The pulse radiolysis of alkyl iodides and oxygen in the gas phase


Material Information

The pulse radiolysis of alkyl iodides and oxygen in the gas phase
Physical Description:
x, 135 leaves : ill. ; 28 cm.
Ramirez, Jorge E ( Jorge Eugenio ), 1952-
Publication Date:


Subjects / Keywords:
Kinetic theory of gases   ( lcsh )
Chemical kinetics   ( lcsh )
Radiochemistry   ( lcsh )
Iodides   ( lcsh )
Oxygen   ( lcsh )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1981.
Includes bibliographical references (leaves 127-134).
Statement of Responsibility:
by Jorge E. Ramirez.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000297666
notis - ABS4041
oclc - 08480127
System ID:

This item is only available as the following downloads:

Full Text







Dedicated to

my parents Luis and Anita

and my nephews Ray and David


The author expresses his sincere appreciation to his research

director, Prof. Robert J. Hanrahan, and Prof. M. Luis Muga for their

advice, encouragement, and friendship throughout this work. He also

thanks Prof. Phillip M. Achey for the many stimulating and candid


Special thanks are given to Kelley St. Charles, Bob Doyle, and Dawit

Teclemariam for their friendship and assistance, especially when "the

going got tough," on and off the field.

Special appreciation is given to John Kurtz, Mike Collins, and

John Pooser for their friendship and all the "red, white, and blue."

His deepest thanks are given to the WGAS gang in Miami, Robert

Pridgen, Nestor Herrera, Martin Miller, Lynn McCullough, and especially

Joe Benitoa for the "best years of our lives."

Above all, the author expresses his deepest love and gratitude to

his father and mother and his brothers and their families whose love

and understanding made this work possible.

Thanks to Frank Ramos for making many of the drawings in this text.

Thanks Wil.


ACKNOWLEDGMENTS ........................................... iii

LIST OF TABLES ................................... ........... v

LIST OF FIGURES ............................. .... ........... vi

ABSTRACT .................................................... ix

I. INTRODUCTION .......................................... 1

A. Forward ........................................ 1
B. Review of Previous Work ..... .................... 2


A. Monochromators and Light Transport Apparatus ..... 10
B. Optical Efficiencies ............................. 27
C. Reaction Cells .................................. 34
D. Spectrophotometry .............................. 43
E. Lamp Systems ............... ............ ......... .. 49
F. Febetron 706 System ............................. 74
G. Transient Recorder and Computer System ........... 77
H. Computer Programs ....................... .......... 80
I. Reagents and Their Purification ................. 82
J. Sample Preparation ............................ .... 84
K. Sample Irradiation ......................... ....... 87


A. Experimental Results ...................... .. ... 89
B. Discussion ....................................... 101

RADIOLYSIS OF OXYGEN .................................. 105

A. Experimental Results .................. ........... 105
B. Discussion ...................................... 110


REFERENCES .................................................. 127

BIOGRAPHICAL SKETCH ......................................... 135


Table Page

1 I Quenching Rates by RI ................................. 95

2 I Quenching Rates by the Medium ........................ 98

3 I Fraction and Branching Ratios ........................ 100


Figure Page

1 Entrance and exit slit system flange adaptation
to the Jarrell-Ash monochromator ..................... 12

2 Coupling arrangement from the lamp flange to the
entrance slit flange (Fig. 1) ......................... 13

3 Exit slit flange adapter flange and window mechanism
for connecting the exit slit flange to the PMT
housing, Model 3262, for end-on PMTs .................. 15

4 Quartz rod light pipe guides and housing .............. 17

5 Nylon guides .................. ..................... 18

6 Lead shields for PMT photocathodes in the Pacific
Precision Instruments PMT housings .................. 20

7 Single-pass cell arrangement for VUV work ............. 22

8 Vacuum extension flange and lens focusing arrangement .. 23

9 McPherson 218 exit window flange and window ........... 25

10 High vacuum pumping system for the McPherson 218
monochromator ................. ..................... 26

11 University of Florida Pulse Radiolysis System ......... 28

12 900 optical configuration using a 10 or 7.5 cm
focal length parabolic mirror ........................ 29

13 Iodine discharge lamp spectrum with nitrogen
flushing of the optical path ...... ......... ..... .... .. 32

14 Iodine discharge lamp spectrum with helium
flushing of the optical path ........................ 33

15 Single-pass cell with six equal flange ports .......... 35

16 Multiple-pass reaction cell arrangement for visible
and UV studies ....................................... 38

LIST OF FIGURES (continued)

Figure Page

17 Multiple-pass reaction cell mirror system with
MgF2 coated first surface aluminum mirrors ............. 39

18 Reaction cell-Febetron 706 face plate coupling
mechanism ......................................... 41

19 Complete assembly of the reaction cell flange
onto the Febetron face plate ........................... 42

20 Photomultiplier tube base schematic used with EMI
9750 QB and 9856B PMTs and Model 3262 housing .......... 46

21 Photomultiplier tube base schematic used with EMI
9783B side window PMT and Model 3150 housing ........... 48

22 Lamp system ............................................ 51

23 Evenson cavity ......................................... 53

24 Microwave discharge lamp bodies ...................... 56

25 Lamp window and coupling flange ....................... 57

26 Iodine emission spectrum from the microwave discharge
lamp using a sodium salicylate fluorescent screen to
cut off stray visible light from other orders .......... 62

27 Iodine discharge lamp spectrum after 5 minutes of
active pumping on the McPherson 218 .................... 64

28 Iodine discharge lamp spectrum after 10 minutes of
active pumping on the McPherson 218 .................... 65

29 Chemical compound absorption interference in the
spectral range 170 to 190 nm (Fig. 26) ................. 66

30 Absorption band for 2.5 torr methyl iodide in
the 200 nm region ...................................... 67

31 Emission spectrum of OH from the microwave
discharge lamp ........................................ 69

32 Mercury lamp system in DC operation mode ................ 70

33 Hg lamp intensity versus time for various input
sensitivities of the transient recorder with the
reaction cell evacuated ...................... .......... 72

LIST OF FIGURES (continued)

Figure Page

34 Hg lamp intensity versus time for two initial
oxygen pressures and vacuum in the reaction cell at
1 VFS transient recorder input sensitivity ............ 73

35 Febetron 706 system ................................. 76

36 Vacuum line and pumping system ....................... 85

37 Graphical display of optical signal versus time
for a single Febetron pulse .......................... 90

38 Graphical display of optical signal versus time
for three Febetron pulses summed into the buffer ...... 91

39 Curve fit of data from Fig. 38 ....................... 93

40 Pulse radiolysis of methyl iodide with 180 torr
argon .......................... .................. 96

41 Pulse radiolysis of trifluoromethyl iodide with
575 torr argon ................................... 97

42 Computer calculation for I* deactivation, intermediates,
and products .. ...................................... 103

43 Ozone absorption growth monitored at 253.7 nm
following the pulse radiolysis of 750 torr 02 ......... 106

44 Dosimetry: Ozone yield in the pulse radiolysis of
oxygen measured at 253.7 nm ........................... 107

45 First-order plot of ozone absorption growth (Fig. 43)
following the pulse radiolysis of 750 torr 02 ......... 108

46 Ground vibrational state ozone formation rate
versus 02 pressure ................................... 1

47 Computer calculation for a=b=0.2 and c=0.6 ............ 113

48 Computer calculation for a=0.1, b=0.6, and c=0.3 ...... 115


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



Jorge E. Ramirez

December, 1981

Chairman: Dr. R.J. Hanrahan
Major Department: Chemistry

A pulse radiolysis facility for the study of fast chemical kinetics

in the gas phase, based on a Febetron 706 electron beam accelerator, has

been established at the University of Florida Radiation Chemistry Labora-

tories. Maximum accelerator specifications are 600 keV electron energy,

8000 amps, and 10 joules per pulse; pulse width is 3 nanoseconds (FWHM).

Hardware has been developed for studies in the ultraviolet and vacuum

ultraviolet regions using absorption spectrophotometry.

Parameters relevant to a linear electron beam initiated atomic iodine

laser, which would lase on the transition 52 P/2 (I*) 52P3/2 (I), have

been investigated by pulse radiolysis. Kinetic data for the parent

compound quenching of I were obtained by variation of the parent compound
pressure at constant buffer gas pressure and observing I decay rates

versus time. Deactivation rate constants for the perfluoroalkyl iodides

were found to be much lower than for alkyl iodides (in cm3/molec s): CH3I,

2.0 0.1 x 10-13; C2H5I, 5.0 0.3 x 10-13; CF3I, 8.8 1.5 x 1016;

C2F5 9.7 1.0 x 10-15; i-C3F7I, 1.7 0.1 x 10-15; C4F91, 1.8 0.1 x 10-14


The extent of population inversion was investigated by measuring

initial excited state and ground state atomic iodine concentrations and

calculating the branching ratio, [IP]o/ []o: CH3I, 2.7; CF3I, 3.8; C2F5I,

2.7; i-C3F7I, 3.2; C4FgI, 1.8. Perfluoromethyl iodide showed the

largest population inversion from electron beam irradiation.

Using formation of 03 from 02 as the dosimeter (G = 13.8 molecules/

100 eV for ozone), energy deposited per electron pulse in 750 torr 02 was

7.3 x 1018 eV/g. Spectrophotometric detection of ozone utilized the

253.7 nm Hg line.

The rate of formation of ground vibrational state ozone in the pulse

radiolysis of oxygen was followed using the 253.7 nm Hg line. The rate

of formation was found to be nearly second-order with a rate constant of

4.1 + 0.4 x 10-15 cm3/molec s. It is formed from direct combination of

oxygen atoms and oxygen molecules and from collisional quenching of

vibrationally excited ozone and/or an electronically excited ozone

precursor, probably ozone (3B2).


A. Forward

The establishment of a pulse radiolysis system was undertaken to

expand the chemical kinetics facilities of the University of Florida

Radiation Chemistry Laboratory. Major efforts were made to design the

system for the greatest versatility for the studies undertaken as well

as for possible future studies.

Pulse radiolysis is a technique for studying the kinetics of species

involved in the fast elementary steps of a reaction process following

sample irradiation by a short pulse of ionizing radiation. In general,

the transient species are of major interest. The direct study of fast

processes has two basic requirements. First, the process must be

initiated in a time scale which is shorter than or comparable to the

process being observed. Second, the process to be observed must have

associated with it some detectable change in one or more physical


To meet the first requirement, the pulse radiolysis system was

developed around a Febetron 706 electron beam (e-beam) accelerator. A

maximum 8 kiloamp beam of maximum 600 keV electrons in a pulse duration

of 3 nanoseconds (FWHM) could be delivered for sample irradiation. The

second requirement must be met by the chemical system chosen for study.

To aid the study of fast chemical processes, a microcomputer-based

data acquisition system was coupled to the pulse radiolysis system.

The computer system was used to acquire, analyze, and store data.

The chemical studies of alkyl iodides and perfluoroalkyl iodides,

relevant to the atomic iodine laser, were undertaken to help elucidate

previous photolysis work. Also, since the flash photolysis and pulse

radiolysis techniques are very similar, previous studies served as

references as to the validity of data acquired during the development of

the pulse radiolysis system.

The oxygen-ozone studies were undertaken to establish a chemical

dosimeter for our Febetron 706 system.

