o

o

':}-

n

M

<— )

U

(XJ

M-l

S-

0)

s_

S-i

o

3

+->

W

'

LU

S-l

r— '

a:

cu

:d

oo

(fl

oo

3

LU

W

q:

V-l

D_

0)

>

O
LT)

e

•H

>i

U
QJ
Q

LT)

OJ

n
0)

Ln

■H

00

t.

o

(su) 3WIi AVDHQ

70

UJ \

3; ■!->

I— I

c
o

O

•H

0)

c
o

•H

+J
ta
-p

•H

u

X

w

0)

3:

-p

■rH
3

e

CM
2

c

U

in

(L)
U
P

fa

(0[e3S ji?oiu[) AiISrJ3ilJI 3\IiV13a

71
measurements were carried out on a 10 to 1 mixture of argon
to nitrogen, the mixture that had been found previously to
give the most light. Figures 36 and 37 show a spectral
scan of the second positive band of N2 which dominated the
visible emission. There was very little emission in the
first negative band.

Looking at the decay curves of the C-^IIu level, we
find a very instructive result. The decay curves show two
components, a fast and a slow one (see Fig. 38) . Quantitatively,
the fast decay approaches 4 5ns, the radiative lifetime
of the C nu level, while the slow decay shows a zero pres-
sure value o f 600 ns. What we are observing is the fast
initial decay of the directly excited nitrogen level and
the slower metastable transfer from the argon 4'P level
which is in resonance with it. As the pressure drops, the
collision frequency decreases, and metastable transfer is
no longer so important vis a vis direct excitation.
5 . 5 Carbon Tetraf louride

One of the peripheral interests in the optical emis-
sion of fission fragment excited gases had been the
possibility of using a gaseous scintillator as an in-core
monitor of the power level in the LMFBR. Early studies by
Pagano^^ indicated that the fluorocarbon CF4 had the dual
advantage of being inert and having high luminosity. Pagano
was not able to determine whether CF4 or some dissociation
product was causing the intense emission which" he observed.
Because of the difficulty of obtaining CF^ (or Freon-14) ,

72

<u

c

c:

fO

S-

-!=

s^

O

o

---^

4->

o

QJ

o

lyi

IJD

O

<c

I^^

. —

c
o

en
1-1
<C

0)
i-i

>1

1-1
a

0)

en
o ■
u

V

14-1

o

c

u
en

ro

O
1-1

■H
Ci-,

(ateos jeauLL) AiISN3ilJI 3AIl\/13y

73

o

o

CXI

^ o

CM

Z

m o

CiJ

x;

•H

c;
o
en

c
u

C/3

m
0)

■H

(3[PDS jeeuLL) AilSIOilJI 'BAIlVIBa

74

o

I/)

QJ

C

o
o

800 torr

k

400 torr

J I L

337. Inm N2

200 torr

100 torr

0 100 300 0 100 300 0 100 300 0 100 300

TiriE (ns)

Figure 38. Fast and Slow Decay in A/N2

7 5
only a few measurements were carried out. Spectral scans
showed a broad continuum in the ultraviolet which had been
previously observed by Walters. ^ A decay curve was made
of the C-^riu level of the nitrogen impurity at 586 torr,
and the decay was a simple exponential with a 1/e time of
2'6 . 2 1.0ns (see Fig. 39). This emission was much too weak to
account for the intense total luminosity emission.

CF^ was the only gas in which the present system was
able to resolve the couble humped fission fragment distri-
bution (see Fig. 15). This is thought to be due to the large
amount of light given out for each fission fragment allow-
ing much better energy resolution.

Just why CF4 should have such a high specific scintil-
lator efficiency is not clear, nor is it known what is
responsible for the emission in the ultraviolet.

We expect the excited states of the molecule to be
rather high due to the tight bonding of fluorine. The
higher the initial excited state is, the greater the proba-
bility usually is of non-radiative deexcitation. ^ Dis-
sociation following excitation has been 'suggested as the
reason for the almost total lack of luminescence in other
inorganic polyatomic molecules.'^ The same reason has been
given for the absence of luminescence in saturated hydro-
carbons. ^^ This reason may account for the observed lack
of luminosity on the part of UFg . ^" The emission from CF^
remains a mystery.

