PLACE IN RETURN Box to remove this checkout from your record.
10 AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 c:/ClRC/DataDue.indd—p. 1 5

 

A COMPUTER-CONTROLLED SPECTROPHOMETER
FOR RAPID REACTION RATE MEASUREMENTS

By

Timothy Gregory Kelly

A DISSERTATION
Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY

Department of Chemistry

2004

ABSTRACT

A COMPUTER-CONTROLLED SPECTROPHOMETER
FOR RAPID REACTION RATE MEASUREMENTS

By
Timothy Gregory Kelly

A computer—controlled transient EPR spectrometer was developed to
follow rapid radical reaction rates. The EPR system consists of a commercial 2
MHz magnetic ﬁeld modulator, a microwave cavity with a an internal modulation
loop, a modiﬁed Varian Fieldial for external control of the magnetic field and a
PDP 8IE computer for data acquisition and instrument control. A ﬂash-lamp
source was used to create the radicals to be studied in situ in the EPR cavity.

Modiﬁcations were made to the electronics of the 2MHz modulator and
phase-sensitive detector to increase the SIN ratio of the system. Additionally,
high modulation frequency cavity problems led to the development of a second
EPR cavity with an internal modulation loop, which was used for low temperature
experiments.

The ﬂash-lamp trigger was interfaced to the computer to give the
computer control over the repetition rate of the ﬂash. Remote optical sensors at
the ﬂash lamp provided feedback information to the computer. The technique of
Actinometry was used to help characterize the ﬂash lamp, optics and light
gathering capability of the EPR cavity by providing a determination of the number
of light quanta/pulse available inside of the EPR cavity.

The computer-controlled EPR spectrometer is capable of recording a

standard EPR spectrum or recording signal intensity as a function of time or a

time-resolved spectrum of a radical in a transient state. This system was applied
to the study of a Na-THF-Cryptate solution. A solvated electron, e'(so.v) , signal
was observed that increased upon intermittent sample illumination and
subsequently decayed to a steady state. The kinetics of the decay of the em.”
was recorded both as a time-resolved spectrum and as e'(so.v, signal intensity as a
function of time at a constant magnetic field. The decay time observed is
anomalously long and is explained in terms of the rate of release of the sodium

cation from the cryptate complex.

To Carol, a.k.a. C. Dors, and Mom and Dad

iv

ACKNOWLEDGMENTS

I wish to express my deepest gratitude to Professor Chris Enke for his
help, guidance, encouragement and friendship during the course of my research

at Michigan State University.

I thank the faculty and staff of the MSU Chemistry Department for the
excellent support they offered. I especially thank Professor J. L. Dye for his
many interesting discussions, Dr. T. V. Atkinson for his assistance and moral

support, and W. Burkhart for his valuable assistance.

I thank my wife Carol and my parents for their interest in my studies and

their unfailing encouragement and moral support.

Finally, I thank the Nuclear Beer Group and the POETS Club for making

my life a bit more exciting.

TABLE OF CONTENTS

LIST OF TABLES ............................................................................................... viii

LIST OF FIGURES .............................................................................................. ix

CHAPTER 1

Introduction ........................................................................................................... 1

CHAPTER 2

The EPR System .................................................................................................. 9
Introduction .................................................................................................................... 9
Basic Elements ............................................................................................................. 9
The Cavity System ..................................................................................................... 10
The Source System ................................................................................................... 16
The Magnet System ................................................................................................... 17
The Modulation System ....................................................... 18
The MSU 2 MHz EPR Instrument ............................................................................ 20
The Source System ................................................................................................... 21
The Cavity System ..................................................................................................... 22
The Modulation System ............................................................................................. 26
The Fieldial .................................................................................................................. 30
The Flash Lamp .......................................................................................................... 34
The Computer ............................................................................................................. 35

CHAPTER 3

Experimental Techniques ................................................................................... 38
Introduction .................................................................................................................. 38
The EPR Spectrum .................................................................................................... 39
System Tuning ............................................................................................................ 50
Time-Varying Data ..................................................................................................... 51
Time-Resolved Spectra ............................................................................................. 57

CHAPTER 4

Applications ........................................................................................................ 63
Introduction .................................................................................................................. 63
The Sodium-THF Sample ......................................................................................... 65
Discussion ................................................................................................................... 78
Recommendations ..................................................................................................... 82

CHAPTER 5

Characterization of the Flash Lamp .................................................................... 84
Introduction .................................................................................................................. 84
The Flash Lamp .......................................................................................................... 85
Characterization of the Flash Lamp ........................................................................ 91

vi

Preparation of the Working Curve ........................................................................... 95

Photolysis Experimental Setup ................................................................................ 96
Photolysis of the Actinometer Solution ................................................................. 100

CHAPTER 6

System Modiﬁcations ........................................................................................ 104
Introduction ................................................................................................................ 104
The Fieldial ................................................................................................................ 105
Fieldial Error Signal .................................................................................................. 112
The Flash Lamp ........................................................................................................ 118
The 2 Mhz Modulator ............................................................................................... 118
The Modulator/Demodulator ................................................................................... 124
The Microwave Cavity ............................................................................................. 131

APPENDIX A

Program Listings ............................................................................................... 137

BIBLIOGRAPHY ............................................................................................... 159

vii

LIST OF TABLES

5-1 Measured Quanta of Light per Pulse Available in the Micro-Now and Varian
Cavities ..................................................................................................... 102

viii

2-1

2-2

2-3

34
as
3-6

LIST OF FIGURES

Block diagram of a typical EPR spectrometer with 100 kHz magnetic ﬁeld
modulation and phase-sensitive detection .................................................. 11

A rectangular TE102 microwave cavity ....................................................... 13

Block diagram of the Micro-Now 509 2Mhz modulator and phase-sensitive

detector ....................................................................................................... 27
Simpliﬁed schematic of a Varian Fieldial magnetic ﬁeld regulator .............. 32
Flow chart for the ﬁeld-step method employed by EPRSCN ....................... 41

A spectrum of the perylene positive ion in sulfuric acid recorded with 2 MHz
ﬁeld modulation. The spectrum is the result of 1024 averages at ﬁeld

increments of 0.024 G ................................................................................. 45
The narrow-line spectrum of e-(solv) in THF recorded with 100 kHz (a) and

2 MHz (b, c) ﬁeld modulation ...................................................................... 47
A spectrum of a Varian standard weak pitch sample .................................. 49
Flow chart for the timed data acquisition program TIMEDA ........................ 52

Computer-acquired data testing the response time of the Micro-Now
spectrometer. The nominal time constants setting are 10, 1, 0.1

microseconds .............................................................................................. 56
Flow chart for recording a time-resolved spectrum with TIME12 ................ 58
A time-resolved spectrum of e-(solv) .......................................................... 62
A spectrum of Na-THF-Crypt recorded with 2 MHz ﬁeld modulation .......... 70

The e-(solv) signal as a function of time after a perturbing illumination pulse
recorded with 2 MHz modulation and a 1 ms ﬁlter time constant ................ 71

A spectrum of Na-THF-Crypt recorded with 100 kHz ﬁeld modulation ........ 73
A time—resolved spectrum of e-(solv) employing 100 kHz ﬁeld modulation .74

Spectra extracted from the time-resolved data set in Figure 4-4 showing the
spectrum at 0.1 ms after the ﬂash and continuing at 10-second intervals...76

ix

443

5-2
5—3
5-4

5-5

5—6

6-4

6—5

6-6

6-8

6-9

The decay of the e-(solv) signal as a function of time recorded with 100 kHz

ﬁeld modulation and a 20 ms ﬁlter time constant ........................................ 77
Circuit diagram showing the ﬂash lamp and associated trigger circuitry ..... 86
The ﬂash-lamp trigger interface to the PDP 8/E computer .......................... 88
The lamp-ﬂashed detector and light integrator ............................................ 90
A digital capacitance meter ......................................................................... 92

The ﬂash-lamp pulse shape with a 10 kV lamp supply voltage. The
abscissa is 2 us/div ..................................................................................... 94

A calibration curve of absorbance of 1,10-phenanthroline Fe(|l) complex as

a function of Fe(ll) molar concentration ...................................................... 97
Diagram of the focused light beam of the ﬂash lamp .................................. 99
The simpliﬁed schematic of the externally controlled Fieldial magnetic ﬁeld
regulator .................................................................................................... 107
Circuit diagram of the modulator enabling external control of the magnetic
ﬁeld ........................................................................................................... 109
The magnetic ﬁeld as a function of the external sweep voltage at a sweep
width of 100 Gauss ................................................................................... 111
The computer interface to the F ieldial internal error signal ....................... 114
Computer-recorded error-signal data as a function of time after a magnetic

ﬁeld increment. (a) A 500 G increment at a 1kG sweep range. (b) A 250 G
increment at a 500 G sweep range. (0) A 5 G increment at a 500 G sweep
range ......................................................................................................... 1 16

The 2 MHz power ampliﬁer and cavity modulation loop with its associated
impedance matching capacitors ............................................................... 117

The Micro-Now internal 2 MHz modulation loop. (a) A side view showing
the dimensions of the loop. (b) A top view showing the addition of the
magnetic ﬁeld from the individual wires .................................................... 121

A plot of calculated Hm(p-p) as a function of the power ampliﬁer supply
current ....................................................................................................... 123

The circuit diagram of the EPR output ampliﬁer capable of driving a 50-ohm
load ........................................................................................................... 126

6-10 The response-time test circuit used to determine the true response time of
the EPR spectrometer electronics ............................................................ 127

xi

CHAPTER 1
Introduction

Electron Paramagnetic Resonance spectroscopy was born in 1945 when
Zavoisky [ZA45] and later Cummerow and Halliday [CU46] detected electron spin
resonance absorption in solids. The fact that the discovery occurred shortly after
World War II is easily understood since a great deal of war-related research was
directed toward the development of radar. The development of high power
microwave generators, sensitive diode detectors, perfection of narrow—band
ampliﬁers, and lock-in detectors to increase the sensitivity of radar systems
advanced the state of the art sufﬁciently to supply the components necessary for
the magnetic resonance spectroscopy.

Early EPR spectrometers relied on DC detection of the absorption signal.
For a strongly paramagnet sample (Cummerow and Halliday [CU46] used 173
grams of MnSO4-4HzO), a change in resonance absorption results in a variation
of incident microwave power at the detector. The detector can be either a
bolometer or a diode coupled to a milliammeter. As techniques and instruments
improved, sensitivity was limited by the ability to maintain the klystron output
power at a constant value. Feher [FE57] explained in a classic paper on EPR
sensitivity considerations that the power of the klystron must be kept constant to
a greater degree than the smallest change in power detectable due to sample
absorptions. The minimum detectable sample size produces a current change of
approximately 10'“ amps. Unfortunately, it is not feasible to maintain the

klystron power constant to the degree required by such small current variations.

To overcome the stringent requirement on controlling the absolute klystron
power, modulation and phase detection techniques were introduced. Modulation
also helps reduce the contributions to low frequency noise by the diode detector.
These sources of noise, which show up as instabilities at the output, are
classified as 1If noise. At the time of Feher‘s experiments, the modulation
frequencies used were in the audio range, 20 to 6000 Hertz (Hz). Higher
modulation frequencies were introduced when it was discovered that the signal-
to-noise ration improved proportionally to the square root of the modulation
frequency. Piette and Landgraf [Pl60] reported a 16-fold improvement in
sensitivity when they changed their modulation frequency from 400 Hz to 100
kHz. They used this increased sensitivity to measure steady-state
concentrations of unstable intermediates photolytically produced in situ in the
EPR cavity. Rate studies of the decay processes of three radicals were
conducted by setting the EPR spectrometer to on resonance line of the radical
and then extinguishing the light. Decay curves were plotted on a recorder. The
results of the experiment prompted Piette and Landgraf to claim that EPR would
be an extremely useful research tool for photochemists who were relying on
product analysis, quantum yields, and trapping and ﬂashing techniques to
determine mechanisms for free-radical formation and decay in the photolysis of
organic compounds. The use of EPR to detect radical reaction intermediates has
an obvious advantage over the use of optical methods since optical methods
merely ascertain the presence of the intermediates but only in a few cases where
the band spectra can be interpreted is I it possible to identify the intermediate as

a free radical. EPR, on the other hand, only responds to systems containing

unpaired electrons, such as free radicals, and thus presents a highly sensitive
was to determine if the reaction intermediate is really a free radical. Quite often,
the free radical can be identiﬁed by its hyperﬁne pattern. The main disadvantage
of using EPR rather than optical techniques when Piette and Landgraf published
their results was the time response of the EPR instrument.

The ﬁrst comprehensive and elegant demonstration of EPR studies of
transient radicals was conducted by Fessenden [FE63] in a study of transient
alkyl radicals created by continuous irradiation with fast electrons. However, due
to the relatively short radical lifetimes involved (about 5 ms), it was not possible
to directly observe the decay of the EPR signals when the electron beam was
interrupted. In 1964, Fessenden [FE64] reported using a modiﬁed 100 kHz ﬁeld
modulation system and a pulsed radiolysis source to determine shortradical
lifetimes (7.3 ms for the ethyl radial). The basic change in experimental setup
was the use of the 100 kHz output signal normally used for the oscilloscope
display as the EPR absorption signal. In 1968, Atkins et al. [AT68a] used Piette’s
technique to study radicals created in ﬂash photolysis. Using a Decca X—band
spectrometer and 100 kHz modulation, reaction-rate data were taken at 640 ps
intervals with a pulse-height analyzer. The analyzer was capable of triggering a
xenon ﬂash tube. Decay curves obtained from several ﬂashes could then be
averaged. This process, repeated at several magnetic ﬁeld values, allowed a full
time-resolution EPR spectrum to be manually reconstructed from the individual
averaged decay curves. Using this technique, Atkins et al. [AY68b] recorded
what is believed to be the ﬁrst high-resolution EPR spectrum of a radical

intermediate taken under conditions other than steady state. He recorded the

hyperﬁne spectrum of a benzophenone radical at intervals of 0.1 Gauss.

Effort to develop EPR as a true kinetic research tool intensiﬁed in the late
60’s. Glarum and Marshall [GL70] modiﬁed the receiver/detector electronics of a
Varian 100 kHz spectrometer to obtain a net response time of 200 us. Bennet
and coworkers [BE67] modified their spectrometer by changing the ﬁeld
modulation frequency to 500 kHz, which resulted in a wide instrumental
bandwidth (50 kHz) and a short response time (6.5 us). In later work they used
the slightly smaller modulation frequency of 465 kHz [BE69].

In what is believed to be the ﬁrst use of 2 MHz as a ﬁeld modulation
frequency, Smaller et al. [SM68] inserted a hairpin loop inside a microwave cavity
and reported a system response time corresponding to a time constant of 1.6 ps.
Subsequently, Atkins et al. [FA‘l70a] modiﬁed their Decca X-band spectrometer
and modulated the 100 kHz coils with 2 MHz. Due to the high impedance
presented by the modulation coils at 2 MHz, the maximum amplitude of the ﬁeld
modulation was limited to one Gauss peak-to-peak. Atkins reported a response
time on the order of 1 us. This response time approaches the limit for obtaining
useful information from EPR determined by the response time of a microwave
cavity. The time required to change the microwave ﬁeld in the cavity by 1Ie is
approximately 20 oscillations [SM77]. For the X-Band cavity with a Q of 5000,
the time constant is about 1 us.

Atkins et al. were able to use their transient response EPR system to
obtain important new magnetic relaxation data pertinent to the mechanism of
reactions in solution [A170a, b]. They reported that immediately after the ﬂash,

they saw emission from 2-chlorobenzaldehyde in liquid parafﬁn decaying with a

time constant of 14 us into an absorption signal. As a consequence of chemical
reaction, the absorption signal then decayed with a half-life of about 2.4 ms. This
is believed to be the ﬁrst well-documented report of Chemically Induced Dynamic
Electron Polarization (CIDEP).

A commercial 2 MHz ﬁeld modulation EPR spectrometer built by Micro-
Now Instrument Company of Chicago became available in the early 1970’s
[H077]. The EPR cavity contained an internal modulation coil similar in design to
the modulation loop described by Smaller et al. [SM68]. Thus, a true transient
EPR spectrometer was commercially available to the scientiﬁc community.

J.K.S. Wan used a Micro-Now system for his published CIDEP studies [WA75].

Huisjen and Hyde [HU74] have taken a unique approach toward
developing a transient EPR spectrometer. A saturating microwave pulse of short
duration (100 ns) was applied to the sample followed immediately by a
monitoring low-power microwave, which allowed observation of radicals as soon
as 0.2 ps after the pumping microwave pulse. The system is used with magnetic
ﬁeld modulation, thereby eliminating side bands caused by the ﬁeld modulation
frequency. To decrease 1ft noise, the pumping pulse is repeated at 6-50 kHz.
The range of lifetimes accessible with this instrument is 250-02 us.

With the availability of instrumentation, which allowed data to be taken 0.2
us after the perturbation pulse, the problem of handling the potentially
tremendous volume of data needed to be considered. Researchers who used
the original technique of Piette and Landgraf[P160] (which was to hold the
magnetic ﬁeld constant and record kinetic data as a function of time) acquired

data with devices such as a standard recorder for long lived radicals [P160], 3

storage oscilloscope [GL70], and transient recorders that eventually sent their
output to an X-Y recorder [F167, Wl68, AT68a]. Atkins et al. [AT68b] used a
‘Biomac 500’ pulse-height analyzer to record and average data at several
magnetic ﬁeld values. After the experiment, they had to manually reconstruct a
time-resolved spectrum of the radical. The volume of data that could be acquired
using the new, fast instruments posed a serious problem for data storage and
analysis.

In an effort to eliminate the need for manually reconstruction a time-
resolved spectrum from multiple decay curves, Bennett et al. [BE67] designed a
six channel sampler system which was triggered by the flash lamp and which
acquired data at ﬁxed time intervals along the decay curve. The output of each
channel was connected to a recorder. The magnetic ﬁeld was linearly varied and
data were recorded during the magnetic ﬁeld sweep. At the end of the ﬁeld
sweep, six spectra had been constructed, each spectrum corresponding to one
of six time intervals after the ﬂash. Holbrook et al. [H072] built a 16-channel
sampler that wrote data on a 16-channel tape recorder. The spectra were later
plotted on an X-Y recorder. Again, each channel corresponded to a spectrum at
a speciﬁc time after the ﬂash.

Another approach for obtaining a spectrum of a transient radical is to add
auxiliary ﬁeld sweep coils [$068, H168, H873] and synchronize the ﬁeld sweep
with both a signal-averaging device and a pulsed excitation source. There are
several problems, however, associated with this technique [BO37] including
nonlinearity of sweep ﬁelds, limited sweep range and decreased ﬁeld

homogeneity as sweep amplitudes are increased. There is also a problem using

a CAT (Computer of Average Transients) to store the entire spectrum since most
signal averaging devices have limited memory, typically 4096 words.

For data acquisition and analysis both small [KL72, GR73] and large
[6072, SE73] computers have been used to greatly simplify data analysis
compared to previous methods. However, to allow optimization of experimental
conditions and to eliminate problems associated with other techniques,
interactive control between the computer and spectrometer is necessary. In
1975, Goldberg and coworkers [6075] designed a computer-controlled EPR
spectrometer, which had the magnetic ﬁeld under computer control. The
software for their system, enabled data to be acquired either in ﬁeld-step or ﬁeld-
sweep mode. In the ﬁeld-step mode, the computer increments the magnetic ﬁeld
and waits for it to stabilize before accumulating data, while in the ﬁeld-sweep
mode, the ﬁeld is continually changed while data are taken. Results of several
studies using this computerized system have been published by Goldberg
[GO76a, GO76b, 60760, 0078].

The purpose of this thesis is to describe the computer interface and the
necessary modiﬁcations, which were made to an EPR system to create a
computer-controlled EPR spectrometer for rapid reaction rate measurements.

Chapter 2 contains an introductory discussion which ﬁrst familiarizes the
reader with EPR instrumentation and then describes the interfacing,
modiﬁcations, and programming that are necessary to construct a functioning
computer-controlled spectrometer. This chapter includes only brief statements of
problems and modiﬁcations made to the purchased Micro-Now 2 MHz

spectrometer. The detailed circuit descriptions are left until Chapter 6 for readers

who desire a more complete knowledge of the electronics.

The various modes of operation of the computer-controlled spectrometer
are described in Chapter 3 and Chapter 4 discusses applications of the
techniques to chemical problems.

The properties of the ﬂash lamp and design of its computer interface for
kinetic studies are discussed and characterized in Chapter 5.

Finally, the appendix to this thesis contains listings of the computer

programs that were essential for completion of this project.

CHAPTER 2
The EPR System
Introduction
In this chapter, a description of a standard EPR spectrometer is
presented. The reader will be familiarized with the components and their
functions in a typical instrument. Then, modiﬁcations necessary to change the
modulation frequency to 2 MHz and allow computer control and data acquisition
are described. A detailed discussion of the system modiﬁcations is presented in

Chapter 6.

Basic Elements

All absorption spectrometers, regardless of frequency range, have four

basic components:

1. Source of electromagnetic radiation

2. Dispersing element

3. Absorption cell containing the sample to be studied

4. Detector
Various spectroscopy techniques differ in the properties of these four elements.
The main characteristic of the radiation source is its frequency, e.g. UV, visible,
infrared and microwave. EPR instruments, which use microwave sources,
operate at one of two common frequencies: nominally 35 GHz (referred to as 0-
band) or 10 GHz (referred to as X-band). The wavelengths for Q and X bands

are 0.85 and 3.0 cm, respectively. The dispersive element in EPR is the

magnetic ﬁeld. However, its action is much different than the dispersive element
in other forms of spectroscopy because it disperses the response of the sample
(Zeeman splitting) instead of the frequency of the source. The absorption cell is
the microwave cavity, which contains the sample. An EPR detector is usually a
silicon diode biased to operate in its linear region with a voltage output
proportional to the square root of the incident power [HY63].

Figure 2-1 is a block diagram of a typical 100 kHz EPR spectrometer
[WE72a]. The block diagram is divided into ﬁve functional groups. The region
labeled source contains all the components, which control and measure the
microwave frequency and intensity. The cavity system is composed of the
resonant microwave cavity that holds the sample and the components that direct
the microwave to and from the cavity. The modulation and detection region
contains the components that encode and phase detect the absorption signal.
The magnet system provides the homogeneous, linearly variable magnetic ﬁeld
that surrounds the sample. Each of these functional blocks will be discussed

brieﬂy starting with what is generally considered the central component of the

instrument — the cavity system.
The Cavity System

The heart of an EPR spectrometer is the resonant cavity containing the
sample. The cavity (1) establishes the spectrometer frequency by its internal
dimensions, (2) contributes substantially to the sensitivity, and (3) may be an

important factor in resolution [AL68c]. A cavity can be characterized by the

10

£28825 o>£80mlommnq use cows—ace: Eva 23:38 5.3 2: .23 8880.589. Maw 3390 a .3 ESch 0305 ”TN onE

 

«I!rtittremtwinmrmm~Mﬂmmmqmartttrtirw

 

 

 

 

 

 

l
-1

13".“
'001331!

 

 

G

_
lg le

>h~>¢o

 

 

 

 

 

turn». hazocx

 

 

 

 

 

 

 

   
 
    

 

 

 

/ 2%
ﬂ

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
 

 

 

 

 

 

 

 

cohoupua cupuxusc:
nzmuum llmmw.
. _ 4% 322.3; 32.3.: 22:23.
:22; 3823 E :32. 3.53

 

 

 

9694
—08~h¢8nxcuh

 

 

“tau.

 

 

 

11

following parameters [AL680]:
1. Mode or standing wave pattern of the microwaves
2. Resonant wavelength
3. Selectivity factor (commonly called 0 factor)

Electromagnetic waves traveling in free space or in an ideal waveguide
propagate with the magnetic (H) and electric (E) ﬁelds in phase. In a resonant
cavity however, the H and E ﬁelds form standing wave patterns with the H and E
nodes separated. This allows a sample to be placed in a region where the
magnetic ﬁeld is at maximum amplitude while encountering a minimum in the
electric ﬁeld strength. Cavities are classiﬁed by the internal ﬁeld patterns that
result from the separated E and H ﬁelds as either transverse electric rnode,
TE.,m.n, where the electric ﬁeld has no Z component or transverse magnetic
mode, TM.,m.n, where the magnetic ﬁeld has no Z component. The subscripts l,
m, n, refer to the number of half-period variations of the E and H ﬁelds along the
X, Y, 2, dimensions of the cavity. Figure 2-2 shows the X, Y, Z axes as deﬁned
for a TE102 cavity. As one can see from the internal ﬁeld patterns, a sample
would be in an area of maximum H ﬁeld and a minimum E ﬁeld if inserted along
the X axis at the center of the Z axis. TEm is the mode of the cavity used in this
research. This mode cavity is commonly used because its width is not important
as long as it is less than one half-wavelength. This allows the construction of
narrow cavities that ﬁt very easily between the pole tips of all magnets. The ﬁeld
patterns in the cavity allow one to out large holes in the sides of the cavity, eg. to
let in light, without affecting the quality of the cavity. The narrow cavity shape

also provides ﬂat sidewalls on which to mount ﬁeld modulation coils. The

12

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 2-2: A rectangular TE102 microwave cavity.

13

fundamental resonant frequency of a cavity is the frequency at which one half-
wavelength of the microwave corresponds to a cavity dimension.- The internal
dimensions of a TEm cavity at a resonant frequency of 10 GHz are 1.5x<1.5x3.0
cm.

The selectivity factor or quality factor, Q, of a cavity is usually deﬁned by
Equation 2-1 [P067a].

_ 27r(energy stored)
(energy dissipated per cycle)

 

(2-1)

In an ideal unloaded cavity (loss due only to ohmic losses in the cavity walls), the
“energy stored” can be expressed as an integral of the power density over the
volume of the cavity. The “energy dissipated” is power released as heat when
the current induced in the walls of the cavity by the varying magnetic ﬁeld ﬂows
through the resistance in the cavity walls. Cavity Q can be approximated by
Equation 2-2 [AL678]

z Volume of the cavity (2_2)
0 _ 6 x (surface area of the cavity)

 

where 6 is the skin depth (depth at which the current density has decreased to
1le of the surface value) of the current at the cavity resonant frequency. For a
given cavity mode, Le. a ﬁxed cavity volume, cavity 0 depends inversely on the
skin depth which in turn inversely depends on the conductance of the surface
metal. Therefore, to minimize resistive losses and maximize Q, the internal
surfaces of the cavity are usually plated with a highly conductive metal such as

silver or gold.

14

In early EPR literature, expressions were developed showing the
relationship of various parameters to the signal. Signal power was expressed as
[AL67a]

_ P0 (47:15] Q0) 2

S 16 (2'3)

 

where P0 is the incident microwave power, 00 is the unloaded cavity 0, x” is the
imaginary part of the susceptibility, and n is the ﬁlling factor. It should be noted
from this equation that signal power is proportional to the square of the unloaded
cavity 0, but only proportional to the ﬁrst power of the incident microwave source
power. A ﬁgure of merit for cavity 0 in a Tsz cavity is about 7000.

The microwave energy traveling through the waveguide is coupled to the
cavity through a small hole called an iris. This hole provides the function of
matching the impedance of the cavity to the impedance of the waveguide. Since
the reﬂected signal observed at the detector should be due only to the sample in
the cavity, reflected power due to a mismatch of waveguide and cavity
impedances must be avoided. When the impedances of the waveguide and
cavity are matched, this source of reﬂected power is at a minimum. An
adjustment screw in the iris permits optimum impedance matching.