B. Review of Previous Work

Pulse radiolysis

At the University of Florida Radiation Chemistry Laboratory there

has been a continuous interest in the reaction kinetics and mechanisms
of simple alkyl halide systems. Of particular interest in these

studies has been the identity and reactivity of important transient


Historically, pulse radiolysis was first developed to study inter-

mediates which had been postulated in radiation initiated chemical
reaction systems.6-8 Though the technique was developed by radiation

chemists, it was quickly employed for the study of transients of broader
chemical interest. 1 The developmental work on pulse radiolysis was

carried out on liquid phase systems. The use of liquid systems was

dictated by the availability of only high-energy low-current accelerators.

Densities of liquids were convenient for sufficient interaction of the

system under study with the ionizing radiation (greater than 2 MeV), in

order to deposit enough energy to provide an observable concentration

of transient species. In order for gases to absorb enough energy from

high-energy electron beams, high pressure systems (10-100 atm) were
initially used. 14 Gas phase work at an atmosphere or less was made

possibly by advances in detection techniques and the development of high

pulse current accelerators like the Febetron 706.14

In principle, pulse radiolysis is similar to flash photolysis. The

photoflash and photodissociation of molecules is replaced by a short

pulse of high-energy electrons with subsequent excitation and ionization

of the molecules by electron impact. Subsequent dissociation can occur

by unimolecular decomposition or by collision with other atoms or

molecules. The important properties of the electron pulse are its

maximum energy, pulse current, time profile, cross-sectional homogeneity,

and divergence.

The maximum electron energy is directly related to the e-beam's

penetration in a reaction cell; therefore, it must be large enough to

sufficiently irradiate the chemical system. The concentration of the

species of interest is directly related to the number of interactions

between the chemical system and the e-beam. This is determined by the

pulse current. For convenient time resolution of the observed signal

the time profile of the pulse must be shorter than the half-life of the

transient of interest. The cross-sectional homogeneity and the divergence

of the beam are important for uniform irradiation of the chemical system.

This helps ensure optical sampling of a homogeneous reaction region.

In its present form, the technique of pulse radiolysis requires a

high level of experimental sophistication.1013-16 Spectrophotometric

detection techniques have replaced less sensitive spectrophotographic
detection techniques.14 The efficiency of the optical system is of
detection techniques. The efficiency of the optical system is of

great importance for observation of useful signals. Although similar e-

beam sources have been used in previous studies, the associated optical

systems were customized for each individual laboratory.

The optical systems were designed with two major considerations.

First, optical throughput to the detector had to be maximized in order

to achieve the best signal-to-noise ratios possible.14,17 Second, due

to sources of noise associated with pulsed electron beam sources, simple

optical concepts took on elaborate physical designs in order to appropriately

shield the photodetector and electronic circuitry.10'14

A desire to carry out studies in the ultraviolet (UV) and vacuum

ultraviolet (VUV) regions further complicated the optical spectrophotometry

systems. Specific attention had to be given to the properties of the

optical materials and detectors at these wavelengths. Optical system

throughput efficiencies were especially important in these regions.18-20

In order to facilitate data acquisition and reduction, on-line

computer systems have been developed at various levels of sophistication.21

These systems have greatly reduced the time required for data analysis.

The quantity and quality of the data processed in a given time were

greatly increased by the use of on-line computers.22

The pulse radiolysis system in our laboratory was designed to carry

out low pressure (atmospheric and below) gas phase studies in the UV and

VUV. Once operational, the system was used to study transients important

in the operation of an iodine atom chemical laser.

Alkyl iodides-atomic iodine studies

Laser action on the atomic transition

5 2Pl/2 (I*) 5 2P3/2 (I), I-1

where I* represents the excited state atom and I represents the ground

state atom, at 1.315 microns, was first reported by Kasper and Pimentel.25

Further studies raised interest in the dynamics of the photolysis of

alkyl iodides.26 Of main importance to the lasing action of the systems

is the production efficiency of the excited state atom, I*, its rate of

deactivation, and the the initial branching ratio. The initial branching

ratio is the ratio of initial concentrations of excited state atoms to

ground state atoms,

S= [I*]o / [I] o 1-2

In the cases where the branching ratio is large and where other threshold

requirements necessary for lasing are met, a high laser gain may be


Since the above transition is electric dipole forbidden, the mean

radiative lifetime of the excited state is relatively long (0.13s).27

Spin orbit relaxation by emission or by collisional quenching competes

with chemical reaction of I*. If collisional quenching and chemical

reaction are slow relative to emission, then a population inversion may

be maintained and the excited atom population can initiate and propagate

stimulated emission in a laser cavity.

Further interest in the alkyl iodide systems was generated when it

was confirmed that the branching ratio was large for the photolysis of

several alkyl iodides.28 These studies were carried out using the

kinetic spectrophotographic detection technique and suffered from parent

alkyl iodide absorption interference in the region of interest, especially
in 29,30
in the VUV.

Several gas phase studies of the photolysis of alkyl iodide systems

have been carried out.346 Also, gas phase radiation chemical studies

of these systems have been carried out.1'23'47 Many of these studies

were directly stimulated by the need to understand the energetic and
kinetics of potential atomic iodine laser systems.'3
36 39,44
Specific interest363944 was further stimulated by the possibility of

initiating thermonuclear fusion by laser irradiation using the photoinitiated
atomic iodine laser.48

In general, most of these studies have used detection of I* by

absorption in the UV (206.2 nm) and VUV (179.9 nm) and by its emission

at 1.315 microns in the infrared (IR). The more sensitive spectrophotometric
17,37 42
technique 7 has replaced the spectrophotographic technique in the

absorption experiments. IR detection43'45 is still a very useful technique
for emission studies40 and compliments the absorption studies.

Mass spectrometric analyses of flash34 and laser32 photodissociated

fragments have also been carried out. The latter study yielded much

information about the energy distribution of the r.scertphotofragments

following laser photolysis of methyl iodide at 266 nm.

Laser gain studies have yielded pressure dependent information

about the atomic iodine laser46 as well as kinetic data on I* deactivation.39

In recent studies,32-3941-46 except where laser photolysis3235

and VUV photolysis33 were used, photoinitiation was carried out with

broadband flash lamps emitting in the 200-300 nm absorption band of the

perfluoroalkyl and alkyl iodides.29'30 The use of high energy linear e-

beams (pulse radiolysis) as the initiating source for the study of

deactivation rate constants has only been reported by this laboratory,23

although electron transverse discharge excitation of the atomic iodine

laser has been previously reported and modeled.53'54

More recently the use of a solar pumped iodine laser for power

transmission in space has been modeled55 and tested.56 These, as well
as the high power lasers, used the perfluoroalkyl iodides as the

source of atomic iodine. Although many studies have been made of alkyl

iodides, far less work has been done on the perfluoroalkyl iodides and

thus, in general, literature values of rate constants used in modeling
of the laser contained many discrepancies.5

Studies of the alkyl and perfluoroalkyl iodides have been carried

out using the pulse radiolysis system in this laboratory. These results

were compared, where appropriate, to other established results.

Oxygen-ozone studies

The reaction of oxygen atoms with molecular oxygen has been of
considerable interest in the study of the atmosphere. 59 Similarly

the dynamics of ozone decomposition has been investigated because of its

importance in atmospheric chemistry.60 The ozone formation reaction is

also important in the radiation chemistry of oxygen-containing systems.

The reaction

0 + 02 + M 03 + M I-3

has been extensively investigated.58'59'6178 Recent methods used in

these studies have included direct observation of ozone formation in

the Hartley band region (200-300 nm)79,80 via fast absorption spectroscopy
61-64,71-73 62,74-77
using pulse radiolysis61647173 and flash photolysis627477; flash

photolysis-resonance fluorescence5978; time-resolved infrared detection

of vibrationally excited ozone74'75; and direct observation of oxygen

atom disappearance via atomic resonance absorption.586566

The oxygen-ozone system was reinvestigated in the course of setting-

up the pulse radiolysis facility in our laboratory, primarily because it

is a convenient gas-phase dosimeter which has been studied with a similar

high dose-rate e-beam accelerator.63 Since the ground state of ozone

could be monitored conveniently with an intense mercury resonance lamp,

it was convenient to carry out some kinetic studies along with the

observations needed for the dosimetry work.

The time-resolved behavior of the absorption spectrum of ozone in
the 200-350 nm region was first observed by Hochanadel et al. They

attributed absorption at wavelengths other than at the normal ozone

ground state maximum near 260 nm to vibrationally excited ground state

ozone formed in Reaction I-3 with molecular oxygen as the third body.

Riley and Cahill observed a second absorption maximum near 315 nm which

they interpreted as a transient species other than vibrationally excited
ozone, because of its simple first-order decay kinetics.72 The kinetics

of vibrationally excited ozone cascading downward through its vibrational

levels was expected to be more complex than the simple process observed

at these wavelengths.62 Bevan and Johnson observed two transient species

absorbing with maxima near 285 and 310 nm and assumed that they were due

to vibrationally excited ozone states, kinetically distinguishable from

the ground state.73 Von Rosenberg and Trainor observed vibrational

excitation in two vibrational modes of ozone formed from Reaction I-
74 75
3.775 Using flash photolysis and time-resolved IR detection they were

able to determine energy partitioning in IR-active products, vibrational

relaxation rates, and recombination rates. They suggested that the

species previously observed near 310 nm72'73 was vibrationally excited

ozone and/or electronically excited ozone, which was far enough below

the dissociation limit to make cascading more probable than dissociation.

They further suggested that the absorption maximum near 310 nm could be the

ozone 3B2 bound electronic state.74

Recent determinations which used oxygen as the third body and

monitored oxygen atom disappearance concluded that the overall rate

constant for Reaction I-3 lies between 5.7-6.9 x 10-34 cm6/molec2 s 58,59

in agreement with earlier studies.65 In studies where ground vibrational

state ozone formation was monitored, the apparent rate of formation was

consistently slower than the rate of oxygen atom disappearance. This

anomaly has been attributed to vibrationally and/or electronically

excited ozone which was not being directly monitored.77 It was suggested

that the proper measurement of ozone should be the integrated absorbance

over the entire Hartley band which would include all ozone species.76,77

Kleindienst et al. modeled ozone formation in the flash photolysis of

oxygen-ozone mixtures by assuming that the 332 state of ozone acted as

the unobserved intermediate and needed to account for the anomaly in

the observed rate of recombination.77 Using the same assumption the data

for ground vibrational state ozone formation monitored at 253.7 nm in the

pulsed electron radiolysis of pure oxygen have been modeled.


A. Monochromators and Light Transport Apparatus

In the early stages of this work Baird Atomic Interference filters

were used for wavelength isolation. These were broadband filters and

only total light changes in a 20-40 nm region around the central wave-

length could be detected. Also, they did not provide useful throughput

of light for our purposes. Only exploratory work was carried out under

these conditions. Subsequently, two types of monochromators were used:

a Jarrell-Ash Model 82-415SP 0.25 meter monochromator and a McPherson

218 0.3 meter monochromator.

0.25 meter monochromator

The Jarrell-Ash monochromator was used during the major part of the

development of the pulse radiolysis system. It was also used for the

preliminary work on the excited atomic iodine state at 206.2 nm.23 The

size of the monochromator, 17.7 x 22.1 x 21.4 cm, along with its specifi-

cations (Jarrell-Ash Division, Fisher Scientific Co.) made it attractive

for use with the close geometry dictated by our laboratory space. The

grating used has 1180 G/mm with 3.3 nm/mm linear dispersion and

effective aperture of f/3.5. A wavelength motor drive of 100 nm/minute

was coupled to the wavelength drive knob to aid gathering spectra on a

Photovolt Corporation Model 43 chart recorder.

