:__:_:_::__ mwm

 

 

Illl”WIIlllllHlllllllllllll‘llHlllllilllHlflllllllllllUl

3 1293 01774 9936 71

’ 3 'V‘ In

LIBRARY

= Michigan State
University

 

 

 

 

This is to certify that the

thesis entitled

THE RADIOFREQUENCY BRIDGE METHOD OF

DETECTING NUCLEAR RESONANCE SIGNALS
presented by

Alfred E. Villaire

has been accepted towards fulfillment
of the requirements for

_'\L_S_Q_ degree in mm:—

Major professor

Date AUEUS‘C 13. 1952

0-169

 

.‘ u— 4- v»— -__.,‘.

_ PLACE IN REI'URN BOX to remove this chedtout from your record.
“[0 AVOID FINES mum on or before date due.
MAY BE RECALLED with earlier due date if requested.

m DATE DUE DATE DUE

 

THE RADIOFREQUENCY BRIDGE METHOD or DETECTING

NUCLEAR RESOMCE SIGNALS 7

By‘

ALFRED EDMOND VILLAIBE, JR. / ;.'

A TIESIS I
Submitted to the School of Graduate Studies of Michigan
State College of Agriculture and Applied Science ’
in partial fulfillment of the requirements "
for the degree of

MASTER OF SCIENCE ‘1'

Department of Phyeice !

1952

wwwfv'
Ar-L""j.l\'

/3’/</’52' fi’

Acknowledgement

I would like to express my sincere
thanks to Professor Robert D. Spence
for suggesting this problem.and for
his guidance in the course of its
completion" Hie "pitch-in” attitude

was ever an inspiration.

%J£ép/%o/.

I.

II.

III.

IV.
V.

VI.

VII.

Table of Contents

The Nature of the Investigation
Nuclear Resonance

Methods of Detecting Nuclear Resonance
The Radiofrequency'Bridge
Construction and Operation

Results and Apparatus Photographs
Summary

Bibleography

NI-F'N

1%
20
21

22

I The Nature of the Investigation

Pauli (i) first suggested the existence of a nuclear magnetic
moment in l92h.. It was later pointed out that a nucleus in a magnetic
field would have an energy characteristic of the orientation of its
magnetic moment. Gfittinger (ii) and Majorana (iii) pointed out. in
the early 1930's. that under the preper conditions the magnetic
moment would reorient itself in.the magnetic field. Since that time.
physicists have been devising instruments to detect the energy absorbed
by such a reorientation. It might be pointed out that the energy
change involved in the reorientation of one nuclear magnetic moment
(say that of hydrogen) is of the order of 10-19 erg. The first suc-
cessful resonance absorption experiment was reported in 19N6 by
Purcell. Torrey and.Pound (iv).

Ibis thesis reports an investigation of such instruments under-
taken in this laboratory. The circuits used, while not original or
new; have Just fairly recently been adapted for this type detection.
Particular attention.has been given to a study of the relative per-

formance of two types of bridge circuits.

II Nuclear Resonance

Associated with a nucleus of non-zero spin is a magnetic moment

(1) e

fizyZMcfi’

where '1'; is the angular momentum of the nucleus, e the proton charge,

 

M the proton mass. c is the velocity of light. and g, called the
gyromagnetic ratio, is a number characteristic of a given nuclear
species in a given nuclear energy state.

If a nucleus of magnetic moment I: is placed in a static magnetic
field Ea. the magnetic moment vector (or the angular momentum vec-
tor) will precess about 30 with the frequency
(2) -

- ._ e
w. — ~97 2 Mc H,

regardless of the angle between E and 31-0. This is called the Larmor

 

precession frequency.
The potential energy of a magnetic moment in a magnetic field

i0 is, apart from an additive constant.
(3) ..
__ ‘ H _ __ ______e " - H .
U - /"l ° y ZMc P °
For the proton, with which this paper will be entirely con-
cerned. the magnitude of the projection of '5 along the field dir-
ection is -'i‘h. The Ji is the value of the spin, I. for protons.