The ﬁnal component of the cavity system is the hybrid Tee. This is a
“magic” device, which will not permit microwave power to pass in a straight line
from one arm of the Tee to the other. The power coming from the source is
equally divided between the cavity arm and an arm that contains a terminating
load and a slide screw tuner. If all the power is absorbed in the two arms of the

Tee, no power reaches the detector arm. Anything done in either arm to change

15

the terminating impedance, eg. a microwave-absorbing sample, will cause a
reﬂected signal to reach the detector. The slide screw tuner in the terminating
arm of the Tee is used to reﬂect the proper phase and amount of microwave to
destructively interfere with any unwanted reﬂections arising from discontinuities

in the waveguide or imperfect matching of microwave components.

The Source System

The predominant source of microwave power used in EPR is a klystron. A
klystron is a vacuum tube, which produces microwave radiation by causing
electrons to oscillate in a resonant cavity by applying an accelerating and a
reﬂecting voltage. The microwave radiation is tunable over a small frequency
range by varying the applied reﬂector voltage. A coarse adjustment of the
frequency can be made by physically changing the resonator cavity dimensions
using a mechanical tuning stub on the klystron. To prevent the frequency of the
klystron from drifting, a small modulation (typically 10 kHz), which is placed on
the reﬂector voltage, results in a modulated microwave frequency at the cavity.
The reﬂected signal from the cavity due to a mismatch of the microwave
frequency and the cavity resonant frequency is phase detected and used as an
error voltage to the klystron power supply to “lock” the klystron frequency to the
resonant frequency of the cavity. This frequency “locking" process describes the
function of the AFC circuit.

An isolator is a device similar to the hybrid Tee in that it will only pass

microwave radiation in one direction. It is used in the EPR source system to

16

prevent reﬂected radiation from the cavity from interfering with the operation of
the klystron.

The wavemeter is a variable length resonant cavity, which can be adjusted
to an integral number of half-wavelengths by a micrometer. It is used in a typical
EPR instrument as a measurement device to determine the actual frequency of
the microwave being used.

The amount of microwave power reaching the sample cavity is controlled
by a variable attenuator. The attenuator is a resistive element, such as a nickel-
chromium ﬁlm on glass that absorbs energy when inserted into the microwave
through an opening in the waveguide. The depth of penetration, and therefore
the amount of energy absorbed, is controlled by a calibrated dial on the front

panel of the instrument.

The Magnet System

The large, constant magnetic ﬁeld (Ho) used to produce the resonant
absorption is provided by an electromagnet and a highly regulated power supply.
Assuming a g-factor of 2 (g for a free electron is 2.00229), at the X-band
microwave frequency; this ﬁeld must be approximately 3400 Gauss. At Q-band
frequencies, it must be about 12,500 Gauss. The ﬁeld must be homogeneous
over the entire sample volume, and should be very stable. Total ﬁeld variations
should be less than 10 mG over the entire sample volume [P067b]. Modern
magnet systems utilize a Hall-effect device in the ﬁeld as a magnetic ﬁeld-to—

voltage transducer. An error signal applied to the magnet power supply is

17

obtained by comparing the voltage signal from the Hall probe to a reference
signal. A scanning system is created by linearly varying the reference voltage
and allowing the resulting error signal to change the magnet current. Therefore,

the actual magnetic ﬁeld at the sample tracks the reference voltage.

The Modulation System

In general, when making a measurement requiring high sensitivity in a
noisy environment, modulation techniques are used. A physical property is
modulated, ampliﬁed by a narrow band ampliﬁer tuned to the modulation
frequency, and phase detected. Only noise near the modulation frequency can
interfere with the measured signal. By looking at the resonance equation for
EPR,

h v = ng (2-4)
one can see two variables that might be modulated, microwave frequency (v) or
magnetic ﬁeld (H). James Hyde [HY63] has shown that due to the narrow
bandwidth of microwave components, it is impractical to modulate the absorption
resonance by modulating the microwave frequency. Therefore, spectrometers
keep the microwave constant, “locked” to the resonant frequency of the cavity by
the AFC, and modulate instead the applied magnetic ﬁeld.

The magnet ﬁeld is modulated by placing a pair of small Helmholz coils on
the sides of the cavity and applying a modulation signal to the coils. The most

common modulation frequency used today is 100 kHz.

18

Detection of the signal is accomplished through the use of a silicon diode
that rectiﬁes the reﬂected microwave signal. Because of the high frequency of
microwaves, capacitance in the detector and cables ﬁlter out the microwave
carrier so that the only signal passed to the detector electronics is the 100 kHz
modulation. This 100 kHz is ampliﬁed by a narrow band ampliﬁer, phase
detected, ﬁltered and displayed on an X-Y recorder with the X axis of the
recorder synchronized to the magnetic ﬁeld sweep. The transient response
which one sees from an EPR instrument depends upon the modulation
frequency, the bandwidth of the narrow band ampliﬁer, the amount of electronic
ﬁltering used to remove noise and the response of the recording mechanism.

A discussion of an EPR spectrometer would not be complete without
considering the parameters affecting the sensitivity of the instrument. The
minimum detectable number of paramagnetic centers, Nmin, in an EPR cavity at a
signal-to—noise (SIN) of unity is [WE72b].

23FckTsr (Fdeb)r/2
Zag ﬂ s(s+1)H,Qu' Po

 

Nmm= <2-5)
Here Vc is the volume of the cavity, k is Boltzmann’s constant, T, is the
temperature of the sample, I' is half of the half-width of the absorption line in
Gauss, Hr is the magnetic ﬁeld at resonance, Qu’ is the unloaded cavity Q, Ta is
the temperature of the detector, b is the bandwidth of the entire detector and
amplifying system, P0 is the microwave power incident on the cavity, and F is a
noise ﬁgure (>1) used to describe noise other than thermal noise. Ideally, F

would be 1. The terms of particular importance for this research are the

19

bandwidth of the system b, the cavity 0, and the noise ﬁgure F. Assuming a
resonance line quarter-width of 1 Gauss, room temperature and a TE102 cavity,
Nmin = 1011 spins.

It is generally accepted that the three components, which limit the SIN, are
the klystron, the detector diode and microphonics (mechanical vibrations or
spurious electrical transients) [AL68b]. Any of these three may be the noise-
limiting component under certain experimental conditions. For example, at low
power levels, 1If crystal detector noise dominates while at high power levels,

klystron noise predominates.

The MSU 2 MHz EPR Instrument

Since the ﬁrst reported use of 2 MHz ﬁeld modulation by Smaller et al. in
1968 [SM68], researchers who had no desire to design and develop the
instrumentation for 2 MHz EPR have been interested in applying this new kinetic
tool to their particular research. In the early 1970’s, a small company named
Micro-Now instruments started marketing a Model 509 2 MHz EPR spectrometer.
The sales brochure advertises, “The Model 509 is an EPR spectrometer which
uses a 2 MHz ﬁeld modulation frequency, which makes it ideal for pulsed type
systems such as those used in ﬂash photolysis and pulse radiolysis. Its very fast
recovery time permits the Model 509 to be used for the study of paramagnetic
species which have half lives as short as one microsecond” [MNOW]. The Model
509 is made up of three subassemblies:

1. 509-1 microwave source system

20

2. 509-2 microwave cavity system

3. 509—3 2 MHz modulation and detection system.
It is sold without magnet or power supply, and thus it is meant be used with an
existing EPR spectrometer.

Since many applications for a fast-response EPR spectrometer existed
within the MSU chemistry department, a system was purchased. The
commercial EPR spectrometer designated to receive the Micro-Now modiﬁcation
was a Varian 4502 Q-band spectrometer. Since the microwave plumbing for 0-
band is not compatible with X-band, an entire 509 system was ordered. The
resultant instrument, the MSU. 2MHz EPR spectrometer (Micro-Now Serial

Number 5) will be discussed brieﬂy.

The Source System

The Model 509-1 consists of the complete source system described above
minus the klystron power supply and AFC. These functions are supplied by the
Varian host spectrometer.

The microwave source is a V153 klystron with a power output of 300 mW
at a maximum power supply voltage of 350 volts. At this voltage, the klystron
must be water-cooled. Fortunately, the klystron could be safely operated at the
stated maximum 350-volt limit since that value corresponded to the absolute
minimum output of the klystron power supply in the host Varian console. Varian
incorporates a higher voltage, quieter klystron that the V153 in their systems.

Varian literature states that the V153 was initially designed for radar applications

21

and has molded leads and base to allow operation at high altitudes without
pressurization. The NMR-EPR service group here at MSU had several V153
klystrons that came as military surplus. It is unfortunate that Micro-Now saw ﬁt to
skimp on a component that is generally accepted as a SIN limiting component;
but to manufacture an inexpensive spectrometer, priced so that researchers
would be willing to gamble on it, manufacturers cut corners wherever possible.
This philosophy bodes ill for the MSU 2 MHz EPR.

The slide screw tuner used to null unwanted microwave reﬂections arriving
at the detector was replaced in the Micro-Now instrument by an RF phase shifter-
bias attenuator to fulﬁll the same function. It was found that the phase shifter
was not capable of shifting the phase quite far enough for proper nulling of the
unwanted reﬂected signal fro the cavity arm. However, no modiﬁcations were

made to this unit.

The Cavity System

The Micro-Now 509-2 EPR cavity is an X-band, rectangular, TE102 cavity
designed speciﬁcally for 2 MHz ﬁeld modulation. The cavity is machined from a
solid block of brass and then is silver plated to decrease the surface electrical
resistance. This procedure is necessary since the cavity 0 is inversely
dependent on the surface resistance. Only the resistance of the internal surface
of the cavity is important for cavity 0 because high frequency electric current is
carried in the skin or surface of a conductor. The “skin depth” for silver (i.e. the

depth at which the current has decreased to 1Ie of the surface current) is 0.64

22

pm at 10 GHz [AM72]. Since better than 99.9% of the current ﬂows within 5 skin
depths of the surface, only the resistance of the top 3.2 pm is important.

The cavity has a small rectangular opening on one side to allow in situ
irradiation of a sample. This opening was quite small and proved to be difﬁcult to
use with a ﬂash lamp. This light intensity problem is discussed in detail in
Chapter 5.

The 2 MHz modulation coil is similar in design to one published by Smaller
[SM68]. The coil was a single loop of wire shaped like a hairpin positioned inside
the EPR cavity. At high modulation frequencies, an internal modulation coil is
preferable because it would be difﬁcult to pass a high frequency magnetic ﬁeld
through metal. The oscillating ﬁeld induces eddy currents in metallic cavity walls
such that the direction of the induced electron ﬂow produces a reverse magnetic
ﬁeld in opposition to the original flux [$174]. The magnetic ﬁeld decreases to 1le
of its incident value in one skin depth. The skin depth in silver at 2 MHz is 45 um
[AM72]. To pass an appreciable amount of the magnet ﬁeld modulation, the
walls must be one skin depth or less thick at the modulation frequency.

However, to keep the microwave radiation inside the cavity, the walls must be
ﬁve skin depths thick at the microwave frequency. This corresponds to 3.2 pm at
10 GHz. At a modulation frequency of 2 MHz, the dimension of the walls is
critical and becomes too thin to properly support the modulation coils. Therefore,
internal modulation coils are commonly used.

The Micro-Now cavity had several drawbacks. First, the sample
irradiation opening in the cavity wall was smaller than it should have been,

unduly restricting the amount of light available in the EPR cavity. Second, the

23

magnetic resonance peak drifted in the Micro-Now cavity. This effect was
noticed when a quartz dewar, necessary for low temperature studies, was
inserted in the cavity. The problem was thought to be caused by the bottom of
the modulation coil loop. The coil is bent to go around the quartz dewar, but the
arc of the bend is not quite large enough. When the dewar is inserted in the
cavity, the coil is pushed out into the microwave ﬁeld destroying the cavity Q by
drastically increasing the losses in the cavity. The AFC does not work well on a
cavity with low cavity 0 because a low Q cavity will support a wide range of
frequencies. Therefore, the AFC modulation will not modulate the cavity out of
resonance and will not receive an error signal back from the system and the
klystron is free to drift. The shape of the bottom bend in the modulation coil is
undesirable because it will produce a component of magnetic ﬁeld perpendicular
to the desired magnetic ﬁeld.

Several modiﬁcation ideas were tried to create a more stable cavity.
Unfortunately time did not permit starting from scratch to build a new cavity so
the modiﬁcations were made to an existing Varian TE102 cavity, which then could
be operated at 2 MHz without the drifting problem inherent in the Micro-Now
cavity.

For room temperature work, the Micro-Now cavity was satisfactory. The
cavity 0 measured by Professor J. Cowen [COWN] was about 4000, a
respectable value for Q. It was because of this moderately high 0 that the
drifting problem was not seen until low temperature experiments were attempted.

Therefore, for much of the work done and spectra recorded, the Micro-Now

24

cavity was used. The maximum modulation amplitude was measured at 3 Gauss
peak-to-peak.

A Varian cavity (0 about 7000) was modiﬁed for the low temperature
studies. Several attempts were made to introduce a 2 MHz ﬁeld modulation into
the cavity but with the restriction that the Varian cavity coil not be permanently
altered. The selected method of introducing modulation into the cavity was to
install a quartz dewar which had a single strand of 28 gauge magnet wire glued
down its front and back. This piece of wire is used as the modulation loop. The
resulting cavity was better than the Micro-Now cavity in two ways. First, the
resonance did not drift during low temperature work and second, the sample
irradiation opening was larger than that of the Micro-Now cavity allowing a
greater fraction of the ﬂash lamp output to be collected in the cavity. The major
drawback is that not as much ﬁeld modulation can be produced in the cavity (less
than 1 Gauss peak-to-peak). This is not a problem for narrow resonance lines
studied in this research because a modulation amplitude greater than 0.2 times
the peak width will distort the line shape [WE72c]. The optimum SIN occurs with
a modulation amplitude 1.6 to 3.5 times the peak width, depending on the line
shape [WA61, SM64].

An unexpected beneﬁt of installing a quartz dewar in the cavity was an
increase in the SIN. The presence of a low-loss dielectric, such as quartz, in the
cavity has the effect of decreasing the apparent cavity size and increasing the
apparent ﬁll factor, n, of the sample [AL68e]. According to Equation 2-3 above,

this will increase the signal power. For example, Depireqx [AL68t] observed that

25

signal amplitudes were increased by a factor of 2 to 3 when the samples were

encased in quartz tubing.

The Modulation System

A block diagram of the modulator/demodulator electronics is show in
Figure 2-3. This unit contains the 2 MHz crystal oscillator/preampliﬁer used to
drive the power ampliﬁer connected to the modulation coil, plus all of the
electronics to amplify and phase detected the resulting signal. The signal
preamp has a selectable gain of 25 or 35 dB and is battery operated for low
noise. The preamp output is used both as the klystron AFC signal and also as
the EPR signal to be ampliﬁed and phase detected. The signal is routed through
a 0-41 dB attenuator to a detector/video ampliﬁer circuit built around an
integrated circuit balanced demodulator detector (Motorola part number
MC15961). The ﬁnal output passes through an offset and time-constant-select
ampliﬁer. The range of time constant selections is 0.001 — 100 um plus a
position marked NONE.

The modulator/demodulator system had several problems that led to a
need for eventually replacing all of the electronics in the system except for the 2
MHz crystal oscillator and phase shifter. The details of the modiﬁcation will be
discussed in Chapter 6 on system modiﬁcations.

After the system had been taken apart and set up a few times, the EPR
would not even produce a spectrum of strong pitch, which is one of the strongest

samples ever used in EPR. The problem was tracked to the lack of 2 MHz

26

.8883. 03:88-33.“— 23 83338 ~52 m cow 3oZ.o.§2 05 me 863% 5.005 ”Wm PBwE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

59:0
59: 3%:
Cowman. Mam
. 82> can}.
. o . 82m .83
0 0:3 8:80 8:2:on
o
. . 23
. . s w J ,
8:580 8535. «Em 82m 835 83
38 Base 82> as 8:20.32 83:80
_ FIIL _ L, — Hasrl+|L
.8880
036 m
«:9: J 38255 W.
E
5:18 3m 0 h
- uoagaoﬁxw lei ’ - 0130
5930 3:6me
0.? v , 853:2
- W . uoaom £2 N

 

 

 

 

 

I 9 l _
8535-5 E

 

27

modulation in the cavity. The Micro-Now 2 MHz power ampliﬁer required a
speciﬁc cable to be used between the power ampliﬁer and the cavity for the
modulation to work. Also, disconnecting the cable while the power was on would
cause the output transistor in the power ampliﬁer to burn out. The basic
philosophy used in the design of the power ampliﬁer was at fault. The
modulation loop in the cavity is a zero-ohm load at the end of a 50-ohm cable.
This impedance mismatch was probably causing a total reﬂection of the 2 MHz
modulation at the end of the 50 ohm cable. To eliminate this problem, the 2 MHz
power ampliﬁer was redesigned such that it would not self-destruct if the cable
was disconnected while the power was on. To terminate the cable properly, the
cavity modulation loop was used as the inductor in the tuned LC circuit with an
input impedance of 50 ohms and thus the coaxial cable was terminated at its
characteristic impedance. The measured 0 of the LC circuit containing the
modulation loop was 20, giving a 20-fold increase in the circulation current in the
modulation loop over the current from the power ampliﬁer. The circuit for the
power ampliﬁer and tuned LC circuit is show in Figure 6-6. A tuned LC circuit
was also built for the Varian cavity, again using the modulation loop as the
inductor. The measured 0 for the LC circuit including the modulation loop for the
Varian cavity was 2. This low Q was probably the result of capacitive coupling
between the wire canying the 2 MHz current into the cavity and the metal neck of
the cavity. To increase the Q of the tuned circuit, it would be necessary to
decrease the capacitive coupling between the wire and cavity which could be
accomplished by drilling holes in the top of the cavity to pass the wires through

as little metal as possible. Without drilling holes in the cavity, a Q of 2 was the

28

best that could be achieved. The resulting maximum peak-to-peak modulation
amplitude, Hm(W, for the Micro-Now cavity was 3 Gauss. For the modiﬁed Varian
cavity, the maximum HM”) was 0.5 Gauss.

The EPR output of the Micro-Now spectrometer drives a 20-foot length of
50-ohm coaxial cable terminated at the computer ADC with an input impedance
of 1 M0. This is not a problem with the slowly varying signal levels observed
when recording a standard spectrum; however, when transmitting a transient, it is
important to match all impedances to prevent reﬂections of the electrical signals
at the interfaces. A 50-ohm terminator was inserted at the interface between the
cable and the ADC. It was also discovered that the output impedance of the
EPR output ampliﬁer was about 10 k0. An operational ampliﬁer circuit based on
a National Semiconductor LH0032C and LH0063C, capable of driving a 50-ohm
load was built The bandwidth of the buffer ampliﬁer is 3.5 MHz, sufﬁcient for the
pulse response of this system.

At this point it was thought that the system was ready to use; however, the
ﬁrst attempt to study a transient reaction proved otherwise. The reaction rate
observed was much slower than could be explained by the chemistry involved. It
was discovered through pulse response tests described in Chapter 6, that me
EPR output ﬁlter time constant was longer by a factor of 10000 than indicated on
the front panel of the instrument. When the ﬁlter was modiﬁed to correct this
problem, system noise rose to a level that totally swamped the detector
electronics. The noise problem led to an immediate need for a study of the
electronics of the EPR ampliﬁer-phase detector circuits, after which everything

except the 2MHz oscillator was rebuilt. Details are discussed in Chapter 6. The

29

ampliﬁers were replaced with home built low noise ampliﬁers and the phase
detector was replaced by a Mini-Circuits Laboratory, ZAD-1, double-balanced
mixer. In building the circuitry, a trade—off between SIN and ampliﬁer bandwidth
had to be made which thereby eliminated the possibility of studying reaction rates
shorter than 5 pm.

Total instrumental time constants were measured by switching the 2MHz
modulation to the cavity modulation loop on and off while the EPR was at the
maximum of a strong pitch resonance peak and recording the corresponding
decay of the EPR output signal. Actual time constants of 3, 10, 100, and 2500
um were determined and the labels on the front panel output ﬁlter switch were

changed to these measured numbers.

The Fieldial

The magnetic ﬁeld to be applied to the sample is selected by using digital
controls on a Varian F ieldialTM Mark II magnetic ﬁeld regulator. The selector
controls indicate ﬁeld density directly in kilogauss. The Fieldial senses the
magnetic ﬁeld through a temperature-controlled Hall-effect transducer lowted on
one of the magnet pole tips and drives the magnet power supply to maintain the
selecmd ﬁeld value. A ﬁeld-sweep range selector on the Fieldial in conjunction
with a ﬁeld-sweep potentiometer driven by an AC motor provides a repetitive saw
tooth sweep of the magnet ﬁeld over the selected ﬁeld range. The ﬁeld-stability

and sweep-linearity speciﬁcations of the unit are excellent; however, there is no

30

direct was to control the ﬁeld sweep externally. There is a potentiometer
connected to the ﬁeld-sweep motor which gives a reference voltage proportional
to the position of the ﬁeld-sweep potentiometer, but his will not allow control of
the ﬁeld; it only lets one follow the position of the F ieldial. One approach to
external control could be to switch the ﬁeld-drive motor 0N while the computer
follows the ﬁeld reference voltage and switch the motor OFF when the
appropriate voltage is present. However, this approach is unacceptable for
several reasons. When the motor is switched off, it continues to turn for a short
period of time after the power is off making it difﬁcult, if not impossible, to stop at
the exact ﬁeld desired. To minimize this overshoot, an extremely slow scan rate
would have to be selected, which would mean that it would take an unacceptably
long time to move from one ﬁeld setting to another setting.

The interfacing approach, which was selected as the best means for
external control allows the computer to randomly select the percentage of ﬁeld
sweep desired. A detailed discussion of the interface is presented in Chapter 6.
An overview will be presented here.

A block diagram illustrating the function of the Fieldial is shown in Figure
2-4. The Hall-effect transducer is positioned in the magnetic ﬁeld by attaching it
to a magnet pole tip. A 1290-Hz reference signal is applied to the Hall-effect
transducer, the ﬁeld-set controls, and the sweep-range controls. The signal from
the sweep-range control is applied to the ﬁeld-sweep potentiometer. The signal
from the ﬁeld-sweep potentiometer is then combined with voltages from the ﬁeld-
set controls and the Hall-erect voltage at the input of an error ampliﬁer. Any

change in the ﬁeld or sweep control settings results in an error signal that is

31

sesame use 8.25% 323.2 88> So 0:823. unease in 8am

 

   

 

 

 

 

 

 

 

emmzmuozcm D ... w . L .......... J.
a x . .
-o v u 9!. 53me
. uozcm.
n ammzw"
moo: .-.--.-.-...
ammnmmw cohuuhmo
...:
323.». a 92:... commm
cuzom 9 1|,

 

 

 

 

 

 

 

 

S7.
36
Io

.IH

1

8018

 

 

 

 

 

 

32

 

phase detected and applied as a control signal to the magnet power supply. The
magnetic ﬁeld changes as a result of the error signal and the Hall voltage
increases or decreases accordingly. As the magnet ﬁeld attains the correct
value, the error signal approaches zero.

By replacing the ﬁeld-sweep potentiometer with a four-quadrant analog
multiplier, the computer can control the ﬁeld by applying a DC voltage from a
digital-to-analog converter to one input of the analog multiplier. The output of the
multiplier (X*Y) is connected to the error-sense board of the Fieldial.

The DC analog voltage is derived from a 10-bit DAC within the computer
giving a resolution of slightly better than 0.1% of the ﬁeld-sweep range selected.
This resolution is adequate for the small ﬁeld-sweep ranges selected in this
research but at a sweep range of 1 k6, the resolution is only 1 G, which is wider
than many EPR lines. The DAC should be replaced with a 12-bit DAC
(resolution of 1 part in 4096) for general use.

The linearity of the ﬁeld sweep was determined by applying a DC voltage
to the analog multiplier and measuring the resultant magnetic ﬁeld with a home-
built gauss meter utilizing the NMR resonance of the proton. The plotted results
are shown in Chapter 6, Figure 6-3. A linearity test at sweep-width settings of
10, 100, 500 and 1000 G resulted in linearity within +I-0.25% of the sweep width.
These values are better than the linearity speciﬁcations claimed by Varian in their
Fieldial literature (0.5% of sweep width).

The short-term stability of the ﬁeld was difﬁcult to measure with the home-
built gauss meter used in this analysis since it was not a tracking gauss meter. It

had to be manually adjusted. The frequency to establish proton resonance was

33

selected using a variable capacitor. During the entire integration period of the
frequency meter one’s hand had to be kept on the adjustment knob to prevent
the frequency from changing, which prevented continuous monitoring of the
magnetic ﬁeld. The short-term stability results are based on ﬁve readings made
over about one minute. At sweep-range settings of 50 and 100 G, with zero volts
applied to the sweep input, the maximum ﬁeld deviation seen was 25 mG. This
change corresponds to 7 parts per million in the measured frequency.
Considering the manual method of adjusting the frequency of the gauss meter,
this error may not be signiﬁcant. The error value is about 25 larger than the
value reported by Glamm [GL73]. Wrth a voltage input of +5 volts, the maximum
ﬁeld excursion detected was 35 mG.

An error-signal monitoring circuit was built and is described in Chapter 6.
The monitoring circuit produces an analog signal proportional to the error signal
in the Fieldial. The output voltage is connected to one channel of the multiplexed
analog-to-digital converter in the PDP 81E computer. The main EPR computer
program calls a subroutine called ERRSUB every time a new magnetic ﬁeld is
selected. ERRSUB monitors the error voltage and returns to the calling program

when the error signal reaches a steady state.

The Flash Lamp

The excitation source selected for this research was a low repetition rate
(4 Hz at 10 kV supply voltage) Xenon ﬂash lamp. The lamp electronics, optics

and housing were built by Photochemical Research Associated in London,

34

Ontario, Canada [PHOT]. The high voltage power supply used for the lamp was
a Hipotronics Model 820-50 variable voltage supply.

Characterization of the ﬂash lamp system is the topic of Chapter 5. Only
the resuits will be presented here.

The duration of the light pulse was measured using a monochromator and
a photomultiplier. The photomultiplier output was recorded on a storage
oscilloscope. The full-width at half-maximum of the light pulse from the ﬂash
lamp is 2 us with the entire pulse over in an absolute maximum of 10 us at all
supply voltages used (8, 9, 10 kV). Actinometry, a chemical photon counting
technique, was employed to estimate the number of quanta of light available
inside of the cavity at three lamp supply voltages (8, 9, 10 W). The quanta/pulse

for wavelengths below 450 nm at 10 W in the Micro-Now cavity was 2x10”.
The Computer

Because of the high-speed data acquisition, instrument control, and event
synchronization necessary to perform the desired EPR experiments, the system
is placed entirely under computer control during data acquisition. During system
tuning prior to an experiment, the computer assists the operator by acquiring,
averaging and displaying results as the operator varies the many parameters
necessary to tune the system. In addition to the control and acquisition function,
the computer performs all the data reduction and analysis necessary to interpret
the results. A number of peripherals are available for displaying and storing the

results. The computer used for control and data acquisition is a Digital

35

Equipment Corporation [DEC] PDP Lab 8IE with 12k words of memory, an
analog and a digital laboratory-peripherals option comprised of a four-channel
multiplexed, 10-bit analog-to-digital converter (ADC), two 10—bit digital-to-analog
converters (DAC), a 12-bit parallel digital input-and output port and three Schmitt
triggers with selectable slope and variable trigger threshold. The Schmitt triggers
can be used to synchronize the computer with external events. This option also
includes a programmable real-time clock used for time synchronization of events.
Additional standard peripherals on the PDP 8IE were an Extended Arithmetic
Element (hardware multiply and divide circuit), an ASR 33 Teletype, a high-
speed paper tape reader, and a single DECtape drive with a TD8E controller.