In all these measurements it was necessary to remember

76

S-

5-

O

c\

+->

^

UD

CO

LO

E

c

■—

^

r^

u_

<r>

t_>

n

Ln

O

o

§

o

Oo

o

IT)

oo. _

O

0*°

CM

o

o<^

>*

cr>

•^

r.

! -

r^

U-l

u

^o O o

o o

O Oo

<?)f

LD -^

>,

.— ZD

+J

z:

■rH

U

_j

0

UJ

a,

CO ^

E

oo -:

M

•"" <=c

re

CM

(_)

Z

o->

O

,

cr>

m

dJ

j_(

1 —

r^

3

tn

■rH

Cm

CXJ

LD

CXI

o

O

CM

o

SiNnOD

77
that the number of excited states and, thus, the intensity
caused by the passage of a fission fragment would depend
upon which fission fragment was observed. The initial
energy of fission fragments varies and so does the energy
deposition. If we set the discriminator level so as to
trigger on only the most energetic half of the fission
fragment distribution, we found as expected, that the in-
tensity per fission fragment increased. Thus, all the
measurements described in this dissertation are for the
mean of all particles, and the discriminator" had been
carefully set to include all the fission .fragments but to
exclude all the alpha particle's. It should be noted that
in a reactor experiment, the observed emission comes from
both the fission fragments and the alphas, as there is no
way to distinguish excited states caused by the one from
the other. Some differences may, therefore, be expected
in experiments done in the two environments.
5 . 6 Population Inversion Study

To determine if a population inversion exists between
two levels, it is sufficient to monitor the relative
spontaneous emission coming from each level when, by cor-
recting for the relative transition probabilities, we can
determine the population ratio. Observation of the steady
state excitation does not always reveal the presence of
pulsed inversions since a long uninverted tail may mask an
inversion which occurred early in time.

The most difficult problem associated with this type
of measurement was finding an allowed transitipn whose

78

upper and lower levels were both within the purview of
the system. Usually, if one of the transitions was in the
visible, the other was in either the ultraviolet or the
infrared. The laser candidate which we studied was the
second positive band of nitrogen (see Fig. 8). Pulsed laser
emission had been observed previously,^" and so this band
provided a good choice for testing the method. In addition,
we had available a computer solution of the rate equations
of nitrogen (Fig. 40) for electron beam excitation." These
solutions were of the form I versus t, exactly the output
of the system. The upper and lower levels of this band
system are the C^lTu levels which could be monitored through
the second positive emission itself, and the B Flu level
which gives rise to the first positive group (Fig. 41) . The
measured populations and the way they changed with time had
the same form as that predicted by the solution of the rate
equations and showed the same decay rate when allowance had
been made for the difference in pressure between the com-
puter solution (3 atmospheres) and the experiment (1
atmosphere) .

79

o
o

CsJ

AiISN3iNI 3AIi\/13y

30

ra

u

113
OJ

O

CO

b\

Ar/N^ (2.5';c)
800 torr

1st Positive
T = SOOnsec

2nd Positive
T = 40 n sec'

_L

J

100 200 300 400 500
TIME (ns)

Figure 41. Measured 1st and 2nd Positive Populations

CHAPTER 6
CONCLUSIONS AND FUTURE WORK

6 . 1 Conclusions

The conclusions arrived at in this research fall
naturally into three categories: conclusions about the
system, conclusions about pure nitrogen, and conclusions
about gas mixtures and fission fragment excited gases in
general .