This factor gives rise to two (21 + l) Zeeman energy levels. Thus,
(ll)

I it
U: izgmeo-

~2—

The factor eh/eMc a No is called the nuclear magneton. and has the

‘value no 2 5.0h9 x 10‘2” erg/genes (1). Finally. we may write
(1“)
U 2 Ir— 2'? y/u. H. -

The lower (negative) energy characterizes protons with TL (and 3
parallel to the field is and the upper energy level represents pro-
tons ”anti-parallel” to the field.

Assuming it is possible to cause the protons in the magnetic
field to absorb energy. protons in the lower state would be raised to
the upper state. The energy necessary per proton to accomplish this
is
(5) AU 1 Ulupper‘l“ UUOWQ“) = 3/4.. Ho °
For 8,000 gauss one obtains AlJ = 2h 1 10‘20 erg.

The detection of the absorption of this small amount of energy
would be indeed difficult. However, the use of a sample of finite
size insures the presence of millions of protons or possible ab-
sorbers.

If one applies the Bohr frequency condition to this transition,
he obtains
(6) l'tvO-zAUr-j/uoy.

Note that this is simply'the Larmor equation (in scalar form)
where 1).: w./21r.
This means that in order to produce transitions one must apply

energy of Just the Larmor frequency. This act of the nuclei, ab-

.3-

sorbing energy at one frequency only. is called nuclear resonance.
For protons in a field of 8,000 gause,v°=. tall/I1 = 37nd. This

is in the short wave radio range of the electromagnetic spectrum.

III Methods of Detecting Nuclear Resonance

There are several methods of detecting the resonance. One of
these is the molecular beam magnetic resonance method of Rabi (2,3) .
Others involving strictly electronic equipment are the bridge method
of Purcell and the 'bridgeless' (oscillator-receiver) methods of
Roberts (1+) and Williams (5).

At The Super—regenerative Receiver

By means of this instrument resonances are observed by noting the
variation in the output of a super-regenerative oscillator when con—
dition (6) is realized. .A super-regenerative oscillator consists of a
vacuum-tube circuit which would execute normal oscillations at a radio-
frequency were it not for an audiofrequency “quench voltage“ which is
applied to one of the tube elements in such a manner as to prevent
the radiofrequency oscillations from attaining their'normal ampli-
tude. The sample is placed in the tank coil of the oscillator and
acts on this coil at resonance.

The output of the super-regenerative oscillator is fed into an
audio amplifier or an aim. receiver and thence to a cathode ray
oscilloscope. The resonance absorption peaks are displayed directly
on the oscilloscope.

We investigated the super-regenerative method for about a year
and a half. In that time many different circuits were tried; the

oh-

rte

rte

 

  
 
   
 

n, - We 35mh at trees
50 a. , .
% L COII Jm . - ' 2
I ' C

'0

WM IO!
—)

 

 

'esere'ohon ostrol

 

c t trons,

 
 
 
 
 
  

I'Z

 

Y O
MDMICY

 

 

medication
controis

 

 

 

 

 

 

 

 

 

39
SW:
cell 85mh
'0
6J5 5K ,
hp q
audio
oscillator .

 

 

 

t
r 851m.”
BO
2 0 K
on cmccdors Inf/.1

 

 

(C)
POW“ SUN-feceneroiwe oscillators (0)8(b) self-quenched,
(c) externally quenched.

three shown in Figure l worked well. Two are "self-quenched” the third
externally quenched (6). One advantage of the super-regenerative re-
ceiver over the bridge method is that the frequency of operation may be
varied over a rather large range by merely tuning the tank circuit. This
greatly simplifies the search for resonances. Such a search involves
sweeping either the magnetic field or the frequency; in general, a smooth
variation in the magnetic field over a large range is difficult to effect.

The spectrum of a super-regenerative oscillator consists of a cen-
tral frequency voplus a large group of side-bands of frequency 1).:an
where n is an integer and.t%is the quench frequency. The relative in-
tensities of the various spectral components depend upon the conditions
at which the oscillator is operated. Any of the frequencies

(7) 7)==TL%:1V11) (Y\::())i,2v- --)

8
is capable of producing a resonance absorption provided condition (6) is
fulfilled. This fact can be used to advantage in search for resonances.
as it is usually easier to recognize a series of absorption peaks cor-
responding to the frequencies in equation (7) than to recognise a single
peak which might be lost in the random noise signals. .A disadvantage of
the side—bands is that it is often quite difficult to pick out one of the
frequencies (7) which is causing a resonance. This is especially true if
He is modulated by a varying field of appreciable amplitude. (A perb
iodic modulation is necessary for oscilloscope presentation.)