Three modifications had to be made on the original design of the

monochromator to adapt it to the resonance lamp, reaction cell, optics,

and photomultiplier housing.

First, the slit arrangement provided with the monochromator could

not be directly adapted to our purposes so a new arrangement was designed,

Figure 1. The flange fits into the hole where the original slit flange

fit into the body of the monochromator, and is bolted to the monochromator

body making an o-ring seal. The slit bodies were designed to be screwed

into the flange using a tool which fits into the two small holes in the

slit body (front view of the slit body). The slits and the flange were

threaded with forty threads per inch to allow small changes in position

per half revolution of the slit body so the best possible focus could be

achieved. In order to preserve the integrity of the threads, the slit

bodies were made of brass and the flange was made of aluminum. A Helium-

Neon laser was used to help determine the best possible optical focusing

arrangement. The entrance slit flange could be aligned and secured to

the reaction cell body by its four through holesand four bolts on the cell


Next, in order to conveniently look at the emission spectrum from

the resonance lamp while directly coupled to the monochromator, an

adapter flange was designed, Figure 2. The flange had a hole (not

pictured) threaded into the flange cavity which, when combined with a

screw whose threads has been ground down on one side and an o-ring for a

washer, served as a flushing outlet port. The flushing system will be

described later.

Finally, another adapter flange was built to allow direct coupling

of the Pacific Photometric Instruments Model 3262 photomultiplier tube




slit holder

Fig. 1 Entrance and exit slit system flange adaptation to the
Jarrell-Ash monochromator. Slit widths of 0.36, 0.56,
0.71, and 1.6 mm x 17 mm high were used.




adapter flange

slit holder

Fig. 2 Coupling arrangement from the lamp flange to the entrance
slit flange (Fig. 1).

lamp flange

(PMT) housing to the exit slit flange, Figure 3. The PMT housing screwed

onto the outside threads of the flange. For proper alignment the adapter

flange through holes matched those of the exit slit flange. Initially, the

flange also served as a window holder. A sodium salicylate coated

window could be used for detection in the VUV; however, this was never

used successfully since the monochromator could not be flushed sufficiently

for transmission of VUV radiation through the monochromator to the


Light pipe system

After initial operation of the system with the PMT directly coupled

to the monochromator, it was quickly realized that the X-ray (X-radiation)

and the radio frequency (RF) noise associated with the electron pulse

and the irradiation induced fluorescent burst from the optics was not

attenuated enough to keep the PMT photocathode from saturating and

overloading the preamplifier preampp) and the dual operational amplifiers

resulting in a dead-time8,81 (recovery time) of tens of microseconds.

Removal of the preamp from the signal circuit did not prevent the subsequent

overload of the op-amps. The use of a window with or without a sodium

salicylate coating in the PMT adapter flange did not aid with radiation

protection of the PMT photocathode.

It was known from the literature that displacement of the PMT away

from the reaction cell and electron source area,8,62,81 along with extensive

lead shielding,82,83 would greatly aid in reducing the pulsed radiation

interference and subsequent saturation of the photocathode. However,

due to the small laboratory space it was decided to keep the PMT and the

monochromator in the same room as the Febetron 706. There was enough



quartz window


J adapter flange

Fig. 3 Exit slit flange adapter flange and widow mechanism for
connecting the exit slit flange to the PMT housing, Model
3262, for end-on PMTs. The window, when used, is
sandwiched between o-rings. The collar is tightened down
with a special tool to make the o-ring seals.

space in the room, though, to displace the PMT about one meter from the

monochromator, which would remain bolted to the reaction cell.

The idea of a light pipe was first tested using Pyrex glass rods.

For work in the UV region, though, quartz rods were necessary. The

throughput was found to be favorable, but the quartz rods produced an

unfavorable flourescent pulse upon irradiation by the X-ray pulse.

Standard 8 x 4 x 2 inch lead blocks were used to make a lead cave for

the rods and the PMT housing.82'83 This reduced the amount of flourescence,

but it was not reduced to a satisfactory level. The X-rays which came

through the unprotected areas and the backscatter were enough to make

ordinary optical grade quartz fluoresce.

Heraeus-Amersil Suprasil 2 radiation grade quartz rods, 3 mm diameter

and 3 feet long, were then acquired to serve as a relatively fluorescence

free light pipe. An apparatus was designed for coupling onto the exit

slit flange, aligning and housing the fragile rods, and for maximizing

the light level at the PMT window, Figure 4. With this arrangement the

dead-time of the detection system was reduced to less than 1.5 microseconds

and thus allowed useful initial microsecond time resolution. The first

definitive work on the excited atomic iodine state, at 206 nm, was
carried out using this geometry.2

Figure 4 depicts the quartz rod light pipe, housing, and guides.

Nylon guides A, B, and F are detailed in Figure 5. In operation, B type

guides were bolted onto A and D in order to center the stainless steel

(SS) oval housing, C. The housing provided both protection for the rods

as well as added rigidity to the apparatus over its 27 inch length.

This arrangement also allowed sheets of plumber's lead to be wrapped


1- r-


I ': 1

-0 a
-- C -0
1- 3r --
ag .

) r-

O 0
0 a1S -
*r-- (

'- ,-
.C -0 r
-- 0-

4-) ro
I- C-

0 Cr C

>- C X


U 0 (3
*O- -4

Sr- .0

a) 0 *.-


C C 4-


-0 C

3 Si C

e > LL.

---o -
E C)

a 0 r
C ro -
(0 > LL.


4-' -
3 C E-
Cn o 0E

0 X0

=3 C3 -

C rm E m
.C ^r- EO

*.- O


N *- X
s.- (NJ o E

rrco Eu

Fig. 5 Nylon guides. A, nylon light pipe guide, with threaded holes
for centering the quartz rods in the PMT housing; B, nylon
guide, with thru holes, for centering the 27 inch SS oval
housing; F, nylon light pipe guide, for centering the quartz
rods in the exit slit flange.

around the housing to aid with radiation protection without putting any

mechanical strain on the rods themselves. Nylon guide A aligned the

rods at the PMT end and centered the light pipe onto the PMT window due

to its designed snug fit into the PMT housing.

Flange D was designed to adapt extender E to a B type guide. The 6

inch extender E had been previously designed to extend the PMT housing

from the exit slit window flange adapter (Figure 3). The purpose of the

extender had been to allow better shielding of the photocathode, but

this, by itself, had proved inadequate. It was therefore integrated

into the apparatus because of its strength and convenient threaded

coupling mechanism. Plumber's lead could also be wrapped around the

extender for additional radiation shielding. The exit slit flange

adapter flange (Figure 3) was used without the window and collar mechanism.

As before, it coupled E to the exit slit flange. Nylon guide F aligned

the rods and centered them onto the exit slit.

The total geometry was designed to allow the rod bundle to make

contact with the exit slit and have up to one inch variability from the

PMT window. In general, the rods were kept about 2 mm from the PMT

window and made contact with the exit slit.

Finally, a lead slug, Figure 6a, was added to the PMT window side

of guide A. The addition of this X-radiation shield halved the previous

radiation burst intensity at the photocathode and brought the detection

system to a net dead-time of less than 1.5 microseconds. This is the

best that could be achieved with this .type of geometry using only lead

shielding. Further reduction of the noise interference characteristic to

pulse radiolysis systems would necessitate RF shielding of the e-beam,

signal lines, and all the related electronics. Optimum noise reduction


a) 2 1/16 inch diameter x t/2
PMT housing, Model 3262.

inch thick lead shield for end-on

b) 1 1/4 inch diameter x 1/2 inch thick lead shield for side
window PMT housing, Model 3150.

Fig. 6 Lead shields for PMT photocathodes in the Pacific Precision
Instruments PMT housings.

could be best achieved by removing all the related electronics, except

the analyzing light source, from the room which houses the e-beam

source.8,62,81-83 Either or all of these methods would have to be

employed for submicrosecond resolution.

0.3 meter monochromator

Studies in the VUV could not be carried out until a McPherson 218

vacuum monochromator was substituted for the Jarrell-Ash. This mono-

chromator has an f/5.3, a 2400 G/mm grating and a 1.33 nm/mm linear

dispersion. It also had to be adapted to the reaction cell and optical


Figure 7 shows the single-pass reaction cell arrangement for studies

in the VUV. The single-pass cell and the lamp system were the same as

used with the Jarrell-Ash monochromator. The extension flange, EF, was

designed for three purposes. First, it had to provide enough space for

the lead cave, PB, to be built around the exit slit and photomultiplier,

P, area. It also had to contain a sliding lens holder used for mounting

a condensing lens. Focusing onto the monochromator entrance slit was

achieved by sliding the holder along the inner walls, Figure 8.

A coupling flange, CF, was designed for dual purpose. A two inch

diameter by one eighth-inch thick Suprasil 2 window was epoxied to the

reaction cell side of the flange. The window separates the reaction

cell from the entire monochromator vacuum system. The flange was bolted

between the reaction cell and the vacuum extension flange, Figure 7.

This linear geometry aligned the lamp axis, through the reaction cell,

with the entrance slit of the monochromator. Fine alignment and focusing

was carried out with a He-Ne laser.



E a)OL
0CL 4-a)

0 a


o CrC)

*u c a) 01 4)

S- + 0a)-
-) ~-

E 0o 0
0 C- -4 0)
cc E n= 0

0 0 a) 0 o 0O


o -

3 Lua
C4-' C

a) C C C a) C-

0/r Lr) *iWCE -1

U/ +>r -E


O 0
S 0

0 E

00- 00








I I,
L '-

C'- 0

0 "


l I

S I I I 1

I |


0 r-
o q--



Next, an exit window flange was designed which would align the

Pacific Precision Instruments side window PMT housing Model 3150, used

with this geometry, onto it and also bring the exit window to near

contact with the PMT window. This close geometry minimized the atom-

spheric absorption path length between the exit window and the PMT

window. Figure 9 shows the exit window flange. The inset in Figure 7

shows how the alignment was made. The one inch diameter by one eight-

inch thick Suprasil 2 window was epoxied into the flange and extended

1/32 inch out from the flange. The PMT housing (Model 3150) butts

against the main flange window. This geometry ensured reproducibility

of the geometry and proper alignment, and also kept both windows from

contacting and damaging each other.

The monochromator vacuum pumping system design, Figure 10, was copied

from the pumping system of another McPherson 218 monochromator in this

laboratory. A useful modification which was made was to enlarge the

liquid nitrogen well. This extended the time between necessary refilling

of the well, for proper operation, from twenty minutes to one hour.

With the main and backing valves closed and the roughing valve

open, the Kinney KC-5 vacuum compound pump will rough pump the monochromator

directly. Once the pressure is below 1 torr, the roughing valve is

closed and the backing valve is opened to rough pump up to the main

valve. With the pressure below 1 torr the liquid nitrogen well is filled.

After about ten minutes the main valve is opened. When the monochromator

pressure is about 0.1 torr the oil (Dow corning 704) diffusion pump is

turned on. The diffusion pump body was cooled by flowing water filled

coils. An automatic safety thermostat switch, in series with the diffusion

pump power cord, monitors possible overheating of the diffusion pump in

exit window


Fig. 9 McPherson 218 exit window flange and window.

to monochromator

main valve


to kinney
high vacuum pump

backing valve

Fig. 10 High vacuum pumping system for the McPherson 218 monochromator.

a loss of coolant accident. The pumping system provided adequate vacuum

for work down to the 160 nm cut off of the windows, lens, and PMT window.