The detection system was similar to existing fluor-
esence decay systems except that no trigger pulse was
necessary, and so sources that were non-periodic, transient,
and randomly occurring could be studied. The system was
extremely sensitive, capable of counting single photons,
and was relatively insensitive to gamma background. The
source, unlike a discharge, did not cause extensive thermal
decomposition or dissociation of the test gas, nor was it
necessary, as with -a flash lamp, that the level of interest
be connected to the ground state by an allowed transition.
The system could measure times as short as 1.7ns and, per-
haps, shorter. It had repeatability of better than 4% and
a minimum resolving time of about Ins. The minimum re-
solving time was limited by the photomultipliers and could
be improved as photomultiplier transit spread is improved.
The results obtained for the C^Ilu level of pure

81

82

nitrogen were the same within experimental error as those
obtained by others in the past (see Table 1) . The gas had
a single decay time and showed no deviation from two-body
quenching until 800 torr. There appeared to be no other
source of excitation to this level other than by the
secondary electrons. The energy deposition by the fission

fragment was found to exhibit an almost perfect square law

2
dependence and is a maximum when tf = t/Z .

Comparing Figures 11 and 13, we may observe that a
pulsed laser might operate best at a differenct pressure
than would a CW laser. Since, in the pulse case, we are
interested in the greatest instantaneous population, we
would choose the peak in Figure 11, while for the steady
state case, we are concerned with the average population
density and would follow curve 13.

Looking at other gases and gas mixtures, we found that
CF4 gave the greatest light output with a 10/1 argon/
nitrogen mixture second. For the mixture neon/nitrogen,
where there are no resonant interactions, whatever level
was being observed showed only small changes from the pure
gas case except that the decay time was determined by the
total pressure and not the partial pressure of the observed
gas. The lifetime of the N2"^ molecular ion showed extreme-
ly large values when nitrogen was a small impurity (less
than 1%). No explanation for this behavior was found. In
argon/nitrogen and helium/nitrogen, the lifetimes were,
dominated by resonant collisions with metastables. In argon.

83

TABLE 1

Lifetime

k2/kjX10^

(nsec)

(torr~^)

40.5±1.3

1.75±0.06

48.0+8.0

1.60+0.16

46.3±1.0

1.70i0.20

49.0±5.0

37.0±3.0

1.49+0.04

38.2±8.0

35.5±2.0

1.50±0.20

48.0 .

1.67+0.29

44.5±6.0

Previously Published Values for C Flu
Reference No . 5j

84
the C IIu showed some enhanccTient while in helium, the B Eu"^
was tremendously increased.
6 . 2 Future Work

Much work remains to be done using the isotope excited
fluoresence decay system to further characterize fission
fragment excited plasmas. The inclusion of an electric field
within the experimental chamber would allow the separation
of effects due to ions from those due to neutrals. In
addition, the electric field could be used to increase the
light output, thus, reducing the demands on the spectroscopic
apparatus. Cooling the chamber would allow the study of
gases which undergo thermal decomposition at room temperature.

In order to separate the effects of the fission fragments
from those of the secondary electrons, a beta source (such as
-^^C) might be included in the chamber. The beta energy dis-
tribution spectrum of C closely resembles that of the
secondary electrons created in fission fragment bombardment.
A shutter over the fission fragment source would allow ob-
servations to be made of the gas with electron excitation
alone or with fission fragment plus electron excitation.

Another use of this system is for studying extremely
electronegative gases. These gases, of which UFg is one of
particular interest, are difficult to excite using electrical
charges, and one must often resort to "seeding" an inert gas
with small fractions of such molecules.

Still another use of this system is in the serach for
possible fission fragment excited laser inversions. Since

85

the system provides time resolution, it is possible to
identify inversions which are quenched at long time and are,
thus, candidates for pulsed operation.

This system would also work well for studying
liquids.
6 . 3 Scanning Mode Recommendations

A major improvement in the scanning technique could be
experienced through the use of a minicomputer or micropro-
cessor. The scanning rate could be controlled in such a
way that when there was no signal (as determined by compar-
ing the count rate against background) , the system would be
stepped quickly. When a spectral feature is reached, the
system would step slowly enough to achieve any required level
of statistical accuracy.

Another interesting system could be developed to com-
bine the features of the scanning system and the fluorescence
decay system. In this mode, the system would accumulate a
decay curve, read it out, step to a new wavelength, and
repeat the process. This mode would be of use where there,
is emission continuum such as the xenon excimer band where
the lifetime varies across the band. A scan of this nature
would allow one to choose the region best suited to lasing.