The great disadvantage encountered was that whenever the receiver

was adjusted so as to be very sensitive. it saturated the sample (7).

-5-

It might be well to point out that certain circuits employing ex-
ternal quench gave signals which showed no saturation effects; how-
ever. this type circuit was abandoned because it seemed rather in-
sensitive compared to the internally quenched type.

The saturation effect may be explained as follows:
Only'protone in the lower (parallel) energy state may absorb energy
to be detected as a resonance. The Boltzmann distribution function
shows an excess of’protons in the lower energy state and there are
several relaxation mechanisms operating'within the sample which re-
turn the nuclei from the upper state to the lower. If. however. we
send in a radiofrequency signal of too high an intensity. the natural
relaxation mechanisms will be unable to compete and the lower level
rapidly becomes depopulated. This. of course. changes the shape of
the absorption line. destroying its symmetry. The effect may even
annihilate the resonance in samples which have no paramagnetic con—
stituents. Since it was the line shape itself which was to be in-

vestigated. another method of detection was sought.

B. The Radiofrequency Bridge

The radiofrequency bridge method of detection employs essentially
three separate components: the oscillator. the bridge (containing
the sample coil). and the detector-amplifier (radio receiver). Since
the oscillator and the detecting mechanism (bridge) are separate.one

may place an attenuator between them. thereby reducing or removing

the saturation effect without appreciably affecting the sensitivity.
-6-

csiée bands. .Also a dispersion signal as well as an absorption may be
obtained from the bridge. The bridge is the essence of simplicity. as
it is Just a passive quadrupole network. .A detailed description is

given in the following section.

IV Radiofrequency Bridge Circuits
A. General Theory

The circuits used are of the "parallel” bridge type (8). A
parallel bridge consists of two circuits which receive the same input
signal and whose outputs are connected. Theoretically, the output sig-
nal of one "arm' of the bridge may be made equal in amplitude to and
1800 out of phase with the output signal of the other ”arm”. In this
manner a null balance is acheived.

The sample to be investigated is placed in one arm of the bridge;
at nuclear resonance the balance is disturbed and a signal is passed
from the oscillator to the receiver through the bridge. This radio-
frequency signal is modulated by the resonance absorption or dis-
persion signal. This amplitude modulation is detected by an a.m.
receiver and then presented on an oscilloscOpe or a strip chart re—
corder.

In general.the parallel bridge may be represented schematically

as follows:

 

 

()—--

 

In order to investigate the balance (null) conditions each arm my be

taken separately and represented thus

 

 

 

 

e1 being the applied voltage and 12 the output current. The general

circuit equations for such a network are

(8)

el: ZIILI'I'zt-le

el- 224 Li + zlev;

By shorting points a and b one obtains the representation

(9) . p
e‘ 2 Eu Lt + 2‘; L1

0 = ZZCL|+ 211 L1.

These may be solved for the ratio SC! yeilding
v.

(10)
+ Zn.-

 

This ratio is called the transfer impedance. Zr. By the reciprocity

theorem 212 a 221; hence.
(11)
i?

= " 2n 22.1 +. 2.1
zit

T

The input impedance 211 is that presented to an observer “looking" into

-s-

the circuit from the left. This is seen to be 21 4-22. The output
impedance 222 is that seen looking into the circuit from the right.
This is 23 + 22. The impedance common to both input and output is
212 s 22. Substituting these into (11) one obtains

(12)

21-: 3.: —'(2‘+23+2——}‘: ).
L; 2.

In a similar manner. one obtains for the other arm of the bridge

(13)

 

In the bridge circuit. figure 2. it is seen that the output terminals

of the two arms are connected. This means 12 and 12' flow in the same

circuit and. since we want a null. their sum must be zero:

(in) .
L1 +L; =0

Employing (12) and (is), one obtains

(15)
£2.
a:

:0 OT‘

 

.2... +
2.

v'
I
33:1- *' Eixr =‘ C) -

Finally. this becomes

 

(16)
z.+23 +Z_——‘23 + .==.'.'+2-3 + 5'1} :O.
2.. 2's.

This is the general equation of balance for any parallel type bridge.
In later sections it will be applied specifically to the two bridges

used.