The computer system, affectionately called the Rolling 8IE, has been
enhanced by several custom and nonstandard peripherals. The addition of a
dual Sykes ﬂoppy-disk drive eliminated the dependence on a single DECtape
unit and was a giant step forward in reducing input-output (lIO) time for compiling
and data storage. Data display was on a Tektronix storage oscilloscope [LAST]
that was modiﬁed to include an ASCII character generator enabling the scope to
be used for high-speed alphanumeric output. The scope was conﬁgured to
emulate a line printer so that standard DEC software could be used to support
the device.

The computer uses DEC’s OSI8 operating system, which includes a
FORTRAN ll compiler. This FORTRAN compiler permits in-line inclusion of
machine assembly language (SABR) statements in the FORTRAN code yielding
an extremely ﬂexible programming tool. One can use the power and ease of

FORTRAN programming when timing is not crucial, but one can also include

36

direct in-line assembly language instructions when timing is critical or when
handling nonstandard peripherals.

All of the programs for the EPR system were written in FORTRAN-SABR
code. Listings of the main programs and subroutines used with this system are
included in Appendix A. Each program will be discussed in the appropriate
experimental sections of the thesis.

In addition to the Rolling 8IE, a second computer was used for some data
reduction and plotting of the results. This second computer is a Digital
Equipment Corporation PDP 11l40 with a Floating-point Instruction Set (FIS), 28
k words of memory, an RK05 disk drive, a Tektronix T4010 graphics-
alphanumeric terminal, a dual Sykes ﬂoppy disk, a surplus 900 line-per-minute
line printer from an old RCA 301 computer [HAHN], and a home-built incremental
plotter [H076].

Data were transferred between the PDP 8IE and the PDP 11l40 using
ﬂoppy disks as the transfer medium. The transfer is accomplished by PIP8
[PIP8], a program running on the PDP 11 I40, which is capable of interpreting the
data structure written by 08/8. A general purpose plotting routine, MULPLT

[MULP] was used to produce the hardcopy output.

37

CHAPTER 3

Experimental Techniques

Introduction

The current research project was spawned by the desire to integrate a fast
EPR spectrometer with a data-logger so that the combined system could be used
to study transients EPR data. The interface was designed to incorporate an
intelligent controller since the use of a computer as a controller provides ﬂexibility
not possible with a hardware controller or data-logger. Data-taking and
experimental monitoring capabilities provided by the system software created for
the interfaced EPR spectrometer are described in this chapter.

The computer interface, discussed brieﬂy in Chapter 2, was designed and
built to allow computer control over the elements of the total system that are
essential for the interactive data acquisition. The interfaced components include
the Varian Mark II Fieldial magnetic ﬁeld controller with a transducer circuit to
monitor ﬁeld stability, the ﬂash-lamp trigger and the lamp monitoring circuits. The
ﬂash-lamp interface is discussed in Chapter 5. Many aspects of EPR system
tune-up, such as adjusting optimum microwave power and the magnet ﬁeld
modulation amplitude, were not placed under computer control, but instead were
left to the experimenter. Having such functions interfaced would be a luxury, not
a necessity, since they add nothing to the capabilities of the instrument.

A computer-controlled EPR spectrometer capable of measuring transient

data is a very versatile instrument. Once the computer interface is built, the

38

ﬂexibility of the instrument depends entirely on the computer programs written for
instrument control, data acquisition, and data analysis. To accommodate a
different experiment it is necessary only to modify or add to existing software, not
redesign the hardware.

The software written to fulﬁll the current research demands allows the
experimenter to acquire an EPR spectrum or to monitor signal amplitude as a
function of time for the kinetic data studies. The functions are combined to obtain
a time-resolved spectrum of a transient radical. Since the hardware is the same
for all three data acquisition methods, each method will be discussed fro the

standpoint of its software implementation.
The EPR Spectrum

An EPR spectrum is a plot of the ﬁrst derivative of the absorption signal of
an unpaired electron as a function of magnetic ﬁeld strength. A ﬁrst derivative
signal is obtained rather than a normal absorption signal as a consequence of
the magnetic ﬁeld modulation, which is used for Isignal enhancement. The
spectrum is obtained by linearly varying the main magnetic ﬁeld, Ho, and
recording the EPR signal output on an X-Y recorder. The X-axis of the recorder
is linked to the magnetic ﬁeld sweep by some electrical method (for example,
Varian uses a follower potentiometer that is ganged on the ﬁeld-sweep
potentiometer to provide an output voltage proportional to the magnetic ﬁeld).
The rate at which the magnetic ﬁeld can be linearly swept depends on several

factors. The magnetic ﬁeld must be capable of tracking the reference signal

39

used to sweep the ﬁeld. This tracking requirement is not usually the limiting
factor for sweep-rate. Sweep rates of 100 Gauss/minute, linear to better than
0.02% have been reported [0075]. The sweep-rate limiting factor is usually the
response time of the spectrometer detector electronics coupled to the X-
recorder. The magnetic ﬁeld must be swept slowly enough to allow the
electronics and X-Y recorder to respond to the true line shape and amplitude of
the sample. The only method of averaging a signal to improve SIN ratio when
using either an X-Y recorder or a data-logger as a readout device is to increase
the time constant of response of the instrument. This prevents the output signal
frOm following the noise signals that are faster than the desired EPR signal. In
an experiment where the signal-to—noise ratio is small, the output ﬁlter time
constant must be large, possibly as much as 10 seconds, to remove random
noise and allow the average signal to pass to the recorder. Such long time
constants require the magnetic ﬁeld scan to be very slow to prevent distortion of
the output signal since it takes ﬁve time constants for the signal to reach 99.9%
of the new true value after a step-function change.

Wrth the MSU computer-controlled spectrometer, an EPR spectrum is
recorded by using a ﬁeld-step method. The computer, using a 10-bit DAC is
capable of dividing the selected ﬁeld-sweep range of the Fieldial into 1024
discrete steps. (Due to the nature of the interface to the Fieldial, the computer is
capable of randomly selecting for its starting point any one of the 1024 discrete
magnetic ﬁeld values.) The ﬁeld-step method of recording an EPR spectrum is

best described by following the ﬂow chart in Figure 3-1.

40

 

EPR scan sroee onrn
a—‘3‘ ( KEPRMON;\_ ._%>___.

(Lou-tum porn
I EPRSCN ( srone )
i SCPLOT i

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

{ENTER ; (ET {ENTER ;
FIELD SCRLE FNRHE
DRTH
UPDHTE ENTER
DIR COMMENT
PLOT
DRTR
{ENTER WRITE
9999". @ YES ‘ FILE
' ND
CRLL
CLOSE
ERRSUB RETURN FILE
BVERRGE
09TH RETURN
INCREH.
FIELD
YES
CRLL
FLOTDB
RETURN]

 

 

Figure 3—1: Flow chart for the ﬁeld-step method employed by EPRSCN.

41

An EPR spectrum is one of the options that can be selected by the
experimenter while the computer is in the monitor routine, EPRMON. A program
listing of EPRMON along with a list of other options available to the experimenter
appears in Appendix A.

When the EPR spectrum option is selected, EPRMON chains to a
program called EPRSCN. Through initial dialogue, the experimenter ﬁrst selects
a starting ﬁeld value for the scan. The default starting ﬁeld value corresponds to
50% of the sweep range below the magnetic ﬁeld value selected on the ﬁeld-set
dials of the F ieldial. However, the sweep may be started at some other value.
Since it may take several seconds for the Fieldial to stabilize if the starting ﬁeld
value is much different than the value currently selected, the computer updates
the magnetic ﬁeld ﬁrst and then continues the remaining dialogue to allow time
for the magnetic ﬁeld to stabilize. The remaining parameters to be speciﬁed are
the resolution and the number of data points to be averaged at each ﬁeld setting.
The resolution is the magnetic ﬁeld sweep-width divided by the number of points
to be taken across the spectrum. The maximum value that the resolution may
take on is 1I1024th of the ﬁeld-sweep width. The number of data points averaged
affects the signal-to—noise ratio. For truly random noise, the SIN improvement in
the averaged signal is proportional to the square root of the number of averages.
Signal averaging is an alternative to instrument response-time ﬁltering used by
conventional EPR spectrometers.

Once the scan parameters have been selected, the computer enters the
data collection loop. The ﬁrst task in the data collection loop is to see if the

magnetic ﬁeld has stabilized. The subroutine ERRSUB, called by the main

42

program, monitors the error signal generated by the error-signal interface in the

F ieldial. The voltage signal produced by the error electronics appears at one
input of the computer's multiplexed ADC. When the error signal reaches a
constant value (which indicates that the magnetic ﬁeld has reached the desired
value), ERRSUB returns to the calling program. When ERRSUB returns control
to EPRSCN, the requested number of data points are acquired at time intervals
equal to the ﬁlter time constant and then averaged. Next, the ﬁeld is incremented
and the process continues until the scan is complete.

The averaged data are stored as double precision integers during data
collection and are converted to ﬂoating point format by a subroutine called
FLOTDB. ERPSCN then returns to EPRMON where the data are auto scaled
and displayed on the computer storage oscilloscope using a program called
SCPLOT. SCPLOT also reports the X and Y-coordinates of the maximum and
minimum points in the spectrum. The spectrum can be expanded for observation
of the certain portions of the spectrum so that the operator can make decisions
about parameter changes before taking additional spectra. If the operator is
satisﬁed with the spectrum, it can be saved on an output device, such as ﬂoppy
disk. Subroutine PUT writes the data ﬁle onto the speciﬁed output device. The
operator is queried for a standard OSI8 “device-ﬁlename” which is used to open
the output ﬁle. The name of the ﬁle, the current date, and all of the scan
parameters are written into the output ﬁle for a permanent record of the scanning
conditions. In addition, the operator can enter a comment that will be written into
the data ﬁle identifying the sample and explaining any special conditions that

were used to acquire the data. The comment may be any length and is

43

terminated by typing a control G. The data are written in a standard ﬁle format
used widely within the Department of Chemistry, Le. a two character tag to
describe the data on the current line, such as ‘; ‘ for a comment, ‘RD’ for real data
format or ‘ED’ for end of data. The data, the scan parameters and the comment
can be later recalled by subroutine GET for plotting or data manipulation using
the data retrieval and manipulation monitor DATMON. The computer is left in
EPRMON after the scan.

Since this project was developmental in nature, the scanning technique
was mainly utilized in characterizing the instrument. Standard samples such as
pitch and perylene were used in the characterization experiments.

Figure 3-2 is a spectrum of the perylene positive ion in sulfuric acid
recorded at magnetic ﬁeld increments of 0.024 G and modulation amplitude of
approximately 0.1 G. Each datum is an average of 1024 samples with an
instrumental time constant of 100 us. The hyperﬁne splitting constants
determined from this spectrum are 4.11, 3.09 and 0.46 G. These values are in
good agreement with the literature values reported by Carrington and coworkers
[CA59].

A normal EPR spectrum should be symmetrical with respect to the central
peak; this perylene spectrum, however, is asymmetric. The cause for the
asymmetry is the presence of strong 2 MHz sidebands in the spectrum. Since
the diode detector is a nonlinear device, its output will contain the sum and
difference of the microwave frequency and the modulation frequency. These
sum and difference frequencies result in sidebands, which are associated with

each resonance line and are spaced at intervals of the modulation frequency,

44

.330 vmod mo $58805 20m 8 wows—26 em? no 258.. 05
mm :53on 2:. deceive": Eon Nag N 5m? @3808 Eom uremia 5 com 2638 one—bog 05 .3 85.58% < “Wm onE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

45

Wm. Conversion from frequency to observed magnetic ﬁeld splitting values is

obtained by direct substitution into the EPR resonance equation

hv = gﬂH (3-1)
or

Wm = g/3H
or

W... = 7”

When the modulation frequency, Wm, exceeds the line width being observed
(yAH), sidebands develop at intervals of Wmlv, in units of Gauss, from the
resonant magnet ﬁeld, Ho. Assuming a g of approximately 2 for the free electron,
the sideband interval is 720 mG at 2 MHz. The 720 mG interval for sidebands is
approximately twice the smallest coupling constant of perylene. Therefore, the
ﬁrst harmonic of the sideband, which is 180 degrees out of phase with the parent
peak, will cause peak broadening and amplitude reinforcement or reduction of
the second peak away from it in the spectrum. The number of prominent

sidebands, n, on each side of H0 is given by

”=7Hm/2W... (3-2)
The number n must be rounded to the nearest integer.

The spectrum of perylene is too complex to use for demonstrating the
inﬂuence of sidebands. Figure 3-3 shows a spectrum comprised of a single very
narrow resonance line attributed to the solvated electron and recorded with 100
kHz and 2 MHz magnetic ﬁeld modulation. The top spectrum, recorded with 100

kHz modulation (sidebands at intervals of 36 m0), is a standard ﬁrst-derivative

46

(a)
Millilllllllﬂlillt IiIlii‘illlliwllll‘WW

(b)

 

(C)

tMi—NMWMMW

Figure 3-3: The narrow-line spectrum of em...) in THF recorded with 100 kHz (a)
and 2 MHz (b, c) ﬁeld modulation.

47

curve with possible broadening due to the 100 KHz modulation. The middle
spectrum, recorded at 2 MHz, shows a pair of prominent sidebands while the
original resonance line almost disappears. By adjusting the reference signal
phase to 0 degrees instead of the normal 90 degrees, the spectrum that would
normally appear as a central line with sidebands will appear only as sidebands
[P067c]. This phenomenon, show in the bottom spectrum, was demonstrated by
Burgess and Brown [BU52] in a study of the NMR of protons in water.

The spectrum showing the separated sidebands is easier to use for
observation than the more complex spectrum of resonance lines plus sidebands;
therefore, all narrow-line spectra and kinetic data were taken with the reference
phase set to 0 degrees.

In Figure 3-3, the two peaks, one positive and one negative, are 1.4 G
apart or twice the modulation frequency expressed in Gauss. lnforrnation
concerning the original resonance peak shape is lost by using 2 MHz modulation
but then for line-shape analysis of extremely narrow lines, audio range
modulation frequencies should be sued rather than 2 MHz modulation.

Figure 3-4 is a spectrum of a Varian weak pitch (coal diluted with KCI)
sample. A Varian standard weak pitch sample is 0.00033% coal, which
corresponds to 1013 spins/cm in the active region of the microwave cavity. The
active region of the Micro-Now cavity is 3.8 centimeters, which corresponds to
approximately 3.8x10'3 spins. The spectrum was recorded with a 1 ms time
constant and is the result of 128 averages. Weak pitch has a measured line
width of approximately 1.2 G; thus the sensitivity of the system at a time constant

of 1 ms is about 3x10“ spins/Gauss. Therefore, to use the 2 MHz EPR system

48

 

 

 

Figure 3-4: A spectrum of Varian standard weak pitch sample.

49

 

successfully, either the number of spins must be very large, or the line width
must be very small. Varian advertises a sensitivity of 5x1010 spins/Gauss with its
100 kHz system at a time constant of 1 s. At a time constant of 1 ms, the

sensitivity of the 2 MHz spectrometer is about three orders of magnitude less.

System Tuning

System tune-up involves optimizing several instrumental parameters in
order to obtain the maximum signal amplitude of the true line shape for the
sample being observed. Since each of the parameters involved in tuning the
instrument affects the peak amplitude or line width, system tuning can be
considered a special case of recording a spectrum. The normal tune-up
procedure involves recording a section of a sample spectrum and then changing
a parameter such as the amount of microwave power. The same section of the
spectrum is again recorded and compared to the previous one. The process is
repeated until the optimum setting is obtained for each parameter. Some of the
parameters involved in the optimization process are the amount of microwave
power, the amplitude of the magnetic ﬁeld modulation, and the phase of the
modulation reference signal, which is sent to the phase-sensitive detector. The
controls for these parameters have not been interfaced to the computer and
therefore, must be manually adjusted by the experimenter. SETUP, a
specialized version of EPRSCN, was written as an aid for optimizing these
parameters. SETUP allows the operator to select the section of the spectrum

used for observation, the resolution desired, and the number of points to average

50

at each ﬁeld setting. SETUP then displays on the oscilloscope an automatically
scaled spectrum and reports the peak amplitude and width. SETUP waits for the
operator to change any one of the tune-up parameters, eg. the ﬁeld modulation
amplitude, and press the carriage-retum key on the terminal. It then displays,
without going through the dialogue again, another partial spectrum so that the
results of the change can be observed. This process continues until the operator

is satisﬁed with the setup and elects to return to the monitor program, EPRMON.

Time—Varying Data

The time-varying data acquisition method developed for use on the MSU
EPR system is similar to the method of Piette and Landgraf [Pl60]. In the
experiment performed by Piette and Landgraf, the EPR spectrometer was
adjusted to one of the lines of the sample and a shutter was closed to stop
illumination for the sample while the EPR output was observed with an X-Y
recorder. Instead of using a continuously operated lamp to build up a steady-
state population of the radical being studied, the MSU system uses a xenon ﬂash
lamp to produce a transient radical population. Data are then collected using an
ADC connected to a PDP 8IE computer. The ADC is triggered by a crystal clock
for precise timed-data collection.

To record time varying data at a constant magnetic ﬁeld strength, the
operator selects the timed data acquisition program, TIMEDA, from the list of
options in the monitor, EPRMON. The experiment can be followed in the ﬂow

char in Figure 3-5. It is assumed that the operator has already chosen the

51

TINED-DRTR / \ srone onrn
acouxsrrrouT\_"“"°" / 3 I

DISPLRY DRTR

SCPLOT

       

 

TJNEDR

 

UPDRTE
DIR

ENTER
PRRRN.

 

N0
RETURN

CRLL
ERRSUB

   

 

SETUP
CLOCK

RETURN

FLRSN
LRNP

REQUIRE

YES

 

A
' D
< a
n 4
3 a
3
O

RETURN

no
r-a
or
qr
o
a

Figure 3-5: Flow chart for the timed data acquisition program TIMEDA.

52

magnetic ﬁeld setting appropriate for observing the desired transient response.

During initial dialogue in TIMEDA, the operator selects the magnetic ﬁeld
appropriate for the sample being observed. As is the case with the ﬁeld-step
method initialization, the computer selects the magnetic ﬁeld immediately to allow
time for the magnetic ﬁeld to stabilize. Additional parameters that must be
entered are the tie between data points, the number of data points, and the
number of transient responses to average. The computer initializes the
programmable clock used to trigger the ADC and waits for the operator to press
any key on the terminal to indicate READY.

The computer triggers the ﬂash lamp and then waits for data from the
ADC. A “lamp-ﬂashed” signal is generated at the ﬂash lamp at the time that the
lamp ﬂashes by an ultra-fast photodiode circuit described in Chapter 5. This
pulse is received by a Schmitt trigger input at the computer, which starts both the
ADC for a time zero point and the programmable clock. From this point on, the
crystal clock starts the ADC at the data rage selected during the initial dialogue
until the requested number of points has been acquired. The nominal conversion
time for a successive approximations converter is 20 us with a sample acquisition
time of 3 us; however, in practice, the shortest time interval that does not yield an
ADC timing error from too-fast sampling is 25 us. This is sufﬁcient time to
average the collected data as the data are taken. The computer averages the
requested number of transient responses, and pauses long enough between
ﬂashing the lamp to maintain the maximum duty cycle recommended for the

lamp.

53

When the averaging process is complete, the data, which are stored as
double precision integers, are converted to ﬂoating-point format by subroutine
FLOTDB before returning to the monitor routine, EPRMON. The data are
subsequently scaled and displayed on the storage oscilloscope to allow the
experimenter to decide if the data set should be kept. If desired, the data can be
written onto a storage device, such as a ﬂoppy disk for later retrieval, plotting and
analysis.

Acquiring kinetic data at a constant ﬁeld value is not possible for all radical
studies since the exact ﬁeld position of the resonance peak must be know before
the experiment starts. If a sample does not have a dark spectrum, i.e. a
detectable concentration of the radical of interest already existing in the solution
before irradiation, the probability of choosing the correct ﬁeld setting for the
resonance peak is very small. Radicals with a dark spectrum allow prior
selection of an appropriate magnetic ﬁeld by the use of the ﬁeld-step method
discussed previously. For example, alkali metals dissolved either in ethers or
amines dissociate to produce a cation and a metal anion [T E74]. Experimental
evidence also indicates the presence of the solvated electron and monomer
radical in these solutions [DY71]. In fact, Friedenberg and Levanon [FR77]
reported observing a dark spectrum attributed to the solvated electron in a
solution of potassium metal dissolved in THF at room temperature. Thus, the
kinetics of the formation and disappearance of the solvated electron peak under
intermittent illumination is a possible application for the time-varying data

acquisition method discussed here.

54

The time-varying data acquisition method was used to characterize the
time response of the MSU EPR spectrometer. In Chapter 2, it was mentioned
that the instrumental response time of the Micro-Now spectrometer was
incorrectly labeled on the front panel. To determine the actual instrumental time
constants, the response-time test circuit show in Figure 6-10, was devised. The
double-balanced mixer was used to switch the 2 MHz ﬁeld modulation which
drives the modulation coils. The signal generated by the computer which
normally triggers the ﬂash lamp, was used to turn off the 2 MHz modulation and
was also used as the “lamp-ﬂashed” signal at the computer. First, the
spectrometer was tuned up on a standard strong pitch sample and an
appropriate magnetic ﬁeld was selected to observe the maximum amplitude of
the resonance line. Then, the 2 MHz modulation was switched off and the decay
of the EPR signal was recorded as a function of time. This procedure was
repeated for each time constant setting on the Micro-Now spectrometer.
Computer-acquired data for three time-constant settings of the Micro-Now EPR
are show in Figure 3-6. The nominal time-constant values of 10, 1, 0.1 us
yielded calculated time constants of 120, 31 and 7 ms respectively. These
values differ by more than four orders of magnitude from the labeled values. As
the selected time constant is decreased, the noise level becomes a major
component of the signal. In the 0.1 us plot (actually 7 ms time constant), the
noise level is already at an unacceptable level.

The computer could not be used to determine the time constants of the
rebuilt EPR spectrometer due to the limited data acquisition rate of the ADC in

the PDP BE (40 kHz). If necessary, the ADC data rate could be increased by

55

1
j: 10 US TIME CONSTANT

INTENSITY

 

 

 

 

 

 

 

..r

<

z

o

0’ 4.::.

>-

'—

2 1 US TIME CONSTANT

LU

...

2'

..J

<

z

o

a)

)-

l" l

‘2 0.1 US TIME CONSTANT

Lu

....

z

_r

<

z

o

to -+iH:::H+:+:rJ,H:::r:
TIME (S)

Figure 3-6: Computer-acquired data testing the response time of the Micro-Now
spectrometer. The nominal time constant settings are 10, 1, and 0.1 us.

56

purchasing a high-speed converter. An 8-bit AID converter with 50 ns conversion
time is currently commercially available. However, the computer is not fast
enough to handle such high data rates. The solution to the data-rate problem is
to use an analog shift register such as the one developed by Last [LA77]. The
analog shift register is capable of being clocked at data rates of 10 MHz for a
total of 100 samples. The sampled data may then be clocked to the output of the
device at a data rate that the computer can handle i.e. 40 kHz. Only 100 data
points can be taken at the maximum data rate, but additional data can be

acquired while data are output to the computer, at the output rate.

Time-Resolved Spectra

The time-varying data acquisition method described above is of limited
usefulness in the study of transient radicals that do not have a dark spectrum.
However, when this method of timed-data acquisition is combined with the ﬁeld-
step method, the instrument has the capability of obtaining a full time-resolved
spectrum of a transient radical by taking three parameter data (signal amplitude
as a function of time and magnetic ﬁeld). The time-resolved experiment is best
described by following the ﬂow chart in Figure 3-7. The control program, called
TIME12, can be selected as one of the options in the monitor routine or it can be
run as a stand-alone program without going through ERPMON. The initial
dialogue to set up the experiment is necessarily the combination of the two
previously discussed techniques. The experimenter selects the starting magnetic

ﬁeld value, which is immediately issued to the Fieldial to allow plenty of time for

57

 

(mm)

ﬂ

 

 

 

 

 

 

 

 

 

 

 

 

mam]

 

ELO

,_i
11H
”2

 

Ta:
030’
H10

RLE
o
sPan

 

 

[.ﬁﬂ

NO DONE?

YES

CLOSE
OUTPUT

 

 

 

Figure 3-7: Flow chart for recording a time-resolved spectrum with TIME12.

58

the ﬁeld to stabilize. Other parameters that are speciﬁed by the experimenter are
the magnetic ﬁeld resolution, the number of data points to be taken at each
magnetic ﬁeld value, the time interval between these points, and the number of
transient responses to average at each magnetic ﬁeld value.

Since the total number of data points collected during a time-resolved
EPR experiment can be vary large, in-memory storage of the data is not
possible. Therefore, the data for each magnetic ﬁeld value are averaged in
memory, displayed on the oscilloscope and then written into a data ﬁle that is
opened during part of the initial dialogue. All of the parameters speciﬁed by the
experimenter are written into the data ﬁle along with the name given to the data
ﬁle, the current date, and a comment entered by the experimenter to help identify
the experiment at a later date. Typing a control G terminates the comment.

When all of the parameters have been selected and the experimenter is
ready, the experiment is started by typing any character on the terminal. First,
the magnetic ﬁeld stability is checked by a call to ERRSUB. When ERRSUB
returns control to the main program, an average baseline value is obtained and
then the ﬂash lamp is triggered. Data collection proceeds in an analogous
fashion to that used in the time-varying method described above. After the
requested number of averages at a magnetic ﬁeld value is complete, the data are
automatically scaled and plotted along with a baseline on the storage scope.
The data are then written into the output ﬁle. The magnetic ﬁeld is incremented if
the scan is not complete and the process is repeated. After the entire

experiment is complete, the output ﬁe is closed. The data acquisition can be

59

halted gracefully at any time during the experiment by typing an “S” on the
terminal.

After the data ﬁle has been closed, the data can be displayed on the
storage oscilloscope on the PDP 81E computer; however, this computer has no
hardcopy facilities. Therefore, the data ﬁle must be transferred to the PDP 11/40
computer for plotting. A translator program, PIP8 (Peripheral Interchange
Program) [PIP8], translates the 08/8 data ﬁle into a ﬁle that is compatible with
the PDP 11 computer operating under DOS/BATCH. The DSTRIP (see Appendix
A) interprets the 08/8 data and creates secondary data ﬁles that are compatible
with the plotting routine MULPLT [MULP]. Data are written in the original OS/8
ﬁle as a series of decay curves, one for each magnetic ﬁeld value. DSTRIP will
extract a series of EPR spectra from the data starting at any point along the time
axis and continuing at the time interval selected by the experimenter. The
maximum number of spectra extracted each time the program is run is 99. The
ﬁle created by SDTRIP can then be plotted, 10 spectra at a time, by MULPLT.
DSTRIP will also write each individual decay curve from the original data ﬁle into
separated data ﬁles to be used for plotting or for data processing in order to
obtain information concerning the decay mechanism. DSTRIP also has an
option allowing digital smoothing of the data by calling subroutine SMOOTH.
This subroutine is a Savitzky and Golay [SA64] Quadratic-Cubic smoothing
routine capable of a 5-to-25 point smooth. However, care must be exercised
when smoothing data because smoothing will distort the information contained in
the data. Smoothing should be used for cosmetic purposes only, since the

process will not yield more information than that contained in the original data

60

[EN76]. DSTRIP has one other function: it will create a data ﬁle that can be
transmitted to the MSU CDC 6500 computer for tree-dimensional perspective
plotting using GEOSYS [GEOS]. Figure 3—8 is a 3-D perspective plot of the
decay of the solvated electron as a function of both time after the ﬂash and

magnetic ﬁeld.

61

5—
==

====E=EE
..EEEEE...
==_=E=== =====_==
:52: :— _==. =_ = _ .

 

 

M
M

62

Figme 3-8' .
' A “me-r930
Ned sPﬁ‘ctru
m Of e.(sorv
).