The vacuum extension flange was evacuated directly through the entrance

slit of the monochromator. In order to speed up its evacuation time,

the slit was opened to near its maximum width of 2 mm and closed down as

needed after a useful vacuum had been attained. Active pumping was

always maintained during an entire experimental session.

B. Optical Efficiencies

0.25 meter monochromator and flushing system

Figure 11 shows the linear geometry used with the Jarrell-Ash

monochromator and quartz light pipe system. Basically, the light from

the lamp, L, traverses the single-pass reaction cell, C, and the monochromator,

M, in an overall linear geometry. At the exit slit the quartz rod light

pipe bundle, Q, carries the light to the photomultiplier window.

Although a simpler optical train could have been used for studies

above 200 nm, Figure 12, this tight geometry had desirable qualities for

studies down to 180 nm. For studies of the ground state iodine atom

useful throughput of the 178.3 nm light would have to be attained. In

order to achieve a useful throughput below 190 nm, without a vacuum, the

monochromator and PMT system would have to be purged with nitrogen or


Nitrogen purging was first attempted. The light pipe and the

reaction cell were omitted for initial testing of the purging system.

The lamp was coupled to the monochromator via the flanging mechanism

shown in Figure 2 and the PMT housing was coupled to the exit slit

flange shown in Figure 3. The PMT housing adapter flange was first used

Fig. 11 University of Florida Pulse Radiolysis System

B Biomation 610B
C Reaction chamber
DO Dual op amp
FP Febetron power supply
G Sample handling system
HV PMT power supply
L Iodine lamp
LS Lamp flow system
M Monochromator
MG Microwave generator
P Photomultiplier
PB Lead cave
PR Printer
Q Quartz rod light pipe
S Oscilloscope
TR Trigger source
TY Teletype
V Video


I m

PMr I I|

Fig. 12 900 optical configuration using a 10 or 7.5 cm focal length
parabolic mirror. This configuration focuses the light
onto the PMT window while keeping the PMT window from
facing the X-ray pulse directly.

L Light source
M Monochromator
P Photomultiplier tube
PB Lead cave
PM Parabolic mirror
C Reaction cell

without a window. A SWAGELOK eighth-inch pipe to quarter-inch tubing

fitting (SS-400-12) was put into the side of the monochromator to serve

as an inlet port for the purging gas. The monochromator cover and

potential areas for leaks were sealed with Apiezon Q sealing compound.

The slits served as flushing outlet ports for the monochromator. Since

the lamp adapter was being used, the half-threaded screw served as one

outlet port on the entrance slit side. The exit slit flushed nitrogen

into the PMT housing where it leaked to the outside. Generally, fast

flushing was first done followed by several hours or days of slow flushing

while the lamp spectra were periodically taken. No useful gain in light

throughput was achieved nor useful wavelength extension below 190 nm.

Unfortunately, at one point nitrogen purging of the PMT area caused high

voltage breakdown of some of the components of the PMT base. The resistors

had to be replaced due to the damage incurred from arcing.

A sodium salicylate coated window was then used in the exit slit

adapter flange (Figure 3) as a fluorescent screen for detection of VUV

radiation. Various methods of deposition and thicknesses of the sodium

salicylate were tested. Access to wavelengths below 190 nm was not

achieved, and a lowering of throughput was observed since the isotropic

nature of the induced fluorescence greatly reduced the intensity at the


Extensive helium purging was then attempted. The helium was first

passed through a sodium hydroxide trap, followed by a large molecular

sieve-silica gel trap, and finally passed through a smaller molecular

sieve trap which was immersed in a liquid nitrogen filled dewar. The

sodium hydroxide was eventually removed from the system as its presence

did not show any useful effect. Various flushing rates were tested. The

system was purged for days and sometimes for weeks while the lamp

spectrum was taken at various times. Although some gain in transmission

of lower wavelengths was observed, as well as an overall gain in intensity

in other wavelengths, no net benefit for the experimental requirements

was achieved. Figure 13 shows the lamp spectrum with nitrogen purging,

Figure 14 shows the spectrum with helium purging.

It was apparent that the volume being purged was too large to

achieve optical efficiency below 190 nm. Because satisfactory throughput

could not be achieved with this close geometry, and because of the

expected parent compound absorption in the wavelength region of interest,

and since it was apparent that the linear dispersion associated with

this monochromator would not suffice for adequate line isolation at

useful slit widths, it was futile to pursue studies below 200 nm with

the Jarrell-Ash monochromator.

0.3 meter monochromator

For studies of the ground state iodine atom at 178.3 nm a McPherson

218 monochromator and vacuum system were borrowed from a UV-VUV spectro-

meter system in this laboratory. Subsequently, after it was determined

that it provided the experimental requirements, a McPherson 218 was

purchased for the pulse radiolysis system and was adapted as discussed


The geometry used with this type of monochromator is shown in

Figure 7. Light from the lamp traversed the reaction cell and was

condensed onto the entrance slit. After being dispersed the light of

the selected wavelength was detected by the photomultiplier. This

250 200 150 nm
250 200 150 nm

Fig. 13 Iodine discharge lamp spectrum with nitrogen flushing of the
optical path. Jarrell-Ash monochromator, 0.356 mm slits, and
50 watts microwave power.

250 200 150 nm

Fig. 14 Iodine discharge lamp spectrum with helium flushing through
NaOH, grade 42 silica gel, lOx molecular sieve, and a
molecular sieve-liquid nitrogen trap. Jarrell-Ash mono-
chromator, 0.356 mm slits, and 50 watts microwave power.
The vertical scale is twice that used in Fig. 13.

geometry could potentially be used for absorption spectrophotometry in

the visible, ultraviolet, and vacuum ultraviolet regions.

A 2 inch diameter plano-convex lens, LE, with a 5 inch focal

length was used as the condensing lens. The throughput from the discharge

lamp, L, provided adequate light intensity for good signal-to-noise

conditions for useful observations. Final studies of the excited state

and ground state iodine atom were carried out with this geometry.

The observable signal was also dependent on the strong absorption

by the 206 and 178 nm atomic iodine states. For work with species with

lower absorption coefficients a multiple-pass reaction cell system with

a useful net optical throughput was needed in order to achieve an

observable signal.

C. Reaction Cells

Two types of aluminum reaction cells were used: a versatile

single-pass cell and a multiple-pass cell. The single-pass cell was

used for the atomic iodine studies. The multiple-pass cell was used for

the ozone studies and the attempts to look for molecular iodine absorption

following pulse radiolysis of alkyl halides.

Single-pass reaction cell

The single-pass cell, Figure 15, had been previously designed in

another laboratory. The convenient multiport system made it very useful

for our purposes. The cell has six 2.25 inch flange ports and a body

cavity of ca. 800 cm3. Two of the ports were on the two large cover

plates. The cover plates made o-ing seals with the cell body. All

attachment flanges were designed with a short barrel in order to fit the


top view
with the
covers off

Fig. 15 Single-pass cell
are made at each
and the reaction

with six equal flange ports. 0-ring seals
port lip and between the two cover plates
cell body.


r i L,

ports and make o-ring seals with the port's lip. Viton o-rings were

used because of their resistivity to chemical attack.

The effective interaction volume was defined by the intersection

of the e-beam with the analyzing light beam. The primary e-beam has a

diameter of 2.5-3 cm, spreading only slightly, but there is also an outer

region due to secondary electrons which adds considerably to the effective

cross section. (The range of 100 keV electrons in air at STP is 6 inches.)84

The light beam enters the cell with a maximum diameter of 2.5 cm, but ca.

90% of the intensity is within a diameter of about 1.5 cm diverging only

slightly, due to the ca. 1 cm ID and collinear cylindrical shape of the

discharge lamp body.

The many detection methods which could be used with the single-pass

cell makes it a versatile apparatus for future studies. The analyzing

light may be used perpendicular to the e-beam and linear with the condensing

lens, as used in these absorption studies. The light traversed the

cell in a region where the energy deposition per unit volume was very

nearly homogeneous. Relatively uniform irradiation was insured since

the range of the electrons was substantially greater than the thickness

of the light beam and e-beam interaction region.185 For pressures

used in these experiments the interaction length is about 2-3 cm, while

the range of the electrons was about one meter.

Emission studies could also be carried out using this cell. Light

could be gathered from four different ports, all perpendicular to the e-

beam. Finally, the resonance fluorescence detection technique could be

used with this cell.86,87 In this very sensitive technique the light

source, e-beam, and the light gathering lens would need to be perpendicular

to each other. The multiport system would allow several geometric

variations of this principle.

Multiple-pass reaction cell

The multiple-pass cell was designed to help observe species with

relatively low absorption coefficients by providing a longer absorption

path length. For wavelengths above 300 nm, multiple-pass cells greatly

increase the signal-to-noise ratio (S/N).14 Below 300 nm the wavelength

dependency of mirror reflectivity plays an increasing role with decreasing

wavelength. Therefore, in designing a multiple-pass system the expected

increase in absorption must outweigh the reflective light losses. A

simple, 39.1 cm, six-reflection multiple-pass cell was designed for

studies in the visible and ultraviolet regions down to 200 nm.

A White cell is typically used to obtain a long path length in a

confined volume.13'14'62'88 Because of the damaging effects on mirrors

by the chemicals in the reaction cell and the expense of UV-VUV grade

White cell type mirrors and coatings, a plane mirror multiple-pass

system was used.

Figure 16 shows the multiple-pass cell optical arrangement. This

geometry was successfully used for studies of the oxygen-ozone system at

253.7 nm. For work in the VUV with this cell, vacuum window flange

adapters would have to be designed in order to couple the light source

and the monochromator to the cell. The monochromator and the PMT

arrangement was the same as that in Figure 7. The multiple-pass cell

was attached to the Febetron using the same flange that was used for the

single-pass cell.

Figure 17 shows the multiple-pass mirror configuration. The

incoming light beam traversed the cell and was reflected off the entrance

mirror which initially directed the light along the plane mirrors. The

exit mirror redirects the light beam through the exit window. The cell

j\ ,\ ,\
I \/
I *I ~




Fig. 16 Multiple-pass reaction cell


arrangement for visible and UV

Multiple-pass reaction cell
Sample handling system
Mercury lamp
Condensing lens
Lead cave



windows were 1 inch diameter by one eight-inch thick Suprasil 2.

The first surface mirrors were MgF2 coated aluminum (coating for UV

#7992, Oriel Corporation). The system pressure determined the e-beam

divergence, which in turn determined the actual interaction path length.

Reaction cell-Febetron 706 coupling flange

Both reaction cells were attached to the field emission tube window

face plate by a simple rotating flange mechanism, Figure 18. The

reaction cell flange fit through the attaching flange and was free to

rotate. Tension was provided by tightening the nuts on the Febetron

face plate bolts, Figure 19. As the attaching flange was tightened

toward the Febetron face plate it pressed the cell flange forward. The

cell flange was thus firmly pressed onto the o-ring on the face plate

and was no longer free to rotate. Four nuts and bolts were used to

secure the attaching flange in order to provide uniform tension. The

pitch of the cell could be changed slightly by differential tightening

of the nuts, i.e., when the upper nuts are tightened the back of the

cell is raised. Thus, with this simple flanging mechanism the cell

could be rotated and leveled for proper optical alignment with the rest

of the optical system.