Relation of the Bridge Signal to the Nuclear Susceptibilities

Consider the tuned circuit containing the sample coil,

I a n

; ti

 

oeqb

' Figure 4.
where R is the equivalent shunt resistance of the coil. The inductance
V of the coil may be represented by
(17)
L‘ = L01: ' + 4“Q(xe+xn)] .
Lo is the inductance of the coil without the sample in place and q is
a filling factor. If the radiofrequency field in the coil were homo-
geneous. q would be the fraction of the coil volume filled by the sample.
Xe represents the various electronic susceptibilities and Xeie the
nuclear susceptibility. The inductance may be rewritten as
(is) .
L=Lo(lue+ Afiq’xn)
where n. a 1 . uan. would be the permeability of the sample if it
had no nuclear moment. The electronic susceptibilities have no re-
sonance properties near the nuclear resonance, so u, may be considered
a constant for this calculation. The nuclear susceptibility has the
value X, for any magnetic field which does not satisfy the resonance

condition (6). X... changes markedly on passing througl the resonance:

~ -7 4; -8

-10.

When the bridge is balanced, the L—C combination is very close to

a resonance

(19) (AC 2: ‘

 

w Lo He
where w is the frequency of the signal generator and where the value of
hug X c has been neglected. This condition does not change appreciably
at nuclear resonance. The input admittance at or near nuclear resonance
is

(20)

 

: r-L- + ' C ~ ‘
Y R J". LU wL.‘Fe+411-¢LX“\

Making use of (19) and the fact that t4ngnl<<i,m. becomes

(21)

+
L
4:.
x
a

~._L
Y"R

I :0
Writing Q for R/wueLo and X. ~JX for Xnthis becomes

(22) u I
Y: {R—[ \ +411QX +j41rc3QX ] .

Notice that X: the imaginary part of the susceptibility. leads to a
real or dissapative term in I. This means that the absorption is pro-

portional to X”.
When the circuit is supplied by a constant current source I. the

potential drop across it is

(23)

V _-: 1/y = IR[\+S4W1QXn]-‘.

Writing '0 for IR and assuming that \4W%anl << \) (23) my

-11..

,be written
(214) ~
v :: Vo( ‘ " 4W%an\ .
This voltage V is Joined with an out of phase signal of approximately
equal amplitude from the other arm of the bridge. Let this second
I . a
voltage be denoted by -V1 and replace X... by X -5'X. : then the sig-
nal fed to the receiver is proportional to
(25)

V = V'—V. = VO-VI ’4W1QV°(K“+3X:\ '

One may thin]: of this output voltage as being the sum of two vectors,
(Yo-V1) and

I! . I
(26) v1: —4w%Q\l.(7L UM.

Vo

V

 

Figure 5. Voltages present at the bridge output terminals.

1— ‘. ._-

 

.-..~—.—.—_ _-.-_—_—

From the cosine law;

(27) wPL-wo—VJ‘Hvxx‘ + zw.—\I.|\v.| may“,
The phase angle of V1 with respect to 70 is 55. This angle is spec-
ified by
(28) x"

m ,6 = __

«'10

From this can be obtained the relations

~12-

(29)

 

 

ms é " I = 1C: amt
\[ \+'t0w\‘ ¢ K
s x. |
SM : __.._ .
95 75

Making use of these and the expansion of cos (6+¢) and neglecting
the very small Nil: one obtains. finally,
(80)

m" 5—: No-v.\"+ Z IV.\\V.,~V.| 41rq,&(7€"ccse at“).
The first term is purely radiofrequency; the second term is the mod»
ulation. The output of the a.m. receiver, assuming it is a square
law detector, is then proportional to
(31) 8’11”? Q Wel Wtht (flame: — X'sim e) ;
this is plotted by the oscilloscOpe. Two distinct cases are readily
distinguishable.
Case I:9=D.c1\'—The oscilloscOpe plots the absorption curve Xflunder
this condition of pure amplitude unbalance. (See Figure 5)
Case II: 6 = 1V; 0" 37g? The oscilloscope plots the dispersion curve
1' under this condition of pure phase unbalance.

for other values of 6 . mixtures of K, and 7("appear. They are
easily recognized by their lack of symmetry.