CHAPTER 4
Applications

Introduction

The observation of blue solutions consisting of alkali metals dissolved in
liquid ammonia was ﬁrst reported in 1863 [WE63]. Since that time, it has been
show that alkali metals also dissolve in amines and ethers [Ql27, 0057, D059].
However, unlike the metal-ammonia solutions that are lacking in speciﬁc optical
bands or EPR patterns, which could be used to characterize the various species
in solution, metal solutions in amines and ether provide abundant information
about distinguishable species. Optical and EPR spectra indicate the presence of
at least three reducing species in these solutions.

Alkali metal-amine solutions have been studied extensively with the aid of
optical and magnetic resonance techniques [DY73, CA75, GA70]. lnfonnation
about metals dissolved in ethers has been obtained mainly by optical [LO72,
EL63, KL71, SE77] and, to a lesser extent, magnetic resonance experiments
[GL70, FR77].

These solutions appear blue in color since they possess an absorption
band between 600-1000 nm. The exact position of the band depends on the
metal-solvent combination [LO72, DE64]. However, it wasn't until the work of
Hurley, Tuttle and Golden [HU68] that the information contained in the optical
spectra was interpreted. They were able to demonstrate that the alkali metal in
solution was exchanging with the sodium present in the Pyrex glassware used

during preparation, thus giving rise to a third, extraneous, absorption band in the

63

spectra of the alkali metals. Therefore, only the remaining two absorption bands
needed to be considered for any given metal-solvent combination [DY73]. These
absorption bands have been attributed to the species M' and e'(so.v) [MA69, Gl70,
KL71]. The frequency of one of the two absorption bands, the IR band, is
independent of the type of metal in solution, hence, is attributed to the presence
of the solvated electron species. In addition to the species identiﬁed by optical
methods, EPR evidence indicates, by the observation of hyperﬁne splitting in the
metal-amine solutions, the existence of low concentrations of a species with
stoichiometry, M [V063, BA64, Ga70]. Dye [DY71] has proposed an equilibrium

mechanism to describe the experimental observations.

2M(s) : M-+M+ (4-1)

M' : Mo+e' (4-2)
—> , _

M.(—M +6 (4-3)

Even though EPR evidence indicates the presence of a monomer species in
solution, the concentration is too low to observe its optical spectrum under static
conditions. However, using ﬂash photolysis coupled with optical absorption
techniques, researchers have observed a third absorption band associated with a
transient intermediate produced in the solution [GA71, KL71]. They proposed
that the intermediate species was a monomer radical, M -.

Alkali-metal ether solutions present an excellent application for a transient
EPR study. The original species is solution, M', is diamagnetic and will not be

observed with EPR techniques. After the ﬂash, a single resonance line for the e'

64

(30M and multiple lies for the monomer, M ~, if present as suggested [MA69, GL70,
KL71], should be detected and will show the same lifetime as was seen in the
optical studies. The EPR spectrum of the monomer should contain 2l+1 lines
where l is the nuclear spin of the alkali metal. The nuclear spin of sodium is 3/2,

producing a quartet of lines.

The Sodium-THF Sample

In order to obtain a better understanding of the species present in alkali-
metal solutions in ethers, a study of Na-THF (tetrahydrofuran) using EPR
techniques was proposed. Sodium was chosen as the alkali metal in order to
complement previous studies which deal exclusively with potassium and higher
alkali metals.

In related work that has been reported, Glarum and Marshall [GA70],
using a modiﬁed Varian EPR spectrometer (response time of 200 us), observed
an increase and subsequent second-order decay of the solvated electron signal
for potassium, rubidium, and cesium in dimethoxyethanen (DME). They
concluded that the recombination kinetics of the solvated electron indicated the
probable presence of an M radical, but they could not explain the absence of a
signal from the monomer in the EPR spectra.

The optical spectrum of Na' in THF shows the presence of Na' and e180")
Absorption peaks [L072]. The Na-THF solution should give rise to a strong EPR
singlet due to e130”) but contrary to the results obtained optically, EPR, being

much more sensitive than optical techniques, might observe hyperﬁne splitting

65

from the monomer. Irradiation of the sample with light containing wavelengths
above 600 nm, the region of the Na' absorption band, should produce an
increase in the concentration of the e'(so..,) and either a sodium monomer or a
sodium cation. Using ﬂash photolysis and optical absorption techniques,
Kloosterboer [KL71] observed an immediate decrease of Na- absorption after the
ﬂash coupled with the appearance of a strong e130“) absorption at its IR band.
As the em“) absorption decreased, a third band, interpreted by Koosterboer as
Na (monomer radical), appeared and subsequently decayed leaving only the
original absorption band. The second-order decay kinetics of the e130”) suggests
that the intermediate is composed of one sodium cation and one e'(soM.
However, it has been suggested that the transient absorption band observed in
this research was actually due to the presence of a potassium contaminant in the
solution, not sodium monomer as reported by Kloosterboer [CA77].

A similar kinetic study experiment yielded results, which are in apparent
conﬂict with the reaction mechanism proposed by Kloosterboer. Huppert and
Bar-Eli [HU70], studying Na—amine solutions optically, interpreted their ﬁndings

by the mechanism:

Na- —) Na 0 + 3‘ (4-4)
and a subsequent second-order recombination of M':

Na 0 + 3’ —+ Na- (4-5)
The addition of Nal did not change the kinetics to pseudo-ﬁrst—order and

therefore, it was assumed that Na~, not Na*, was formed upon sample

irradiation. However, they did not observe a Na absorption peak. Similar results

66

were reported by Gaathon and Ottolengthi [GA70] studying Na-amine solutions
and again the monomer was not detected. They reported that the mechanism for
sodium appeared to be different than that generally accepted for potassium.
Potassium solutions decay to their original state through a transient intermediate,
which is thought to be a monomer radical.

The introduction of organic macrocyclic complexing agents for alkali-metal
cations, increases the stability and the solubility of metals in solution and opens
many interesting possibilities in the study of alkali-metal solutions [L072].
Friedenberg and Levanon [FR77] reported an EPR study of potassium, rubidium,
and cesium in THF with and without the presence of crown ether. Without the
addition of crown, CR, the only paramagnetic species, which was observed, was
the solvated electron. With the addition of a small amount of CR, a weak signal
corresponding to the monomer was detected. Due to poor SIN ratio of the
monomer signal, the kinetic study was based on the observation of the decay of
the e'mm resonance line. They reported pseudo-ﬁrst-order rate constants for
potassium, rubidium, and cesium in THF at different CR concentrations and
found them to be CR concentration dependant.

For the present study of the Na-THF system, the cation-complexing
bicyclic diamine ether (2,2,2-crypt), a more powerful complexing agent than
crown, was added. The complexing agent, 0222, is expected to enhance the
concentration of Na' and solvated electrons according to the equilibria suggested

by Dye [DY73]:

2M(s) : M‘+M‘ <4-6)

67

M5 : M °+e' (4-7)

-) , _
Moe M +3 (4-8)
M"+C: M*C (4-9)

Additionally, there may be a small increase in the concentration of monomer,
which might yield a detectable monomer dark spectrum. By this effect,
Friedenberg and Levanon [FR77] were able to observe a dark spectrum of
potassium monomer with K-THF solutions when a small amount of crown either
was added as a cation-complexing agent to their solutions. However, adding
large amounts of complexing agent may produce other reactions, such as ion-
pair formation, which cause the monomer to be undetectable [FR77]. Care must
be taken in sample preparation since, only as long as crypt, C, is in excess, is it
certain that both Na- and e- will be observed. For a saturated solution of sodium

and 0222, the em.” concentration depends upon the equilibrium:

Na-+C: Na+C+Ze- (4-10)

Experimental evidence indicates that the equilibrium constant for this reaction
heavily favors the formation of Na'. Dye [DY71] reported observing the
disappearance of infrared absorption band of em”) in a saturated solution while
observing an off-scale absorbance at the Na' absorption band. Lacoste [LA76]
has estimated the equilibrium constant for the equilibrium in Equation 10 as
approximately 10'”. However, it is difﬁcult to totally saturate the Na-THF-C222

solution because of the time required to complete the saturation process.

68

Therefore, there will be both an excess of 0222 in the solution and an e'(so.,,)
concentration above the em], concentration.

The details of the preparation of reagents, glassware and the samples
used in this study are discussed elsewhere [DA79, L072]. To retard sample
decomposition, after preparation, the Na-THF solution was kept at liquid nitrogen
temperature until the EPR instrument was set up. The experiments were
performed at -20 0, again to retard sample decomposition. The sample
temperature was maintained by passing a regulated stream of cold nitrogen
around the sample which was in a quartz dewar in the EPR cavity. The
temperature was measured using a copper-constantan thermocouple.

Figure 4-1 is a dark spectrum of the Na-THF-0222 solution recorded with
the 2 MHz EPR spectrometer. The sweep range is 10 Gauss. The observed
singlet which is distorted by the 2 MHz modulation frequency (see Chapter 3), is
attributed to the em”). In an attempt to observe hyperﬁne splitting due to the
sodium monomer, the spectrum of Na—THF was scanned in increments of 10
Gauss, over a 50 Gauss region 0 either side of the e130", resonance line. The
presence of a monomer escaped detection. Based on the reported coupling
constants for K and Rb in THF (25 and 150 G [FR77]), resonance lines for the
sodium monomer should fall within the region investigated.

An experiment to determine the kinetics of the photoelectron by monitoring
the em”) resonance line was undertaken using the time-varying data acquisition
procedure discussed in Chapter 3. The magnetic ﬁeld was adjusted to allow
observation of the e130») resonance line near its maximum absorption and then

the computer, running program TIMEDA, ﬂashed the lamp and acquired the time-

69

55388 2% E: N a? Bass: “$05532 .s 558% < ”3 05mm

 

 

 

 

70

INTENSITY

SIGNAL

 

0.000 1.000
TIME (SEC)

x1e “1

Figure 4-2: The e’mm signal as a function of time after a perturbing illumination
pulse recorded with 2 MHz modulation and a 1 ms ﬁlter time constant.

71

varying data. Figure 4-2 shows a decay curve obtained from 16 averages using
a response time constant of 1 ms. The disappearance of the photoelectron signal
was extremely slow making averaging (the only method available to increase the
SIN lost when the response time of the instrument was reduced) a very lengthy
process. The lamp-trigger signal was delayed an additional 10 seconds after the
data were collected, but this time interval was not nearly long enough to allow the
total decay of the e130.” line. The long decay time of this signal is not best suited
for study using a 2 MHz instrument because of the need to average to obtain a
reasonable SIN (for random noise, SIN improvement is proportional to the
square-root of the number of averages). The SIN ration is needlessly reduced in
this kinetic experiment by using the wide bandwidth of the 2 MHz spectrometer.
Since the microwave cavity being used was a Varian cavity with 100 KHz
modulation coils on the sidewalls and a 2MHz internal modulation loop, the
Micro-Now 2 MHz EPR ﬁeld modulation and phase detection electronics were
replaced with the Varian 100 kHz modulator and phase detector and the
experiment was repeated using the analog output normally used for the 100 kHz
oscilloscope display. The time constant for this signal output is 300 us; however,
additional RC ﬁltering (20 ms) was used. The 0.3 and 20 ms time constants
were experimentally veriﬁed.

A dark spectrum of the Na-THF solution at —20 0 recorded with 100 kHz
modulation is show in Figure 4-3. Because of the decreased modulation
frequency (sidebands at 36 mG intervals instead of 720 mG), the ego“,
resonance line appears as a conventional ﬁrst derivative signal. Again, scanning

around the em") signal did not reveal the presence of the monomer.

72

.8332: 22.“ N5 2: a? 8282 aboiaazs 58% < ”3. 05mm

 

 

 

73

..-_____..—-
__’_—

                            
                 

  

as .

k1 \VN‘vax =~
: ==ﬁ==== /////
///// 203% ‘—
S— a
85—. ...-
u-

. =\=

  

  

/////// //
...; , //// 44/////////46z%%%§5§.¢v~.== xx
x/ a”? ...: \\\\
2,2, _\_\.\___xvvmwﬁﬁmmwﬁﬁwwwwwwﬁﬁE.-n
w\\\\\\
. \\\\\\§\\\\ RR ‘ihiih
s§§§§.ES~==--------n-u-n-nu.-nun-Ill..-
=== ...—....nnnnun I

        
  

///¢////
5::

.....wavv$\\\\\
.. _ \\\\\\ \\
\\ \\\\\\\““\\\\\\\\

Msswxxxxssx E:
\EESE.======_ ==
...—5:...“ ...—......—
“ ...-E...“
III I

\

 
 

 

     

   

   

  
            

.. \..u§\\v
. M. _\\\\\\$\\\\\\\\\\\\
Sc\SSS‘SS:E.&.~.~.~.~..—~__.=...-
==§_.=_......uﬁ......=.==.====n======
____ =_====_===.== ======. ====-
==_====_.....====_...—.5...- =======I=
_. , ... E. .
ill- \lllilllil.‘ . .

 

  

1“

 

.2. ===

FiQUre 4.4.
Olved Spectru
m of

mOdUIation.

e'
’ (solv) employing 100 kHZ ﬁeld

74

A time-resolved spectrum of the 6mm resonance was recorded using
computer program TIME12 and is shown in Figure 4-4. The data were recorded
with a time constant of 20 ms and at intervals of 0.1 s (5 times the time constant).
The ﬂashing of the lamp was delayed an additional 20 s after the data were
collected to allow for a total decay of the signal. No averaging was done, but the
data were digitally smoothed before plotting. Another method of displaying the
information encoded in the time-resolved spectrum in Figure 4-4 is to extract
information from the time-resolved data set to reconstruct an EPR spectrum at
various times after the ﬂash. Program DSTRIP will remove this information as
described in Chapter 3. Figure 4-5 contains six such spectra. The spectra show
the em") resonance line starting at 0.1 s and continuing at 10 s intervals after
the ﬂash. Figure 4-6 shows the decay of the e'mw) signal recorded near the
maximum of the resonance under the same experimental conditions as the time-
resolved spectrum in Figure 4-4 except that this plot is the result of 16 averages.
These data have not been smoothed. The calculated tm for this decay process
$25s

Because the decay of the photolytically produced species is slow, an
attempt was made to observe the monomer by scanning the magnetic ﬁeld while
manually triggering the ﬂash lamp every few seconds. For a long-lived species,
this process should build up an observable population of the radical; however,

the experiment produced negative results.

75

 

Figure 4-5: Spectra extracted from the time-resolved data set in Figure 4-4
showing the spectrum at 0.1 ms after the ﬂash and continuing at 10 5 intervals.

76

INTENSITY

SIGNAL

0.000 5.000
TIME (SEC)

x19 '1

Figure 4-6: The decay of the em...) signal as a function of time recorded with 100
kHz ﬁeld modulation and a 20 ms ﬁlter time constant.

77

Discussion

The absence of a hyperﬁne multiplet for sodium, as well as for some other
alkali metals in THF, has been previously observed. Catterall et al. [CA77] stated
that EPR spectra indicate the higher alkali metals in THF will yield a singlet from
the solvated electron and a hyperﬁne multiplet from the monomer while spectra
of lower alkali metals will be a single resonance line. Caterall described this
phenomenon as a time-average of the two possible environments for the
unpaired electron. He indicated that even spectra of potassium solutions have
been reported as a single line. However, since that article, EPR spectra
containing hyperﬁne splitting attributed to potassium monomer have been
reported [FR77]. Hyperﬁne splitting of K in THF was observed under continuous
illumination without the addition of a cation-complexing agent and was also
observed in the dark spectrum with the addition of a small amount of crown
ether, CR, as a cation-complexing agent. However, at CR concentrations above
10" M, a monomer spectrum was no longer detected. Friedenberg and Levanon
[FR77] suggest that the disappearance of the monomer spectrum is due to ion-
pair formation, which exchanges with the monomeric species yielding a single
average resonance line. When documenting the detection of the potassium
monomer, F riedenberg and Levanon [F R77] included a spectrum of a K-Na-THF
solution, which showed only the same lines as the lines observed in the pure K-
THF solutions. Even under conditions that produced a K spectrum, Na was not

observed. This does not, however, imply that the same results would be

78

obtained with the addition of cryptate, a much stronger cation-complexing agent.
Under the proper reagent concentrations, the monomer may be observed.

If hyperﬁne splitting from sodium is to be observed, the chance of
detecting it would increase by studying the solution at room temperature or
above rather than at —20 C (253 K) which was used during these experiments.
Friedenberg and Levanon did not observe the potassium monomer below 230 K,
while at room temperature; they observed hyperﬁne splitting in a dark spectrum
of K-THF with crown ether added.

There is a limited amount of reported research on the kinetics of sodium-
ether solutions. Studies have been hampered by the limited solubility of sodium
in these solvents. Glarum and Marshall [GL70] reported a kinetic EPR study of
potassium, rubidium and cesium; however, they were not able to dissolve sodium
in DME. DME left in contact with sodium for several days, was pale green in
color and absorbed light at 480 nm instead of the usual 670 nm Ma- band.
However, the solvated electron signal was observed and its amplitude increased
upon illumination; but they believed that the species in solution were
decomposition products rather than the desired Na’.

To prevent irradiation of the 480 nm absorption band attributed by Glarum
and Marshall to decomposition products, a red ﬁlter was inserted between the
ﬂash lamp and the sample. The addition of the ﬁlter had no effect on the
observed results. The calculated tm for the decay of the e150") resonance line in
the Na-THF solution reported in this study is 25 seconds - much longer than was
expected. However, the sodium solutions previously studied did not contain a

complexing agent. Friedenberg and Levanon [FR77] studied potassium,

79

rubidium and cesium in THF with crown present and reported that the
recombination rate constants were dependent on the crown concentration. They
proposed the process below to account for the decrease in the rate constant with

an increase in the CR concentration:

e‘+MCR‘ ->elMCR* (4-11)
This process competes with the recombination process:
6‘ + M+ -—» M (4-12)

The proposed ion pair (e', MCR“) would have a broad EPR signal that would be
masked by the solvated electron signal.

Friedenberg and Levanon [FR77] reported a factor of 13 decrease in the
observed rate constant for potassium in a 2x104 M CR solution compared to a
solution without CR yielding a t"; of 0.1 s.

Since the reaction rate recorded for Na-THF is still signiﬁcantly longer than
that reported for potassium above, it is possible that we are observing a different
phenomenon. The cryptate, 0222, is in excess in the Na-THF solutions. Thus,
the species in solution are Na‘C, Na', 0, e130“), Irradiation with light containing
wavelengths around 670 nm, the region of that Na- absorption band, will remove
one electron [HU70, GA70]. The monomer produced is in an excited state and
will decay by losing another electron. The newly created cation has two reaction

paths to follow as shown below:

80

Na
JrhV

Na+=2e' : Na' (4-13)
+

C
it

Na+C

The complexation reaction between Na+ and 0222 is known to be fast. An
estimate for the recombination rate for sodium can be obtained from the ﬂash
photolysis work of Kloosterboer et al. [KL71]. They optically observed a tug
approximately equal to 0.15 s for the disappearance of the e130,.) signal. Even
though it is generally accepted that the product being formed was K' due to
potassium contamination rather than Na, the observed lifetime can be used as an
upper limit for the sodium recombination since K‘ formed more rapidly than Na-.
This number is only an estimate because both the solvent (DME) and the
temperature (-60 C) in Kloosterboer’s study were different that in the current
work. The e'(so.,,) will not displace the Na+ from the crypt; therefore, the
disappearance of the e130”) signal would depend upon the rate of release of the
Na+ from the crypt. This release rate via the proposed mechanism below has

been measured using Na23 NMR [01577].

+ -* +
Na C ‘_ Na +C (4-14)

The release rate of Na+ from 0222 was determined at 298 K for the Na-THF-

0222 solution and was reported along with the thermodynamic parameters AHJ,

81

A35, and AGOI. Using these parameters, AGO¥ and the release rate constant (k-

1) and tug at a different temperature can be calculated by Equations 4-15, 16,17:

AG6=AH6-TASB (445)
T u
K_1= Ef—e‘AGo/ RT (4-16)
0.693
= — (4-17)
T1 / 2 K_1

The calculated Tm at 253 K is 6.5 s. This number is respectably close to the
measured 25 s Tm. The reaction rate measured was probably the rate of

release of Na+ from 0222, not the recombination rate of Na. and 2e'.

Recommendations

The complex chemistry of alkali-metal solutions in ethers and amines is
not completely understood. There is still much controversy about the nature of
the species in solution. Many mechanisms have been proposed to explain the
observed and unobserved experimental evidence. A systematic transient EPR
investigation of sodium solutions (the alkali metal lease studied) could possibly
reveal some very important data on the species in solution. A series of sodium
solutions, with varying concentrations of cryptate, could either allow the
observation of the monomer under the “proper” reagent concentrations or provide
very strong negative evidence that would support ion-pair formation and

exchange mechanisms.

82

The study of alkali-metal solutions could be greatly enhanced by the
addition of an optical spectrometer to the EPR system to allow direct comparison
of EPR and optical kinetic data. A simultaneous optical and ESR spectrometer
(SOESR) [WA76] could yield the correlation between optical and EPR data
necessary for understanding the chemistry involved in alkali-metal solutions.

The 2 MHz EPR spectrometer could be enhanced considerably by adding
a pulsed laser source for the kinetic studies. The biggest problem with the 2
MHz spectrometer is the signal-to-noise ratio. The current experiments were
performed under conditions near the limit of detection for this instrument. The
addition of an intense, collimated source might produce a large enough signal to

make the 2 MHz instrument a viable kinetic tool.

83

CHAPTER 5

Characterization of the Flash Lamp

Introduction

Since the mid 60’s, EPR has been used to investigate transient radicals
generated by photochemical, electrochemical, and electron irradiation processes.
More recently, it has become possible to obtain kinetic information on radical
reactions, which can be carried out in the EPR sample cavity. Modiﬁcations to
existing EPR instrumentation have enabled experimenters to follow rapid
changes in radical concentrations as a function of time. Free radicals
investigated in this way have been produced by ﬂash photolysis [BE67, AT70,
Fl67], pulsed electrolysis [6071, G074], and pulsed radiolysis [FE73, SM71].
The technique involves applying a perturbation pulse to the sample and
averaging the transient response obtained from the EPR.

Early ﬂash photolysis experiments were carried out using a continuously
operated light source in conjunction with a rotating slotted disk to provide
intermittent illumination of the sample [GA70, HAM70]. The main limitation to this
approach is the mechanical problem associated with the very high-speed rotating
disk, which is necessary for observation of radical lifetimes on the order of ms
and below. Since the perturbation pulse must be signiﬁcantly shorter than the
lifetime of the radical under study, the rotating disk method is restricted to studies
of long-lived species. Several investigators have overcome this problem by

using lasers [AT73, HAL72] to produce intense light pulses of short duration (<1

84

us). These short, intense light pulses make a laser an ideal irradiation source,
but unfortunately, lasers are only available in a relatively narrow range of
wavelengths, which does not include much of the ultraviolet, a region in which a
great number of photochemical reactions occur.

This chapter is devoted to the approach taken by this investigator — a fast-
extinguishing ﬂash lamp that has a spectral output from the ultraviolet to the
infrared. In this chapter, the ﬂash-lamp trigger and ﬂash-detection circuitry are
presented and the ﬂash lamp is characterized. Finally, the chemical photon-
counting technique called actinometry, used to measure the ultraviolet light

intensity available in the EPR cavity, is discussed.

The Flash Lamp

The ﬂash lamp was purchased from Photochemical Research Associates,
University of Western Ontario, London, Ontario.

The circuit in Figure 5-1 shows the ﬂash lamp with the associated high
voltage supply and trigger circuitry. A thyratron was placed in series with a
Xenon Corporation Suntron 6C ﬂash lamp to give control over timing of the ﬂash
and to allow the ﬂash lamp to be operated at voltages above the normal hold-off
voltage of the lamp. In this circuit, capacitor 03 is charged to 400 Vdc by the
voltage doubler made up of diodes D1, 02 and capacitors C1, 02. A 5-volt pulse
at the trigger input of the lamp causes the SCR, S1, to turn on, discharging

capacitor 03 through me primary of transformer T3. The secondary of T3 is

85

.3580 Emmi“ 3308mm use 952 amen 65 330% :8..me 3:85 ”Tm oBmE

 

 

     

 

 

 

 

 

 

 

 

 

 

 

 

 

 

zoZomzzoo Ar 9&3 a 2
35...; 131‘ be; u 2.
«3&3 u C
2325:: 0 mcmxcoemzée
ozoxz<z
sooez. u mo._o
e:~o.o
. 1%:
m 7M
“Y
o I. H0 .2203
we muooame
...as a.
3.
92 22.. u. 3H3 mo
M... 1H:
.33.on user—Lo 8; m2
aims» - xii/M. - Z
n. em om: m 3

86

connected to a wire wound a couple of times around the body of the ﬂash lamp.
This coil starts the ionization of the ﬁller gas in order to insure a reliable trigger of
the ﬂash lamp. At the same time, a trigger pulse to the thyratron causes it to turn
on and discharge the main storage capacitor, 04, through the ﬂashtube.

The Suntron 60 ﬂashtube is a fast-extinguishing lamp that will produce a
spectral output, which is an almost pure continuum above 3000 Angstroms when
operated at high current density [XENO]. The recommended operating condition,
taken from manufacturer’s literature, is 10 W with a 1 [IF capacitor. The
maximum average power that the lamp will dissipate safely in free air is 100
watts. The ﬂash-lamp housing contains a forced-air cooling system to help
dissipate heat generated by the lamp.

The computer controls the timing of the ﬂash lamp through the circuit
shown in Figure 5-2. A trigger command is issued by the PDP 8IE computer by
moving a digital word which contains a 1 in bit 11 to the digital output buffer
provided on the laboratory peripherals option of the computer. The rising edge of
bit 11 triggers a monostable wired for a 20 us pulse. Twenty microseconds is an
experimentally determined minimum time, which produces a reliable ﬂash-lamp
trigger. The monostable pulse is then driven to the receiver box at the ﬂash lamp
by a noninverting buffer. In the receiver box, an optical isolator receives the
pulse and causes the ﬁnal output transistor to conduct the current pulse
necessary to the trigger transformer in Figure 5-1. An optical isolator was
employed in order to isolate the power and ground in the computer from the

electrically noisy ground of the ﬂash lamp.

87

6:388 m3 can 65 o. ooﬂcﬁﬁ Emmi“. cps—Ame 2:. ”NR oSwE

 

 

 

P P
1:205
macaw—E. .Ai > 2.02 TN“ .... mamﬁmozoz
Edi maﬁa Jr .0 O _< . 2b:—
225! .A II... .C H .....o
.47 . . m: 2
EN
_ m+ m+ m+

m+

88

Since the timing of the data acquisition should start when the lamp
actually ﬂashes instead of when the computer requests a ﬂash, a “lamp-ﬂashed”
signal is generated at the lamp by the circuit in Figure 5-3. An ultra fast response
(<1.0 ns) PIN silicon photodiode is mounted on the outside of the ﬂash-lamp
housing. A small hole was cut in the lamp housing allowing some of the light
from the lamp to strike the photodiode. When the photodiode starts to conduct, a
TTL signal is sent to a Schmitt trigger input on the analog peripheral option of the
PDP 8IE. The analog control board in the computer was modiﬁed so that a
Schmitt trigger pulse could be used to stat the crystal controlled clock without
requiring software intervention. DEC literature indicates that this function is
available on the lab peripherals option; however, it was not previously
implemented on this particular computer. It now works as stated in DEC’s
documentation.