APPENDIX 1
ENERGY DEPOSITION

The whole question of how many excited states are
produced by a fission fragment revolves about the exact
form of the energy deposition equation.

The Bohr energy loss equation describes the slowing

down of a fission fragment due to electron collisions as

dE ^ 4TTN(Zeff)^e'^ZN

dr M^v2 ^^r

where e = electronic charge,

ZgjTf = effective charge of. fission fragment with
velocity V traversing a medium of atomic
number Z^,

Lg = logarithmic summation term, and

V = velocity.

A semi-empirical expression for Z^^f, also due to Bohr, is

Zeff = Zff^ (V/Vq),

where Z^^ = the atomic number of the fission fragment,
and

Vq - the velocity of the electron in the first
Bohr orbit.

dF ' *?

For a gas, ^ = -kV. Integration gives E = En(l-r/R)'^.
dr r> ri u

This expression has been described as giving poor agreement
with experimental data.^ From energy loss measurements in
gases, the exponent has been assigned a value from 1.5 to
1.7. Axtmann^° has derived an expression for the intensity

86

37
of the fission fragment exci+ ad gas in terms of the exponent in
the energy deposition equation. Let

E^ = Eq - E = Egd-Cl-r/R]") (1)

be the energy dissipated in the gas.

A binomial expression of Equation 1 gives

E^ = EQn6[l-B(n-l)P/2]P, ^2)

where g = r/RgPo/ and 0 refers to one atmosphere. Now the
measured intensity equals

ki
I = (eQEd) F[,^^^], (3)

where Q is the number of excited states formed per unit

energy loss in the gas; e is the collection efficiency; F

'^l
is the number of fission fragments per second; is

5^1+^2?
the probability of light emission by an excited molecule;

P is the pressure in torr; k-^ and k2 are the radiative and

collisional decay constants with units of sec"-'- and mm"-'-

sec~ respectively.

Substituting Equation 2 into Equation 3, differentiating

J T

and setting -Tp- equal to zero gives

n3(l+j^P ) (1-SPn,)"'^

l-(l-Pj^)" =0,

k2/ki

where n is obtained implicitly in terras of 6, k2/k-]_ and Pj^

the pressure of maximum luminosity. For this experiment

6 = —ijJ-i — = 2.335x10"^, and P^ = 300 torr. Thus,
2.2x800 ^

l-(l-2 34x10-^300)--- ["^2. 34xl0"^l+. 018x300) (1-2. 34x10-3x300) ""^_Q

.018

88

A fit of our experimental data gives n = 1.98 within 1%,
sufficiently close to two to justify using the expression,
E = Eq (1-r/R) 2.

■ ■ ■ . APPENDIX 2

NE-111 AS A CALIBRATION SOURCE

After the delayed coincidence single photon counting
system had been assembled and tested, a search was made
for a fast, low level light source to calibrate the timing
measurement. The characteristics required of the source
were :

1) The pulse should have as small a half-width
as possible and resemble in shape the
fission fragment pulse.

2) The pulse amplitude should be as near that

of the fission fragment pulse as possible,

' ft

on the order of 10 photons.

3) The source should be small so that it would
fit in the experimental chamber. The opti-
mum solution would have the source small
enough that it could be inserted in the
chamber through one of the gas access ports
so that the calibration could be done with-
out disturbing the system.

4) The source should have near 4it geometry so
that it could be observed by both photomul-
tipliers simuntaneously .

Normally, single photon counting systems are calibrated

89

90

using a tungsten ribbon lamp and filters. This was
clearly not possible in our case since we needed a pulsed
source. Choppers were briefly considered, but the rotation
speeds necessary to achieve chopping speeds of Ins were
considered excessive.

Fluorescence decay systems are normally calibrated
using a quenched spark source. This source was too large,
too bright, and too expensive to fit easily into this
system.

The solution was a small chip of NE-111 , a fast
plastic scintillator mounted on the end of a short length
of copper tubing brazed to an 1/8-inch pipe plug which
could be screwed into the gas system access port. The NE-111
was secured to the end of the copper tubing by heating the
tubing and pressing the plastic into the end. The length
of tubing was so chosen that when the plug was screwed all
the way in, the source was in the center of the field of .

view.