The cell flange also served as the cell e-beam window holder,

Figure 18. A 1.25 inch diameter by 0.075 mm thick titanium window was

sealed to an aluminum disk with Varian Associates "Torr Seal." The

window disk rested on an o-ring which made a vacuum seal. Pressure was

applied to the disk by tightening the threaded ring. The ring was made

of brass and the flange of aluminum in order to preserve the integrity

1 1

window securing

aluminum disk and
e-beam window holder

reaction cell flange

attaching flange

L -I--

Fig. 18 Reaction cell-Febetron 706 face plate coupling mechanism.
The cell flange contains the e-beam window and mechanism.

attaching nut

face plate
and bolts

Fig. 19 Complete assembly of the reaction cell flange onto the Febetron
face plate. The reaction cells were attached to the reaction
cell flange.

of the threads. Care was taken to make sure that the window was not

deformed while the epoxy cured, so as to minimize any possible

defocusing effects by the window on the e-beam. The window was removed

from the field emission tube window enough to prevent possible damage to

the tube window from spallation which could occur from the electron

impact on titanium.

D. Spectrophotometry

Light source

The technique of absorption spectrophotometry requires a light

source, an absorption path, a monochromator, and a photomultiplier. All

other related equipment support these and aid in achieving an observable


The light source used could be a standard commercial lamp suited to

the spectral region of interest or it could be of novel design. Since

the signal-to-noise ratio is directly proportional to the square root of

the light intensity, it is important that the source be relatively

intense. High power spectral lamps that emit a continuum distribute

their over-all power over a spectral region. Their intensities per

wavelength interval may not be great enough to carry out studies of

discreet line absorption. High power light sources, unless filtered,

could also photolyze the chemical system under study and complicate the

observed signal. Less expensive low-power line sources may often be

more intense over a narrow spectral band of interest than high power

continuum lamps. Usually in order to produce radiation at discreet

wavelengths an appropriate atomic or ionic state must be produced. Low

power (less than 125 watts) electrodeless discharges have been shown to

produce intense radiation primarily from neutral atoms.89

There are various methods of coupling microwave power to sustain a

discharge, but the most effective method is through a tuned cavity.90

A tuned cavity microwave discharge light source was therefore chosen for

the kinetic spectrophotometry system. The lamp system will be detailed


Absorption path and monochromators

The reaction cells described previously, coupled with the Febetron

706 high pulse-current irradiation system,produced concentrations of

transient species in the cells which were high enough to allow direct

observation by absorption spectrophotometry.

Of the two monochromators used, the McPherson 218 best exemplified

the versatility needed for spectrophotometry over a varying wavelength

region. A major factor in assessing a monochromator is its light

throughput. The light gathering power, LGP, best assesses this quality.17

LGP is defined by

LGP = (h/f2) m, II-1

where h is the slit height (mm), f the effective aperature (f#), and m

the linear dispersion (nm/mm). An LGP of about 0.3 or greater has been

recommended for fast kinetic spectrophotometry.7 For this monochromator

f = 5.3, m = 1.33 nm/mm, and h = 6 or 8 mm. Thus, the LGP for this

monochromator, as used, was either 0.28 or 0.38 nm.


Photodetectors used in fast absorption spectrophotometry must be

sensitive enough to measure small changes of light intensity in the

spectral region of interest, and they must have a fast frequency response

relative to the half-life of the transient of interest. Two types of

photomultipliers were used under different conditions with the two

different monochromators. Two EMI end-on 50 mm window PMTs Models

9750QB and 9856B were used with a Pacific Precision Instruments Model

3262 housing, Figure 11. Because of its lower noise sensitivity, an EMI

Model 3150B 39 mm side window PMT in a Pacific Precision Instruments

Model 3150 housing was eventually used for the spectrophotometry system,

Figure 7.

Initial studies in the UV employed the Model 9750QB PMT. The

window was Spectrosil quartz with an optical range of 165-650 nm. The 2

mm thick window had 90% transmission down to 190 nm but dropped to 50%

by 170 nm. The bialkali photocathode had a quantum efficiency of 22% at

200 nm. The 9856B PMT was used for tests in the visible region. It had

a soda-lime window with a spectral range of 320-650 nm. The 2 mm thick

window had 88% transmission down to 350 nm and fell off to 38% at 320

nm. The maximum quantum efficiency of 20-28% was between 470 to 400 nm.

Standard photomultiplier tube base electronics have been customized

in order to achieve desired rise times, gains, and linearities in their

respective systems.1791'92 Figure 20 shows the diagram of the base

components used with the end-on PMTs. Although this circuit provides

linearity under usual spectroscopic conditions, it deviated from linearity

after exposure to the large electromagnetic radiation pulse. In order

to maintain dynode voltages when high currents were drawn, capacitors in

the range of tenths to several microfarads were needed. It can be

seen in Figure 20 that the capacitors were relatively small in value and

also, that they were not on all the dynodes. Also, when high currents




L) 01
o" -4


ot- C
C",) U,,



0 3


(n 4-)

E 3

'-4 rO
> '-4
S- S- LI) oLL S-
4-) 4-) 0 --
*.- *r- 4- CO

r- r-- r
UL) V)

< 0 c /)
O /


00 0ra










r-- C







were drawn the voltage drops could be substantial across the large

resistors. Both of these factors combined to produce nonlinearity in the

signal, especially during the first few microseconds after the radiation


The 9783B side window PMT was acquired to help alleviate some of

these and other problems. The side window PMT was used in all the

spectrophotometry geometries, after its initial tests, because of its

more favorable properties under the harsh electromagnetic interference

pulse conditions. The PMT window was 2 mm Spectrosil quartz. The

quantum efficiency reached a maximum of 22% at 300 nm and fell off to

18% at 200 nm. Figure 21 shows the diagram of the base components. The

value of the capacitors on the last stages was increased and the resistor

values were decreased relative to Figure 20. This combination greatly

helped reduce, but not eliminate,, nonlinearity after the radiation

pulse. Further improvement could be achieved by using capacitors to

stabilize the voltage on each dynode.17'91 Also, since each dynode

contributed to the nonlinearity, the minimum number of dynodes necessary

for a useful signal should be used.92 This PMT used one less dynode in

the circuit than the other two PMTs.

High beam-current machines, such as the Febetron 706, produce large

electromagnetic disturbances which interfere with the PMT circuitry.14

Although the interference is initiated during the e-beam pulse, the

overload and subsequent recovery could last several microseconds.

Several procedures were used to reduce noise which affected the PMT and

its circuitry. The electrons were always absorbed in and by the reaction

cell and returned to ground. The RF interference was reduced equally




S- 0
U O) I--i


-- r
> a) c

- c O--

> LO C:
n in c

*, .0- -L *:

r. C 4-

U U S- w-

: C











O r-








a) -
0 .C

whether the reaction cell was directly grounded or grounded through the


Cerenkov radiation is not a problem with either reaction cell since

the light paths in both cells are transverse to the e-beam. The Cerenkov

light angle is very small and thus travels in nearly the same direction as the

e-beam. Collinear optical systems suffer more from Cerenkov light

interference than transverse optical systems.

The X-ray pulse is by far the largest contributor to interference

and subsequent PMT overload. Shielding of the X-rays with lead greatly

attenuated the pulse intensity at the PMT photocathode. Unless the

exposed photocathode is removed several meters from the Febetron area

behind proper shielding, gross overloading of the PMT will occur. While

using typical experimental geometries, tests showed that encasing either

PMT in lead bricks and masking the photocathode with a half-inch lead

slug produced a low intensity noise pulse of approximately 80 ns duration

and 30 ns FWHM. In typical operation, though, where the photocathode

was exposed, microseconds noise pulse durations were observed due to the

higher intensity of the X-ray pulse at the photocathode. The McPherson

218 monochromator offered some advantage toward reducing the X-ray

problem since the photocathode did not face the reaction cell area

directly as did the PMT geometry with the Jarrell-Ash monochromator.

Backscatter interference, though, was still evident and substantial.

E. Lamp Systems

Because the signal-to-noise ratio is proportional to (I/bw)2,

where I is the light intensity at the photocathode and bw is the amplifier

and circuitry bandwidth and since the bandwidth is usually large and

fixed by the electronics, S/N may be increased through increasing light

intensity, as long as PMT linearity is maintained.

The basic factors involved in designing the lamp system were size,

power, intensity, emission spectrum, and stability. These factors were

carefully considered to ensure applicability for the chemistry being


Iodine lamp

In order to observe a signal for low concentrations of transient

species by fast absorption spectrophotometry, a stable intense light

source7 rich in the wavelengths of interest was needed. Hartek et al.

first developed an electric discharge iodine lamp.93 An argon carrier

flow system was used, as it was found to reduce buildup of impurities in

the lamp. Strong emission from OH radical bands showed that water

vapor, when present, was the most troublesome impurity in their system.

Since the vapor pressure of iodine was high, but less than 1 torr at room

temperature, it was ideal for use in a discharge lamp. Similar intense

atomic line sources using microwave excitation of flowing mixtures89

including iodine-argon mixtures36-38 were eventually developed.

For our purposes an iodine lamp system was developed and powered by

a Raytheon Model PGM-1OX1 microwave generator, Figure 22. The continuous-

wave microwave energy at 2450 MHz was generated by a magneton oscillator.

A Microwave Devices Incorporated Model 752.3 bi-directional RF power and

standing wave ratio (SWR) meter was attached to the output coupler on

the front panel of the Raytheon. The front panel also had a relative

S1 S2

gas tank


a) Iodine lamp flow system.

outlet t=i_
to trap

to lamp


sieve trap

c polyethylene

b) Iodine lamp. SWAGELOK connections couple 1/4 inch
quartz tubing to 1/4 inch polyethylene tubing.
The knurled brass coupler attaches the lamp to the
lamp flange (Fig. 2).

Fig. 22 Lamp system.

power meter which indicated a percent-of-power output. Maximum power

output was about 85 watts, with less than 1% ripple. In typical operation

only 50 watts was used. An RG 214/U flexible coaxial cable carried the

microwave energy from the meter to an Evenson cavity.90

As previously mentioned, tuned cavities are the most efficient
method of delivering microwave power to sustain a discharge.90 Figure

23 shows the Opthos Instruments Company Evenson-type cavity used in

these studies. The cavity was mounted onto the lamp by sandwiching the

lamp body between its main body and the bottom plate. When the standing

wave ratio was maximized the most efficient power coupling to the plasma

was achieved. The bi-directional power meter was used to ascertain a

properly tuned cavity.

With the lamp system assembled and with a stable gas flow the power

was turned to about 50-60 watts. The discharge was initiated by a Tesla

coil, which provides initial ionization of the gas mixture. When the

pressure in the flow system was below 5 mm Hg,a discharge was sustained.

The screw adjustment was turned, as required, to minimize the

reflected power (maximize the delivered power) indicated by the power

meter. The sliding tuner was also adjusted to further reduce the reflected

power. In general, intense and narrow atomic emission lines were best
produced when the pressure in the flow system was kept around 1 torr.3934

Once the cavity was tuned it remained in good adjustment for several

days if left undisturbed. Once tuned, the SWR could be maximized daily

in a few seconds.

Under these conditions, and with proper caution, the microwave

equipment posed no hazards to laboratory workers. It should be noted,

tuning screw

cooling air port

variable sliding tuner


bottom plate

Fig. 23 Evenson cavity. The lamp body was designed to fit between
the cavity body and the bottom plate. Proper tuning of
the cavity is achieved with both variable tuners.

to power


however, that these types of discharges are rich in UV and VUV radiation

and thus proper eye protection should be worn.