A second photodiode is used to obtain the integrated intensity of the light
pulse so that the variations in each light pulse can be taken into account in the
data manipulation. The circuit is show in Figure 5-3. The capacitor on the input
of the integrator is used to prevent the negative operational ampliﬁer input from
saturating and thereby losing the intensity information. A sample-and-hold
ampliﬁer samples the integrator output after 70 us to allow time for the integrator
to drain all the charge from the input capacitor. The sample-and-hold ampliﬁer
“holds” the voltage output of the integrator until the computer has time to sample
the signal. The integrator is discharged by a resistor across the integrating
capacitor of the operational ampliﬁer. The time constant of the discharge, 50 ms,

is long enough so that it will not affect the integrated light pulse signal before the

89

0\c Oh

0883:: Em: can 883% «scams—«68$ 2E. ”Wm 2:me

nuruc:
DJOIIOZCIMszco

 

 

.Alll

 

unused

 

[4ID>

Imam: c... A
hhnxzom or

f

 

 

coma: I no

 

040: u a
muaxcm a o

NJGChmozo:

 

mos

2%

n+

ﬂuid

 

 

 

 

«roux;
+
.519»
/_ H
“330.9 3+ .4
2.
ill—o
h,
good

90

sample-and-hold ampliﬁer captures the signal (70 us). Since the maximum lamp
repetition rate is 4 Hz with a 10 kV supply voltage, the discharge resistor has
sufﬁcient time to discharge the integrator (5 time constants). The discharge
resistor approach was taken to eliminate the need for another control line from
the computer to the ﬂash lamp to control a shorting switch across the capacitor.
The sample-and-hold ampliﬁer timing is controlled by the ﬂash detection

photodiode.

Characterization of the Flash Lamp

The total energy dissipated by the lamp during a ﬂash is determined by
the capacitance across the tube and the supply voltage to the capacitor

according to Equation 5-1

E=1/2CV2 (5-1)
where E is the energy in joules, C is the capacitance, and V is the applied
voltage. The capacitance of the storage capacitor was determined by using it as
the integrating element of an operational ampliﬁer (O.A.) integrator. A Heath
[HEAT] voltage reference source (VRS Model EU-80A) with an optional current
probe was used as a constant current source to charge the integrator. The test
circuit is shown in Figure 5—4. The OFF-ON switch controls the current to the
integrator and the start and stop of a Heath Universal Digital Instrument (UDl) in

timer A-B mode. The capacitance is calculated using Equation 5-2

91

 

 

.8on 853633 :3me .4. 31m onE

 

 

 

 

 

0\z >¢Jm¢

Lb?

 

 

I
l

 

 

 

ﬁe.

 

 

~\> mo

0mm
CCU
but
Jib:
OHIO
>¢¢n

 

92

C=Q '

_=l

V u;— (52)
where l is the current, t is the time and V is the voltage of the integrator
measured at the end of the integration time, t. The measured capacitance was
0.540 [JR The maximum recommended operating voltage of the lamp is 10 kV;
therefore, the maximum energy dissipated in the lap is 27 joules/pulse. This,
unfortunately, is below the 100 joules per pulse recommended by the
manufacturer for optimum spectral output. The small capacitance is a trade-off
that had to be made to decrease the ﬂash pulse-width to the desired <10 us
value since the size of the capacitance affects the length of the ﬂash.

The characteristic of a ﬂash lamp, which limits the observable reaction
rate is the pulse width. This was determined by placing the lamp in front of the
entrance slit of a Heath (Model EU-700) monochromator and observing the
photomultiplier voltage output on a Tektronix 7623A storage oscilloscope. This
measurement was made at several wavelengths and supply voltages. The
results are shown in Figure 5-5. The full width at half maximum for the pulse is
about 2 us with the entire ﬂash over in an absolute maximum of 10 us at all
supply voltages used.

The most important characteristic of a ﬂash lamp, optics and sample
chamber system is the number of photons of an appropriate energy that arrive at
the sample. This number depends on the intensity of the ﬂash lamp, the
collection efﬁciency of the optics, and the size of the opening in the microwave

cavity. This number is most easily determined using a photon counting solution

in the EPR cavity. The chemical photo-counting technique is called Actinometry.

93

Measured at 360 nm

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Measurd at 450 nm

 

 

 

 

 

 

 

 

 

 

 

 

..f ..-_.._-:-
. ..rf .y "QR-u.-

2 us/D
lamp pulse shape

5. #“v. "

v."

 

 

 

 

 

.m: -, ,.
:’(‘{':II;L“V'

 

 

iv

with a 10 kV lamp supply voltage. The

The ﬂash-
abscissa is 2 ps/div.

Figure 5—5

94

A potassium-ferrioxalate actinometer was used as the photon counting solution.
The ferrioxalate actinometer, rated as the best solution-phase actinometer
available today [CA66], is easy to use and sensitive over a wavelength range of
250-480 nm. A detailed discussion of the actinometer, including its nature,
preparation, puriﬁcation and use, has been published by Hatchard and Parker
[HA56, PA53]. However, a brief discussion will be presented here to acquaint
the reader with the technique. When sulfuric acid solutions of K3[Fe(C204)3] are
exposed to light in the range of 250-480 nm, a simultaneous oxidation of the
oxalate and reduction of iron to Fe(|l) occurs. The quantity of Fe(ll) produced
can be determined by ﬁrst complexing it with 1,10-phenanthroline in an acetic-
acid, sodium-acetate buffer of pH 3.5 and then measuring the absorbance of the
complex at 510.0 nm. The high molar absorptivity of the Fe(ll) complex,
E=1.11x10‘M"CM“, makes it possible to determine very low photon levels.
Stock solutions were prepared, as recommended [HA56], from solid
K»,[Fe(C204)3]-3H20 which had been recrystalized three times from deionized
water. Weighed quantities of the recrystalized solid were dissolved in 0.1 N
H2804 and diluted to volume in volumetric ﬂasks. Since the actinometer is highly
sensitive to normal room light, all solutions were prepared in a dark lab and

stored in a closed cabinet until use.

Preparation of the Working Curve

A calibration curve for the 1,10~phenanthroline iron(ll) complex was

prepared using ferrous ammonium sulfate [C162] as an iron standard instead of

95

using the iron standardization procedure published by Hatchard ad Parker
[HA56]. The Fe(ll) stock solution was prepared by dissolving the ferrous
ammonium sulfate in 0.1 N H2804. A standard 5.0 x 10‘4 M Fe(ll) solution was
prepared from the stock solution. This procedure requires all of the iron in the
standard solution to be in the Fe(ll) oxidation state, so hydroxylamine
hydrochloride was added to reduce any Fe(lll) present. Another departure from
the method of Hatchard and Parker was a 20-fold excess of 1, 10-phenanthroline
used to ensure total complexation of Fe(ll).

The following volumes of the Fe(l) solution were added to eleven 25 ml
volumetric ﬂasks: 0 (blank), 0.5, 1.0, 5.0 ml. Added to each ﬂash was 5 ml of
buffer, 4 ml of 1, 10-phenanthroline, 1 ml of hydroxylamine hydrochloride and
sufﬁcient 0.1 N H280, to bring the total number of milliequivalents of acid to 1.25.
The solutions in the ﬂasks were diluted to volume, mixed and allowed to stand at
least one-half hour. The absorbance was measured using a McPherson Model
EU-700 Spectrometer with an oscillating sample cell. Figure 5-6 shows the ﬁtted
results of the working-curve data. The least squares ﬁtted slope of the working
curve, E, was 1.11x104M'1CM", numerically equal to previously reported molar

absorptivity values [HA56, H069].

Photolysis Experimental Setup

The EPR cavity was removed from the microwave bridge of the EPR
instmment and clamped to a ring stand on a lab bench for easier access. The

task of focusing the lamp on the sample in the Micro-Now EPR cavity was a very

96

ABSORBANCE

 

 

0.00001' 0.0 0001b
FE(II) MOLARITY

-ru-
.4
o-l-

Figure 5-6: A calibration curve of absorbance of 1,10-phenanthroline Fe(ll)
complex as a function of Fe(ll) molar concentration.

97

difﬁcult one. According to manufacturer’s literature, previous Micro-Now cavities
were designed for coupling with a linear accelerator beam. Such a well-
collimated source does not require a large opening in the sample cavity wall.
The actual opening in the cavity wall of this micro-Now cavity is a rectangle 0.6
cm wide by 0.2 cm high. To prevent the microwave in the cavity from leaking out
through the hole, a cylindrical waveguide “beyond cutoff” was welded around the
rectangular hole in the cavity wall. The cylindrical waveguide is 0.6 cm inside
diameter and longer than one half wavelength, 3.0 cm. It is too long and too
narrow to support a standing wave at the frequency of cavity resonance and thus
no microwave will leak out of the opening.

This small entrance would not present a problem for a collimated source
such as an electron or laser beam; however, the ﬂash-lamp source used in these
experiments is not collimated. Neither is the lamp a point source, but instead a
column and the optics of the lamp focus to give a reduced size image of the
column at 3.5 cm from the end of the lamp optics. This would not be a problem
in an EPR experiment since the height of the focus image, 1.2 cm, is shorter than
the active sample region in the microwave cavity. However, the geometry of the
opening in the Micro-Now cavity prevents optimum focusing. The problem is the
small area presented by the end of the cylindrical waveguide.

Observation of the shape of the light beam at various distances from the
end of the lamp optics revealed that at 1.5 cm, the beam had converged to its
smallest diameter before diverging into a focused image. A diagram of the light
beam is shown in Figure 5-7. The ﬂash lamp was placed so that the entrance to

the waveguide is 1.5 cm from the optics. A considerable amount of light was lost

98

dEm. can: 05 .6 Emma Em: 63:02 on. ..o E985 £6 2:9“.

 

Eu m;

 

 

 

 

 

-h

 

 

 

 

 

“Eon—
Room

 

 

 

 

 

 

99

because the minimum beam spot is still not as small as 0.6 cm. As the light stats
to diverge in the cylindrical waveguide, the internal surface of the waveguide
should act as a light pipe and reﬂect the light toward the sample. Once the light
is in the waveguide, the limiting parameter is the height of the slit presented by
the rectangular opening in the wall of the cavity. The slit width, 0.2 cm, will only
allow about 40 percent of the light in the waveguide to enter the cavity (based on
areas). The study was also conducted in a second EPR cavity to determine the
amount of light lost by using the Micro-Now cavity. This cavity was a Varian
transmission cavity, i.e. openings in both front and back walls of the cavity for
light to pass. The opening into the waveguide and cavity were rectangles 1.2 cm
high by 0.6 cm wide. This made it much easier to focus the lamp and, as it

turned out, allowed more light to enter the cavity.

Photolysis of the Actinometer Solution

The wavelength sensitivity of the potassium ferrioxalate actinometer is
concentration dependent. Plots of absorbance and quantum yield of Fe(ll)
versus wavelength for several concentrations are show in a paper by Hatchard
and Parker [HA56]. Selection of the concentration of the actinometer gives a
natural cut-off ﬁlter. A concentration of 0.15 molar was selected which according
to Hatchard and Parker absorbs all light below 450 nm.

The actinometer solution was placed in a quartz NMR tube and inserted in

the EPR cavity. Flash-lamp supply voltages of 8, 9, and 10 kV were used. Each

100

sample was ﬂashed 50 times to build up an Fe(ll) concentration large enough to
optically measure. The sample was stirred after every ten ﬂashes.

The analytical procedure used for the unknown was identical to that used
in preparing the working curve, except the 4 ml of photolyzed solution replaced
the standard Fe(ll) solution and the hydroxylamine hydrochloride was excluded.

The quanta of light absorbed can be calculated by the following equation
[HA56]:

Quanta = AS V1 V3 (6'023x1 0 ) (5_3)

ELV.

 

where
V1 = volume of actinometer solution photolyzed (5 ml)
V2 = volume of aliquot taken for analysis (4 ml)
V3 = the ﬁnal volume to which V2 was diluted (25 ml)
A. = the measured absorbance of the solution at 510.0 nm
L = the path length of the cell (0.9 cm)
E = the molar absorptivity of the Fe(ll) complex (1.1 1x10‘M'1cm")
Q = the average quantum yield for Fe(ll) formation for light in the range of
analysis (1.2)
The results of this study for both the Micro-Now and Varian cavities are
shown in Table 5-1.
At the time these experiments were performed, the Varian cavity was not
being considered as a possible alternative to the Micro-Now cavity; therefore, a

complete set of supply voltages were not used. The only reason for the Varian

101

Table 5-1: Measured Quanta of Light per Pulse Available in the Micro-Now and
Varian Cavities.

 

 

 

 

 

Supply Voltage Micro-Now Cavity Varian Cavity
(kV) Quanta/pulse) x10'15 Quanta/pulse) x1 0'15
8 1.02 +/- 0.29 3.14 +/- 0.37
9 1.46 +/- 0.54
10 2.05 +/- 0.48

 

 

 

cavity's inclusion was to determine the effect the small entrance to the Micro-
Now cavity had on its total light gathering capability. As expected, the Varian
cavity collected more light than the Micro-Now cavity but the difference is more
than can be explained by comparing the areas of the entrance waveguides of the
two cavities. The rectangular waveguide in the Varian cavity captures a larger
fraction of the total light because the shape is similar to the focused image of the
lamp.

The number of photons per pulse is a disappointingly small number for
both cavities considering the range of wavelength absorbed by the actinometer
solution. Fortunately, most molecular absorption bands are very broad and will
absorb light in a range of wavelengths.

Even though this actinometer study did not yield any speciﬁc data about
the quanta of light of wavelengths of interest for the photochemical experiments
done in this research, it is important for several reasons. The results indicate
that the focusing of the lamp on the sample cavity opening was not as critical as
was ﬁrst thought. The experiments were performed with the lamp at several

slightly different positions with very similar results. This is important because

102

 

when one tries to focus the lamp on the cavity when the cavity is between the
EPR magnet pole tips, it is difﬁcult to see if it is perfectly focused. Secondly, the
ultraviolet region is of importance in a great number of photochemical reactions.
The results of this experiment suggest that such reactions could be studied with
this lamp and cavity system. Also, from the spectral response curve of the ﬂash
lamp, an estimate of the quanta of light in other wavelength regions can be
made. The comparison between the openings in the Micro-Now and Varian
mv'rties suggest that if a new cavity were to be built, an opening similar to the

one in the Varian cavity should be used.

103

CHAPTER 6

System Modiﬁcations

Introduction

The original goal for this research project was the design and fabrication
of a computer interface to a fast EPR spectrometer with a pulsed light source,
which would allow EPR data to be acquired either as a conventional absorption
spectrum or alternately, as a function of time for reaction-rate studies. Computer
programs to synchronize events and to record, average and manipulate the data
were to be written. The resulting system would be applied to the studies of
radical reaction rates in situ in the EPR sample cavity. When it was discovered
that the Micro-Now 2 MHz EPR spectrometer was not appropriately designed for
the fast reaction-rate studies intended, a major change in emphasis for this
research project became necessary. The new goal was to develop a functioning
transient EPR system.

In this chapter, modiﬁcations made to the EPR system will be discussed in
detail. First, the interface design, which was necessary to complete the research
project as originally conceived will be presented. Finally, the modiﬁcations made
to the Micro-Now system in an attempt to create a working spectrometer will be

reported.

104

The Fieldial

To use an EPR spectrometer as a kinetic took, it is necessary to maintain
a constant magnetic ﬁeld at the point of resonance, apply a perturbation pulse to
the sample and record the transient response from the EPR. To obtain an EPR
absorption spectrum, the magnetic ﬁeld is varied and the EPR output signal is
recorded as a function of magnetic ﬁeld strength. The two experimental
techniques have been combined to give an EPR spectrum of a transient radical.
This has been accomplished by both allowing the magnetic ﬁeld to scan slowly
while the lamp is ﬂashed and the data are acquired [AT68b, F I67, H072] and
also by adding auxiliary ﬁeld-sweep coils to the magnet pole tips [Hl68, H873].
Data collected using the former technique were acquired without signal
averaging on multichannel X-Y or magnetic tape recorders. Using ﬁeld-sweep
coils, data are collected using a signal-averaging device, which is synchronized
with both the ﬁeld sweep and an excitation pulse to produce data collected as a
function of time over a narrow magnetic ﬁeld range. There are several problems
with the latter technique [3039]: 1) the sweep ﬁelds are often nonlinear with
time, 2) the auxiliary ﬁeld may disrupt the static magnetic ﬁeld, 3) l-r heating of
the coils limits the sweep ﬁeld to a range of +I- 25 to +I-100 Gauss from the static
ﬁeld, and 4) ﬁeld homogeneity deceases as the applied ﬁeld increases. The
technique is also limited in the rates of reactions it can follow.

The system discussed above does not beneﬁt substantially from the use

of a computer because there is no provision for allowing interactive control of the

105

EPR spectrometer. The addition of a computer does not eliminate any of the
limitations, but does, of course, simplify data acquisition and reduction.

More recently, Goldberg et al. [G075] reported an interactive computer
controlled system, an 0. S. Walker magnet in conjunction with an FCC-4
controller, which has an external analog voltage input that allows remote control
of the magnetic ﬁeld. Varian, in fact, offers an accessory for their Fieldial that
allows external control of the magnet ﬁeld sweep rate; however, it requires that
the Fieldial be modiﬁed at the factory. S. H. Glarum [GL73] reported a circuit that
he installed in a Varian Fieldial Mark I allowing replacement of the original
Fieldial potentiometer with a more accurate external potentiometer. This
published circuit was modiﬁed to enable input from a DAC to replace the
potentiometer. Some of the electronics were updated and the circuit was
installed in the MSU Fieldial Mark II.

The basic circuit for the externally controlled Fieldial is show in Figure 6—1.
A 1290 Hz oscillator excites a loop, which contains a Hall-effect transducer, the
ﬁeld selector resistors, and a transformer, T1, that drives the modulator circuit.
The modulator circuit is basically an analog multiplier with the 1290 Hz sine wave
at one input and the DC sweep-input voltage applied to the second input through
a DAC. The product of these two signals appears at the output, passes through
the sweep-range selector and is combined with the outputs of the Hall-effect
transducer and ﬁeld selector to produce an error voltage that is used to maintain
the desired magnetic ﬁeld.

The modulator was installed in the Fieldial Mark II at the corresponding

position chosen by Glarum [GL73] for the Fieldial Mark I circuitry. The modulator

106

cog—$2 22m 0:0:me 3203 3:888 33585 05 mo oumﬁonom 3538mm 25. "Te 2&5

umzcc awmzm

0‘

 

 

 

moo—...

 

 

    

 

 

 

  

 

mmumnamxc
macaw 0...

Al...

 

 

«00¢...

   

 

 

 

 

«00‘...

 

eaezn
muuxm
_|||l0|r._¢zmuexm
1. H
" Jocezog
" 04m;
muczcmexy
w . E
m m haezu
u unocau~z emu.
Hm HUN...“ 1
m muomezoom
u hum"
“ cause"

   

 

 

107

circuit with transformer T1 is shown in Figure 6-2. The heart of the modulator is
a Burr Brown 4204 four-quadrant analog multiplier. The output from the digital-
to—analog converter (range = +I-5 volts) is sensed by an Analog Devices A0521J
instrumentation ampliﬁer conﬁgured with a gain of 2. The resulting output (range
= +l-2 volts) makes full-scale use of the X analog-multiplier input. The 1290 Hz
reference signal is connected through a double-pole double-throw switch to a a:
step-up transformer and a reference-level potentiometer to adjust the reference
signal to 20 volts peak-to—peak. However, because of differences in the circuits :-

of a Mark I and Mark ll Fieldial regulator, this signal was a maximum of only

 

about 10 volts peak-to-peak. The output of the analog multiplier (X*Y/10) passes
through a phase-shifter ampliﬁer to the output ampliﬁer. The output ampliﬁer is
connected to the sweep-range selector by a second double-pole double-throw
switch. The two switches were inserted in the circuit to allow the selection of the
standard Fieldial arrangement. This was necessary because the computer used
as the controller in this research is used by many other researchers in the
department and could not be dedicated to the EPR spectrometer. Without the
switches to return the Fieldial its original condition, the EPR would be inoperable
whenever the computer was needed elsewhere. The circuit diagram in Figure 6-
2 does not include the trim potentiometers necessary for the analog multiplier or
offset and gain-adjust potentiometers for the various ampliﬁers.

Connections were made to the Mark II F ieldial by inserting the modulator
circuit between point 1 of the CB1101 oscillator and phase detector reference
assembly and point B of switch $1007 and point F of J1001 and the wire that

eventually goes to terminals 6, 6 of transformer T1003. The labels here and in

108

.20: ozocmmE 05 B .9280 .mEowxo @5595 .oﬁSuoE 9: ho Ema—av «Soto ”N6 939....

c— H ~—
—N@VZN ﬂ NO
m—avZN H —0
mm- Saar—Z: .Zéhtcme—z ﬂ uh.
Hunm .>HG UOA<Z< ..l.. <—
x m. NV
Anne“: 8. .35 p

    

A .h
WW7:

 

 

 

 

 

imam
53$ 9 es
134 v.2 chaN—
m A1 . as
a a
5 :2 8.
:2
3+
5. 9 5.2.
2x

 

mmmamunAm—E

109

Figures 6-1 and 6-2 refer to the labels on Varian publication number 87-165-708
for the Mark ll F ieldial Regulator. Once inserted in the Fieldial, electrostatic
balance, signal phasing and reference tuning were performed according to the
procedures outlined in the Varian Fieldial manual. Adjustment of the analog
multiplier was performed following the manufacturer’s instructions.

The DC analog voltage is derived from a 10-bit DAC within the computer
and gives a resolution of slightly better than 0.1% of the ﬁeld-sweep range
selected. This resolution is adequate for the small ﬁeld-sweep ranges used in
this research, but at a sweep range of 1 k6, the resolution is only 1 G, which is
wider than many EPR lines. The DAC should be replaced with a 12—bit DAC
(resolution of 1 part in 4096) for general use.

The linearity of the ﬁeld sweep was determined by applying a DC voltage
to the ﬁeld-sweep input of the analog multiplier and measuring the resultant
magnetic ﬁeld with a home-built gaussmeter which utilizes the NMR resonance of
the proton. The results are show in Figure 6-3. The nonlinearity in the center of
Figure 6-3 is due to the nonideal behavior of the analog multiplier. When the
sweep-input voltage is zero, the analog multiplier output should be zero;
however, a small amount of the signal at the second input feeds through to the
output (maximum of 10 mVp_p for this multiplier). The nonlinearity could be
reduced by replacing the Burr Brown analog multiplier with a unit having a better
feed-through speciﬁcation. A linearity test at sweep-width settings of 10, 100,
500 and 1000 Gauss indicated that the linearity within =l10.25% of the sweep
width. This value is better than the linearity speciﬁcation quoted by Varian in its
F ieldial literature (0.5% of sweep width).

110

GAUSS

 

lJJlIJIlJlJlI
T II

I I I I I l I l l

VOLTAGE

T I I 1 l I : l T
Figure 6-3: The magnetic ﬁeld as a function of the external sweep voltage at a
sweep width of 100 Gauss.

111

 

The short-term stability of the ﬁeld was difﬁcult to measure with the home-
built gaussmeter used in this analysis since it was not a tracking gaussmeter and
had to be manually adjusted. The frequency to establish proton resonance was
selected using a variable capacitor. 0ne’s hand had to be kept on the
adjustment knob during the entire integration period of the frequency meter to
keep the frequency from changing, and this prevented continuous monitoring of u-
the magnetic ﬁeld. The short-term stability results are based on ﬁve readings
made over about one minute. At sweep-range settings of 50 and 100 Gauss,

with zero volts applied to the sweep input, the maximum ﬁeld deviation seen was ,

 

25 mG. This change corresponds to 7 parts per million in the measured
frequency. Considering the manual method of adjusting the frequency of the
gaussmeter, this error may not be signiﬁcant. The error value is about 25 times
larger than the value reported by Glarum [GL73]. Wrth a voltage input of +5

volts, the maximum ﬁeld excursion detected was 35 mG.

Fieldial Error Signal

After a new magnetic ﬁeld has been selected by the computer, it is
necessary to know when the Fieldial has established the correct ﬁeld. The best
approach to this problem is to insert a tracking gaussmeter into the ﬁeld and
monitor the ﬁeld strength until it stabilizes; however, such a gaussmeter was not
available. An indirect approach to determining the ﬁeld stability is to monitor the
Fieldial internal error signal. The error signal was experimentally determined to

be an AC signal with an average voltage near zero when the Fieldial is at null.

112

The small AC signal and its common are superimposed on a constant -58 volt
signal. When the ﬁeld is changed, information appears on both the common and
signal lines so it was necessary to compare both signals with a differential
ampliﬁer to obtain a valid result. The circuit built to process the error signal to
produce a signal useable by the computer is show in Figure 6-4.

The differential ampliﬁer is an Analog Devices AD521J instrumentation
ampliﬁer. The error signal and common are capacitively coupled to the inputs of
the ampliﬁer to remove the —58 volt offset that is only present when the Fieldial

power is switched 0N. Diodes to the power supply of the instrumentation

ampliﬁer protect its inputs when the Freldial is being used without the computer
for control. The output is connected both to a comparator and to a half-wave
rectiﬁer circuit. The output of the half-wave rectiﬁer is divided to ensure that the
voltage does not exceed the maximum limit of the ADC input (+I-5 volts). The 1
M0 resistor shown connected to the output terminal is the terminating input
resistance of the ADC in the computer. The analog output is not ﬁltered enough
to be a smooth DC signal but contains a component of the 1290 Hz excitation
signal that is used to generate the error signal. Therefore, random sampling of
the signal with the analog-to-digital converter might never yield several points in
a row with the same value. It was necessary to add a comparator with a variable
trigger threshold to the ampliﬁer output to generate a clock pulse to control data
acquisition. The period of the clock signal is approximately 0.38 ms. To keep
the comparator from ringing as it changes states, it was necessary to add
hysteresis to the comparator. The comparator output is connected to a Schmitt

trigger input on the laboratory peripherals panel of the computer and used as a

113

..mcmﬁ .26 $595 66.3”. 05 o. canto: ESano on...

     

.1

v— EN

 

¢OP<¢<mEOU
:m E.—

 

v.2

 

b

 

”v.0 9:9“.

 

 

 

 

 

 

v— m v— b.~
>- (((zl(/\(r{((z >+ 332— u :—
b 0.2: n 3
hate... u G
.>I
523.54.. _ a ﬁn _ a
20.5.5sz352. :33.
:11A
.. 3. ..u
t
r M E S. .o
A “TIA
a

 

114

hardware start for the ADC. To display the error signal when the magnetic ﬁeld
is suddenly changed, a minor variation of the program TIMDA (T lMed Data
Acquisition) changes the DAC output, which causes the magnetic ﬁeld to change
and immediately starts to acquire data from the error signal output. Three
examples of the error signal resulting from an instantaneous change in the
selected ﬁeld are shown in Figure 6-5. Figure 6—53 was recorded after a 500 G l‘
increment while a scan-width range of 1 kG was selected. The noise level at the I
1 k6 scan width is considerably higher than the noise at the 500 G scan width

selected for the other two plots. This is probably due to noise on the DAC sweep

 

signal and modulator electronics since at the 1 k6 setting, the Fieldial is more
sensitive to small changes in the reference signal than it is at smaller scan
widths. Figure 6-5b is the result of a 250 G increment at a 500 G sweep-range
setting. The ﬁnal plot (6-5c), is a 5 G change at the 500 G sweep-range setting.
A 5 G increment is larger than a normal change of ﬁeld when recording a
spectrum. The large change in the ﬁrst two plots could be the result of jumping
from one end of the sweep range to the other to repeat a scan or to initialize the
ﬁeld before starting a scan. A subroutine called ERRSUB (ERRor SUBroutine) is
called by the main program before EPR data are taken to ensure that the ﬁeld
has settled. When called, subroutine ERRSUB acquires groups of 100 points,
divides by 100 and compares the result to previously acquired results until three
consecutive results agree to +l-1. When null has been determined, ERRSUB ,
returns to the calling program. The minimum time for null to be declared is 0.11
seconds. The ERRSUB approach may fail if the magnetic ﬁeld change is large.

