A BIPOLAR PULSE TECHNIQUE FOR -
FAST CONDUCTANCE MEASUREMENTS
Thesis for the Degree of .Ph’. D.
MiCHiGAN STATE UN‘S‘IERSITY
DONALD E._JOHNS_0’N
I 9 79
LI B R A R Y
Michigan Sum
University
,HESIQ
This is to certify that the
thesis entitled
A BIPOLAR PULSE TECHNIQUE
FOR FAST CONDUCTANCE MEASUREMENTS
presented by
has been accepted towards fulfillment
of the requirements for
Ph.D. degree in Chemistry
4.4
\_/ r7
Malcor professor
Date February 27, 1970
0-169
mesme- °
ABSTRACT
A BIPOLAR PULSE TECHNIQUE
FOR FAST CONDUCTANCE MEASUREMENTS
By
Donald E. Johnson
The applications and limitations of several a-c bridge
techniques and models are analyzed, especially with reSpect
to polarization, the parallel (Cp) and series (Cx) cell
capacitances. and the frequency required. A complex model
of the conductance cell is developed, and the necessary ex-
perimental parameters needed in order to use a simplified
model are given. It is shown that a-c techniques are accurate
only over a relatively narrow frequency range. This range is
especially dependent on Cx and C for all of the a-c bridge
P
techniques. In the case of using a phase-angle voltmeter as
a null detector, the ideal frequency is shown to be propor-
tional to (CPCX)7§.
Recently. there has been considerable interest in ex-
tending conductance measurements to new areas. Several
attempts to measure rapid conductance changes are cited.
These techniques use unbalancedbridges and assume that any
change in conductance is due purely to the resistive com-
Ponent. At best, this is only a very rough approximation.
Other areas 0
high conducta
very low conc‘
heasurements
pecialiy with
The bi;
overcome many
niques. This
secutive cons
OPPOSite pola
current/yo 1m
Pulse to give
Combines the
3-0 techniQue
Nsec. the eff
Same as When
Slightly p01;
is Constant
C913 the sa
there is no
on Cx’ Where
A Pro1
aPplicetions
make aCCUra.
lMOhm). T
“Lance on C
titration g
n-w _
”Wt '13 Q
Donald E. Johnson
Other areas of recent interest include solutions with very
high conductance (such as molten salts) and solutions of
very low conductance (such as many non-aqueous solvents).
Measurements on these systems take special precautions (es-
pecially with respect to C and Cx)°
P
The bipolar pulse technique has been developed to
overcome many of the limitations of the traditional tech-
niques. This technique consists of applying two con-
secutive constant voltage pulses of equal magnitude but
opposite polarity to a standard conductance cell. The
current/voltage ratio is measured at the end of the second
pulse to give the conductance directly. This technique
combines the best features of both high and low frequency
a-c techniques. Since the pulses can be as short as ten
usec, the effect on Cx and the Faradaic impedance is the
same as when using very high frequency, since Cx is only
slightly polarized. On the other hand, since the potential
is constant by the end of the second pulse, the effect on
Cp is the same as when using very low frequency. In theory,
there is no dependence on C and a (t/Rxcx)2 dependence
P
on Cx' where-t=pulse width, and Rx=solution resistance.
A prototype instrument was built to test the theory and
applications of this technique. This instrument is able to
make accurate (0.01%) measurements over a wide range (100 ohm-
1 Mohm). It was found to have excellent linearity. no depen-
dence on Oh. and only slight dependence on Cx' An acid-base
titration gives a typical titration curve (continuous analog
output is obtained through a sample and hold amplifier). The
rate of etl
different '
results by
buffered I!
point chat
me of 1+1
Ano
prototype
has a one
Dr
..0ve
Donald E. Johnson
rate of ethanolysis of acetyl chloride is studied at three
different temperatures and is found to be consistant with
results by others. An EDTA titration of Zn2+ in a highly
buffered medium is followed conductometrically and an end-
point change of approximately #0 umhos out of a total conduct-
ance of #0.000 umhos is recorded.
Another instrument, which has many advantages over the
prototype instrument. was then built. This new instrument
has a one MHz crystal oscillator to control the pulse width
and the pulse repetition frequency. In addition. it has a
high absolute accuracy as 0.005% resistors are used. Also,
it has provisions for automatic temperature compensation over
a narrow (lo) temperature range. Many tests have been run on
dummy and real systems. It has an absolute accuracy of better
than 0.05% over the range from 10 ohm to l Mohm. falling to
0.1‘ at 10 Mohm. It has a maximum sensitivity of 1 ppm at
1 kohm. This instrument is nearly independent of Cp and Cx’
It also can be completely independent of the effects of lead
and contact resistance (using a four-lead system). Measure-
ments of solution conductances with this instrument and with
a Wayne - Kerr Bridge agree within the accuracy of the
bridge (0.1%).
Several applications of this technique are mentioned.
Since this technique permits fast, accurate conductance
measurements which are independent of ex and Cp, it should
prove extremely useful under a wide variety of experimental
conditions.
A BIPOLAR PULSE TECHNIQUE
FOR FAST CONDUCTANCE MEASUREMENTS
By
. \
Donald E: Johnson
A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Department of Chemistry
1970
Olav
; Av Arl/
II I o
R l
.yfldrt‘
Unisex"
a
on",
\vc
uh x u ‘ -
l y .s u. A
YA
.m A... e.
0v 5 n. V“'
~ 4.. 2‘ “U.
. t r” a;
a; a}. n.
6627??
74-70
ACKNOWLEDGMENTS
The author would like to express his appreciation to
Professor C. G. Enkc for his help and encouragement throughout
the course of this study.
He also gratefully ackrowledges a National Science
Foundation Traineeship, which supported him during his tenure
at Michigan State University.
The author wishes to rive Special thanks to his wife,
Sharon, for all the help and encouragement she provided, and
especially for typing and proof-reading the drafts of this
work.
ii
II.
TII.
IV.
TABLE OF CONTENTS
INTRODUCTION . . . . . . . . . . . . . . . . . .
HISTORICAL DEVELOPMENT OF CONDUCTANCE TECHNIQUES
A. 0-0 Techniques .,. . . . . . . . . .
Bo A-C TeChnlqueS o o o o o o o o o o o
1. Traditional Bridge Techniques. .
a. Typical Solutions. . . . . .
b. Solutions of High Conductivity
2. Unique Bridge Techniques . . . . .
3. Phase-Sensitive Bridge Techniques.
4. Techniques Used to Follow Conductan
Changes. . . . . . . . . . . . . .
a. Slow Changes . . . . . .
b. Rapid Changes. . . . .
c. Conductivity Detectors .
6‘
00000000000.
THEORY OF A-C BRIDGE TECHNIQUES. . . . . . . . .
A. General Bridge Equation. . . . . . . . .
B. Kohlrausch Bridge Balance. . . . . . . .
C. Phase-Sensitive Detector Balance . . . .
D. Applicability of Traditional Techniques.
THE BIPOLAR PULSE TECHNIQUE. . . . . . . . . . o
A. Theory . . . . o . . . . .
B. Theoretical Error Analysis
C. The Instrument . . . . . .
D. Instrumental Tests . . . .
l. Linearity Test . . . .
2. Dependence on Parallel Cell Capacita
3. Dependence on Series Cell Capacitanc
(4. A‘C Bridge Comparison. 0 o o o o o
E. Instrument Tests on Chemical Systems . .
1. Acid - Base Titration. . . . . . . .
2. Reaction Rate Studies. . . . . . .
3. Conductometric EDTA Titration of Zin
A NEW INSTRUMENT . . . . . . . . . .
o
o
o
o
.
C0
.
.
.
o
o
o
C
A. Improvements . . . .
B. The New Circuit. . .
1. Analog Circuit .
2. Digital Circuit.
iii
70
71
71
75
.
.
FWD 9.
i’.
C :
'O
L.
V. C. Temperature Compensation . . . . . . . . . . 78
D. Obtaining Absolute Conductance . . . . . . . 82
E. Instrument Tests . . . . . . . . . . . . . . 83
l.‘ AbSOlUte Accuracy. 0 o o c o o o o o o o 83
2. SenSitiVity. o o o o o o o o o o o o o o 85
3. Dependence on the Parallel Cell
CapaCitance. o o o o o o o o o o o o o o 86
U. Dependence on Series Cell Capacitance. . 86
5. Use of Four Cell Leads . . . . . . . . . 86
6. A-C Bridge Comparison. 0 o o o o o o o o 89
7. Temperature Compensation . . . . . . . . 90
VI. CONCLUSION O O O O O O O O O O O O O O O O O O O 92
REFERENCES 0 O O O O O O O O O O O O O O O O O I O O O 95
APPENDICES. O 0 O O O O O I O O O O O 0 O O O O O O O 99
A. PROGRAM FOR CALCULATING PSEUDO FIRSTEORDER
RATE CONSTANTS . . . o . o . . . . o o o o o 99
B. CIRCUIT DESCRIPTION OF IMPROVED BIPOLAR
PULSE CONDUCTANCE INSTRUMENT . . . . . . . o 105
iv
7.
C O
.2.
Table
l.
2.
3.
u.
5.
6.
7.
8.
9.
10.
ll.
12.
13.
14.
A-l.
LIST OF TABLES
Complex Conductance Cell Symbols . . . . . . .
Phase-Angle Voltmeter Dimensionless Parameters
Examples For Using Table 2 . . . . . . . . . o
Constants for Calculating the Ideal Frequency.
Phase-Angle Voltmeter Frequency Dependence . .
Dependence on Cx . . . . o . . . . . . . . . .
Reaction Rate Studies -- Acetyl Chloride and
EthanOIoo00000000000000...
Absolute Accuracy Test . . . . . . . . . . . .
Sensitivity gs. Resistance . . . . . . . . . .
Dependence on Cp . . . . . . . . . . . . . . .
Dependence on Series Capacitance . . . . . . .
Elimination of Contact and Lead Resistance . .
A-C Bridge Comparison. . . . . . . . . . . . .
Temperature Compensation in KCl Solutions. . c
Three-Parameter, Least-Squares, Curve-Fit
Programoooooooooooooooooo.