Various theoretical considerations (9.,10) lead to the curves of

I I
Figure 6 for the expected shapes of 2: and x .

-13-

 

nsu'eto “mandates-n19.“
"T”°'"'-U-"°-~IMquu—suuuunu
neon-museum. sum,»
wm~r,u~”m~

Ina-ecu nets a.m. cum m fie lees-nee
value.

. ‘.—-_; .-.—..._._

B. Twin 'T' Bridge
Using the notation of section IV.L and Figure 7. the various in-

pedances for the twin ”T' bridge are

(32)

 

. I ' _.__L_. _...'
Z.:—3/wc‘) £1=E+ 3(wC wLili3_ 446;

l ‘ I | _ . ’ ~
2.3—ch.) Elf.- ch) 23... R .

where B is the equivalent shunt resistance of the coil. Substituting
these values into equation (16) and separating the resulting equation

into real and imaginary parts, one obtains that at balance

(38) t __. C+C‘+Cq(\+ 9;) Phase \aaAance
w‘L. C:

 

J... : wIC‘Cz( l+_§_'.\ RI chasm-v.39 Edema.
R C"

Notice that phase and amplitude are adJusted independently by
means of the tuning condensers C and C' respectively. (See section
V for a numerical calculation.)

C. Purcell Bridge
Using the notation of section IV.L and also that of Figure 8.

the impedances for the Purcell bridge are
410-

(3h) . _t__ __ . .
z'z’J/Coc‘) Zl-t-‘l‘JhflC- fi)) 23:—5/wcz_

a - _l. I ,
ziz—J/wc\l) —wL’3’ Zzz-J/wct’

.-\
H
I-

J.
E“
(3

Equation (16) must be transformed to accomodate the half wave
length line which appears in Figure 8. This line acts as a trans-
former of ratio -l:l. This action necessitates replacing by its
negative the transfer impedance of the arm of the bridge in which

the line appears. Thus. equation (16) reads

(3:) 22 / I] I '
2.+Z;+—‘——2—2.-2 —i“i‘:— 0.
=2. 3 "‘2‘: “

Substituting the values of (3M) into (3‘), the balance con-

ditions are

 

 

(s6) ‘ (C C.+C— ‘ = ' ' ' '~—L-
cc. ‘+ wit...) c.‘c:(c'+c’+c gnu)
Phage: Balance:
R C ‘CI ‘- Rf C: C1] 0m “{“Le kAdvice ,

Notice that there is no convenient separate adjustment for ampli~
tude balance, as ordinarily only C and C' are accessable for tuning.

A numerical calculation is given in section V.

V Construction and Operation

A Twin "'1'- Bridge

The twin “T" bridge (11) is, except for the sample coil, en-
tirely encased in an aluminum box. The top is readily removable for

adjustment of the series condensers. These. however. are set only

-15-

R. 50:: c, any

R IOO n c: sir/if
Cu 'Iupf C.C' 65/41} with
L See text U. 5 ,uptrimmers.

 

Figure 7. Radio trequency "Twin-T" bridge (otter Anderson.)

once for approximate null balance and then remain fixed. All orb
dinary balance adjustments are made by means of the parallel tuning
condensers and their trimmers which can be manipulated from the out-
side by means of the four knobs (see Figure 1N.) The sample coil is
connected to the circuit by means of’about ten inches of half-inch
brass tubing. This tube is grounded to the box and axially through
it runs the lead to the tuning condenser C. The sample coil itself
consists of about seven turns of number 1“ plated copper wire.

The three series condensers Cl. 02 and 01' are ceramic disc
type trimmer condensers which are adjustable from 5 - SO upf. If
smaller values are necessary, they are made up by placing several
of these trimmers in series. The values shown in Figure 7 are the
approximate settings for 01, 02 and 61' at balance. Due to their
small size, accurate measurment is difficult.

The tuning condensers C and C' are of the midget variable type.
Their range is about 101- 65 puf. The trimmers on these are ofta
special type. That on O is made up of two brass discs: one is on a
micrometer screw and may be advanced toward the other giving a very
fine adjustment. (This is necessary due to the extreme sharpness of
the null.) The trimmer on Gt is a two plate variable condenser and
obtains its slow action'by means of a worm gear drive.