In Figure 6-5a, sections of the curve are ﬂat because the electronics in the error

115

FIBER...“ Cantu... [III—al.- L.

6ch @826 0 com a 8 “coached O m < 3 emcee 30% 0 com a E 808205 0 omm < 3v .owqﬁ 303m 00—
a a E 82565 0 com < $3 #58805 20m 030% a Sam 08% mo conga a mm 83 gasses 32008.85Q8o0 6.0 0.5mm“—

Amv msz Amy mzﬁk
..m od 3 _ C .
1 . q _ . .

‘1—0

p P F p
q

. _ _
d1 a - d H a d - - 0 u q 1

4+0

.o

P
.-

 

P p L b — P P . p p p P P b h -
u q 1 dr 4 1 ﬂ 1 .—r a a

_ .eb
dﬁ-jiﬁ‘

 

 

 

E

E

 

I J L l l l l L ‘

WVNQIS

ALISNBlNI

116

 

382038 32038 3:323: 333683 a: 5MB @8— coumivoﬁ $23 98 comm—nae .538 NE N 22. 6-9 95mm;

 

.....q

£004
ZOHFH ..H. has I§om~

no:
><o mma.o

NQH

117

 

detector circuit have saturated due to the large error. These ﬂat sections may be
incorrectly interpreted by ERRSUB as null. To prevent this from happening, the
starting ﬁeld-strength value is always determined ﬁrst through the introductory
dialogue and updated immediately to allow time for the ﬁeld to stabilize while the

remaining parameters are speciﬁed by the operator.

The Flash Lamp

The details of the interface to the ﬂash lamp and the tests performed on
the ﬂash lamp are discussed in detail in Chapter 5 and will not be reiterated in

this chapter.

The 2 MHz Modulator

The modulation system consists of a 2 MHz crystal oscillator/preamp, a
power ampliﬁer and the modulation coil inside the microwave cavity. In Chapter
2 it was mentioned that problems developed with this system. After the Micro-
Now EPR had been taken apart a few times to allow the Q-band EPR to be set
up (the 2 MHz spectrometer shares the magnet, AFC, and power supplies with
the Q-band instrument), it would not produce a spectrum of strong pitch. The
problem was diagnosed as a lack of 2 MHz modulation in the cavity. The 2 MHz
power ampliﬁer, including the fusible output transistor, was operating; however,
no measurable modulation was detected in the EPR cavity. The problem was

blamed on improper termination of the coaxial cable carrying the 2 MHz

118

 

modulation to the cavity. The modulation loop in the cavity was approximately
zero ohms and the coaxial cable carrying the modulation was 50 ohms. This
impedance mismatch could cause a total reﬂection of the modulation at the
junction of the cable and the modulation loop. Micro-Now dealt with this problem
by supplying a speciﬁc cable to use between the power ampliﬁer and the
modulation loop that worked with the system. While it was believed that the ...
cable being used was the original cable, something in the system had changed.
Rather than experiment with Micro-Now’s power ampliﬁer and cable, the

modulation system was redesigned.

 

rr‘

The power ampliﬁer was redesigned as show in Figure 6-6 and will no
longer fuse if the output cable is accidentally disconnected while the power is
ON.

To terminate the cable properly, the cavity modulation loop was used as
the inductor in the tuned LC circuit shown in Figure 6-6. The capacitors establish
the resonant frequency of the tuned circuit and also the input impedance of this
terminator. High-Q capacitors were selected to present approximately 50-ohm
impedance at the input thus terminating the coaxial cable at its characteristic
impedance. The resonance frequency of the LC circuit was established by using
a Texscan Model VS-40 RD sweep generator as the frequency source to drive
the LC circuit. The output of the modulation coil was picked up by an antenna
connected through the sweep generator to the vertical input of a scope. The
sweep generator allows a marker frequency to be combined with the antenna
signal to establish an absolute frequency index on the scope display. The 2 MHz

crystal output from the Micro-Now EPR was chosen as the marker. The

119

horizontal scope input was a voltage proportional to the frequency being put out
by the sweep generator. The resulting scope display looks like a bell-shaped
curve with the amplitude of the antenna-coupled signal increasing to a maximum
at the resonant frequency of the LC circuit. The capacitors were had selected to
center the maximum of the response curve on the 2 MHz marker. The Q of the
LC circuit was measured using a variable frequency marker connected to a
frequency meter and measuring the width, in Hertz, of the response peak at the
half-power pint (0.707 Vp) and dividing the central frequency by the width of the
resonance peak. The measured 0 was 20. Since 0 is the energy stored divided
by the energy dissipated per cycle, the circulating current in the tuned LC circuit
will be 20 times the current arriving at the input of the circuit.

Based on the current circulating in the LC circuit, a theoretical modulation
amplitude can be calculated. The modulation loop, which acts as the inductor in
the LC circuit, is shown in Figure 6-7. To calculate the magnetic ﬁeld intensity at
the sample in the center of the modulation loop, each side of the loop can be
treated as a long straight wire. The magnetic ﬁeld at a distance, r, from the wire

is given by [HA62]

 

,6: W (64)

270'
where yo = 4n x 10'7 webers/amp-meter
r = distance in meters (6.4 x 10'3 M)
l = current in amps
Because the current is ﬂowing in a different direction in both wires, the magnetic

ﬁeld at the sample is the sum of the magnet ﬁelds from both wires. This is show

120

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I . 1.27 cm ‘ I
n—n :
U U A I
I
g I
I
I
I
: Modulation
: Loop
I
r
3.10 :
cm I
I
I
I
I
I
I
I
I
I
I
I
I
I
v 1
i
' s 1
I amp e
1' I Position
I
(b) .
Modulation Loop
Viewed ﬁ'om Top

 

 

 

Figure 6—7:The Micro-Now internal 2 MHz modulation loop. (a) A side view
showing the dimensions of the loop. (b) A top view showing the
addition of the magnetic ﬁeld from the individual wires.

12]

 

L’TWi- " --. .

in Figure 6-7 where a black dot represents current ﬂowing into the page.

The current ﬂowing through the wire was determined by measuring the
voltage across capacitor C1, in Figure 6-6, with a high impedance (10 M0)
oscilloscope and dividing by the input impedance of the circuit (50 ohms). Since
the Q of the circuit is 20, the true circulating current is 20 times the current
calculated on the basis of the 50 ohm input impedance.

An ammeter was placed between the power supply for the modulation
power ampliﬁer and the output stage of the ampliﬁer. The theoretical peak-to-
peak modulation amplitude, Hm”), in the cavity was calculated by the above
method for many ammeter readings and a plot of calculated modulation as a
function of power-supply current was made. The plot, shown in Figure 6-8, can
be used to estimate the modulation amplitude in the Micro-Now cavity.

The actual modulation amplitude in the cavity can be determined by
observing the effect of the modulation on the shape of a resonance line. A
detailed analysis of line shapes for Lorentzian [WA61] and for Gaussian lines
[SM64] revealed that a line will be broadened by modulation amplitudes greater
than 0.2 times the true peak width of the resonance line. As the modulation
amplitude is increased, two phenomena occur: the signal amplitude passes
through a maximum when Hm(w) is approximately 3.5 times the peak width for
Lorentzian lines (1.8 times for Gaussian lines) and the peak width broadening
increases. When the modulation amplitude is much larger than the true peak
width, the measured peak width is equal to the modulation amplitude. Since the
maximum calculated modulation amplitude for the Micro-Now cavity is

approximately 2 Gauss, the condition requiring Hm to be much larger than the

122

 

 

2.2001:
A ‘i
0- I
I ..
o. +-
v ‘:
Z .-
O ._
H
l— I
< ..
_] -.
2:) T
D 4..
O +
Z ..
«F
9-900‘::::{::::{::::+:44:j‘%;:ﬁ
0.000 3.000
URRENT (MA)
x10 ‘2

Figure 6-8: A plot of calculated Hm(W) as a function of the power ampliﬁer supply
current.

123

peak width is difﬁcult to fulﬁll. Narrow line spectra are distorted by the presence
of prominent sidebands at intervals of the modulation frequency, 720mG, as
discussed in Chapter 3. Lines that are broad enough not to show the effects of
modulation sidebands are too broad to fulﬁll the condition stated above.
However, attempts to over modulate diphenylpicrylhyrazyl (DPPH) with an
experimentally determined line width of 1.5 G, produced peak broadening that y.—
can be used to estimate the modulation amplitude from the equations developed
by Wahlquist [SW61] which evaluated the dependence of line broadening on

modulation amplitude for Lorentzian line shapes. The estimated Hm(p-p) is 3.0 G,

 

slightly more than the 2.1 G modulation amplitude calculated with Equation 6-1.

The Modulator/Demodulator

The problems associated with the Micro-Now modulator/demodulator are
discussed in Chapter 2. This section of Chapter 6 will discuss modiﬁcations
made to the system with only brief descriptions of these problems.

Originally, the Micro-Now EPR output was connected to about 20 feet of
50—ohm coaxial cable and was terminated at the end by the input impedance of
the PDP 8/E’s ADC, which is 1 MD. This is not a problem with slowly varying
signals observed when recording a standard EPR spectrum; however, when
trying to observe a transient signal, it is important to match all impedances to
prevent reﬂections at the interfaces which could result in erroneous results. A
feed-through 50 ohm terminator was inserted at the connection between the

coaxial cable and the ADC input to properly terminate the signal at the computer.

124

The result was that the signal acquired by the computer was reduced to
practically nothing. A quick look at the circuit diagram of the source revealed a
source impedance of 10 k0 in the EPR output ampliﬁer. The ampliﬁer circuit
show in Figure 6-9 was inserted at the EPR output to correct this problem.

The FET input operational ampliﬁer, (National Semiconductor LH0032C)
which functions as the input and ampliﬁcation stage, is an ultra-fast operational
ampliﬁer conﬁgured for a gain of 23, the largest gain possible while still
maintaining the required bandwidth. The National Semiconductor LH0063C is a
high speed F ET input voltage follower capable of continuously driving a 50-ohm
load with excellent phase linearity up to 20 MHz. The 3 dB roll-off point of the
combined ampliﬁers was 3.5 MHz. Decreasing the gain of the ampliﬁer stage
could increase the bandwidth; however, this response is sufﬁcient for this
system.

The transient response of the EPR spectrometer was suspect because the
results of an experimentally determined reaction rate could not be explained by
the chemistry involved. The response of the EPR was determined using the
experimental setup in Figure 6-10. The ERP spectrometer was tuned and the
magnetic ﬁeld adjusted to the maximum of the resonance line of a standard pitch
sample. The output of the 2 MHz power ampliﬁer was connected to the local
oscillator (L0) input of a Mini-Circuits Laboratory, ZAD-1 double-balanced mixer.
A debounced switch applied a high or zero level signal at the IF port of the mixer.
The switch also generated a time signal that could be connected to the Schmitt
trigger of the computer or the trigger of a storage oscilloscope. The signal

appearing at the RF port is a mixture of the L0 and IF ports. When the IF signal

125

.002 £393 a magic mo 03093 05380 3930 yam 0.: mo Ewan. “30:0 05. 5-0 03E

 

 

 

m3m~220
m\m moo.—
On I UNmQQI... QNNOQIA
1 mG.—.owhmo mmCIm
- — NIIN tom“—
F 00¢ — r

 

 

 

¥NN v:

126

00680020 080880090 Mam 0% .«o 08% 00.80000 00.: 05 038.000 3 00»: «39:0 83 08%-033000 0E. 570 0.5mE

      

 

 

 

 

 

 

 

Wm n 5:32;
NI: a game
s-o¢N
lame.
X10 0..
398
omens
_ ego
_ x"
zoceccnmmu , n+
>e_>¢o c» on x”
r 6
20

mmoecme 0+

hhuzzuw o...

127

is zero, a zero signal appears at the RF output. The 2 MHz local oscillator signal
is zero, a zero signal appears at the RF output. The 2 MHz local oscillator signal
had to be kept below 1 volt (p—p) for the double—balanced mixer to work properly.
The fall time of the signal at the RF port of the mixer after the 2 MHz modulation
was switched off was 0.1 ps. The measured time constants were greater than
10000 times longer than indicated on the front panel of the instrument (0.12 s for
the 10 us ﬁlter position). When the capacitors in the EPR output ﬁlter were
changed to select the time constants labeled on the front panel, the noise level
totally swamped the demodulator electronics.

The manufacturer provided a copy of the test data from the MSU EPR
take before delivery. The test data reported noise levels and signal gain at two
ﬁlter values: NONE and 0.01 ms. No time response data were provided.
Included with the data was a list of the eight customers who purchased the
Micro-Now 2 MHz EPR. Inquiries were made to see what other customers were
doing with their spectrometers. The spectrometer sold to Argonne National
Laboratory eight months after the MSU model was not being used [ARGO]. F. P.
Sargent [SA77] changed the cavity, extensively modiﬁed the electronics, used a
higher gain preamp, changed the modulation ampliﬁer, and has given up using it
because he could not get funding for more modiﬁcations. J. K. SA. Wan [WA77]
modiﬁed the system extensively but is now able to use it only for very large
amplitude signals.

The circuits in the Micro-Now demodulation electronics were created on
Vector board with wires soldered to the components. The wires were two and

three deep in some sections of the circuit boards which is poor construction

128

practice for electronics operating in high frequency regions because of cross talk
between wires. Decoupling capacitors, 0.1 pF, were inserted in the power-supply
lines in an attempt to remove the 2MHz signal picked up by the power supplies.
This failed because the power-supply common already had 2 MHz on it.

Attempts to clean up the common signal by connecting it directly to the chassis of
the instrument were only partially successful.

Rather than attempt to rectify problems in the existing electronics, the
decision was made to replace the Micro-Now modules with commercial and
home-built equipment and compare SIN ratios to those obtained with the system
conﬁgured using Micro-Now components. The noise level was determined by
connecting a 50-ohm terminator to the input in place of the crystal current and
measuring the noise at the output with an oscilloscope. The 2 MHz modulation
output from the oscillator circuit was connected through a 0-81 dB attenuator to
the crystal current input of the electronics for gain and SIN measurements.

First, a Low Noise Ampliﬁer, LNA [A109], was compared to the Micro-Now
preampliﬁer with the result that the signal-to-noise increased using the LNA. This
was expected because the Micro-Now preamp has a frequency range of DC to
greater than 7 MHz while the LNA range is limited at the low end to 70 kHz. A
look at a typical noise power spectrum shows that major components of noise
are 1/f and discrete nose and they are largest at low frequencies. The LNA is a
high-pass ﬁlter and removes much of the low frequency noise. This however, did
create a problem with the 10kHz AFC signal and made it necessary to parallel
the two preampliﬁers at the crystal current input. The Micro-Now preamp was

used for the AFC signal while the LNA was used for the EPR signal. The gain of

129

the LNAS is 37 dB, slightly greater than the Micro-Now preamp. The next
module inline was a video amp. A video amp was originally used in the circuit to
allow external blanking of the EPR signal via a PIN diode switch. The video amp
was not really needed because the blanking option was not included in the
purchase; therefore, it was replaced with a 2 MHz narrow band ﬁlter and a 55 dB
gain tuned amp, both designed locally [A105]. Placing tuned ﬁlters and
ampliﬁers in the circuit limits the rise time of the electronics (rise time equals
0.35Ibandwidth) but in this case, the imposed limit is a worthwhile sacriﬁce.

The integrated circuit demodulator, Motorola MC1596L, and associated
electronics were replaced by a Mini-Circuits Laboratory, ZAD-1 double-balanced
mixer. This phase detector requires the input signal amplitude to be less than 1
V”) to prevent clipping, which imposes a deﬁnite limit on the amount of
ampliﬁcation that can be achieved while the signal is encoded in the 2 MHz
carrier. Ampliﬁcation at the carrier frequency allows discrimination of noise at
other frequencies at the time it is phase detected since only signals near and in
phase with the reference signal will pass the phase detector. To obtain the same
size signal with the new demodulation as compared to the integrated circuit
demodulator, DC ampliﬁcation had to be added to the ﬁnal output stage (see
Figure 6-9). The local oscillator input to the double-balanced mixer is supplied by
the Micro-Now 2 MHz reference and phase shifter circuits with minor
modiﬁcations to improve the rise-and-fall times of the integrated circuit phase-
lock loop in the output of phase-shifter circuit.

The results of ﬁnal tests of the electronics revealed that an input signal

level of 3.4 pV(p.p) produced a 1 V output with a SIN of 1 using a true 10 ps ﬁlter.

130

The SIN of the 3 and 100 ps ﬁlters are 0.8 and 2.5 respectively at the same input
signal level. The ﬁnal time constant test revealed a minimum time constant of 3
us. This severely limits the usefulness of the EPR for transient radical studies.
These results are based only on electronic noise and do not take into account

any noise generated in the klystron, diode detector, or microwave cavity.

The Microwave Cavity

The Micro-Now cavity has several drawbacks that were discussed in
Chapter 2. The most serious problem is that the magnetic resonance peak drifts
slowly during low temperature experiments using a quartz dewar. The problem is
believed to be caused by the cavity’s internal modulation loop. If an internal
modulation loop is used in a microwave cavity, extreme care must be taken to
avoid interactions between the metal wire and the electric ﬁeld component of the
microwave. Such interactions would cause serious loss of microwave energy
and according to Equation 2-1, a drop in the cavity Q. A decrease in cavity Q not
only causes a decrease in signal power (Equation 2-3), but also decreases the
effectiveness of the AFC circuitry.

To minimize the effects on cavity Q, internal wires must be oriented
perpendicular to the E ﬁeld or must pass through a region where the E ﬁeld is
zero [ALSBg]. A diagram of Micro-Now's modulation coil is show in Figure 6-7.
The coil is a single piece of wire passing perpendicularly through the E ﬁeld to
the bottom of the cavity where it is bent out into the E ﬁeld to go around the

sample and a quartz dewar and then passes back up through the E ﬁeld. The

131

 

position of this wire does not have a detrimental effect on cavity Q, which was
measured by Professor J. Cowen [COWN] as approximately 4000. However,
when a quart dewar is placed in the Micro-Now cavity, the bottom of the
modulation coil is pushed into the cavity because the bend at the bottom of the
coil is not large enough to go around the dewar. The modulation loop not only is
not longer'perpendicular to the E ﬁeld, but it is also no longer perpendicular to
the sample.

When working with resonance lines that are 200 to 300 m6 wide, such as
the solvated electron peak, it is important that the resonance peak does not drift
during the entire period of the experiment. To determine if the Micro-Now cavity
was the cause of the signal drifting, another 2 MHz cavity was constructed.
Rather than start from scratch, permission was obtained to modify a Varian
Model E-236 Bimodal cavity (Q of 7000) as long as nothing irreversible was done
to it.

Smith et al. [SM77] had success reconnecting the modulation coils of a
Varian E4531 cavity in parallel from the original serial conﬁguration and series
resonating the coils at 1 MHz. This approach was not taken because of the need
to cut and solder wires connecting the modulation coils. The Varian coils could
not be resonated at 2 MHz in the standard conﬁguration. The inductance of the
Varian modulation coils, measured with an inductance bridge (Boonton Radio
Corp. Model 260-A), is 3.7 mH. It would require a series capacitance of 1.7 pF to
resonate such a coil at 2 MHz. Stray capacitance in the cables is much greater

than this value.

132

Since the modulation coils could not be used at 2 MHz and the side walls
of the Varian cavity were too thick to allow 2 MHz electromagnetic radiation to
pass through, they were removed. The side walls were replaced with pieces of
G-10 circuit board with a 0.2 mil copper foil epoxied to the inside surface. The
skin depth of 2 MHz is copper is 1.8 mils; therefore, 0.2 mil copper foil should not
attenuate the 2 MHz modulation to a great extent. The copper foil is, however,
greater than the required 5 times the skin depth of the microwave frequency and
thus prevents the loss of the microwave through the walls. Copper is an
excellent conductor; therefore, the resistive losses in the cavity walls will not
increase and the cavity Q should not be affected by changing the walls. To verify
this, spectra of a standard sample were obtained with the Varian cavity and 100
kHz modulation before and after the cavity walls were changed. No noticeable
degradation of the signal amplitude occurred. Helmholz coils were wrapped in
Plexiglas supports and tuned to 2 MHz. The magnetic ﬁeld, B, in the center of a
set of Helmholz coils is given by [Wl55]

#0” r2
= 024913)?”2

 

l3 (6-2)

where lie is the permeability of free space (1.26 x 10‘6 Weber/amp—m), l is the
current in amps, r is the radius of the coil, N is the number of turns, and x is the
distance to the center of the Helmholz coil pair. Unfortunately, this coil
arrangement did not produce a spectrum of a standard sample, pitch.
Professor K. M. Chen [CHEN], a microwave specialist in the MSU
Electrical Engineering department, suggested that the skin depth of the cavity

walls is not the only consideration when you have a ﬁnite wall size. The flux lines

133

passing through the coil must return outside of the cavity and since the cavity is
heavy metal everywhere else, the metal outside of the cavity probably attenuated
the modulation. Even though Helmholz coils provide the most homogeneous
magnetic ﬁeld of any coil arrangement, they appear to be of little use at 2 MHz
when a metal cavity is used. Possibly constructing a non-metallic cavity for use
with Helmholz coils would solve this problem. Although there was interest in
building a non-metallic cavity from a machinable ceramic [P067d, LA59, C064],
time did not permit such a task. The only remaining coil conﬁguration was an
internal modulation loop. The best approach to implementing internal modulation
coils is to drill a couple of holes in the top and bottom of the cavity and pass a
heavy-gauge wire straight through the cavity without bends. The loop would be
continued by connecting the two wires outside of the cavity, avoiding interactions
with the microwave ﬁeld. The wires are supported at both ends of the cavity,
which helps to eliminate microphonics that would be present with a coil
supported only at one end, such as the loop in the Micro-Now cavity.

Because EPR cavities are extremely expensive, drilling holes in an excellent
Varian cavity was forbidden. The only opening already in the cavity that could be
used to ring wires into this cavity is the cylindrical opening at the top of the cavity
provided for the introduction of a sample and dewar. Since a dewar was required
for most of the experiments that were planned, a quartz dewar was used as the
support for the coil. Since there is not much room between the dewar and the
cavity, a small gauge magnet wire had to be used for the modulation loop. The
wire was glued down the front and back of the dewar, and the entire assembly

was inserted into the Varian cavity. There are, unfortunately, two problems with

134

 

this arrangement. A large modulation amplitude requires a lot of current in the
wire, but 28-gauge wire has a signiﬁcant amount of resistance. Also, the wire is
in physical contact with the metal cavity for the entire length of the neck of the
cavity (1.25 inches), which results in capacitive coupling of the 2 MHz signal to
the cavity and attenuation of the modulation signal. Because of these two
problems, the Q of the tuned LC circuit containing the modulation loop was
measured as 2. The maximum calculated modulation at the sample was 0.5
G(p—p). Even though the modulation amplitude in the modiﬁed Varian cavity is
signiﬁcantly less than the modulation in the Micro-Now cavity, the problem of the
drifting resonance peak was removed. The modiﬁed Varian cavity was used for

all of the low temperature experiments.

135

APPENDIX

136

APPENDIX A

Program Listings

The computer programs for instrument control and data acquisition are
just as important as the hardware that the computer controls. Thus Appendix A
contains commented listings of the main programs that were used to control the
instrument and acquire and display the data.

The programs are written in FORTRAN II with in-line inclusion of machine
assembly language (SABR) statements. The program statements with an “S” in

column one are SABR instructions and are ignored by the FORTRAN compiler.

137

mmmwmmmmmmmmmmmmmmmmnonnnonooooooaon

OPDEI'.SB
FILE NAME: OPDEF.SB

OPERATION DEF IN ITION IDDULE

THIS MDULE CONTAINS DEFINITIONS OF THE NONSTANDAHD
OPERATIONS THAT ARE USED IN ALL OF THE 08/8
PHOGNAIS. AN OPDEI" DEFINE THE MNEI‘DNIC SYMBOL
THAT FOLLOWS AS AN OPERATE COMMAND AND ASSIGNS

THE SYMBOL THE I/O OCTAL CODE THAT FOLLOWS THE
MNENOMIC. A SKPDEF DEFINES THE MNEPNJNIC THAT
FOLOVS AS A IKIP-ON-FLAG INSTRUCTION AND ASSICNS
THE MNEMNIC TO THE OCTAL I/O CODE THAT FOLLOWS.

THIS FILE MUST BE INCLUDED WITH ALL THE EPR
PNOCRAPB.

OPDEF ADCL 6530 /CLEAH ALL - A/D

OPDEI" ADLM 6531 /LOAD MULTIPLEXER - A/D

OPDEF ADST 6532 /CLEAR A/D DONE AND TIMING FLAGS
OPDEI' ADHB 6533 /READ A/D BUFFER

SKPDF ADSK 6534 /SKIP ON A/D DONE

SKPDP ADSE 6535 /SKIP ON A/D TIMING ERNIE

OPDEI" ADLE 6536 /LOAD A/D ENABLE REGISTER, CILAH AC
OPDEP DILX 6053 /LOAD X CHANNEL D/A

OPDEI" FLASH 7000

OPDEF SWAB 7431 /SWI'ICII FROM MODE A '10 HIDE B—EAE
OPDEF MOL 7421 /AC 10 m. 0 11) AC

OPDEI" CLZE 6130 /CLEAH CLOCK ENABLE REGISTER PER AC
SKPDF CLSK 6131 /SKP 0N CLK OVENFLOW

OPDEF CLOE 6132 /SET CLOCK ENABLE NECISTﬂl PEN. AC
OPDEF' CLAB 6133 /AC TO CLOCK BUFFER

OPDEF CLSA 6135 /CLEAR CLOCK OVERFLOV

OPDEF CLBA 6136 /CLK BUFFER TO AC

OPDEF CLCA 6137 /CLEAR AC, CLOCK COUNTER TO AC
OPDEF IDCA 3400 / INDIRECT DCA TO FOOL SABR

OPDEF ITAD 1400 /INDIBECT TAD TO FOOL SABR.

138

000000000000000

-0
8

101

2000—0000!” mama—nonu—
° as

0000 $000 03000
O O O

0000

FILE NAME: EPRIDNJ'T

FILE NAME: EPRIDNJ'T
ESR DATA AMUISITION MNITOR

ALL PROGRAIB ARE WRITTEN AS SUBROUTINIS GORESIDENT IN CORE

AND CALLED BY MNITUR AFTER SELECTING THE OPTION YOU WISH

TO USE. IF TTIE PROGRAM IS TO LARGE TO BE COREIDENT LE.
DATmN.PT. IT IS LOADED BY A FORTRAN CALL TO CHAIN. THE
PROGRAM MUST BE IN A SAVED FILE ON THE SYSTEM DISK ($.SV).

THE VARIABLE IN COMMN IS NOT DISTURBED IN THE CHAIN OPERATION
SINCE ALL PROGRAIS HAVE THE SAME COMIIIN STATEMENT.

PROGRAMMER: TIINJTHY G. KELLY OCT. 16. I975
IDDII’IED JUNE 13.1976

COMINJN IADRES.NUMINC. IDATPT. NSCAN. IDELAY.NPT‘S. ITIIE. NAVE. BASE.
ISIGMA. IKINSP.YDATA.TIMPPT. NBASE. ISTART. INCREM
DIMENSION YDATA( 1024)

FORMATt 1X.'I = EXIT. 2 I ACQUIRE. 3 3 DISPLAY. 9 3 STORE.
15 3 FETCH'I' 6 = SETUP. 7 = DA'I‘IIIN’)

PORMAT(' 3 ’.I5)

FORMAT(1X. ‘ 1=ESR SCAN. 23TIMED DATA AQUISITION')
FORMAT('PROMPT? ’.A1)

ENTRY POINT

READ( I. 103) IY

TAD \IY

TAD CIAY

SZA CLA

JMP \2

WRITE(1. 100)

READ( I. 1011 IA

00 TO ( 10.20.30.40.50.60.90) IA
EXIT TO (BID MNITOR

CALL EXIT

CALL DATA ACQUISITON PROGRAM
WRITEI 1.102)

READ( I. 101) IKINSP

GO TO (70.80) IKINSP

DISPLAY DATA ON STORAGE SCOPE

CALL SCPLOT
com 1

OUT'PUT DATA

CALL STORE(2)
GO TO 1

READ DATA FILE INTO CORE

CALL PETCIRI)
0010 1

SETUP.” IS A RAPID SCANNING. FAST DISPLAYING PMGRAM TO
HELP OPTIMIZE INSTRUMENTAL PARAMETEIB

139

am. PT. CONT’D.