Input and Output of Least—Squares Program. 0 o
Page
ll
26
28
31
52
57
84
87
88
LIST OF FIGURES
A-C Conductance Bridge. . . . . . . . . .
Complex Equivalent Circuit of Conductance
A-C Bridge Analysis . . . . . . . . . . .
A Bipolar Pulse Conductance Device. . . .
The Prototype Bipolar Pulse Instrument. .
Timing Sequence During Measurement. . . .
Constant Potential Generator. . . . . . .
Linearity Test. . . . . . . . . . . . . .
Acid - Base Titration . . . . . . . . . .
Reaction Rate Studies . . . . . . . . . .
Temperature Dependence of Reaction Rate .
Temperature Profile During Reaction . . .
Conductometric EDTA Titration of Zinc . .
The Improved Bipolar Pulse Conductance Instrument
Digital Logic Circuit for Improved Instrument
Waveforms Of Logic CirCUit. . o o o o o 0
Circuit Diagram of the Improved Conductance
InStrumentocoo-000000000.
The Bipolar Pulse Instrument Front Panel.
vi
Page
106
110
a‘ I
e
and the
:4
VA‘
.1.
I. INTRODUCTION
Conductance measurement is one of the oldest of the
accurate electrochemical techniques. The basic a—c bridge
technique was developed nearly a hundred years ago, and
hasn't changed significantly since that time. Throughout
the history of conductance techniques, minor modifications
have been made in the methods and the models in order to
more accurately determine and characterize the conductance
and the equivalent circuit of the conductance cell. The
conductance techniques which are used today are quite limited
in application since each was developed for a particular
model of the conductance cell. The conductance measurement
is valid only as long as the model truly represents the cello
The most appropriate model, in turn, depends on the experi-
mental conditions used, especially the cell design, resis-
tance range, and frequency.
In discussing the history of conductance techniques
in Chapter II. a complex model for the conductance cell will
be developed. Several examples of what has been done to
simplify the cell model will be cited. and the best model
for normal conductance work will be critically analyzed in
Chapter III. It will be shown that the traditional Kohl-
rausch method and the new phase-sensitive techniques have
. c'v
c.
S
xyfiv‘i.’
VSa'
'
a
“3’0““
M-‘
CI'
3'
v. e e w... t C t t i :u a» I
AM“ w...“ .r» A» V!‘ s y .1 cl 9,.“ .‘ st. :
2
very distinct frequency ranges in which the results fall
within reasonable error limits. The frequency range is
especially dependent on the parallel and series cell capa-
citances.
Recently. there has been wide interest in systems
where the traditional conductance techniques cannot be used,
or can be used only with great difficulty. These systems
include those with very high resistance (such as much non-
aqueous work), with very low resistance (such as work in
molten salts), or where platinized electrodes cannot be
used (if surface adsorption is a problem). Furthermore.
traditional bridge techniques are not readily applicable to
systems which require continuous or instantaneous conductance
measurements (such as in following conductance changes in
reactions. titrations. flow systems. or ion exchange).
Several attempts to measure conductance under these condi-
tions will be cited.
A new technique, which is less restricted, will then
be introduced in Chapter IV. This technique uses the best
characteristics of both high and low frequency measurements.
It is demonstrated on dummy and real systems and is found
to be very fast. accurate, and wide-ranged. In addition.
it is much simpler to use than traditional techniques as
it requires no balancing (either capacitive or resistive).
Finally, a new instrument that is extremely accurate
will be described. This model includes a crystal oscillator
timing device. provisions to completely eliminate contact and
lead resistance, and automatic temperature compensation over
a narrow temperature range.
*7
.5.
ductanc
a ConS'
{if
H
‘0
L‘
l n
(/l
1.)
H
.4
r‘
II. HISTORICAL DEVELOPMENT OF CONDUCTANCE TECHNIQUES
A. D-C Techniques
Thus far. there has been only one technique which can
compete with the a-c bridge techniques to accurately obtain
conduCtance. This is the d-c technique. introduced by
Newberry (1) in 1918. He showed some of the disadvantages
of a-c methods and presented a cell arrangement for d-c con-
ductance measurements. The d-c technique consists of applying
a constant current through a cell and measuring the potential
drop across a certain portion of the solution with two elec-
trodes which carry no current. Newberry used calomel elec-
trodes for all four electrodes. and made sure that each pair
had exactly the same open circuit potential. He compared
his measurements of the specific conductance of KCl solutions
to those of Kohlrausch, and found the Kohlrausch values to
be about 0.5% lower.
Eastman (2) extended this d-c technique and found the
Kohlrausch values to be about 0.2% low. He built a bridge
arrangement which could be used with either d-c or a-c tech-
niques. He concluded that conductivities at zero frequency
are about 0.02 - 0.03% higher than those at frequencies
between 1 - 2 kHz. Fuoss and Kraus (3) found the d-c tech-
nique necessary for the measurement of the conductance of
solutions with resistances over 100 kohm. In their
3
:onumenta
constant
103 Re tr“
Pl
1 '1
m (D
H (.2.
F10
N
fl
1"
a
monumental work to determine the influence of dielectric
constant on conductance. a-c techniques were used up to
100 kohms. and d-c techniques above that value.
Gunning and Gordon (4) showed that the d-c technique
can be a serious competitor to the a-c bridge techniques
(0.02% accuracy). They pointed out the necessity for small.
reversible electrodes. and they improved the cell design.
Elias and Schiff (5) modified this technique even further by
using a liquid junction to remove the measuring electrodes
from the solution. This solved the problem of finding rever-
sible electrodes for some non-aqueous solvents.
Gavis (6) rigorously derived the current-voltage rela-
tionships in d-c conductance work. He gives equations for
the proper cell design. and shows under what conditions a-c
methods may be used. The determining factors include dielec-
tric constant. resistance. parallel cell capacitance. and
frequency. Others (7. 8) have also had success using d-c
techniques. especially with very low conductance solutions.
B. A-C Techniques
1. Traditional Bridge Techniques
a. Typical Solutions
From the very beginning of conductance measurements.
it was realized that an a-c signal was necessary in order to
reduce polarization of the electrodes. If a d-c potential
were applied to the cell. the electrodes would become polar-
ized. and would quickly reach a potential at which a reaction
would take place (e.g. - hydrogen liberation). Applying an
c
h
'1.
ith t
.;U'.
half-c
W
a flu b...
V O a
.l R
Q s u e
H... Ox; “I
Cc m... a .
5
a-c signal to a Wheatstone bridge with the cell in one arm
did cause a problem: the bridge could not be brought to a
null. The electrodes were slightly polarized during each
half-cycle. giving rise to an effective capacitance (the
double-layer capacitance) in series with the solution
resistance.
Kohlrausch (9) was first to put a capacitor in parallel
with the standard resistor in order to compensate for the
reactance caused by the double layer capacitance. For con-
ductance work prior to this. see the literature cited in his
book (10). Figure 1 shows a typical conductance bridge
where Rx is the solution resistance. Cx is i the double layer
capacitance (for identical electrodes), and Cp is the inter-
electrode and stray parallel capacitance. The Kohlrausch
equivalent circuit is shown in Figure 1 with C O. This
p:
method introduces a relative error of (RXCXu)'2, where a is
the angular frequency (27f), as will be shown later. Kohl-
rausch recommended several precautions to cut down on this
error which is due to polarization of the electrodes:
1) platinize the electrodes
2) use frequencies > 1000 Hz
3) use large electrodes
4) design the cell so that RX > 100 ohms
Washburn and Bell (11) extended the Kohlrausch method
by using a "pure" signal source and a tuned telephone detec-
tor. They also stressed more careful capacitance balancing.
Washburn (12) discussed the theory of cell design and showed
\
that at high frequency. Cx can be neglected, but that
Figure l
A—C Conductance Bridge.
1191 <
I‘d
“F
~-
0
mi
3110...
flag?
y VA
m) alsc
(h
cOI
'rr‘fi
e U
3
U. gun
v
conduc
ta
IRE TERI
a
V
phase 2:
‘wo pan
.
U
9..
.e L.
03011‘
‘
v0
‘8.
‘3‘
P
9.
S
9
b.
l e
l .3 F
9 h. 3
h. A: nu
8
parallel cell capacitance causes interference. Hall and
Adams (13) extended the telephone null detector by using a
thermionic amplifier to increase the sensitivity. Gieringer
(1#) also developed an improved a-c null indicator.
Jones and Josephs (15) developed a direct-reading a-c
bridge technique using a modified Wagner ground that extended
conductance measurements to solutions of low conductance (up
to about 60 kohm). They demonstrated that for bridge balance,
the reactance in any arm must be balanced in another arm.
Thus. the relationship R
Rx = R R3. is valid only if the
1 2
phase angle between the voltage and current is the same in
two pairs of adjacent arms of the bridge. They concluded
that the method of Kohlrausch was the best approximation to
this balance. and that R1 and R2 should be equal to minimize
the chore of balancing reactances. They discounted the method
of Taylor and Acree (16), who used an inductor in series with
the cell. They also discarded as impractical the idea of
putting a large capacitor in series with R8.
Jones and Bollinger (17) made refinements in both the
oscillator and detector portions of the a-c bridge. Their
modifications reduced polarization because of a lower signal
voltage. They were first to realize the importance of the
parallel cell capacitance (18). Their model consisted of a
series resistance (Rt) and capacitance (Ct) across a cell of
pure resistance (Rx)' They were working with a capillary
cell which had large filling tubes (the source of Rt and Ct)’
and by its nature, a large Rx' which allowed CX to be ig-
nored. They used R and Ct to eXplain the Parker effect
t
which 0
censtan
problem
imity c
(I)
ideas a
itprove
I.
:D
’ 1
:3
(D
(D
m I
0 H0
:T H' 73
:1 a H
0
9
which other authors (19, 20, 21) reported as a changing cell
constant with increasing resistance. Their solution to this
problem was to design the cell in such a way that close prox-
imity of parts of opposite polarity would be avoided.