The shape of the light pulse which NE-111 emits when
struck by a gamma ray is well known, having a full width
at half maximum of 1.54ns and a decay time of 1.7ns. The
output peaks at about 370nm, a wavelength well suited to
most spectroscopic systems.

In use, the calibration source was placed in the cen-
ter of the experimental chamber and viewed by the detection
system. Since both photomultipliers were seeing the same
light pulse, the measured time interval, corres^ponding to

91

no time difference, was subtracted from all decay measure-
ments. This zero time inteival was caused by differences
in cable lengths, light paths, and photomultiplier transit
times.

APPENDIX 3
GAMMA VERSUS TLP FOR START PULSE

As described in Chapter 3, a test was made to deter-
mine if using the prompt fission gammas gave the same
decay curve as using the output of the TLP. Figure 42
shows a pair of measurements of the decay of the C Ilu
state of nitrogen. In both cases, the stop pulse vvas
supplied by the SPL while tlie start pulses were supplied
by the TLP (the ''s) or the gamma detector (the A). The
count was continued in each case until 512 counts had
been accumulated in the peak channel. The decay times
determined from the tv;o curves are well within the 3%
statistical precision of the experiment. Thus, from
the standpoint of accuracy, there was no distinction be-
tween the two methods of timing; however, using the TLP
provided a count rate about 1000 times greater than using
the gamma detector. For this reason, all measurements
were made using the TLP.

92

93

o

<•

+->

Q. S- CL

4-> O fd O

S_ +J +J 4->

+->

_J _l _J

C3
CVI

1/1

UJ

M

w

CM
0)

>

^ •

(T3

o

en
C
•H
in

ID

<•

• 4

0)
-H

O

o

o

o

o

o

o

o

o

o

o

un

■^

m

CXJ

. —

SiNnOD

APPENDIX 4
GAS PURLTY

One question which had to be addressed each time a run
was made was "How long must the vacuum system be pumped on
so that the residual impurities and the impurities which
leak in and out gas off the walls during the course of the
experiment will not affect the results of the measurements?"

Clearly, if the purity of the test gas was low, the
system need not be pumped so long as if research grade gas
were being used. As an example, assume that the test gas
has a purity of 99.9995% or 5ppm impurity which sets the
lower limit on gas purity. Suppose the impurity that the
experimental apparatus contributes to the gas is required
to be at most 10% of that figure or 0.5ppm. Then, if the
chamber is filled to 240 torr pressure, the equivalent
impurity pressure from the system must be no more than 24 Ox
0.5x10"^ torr or 0.12 microns. However, if the count is ex-
tended one, and the impurity level is to be no greater than
this at the end of the measurement, the system must be
pumped farther down before admitting the test gas. It was
found that when the ' system had been pumped to its ultimate
pressure of SxlO-^mmHg, it required 72 hours for the pressure
to rise to 0.1 micron. Therefore, for those experiments re-
quiring the greatest degree of gas purity, namely low

94

pressure ultra pure gas, the count was limited to 72 hours
On the other hand, for tank gas of 99.995 purity at one
atmosphere, the required impurity pressure was about 3.5
microns, a pressure which the system would maintain indefi-
nitely.

APPENDIX 5
FISSION FRAGMENT TRANSIT TIME

One surprising result of this study was the near
constancy of the fission fragment transit time. The analy-
sis of the decay curves was constrained by two conflicting
requirements. The least square fit should start as early in
time as possible since the nearer the peak, the greater
the number of counts and, hence, the larger the statistical
accuracy. However, if the fit was started too early, before
the fission fragment had passed out of the field of view,
the population would not be decaying with a simple exponen-
tial decay, and so the fit would be distorted.