O
O

0000 000N000
8 0

CALL CHAIN( ’SETUP’)
GO TO I

EPIBCNJ'T IS USED TO OBTAIN EPR SPECTRA
CALL CHAIN( ’EPRSCN’)
TI-A.I"T IS A TI- DATA ACQUISITION PROGRAM

CALL CRAIN( ’TI.A’)
GO TO I

DATIDNJ'T IS ANOTHER mNITOR LIKE EPRMN." WED FOR DATA
MANIPULATION

CALL CHAIN( ’ DATHIN' I

4640 IASCII CODE FOR Y IN A1 FORMAT + CIA
END

140

FILE NAME: EPRSCN."

FILE NAME: arm."
IR ACQUISITION PMGRAM

THIS PROGRAM IS D-IGNED TO RECORD AN ER SPECTRUM.
YOU DEFINE THE RESOLUTION OF THE MAGNETIC FIELD (UP
TO 1024 POINTS/SPECTRUM) AND THE NUMBER OF
INTEGRATIONS AT EACH POINT. THE MAGNETIC FIELD

IS INCREMENTED USING THE X-AXIS OUTPUT ON THE FMNT
PANEL OI" THE PDPB/E. AFTER THE FIELD HAS BEEN
INCREMENTED. SUBROUTINE ERRSUB IS CALLED TO

DETERMINE WHEN THE MAGNETIC FIELD HAS REACHED

ITS CORRECT FIELD STRENGHT. N POINTS ARE TAKEN

AND AVERAGED. THE DATA ARE STORED AS DOUBLE PRECISSION
INTEGERS IN THE LAST TWO THIRDS OF YDATA( I) IN COMIDN.
THE NUMBERS ARE FLOATED BEFORE RETURNING TO EPRINJN.

TIHTIHY G. KELLY OCT. 16. 1975
MDIFI- MAY 30. I976

000000000000000000000

COMI'DN IADRES.NUMINC. IDATPT. NSCAN. IDELAY.NPTS. ITIm.NAVE.BASE.
ISIGMA. IKINSP. YDATA.TIMPPT. NBASE. ISTART. INCREM
DIMENSION YDATA( 1024)

 

C
C
C DEFINE ABSOLUTE IDCATIONS
C
S ABSYM XPONTR 130 /PAGE 0 ADDRESS
8 ABSYM POINTR. I31 /PAGE 0 ADDRESS
8 ABSYM STORE I7 /AUTO INDEXING REGISTER
S ABSYM ISWRD 16 IAUTO INDEXING REGISTER
C
C START
C
200 FORMATT ’NUMBER OF POINTS/SPECTUM (MAX 8 1024) ’. I5)
8 TAD (7000 /INITIALIZE MAGNET-5WLTS
DILX
S CLA CLL
S TAD DATFLD /SETUP RETURN TO CORRECT DATA FIELD.
8 AND {7707
8 RD!" /OR DATA FIELD INTO BITS 6-8 OF AC
S DCA DAT‘FLD
10 READ( 1 .200)NPT‘S
C
C DETERMINE THE NW TO BE 08. TO INCREMENT THE MAGNETIC FIELD.
C
S TAD \NPTS /SIZE OF INCREMENT 8 102va
S DCA DIVISR
S CAM /0 TO AC. 0 TO III
S TAD (2000 /1024 - 10 BIT CONVERTER
8 ML IAC TO m, 0 TO AC
8 DVI
SDIVISR.0
S /TmT FOR DIVIDE ERROR
S JI‘B ERR2
S SNA ICHECK FOR REMAINDER
S JMP NOREM INO REMAINDER
C
C TO COVER THE ENTIRE MAGNETIC FIELD RANGE SET ON THE FIELDIAL.
C ITISNECUSARYTOCHOWEANPTSTHATISAPONEROFZ.
C
WRITEt 1 . 100)
100 FORMATC 'SELEC'T A NEW ”TE-POW“ OF 2')

141

EPIBCNJ'T. CONT“ D

GO TO 10
SRO-I. I‘HA CLA
S DCA INCR

READ( 1.300) NAVE
FORMAT( ’NUMBER OF AVERAG-/POINT ’ . 15)

TAD \NPT‘S

C IA

DCA \ I DATPT
TAD \NAVE

DCA IAVE
TAD (2216
DCA \IADRES
TAD \IADRES
DCA STORE
TAD \IDATPT
DCA NUMPT‘S
TAD XCOUNT
JMP FIELD
INCFLD.TAD MAGFLD
INCR
FIELD. DILX

DCA MAGFLD
DCA HIGH

DCA LOW

ILOAD m INTO AC

/SET UP POINTS COUNTER

/IDATPT IS -NPTS
/SET UP AVERAGE COUNTER

/IAVE IS -NAVE
/ADDRFSS OF LAST 2/3 OF YDAT( I)

/GET ADDRESS POINTER FROM COMMN
/GET -NPTS FROM COMIDN

M” X CHANNEL DAC TO CONTML
/MAGNTIC FIELD

/INITIALIZE LOCATIONS HIGH AND LOW

'CAuERIISUB-MAGNETNULLDETECTIONPROGRAM

CALL ERRSUB
CLA CLL
ADCL
TAD (3
ADLM
TAD IAVE
DCA NUMAVE
SGE'I'H)R.ADST
SABC. ADSK
JMP Am
ADRB
ADSE
SKP
JPB ERR
SMA
JMP NOTHEG
TAD LOW
DCA LON
SZL
JMP TOOMP
CLA CMA
TAD HIGH
DCA HIGH
JMP TCOPH’
OTNEG.TAD LOW
DCA LOW
SZL
ISZ HIGH
, 7000
CLL
ISZ NUMAVE
JMP GETIDR
6211
TAD LOW
IDCA STORE
TAD HIGH
IDCA STORE
ATFLD,6201
ISZ NUMP'TS

mmmmmm anomommmmwmmmmmmmmmmmmmum
E

mmmmmmmm mmmgmmmmmwmmmmmmmmm

"8

/BE SURE AC IS ZEN)
/CLEAR ADC FLAGSAND norm

/LOAD MULTIPLEXER 3. CLEAR AC
/GET -NAVE FMM COMIDN

/SKIP ON A/D DONE

/READ A/D BUFFER .CLEAR DONE FLAG
/SKP ON ADC TIMING ERROR

ICHECK FOR NEGITIVE NUMBER

INO

/CHECK FOR OVERFLOU

/SIGN EXTEND THIRNJGH UPPER WRD

/NUMBER IS MITIVE
ICHECK FOR OVERFLOW
/PLACED HERE TO NULLIFY ISZ

/CHANGE TO DATA FIELD 1

/AUTO INDEX REGISTER

/CHANGE TO DATA FIELD 0
/INCREMENT THE MAG FIELD?

142

EPRSCN. FT. CONT' D

S JMP INCFLD /YEB
CALL FLOTDB
TIMPPT=0.
CALL CHAIN( ’EPRIDN’)
SCONVRT. 1000

SNUMAVE.0
SNUMPTS.0
SIAVE. 0
SXCOUNT. 7000 /-5 VOLTS FOR 10 BIT DAC
S INCR. 0
SMAGFLD.0 /INCREMENTED NUMBER FOR MAG FIELD
SADCEN. 0000
8 HIGH. 0
8 L011. 0
S ERR. 0
VRITE¢ 1 .500)
500 ' FORMAT( 1X.'A/D TIMING ERROR‘)
CALL EXIT
S ERR2. 0
WRITE( I .600)
600 FORMAT( 'DIVIDE ERROR’)
CALL EXIT
END

143

000000000000000000000000000

mmmmmmmmmgmmnoommmnmnnn

F ILE NAIE: DATA. F'I'

SUBROUTINE DATA
FILE NAME: DATA."
ESR ACQUISITION PROGRAM-KINETIC DATA

THIS SUBMUTINE NE WRITTEN TO PERFORM DATA ACQUISITION

UNDER THE CONTmL OF THE DKB-EP PROGRAMMABLE REAL-TIME CLOCK.
WHEN THIS ROUTINE IS CALLED. IT TAKES NUMPTS (VARIABLE NAIE)
NUMBER OF POINTS AT A TIME INTERVAL SET BY THE CLOCK.

THE MINIMUM TIME BETWEEN POINTS IS 25 USFB USING THE

ADO-EA A/D CONVERTER.C THE DATA POINTS ARE DEPIBITED

IN THE ARRAY YDATA( 1024) IN COMMN.

ALL INDEXES AND PARAMETERS MUST BE SETUP BY THE CALLING PMGRAM.

ALL ARGUMENT‘S ARE PASSED THROUGH COMMN
SPECIAL LOADING INSTRUCTIONS!

THIS SUBROUTINE MUST BE LOADED INTO FIELD ZERO.
THIS IS REQUIRED BECAUSE THERE MAY NOT BE TIME TO
READTHEDATAFIELDANDSETUPARETURNTOTHAT
DATA FIELD BEFORE THE A/D CONVERTER HAS THE
FIRST DATA POINT: THEREFORE. THE ROUTINE ASSUME
THAT IT IS IN FIELD 0.

PROGRAMMER: TIIMYIHY G. KELLY AUG. 8. 1976

COMEDN IADRES.NUMINC. IDATT’T. NSCAN. IDELAY.NPTS. IT'IE. RAVE. BASE.
ISIGMA. IKINSP.YDATA.TIMPPT.NBASE. ISTART. INCREM

DIMENSION YDATA( 1024)

ABSOLUTE SYML DEFINITION

ABSYM NUMPTS 17‘! /PAGE 0 ADDRESS
ABSYM XDOUNT 175 /PAGE 0 ADDRESS
ABSYM XPON'IH I30 /PAGE 0 ADDRESS
ABSYM POINTR I7 /AUTO INDEX REGISTER
ABSYM YPOINT I6 /AUTO INDEX REGISTER

ENTRY POINT

6211 /CHANGE TO DATA FIELD 1
CLA CLL
MILADSK /SKIP ON A/D DONE

JMP GET'H)R
ADRB IREAD A/D BUFFER. CLEAR DONE FLAG
ADSE ISKIP ON A/D ERMR FLAG
SKP
JIB ERR
IDCA POINTR /DEPOSIT DATA IN FIELD 1
I82 NUMPT‘S /HAVE I TAKEN ALL THE DATA POINTS?
JMP GETTDR /NO. GET'MRI
6201 /RETURN TO DATA FIELD 0
RETURN

SERR. 0
URITE( 1 .500)

500 FORMAT( 1X. 'A/D TIMING ERMR')

CALL EXIT
END

144

000000000000

mmmmmmmmmmmmmmommmmwmonomuwmwmoonn

SUBBOUT‘INE ERRSUB

FILE NAHE: ERRSUB.”
PROGRAIIIIER: TIIIYIHY G.

DATE 3 24-JUN-76

FILE RAE! mun.”

KELLY

THIS PROGRAII VAS WRITTEN AS A SUBMUTINE TO DETERHINE HAGNETIC
FIELD NULL. THE ERROR SIGNAL 1. SAIIPLED AT HUTIPLEXER INPUT
I AT INTERVALS SET BY THE ERMR SIGNAL AT SCHIIIT'I' TRIGGER O.

1” POINTS ARE AVERAGED.

NULL IS DIXILARED WHEN 3 CONSECUTIVE

AVERAGES ARE WITHIN +/-I.

COIIIIJN IADRES.NUIIINC. IDATPT.NSCAN. IDELAY.NPTS. ITIIIE.NAVE.FELDRG.
ISCAN'ITI. IKI NSP . YDATA. TIIIPPT. NBASE
DIIIENSION YDATA( 1024) .NUII( 10)

ABSOLUTE SYMOL DEF INITIONS

ABSYM CHKFLG 134
ABSYII NUNPT'S 133
ABSYN XCOUNT 132
ABSYM XPONTR I31
ABSYR POINTR I.
ABSYM TIIIEPT 1 1

IPAGE O ADDRESS-WED FOR DIRECT ADDRESSING
/PAGE 0 ADDRESS

/PAGE 0 ADDRESS

/PAGE 0 ADDRESS

/AUTO INDEX REGISTER

IAUTO INDEX REGISTWI

/CLEAIB CIDCK ENABLE REGISTER

/SET BIT ZERO OF CLOCK ENABLE REGISTER
/1NITIALIZE CLOCK FLAG

/CLEAR ALL A/D FLA“

/I‘1ULTIPLEXER I

/LOAD NULTIPLEXER FROH AGO-II AND CLEAR AC
/SET ADC FOR CLOCK EXTERNAL START
/LOAD ENABLE AND CLEAR AC

/INITIALIZE INDEX I

/INCREIIENT I BY 1

/IS IT WITHIN BUFFER SIZE OF 5?

/YES IT IS
/NO. SET ['1

IDIVSORiNPTS TO BE AVERAGE!)

/NUNAVE

/SET DELAY TIRE OF 1 UIEC

145

ERIBUB.FT. CONT’D.

TAD CLEENB

CLOE

CLA CLL
vm

gunman

. CLSK
JMP OVER2
TAD (3777

CLZE

CLA CLL
ADC. 182 cm

SEP

JIIP CONTIN

ADSK

€30UbgquﬂbmODQODOUDQUDQODGUDQUDQODOUDQ
3

E

"U
:3
s
D
/
L.

gammﬁmmmmwmmmm
S 5%

E
E
2

NO. 182 mm

mmmmm
"I
E
3

/IDAD CIDCK ENABLE WRD INTO AC
ISET AC INTO CLOCK ENABLE REGISTER

ISTOP CLOCK AND CLEAR CLK
/EXT. START ENABLE

/THIS SELF RETARTING OPTION IS NEEDED
/TO OVERCOIIE INADAQUICIES IN DEC
/HARDWARE

/SKIP ON A/D DONE

/READ A/D BUFFER .CLEAR DONE FLAG
/SKIP ON A/D ERROR FLAG

ICONVERT I. BIT 2'8 CORLIIIENT TO BINARY
IDOUBLE PRECIIIION ADD

/INCREPIENT HIGH cam-m BITS

/HAVE I AVERAGED ENOUGH POINTS?

/NO

/CLEAR LINK-RUST BE CLEAR BHORE DIVISION
/AC TO m

ll‘NJST SIGNIFICANT BITS

/DIVIDE

/DIVISOR' IO.

/T'EST FOR DIVIDE ERROR

/AC TO m. in TO AC
/SET AC3-1

/I-18O?
INO

/AVERA(E INST BE WITHIN +/-1

/-1
IAC NAS'I OR 0. NEETS REQUIRMTS
/OUTSIDE OF LIHITS

/SET' AC3!
/DECREIIENT J

/NUST HAVE 3 AVERA. EQUAL

/CONVERT IO BIT 2'S COH’LIIIENT TO BINARY
/OVERFI.0V+PDDE 00+: HHZ+EXI'. START
/AND CLOCK INHIBIT+EVENT1

146

ERISUB. FT. CONT‘D.

SADCENB. 0200 /ADC ENABLE EXTERNAL START
S ERR. 0
WRITER 1 , 500)
500 FORIIAT( 1X. ’A/D TIIIING ERROR’ )
CALL EXIT
SERR2. 0
WRITER 1 .600)
600 FORIIATT 1X. 'DIVIDE ERROR’ )
CALL EXIT
END

147

FILE NAIE: FLOTDB.”

SURMUTINE FLOTDB

§EE

TAD
DCA
TAD
DCA

gas:

9

E

58§§

DCA
TAD
SNA
J RP

B$§§E§E

IAC

mmmmmmmmmwwwmmmmmmmmgmmummmmmmmnnnmmmmoono 00000000000000

8
a.

I. SCA

0
—
b

nun:
F!
B

TAD

PRECISION

BITS 1-8 8
BITS 9-35

CONN IADRES.NUHINC. IDATPT. NSCAN. IDELAY.NPTS. ITII.NAVE. BASE.
ISIGNA. IKINSP. YDATA.TIPIPPT. NBASEISTART. INCREH
DIMENSION YDATA(1024)

( 2 16
XCOUNT
( 22 16
XPONTR
DATFLD

DATFLD

I TAD XPONTH
ITAD XPONTH

TEHPI
TEHPI
(4000
WORDI
WORDI
CLA
P081
TEPIPI

TEHPI

(27
(200

SUBROUTINE FLOTDB."

THIS PMGRAI‘I WAS WRITTEN TO FLOAT DOUBLE PRECISION 1mm.
USING THE EXTENDED ARTHPIETIC ELEIIENT (EAE). THE DOUBLE
INTEGERSARESTORED INTHELAST2/3OFTHE

ARRAY YDATA(1024). THIS ROUTINE REIDVES THE DOUBLE
PRECISION INTEGER. FLOATS IT. AND STORES IT AS 3 WORDS
STARTING AT THE FRONT OF THE ARRAY.

REAL DATA ON THE B/E IS STORED IN 3 Home (36 BITS).

BIT 0 I SIGN BIT (0 IS PwITIVE. 1 IS NEGATIVE)

EXPONENT + 200 OCTAL

' NANTIISA

ABSOLUTE SYMBOL DEF IN ITIONS

132 /PAGE 0 ADDRES
I? /AUTOINDEX REGISTER
16 /AUTOINDEX REGISTER
131 /FAGE 0 ADDRESS

/SET UP LOOP

IADDRESS OF YDATA IN comm-1
/ADDRESS OF 1m IN COM-1
/OR DATA FIELD INTO AC

ISETUP DATA FIELD FOR RETURN
/CHANGE TO DATA FIELD I

IAC TO no.0 TO AC
/NOR!IALIZE-EAE INSTRUCTION

/GET SIGN OF mum

/Cm FOR NEGATIVE NUHBER

/YES. IT IS NEGATIVE

/CREATE 2’s COHPIIIENT OF THE DOUBLE
/PRF£ISSION INTEGER

/CIIECK FOR A CARRY FMII CIA OF LOW WRD
llNCREHENT HIGH WRD

fSTEP COUNTER CONTAINS NUIIBER OF SHIFTS
IEIIPONENT IS 27(OCTAL) - RUBBER OF SHIFTS

148

FLOTDB.". GONT’ D.

AND (377
CLL
RTLzRAL
TAD WORD!
DCA WORD!
'I‘AD 'I'EIIPI
AND (3400
RSV
CLL
RTR
TAD WRDI
6211
IDCA XCOUNT
6201
TAD 'IENPI
AND (377
RTLsﬂTL
DCA HORDI
SWP
DCA TEIIPI
CLL

082. TAD TEHPI
AND (7400
new
R'I'R
TAD WORD!
6211
IDCA XIIOUNT
6201
TAD TEHPI
AND (0377
RTLzllTL
6211

gmammmmmmmmmmm

mammmgmgummmgmmmmmmm

IEXPONENT IS STORED AS EXPONENT‘P200(8)
/IIAKE SURE LINK = O

IVORDI NON RAS SIGN AND EXPONENT
IGET FIRST 3 BITS OF HANTISSA
IHAKE SURE LINK=O

/AU'IOINDEX REG.

/GET RENAINING 8 BITS FRO}! WRDI
/LEFT JUSTIFY BITS

/AG TO m. 1!: TO AG

/GET' 4 BITS TO COIIPLETE WORD 2
/HWE BITS 0-3 INTO MITIONS O-II

/GET REMAINING 8 BITS
ILEFT JUS'I‘IFY

149

0QUOQQ0000000000000000000000000000000000

100
101

130
160
201

500
600
700

900
1000
1100

FILE NAIE: STORE."

SUBMUTINE STORE(H)DE)
FILE NAIIE: STORE."

SUBROUTINE TO OUTPUT A DATA FILE TO STORAGE DEVICE IN
C.G.E.CO. STANDARD FILE FORRAT. FIIST m CRARACTEIE
ARE CONTROL CRARACT‘EIB.

COMMENT

READER (FILE NAIIE. DATE. YOUR NAIIE)
INTEGER DATA ( I5 FORIIAT)

READ DATA (E13.3 FORIIAT)

ASCII DATA (A6 FORIIAT)

END DATA FILE

DATA ARE WRITTEN WITHOUT THE X DATA POINTS IN ALL 11mm FORRATS
IN ORDER TO SAVE SPACE ON THE OUTPUT DEVICE. THE X DATA POINTS
WILL BE ADDED TO THE DATA FILE ON THE PDP “/40 WHEN IT IS READ
IN USING DAT‘IDN.LDA UNDER DOS/BATON.

EEEEE”

AN 'AS’ DATA FILE REQUIRES THE LEAST SPACE ON THE STORAGE DEVICE
AND IS BY FAR THE FASTEST FORII OF OUTPUT'. HOWEVER. THIS FILE
FORRAT HAS THE DISADVANTAGE OF HAVING TO BE TRANSLATED BEFORE
YOU CAN READ THE VALUE OF THE POINTS.

THIS PROGRAII IS WED IN 1110 muss.

IDDES
II OPEN DATA FILE AND RETURN TO CALLING PMGRAII
'2 OPEN DATA FILE AND OUTPUT DATA

PROGRAM: TIWIHY G. KELLY JAN. 9.1976
IDDIFIED JULY 29.1976

DEFINITIONS:

OPDEF RDF 62I0 lREAD DATA FIELD INTO AC BITS 6-3
OPDEF ITAD 1400 /INDIRECT TAD TO FOOL SABR

OPDEF BSW 7002 /BYTE SNAP IN AC

ABSYII DIN)” 0133 IPAGE 0 momma

ABSYR POINTH I0 /AU'IO INDEXING REGISTER

CONMN IADRES.NUIIINC. IDATPT. NSCAN. IDELAY. NPTS. ITIIIE.NAVE. BASE.
ISIGIIA. IKI NSP. YDATA.TIIIPPT. NBASE. ISTART. INCRER
DIIIENS ION YDATA( 1024) . IYNAIE( B)

FORMATI 'ENTER COINIENT TEXT’I'OONTROL G TO END'/)
FORIIA'N ’WRIT'E ENABLE ’A4l)

FORRATIAI)

FORMAT( 'IID' .A6.2I2, II.8A2.7I3.FIN.6.2IU)
FORPIA'I‘t A2)

FORHATT 'ENTER FILE NAME '.A6)

FORIIAT( ‘ENTER YOUR NAllEz ’ .8A2)

FORRATT 'OUTPUT DEVICE: ’M)

FORPIAT( ’ IN' 15)

FORMAT? 'ED')

FORIIAT( ‘RD'EIU.8)

FORIIATT 'AS'AG)

FORMAT(’I 3 IN. 2 = RD, 3 ' AS: 'II)
FORRAT( '8 FIELDIAL SET (KG): 'FIO.3)
FORIIAT( ’3 RANGE (KG): 'F10.3)

ENTRY POINT
READI I .000)DEVICE

150

STONE.FT. CONT" D.

QQQQQMQQ 000

0C50f5 Inuunwanmanmanmlam

IDWGDQUHDO MUHDU MUDWUDQUD

WRITE( I . 101) DEVICE

READ( 1.2011FNAIIE

CALL OOPEN1 DEVICE. FNARE)
READ( 1.300) IYNAIIE

WRITE HEADER ON OUTPUT DEVICE

Ill) '0

IDAY‘O

1YR80

TAD (7666

DCA DMNYR

RDF IREAD DATA FIELD TO SETUP RETURN
TAD DAT'FLD

DCA DATFLD

6211 /CHANGE TO DATA FIELD 1

ITAD DTDNYR /DA'I'E IS IN LOCATION 17666

ATFLD.6201 /RETURN TO DATA FIELD N

DCA AG

TAD AG

AND (0007 /LAST DIGIT OF YEAR 18 IN BITS 9-11
DCA \IYR

TAD AG

RTRIRAR

AND (37 /DAY 18 IN BITS 4-8

DCA \IDAY

TAD AG

RTLsRT'LsRAL

AND (17 mm 18 IN BITS 0-0

DCA \IID

WRITE(4. 130) FNAIE. 1m. IDAY. IYR. IYNARE. IDATPT.NSCAN. IDELAY.NPTS.
1 ITIHE.NAVE. IKINSP.TIRPPT.NUNING.NBASE

WRITE CONTENT ON OUTPUT DEVICE IN C.G.E.GO. STANDARD FILE
FORE-CONTROL CHARACTERS ARE 8 .

READ( 1 ,1000)SET
READ( 1 . 1100) RANGE
WRITE(4.1000)SET
WRITING. I 100)RANGE
WRITE( I. 100)

TAD CRZIU

DCA AG

NWLINE.TAD (7340 /: 'GOIIENT (IN A2 FORHAT)

DCA \ICHAR

WRITE(4. 160) ICHAR.

TAD (7671 /LINE LENGTH OF 72 CHARACTEIU

DCA LINE /SET UP A LINE COUNT
/CORRESPONDING TO WIDTH OF TTY

GETOHR /GET CHARACTER FRO]! T'TY

TAD \ICHAR ICHARAC'I‘ER HIST BE SIX BIT ASCII

/FOR FORTRN FORMAT

c.
i

BSW

AND (7700

TAD (60

DCA \ICHAR

WRITE(4.140)IGRAR.

ISZ LINE

JIIP cannon

TAD CROW /CR 18 WRITTEN A8 A SPACE
DCA \ICHAR

WRITE(4. 140) ICHAR

TAD CR213

JPS PUNCH

TAD LF212 /LINE FEED

JIB PUNCH

TAD AG

SZA CLA an» ONLY AFTER GONTTIDL G
JNP NWLINE

151

STORE. FT. CONT' D.

136

@000

meﬂbmﬂbgfi0t5ﬂﬂbwQDQIIQODQODCHDUDQODO
.33

GO TO(250. 136)MDE
READ( 1.900) IA
GO TO( 190.200.210) IA

SABR SUBROUTINE TO GET CHARACTER FMN TT'Y

GE'ICHR.0

CLA CLL

NSF /TEST TT'Y FMG

JIIP TT'Y2

KRB /READ CHARACTER AND CLEAR FLAG
TLS /ECHO GRARACETR

DCA \ICHAR

TAD \ICHAR

TAD CIA213 ITEST IF CR

SNA CLA

JPIP CRLF /YIB. CHARACTER WAS CR

TAD \ICHAR

TAD CIACG /TFST IF CONTIN”. G

SZA CLA

JIIP I GETOHR

DCA AG /YES. CHARACTER WAS CONTROL G
JPIP CRLF

SABR SUBMUTINE TO TYPE CHARACTER ON TTY

0

T‘SF /SK1P IF TT'Y FLAG =1

JIIP TT'YI

T'LS ILOAD T'I'Y BUFFER AND PUNCH CHARACTER.
CLA CLL

JMP I PUNCH

SGIA215.7563 /CR 0215 PLUS CIA
SCIA212.7566 /LINE FEED 0212 PLUS CIA

SCRZIU.
SLF212.
SLINE.
S AG.
SCIACG.