Shedlovsky (22) was able to obtain very accurate meas-
urements (0.001% sensitivity) by using a screened bridge to
measure conductance. Fey and Martell (23) used Shedlovsky's
ideas along with suggestions by other authors to build an
improved conductance bridge. Cole and Cross (24) developed
a bridge that could accurately measure both conductance and
capacitance over a fairly wide range. Hladek (25) designed
a similar bridge (0.04% accuracy).
b. Solutions 32 High Conductivity
When working with solutions of low resistance, one
must be concerned not only with Cx and Rx' but also with the
lead and contact resistance. RC, and the Faradaic impedance,
ZF, if there is any possibility for a reaction to take place.
A complete conductance cell can be represented by the equiva-
lent circuit shown in Figure 2, with the symbols described
in Table l. ZF consists of several terms, including the
frequency dependent Warburg impedance (26), which varies in-
versely as the square root of the frequency. Winterhager and
Werner (27) have used a high frequency {50 kHz) bridge to
eliminate the importance of ZF, since at high frequency. the
admittance of Cd is infinite with respect to that of ZF' No
technique has been used to compensate for ZF; the experimen-
tal conditions are adjusted to minimize its effects (high
frequency. low es. etc.). Feates. Ives, and Pryor (28) have
10
Figure 2
Complex Equivalent Circuit of Conductance Cell
cl
~
‘mn
158‘.)
,‘U
+9
fly
.3
11
TABLE 1
Complex Conductance Cell Symbols
Name
solution resistance
double layer capacitance
Faradaic impedance
contact and lead resistance
contact and lead capacitance
inter-electrode capacitance
filling tube capacitance
filling tube resistance
Value in Simplified Model
(Figure 1)
remains the same
forms Cx
>> impedance of C
d
reduced to zero
combine to form Cp
low or non-existent
sta
res;
o
N
A
1
call;
<
l
0"n
.A.
’,
i
12
stressed the importance of eliminating ZF. They also have
devised a method to eliminate the effects of lead and contact
resistance by using a special switching arrangement in their
bridge. They used a four-lead, double-cell technique.
Metcalf (29) points out that the double layer becomes chemi-
cally ill-defined if ZF is important. He also points out
that ex is normally 10”
- 106 times Cp. Probably the best
method to date for measurements on solutions of low resis-
tance that have low Cd is that recently developed by Robbins
and Braunstein (30). They use a specially designed bridge
with a series (instead of parallel) RS and CS. and have ob-
tained an accuracy of better than 0.5% (RX = 100 ohm,
Cx 8 20 pF).
2. Unigue Bridge Techniques
Bertram and Cruse (31) have developed a unique R-C
method for conductance measurements. The basis of the method
is that the frequency and magnitude of an R-C signal genera-
tor is dependent on the characteristics of the feedback loop.
They point out the frequency and sensitivity limits of their
method over its entire range of conductivity measurement
(10'12 - 10 ohm'lcm'l). Also. the power of this method is
pointed out (32) and the method is compared to d-c methods.
Cruse and Bahr (33) use the R-C method with a three-electrode
system to follow reactions with half-lives of several hun-
dred seconds.
Wershaw and Goldberg (34) used the equivalent circuit
shown in Figure l for the cell model and devised an ingenious
technique using an inductive divider bridge to obtain very
must be
arsed pr
tcrs mus
\u ‘-
SuppOI‘.
30 ECELE
zillion.
13
precise measurements of conductance changes. They have a
variable capacitor in each of two bridge arms which is
varied to balance the reactance of Cp or Cx' The frequency
must be varied to ensure that each cell capacitance is bal-
anced precisely. The main disadvantage of this method is the
length of time required for a measurement (RS and two capaci-
tors must be balanced and the frequency varied). Strong
support for the simple model is indicated, as they were able
to achieve resistive balance sensitive to 0.2 parts per
million.
There have been many other attempts made to obtain the
conductance more easily or more accurately using a-c bridge
techniques. Warshawsky (35) developed a multiple-bridge
circuit. Griffiths (36, 37) deve10ped a contactless, four-
electrode system (using a three-terminal transformer as the
ratio arms) that uses relatively low frequency. The four-
electrode electronic half-bridge system of Ferris (38) and
the very low frequency (0.8 Hz) method of Hanss and Guermon-
prez (39) are just two more examples.
3. Phase-Sensitive Bridge Techniques
Recently. several phase-sensitive devices have been
applied as null detectors in a-c bridge arrangements. Janz
and McIntyre (40) used a modified impedance comparitor to
separate the difference signal into in-phase and quadrature
components. They were able to achieve very precise bridge
balance (0.002%) over a wide resistance range (2 ohm -
20 Mohm) using an inductively coupled ratio arm and four cell
leads.
Wal
observe a
resistive
an accura
nique (#2
conductor
half-life
bfidge((
Sc!
can be us
fully wi‘
greatly 1
hsted 1;
later).
L'o Tech
\'
3.
Cf
l4
Walisch and Barthel (41) have developed a technique to
observe a rapid phase-shift in the difference signal near
resistive balance. They claim a sensitivity of 0.0001% and
an accuracy of 0.03% for their method. They used this tech-
nique (42) to follow the titration of Fe2+ with Cr2072‘
conductometrically. They also studied a reaction with a
half-life of approximately one minute using a self-balancing
bridge (0.1% accuracy).
Schmidt (43) has shown that a phase-angle voltmeter
can be used in a conductance bridge circuit quite success-
fully without any capacitance balancing. Although this
greatly simplifies bridge balance. the frequency must be ad-
justed if one expects accurate results (as will be shown
later).
4. Techniques Eggg 32 Follow Conductance Changes
a. §l2g Changes
Often, it is not necessary to know the absolute con-
ductance, but merely to follow a change in conductance. Most
often. under these circumstances. it is desirable to be able
to follow the change without further manual intervention
(such as balancing the capacitance and resistance). Some-
times there is time to allow semi-balance or even complete
balance while following such changes. For example, Hladek's
method (25) allows separate capacitance and resistance balance
with little interference between them. Robertson (44) used a
different method to obtain separate resistance and capaci-
tance balance. He studied some reactions (44, 45) with half-
lives on the order of 1000 secs and obtained an accuracy of
4 a
8057‘: 1n
4593 18
to withi
Si
afiplifie:
method h;
total) 0'
uEVt
(l - 2,4 <
1’10 Lepra‘
100 k‘z .
as the
o._ S‘
4
Eu:
WOrk.
fluetifl _
15
0.5% in k (pseudo first-order rate constant). The method
used is unique in that two arms of the bridge consist of
identical cells: one for the reaction and the other for a
titration. In this way. he could titrametrically determine
the concentration Kg; time. Murr and Shiner (46) developed
an improved technique. and were able to obtain rate constants
to within 0.03%.
b. Rapig Changes
Sirs (47) used a 50 kHz oscillator with a rectified and
amplified difference signal to obtain a d-c output. His
method has an estimated relative accuracy of 5% (0.5% of
total) over a 10% conductance change. and is capable of fol-
lowing reactions with half-lives of about 30 msec. Prince
(48) developed a similar method using a 20 kHz signal and was
able to follow reactions with half-lives as small as 10 msec
(l -‘2% conductance accuracy). Strehlow and Wendt (49) were
able to follow very fast reactions with relaxation times as
short as 50 psec, using a "pressure-step" technique. They
photographed the entire relaxation curve envelope (50 or
100 kHz signal observed on the time scale of the experiment)
as the system returned to bridge balance.
Eurato and Leimu (50) used a system that was based on
a working curve for bridge off-balance in order to obtain
kinetic results with half-lives in the seconds range. Dusek
and Lesek (51) developed an improved four-electrode technique
for measuring fast conductance changes. Tregloam and Laurence
(52) have devised a very sensitive bridge arrangement using
a tuned a-c detector followed by an a-c to d-c converter.
16
They can follow a 1% conductance change within 0.5% (0.005%
of total) with a response time of 1.5 msec.
c. Conductivitnyetectors
In addition to their use in kinetic studies. there has
been much interest in the past few years in developing good
detectors for ion-exchange and liquid-liquid chromatography.
Conductivity methods have even been applied to paper chroma-
tography (53). Duhne and Sanchez (54) developed a four-cell
conductivity bridge for use as an ion-exchange detector.
They used a 20 V, 20 kHz signal source. Pecsoc and Saunders
(55) developed a very sensitive, low-volume conductivity
detector for liquid chromatography. They used this method to
determine the distribution coefficients in inorganic gel
chromatography (56). Knudson, Ramaley, and Holcombe (57)
suggest two circuits for general conductivity detectors (es-
pecially for chromatographic use) using operational amplifier
bridge networks. One circuit is extremely sensitive (0.001%)
and is linear for small changes. The other circuit is not
as sensitive but is linear for even large conductance
changes.
Additional information on conductance methods is given
in several review articles. Gerischer (58) has written an
excellent article with 203 references. He critically ana-
lyzes various bridge networks for conductance measurements.
His article includes balance equations. bridge equivalent
circuits. effects of Z Cp. Cx' lead inductance, and cell
FD
design. Cruse (59) also has a very good article, with a
review of the classical bridge techniques, and with Special
I
l
emphasis c
consulted.
17
emphasis on high frequency methods. A review of the theory
of electrolytic solutions is given by Barthel (60) with 248
references. There are also many books (61 - 64) on the
subject of conductance theory and methods that can be
consulted.
n
U
~
g
OI‘Q
.er. dev
sated to
heref
fit an al
:siel i
y.
b
as;
o‘c‘
III. THEORY OF A-C BRIDGE TECHNIQUES
A. General Bridge Equation
A complex model (Figure 2) of the conductance cell has
been developed in Chapter II. This model is far too compli-
cated to use realistically in practical applications.
Therefore. the experimental parameters must be adjusted to
fit an acceptable simplified model. The best simplified
model is shown in the bridge network of Figure 3. This model
assumes that the lead and contact resistance is negligible
and that the frequency is high enough so that ZF can be con-
sidered infinite. These are reasonable assumptions since
they are necessary conditions for accurate measurements.