When the bridge is set for a null at 33.5 mc.. C is about bsuuf
(including about lZuuf of distributed capacitance in connections to
the coil) and c' is about :59 Me

If one inserts into equation (33) these values of C and C' and

the values shown in Figure 7 for the other components he obtains

-16-

L.5E f3illpch
Rs \3 an.

~JL-s
Q: at, 190.

Computation of these values employed w‘: (211 1) )2= “.163 x 1016

2 which corresponds to the frequency 1) = 38.5 me. at which the

leec
bridge was balanced.

The resistor B.is used to present an impedance match to the 5011
co-ax line which connects the bridge to the oscillator.
B. Purcell Bridge

This bridge is similar to one used by Bloembergen, Purcell and
Pound at Harvard (12) . Its physical construction. aside from the
circuit differences. is very similar to that of the twin 'r' bridge.
Except for the sample coil, the entire circuit is enclosed in a cop-
per box. Ceramic trimmers are used for the series coupling con-
densers: the values indicated in Figure 8 for Cl, 02. 01' and 62'
are for a null near 3M me.

The sample coil and connecting apparatus is the same as that
used for the other bridge. The "dummy” coil L'. in the other arm
of the bridge is similar to the sample coil but is contained inside
the case. This coil is loaded with a paraffin block to reduce its
Q to a value comparable to that of the sample coil.

The tuning condensers C and C' are of the 65unf midget variable
type. The trimmers paralleled across them are three plate variables

and constitue the fine tuning arrangement.

-17.

 

 

 

 

 

 

 

 

 

 

Ch h1 ;C2
- at u %-~--—~
1 " lac E f
.41? : :
S'QLHMMA l x I I
‘ ”SEC ::
C: If 1/062 5 ERecsivsr
/\ m
R 50 n (3.6 65lu+Lf with
C.,Q,C; 4};qu t,t lolu/lf trimmers
C'. 20,u.lu.f L,L' See text.

 

Figure 8. Radio trequsncy bridge (after Bloemberqen, Purceu, and Pound.)

 

50K 0— IOO
OOI
F _

Figure 9. The detector used

 

to balance the bridge.

 

For a null at 38.5 mo.. 0 is set at 68nd: (including about loud!
in connections to the sample coil) and C' is set at hhppf. Using
the values indicated in Figure 8 for 01. C?- and 02' and changing the
value of CI' to Tupi, equation (36) yeilds

L’ =’-=- .4.th
R 5-; 7.4 KIZ
Q': 3;, 2' 53 .
ad.
Computation of these values employed the results obtained for the
sample coil (11,13) in the twin 'T" balance conditions.

It was necessary to calculate the length of cable needed to
form a half wave length line. The relation between the wave length
in vacuum and in a cable with dielectric constant K is

(37) >3 :: 2L!!!“
{E

if one neglects the magnetic susceptibility. The coaxial cable em-
ployed had a polyethylene dielectric. This substance has a dielectric
constant of 2.26 (13) . For a frequency of 30 mc.. has: 10.0 meters
and
x:
.——::. —% ———————-‘O‘0 Y“ :1 3.33 m.
Z , Jzzc
The cable actually employed was 3.1a meters long. The bridge
balanced over the range 30 - 314 mc.. showing that the length is not
too critical.
The experimental set-up is shown in Figure 10. The output of

the signal generator was taken through a variable attemator and was

then fed into the bridge. Here the radiofrequency signal could act

-18—

 

 

 

 

 

 

 

 

 

 

 

 

MAGNET
* e
SAMPLE$ MODULATION
f COIL MODULATION COILS CONTROLS
9 J? LMAGNET 3
D ' VD

 

 

 

 

 

so we. as. PRE-
Eneano smos s assumes A p
1 1

RECEIVER

 

 

 

 

MILLIVAC
RECORDER

 

 

FIgureIQ Block dlegrsrn of the experimental arrangement.

 

 

 

 

      

IOO' 4e FINE newsman a MAGNET

ANAL mDULATION

MANUAL MANOR
g f N NOMATION
CURRENT CWTROL

II? ' ac. I

g

 

“‘ w” MODULATION COILS
»

 

 

Figure". learner and modulation controls.

 

 

on the sample, producing the resonance.