The lamp flow system was constructed of Pyrex and polyethylene

tubing, Figure 22. Fisher and Porter teflon stopcocks and Ace Glass

Incorporated teflon stopcocks with Ace Glass Incorporated brand "FEFTE"

o-rings were used. Pyrex joints #15 were used in order to facilitate

coupling of different segments of the system, as well as to facilitate

disassembly and cleaning. One quarter-inch OD, polyethylene tubing, was

used to allow the greatest mobility when coupling the lamp to the reaction

cell from the nearby vacuum rack.

The gas flow system consisted of a three stage pressure reduction

area between the argon carrier gas tank and the molecular sieve trap. By

adjusting stopcocks S1-S3 (Figure 22), the pressure was reduced from

just above atmospheric to about 1 torr by a constant pumping rate with a

Welch Duo-Seal mechanical pump. On the high pressure side of the molecular

sieve trap, the approximate pressure of"the gas flowing through the lamp was

indicated by a Kollsman Instrument Corporation absolute pressure gauge.

The lOX-molecular sieve prevented fast back diffusion of molecular

iodine into the guage. In operation, six inches of the molecular sieve,

in 10 mm OD thin wall Pyrex tubing, would become saturated in one month.

The trap contained about fifteen inches of the sieve. When one third of

the trap was contaminated, that third was removed and the trap was

refilled from the opposite end. The contaminated sieve was not regenerated,

but instead was properly disposed.

At room temperature the vapor pressure of iodine was less than 1 torr.

With stopcock S4 open, a constant and stable source of molecular iodine

was provided from the reservoir to the lamp area. The length of the

glass and polyethylene tubing from the iodine reservoir to the inlet of

the lamp was about one meter. Proper mixing occurred over this length.

Quarter inch stainless steel SWAGELOK fittings were used to make all the

glass to polyethylene tubing connections.

The outlet of the lamp went to a Pyrex 29/42 trap in series with

the mechanical pump. A carbon dioxide-isopropanol bath was used to

efficiently sublime and trap the iodine. The argon carrier gas passed

through the trap and was pumped away. The pump was exhausted directly

to the outside of the laboratory building.

The lamp body eventually used in the majority of the alkyl halide

studies was made of a #22 glass o-ring joint, a two inch Vycor graded

seal, and a quartz main body, Figure 24b. The inner tubulation is also

quartz. The geometry was designed such that the Evenson cavity mounted

over the quartz-rich side of the graded seal. Quartz was needed because

the local heating near the cavity could deform regular glass. Local

cavity cooling was achieved by rushing filtered air through the Evenson

cavity cooling port, Figure 23.

Figure 24a shows the type of microwave lamp which previously had

been used for photochemical work in this laboratory. The lamp body was

made the same as previously described. The front tubulation, though,

was made of glass. This lamp was first used for this work, but several

disadvantages were quickly realized.

A teflon collar, TC, was fitted over the neck of the o-ring joint.

A brass coupler, BC, screwed down onto a flange, which was bolted onto

the reaction cell, Figure 25. At the same time the coupler pressed on




a) Original lamp body
TC, teflon collar
BC, brass coupler

outlet quartz body

t inlet

b) Improved lamp body.

graded seal


Fig. 24 Microwave discharge lamp bodies.

brass coupler

teflon collar

reaction cell and lamp


window holder


tightening screw

Fig. 25 Lamp window and coupling flange. The lamp body is coupled
to the flange by the brass coupler and the teflon collar.
The lamp window is sandwiched between o-rings and kept
snugly in place by the pressure provided on the window
holder by the tightening screw.

I !


the teflon collar which centered and tightened the o-ring joint and made

the vacuum seal. A Heraeus-Amersil Suprasil 2 window, which was sealed

by o-rings inside of the flange, separated the lamp system from the

reaction cell cavity. This window system was preferred to the sealed
type where the window is epoxied directly to the lamp,89 since window

material could easily be varied for different spectral ranges and conditions

without manufacturing another lamp body. It also facilitated cleaning

and annealing of the windows for optimum performance.95

The problems encountered with this first lamp were twofold. First,

it may be seen from Figure 24a that the Evenson cavity could not be

placed closer to the window than the region directly behind the front

tubulation, about 10 cm. In the 0.5-1 torr pressure range, the discharge

plasma varied from about 5-2 cm, respectively, in total length. Second,

two layers (volumes) develop in the lamp which are important. The

principles involved with these types of lamps have been extensively

discussed elsewhere.96 An emitting layer, powered by the microwave

cavity,contains radiating and absorbing atoms. Absorbing atoms are

removed by diffusion to the walls or possible chemical reactions. A

reversing layer, which contains absorbing atoms, develops in the volume

between the discharge and the lamp window (the volume in the o-ring

joint area). The larger the reversing layer, the more self-absorption

and subsequent reduction in transmitted intensity. The diffusion rate,

deactivation rate, and the efficiency of recombination at the walls

determine the concentration gradient of the absorber atoms. Recombination

efficiencies on glass and silica surfaces for ground state iodine atoms
are high,97 and thus are helpful in reducing the ground state absorber

atom concentration. Ground state iodine atom recombination with argon


as a third body is slow.98 Therefore, a substantial amount of ground

state iodine atoms capable of absorbing light emitted from the discharge

can be expected near the axis of the lamp rather than near the walls. A

much higher concentration of ground state atoms is expected than any

other state. Minimizing the reversing layer would therefore be advantageous.

The steady state concentration of the excited state atom, I*, was expected

to be very small due to its rapid deactivation by molecular iodine.38

Clear synthetic fused silica windows, Suprasil 2, were used in the

lamp flange because of their high discoloration and degradation resistance

to VUV, X-ray, and electron radiation. Suprasil 2 is specified to have

90% transmission at 200 nm, 85% at 170 nm, and 20% at 160 nm, for a 1/8

inch thick window. Suprasil 2 is practically free of fluorescence. As

previously discussed photomultiplier tube saturation and overload from

window fluorescence after electron and X-ray bombardment had been a

gross problem in the early stages of development of the pulse radiolysis


Even with the high radiation resistance of the window, it was

impractical to run the discharge against the window in order to minimize

the absorbing layer. The o-ring on the lamp side of the window could

become a source of contamination upon exposure to the discharge. With

sealed lamp systems the epoxy becomes a source of contamination when
exposed to the discharge. Thus, there is no practical way of eliminating

the reversing layer, but it was possible to reduce it from the size that

it was with the lamp in Figure 24a.

The long reversing layer, the efficient atomic recombination on

glass and silica, and the temperature differential between layers may

have been the causes for substantial deposition of thin films of iodine

and iodine oxides on the walls and window in front of the discharge region.

This region was much cooler than the region near the plasma, and once

condensed, iodine (and its oxides) would remain for long periods during

regular operation. Overnight pumping of the lamp system reduced the

amount deposited on the walls, but this cleansing process was not practical

for normal operation. This posed a serious problem since the transmission

of the window was markedly reduced after several minutes to a few hours

of routine operation. Switching the direction of the flow so that the

outlet was in front, Figure 24a, helped increase the time before substantial

window contamination occurred by providing direct pumping between the

window and the plasma, but this did not solve the problem.

Figure 24b shows the lamp which was designed to eliminate several

of the disadvantages of the first lamp. An annular tubulation was used

for the outlet port. This provided active pumping on the lamp axis to

help remove substances from the front of the discharge. The Vycor

graded seal was blown as close to the o-ring joint as possible. Since

there was no front tubulation, the Evenson cavity could then be butted

against the brass collar. The center of the cavity was then about 4 cm

from the window. With this setup the discharge extended to within 1.5

cm of the window. The path length of the reversing layer was greatly

reduced and the window region remained warm during operation and thus

prevented gross deposition of iodine. This helped solve the physical,

and some of the chemical, problems.

With the new lamp the spectral line intensities were increased and

contamination of the window was noticeable only after several days of

routine operation. As with the previous lamp, the entire lamp system

had to remain relatively oxygen free. Iodine oxides (possibly 102 or

1204 and 1205) readily formed in the discharge area and were sources of

window contamination. Water vapor was a constant contamination in the

lamp system, although its emission around 310 nm was not a problem in

the spectral region of interest. The water vapor did not cause any

instability in the operation of our lamp.93 A relatively constant light

intensity was reached within seconds of the plasma formation and

would remain stable for at least ten minutes. In practical operation

the lamp was rarely operated for more than one minute during one pulse


This lamp provided very stable, intense line emissions. It had

several advantages not found in other lamps used heretofore. First, it

reduced the length of the absorbing layer in the lamp. This led to

higher intensities. Also, the iodine reservoir stopcock, S4, did not

have to be open very much to provide sufficient molecular iodine for

useful intensities. Thus, it allowed use of the lower parent to carrier

gas ratios, which helped minimize possible parent contamination of the

lamp window. The flow system pressure was kept at ca. 1 torr to provide

an intense and stable discharge. Finally, it allowed the quick exchange

and cleaning of the windows and the lamp body.

Figure 26 shows the spectral region of interest scanned on the

McPherson 218 0.3-meter monochromator with 0.15 mm slit widths. Line

reversal could not be determined at this slit width.96 In operation the

optical system was always optimized for maximum throughput.

The excited atomic iodine state, 52P was easily studied using the
intense 206.2 nm line. Efficient absorption of this line along with its

intensity made it the most useful for studying the excited state atom.

a. 6P3/2 521/2 206.2 nm

b. 64P1/2 52 1/2 187.7 nm

a c. 64P3/2 21/2 184.4 nm
d. 64P5/2 52 3/2 183.0 nm

e. 62P1/2 52P1/2 179.9 nm

f. 62 3/2 52 3/2 178.3 nm

b c



Fig. 26 Iodine emission spectrum from the microwave discharge lamp
using a sodium salicylate fluorescent screen to cut off
stray visible light from other orders.

Efficient absorption of the 178.3 nm line made it the most desirable for

studying the ground state iodine atom.37'38 It was for this reason that

efficient VUV spectrophotometry was necessary. Useful absorption signals

were not detected for the other lines with either monochromator or

optical system.

Although the lamp emitted light in the VUV, several lines below 200

nm could not be detected until a vacuum monochromator was used. The

importance of the vacuum monochromator can be seen in Figures 27 and 28.

The spectrum is the same as Figure 26, but with different resolution and

scanning speed and with no fluorescent screen. It can be seen that

after ten minutes of active pumping good optical transmission was achieved

at wavelengths below 180 nm. In operation, the monochromator was pumped

for one hour prior to a set of experiments, at which point useful

transmitted light at 178.3 nm could be observed. In Figures 26-28 the

reaction cell was evacuated.

A major reason an intense light source was needed was due to the

expected parent alkyl iodide absorption in the spectral region of interest.

Figure 29a shows the lines b-f under high vacuum in the reaction cell.

When 2.5 torr of methyl iodide was introduced into the reaction cell,

Figure 29b, immediate loss of intensity for lines d-f was observed.

Figure 29c and 29d show the line intensities returning for lines d and e

after thirty seconds and five minutes, respectively, of active pumping

of the reaction cell. The full intensity of line f did not return until

after about ten minutes of active pumping. Figure 30 shows the absorption

bands for 2.5 torr methyl iodide. The lamp system, though, as designed

and optimized was able to transmit useful light intensity for work with

the 206.2 nm line using 10 torr or less of alkyl iodide parent compound

210 200 190 180 170 nm

Fig. 27 Iodine discharge lamp spectrum after 5 minutes of active
pumping on the McPherson 218. 50 watts microwave power
and 0.4 mm slits.