When the experimental parameters make this model inappropriate
accurate conductance data is extremely difficult, if not im-
possible. to obtain. Therefore. the following mathematical
theory of a-c conductance measurements will be based on this
simplified model.
Assuming a sinusoidal signal source for the bridge net-
work in Figure 3,
eS = V sin at
(all voltages with respect
e2 = A sin at + B cos at to the reference point)
e = D sin at + E cos wt
18
19
Figure 3
A-C Bridge Analysis
Cell consists of R . C , and C
x x p
where y
D
.
.
03‘
C
p
= C
.
' I
Rd _
F
wastitu‘t
and Solvi
and
21
From Figure 3,
1 eZ-eC - C deC
2 ‘ 'R;' ‘ x dt
or (A-D) sin wt + (B-E) cos at = RXCxw(D cos wt-E sin wt)
(1)
Solving Equation 1 for D and E and letting Ky = wRXC
where y = x. s. or p. yields:
D
(A + ka)/(1_+ 183') (2)
E
. 2
(B - KXA)/(l + KX ) (3)
If no current flows through the measuring device between
e1 and e2,
. dez e2-eC deS de2 eS-e2
..det + RX =Cs
> Cp/Cx. this becomes 1
.1. ..
kX a (ox/cp)[(1 + 122/121); — 1]2 (26)
Equation 23 can be rearranged to get:
a 2 2 2
Kx (op/0x) + kX [(Cp/CX) + zap/cx -
Q(1+R2/R1+Q R2/R1)] —(1+0)2 RZ/Rl = 0
(27)
which is quadratic in sz. Table 2 shows Kx as a function
of RZ/Rl' Cx/Cp. and Q. The estimated value of XX is calcu-
lated using Equation 26, the exact value using Equation 25,
and the Kx for the errors indicated using Equation 27. Two
26
00000.0 00500.0 00000.0 00500.0 00050.0 00050.0 00.5 0.000
00000.0 05500.0 50000.0 00500.0 00050.0 00050.0 00.0 0.000
55000.0 00500.0 00000.0 00500.0 00050.0 00050.0 00.0 0.000
00000.0 00000.0 50000.0 00500.0 00050.0 00050.0 00.0 0.000
00000.0 00050.0 00000.0 00050.0 00000.0 00000.0 00.5 0.00
00000.0 00050.0 00000.0 00050.0 00000.0 00000.0 00.0 0.00
00000.0 00050.0 00000.0 50050.0 00000.0 00000.0 00.0 0.00
00000.0 00000.0 00500.0 00050.0 00000.0 00000.0 00.0 0.00
00000.0 00000.0 00000.0 00000.0 00000.0 00000.0 00.5 00.0
00000.0 00500.0 00000.0 00000.0 00000.0 00000.0 00.0 00.0
00000.0 00000.0 00000.0 00500.0 00000.0 00000.0 00.0 00.0
00000.0 00000.0 00000.0 00500.0 00000.0 00000.0 00.0 00.0
00000.0 50050.0 00000.0 00050.0 00000.0 00000.0 00.5 00.0
05000.0 00050.0 00000.0 00050.0 00000.0 00000.0 00.0 00.0
00550.0 50000.0 05000.0 00050.0 00000.0 00000.0 00.0 00.0
00000.0 00000.0 00550.0 00000.0 00000.0 00000.0 00.0 00.0
05000.0 50500.0 00000.0 50500.0 00000.0 00000.0 00.5 00.0
05000.0 50500.0 05000.0 00000.0 00000.0 00000.0 00.0 00.0
00000.0 50000.0 00500.0 00000.0 00000.0 00000.0 00.0 00.0
00000.0 00005.0 00000.0 50000.0 00000.0 00000.0 00.0 00.0
e00.0+ 000.0- e0.0+ 00.0- pomxm .000 (mwxxo 0
x. woq m\ m
A 00 000
mpmpmsmuem 0000:00mcw50a smpmsp0o> m0ws<:emmnm
N mqma
e.eeed
c.voeo.
e.eeeu
b.0000
m.eeeu
o.eeeeu
_
o.oeeeu
eu
”.meaq.s-
e.aea
n.qeaa
9.0::
~.eoo)
e.ee=u
u.aav$
9.3::u
e.eeeeu
a.:==u
e.ee:eu
e.:=ou
=.==euu
e.a:a
n.o=oa
9.733
n.23a3
e.e==u
u.oaaa
=.ee=u
0.33:1
e.:=e4
e.ee=eu
e.eeeu
e.eeeeu
c.333
o.ee>o
9.>¢)
~.eoa>
e.eeeu
u.:oo:
e.eeeu
9.009)
e.eeeu
e.ee=eu
v.e=eu
v.eeeeu
e.aaa
n.ne9e
9.099
«.9993
a.=eeul..
Muooo.
e.eeeu
e.eeeeu
e.~eeu
n.»es.u
mqnmefl.
o.oneeu
n
e.ceeu
u.cooo
e.eecu
0.0993
e.eueu
e.eeesu
e.eeeu
e.eeeeu
e.eeeu
e.eeeeu
e.eeeu
e.eeueu
e.eecu
u.ooo)
e.ceeu
o.oooo
e.eeeu
e.eeeeu
e.eeeu
e.eeeeu
e.eeeu
e.eee=u
u.eesu
u.eseeu
e.eeeu
u.oooa
e.es=u
c.oooa
e.eeeu
e.eeeeu
e.eeeu
e.eeeeu
m.eeeu
o.eeeeu
u.oeeu
e.eueeu
e.eeeu
u.oooo
e.eeeu
o.oeoe
.e.deed;.
e.eeeeu
e.eeeu
oweeeeu
v.oveu
e.uoeeu
eJNomu
u.ouoeu
u
e.eeeu
c.oooe
e.eeeu
o.oeoo
e.eeeu
e.eeeeu
e.eeeu
e.eeeeu
e.eeeu
e.eeeeu
e.eeeu
e.eeeeu
e.eeeu
c.9900
e.eeeu
o.oeoo
e.eeeu
e.eeeeu
e.eeeu
e.eeeeu
e.eeeu
u.eeeeu
e.eeeu
c.eeeeu
e.eeeu
c.oooo
c.:oou
o.oooo
e.eeeu
e.eeeeu
u.eeeu
n.eeeeu
~.neeu
u.neeeu
c.cneu
n.~oeeu
e.eeeu
c.0000
e.eeeu
o.oooo
u.eeeu
m.eeeou
¢.meeu
v.uneeu
e.eumu
e.ecmeu
¢.m~en
v.cunvu
Q.
o.roeu
e.aeeeu
e.eeeu
e.ueeeu
e.eeeu
e.eeeeu
e.ceeu
e.eeeeu
e.c004
e.ceeeu
u.eeeu
u.cooou
e.eeeu
e.eeeeu
e.eeeu
e.eepem
e.eeeu
c.cocod
e.eeeu
e.eeeeu
m.ceeu
o.cooeu
u.oeeu
e.euesu
e.eeeu
e.eeeeu
o.ro:u
e.ceenu
e.ceeu
u.eeeeu
o.eeeu
e.meeeu
v.ovou
n.uoeeu
e.eamu
v.ouoeu
e.eeeu
e.eeeeu
e.eeeu
u.eeeeu
e.ueeu
~.oeeeu
n.uueu
~.voveu
v.0nnn
e.enonu
ouonWau;
n.1sunn
u.
c~.~
u~.u
c..~u
nu.o
u~.sc
o~.vu
cm.~vn
oe.vcu
eo.ec¢u
o~.~oc
uo.~eov
um.o¢mu
me.u
ve.n
uv.oe
uc.u~
~u.ou~
on.oeu
ee.neeu
oe.emv
ue.ecmv
c..~unu
o~.vwhma
oo.omen
ev.~w
cv.~u
on.vmu
u~.~c
uu.~eo
on.~vn
uo.n«vn
eo.ovvu
no.movvu
uo.~nov
cu.o~uov
e~.noonu
~0.os
ev.on
00.non
~«.vu~
e~.«vd~.
vo.n¢en
on.nnoou
no.9cnv
so.oeenv
ou.¢«unu
un.~¢~umu
wo.oc~en
Aumovu
ee.e
eu.e
ne.e
ne.e
~n.e
en.e
co.»
n~.~
um.-
ee.o
ee.mo
oc.un
ne.e
ee.e
~n.e
_~uum
uu.u
me.»
ou.en
u¢.-
me.v~«
eo.mo
oe.smo
eu.eup
~n.o
~u.°
eu.u
u..n
u«.un
e~.en
oo.~en
ne.-n
no.a-~
o¢.~vo
«ou.~.e
-..e.n
cu.»
nu.»
ue.un
eo.un
oe..uu
no.uew
oo.eeo~
on.uce~
on.nn.e«
ne.euno.i
ov.eeu~o
uu.~eco~
Aauavu
mflnmuzw
u~.~
v~.u
ov.~u
uu.o
u~.no
w~.vn
uo.~vn
eo.vvu
en.evvu
vm.~ov
em.ueoc
w~.ounu
se.~
vo.n
ev.on
u'wum
~u.¢u~
nn.ceu
an.noeu
ce.onv
ou.eemv
.ov.~nmu
~u.¢wsau
mm.o~en
on.-
cv.~u
om.v~u
us.~o
uu.~eo
cm.~vn
uo.n~vn
on.uvcu.
no.acoou
uo.~nov
co.muuev
om.noeeu
~o.ee
e...»