The sample coil was placed between the poles of a large electro-
magnet (1h). The field.produced by this magnet is variable and
ordinarily was set at about 8,000 gauss. A.current of 12 amperes was
necessary to produce this field in a gap of 2 inches. 0n the pole
pieces and immediately adjacent to the gap were two Helmholtz coils
(the modulation coils.) A 60 cycle alternating voltage was im-
posed on these coils. If the static field Ho was set near resonance
value. the coils would cause the field at the position of the sample
to be swept through the resonance value 120 times a second. In this
manner, the resonance occurred periodically and could be viewed on
an oscilloscope. The output of the bridge was fed to the receiver,
a Hallicrafters 51-62. Sometimes a preamplifier was inserted be-
tween bridge and receiver to boost the signal strength and the sig-
nal to noise ratio. From the receiver, the signal was applied to an
oscilloscope with sweep synchronised with the field modulation. The
oscilloscOpe had a Poloroid Land Camera attachment with which the
photographs in the following section were taken. Occasionally the
receiver output was placed on a.Millivac-Sanborn strip chart re-
corder. For this type record the field was modulated by manually
Operating the rheostat (see Figure 11) in the magnetic field circuit.
This caused the field to pass slowly through the resonance value.

The critical problem of balancing the bridge was solved in two
ways. The output of the Purcell bridge could be fed into the detec-
tor shown in Figure 9. The tuning condensers on the bridge were
then adjusted first for a maximum reading on the microammeter as one

of the circuits was tuned to resonance and then for a minimum on the

-19-

 

 

 

 

 

 

 

 

50

r].

 

 

 

80 VR ISO

 

 

 

 

 

+I50'

 

FiguelZIne 30 megocycle signal generals! and power supply.

 

.333 332. one caries... us. On 25.6359...

a wagon

all
>com +

hm

 

 

 

 

 

 

 

3:9. 3.3850 36E:
all c. 30.688 :4

a

 

 

 

 

 

 

 

On 0.:
a.
on

 

 

 

 

 

ho<m

 

 

 

 

 

CNN

 

meter as the other circuit was tuned to resonance thus balancing out
the output of the first circuit.

Theoretically, this method could be employed with the twin “T”
bridge; however, because of the very small series coupling capacitors
the output of this bridge was small and a somewhat more sensitive
detector would have been necessary.

The other method employed for balancing either bridge was to
observe the pattern on the oscilloscope. The trace was rather smooth
unless the bridge was balanced; at balance. the characteristic ran-
dom noise appeared on the screen. The resonance could usually then

be observed by adjusting the field to the proper value.

VI Results and.Lpparatus Photographs

The following set of photographs indicates the type of results
obtained with the two bridges.

It should be noticed that the signals are reasonably'like the
shapes indicated in Figure 6. The performances of the two bridges
are comparable as regards sensitivity and signal to noise ratio.

ill the photographs are for protons except Figure 19 which
shows the resonance in Fluorine. This is included to show that the

instruments will detect other than proton resonances.

Also included are several photographs of the apparatus used.

These photographs indicate construction details and the overall

set-up.

-20—

t O o
0 e 0 ¢ g

. 3..” shaman

 

 

 

is? seasoned shaman .eeesea ea. 2&3 ocean as seen Jeans 2.88 one

. ma gnu:

 

 

 

’ |
-

‘ n
V
Q

1'. '- -- all

 

Figure 15. General view of apparatus.

.aaea o» unmua Beau nu aest- eaooeoH
twaomo 0:9 .euoeaae coausxsaea assuage ousheaeooa o» endaoHSo eaaaeu
sues peasants seen as: canese one chops: an accused Bonn ens enacwue

0:9 .eapann :9: earn 0:» scat cecaeuno ens:m«e codeaeaeuc aneuaa>

.bH shaman

 

.e oasmaa co can one seat checaefiaso ea
uueaae ease ensueseaoec maeaa use wsuemoae
is“ n~e«a.racouuoeaam cs» :a eocscoeea on»

canons» cheese Cassy ceausasoos 0:9 .anu
cu wannabes one cascade acoesncs es» vane
oeuuoz .Hscmue :Odnaeauao :ouoaa tracked

emu

 