In examining the early intensity curves, it was noted
that the rise time was equal to about 4ns independent of
gas pressure or molecular weight; a result v^?hich was thought
to illustrate some limitation of the system. When the
transit times were calculated using the expression in
Chapter 3, it was found that in going from helium at six
torr to argon at 800 torr, the transit time only changed,
from 4.28ns to 4.12ns. This result, while at first sur-
prising, can be explained as follows: When the fission
fragment first entered the gas, it was very highly charged
and was traveling very fast; fast, that is, with respect
to the orbital velocity of the electrons bound to the

96

97

target gas. If the velocity of the fission fragment com-
pared to that of the bound electron is such that v2/u^>>l,
where v is the velocity of the fission fragment and u is
the vlocity of the bound electron, then the electron
cloud around the atom or molecule does not have time to
adjust to the rapidly passing microfield and inelastic
collision occurs. The initial velocity of the fission
fragment is 1 . 04xl0^cm/sec for the heavy fragment of 80
MeV and 1 . 37xlo"cm/sec for the light fragment of 106 MeV,
which is to be compared with a typical orbital velocity
of 4 . 4xl0~^cm/sec for the IS^ state of helium

Thus, v^/u^ equals 5.6 for the heavy fragment and 9.7
for the light, both v\?ell above unity. When the fragment
has slowed to the point where v^/u^ was equal to unity,
the velocity of the fragment would still be 4 . 4xl0^cm/sec
(about 10 MeV for the light and 30 MeV for the heavy frag-
ment) , but it would no longer undergo inelastic collisions
with the electron cloud, and, hence cause no more excited states.
At that velocity, the heavy fragment would have lost 50 MeV
of its original 80 and the light, 95 MeV of its original 105
MeV. The significant result here was that even after the
fission fragment had lost all the energy it could to the electrons,
and so for the purpose of this experiment it could be said
to have completed its transit, it was still traveling at no

98

Q

mean speed, 4 . 4xl0°cm/sec and crossed the chamber very
rapidly. This then explains the near constancy of the
intensity curve rise time and the calculated transit times:

1) Because the excitation cross-section is
determined by velocity rather than energy
(the Bethe-Born approximation) .

2) Because the velocity goes as the square
root of the energy.

3) Because even when the fission fragment has
lost all the energy to the electrons that
it can and is, therefore, no longer observ-
able, it is still one speedy little mother.
Thus, whatever happens to it in crossing the
chamber affects the transit time hardly at
all.

APPENDIX 6
252cf CHARACTERISTICS

7 S 7
The Cf isotope used in this experiment was an

open source made by evaporating an HCl solution of the

metal onto a thin nickel disk. This isotope layer was

only a few atoms thick and, thus, the energy spectrum of

the emergent fission fragments was essentially tlie birth

distribution and showed the familiar double humped shape

(Fig. 14). The source had a nominal activity of 5pCi

252
and weighed O.Oliig. The main characteristics of Cf

are shown in Table A.

Using these numbers, we may calculate the relevant
characteristics of the actual source. These are shown in
Table B. Measurements of the fission fragment and alpha
rate, which were the only ones attempted, gave nearly iden-
tical numbers.

99

100

TABLE A
MAIN CHARACTERISTICS OF 252^^

Half-life:

2.646 y

Fission Fragment Energy

: Light, 105.71 MeV
Heavy, 80.01 MeV

Fission Fragment Ma.ss :

Light, 108.39 AMU
Heavy, 14 3.61 AMU

Fission Activity:

16.8 uCi/ug

Fission Half-life:

85.5 y

Alpha Activity:

526 viCi/ug

Alpha Half-life:

2.731 y

Alpha to Fission Ratio:

31.3:1

Average Alpha Energy:

6.117 McV (02':.)
6.08 MeV (15%)

Neutrons Per Fission:

3.76

Neutron Activity:

63.2 (jCi/yg

Neutron Emission Rate:

2.34x10^2 sec"^g"^

Average Neutron Energy:

2.34 8 MeV

Gamma Energy:

4 3 keV

Beta Energy:

22 keV

01

TABLE B
CHARACTERISTICS OF THE SOURCE

Neutron Rate

2.34x10'* sec~^

Gamma Rate :

1.3x10^ sec ^

Alpha Rate

1.95x10^ sec"l

Fission Rate

6200 sec

-1

BIBLIOGRAPHY

1. R. A. Walters, Excitation and Ionization of Gases by

Fission Fragments, Ph.D. dissertation. University
of Florida (1972) .