co.non
~u.vu~
.e~.ucu~
em.neeu
00.nnneu
no.90nv
um.eeeno
-.v~uau
~u.uo-ou
mn.oc~en
as:
o»..oru re.eeeeueoceeeeu
ew.«ec .e.eeeu ceeeeeeu
Co.r0c0 no.aroa cacoccu
v¢.oomu co.ecu eoeeeeu
no.00co. re.eeu coeoeu
oo.eeun we.eu eeeeeu
m¢.ooe- ce.ev eeeeu
on.eenua re.u eeeeu
nc.escma~ ce.u eeeu
o...eoo.u -nuue soeu
o¢.oucovou uu.a ecu
on.~m00cnu .o.o ecu
ou.uoru ne.eeeeueeeeeeeu
or.to: co.oco« eoooeoou
no.rncn re.eeeu ceeeeeu
u...oru «e.eeu coseceu
ou.coon. re.eeu eeeeeu
~u.¢son .e.eu eeeeeu
«w.u¢s~r ro.ou eoocu
uu.uov~e ce.fi eeeeu
eu..s¢v«~ ce.u eeeu
n¢.e~¢omu wu.e - .eeeu
ac.cewo¢ou mu.e eeu
no.t0r0Lnu .o.o eou
oe.uoru ce.eeeeueoeeeeeu
or.rer ce.eceu eoeeeeeu
n1.oncm eo.ocou cooouou
np.ooru .e.eeu eeeeeeu
vn.roon. ce.eeu eeeeeu
or.oonn ce.eu eoeeeu
n¢.ro¢~r re.eu eeeeu
uc..o«~. we.u ease.
as.rv¢~a~ .e.u eeeu
oe.m~¢n.u eu.e eeeu
=.wn.n~eomou ru.e. eeu.
on.uure¢nu ue.e eeu
on..oru ce.eeeeueoeoeeou
em.neo ce.eeeu eeeeeecu
ve.rnco ce.eeeu eeeeeeu
os.ooau ce.eeu eeeeeou
no.~dmfld- ee.eeu._queeu
uv.uomn re.eu|. eeoeeu
.e..u.u. .e.eu eeeeu
ac.uoaea we.u eeeeu
eo.¢s.«.~ ee.u eeeu
eo.eeunnu vu.e eeeu
vnaNNqumauieuqel.. ecu .:
ou.¢ecn.~u ue.e eeu
Axussu u¢\- xx
I
X fiddfldfl'dddfldd
eeeu
econ
eeeu
eceu
eeeu
eeeu
eceu
econ
eeeu
eeeu
eceu
eeeu
ecu
eeu
ecu
ecu
eeu
ecu
ecu
ecu
ecu
eeu
eeu
eeu
eu
eu
eu
eu
eu
eu
eu
eu
eu
eu-:
eu
eu
It- a
o
a?
cu
cu
eu
cu
cu
9.
ea
9'
an
cu
or
or
cu
nu
an
on
eu
cu
on
an
3.
cu
eu
cu
cu
cu
:
cu
cu
eu
cu
cu
eu
3
on
3
cu
cu
eu
3
ea
as
on
cu
-eu
2
.eu
a
u
:32
40.~o\
0.n~.n
0.».0
”Or.“
u.o00
d.noo0
0.00.
w.::.:
e.ceeu
c.....
E.eeeu
e.eeeeu
T:
..c~.a
r.co0
0.000.
H.000
:.noo0
0.0.0
n.eo00
e.eeeu
1....0
n.eeeu
..eeeeu
0.~o\
0.90.»
c.co0
u.nno~
u.c00
c.noo0
0.000
"dooaoO
e.eeeu.
w.....
u.eeeu
1.....
”n.ho\
..c~.q
c.no.
“n.0eoc
u.000
.0.nuo.
m.....-
.0.c~00
un.ceeu.
...c...
c.0eeu_c
Mu.ceeeu
0H
e.~u0
o.cemo
n.e00
n.c\u.
0.000
o.uu00
0.000
u.0000
e.eeeu
0.0000
eteeeu.
e.eeeeu
e.~u0
c.0euo
nae..-
n.c~u.
0.000
o.uu0.
0n...
e.n.00
e.eeeu
0.0.00
e.eeeu
e.eeeeu
e.\u0
o.ne~o
n.e00
n.o\u.
0.000
s.uu00
0.000
0.000.
e.eeeu
.o......
c.eeeu
e.eeeeu
e.\u.
c.ce~c
n.e..
c.cuu.
e....
c.uu..
0....
c.c...
c.ueeu
-0 '0
eceeeea:;zw.~oeau.
c.cceu
e.onpbu!azm.0u0efi
n
u.c..
..ccc.
c....
~.uc..
e.eeeu
c.c...
e.eeeu
0.0...
e.eecu
e.eeeeu
e.ceev
e.eeeeu
u.c..
..ccc.
c....
~.uc..
e.eeeu
..c...
e.eeeu
c.....
e.eeeu
c.eeeeu
u.eeeu
u.eeeeu
u.c..
..ccc.
c....
~.uc..
e.eeeu
..c...
e.eeeu
......
m.eeeu
c.eeeeu
u..eeu
c..eeeu
u.c..
.Jccc.
c....
l~.uc.m
e.eeeu
u.c...
..ceeu
..~eeeu
e...eu
0.00mu
u
0....
..en..
...0.
c.c...
e.eeeu
0.0000
e.eeeu
e.eeeeu
eueeeu-
e.eeeeu
e.eeeu
e.eeeou
.4...
e.\»..
0....
c.c...
c.eeeu
c.0000
e.eeeu
e.eeeeu
e.eeeu
e.eeeeu
c.eeeu
c.eeeeu
0.000
e..»..
e....
u.n000
e.eeeu
0.0000
u.eeeu
u.eeeeu
u.neeu
c.0eeeu
o.vmeu
..uceeu
0.000
0.xn00
...0..
..n000
u.eeeu
c.....
c.ceeu
o..~eeu
c...~u
c.c00eu
~.muen
A!
.cmeuccu_
Q.
e.eeeu
u.¢000
e.ceeu
c.0000
e.eoeu
e.eeeeu
e.eeeu
e.eeeeu
enceeu.
e.eeeeu,
«.ooeu
u.eeeeu
e.eeeu
«.0000
e.eeeu
0.0000
0.eoeu
e.eoeeu
e.eeeu
e.ooeau
n.eeeu
0.eeeeu
u.0eeu
c.0uoau
e.eoeu
«.0000
o.coed
c.0000
c.eoou
e.eoeeu
0.neou
0.ceeou
v.0vcd
e.u0eeu
c.00nu
u.0u0:u
e.eeeu
u.e000
e.eeou
0.0000
e.uoefl
o.ceeeu
n.u0eu
n.~00eu
c.09nn
e.co.nu
n.0Nmeu
c.euuun
a.
ue.e
0n.e
c0.»
cu.~
uv.u0
ne.eu
en.0eu
ue.0v
.0uu000
om.emu
00.~00u
so.~en
00.0
0«.u
ov.~u
.~.o
u~.uo
00.00
on.~0n
00.00u
eo.ovcu.
ou.~00
uo.~000
.m.0¢nu
ce.u
v0.»
ev.0n
uv.u.
~u.0u~
nn.oeu
vn.neeu
0e.onv
cc.eonv
ov.~smd
00.0«snd
cc.c~ec
e..-
c..~u
...c~u
uu.uc
uu.ucc
cm.~vn
uc.c~cc
ec.cc.u
s0.movvu
uc.~..v
vu.ows0v
ec.c.c0u
Aucuvu
AcmscuPCOOV
en.e u..e
ee.e 0n.e
ce.e c0.»
ne.e cu.~
mn.e u0.u~
en.0 nc.eu
neun cnqoem-2
n~.~ uo.nv
ec.- 0ecomv-
00.0 o~..nu
ecuc. .c.~.0u
uo.un oo.~en
acne v~.~
ne.e cm.u
~c.e - a cc.~u
.~n.e .~.o
uu.c u...c
ac.» 00.0»
0~.en em.~cn
c~.- cc.ccu
ccuuuc ec.ccmu
Nu... cu...v
00.~00 uo.~s00
~c.euu cm..ccu
mu.e ce.u
~n.e v0.n
cu.c e...»
ou.n uv.u~
ec.un ~u.vu~
u...~ cn.oeu
ec.u.~ cc.nceu
cc.ee~ ce.cmc
on.u0c~ 00.eonv
.e.u~¢ svuwsnu
a..eu.o u..v~.ou
u~.eoc~ co.o~en
ou.n e0.-
uu.» cc.~u
.c.un 00.00u
a...» ue.cc-
nu.cec eu.u.o
m..uv~ om.~0n
u~.0u0~ 05.0009
on.on0u no.000u
outnonuu 00.novou
_ec.c~vc cc.~u.w
~0.~n~00 00.nwe00
u...eecu- u....m.m;,
Aeeeve Axoeu
mu Mummers
-. 'lll'lI." '1‘ l
-
cu..me
or.¢m
00.n0v
00.00.
«c.000u
uc..~0
-mu.c.wn
c.ce.~
uwucovuct
oc.~ocou
n0.cucmcu
n...»c.ru
cu..cw
oc.cn
00.0».
e..0nu
cc.e.0u
c..ca0
00.0.00
ec.c~.~
sc0.cc.~m .
cc.ucc0.
cu.~ccncu
«u.cm.cuu
00.00.
nu.cm
er.nex
cc.em.
0..<00u
vu.0u0
uc....0
00.0«cu
oc..¢neo
sc.u.enc-
m...ecnwu
u..ceucwu
cu..n.
oc.cn
.0.ce0
00.00ul.
cc...0u
cc..~0
cc.cuun
en..c.u
~e.r0¢00
0u.uo~n i
0n.uucsr
u...cccw.iz.epe epu
Axmavu
.c.eeecuceceeeeu eceu ecu
re.eceu ceceeeeu eceu ecu
.e.eeeu ceceeeu. eceu epu
.e.eeu ceceeeu eceu ecu
ce.ecu ceceeu eceu ecu
.e.eu cececu eecu ecu
.e.eu ceeeu eceu ecu
ce.u eeceu eceu eeu
ce.u eucu eceu.eeeu
mu.e c cu eeeu ecu
.u.e- eeu eeeu. ecu
me.e ceu eceu ecu
ce.eceeueeceeeeu ecu ecu
ce.eecu ceeeeeeu ecu ecu
we.ecem ceceeeu ecu ecu
ce.eeu .eeeeeeu ecu ecu
ce.eeu eececu eeu eeu
.e.eu eeeeeu ecu ecu
.e.eu coeeu ecu ecu
ce.u eeeeu eeu eeu
ce.u eecu ecu ecu
:uucq .. eeeu. ecq ecu
cu.e ceu ecu ecu
we.e ceu ecu ecu
.e.eceeucecececu eu ecu
cc.eeeu ceceeeeu eu ecu
ue.eceu ceceeeu eu ecu
ce.ecu ceceeeu eu ecu
ce.ecu ceeeeu eu ecu
ee.eu eeeeeu eu ecu
ce.eu eeeeu eu ecu
ce.u ceceu eu ecu
ce.u eecu e. ecu
uu.e ---eeeu eu ecu
.u.e eeu eu ecu
we.e eeu eu ecu
ce.eeeeueeeeceeu u ecu
ce.eeeu eeceeeeu u eeu
ce.eceu eeeeeeu u ecu
ne.eeui eeeeeeu u. eeu
ce.eeu eeeeeu u ecu
ce-eu . ceeeeu u - ecu
ce.eu eeeeu u ecu
cu.u : ceceu u ecu
ce.u cecu u ecu
mu.e .n:ceeu.:;. u .:-:qeu
taco can u ocd
u ,mwu
ue\ue xx x0 0
133
..u\m
m.\n0
.cvo
\.0u.v
o.\u0
e.\eoc
~.u00
n.:vn0
0.000
c.u000
e.eeeu
0.0000
\.u\~
~.\n0
0.c~o
..0u.v
c.\u0
c.\ev 33:0
..mmm.
mmm.