210 200 190 180 170 nm

Fig. 28 Iodine discharge lamp spectrum after 10 minutes of active
pumpimg on the McPherson 218. 50 watts microwave power
and 0.4 mm slits.

d a) b)


c) d)

Fig. 29 Chemical compound absorption interference in the spectral
range 170 to 190 nm (Fig. 26).

a) Evacuated reaction cell.
b) 2.5 torr methyl iodide in the reaction cell.
c) ca. 30 seconds of active pumping on the reaction cell.
d) ca. 5 minutes of active pumping on the reaction cell.


260 240 220 200 180

Fig. 30 Absorption band for 2.5 torr- methyl
region. Upper line is maximum light
an evacuated reaction cell from a 60

160 nm

iodide in the 200 nm
intensity through
watt deuterium light

or 3 torr or less for the 178.3 nm line. The perfluoroalkyl iodides did

not absorb as much in this region as the alkyl iodides.

In general, the lamp was found to be very versatile in operation.

After overnight pumping and change of the molecular sieve and parent

compound, various emission spectra could be observed. Among these,

Figure 31 shows the OH radical emission, with argon carrier gas, from

the microwave lamp.99

The microwave lamp system used in these studies has several other

advantages. Since no internal electrodes were used, the lamp body is

simpler in design and construction. The use of an efficient resonant
cavity also produces a high degree of molecular dissociation,9 and

therefore produces an intense emission spectrum. An added advantage of

the electrodeless lamps was that no high voltage points existed outside

of the power source. Only microwave radiation shielding must be used.

Finally, the flow system had the greatest advantage of being able to

handle various carrier gases and parent compounds, enabling the whole

system to be a versatile light source.

Mercury lamp

The monitoring light source used for the oxygen-ozone studies was an

intense low pressure electric discharge mercury lamp, Figure 32. The

lamp system, already available in our laboratory, had been originally

designed for AC operation. The lamp was converted to DC by rectifying

the AC and stabilizing its steady-state operation with a large, high

voltage capacitor. DC operation was necessary since sixty-cycle fluctuation

was readily detected in the time scales used in the experiment, and the


I i I I I I ,
305 310 315 320 nm

Fig. 31 Emission spectrum of OH (A2 + X2R) from the microwave
discharge lamp. Lamp pressure ca. 1 mm argon and water

power resistor
5.2 K
10 W

diode rectifying


Fig. 32 Mercury lamp system in DC operation mode.

DC provided a constant baseline for the experimental measurements. A

power resistor was used in series with the lamp to serve as the current

limiter when the lamp was in operation. The lamp was always started

with a Tesla coil in order to provide initial ionization. The lamp was

used with the experimental configuration shown in Figure 16.

The lamp did not have the stability observed with the microwave

discharge lamp. The light intensity versus time changed over the course

of its warm-up which took about 90 seconds. After this time the light

intensity would gradually decrease mainly due to buildup of ozone near

the lamp window and around the optical path outside of the reaction cell

(Figure 16) as well as in the reaction cell. The ozone was formed from

the photolysis of oxygen by the 184.9 nm mercury line. In order to

minimize the ozone buildup by photolysis of the oxygen in the reaction

cell, the pulsed experiment was always carried out 35 seconds after the

lamp was started. This time was chosen because between 30 and 60 seconds

after the lamp was started it showed consistent and reproducible light

intensities, and it was early enough so that internal and external ozone

buildup was a negligible factor.

Graphs were made of the buildup of the lamp intensity so that

maximum intensities at the time of the electron pulse would be known.

Since the experiment measured intensities, the initial intensity was

needed for the calculation of optical densities. Figure 33 shows the

lamp intensity, in relative units, versus time, with the reaction cell

evacuated. Figure 34 shows the lamp intensity versus time at various

oxygen pressures in the reaction cell. The use of these curves to

compute the optical densities for the experimental conditions will be

described later.




.-- 20


0 illill

0 20 40 60
Time (seconds)

Fig. 33 Hg lamp intensity versus time for various input
sensitivities of the transient recorder (volts
full scale) with the reaction cell evacuated.
The slit widths were 0.8 mm.


80 I



S_ 400


n torr
7- -

50 -

0 20 40 60
Time (seconds)

Fig. 34 Hg lamp intensity versus time for two initial
oxygen pressures and vacuum in the reaction cell
at 1 VFS transient recorder input sensitivity.
The slit widths were 0.8 mm.

F. Febetron 706 System

The electron beam source is the apparatus around which the pulse

radiolysis system was designed. In principle the electron beam could be

used as a direct sample irradiation beam or it could be used to generate

an X-ray irradiation beam. The efficiency of conversion is low and the

deposition of X-ray energy in a sample is usually low also. For example,

a 2 MeV electron may convert about 2% of its energy into X-radiation. A

small percentage of this X-radiation would be absorbed by the sample.

Thus, the dose per pulse from direct electron beam irradiation of the

sample would be about 104 times larger than from converted X-rays.85

Conversion efficiencies for electrons below 1 MeV are even less.100

Therefore, in practical operation for fast kinetic spectrophotometry,

where a large concentration of transients was desirable, it was advantageous

to use the electron beam directly. The electron beam can deliver 4 Mrads

of surface dose in a thick aluminum absorber. With an efficient X-ray

target (converter) the electron beam could produce an X-ray dose of 100

R at the target surface. This X-ray dose 30 cm from the target, through

air, is 80 mR. Larger Febetrons, like the 705, can deliver a dose, at

this distance, of 25 R. For our purposes the e-beam was utilized directly.

As previously discussed, low pressure (atmospheric or less) pulse

radiolysis requires a large pulse current for production of observable

concentrations of transients. For fast kinetic spectrophotometry the

initiating pulse should ideally be shorter than the rate process being

observed. The range of lifetimes which can be studied is extended by a

short pulse. The Febetron 706 easily met these two basic requirements.

Equipped with a Model 5515 high vacuum field emission tube it could

deliver a beam energy of 10 Joules, a peak current of about 8 kiloamperes,

and a maximum electron energy of 600 KeV in a pulse duration of 3

nanoseconds, FWHM101. The pulse-to-pulse repeatability was specified as

5%. This was observed to be fairly consistent, although a 5-10% difference

in pulses was observed for about 50% of the pulses during normal

operation under equal conditions.

Details of the Febetron 706 system may be found elsewhere.00102

Figure 35 shows a diagram of the system. Briefly, 15 modules in the

Marx-surge pulser circuit were charged from the high voltage power

supply in parallel, with their central spark gaps in open circuit. A

triggered voltage breakdown of the spark gaps discharged the modules in

series. The high voltage output to the short pulse adapter was thus

created by addition of the charging voltages. The high voltage, coaxial,

oil-dielectric capacitor in the short pulse adapter received the high

surge current pulse. After a predetermined voltage was reached and

subsequent breakdown a shorter pulse was discharged into the field

emission tube and passed out the tube window.

The e-beam tube incorporated a cold cathode electron source and a

thin window of low density material. The high density e-beam was

accelerated through the thin window.

Since the energy density of the e-beam was high enough to cause

melting, evaporation, or spallation of metals the titanium reaction cell

window was designed to be several centimeters from the tube window. This

design insured that melted, evaporated, or spalled materials could not

condense on or strike the tube window causing subsequent damage.

In our experience the Febetron 706 was found to be a highly reliable

low maintenance system. Its various operating requirements are simple.

A compressed air tank is used for pressurization of the module chamber,


(U E

) ,- I-

O r- 0 0

,-- z


QL 0
UL) 5-
,-- L.I.


4-) 0

> I

*r- 0
r. a



4- E

3 -

4r) r
- I3

S- 0


0 S

00 ir-
I v1
0 CL




X 0

S- E
TO "---


0 E
-o ta

at a maximum 80 psi, and the short pulse adapter, at a maximum 240 psi.

Each chamber had its own pressure regulator and gauge. Freon 12 was

used in the pulser case at a constant 15 psi pressure.

The transformer oil in the short pulse adapater was circulated from

a large reservoir by an electric pump. During normal experimental

operation the oil was circulated once daily. The lifetime of the Model

5515 field emission tube was extended beyond 3500 pulses by using it

below its maximum output potential. In typical use the total output

energy used was about 7 Joules, which corresponds to about a 510 keV

maximum electron energy.

G. Transient Recorder and Computer System

Early pulse radiolysis studies used the kinetic spectrophotographic

method for the detection system.8'9 Transient signals were either

photographed from a spectrophotometer directly or from an oscilloscope.

Expensive high speed photographic film was generally used, followed by

manual methods of data analysis. Methods for automatic digitization and
analysis of the transient signal were quickly developed.2122103

These methods not only facilitated data acquisition, but greatly reduced

the time for analysis of large amounts of data. Substantial savings

could also be realized by reducing the use of high speed photographic


Present methods employ transient recorders to obtain and store a

digital record of the transient signal as a function of time. Such

recorders can be directly interfaced to a computer, oscilloscope, or

graphic recorder for data display and analysis.

The data acquisition and analysis system developed for the pulse

radiolysis system has been described previously.23 A Biomation Model

610B transient recorder (Biomation) was used to digitize the signal
which came from the PMT through two amps in series, Figure 11.1

The amplifier's bandwidth extended from DC to about 5 MHz.

The Biomatation has an input range from 50 mV to 50 V full scale.

The analog bandwidth, though, is directly limited by the input sensiti-

vity. The higher sensitivity scales have lower bandwidths. From 50 mV

to 5 V the bandwidth changes from 100 kHz to 2.5 MHz. The analog circuitry

preceding the digital circuitry may not be fast enough, under certain

conditions, to fully utilize the digital circuitry. The digital circuitry

could store a new six-bit word into its 256 word memory as slow as every

50 ms or as fast as every 100 ns. Thus, the analog risetimes could only

be neglected when the time between sampling was greater. Therefore, to

achieve useful microsecond sampling rates the Biomation had to be used

in the 0.5 to 2 V input scales. This necessitated the use of high lamp

intensities and high PMT gain. Fortunately, the study of the fast decay

of the excited state iodine atom associated with the parent compound

pressures used was also associated with a high light intensity. No

mismatching of the analog and digital circuitry was observed. For the

ground state atom studies, lower lamp intensity necessitated the use of

a higher input sensitivity. This, though, did not present a problem

since longer sampling rates were used due to the longer lived signals.

The ozone studies posed no problems either since the light was intense

and the signal was long-lived.

If a signal were to require a high sensitivity scale, the transient

lifetimes involved could usually be extended by using lower pressures.


This would help insure faithful digitization of the signal.

In operation the Biomation was used in the single-sweep-delayed-

sweep mode. This mode allowed a one shot sampling of the pulse experiment

at a chosen delay interval after the trigger signal. Once taken the

data could not be lost, changed or disturbed by any subsequent trigger

signals until the user armed the input for the next signal. In practice,

the trigger delay on the Biomation was set at zero and instead the

Febetron trigger pulse was delayed. The Febetron trigger amplifier had

a ten turn graduated helipot in conjunction with a decade scale from 10

microseconds to 100 ms. This device was far more sensitive and

reproducible than the ungraduated one turn trigger delay on the Biomation.

This triggering system would trigger the Biomation to start sampling

first and after an appropriate delay-time the Febetron was triggered,

all from one initial manual trigger. In one sweep the Biomation could

record the initial intensity before the electron pulse (baseline), the

pulse, and the subsequent signal. By varying the sampling rate and the

trigger delay the entire signal could be analyzed with any resolution

desired that was not analog bandwidth limited.