2. R. N. Davie, Spectroscopy of Fission Fragment Excited

Gases by Fission Fragments, Ph.D. dissertation.
University of Florida (1975).

3.' J. C. Guyot, Measurement of Atomic Metastable Densities
in Noble Gas Plasmas Created by Nuclear Radiations,
Ph.D. dissertation, University of Illinois at Urbana-
Champaign (1971) .

4. P. E. Theiss, Optical Emission and Kinetics of High-

Pressure Radiation-Produced Noble Gas Plasmas,

Ph.D. dissertation, University of Illinois at Urbana-

Champaign (1971).

5. J. M. Calo, R. C. Axtmann, and R. G. Persing, Rev. Sci.

Inst. , 4^, 11 (1970) .

6. L. G. Christophorou , Atomic and Molecular Radiation

Physics , Wiley Interscience , N.Y. (1971) .

7. R. H. Lo and G. H. Miley, IEEE Trans, on Plasma Science

(1974) .

8. G. Herzberg, Molecula.r Spectra and Molecular Structure:'

I. Spectra of Diatomic Molecules, 2nd Ed., D. Van
Nostrand, Princeton, N.J. (1950).

9. G. H. Miley and P. E. Theiss, Nuclear Applications, ^,

(1969) .

10. R. C. Axtmann and J. T. Sears, Nuc. Sci. &< Eng., 23

(1965).

11. D. F. Keefer, Private Communication (1976) .

12. W. K. MacGregor, Private Communication (1968).

13. M. E. Jones, Private Communication (1970).

14. R. A. Hefferlin, Private Communication (1975).

102

103

15. J. F. Davis, Kinetic and Experimental Study of Argon

and Argon-Nitroger Mixtures Excited by Fission
Fragments , Ph.D. dissertation, University of Flor-
ida (1976) .

16. R. Pagano and J. W. Wethington, Jr., Trans. Am. Nucl.

Society, 14 (1971) .

17. P. Pringsheim, Fluorescence and Phosphorescence, Inter-

science Publishers, N.Y. (1949).

18. F. Hiroyama and S. Lipskey, J. Chem. Phys . , 5J^ (1969).

19. R. N. Davie, et al , Third Symposium on Uranium Plasmas,

Princton, N.J. (1976) .

20. E. T. Gerry, Appl . Phy. Lett., 1, 6 (1965).

21. R. A. Heustis, et al , Stanford Research Institute Tech-

nical Report 4, "Project PYU-1925 (1975).

BIOGRAPHICAL SKETCH

George Robert Shipman was born January 3, 1944, in
Bridgeport, Connecticut. He moved with his family to
Orlando, Florida, in 1953. He attended Fletch Academy,
Fletcher, North Carolina, and Forest Lake Academy, Maitland,

Florida.

He received the Bachelor of Science degree at the
University of Florida with a major in Physics in June, 1966,
and was elected a member of Sigma Pi Sigma.

In 1968, he received the Master of Science degree in
Astronomy with a minor in Physics from the University of
Florida. He has held an Atomic Energy Commission trainee-
ship and a departmental assistantship while in the Depart-
ment of Nuclear Engineering Sciences.

104

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.

lairman

Associate Professor of Nuclear
Engineering Sciences

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of DoctorxTsf Philosophy.

W

•Jt ^^hjLicA..

R. T. Schneider

Professor of Nuclear Engineering

Sciences

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality.

as a dissertation for the degree of Doctor of Philosophy

E. E. Carroll

Professor of Nuclear Engineering

Sciences

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.

T. L. Bailey

Professor of Physics and

Astronomy

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.

tl,:

X w

K. Y.UChen

Associate Professor of Physics

and Astronomy

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy,

R.' Lebc

Assistant Professor of Physics
and Astronomy

This dissertation was submitted to the Graduate Faculty of
the College of Engineering and to the Graduate Council, and
was accepted as partial fulfillment of the requirements for
the Degree of Doctor of Philosophy.

December, 1976

Engineering

Dean, Graduate School