000. _
annolea
El-A
51
gain differentiator with a l Mohm resistor in the feedback
loop. For example, with Rx = l Mohm and Cp = 1000 pF, the
relative scatter was 0.2%. This is, however, a condition
which has no physical significance because Cp would never
be that high, and especially not for a l Mohm solution
resistance.
3. Dependence 23 Series Cell Capacitance
A dummy cell consisting of a capacitor (0.05 - 10 uF)
and a resistor (l kohm - l Mohm) in series was made to test
the dependence of the conductance measurement on the series
capacitance. It should be noted that the largest capacitor
(10 pF) used is still much smaller than t the typical double
layer capacitance. The results of the experiment are shown
in Table 6.
From Table 6, it is obvious that there is a trend
toward larger errors as Rx and Cx decrease and as t increases
(a completely expected result). For large resistances, the
pulse length is increased to allow the amplifier (A2) to
reach its final voltage before making a measurement. For
the low conductance ranges, the feedback resistor (R2) is
either 100 kohm or 1 Mohm, and 20 usec may not be enough
time to achieve the final voltage. It should be noted that
in these conductance ranges, very little current flows
through the cell, so Cx still does not charge significantly.
If one uses a typical conductance cell in which Cx > 100 pF,
the errors would all be negligible, even for lower resis-
tances. In a situation in which unplatinized electrodes must
be used (such as in some non-aqueous work), the resistance is
52
TABLE 6
Dependence on Cx
10
50
100
50
100
500
kn
kn
kn
kn
kn
ka
kn
k0
Mn
% Error for Cx in uF
10 1 0.5 0.2 0.1 0.05
0.050 0.861 1.930
0.007 0.059 0.091 0.197
0.005 0.046 0.080 0.163 0.423 t = 20 usec
0.001 0.020 0.042 0.103 0.206
0.000 0.003 0.010 0.036 0.085 0.252
0.006 0.024 0.042 0.137
t = 0.2 msec
0.005 0.008 0.028 0.069
0.001 0.004 0.026 0.058 0.122 0.300
0.000 0.001 0.005 0.017 0.042 0.133
53
normally high so that the decrease in Cx is compensated for
by an increase in Rx' Therefore, the error remains small.
In any case, the error involved using this technique is
smaller than the error involved when using a bridge tech-
nique on the same system.
4. 5:9 Bridgg Comparison
The conductance of three solutions was measured using
both this instrument and a modified Wien bridge (65). Below
10 kohm the two methods agreed to better than 0.1%, but at
60 kohm, the bridge could not be nulled and was about 2%
high.
E. Instrument Tests 93 Chemical Systems
1. Acid - Base Titration
As a first test, the chemically uninteresting reaction
between an acid and base was followed conductometrically. A
solution that was 10"3 M in HCl was titrated with 5 M NaOH
using a constant delivery burrette. Delivery was made
through a capillary tube with a 15-foot hydrostatic head to
provide constant pressure. A typical curve is shown in
Figure 9. The experiment was repeated several times, and
in each case. the same shaped curve was obtained. When the
amount of HCl was varied, the endpoint was proportional to
concentration, but the magnitude of the minimum remained
unchanged. From this minimum, one can calculate the ratio
of the equivalent conductances of NaCl to HCl of 0.296. This
compares very favorably with the Kohlrausch ratio of 0.297.
Figure 9
Acid - Base Titration
0.001 M HCl with 5 M NaOH
SS
56
2. Reaction Rate Studies
The pseudo first-order reaction between ethanol and
acetyl chloride was studied at three different temperatures.
The ethanol was dried by the method of Lund and Bjerrum (66),
and the acetyl chloride was used without further purification.
Approximately 3 p1 of acetyl chloride was injected into 125 m1
of ethanol for each experiment. Several experiments were run
in the same solution (using the current offset to keep the
same sensitivity), and no significant deviation was found
between runs. The conductance change was measured using a
Heath Universal Digital Instrument (UDI) in the digital
voltmeter mode (0.1 sec aperture), with a period of 1.500 sec
between measurements. Every fourth measurement was manually
recorded (with another UDI keeping track of counts) so that
a data point was obtained every six seconds. The rate con-
stants were obtained using the three parameter least squares
curve-fit computer program, which is included in the appendix.
This program was written for the PDP 8/i computer, but could
easily be revised to adapt to other systems. It minimizes
the general equation:
E =
i
"MS
1 [Y1 + B'exp(-kXi) -A]2 (7)
where Xi and Yi are the data points, and A, B, and k are the
curve-fit constants (k = rate constant). To guarantee con-
vergence, a modification of the Gauss - Newton successive
approximation method was used. It is similar to the modifi-
cation by Hartley (67), which Daum (68) used in his treatment
of current impulse data. The results are given in Table 7.
TABLE 7
Reaction Rate Studies -- Acetyl Chloride and Ethanol
Temperature k (sec'l)i std. dev. Number of Runs
4.740C‘ 0.03oéio.oooz 7
6.4o°c 0.0346i0.0002 5
15.01°c 0.07ooto.ooo7 15
Typical plots of (n (G. - G)/G. lg; time are shown in
Figure 10. The standard scatter for a typical experiment is
less than 0.1%. The data obtained agree very well with data
obtained by Euranto and Leimu (50), who used a rather com-
plicated, fast-quenching titrametric method. Their data and
these data obtained conductometrically are shown in the
Arrhenius plot in Figure 11.
It should be mentioned that constant temperature was
assumed during the reactions, although this is not strictly
true. The temperature of the bath was constant, but the heat
of reaction, and to some extent, the original higher tempera-
ture of the acetyl chloride, caused the temperature to
increase by about 0.01°C during the reaction. A temperature
profile is shown in Figure 12 for a 5 - 6 times greater
concentration of acetyl chloride than was used to determine
the reaction rate. These data were obtained by using the
bipolar pulse conductance instrument to measure the conduc-
tance of a calibrated thermistor in place of the conductance
cell. This instrument permits a rapid, accurate measurement
Figure 10
Reaction Rate Studies
Acetyl chloride (3 pl) and ethanol (125 ml)
Consecutive reactions at 6.4000
._ 1———
59
938-99 5-
Time (sec)
60
Figure 11
Temperature Dependence of Reaction Rate
Circles obtained conductometrically
Dots obtained by another method
-ln K
61
l05/T
L—
F'
J.I I...Jil.
330 340 350 360
62
Figure 12
Temperature Profile During Reaction
18 pl acetyl chloride into 125 m1 ethanol at 6.420C
63
'6A7r
r I
6.46 '-
6.45 "
T°C
6A4“-
6.43 '"
6.42%
l 1
l 4 1 l 1 j I
I00
200
Time (sec)
300
64
of temperature with extremely little solution heating.
The half-lives of these reactions were between 10 - 20
seconds. Much faster reactions could be studied if appropri-
ate readout and recording were used. Reactions with half-
lives of less than 1 msec could easily be followed using
stopped-flow or other techniques. Since the output is pro-
portional to concentration, the results are very easy to
manipulate.
3. Conductometric EDTA Titration 9f Zinc
Known volumes of 0.04 M Zn(N0 and pH 10 buffer
3’2
(0.55 M NH3 and 0.10 M NHuCl) were pipetted into a beaker
and diluted to 125 ml. The resulting solution was titrated
with a 0.1007 M EDTA solution using a Sargent Automatic
Burrette (10 ml). The conductance cell consisted of two
unplatinized platinum disks 1 cm in diameter and 1 cm apart.
The solution resistance varied between 21 and 35 ohm, depen-
ding on initial concentrations of zinc and buffer, and the
change at the endpoint was only about 0.02 ohm (0.1% of
total). This is truly a worst case situation since the
conductance instrument was not designed to operate below
200 ohm! The noise level was under 0.01% of the total con-
ductance, even at this very high conductivity. The change
in conductance was recorded on a strip chart recorder at
5 in/min with a full-scale sensitivity of 175 pmhos (about
0.5% of the total conductance). The offset current source
was used to "buck out" about 99.5% of the total cell current.
Typical endpoints are shown in Figure 13 for 10 ml of
the zinc solution and varying amounts of buffer. Notice that
f...“ “r.- (:01
.i
65
Figure 13
Conductometric EDTA Titration of Zinc
(a) 10 ml buffer (0.0281 0‘1)
(b) 15 ml buffer (0.0377 n'l)
(c) 20 ml buffer (0.0476 0'1)
Full scale = 175 “0‘1
Arrows indicate endpoint
Sharp breaks due to changing the offset current
INCREASING CON DUCTANCE
I4—Iml —->I _
— INCREASING VOLUME
mi
67
this is not a "typical" conductance titration curve in that
there are actually two breaks near the endpoint. The first
break is taken as the endpoint for the following reasons:
(1) ‘The first break is proportional to the amount of
zinc added (whereas the second break is not.)