The IMSAI 8080 S-100 bus microcomputer system, which was developed

elsewhere in our laboratory, 105 is based on the Intel 8080 microprocessor

chip, Processor Technology Company 3P+S parallel and serial input-output

(I/O) board, is similar in concept to larger more versatile systems
21 22
based on the PDP 11 family hardware.21,22 However, ours was the first

full-featured fast pulse data acquisition and analysis system based on

an 8-bit microprocessor to be described in the literature.23106

Operation of the IMSAI 8080 microcomputer required several levels

of supporting software. Programs were loaded via an Innovex brand 8

inch floppy disk system controlled by a Peripheral Vision Company interface

system and disk operating software. Interaction with the console keyboard

and video display was supported by a ROM firmware package associated

with the "Merlin" video controller board. This unit, which plugged into

the S-100 bus of the computer, also provided hardware and software for

good quality graphical display of the experimental results. However, an

additional program written in assembly code was necessary to provide the

specific graphic features required for the display of pulse radiolysis

data. Additional programs written in assembly code provided for transfer

of data from the Biomation transient digitizer to the computer, and for

transfer and display of the video screen image to the Integral Data

Systems Model 225 printer. The interface between all of the computer

facilities and the user was provided by a program in BASIC language,

running under a considerably modified Processor Technology Company "5K

BASIC" interperter. Modifications to the latter included provision to

work with a North Star Company floating point mathematics board, to

speed up computation; provision of additional built in functions including

LOG (natural logarithim), EXP exponentiationn), and others; and the

ability to transfer 8-bit bytes and 16-bit words from main memory into

BASIC's work space using PEEK and POKE functions. Work on all of the

systems programs was done by Mr. Scott Crumpton and Mr. Mark Adler of

this laboratory.

H. Computer Programs

The BASIC programs allowed full interaction between the operator

and the complete computer system. Program listings are provided in

Appendix I. *Both programs have similar commands and executions with two

exceptions. Only the experimental program (EXP), used during the pulse

experiment, had a command (11) which made a hard copy of the data on

paper tape, for subsequent data reduction. Only the analytical program

(ANA), used solely to reduce data, had a command (8) which was used to print

experimental data on the teletype as well as an option for interpolation

or extrapolation of the linear least squares (LLS) calculated curve.

For EXP the data from the Biomation was transferred directly to the

computer. For ANA the data were transferred to the computer through the

teletype from paper tape.

In typical operation an EXP program was used to acquire data from a

set of pulse experiments and store the data on paper tape. Its data

reduction capabilities, Commands 1-7, were rarely used for on-the-spot

data analysis due to the time consuming attention required by the

operation of the entire pulse radiolysis system by a single operator.

Some data were analyzed on-the-spot but only to observe general trends in

a set of experiments. Instead, at a more convenient time, the analytical

program was used to thoroughly analyze the data.

The programs could input data into memory via command 0 for

subsequent manipulation. Commands 1-3 transfer information to a buffer.

Command 1 clears the buffer and stores a new signal into memory. Command

2 sums and averages a signal into the buffer for signal averaging, thus

achieving better S/N. Command 3 could subtract a signal from the buffer.

This was used when a compound noise and physical signal was observed

during a pulse experiment. The noise signal, which could usually be

observed with the analyzing light turned off, could then' be subtracted

to yield the signal of the phenomenon under study. For the systems

studied this was only necessary when the sampling rate was less than or

equal to 2 microseconds per word or the volts full scale the Biomation

was 1 VFS or less with a 50 Ohm terminator and 700 volts on the PMT.

These sampling rates and conditions would always contain the overload

noise pulse within the experimental signal.

Commands 4 and 5 allowed linear least squares fit of a single-

experiment signal and the data in the buffer, respectively. Left and

right hand cursors could be adjusted to define the time interval over

which LLS fit would be carried out. Print-outs of the curve fit and

pertinent LLS fits were taken and plotted over the data curve in

order to achieve a visual, as well as a numerical, best fit format.

Commands 6 and 7 were used to print raw data values of a single-

experiment and the buffer, respectively, on the video. In conjunction

with a command to print the data being displayed on the video, a printout

of the raw data could also be attained.

Commands 9 and 10 input information which was used for the graphic

hard-copy and simplification of the video display, respectively. The

graphic output will be detailed later.

I. Reagents and Their Purification

Methyl iodide

Eastman and Mallinckrodt Analytical Reagent grade methyl iodide

were initially purified by removing 12 with Fisher Certified Reagent

grade sodium thiosulfate.07 The liquid was then degassed in the vacuum

line for three freeze-pump-thaw cycles. It was then distilled at room

temperature into a vessel in a carbon dioxide-isopropanol bath through a

Mallinckrodt Analytical Reagent grade phosphorus pentoxide drying tube.

The sample was again degassed in a freeze-pump-thaw cycle. Finally, the

sample was again distilled in vacuo, without the drying tube, into a

storage vessel immersed in a carbon dioxide-isopropanol bath. The first

20 percent and the last 20 percent of the volume were discarded. The

storage vessel, which was fitted with a Teflon stopcock, was stored in

the dark and the sample was always degassed before use.

Ethyl iodide

Reagent grade ethyl iodide obtained from Baker Chemical was purified

and stored using the same method as described above for methyl iodide.

CF3I and C2F5

Perfluoroalkyl iodides obtained from PCR, Incorporated were used

without further purification.

C4FgI and i-C3F7I

Perfluoroalkyl iodides obtained from PCR, Incorporated were degassed

and used without further purification.


Hexafluoroethane obtained from PCR, Incorporated was used without

further purification.


Fisher Certified Reagent grade iodine was degassed at 00 C. It was

then distilled in vacuo into a small storage tube in a carbon dioxide-

isopropanol bath. The tube, fitted with a Teflon stopcock, was on the

lamp system vacuum line and was used at room temperature (see Figure 22).


Airco, Incorporated oxygen was used without further purification.

Carrier and flushing gases

Airco, Incorporated argon, helium, and nitrogen, were used without

further purification. Sporlan Valve Company Catch-All line filters were

used on all the gas lines. The Catch-All filters could be reactivated

by heating at 1500 C while purging with nitrogen.

J. Sample Preparation

Vacuum system

All samples were prepared in a vacuum line, Figure 36. The pumping

system consisted of a Welch Duo-Seal mechanical pump connected through a

liquid nitrogen trap to a stainless steel oil diffusion pump. The

pumping system was connected to the main manifold through a second

liquid nitrogen trap and a 10 mm Teflon stopcock with "FETFE" o-ring


A three-way Eck & Krebs stopcock was used to bypass the diffusion

pump, for rough pumping, or direct the flow through it for high speed


The main manifold was designed with four stopcock ports. Two 3 mm

stopcocks were used as buffer gas inlets. Pyrex o-ring joints (#9)

connected the stopcocks to the tank gas lines. One 3 mm stopcock was

used to isolate the main manifold thermocouple (TC) tube. A 10 mm

stopcock separated the main manifold from the submanifold and the Wallace

& Tiernan differential pressure gauge.



4-) U

-TD G)

' 0 0-
0- S.- -
*r-- 4 S-


10 /)

tin *r-
t/ "I=


O 0

r- i- E
*- .- 30-


C- C








The submanifold was designed with five stopcock ports. Three 3

mm stopcocks served as sample inlet ports as well as a convenient arrangement

for in-vacuo distillation. One 3 mm metering stopcock was used as a

sample filling port to the reaction chamber. A 10 mm stopcock was used

to isolate the submanifold from the main manifold and the Wallace &

Tiernan differential pressure gauge.

A glass tube was used to connect the submanifold with the reaction

cell. A 3 mm stopcock at the reaction cell was used to isolate the cell

from the rest of the vacuum system as well as serve as the inlet and

outlet port for the cell.

All major connections were made with #9 Pyrex o-ring joints. The

total design of the vacuum system allowed the system to be taken apart

at the o-ring joints when major work had to be carried out on a section.

Metering and filling of the reaction cell

High vacuum was monitored with CVC thermocouple tubes, GTC-004, and

a thermocouple gauge. Sample pressures were measured with a Wallace &

Tiernan Model 62B-4D-0800 differential pressure gauge (range: 0-800 mm

Hg). The case of the gauge was always pumped to a high vacuum and

isolated with a 3 mm stopcock. The gauge sample capsule was opened to

the chemical sample or isolated through a 3 mm stopcock. The gauge

could be used to monitor either glass manifold as well as the entire

vacuum system.

All alkyl iodides were expanded into the submanifold, the differential

pressure gauge, and the reaction cell. Once the desired pressure was

reached the reaction cell was then isolated from the rest of the system.

The rest of the vacuum system was pumped down and the buffer gas was

introduced through the main manifold into the submanifold and the

pressure gauge. To promote mixing of the gases the buffer gas was metered

into the reaction cell by three or four expansions.

Low alkyl iodide pressures were attained by volume ratioing between

the cell and the submanifold.

For the oxygen-ozone studies, oxygen was expanded into the reaction

cell from the main manifold through the submanifold. No mixing technique

was required.

When used, the hexafluorethane buffer gas was metered into the

submanifold and expanded into the reaction cell.

K. Sample Irradiation

Radiation source and reaction cell

The Febetron 706 e-beam accelerator and the two reaction cells have

been described.

The electron beam exiting the field emission tube traverses about

2.5 cm of air before entering the reaction cell through a 2 mil titanium

window. The reaction cells were designed so that relatively uniform

irradiation was ensured in the region that was sampled by the analyzing

light beam. The reaction cell coupling collar to the Febetron allowed

reproducible positioning of either reaction cell.

Typically, the parent compound of interest was introduced into the

reaction cell at the desired pressure or reduced to a useful pressure by

volume ratioing. The buffer gas was metered into the cell so as to

promote mixing of gases through turbulence. With the pulse radiolysis and


microcomputer system ready, the Febetron was energized and triggered.

Due to the build-up of products in the reaction cell, each chemical

sample was rarely used for more than two pulses.


A. Experimental Results

I* deactivation by RI

The pulse radiolysis of the alkyl iodides was carried out at room

temperature. The excited state atom was monitored at 206.2 nm using

the iodine discharge lamp and atomic absorption spectrophotometry. The

pulse radiolysis was carried out over the parent pressure range of 0.5-

30 torr, depending upon the extent of the parent compound absorption

and its effect upon signal quality. For example, above 9 torr methyl

and ethyl iodide absorption interference hindered achieving satisfactory

S/N ratios. The argon buffer gas pressures ranged from about 150-600

torr; at these pressures isothermal conditions in the reaction cell

following e-beam irradiation were ensured.36 The buffer gas absorbed

energy from the e-beam and transferred energy to the parent compound

which subsequently dissociated.

All the computer fits were carried out on the buffer with at least
two averaged pulse signals.23 Since Febetrons are low pulse repetition

rate accelerators and due to the complexity of changing reaction samples,

rarely were more than four pulses averaged in the buffer for identical

experimental conditions. Good signal averaging was achieved with at

least three pulses. Figure 37 shows the absorption signal at 206.2 nm

for a single pulse experiment. Figure 38 shows the averaged optical

signal of three pulses. Significant curve smoothing may be observed.

Curve fit 4 2


Sc ie: 4 mricro-seconds Per X division
0.0731 volts Per. Y division ( 1 volts PY- )

Fig. 37 Graphical display of optical signal versus time for
a single Febetron pulse. Conditions: 3.2 torr methyl
iodide and 180 torr Ar.

normal run

Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E35P56XYE_8BUPRU INGEST_TIME 2012-03-08T21:53:49Z PACKAGE AA00003872_00001