.Ana

V nanosecond reenmwus: on» no consumes

unease ea emee mundane» on» so asap one

.aeuss :« encased Beau Hecmue couuaaoep<

shaman

inseam: “yaw“: caeuss.:« ecouoaa loan
Hacwue coaeaeaeuc ccs neuuaaoep< “uuou
.excaap ”neoasm on» Beau eascm«m

 

oemoaan :9: :«s» 0:» Beau unwaa
.eacaap aaeeazm on» soak uuea .ecaaosH; Beau euccwue ceauaaoep<

.aH easwam

 

.aopsa as: cause. one .ao\>1 0mm «send me

haabaaasnoe housooea a you and seeds» enema .eadnmuu nodaaaonnm ”unwam .namann nofinaoaoflu

”whee .ueuaooeu chopnsmiompaaaaz on» so owcaap can has» on» mosh menadpno maenwnm

.om unawah

{sin};

 

VII Summary

The difficulties encountered in the course of obtaining the
foregoing pictorial data bring to mind certain obvious suggestions
for improvement.

With regard to the associated equipment, the main criticism is
noise and instability. Improvement could be effected in the signal to
noise ratio by using a tuneable preamplifier of good quality. Re-
duction in the noise level could be brought about by using a non-
micrOphOnic. crystal controlled oscillator and by using battery power
for all of the electronic equipment.

With regard to the bridges themselves. it is not certain that
optimal balance conditions have yet been.acheived. Obtaining the
best balance involves the tedious process of trial and error adjust-
ment Of the series condensers.

If one looks at the dispersion curves of Figure 17 (right),
he will notice that there is not the reversing effect explained in
Figure 18. This may be evidence of saturation in the twin 'T” bridge.
There is no resistive path to ground in the non-sample arm of the
twin 'T' bridge. It is possible that this circuit is not diseapating
much energy and, as a consequence. the radiofrequency power level is
too high in the sample arm. This strange effect demands further in-

vestigation if the patterns of this bridge are to be counted as re-

liable.

-21.

11.

iii.

iv.

10.

11.

12.

13.

Bibleography

w. Pauli. Naturwissenschaften 12. 7nl (192M),

P. Gilttinger. 2. Phys. 73. 169 (1931).
E. Majorana. Huove Cimento 9.N3 (1932).

 

E.M¢ Purcell. H.C. Torrey and R.V. Pound. Physical Rev. 69.37
(19u6).

G.E. Pake. Am. J. Physics 18, M42 (1950).

 

D.R. Hamilton..Am. J. Physics 9, 319 (lghl).
J.B.M. Kellogg and s. Millman. Rev. Mod. Physics 18. 323 (19u6).
A. Roberts. Rev. Sci. Inst. 18. ENS (l9h7).

J.R. Zimmerman ahc.n. Williams. Physical Rev. 73. 9M2 (19kg).

 

5.1. Tahsin. The Theory and Use of the Super—regenerative Receiver
for the Detection of the Nuclear Magnetic Resonance 15. 22 Masters
Thesis. M.S.C. (1981).

N. Bloembergen. Nuclear Magnetic Relaxation (Martinus NiJhoff,
lghs) 72,

w.N. Tuttle. Proc. I.R.E. 23. 23 (lguo).

 

C.J. Garter. Paramagnetic Relaxation (Elsevier’Publishing 00.,
Amsterdam. New York. 19h7),

P. Debye. Polar Molecules (Chemical Catalog 00.. New York 1929)

 

Chap. V.
3.1.. Anderson. Physical Rev. 76. 1162 (19149).

N. Bloembergen. E.M. Purcell and R.V. Pound. Physical Rev. 71.
u66 (19u7).

 

Fed. Radio and Telephone Corp.. Reference Data for Radio Engineers
(Knickerbocker. N.Y. 19h9) so.

-22-

lhw A. Luck. Design and Construction of a Laboratory Electro-

 

magget Masters Thesis. M.S.C. (1950).

15. B. Jacobsohn and R. Wangness. Physical Rev. 73. 9M2 (19MB).

' t ‘ .
.~.4' -

' 'lv t.
2.0. {.143
I .‘ . ‘ . ~

‘s ".i 1

I ' 5.7:} [k "'1'"
. ,‘, - . .\ “ I '1; '5
.V' ‘~‘_' ‘x. . .é‘.‘ .

 

ll\||“NIllllllllllflmilul\HliHIUIHIIHIlUlNHlIWl

 

31293017