(2) There is a sharp conductance increase which
levels off to a more moaerate increase even
when no zinc solution is added (only 15 m1 of
pH 10 bUffer) e
(3) The first break agrees within 0.2% with an
Eriochrome Black T titration on the same system.
While it is not the purpose of this thesis to enter into a
detailed study of equilibria, a few observations of the
behavior near the endpoint should be mentioned. The pH is
constant to within 0.001 pH unit throughout the titration, as
recorded using a Heath eXpanded scale pH meter/recorder. It
is not a slow, rate-controlled process, but rather an equi-
librium situation. The same type of endpoint behavior is
obtained when calcium is used instead of zinc. The difference
between the initial and intermediate (between breaks 1 and 2)
slopes is independent of both the zinc and buffer concentra-
tions. This slope is:
3.2to.2 ver div/hor div x 5 hor div/m1
x 17.5 pmhos/ver div = 270 nmhos/ml of EDTA.
Assume that the conductance is proportional to the
total normality, then G = K'N where K is a constant. Before
the endpoint, the reaction is:
2-
4.
2+ 2 Na , HZY
2 NH3 + Zn 7‘ ZnY2-+2NH1++ + 2 Na+ (8)
yielding an increase in ions of 0.4 meq/ml of EDTA. 1n the
case of 15 m1 of buffer, the conductance is practically
68
constant before the endpoint. Thus, the increase in ions
is almost exactly offset by the dilution with titrant.
Therefore, in this case,
N 0.4 meq/ml
and
-1
_. 000377“ - '1 “1
After the endpoint, the reaction (at pH 10) is
2-
4 Na+, 2 H Y _ _
3 NH3 2 as 3 NH4+ + HY3 + Y4 + 4 Na+ (9)
yielding an increase of 0.7 meq/ml of EDTA. Of this,
0.4 meq/ml is needed to offset dilution and the remaining
0.3 meq/ml increases the conductance of the entire solution.
Thus, the slope after the endpoint should be:
0.3 meg(ml EDTA
129 m1 x 0'1 ml 9.1 = 230 uthS/ml EDTA
which is very nearly the experimental result.
These data tend to point toward the formation of a
complex between EDTA and NH4+ after the EDTA concentration
is built up following the endpoint.
The endpoint is easily detectable, even though it must
be continuously recorded rather than obtained by making a
few measurements and drawing intersecting lines. The result
of seven runs with different volumes of zinc and buffer
solutions is:
2.13210.003 inches/m1 of Zn2+
A volume reading of 6095 on the burrette was found to
be equivalent to 42.31 inches of chart paper. The eXperiment
69
was repeated using 10 ml of Zn2+, 15 ml of pH 10 buffer, and
5 drops of Eriochrome Black T indicator. The result of four
such titrations is:
reading = 310614
blank 2 0029
volume = 307714
To correlate the results:
+ .
31%Zfi1 X Egggégifl = 2.136io.oo3 in/ml
Thus, there is excellent agreement between the two
,1
methods. It should be obvious that since this method is
capable of precisely following very small changes even in
highly conducting solutions, it makes possible conductometric
analysis on other systems (such as redox reactions and ion
exchange monitoring) in whiCh it has previously been im-
possible or very impractical to make conductance measurements.
V. A NEW INSTRUMENT
A. Improvements
The prototype instrument worked so well that it was
decided to build another instrument with many added features.
In the new model, the pulse width is controlled by a 1 MHz
crystal oscillator. This makes the pulse widths much more
stable than the monostable multivibrator pulses used in the
prototype. This, in turn, results in higher accuracy, since
the pulses are much closer to being complementary. In addi-
tion, the pulses are only 10 usec long for the shortest
pulse width. Thus, the error due to polarization will be
only half that of the prototype. The same crystal oscillator
also controls the pulse repetition rate so that experiments
may be carried out on a very accurate time scale.
As an additional improvement, 0.005% precision resis-
tors are used so that a high absolute accuracy may be ob-
tained. The offset current along with the sensitive null
meter on this instrument allow it to make accurate measure-
ments as a self-cantained conductance ”bridge" (with a
linear, calibrated offset). The circuit was completely
re-designed so the solid-state switch resistances no longer
affect the accuracy. It can also operate with four cell
leads so that all lead and contact resistances are
unimportant.
70
71
Finally, an automatic temperature compensation circuit
was added. This allows the temperature of the solution to
vary by as much as 1°C without having a noticeable effect on
the accuracy of the conductance readout. The output is
automatically corrected to the value it would have had if
the temperature of the solution had not changed.
B. The New Circuit
1. Analog Circuit
The block diagram of the new instrument is shown in
Figure 14. The switches are in the positions as shown until
condition A, B, or 0 exists, in which case, the appropriate
~ switch changes state. The conditions occur as follows:
A - during the first pulse, B - during both pulses, and
C - during the sampling time (the last 8/10 of the second
pulse). ,The function of the operational amplifiers is
given below.
A; controls the top electrode of the cell to be
precisely at +5 RZV/lo kS'l during A and -5 RZV/lo k9. during
the second pulse; iggé A-B ("A“ means “not the condition A”
and ”a“ is the logical ”and” symbol).
Ag furnishes an offset current to compensate for the
cell current. R1 is a lOOlohm, 5 decade resistance box
(0.005%). If R1 and R3 are adjusted so that the current
from A2 is exactly opposite from the cell current at the
end of condition B, then a ”balanced-bridge" condition exists.
5} provides an output voltage proportional to the
conductance (with intercept determined by the offset current).
72
Figure 14
The Improved Bipolar Pulse Conductance Instrument
74
During C, A3 controls the bottom electrode to be at virtual
ground and it charges the holding capacitor (Ch) so that
at the end of the second pulse, the voltage across Ch is
proportional to conductance. During 0, Ch is connected to
virtual ground, so that eo remains proportional to the
conductance.
54 keeps the bottom electrode at virtual ground during
0. This further ensures complementary pulses, as was
pointed out as necessary on page 46. A4 must sink the
current supplied by A1 and A2 during C.
55 yields an output which is proportional to the total
conductance (with an intercept of zero). This is used for
temperature compensation.
A6 yields an output which is approximately proportional
to the temperature deviation. RT is a thermistor and RA is
the temperature adjustment. At the start of an experiment,
RA is adjusted to equal RT (ed 2 0). If the temperature
changes, ed will no longer be zero. This voltage and the
output from A5 are multiplied together so that a temperature
correcting current (which is proportional to the product of
the conductance times the temperature deviation) is supplied
to the summing amplifier, A3.
It is important to observe that the electrodes them-
selves are at precisely controlled potentials. The top
electrode is connected to the inverting input of the voltage
follower A1, and the lower electrode is connected to virtual
ground of either A4 (during 0') or A3 (during C). The
outputs of the operational amplifiers control these inputs
75
(gig the feedback loop) to be at the same potential as the
non-inverting inputs (ground potential in the case of
A3 and A4). Thus, both electrodes are controlled to be at
precisely defined potentials regardless of switch, lead,
contact, or cell resistance or inductance.
Amplifiers A1, A3, and A4 are very fast, settling to
Within 0.01% in 1 usec (69), so that rapid_measurements can
be made. The other amplifiers are not as fast, but since
they are not involved with the pulses, speed is not a re-
quirement.. They do have low temperature drifts, so accuracy
may be retained over a reasonable ambient temperature range.
The multiplier has a rated accuracy of 1% and an output of
exey/10V.
2. Digital Circuit
The digital logic circuit is shown in Figure 15. The
output of a 1 MHz crystal oscillator is scaled down to con-
trol both the pulse width (PW) and the pulse repetition
frequency (PRF). The first three decades are used for the
PW, and are gated so that the output of N1 is high during
the last 8/10 of a period (the pulses are each one period
long). 81 and 82 are LATCHED bistable miltivibrators (J-K type),
with the Q outputs going low only by resetting. The first
law transition following resetting causes B1 to change states,
but since the J input of B2 is connected to the Q output of
B1, B2 will not change states until the second pulse. Since
the K inputs are connected to the Q outputs, the Q outputs
cannot return to a low state (until reset). The logic levels
A, B, and C are high during the first pulse, both pulses,
76
Figure 15
Digital Logic Circuit for Improved Instrument
77
lMHz
SCALER RESET-
,‘qO -I-lOO -;-|ooo +l0+ -i-l<35 -E- to6 €- l07
d. d. ct d.
.L____. M:
2.8733 . ____.,
5‘ 7
C
‘1] $5
cxs m
N6 LED‘H ' '
N8 N9 N” '
LL
[ I
J' R a
T EL ._
K Q )Av
Arl
I £2 Q
7' £3
-+K 2 " B
3.45
78
and sampling time, respectively. These signals are con-
verted to +5V/-15V levels in order to drive the switches.
N8 and N9 functknias a gated reset monostable so that
if the input from SS to N8 goes low while logic 8 exists,
the reset cycle will start. The gated reset flip-flop
(consisting of N10 and N11) then causes the sealer, B1, and
B2 to be reset. Notice that a new cycle cannot be started
until the previous cycle is completed. When in the exter-
nal trigger mode, N6 — N11 synchronize the trigger signal to
the PW. Typical waveforms are shown in Figure 16 for the
three trigger modes.
A complete circuit diagram, along with a further
explanation of the components' functions and characteristics,
is presented in the Appendix.
C. Temperature Compensation
Assume that over a narrow temperature range the con-
ductance of a solution and the conductance of a thermistor
vary approximately linearly with temperature; i.e.