0215 /CR

0212 /LINE FEED
0

o .
7571 /CONTROL G PLUS CIA

SCW.“§0 /SPACE IN AI FORMAT
8802306.2304 ISD IN A2 FORBAT

C

—
'0
0

‘0
a

00000 nanom—

§§§

OUTPUT DATA IN I5 FORIIAT. FIRST Tm CHARACTEIS ARE
’IN' = INTEGER FILE

CONTINUE

DO 195 I=I.NPTS
IYDATA3 YDATA( I)
WRITE( 4.500) IYDATA
JI‘IP \220

’RD' OUTPUT FILE. Y IS WRITTEN TO THE OUTPUT DEVICE
IN E15.B FORl'lAT.200 CONTINUE

WRITE(4.700) (YDATA( I) . I: 1.NPTS)
JIIP \220

'AS' DATA FILE. Y VALUES WRITTEN AS AN ASCII CHARACTER IN
A6 FORMAT.

CONTINUE

DO 215 I'I.NPTS
\‘RITE14.800)YDATA( I)
WRITE(4.600)

CALL OCLOSE

RETURN

END

,152

000000000000000000000000000000

000

nun-u
u-O

mmmmmmmmmmmmmmigoon §

FILE NAIR: FETCH."

SUBROUTINE FETCH(H)DE)
FILE NAHE: FETCH.”
SUDROUTINE TO READ A DATA FILE mu STORAGE DEVICE mum IN

C.G.E.CO. STANDARD FILE FORIIAT. FIRST THO CHARACTEIB
ARE CONTROL CHARACTERS.

3 3 COHHENT

HD 3 HEADER (FILE NAHE. DATE. YOUR NAHE)
IN 1' INTEGER DATA (15 FORI‘IAT)

RD 3 REAL DATA (EI5.8 FORMAT)

AS '3 ASCII DATA ( A2 FORMAT)

DK ' DATA FOR KINETICS

EF 3 RETURN TO CALLING PROGRAH

ED 3 END DATA FILE

THIS PROGRAH WILL OPERATE IN FOUR mums. THE CALLING PROGRAH
RUST SPECIFY THE IDDE OF OPERATION RY PASSING A VARIABLE TO
SUBMUTINE FETCH.FT.

MOE:

1 ASK FOR FILENEIE AND OPEN FILE
2READDATAUFTIL ITMANHOR-
3 FETCH ONLY SELECTED POINTS PM]! THE DATA FILE
4 FETCH ONLY SELECTED DECAY CURVE

PROGRAIDIER! TIINJTHY G. KELLY JAN. 9.1976
mum- AUG 20,1976

OPDEF RSV 7002 lBYTE SWAP IN AC

COIDNJN IADRES.NUHINC. IDATPT, NSCAN. IDELAY.NPT‘S. ITIIE.NAVE.RASE.
ISIGHA. IKINSP.YDATA.TIHPPT.NBASE. ISTART. INGREDIENT

DIHENSION YDATA( 1024) , ICHAR( 72) . IYNAHE( 8)

ENTRY POINT

J‘O

GO TO( 1.20. I 1.. 1301mm:
ICOH=-I

READ( 1 .200)DEVICE

FORMAT( ’ INPUT DEVICE: 'AQ)
READ( 1.300) FNAIE

FORPIATT ’ENTER FILE NAIIE ’ .A6)
CALL IOPEN( DEVICE,FNAHE)

READ 2 CONTRIL CHARACTEIB FMN T'HE INPUT FILE
READ(4.400) ISVICH.

FORHAT(A2)

TAD \ISVICH

TAD (6776 m FOR Im ’HD’ 8 HEADER

SNA CLA

JHP HEAD /YF8. READ HEADER

TAD \ISWICH

TAD(0“0 MFORLETTEIB'O’=COHENT
SNA CLA

JHP \50 IYES. READ com

TAD \ISWICH

TAD (6662 /TEST FOR LET'IEIB ’IN' 3 INTEGER FILE
SNA CLA

JNP \60 IYFS. READ INTmER DAWN-I5 FORHAT
TAD \ISWICH

TAD (5576 /TEST FOR LETTERS ‘RD' 8 REAL DATA FILE

153

FETCH. FT. CONT’D.

mmmmmummmmmuwwmmmumm

SNA CLA

JHP \70

TAD \Iswlcn
TAD (76!!!!
SEA CLA

JHP \80
TAD \ISWICH
TAD (7365
SNA CLA

JHP \30
TAD \ISWICI!
TAD (7272

SNA CLA

JMP \40

TAD \1 SWICH
TAD (7276

SNA CLA
JHP ENDF IL

/YFB. READ REAL DATUH—EI5.8 FORIIAT
/TEST FOR LETIEIB 'AS‘ I ASCII FILE
/YFB. READ ASCII DATUN-A6 FORMAT
/TEST FOR LETTERS ’DK' 8 KINETIC DATA

m FOR LET'I'EIB ’EF’ ' END
IOF DATA FILE

/RE'I'URN TO CALLING PMGRAR

/TEST FOR LETTERS 'ED' 3 END OF
/DATA FILE

IYES. RETURN TO CALLING

154

0000000000000000

000§§OQOQQO

a 000%!)

mung
_
0

unmummanmunguomaog 8

FILE NAI‘IE: TINEDA. FT

FILE NAIIE: TIIIEDA.FT
ESR ACQUISITION PROGRAH

TIMEDA RECORDS THE VARYING DATA AT ONE
RAGNETIC FIELD VALUE. YOU ENTER THE THE
INTERVAL AND THE NUHBER OF AVERAGES. THE
MAXIHUN DATA RATE. DETERHINED BY THE A/D
CONVERTER. IS 60 KHZ; HOWEVER. T'HIS PROGRAH
HAS BEEN WRITTEN TO USE THE FREE TIDE AVAILABLE.
DATA ARE AVERAGED AS DOUBLE PRECISSION INTEGEIG
AND ARE CONVERTED TO FLOATING POINT FORMAT
BEFORE RETURNING 'IO EPRHON.

PROGRAHIIER: TIIDTHY G. KELLY OCT. 16. I975
COPIHON IADRES. IXADIB. IDATPT. NSCAN. IDELAY.NI’T‘S. ITIE.NAVE.FELDII3.

ISCANTI‘I. IKINSP. YDATA. TIHPPT
DIHENS ION YDATA( 1026)

ABSYII NUI'IPTS I76 /PAGE 0 ADDRESS
ABSYII MOUNT 175 /PAGE 0 ADDRESS
ABSYII XPONTR 130 /PAGE 0 ADDRESS
ABSYM POIN'I‘R 17 lAU'IO INDEX REGISTER
ABSYII YPOINT 16 lAU'I‘O INDEX REGISTER

FORIIAT('TII‘IE BETWEEN POINTS IN SEC (F10.6) '.F10.6)
FORI‘1AT('NUI‘IBER OF POINTS (MAX 3 1024) '.I5)

ENTRY POINT

READ( 1. 1001TIRE

TINPPT‘3TIIE

READ( 1.200)NPTS

READ( 1.300) NAVE

TAD (7000

DILX

CLA CLL

READ( 1.310) INCREII

FORMATUSIZE OF INCREHENT': '15)
FORNA'N 'NUI‘IBER OF AVERAG ' .15)
TAD \INCREH

DCA XPONT‘R

CONST = 0.4095E—2

D0 10 I=I.5

IF(TIIIE-CONST)11. 11. 10

CONST 3 CONS‘I‘JIO.

N 3 7-1

XTINE 3 T1m10.88N

TEST 3 HIRE-2048.
IF(’I'I:‘ST)20.25.25

ITIIIE = IFIXUITIIIE)

GO TO 30

ITIHE = IFIX(TEST)

TAD (4000

TAD \ITIHE

DCA \ITINE

CONTINUE

TAD \N ISELECT CIDCK FREQUENCY
RT'L: RTL: RTL

DCA \N

TAD \ITIIIE

CIA

DCA \IDELAY

TAD CLOCEN IADD CL! ENABLE ”RD

155

TI-A. FT CONT‘ D.

AND (7077
TAD \N

DCA CLOCEN
TAD (216
DCA \IADRIS

DCA \ IDATPT
TAD \NAVE

DCA \NSCAN
TAD \ I ADRFS
DCA PO I NTR
TAD \ IDATPT

TAD \NSCAN

CLA CLL
TAD \IDELAY
CLAB
CLA CLL
ADCL
ADLN
CLA CLL
TAD ADCEN
ADLE
TAD CLOCEN
CLOE
CLA CLL
CLSA
6211
FLASH
CLSK
JIIP OVER
TAD (7000
TAD XPONTR
DILX
CLA CLL
SGETmR.ADSK
JIIP GETTDR
ADRB
ADSE
SKP
J18 ERR
IDCA POINT'R
ISZ NUI‘IPTS
JIIP GETTDR
ISZ NUIBCA
S HI..T
6201
TAD \IADRES
DCA POINT‘R
TAD (2216
DCA YPOINT
TAD \IDATPT
DCA NUHPTS
6211
SCONTIN. ITAD POINTR
SNA

mmwmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmwwm
O
.5

mmmmmmmmnuwmmmmmmw

JHP POS
IDCA YPOINT
CI‘IA

IDCA YPOINT
JI‘IP TST
IDCA YPOINT
IDCA YPOINT
ISZ NUIIPTS

mgmmmmmm

f1

IST'RIP PREVIOIB CLOCK FRMUENCY
IADDCLKFREQ. TOCLKENABLEWRD

xmnms OF IYDATA IN comm -1
/SET UP POIN'IS COUNTER

llET UP AVERAGE COUNTER

/GET ADDRBS POINTER FRON COMEDN
IGET NUHBER OF DATA POINTS FROII MN
IGET NUHBER 0F SCANS FROII COM

ICLEAIB CLOCK ENABLE REGISTER

/GET DELAY FROII CORINTN
/AC TO CLOCK BUFFER

ICLEAR ALL A/D FLAGS
ILOAD HULTIPLEXER FROH ACE-11 AND CLEAR AC

/LOAD A/D ENABLE WORD INTO AC
/LOAD A/D ENABLE REG. AND CLEAR AC

/LOAD CLOCK ENABLE WORD INTO AC
ISET AC INTO CLOCK ENABLE REGISTHI

/CHANGE TO DATA FIELD I
/FLASH LAN?
MT FOR OVERFLOW DUE TO SCHINIT TRIGGER

ISKIP ON AID DONE
/READ A/D BUFFER .CLEAR DONE FLAG
ISKIP ON A/D ERMR FLAG

IDEPOSIT DATA IN FIELD I
/HAVE I TAKEN ALL THE DATA POINTS?
/NO . GETTDRI

/CHANGE TO DATA FIELD 0
IADDRESS OF DATA IN com— 1

IADDRES OF DOUBLE PRMISSION “3RD
IIN COPONJN-I

IPOINTS COUNTER
/CHANGE TO DATA FIELD 1

ITEST FOR NEGATIVE NUMBER

INEG. POINT NWT BE SIGN EIU‘ENDED

/P(BITIVE NUMBER SIGN EXTENDED

156

TINEDA.FT CONT’ D.

S JMP CONTIN
S 6201
CALL FLOTDB
CALL CRAIN( ’EPRmN’)
8 BLT
S ADCENJDZOO /A/D ENABLE WORD-EXTERNAL START
S CONVRT. 1m
SNUI‘BCA.O
SCLOCEN.5661 /mD 01+IHRZ+OVERFLON START A/D-PCLOCK
S /INRIBIT+EVENT1
S ERR. O
WRITE( 1,500)
500 FORHATT IX. ’A/D TIEING ERROR’ )
CALL EXIT
END

157

BIBLIOGRAPHY

158

[AL68a]

[AL68b]

[AL68c]

[AL68d]

[AL68e]

[AL68f]

[Al-689]

[AM72]

[ARGO]

[AT68a]

[AT68b]

[AT68c]

[AT70a]

[AT70b]

BIBLIOGRAPHY

A

R.S. Alger, “Electron Paramagnetic Resonance: Techniques and
Applications”, lnterscienoe, New York, 1968, p 71.

RS. Alger, “Electron Paramagnetic Resonance: Techniques and
Applications”, lnterscienoe, New York, 1968, p 82.

RS. Alger, “Electron Paramagnetic Resonance: Techniques and
Applications”, lnterscienoe, New York, 1968, pp 106-123.

R.S. Alger, “Electron Paramagnetic Resonance: Techniques and
Applications”, lnterscienoe, New York, 1968, p 109.

RS. Alger, “Electron Paramagnetic Resonance: Techniques and
Applications", lnterscienoe, New York, 1968, p 1520.

RS. Alger, “Electron Paramagnetic Resonance: Techniques and

Applications”, lnterscienoe, New York, 1968, Deprieqx, Laboratory
reference 29, p 150.

RS. Alger, “Electron Paramagnetic Resonance: Techniques and
Applications”, lnterscienoe, New York, 1968, p 149.

American Institute of Physics Handbook, Third Ed., McGraw-Hill,
1 972, p 5-1 05.

Argonne National Laboratory, Batavia, Illinois, Trifunca.

P.W. Atkins, K.A. McLauchlan, and AF. Simpson, Chemical
Communications 179 (1968).

P.W. Atkins, K.A. McLauchlan, and AF. Simpson, Nature 219, 927
(1 968).

P.W. Atkins, K.A. McLauchlan, and AF. Simpson, Molecular
Spectroscopy, 1 77 (1 968).

P.W. Atkins, K.A. McLauchlan, and AF. Simpson, Journal Physics
E: Scientiﬁc Instruments 3, 547 (1970).

P.W. Atkins, LC. Buchanan, R.C. Curd, K.A. McLauchlan, and AF.
Simpson, Chemical Communications 513 (1970).

159

[A171]

[A173]

[A1 05]

[A1 09]

[BA64]
[BE67]
[BE69]

[8039]

[BU52]

[CA59]

[CA66]

[CA75]

[CA77]

[0577]

[CHEN]

P.W. Atkins, R.C. Curd, K.A. McLauchlan, and AF. Simpson,
Chemical Physics Letters 8, 55 (1971).

P.W. Atkins, K.A. McLauchlan, and P.W. Percival, Molecular
Physics 25, 281 (1973).

Design number A-105, Instrument Service, Michigan State
University, East Lansing, MI 48824.

Design number A-109, Instrument Service, Michigan State
University, East Lansing, MI 48824.

B
K. Bar-Eli and TR. Tuttle, Journal of Physics 40, 2508 (1964).

T.J. Bennett, R.C. Smith and T.H. Wilmshrst, Chemical
Communications 513 (1967).

T.J. Bennett, R.C. Smith and T.H. Wilmshrst, Journal Physics E:
Scientiﬁc Instruments 2, 393 (1969).

JR. Bolton and J.T. Warden, Publication 39, Photochemistry Unit,
Chemistry Department, University of W. Ontario, London, Ontario,
Canada.

J.H. Burgess and RM Brown, Review of Science Instruments 23,
334 (1952).

C

A. Carrington, R. Dravnieks, and M.C.R. Symons, Journal
Chemistry Society 1959, 947 (1959).

J.G. Calvert and J.N. Pitts Jr., “Photochemistry”, John Wiley and
Sons, New York, 1966, p 783.

R. Catterall, and P. Edwards, Journal Physical Chemistry 79, 3010
(1975)

R. Catterall, J. Slater and M.O.R. Symons, Canadian Journal of
Chemistry 55, 1979 (1975).

J.M. Ceraso, P.B. Smith, J.S. Landers and J.L. Dye, Journal
Physical Chemistry 81, 760 (1977).

K.M. Chen, Department of Electrical Engineering, Michigan State
University, East Lansing, MI 48824, personal communications.

160

[C064] A.R. Cook, L.M. Matarrese and JS Wells, Review of Scientiﬁc
Instruments 35, 114 (1964).

[COWN] J. Cowen, Physics Department, Michigan State University, East
Lansing, MI 48824, measured Q of Micro Now and Varian cavities.

[CU46] R.L. Cummerow and D. Halliday, Physical Review 70, 433 (1946).
[C162] CEM 162 Laboratory Manual, Quantitative Analysis, Experiment #6,
Department of Chemistry, Michigan State University, East Lansing,
MI 48824.
D

[DA79] M.G. DaGue, Ph.D. Thesis, Michigan State University, East
Lansing, MI 48824, 1979, pp 25-37.

[DEC] Digital Equipment Corporation, Maynard, Mass. 01754.

[DE64] R. Dewald and J.L. Dye, Journal of Physical Chemistry, 68, 121
(1964).

[D057] J.L. Down, J. Lewis, B. Moore and G. Wilkinson, Proceeding of
Chemical Society 1 957, 209 (1957).

[D059] J.L. Down, J. Lewis, B. Moore and G. Wilkinson, Journal Chemical
Society 3767 (1959).

[DY59] J.L. Dye and RR. Dewald, Journal Physical Chemistry 68, 135
(1964).

[DY71] J.L. Dye, M.T. Lok, F.J. Tehan, R.B Coolan, N. Papadakis, J.M
Caraso, and MG. Debacker, Ber. Bunsenges Physical Chemistry
75, 659 (1971).

{DY73] J.L. Dye in “Electrons in Fluids,\", J. Jortner and CR. Kestner, Ed.,
Springer-Verlag, W. Berlin, 1973, page 77.

E

[EL63] J. Eloranta and H. Lischitz, Journal of Chemical Physics 38, 2214
(1963).

[EN76] C.G. Enke and TA. Nieman, Analytical Chemistry 48(8), A705
(1 976).

161

[FE57]
[FE63]

[FE64]
[FE73]

[Fl67]

[FR76]

[F R77]

[GA70]

[GA71]

[GEOS]

[GI26]

[G L70]

[GL73]
[GO71]

[G072]

F
G. Feher, Bell System Technical Journal 36, 447 (1957).

R.W. Fessenden and RH. Schuler, Journal of Chemical Physics
39, 2147 (1963).

R.W. Fessenden, Journal of Physical Chemistry 68, 1508 (1964).
R.W. Fessenden, Journal of Chemical Physics 58, 2489 (1973).

E.W. Firth and D.J.E. Ingram, Journal of Scientiﬁc Instmments, 4,
821 (1967).

A. Friedenberg and H. Levanon, Chemical Physics Letters 41, 84
(1976).

A. Friedenberg and H. Levanon, Journal of Physical Chemistry 81,
766 (1977).

G

A. Gaathon and M. Ottolenghi, Israeli Journal of Chemistry 8, 165
(1970)

G. Gabor and K. Bar-Eli, Journal of Physical Chemistry 75, 286
( 1971).

GEOSYS, A Computer System for the Description and Analysis of
Spatial Data, written by R.l. Wittick, Computer Institute for Social
Science Research, 1974.

GE. Gibson and TE. Phipps, Journal of American Chemical
Society 48, 312 (1926).

SH. Glarum and J.H. Marshall, Journal of Chemical Physics 52,
5555 (1970).

SH. Glarum, Review of Scientiﬁc Instruments 44, 752 (1973).

I.B. Goldberg and A.J. Bard, Journal of Physical Chemistry 75,
3281 (1971).

T.E. Gough and F .W. Grossman, Journal of Magnetic Resonance
7, 24 (1972).

162

[(3074]

[6075]

[GO76a]

[GO76b]

[G0766]

[(3077]

[GR73]

[HAHN]

[HA56]

[HA62]

[HA70]

[HA72]

[HEAT]

[HI68]

[H069]

l.B. Goldberg and A.J. Bard, Journal of Physical Chemistry 78, 290
(1 974).

l.B. Goldberg, H.R.Crowe, R.S. Carpenter ll, Journal of Magnetic
Resonance 18, 84 (1975).

LE. Goldberg, H.R.Crowe, G.S. Wilson, R.S. Glass, Journal
Physical Chemistry 80, 988 (1976).

I.B. Goldberg and H.R.Crowe, Journal Physical Chemistry 80, 2407
(1976)

LE. Goldberg and H.R.Crowe, Journal Physical Chemistry 80, 2603
(1976).

l.B. Goldberg, H.R.Crowe and K0. Christe, Journal Physical
Chemistry 17, 3189 (1978).
K. Gruber, J. Forrer, e. Aopﬁ and H.H. Gunthards, Journal Physics
E: Scientiﬁc Instruments 6, 666 (1973).

H
Controller design B.K. Hahn; Interface design T.G. Kelly,
Department of Chemistry, Michigan State University, East Lansing,
MI 48824.

C.G. Hatchard and CA. Parker, Proceedings of the Royal Society
(London), A235, 518 (1956).

D. Halliday and R. Resniek, “Physics: Part II”, John \Niley and
Sons, New York, 1962, p 853.

E.J. Hamilton, Jr., E. Wood, and GS. Hammond, Review of
Scientiﬁc Instruments 41, 452 (1970).

B.J. Hales and JR. Bolton, Journal of the American Chemical
Society 94, 3314 (1972).

Heath Company, Benton Harbor, MI.

R. Hirasawa, T. Mukaibe, H. Hasegawa, Y. Kanada and T.
Maruyama, Review of Scientiﬁc Instruments 39, 95 (1968].

JD. Hoeschele, Ph.D. Thesis, Michigan State University, East
Lansing, MI 48824 1969.

163

[H072]

[H076]

[H077]

[H873]

[HU68]

[HU70]

[HU74]

[HY63]

[KL71]

[KL72]

[LA59]

[LA76]

[LA77]

J.G. Holbrook, T.H. Wilmshurst, and RR. Simmons, Journal
Physics E: Scientiﬁc Instruments 5, 555 )1972).

F.J. Holler, T.G. Kelly, T.V. Atkinson, C.G. Enke, American
Laboratory 8, 9 (1976).

RE. Hoguet, Chief Engineer, Micro-Now Instrument Company,
Chicago, Ill, personal communication.

E.S.P. His, L. Fabes, and JR. Bolton, Review of Scientiﬁc
Instruments 44, 197 (1973).

l. Hurley, T.R. Tuttle, Jr., S.R. Golden, Journal of Chemical Physics
48, 2918 (1968).

D. Huppert and K.H. Bar-Eli, Journal of Physical Chemistry 74,
3285 (1970).

M. Huisjen and JS Hyde, Review of Scientiﬁc Instruments 45,669
(1974)

J.S. Hyde, “Principles of EPR Instrumentation”, 7‘“ Annual NMR-
EPR Varian Workshop, Varian Associates Instrument Division, Palo
Alto, CA 1963.

K

J.G. Kloosterboer, L.J. Giling, R.P.H. Rettschnick, J.D.W.
VanVoorst, Chemical Physics Letters 8, 462 (1971).

J.G. Kloosterboer, P. Jost, O.H. Grtﬁlﬁth, “Computers in Chemical
and Biochemical Research”, Volume 1, CE. Klopfenstein and CL
Wilkins, Editors, Academic Press, New York, 1972, p 175.

L

J. Lambe and R. Ager, Review of Scientiﬁc Instruments 30, 599
(1959).

Jacques Lacoste, Ph.D. Thesis, Universite Des Sciences Et
Techniques Du Langnedoc, France, 1976, pp 23-29.

T. Last and CG. Enke, Analytical Chemistry 49, 19 (1977).

164

[LAST]

[L072]

[MA69]

[MNOW]

[MULP]

[PA53]

[PHOT]

[PI60]

[PIP8]

[P067a]

[P067b]

[P067c]

[PO67d]

[SA64]

Interface design by TA. Last, Department of Chemistry, Michigan
State University, East Lansing, MI 48824.

M.T. Lok, F.J. Tehan, J.L. Dye, Journal of Physical Chemistry 76,
2975 (1972).

M

Matalon, S. Golden, M. Ottolenghi, Journal of Physical Chemistry
73, 3098 (1969).

Micro-Now Instruments Company, Inc., Chicago, Ill, 60646, Sales
literature.

Program MULPLT, written by T.V. Atkinson, Department of
Chemistry, Michigan State University, East Lansing, MI 48824.

P

C.A. Parker, Proceedings of the Royal Society (London) A220, 104
(1953)

Photochemical Research Associates, University of W. Ontario,
London, Ontario, Canada, James Bolton, Pres.

L.H. Piette and WC. Landgraf, Journal Chemical Physics 32, 1107
(1960)

Program PIP8 written by T.V. Atkinson, Department of Chemistry,
Michigan State University, East Lansing, MI 48824.

C. Poole, “Electron Spin Resonance, A Comprehensive Treatise on
Experimental Techniques”, lnterscienoe, New York, 1967, p 259.

C. Poole, “Electron Spin Resonance, A Comprehensive Treatise on
Experimental Techniques”, lnterscienoe, New York, 1967, p 343.

C. Poole, “Electron Spin Resonance, A Comprehensive Treatise on
Experimental Techniques”, lnterscienoe, New York, 1967, p 413.

C. Poole, “Electron Spin Resonance, A Comprehensive Treatise on
Experimental Techniques”, lnterscienoe, New York, 1967, p 321.

S

A. Savitzky and J.J.E. Golay, Analytical Chemistry 36, 1627 (1964).

165

[SA77]

[SE73]

[SE77]

[SM64]
[SM68]

[SM71]

[SM77]

[8068]

[$774]

[TE74]

[V063]

[WA61 ]

[WA75]

[WA76]

F.P.Sargent, Atomic Energy of Canada, Ltd., Whiteshell Nuclear
Research Establishment, Pinawa, Manitoba, Canada, personal
communication.

T.O. Seim, T.B. Mele, and S. Pyrdz, Journal of Magnetic
Resonance 9, 175 (1973).

WA. Seddon, J.W. Fletcher, F .C. Sopchyshyn, and R. Catterall,
Canadian Journal of Chemistry 55, 3356 ( 1977).

G.W. Smith, Journal Applied Physics 35, 1217 (1964).

B. Smaller, J.R. Remko and EC. Avery, Journal of Chemical
Physics 48, 5174 (1968).

B. Smaller, E.C. Avery and JR. Remko, Journal of Chemical
Physics 55, 2414 (1971).

GE. Smith, R.E. Blankenship and MP. Klein, Review of Scientiﬁc
Instruments 48, 282 (1977).

J. Sohma, t. Komatsu and Y. Kanada, Japanese Journal of Applied
Physics 7, 298 (1968).

R.L. Stoll, “The Analysis of Eddy Currents”, Clarendon Press,
Oxford, 1974.

T

F.J. Tehan, B.L. Barnett and J.L. Dye, Journal of American
Chemical Society 96, 7203 (1974).

V

K.D. Vos and J.L. Dye, Journal of American Chemical Society 38,
2033 (1963).

W
H. Wahlquist, Journal of Chemical Physics 35, 1708 (1961).

J.K.S. Wan, H.M. Vyas, S.K. Wong and B.B.Adeleke, Journal of
American Chemical Society 97, 1385 (1975).

J.T. Warden and JR. Bolton, Review of Scientiﬁc Instruments 47,
201(1976)

166

[WA77]

[Wl531

[WE63]
[WE72]

[WE72b]

[WE72c]

[Wl55]

[W072]

[XENO]

[ZA45]

J.K.S. Wan, Queens University, Hamilton, Ontario, Canada,
personal communications.

T.H. Wilmshurst, T.J. Bennett and RC. Smith, Proceedings of the
Royal Society A302, 305 (1968).

W. Weyl, Ann. Physik. 121, 601 (1863).

J.E. Wertz and JR. Bolton, “Electron Spin Resonance-Elementary
Theory and Practical Applications”, McGraw-Hill, New York, 1972, p
24.

J.E. Wertz and JR. Bolton, “Electron Spin Resonance-Elementary
Theory and Practical Applications”, McGraw-Hill, New York, 1972, p
451.

J.E. Wertz and JR. Bolton, “Electron Spin Resonance-Elementary
Theory and Practical Applications”, McGraw-Hill, New York, 1972,
pp 452-454.

R.P. \Mnch, “Electricity and Magnetism”, Prentice-Hall, 1955, p
506.

S.K. Wong and J.K.S. Wan, Journal of American Chemical Society
94, 7197 (1972).

X

Xenon Corporation, Medford, MA, Manufacturer’s literature.
Z

E. Zavoisky, Journal of Physics USSR 9, 211 (1945).

167

   

I“[llj‘1ljlj’l‘lj[ljlﬂl[)1]