' _ 1 _ 1
for the thermistor, GT - fi- _ RU (1 + FTAT) (1)
T T
' = 4L.=
and for the solution, Gx Rx fit (1 + FXAT) (2)
x
where FT and Fx are the temperature coefficients of con-
ductance. From the block diagram in Figure 14,
d- (g1 .. g1) 500kfl
A T
(D
I
01' ed
( o . _ o
—)v FTAT 500 kQ/RT If R _ RT (3)
P.1L’A “ m1 -
_
79
Figure 16
Waveforms of Logic Circuit
Logic Levels in Improved Instrument
Pulse width setting is 10 psec
Mode l 2 Pulse repetition on even decades
Mode 2 = Pulse repetition every 2.8 periods
(dashed lines)
External triggering mode
ll
Mode 3
80
.SggscmsmmaommsememmHommsememmfiosa
81
edet 5v R 5v R2
Also. 90 = m + 3 - TUE-W; R147 (’4')
d 60 5V R1
in et = " R; * item-ii; R5
e e 5V R
_ d t - 2
°r et ‘ [lav—Kg “ram-R;le (5)
Combining Equations 2, 3, and 5:
-5v FTAT 500m et 5v R2 6
et = a; 10V R6 " m (“PX”) R5 ( ’
In order to have correct temperature compensation,
et=e.c = W 2 (7)
X
Therefore, in order to have correct compensation, from
Equations 6 and 7,
o 'I
5V FTébT SOOKQ et + 5V R2 bxiT = 0 (8)
RT 10V R6 10kfl Rx
Combining Equations 7 and 8 again yields:
2R°R F
R =
5 500 kQ.FT
Since R6 = 5 kfl,
0
RF
ReTX (10)
5 3U'F;
Therefore, from Equation 10 it is obvious that temperature
compensation is possible providing that either the tempera-
ture coefficients are constant or that they vary in exactly
82
the same way. They should be nearly constant over a 1°C
temperature range (the range for which this instrument was
designed). The temperature adjustment is made by varying
RA (a 10 turn, 100 kflpot) until e is zero. At this point,
d
RA = R;, so if the ratio of the temperature coefficients is
known, R5 (a 10 turn, 2 k0 pot) can be calculated directly.
Armitage and French (70), however, have shown that it is
very difficult to make an a priori temperature—coefficient
prediction. In practice, therefore, R5 will be experi-
mentally set by varying the temperature of the solution and
adjusting R5 until eo = eg.
D. Obtaining Absolute Conductance
If proper compensation is made, the combination of
Equations 2, 4, and 4 yields:
5V R4 10 82 R1
eo=ioomT-§; (11)
X
Equation 11 is the main equation for this instrument, as it
allows the accurate determination of conductance. The
voltage eo is available at the front panel and also to the
null meter. It is evident that if eO = 0, then:
3156 - Tar-n— <12)
X
Therefore, this instrument can be used as a conductance
“bridge“ with no capacitance balancing. If the offset cur-
rent is not used, the output, e0, is a linear function of
conductance with a zero intercept. If R3 f-, the intercept
will change and a "bridge“ with a linear off-balance is
'73
F- '
I
1 6. .‘II’I- ‘1 :
83
obtained. Equation 11 can be written:
100k0 e° R
T‘SV‘Q+R'1'
4 3
1 1 1
fit we; (13)
X
Equation 13 suggests an extremely simple readout. Define
five dimensionless variables:
a = R /10 k0 (read one decade to the left of
1 decimal point)
b = 100 RSI/R3 = 100, 10, 1, 0 for R3 = 1, 10, 100;" kg
c = 10 kQ/RZ = 100, 10. 1
for R2 = 0.1. 1, 10 k9
d = 200 kQ/Ru = 100, 10, 1
for R4 = 2, 20, 200 k0.
e = eo/IV (read output in volts)
Equation 13 then reduces to:
%a = (ab + de)c “9‘1 (14)
X
E. Instrument Tests
1. Absolute Accuracy
The absolute accuracy of the instrument was tested over
its entire conductance range, using precision resistors in
place of the cell. The results are summarized in Table 8.
There appears to be a systematic error that leads to
a measured conductance which is approximately 0.02% low. The
tolerance of the input resistor of A2 is temporarily only
0.01% (a 0.005% resistor will replace this resistor as soon
as it arrives). The rest of the error can be attributed to
the lack of optimization of the circuit. In the present
work, for example, the amplifiers are over-compensated with
1
‘4.'l:'.
.
L
84
TABLE 8
Absolute Accuracy Tests*
Resistance
Measured G
Rx(ohm) Tolerance (%) (p971) % Error
10 0.05 99,906 -0.094
50 0.01 19.998 -0.010
100 0.01 9.995-5 —0.045
500 0.01 1.999.“ -0.030
1 k 0.01 999-7 -0.030
5 k 0.01 199.96 -0.020
10 k 0.01 99.98 -0.020
50 k 0.01 19.991 -0.045
100 k 0.01 9.9980 -0.020
500 k 1 2.0057 +0.285
1 M 1 1.0038 +0.380
5 M 1 0.1990 -0.500
10 M 1 0.0990 -l.000
30 M 1 0.0332 -0.400
I:
Unless stated otherwise, all measurements were made with
a pulse width of 1000 usec and with the smallest "c"
possible.
85
respect to high-frequency oscillation. This inhibits their
response times and can cause error (especially for high R4
and short pulse widths). There is also the switching charge
(the switching signal is coupled through the FETS’capacitance
to Ch) which has not been fully compensated for in the sample
and hold portion of the circuit. These problems are solvable,
but their solution will take somewhat more time. Until the
circuit is optimized, 0.05% absolute accuracy is all that can n“
be Claimed e
2. Sensitivity
The sensitivity of this instrument was very good over
its entire range. The limit of detection is shown in
Table 91(limit is 5 0.1 scale division).
TABLE 9
Sensitivity Kg. Resistance
Resistance “1) Detection of 1 part in
10 60,000
100 200,000
1 R 1,000,000
10 k 100,000
100 R 20,000
1 M 8,000
10 M 1,000
The detection below 1kg is limited by noise (the 50 mv
and 0.5V pulses have relatively more noise than the 5V pulse).
Above lld2the detection limit is due to the closed-loop gain
of A3.
86
3. Dependence pg the Parallel Cell Cgpacitance
The dependence on the parallel cell capacitance is
shown in Table 10. These data were obtained using a resis-
tance substitution box for RX and placing capacitors across
the terminals.
From Table 10 it is obvious that the output is quite
independent of Cp. There is some error as Cp increases,
especially for large Rx'
4. Dependence pg Series Cell Capacitance
A series combination of a resistor and capacitor was
used to determine the influence of the double-layer capaci-
tance using this instrument. An ESI 0.01% decade resistance
box was used for RX. The results of this test are given in
Table 11.
From these data, it can be seen that the dependence on
Cx is not large for small t/RXCX. It should be pointed out
that the errors for small t/RxCx are probably due to the
error in the measurement of Rx (see Table 8) rather than the
error caused by Cx' Again, it should be pointed out that
the largest capacitor tested is still much smaller than the
typical Cx of real cells.
5. Egg pf Four Cell Leads
The use of four cell leads to eliminate the influence
of contact and lead resistance on the conductance measurement
was explored. The conductance of two resistors was measured
in three configurations: (1) with two leads, (2) with four
normal leads, and (3) with a one ohm resistor in series with
each lead. The results are listed in Table 12.
87
TABLE 10
Dependence on C
"a” term for balance with Cp in pF
Rx(0¢10%) 0 100 570 1070
1 k 9.0324 9.0321 9.0322 9.0324
2.2 k 4.3352 4.3352 4.3354 4.3358
4.7 k 2.0529 2.0528 2.0528 2.0529
10 k 1.0253 1.0253 1.0254 1.0259
22 k 4.1332 4.1334 4.1335 4.1341
47 k 2.0137 2.0118 2.0097 2.0110
100 x 1.0370 1.0391 1.0380 1.0398
220 k 4.8760 4.8770 4.8784 4.8811
470 k 2.0280 2.0250 2.0272 2.0299
1 M 1.1071 1.1036 1.1064 1.1085
2.2 M 0.4330 0.4336 0.4350 0.4380
4.7 M 0.2290 0.2299 0.2310 0.2340
10 M 0.1089 0.1099 0.1110 0.1133
88
TABLE 11
Dependence on Series Capacitance
measured G for Cx
Theoretical C
Rx (nflrl) l uF’ 10 pF .1
10 100,000 86,100 99,070
50 20,000 19,220 19,996
100 10,000 9,906 9.999 t = 10 usec
500 2,000 1,998.3 1.999.8
1000 1,000 999.8 999.91
5000 200 199.96 199.92
50 20,000 10,100 19,187
100 10,000 8,200 9,910
500 2,000 1,918.0 1,998.3 t = 100 psec
1000 1,000 990.1 999.8
5000 200 199.82 199.96
89
TABLE 12
Elimination of Contact and Lead Resistance
Measured G (ma-1)
Rx Tolerance of Rx 2 Leads 4 Leads 4 Leads + 29
10 0.05 99.904 99.907 99.9
50 0.01 19.998 19.999 20.0
Although the insertion of resistors in series with the
leads reduced the sensitivity (because of noise), the fact
that an accurate measurement can be made with a lead resis-
tance that is 2/10 the value of Rx proves the value of the
four-lead technique. It should be mentioned that long leads
should be avoided unless proper compensation is provided to
Al, A3, and A4; otherwise they are unstable.
6. A;Q Bridge Comparison
The conductance of two solutions of K01 and of some
conductivity water was measured using the bipolar pulse
instrument and a Wayne - Kerr Bridge (Model # B 221). An
experienced operator obtained the bridge measurements. The
results are given in Table 13.
TABLE 13
A-C Bridge Comparison
Bipolar Pulse Technique A-C Bridge Techniqpe
2.366 1.04 (PW = 1000) 2.36 110'1
-1 _
11.295 n0. (PW = 1000) 11.30 p9 1
2.8000 ma'l (PW 100) 2.795 mn'l
90
7. Temperature Compensation
A variable temperature bath was set up to test the
automatic temperature compensation. The temperature was
varied by adjusting the ratio of hot to cold water flowing
into the bath. The conductance of a KCl solution was followed
Kg. temperature with and without compensation. Without
changing the temperature adjustment or the T.C. adjustment,
more KCl was then added to the solution, and the experiment
was repeated. The results are shown in Table 14.
From Table 14, it is obvious that temperature compen-
sation by this method does work. The accuracy is increased
by 1 - 2 significant figures. It should be noted that the
zero temperature adjustment is slightly inaccurate in these
examples. The temperature of the bath was easy to vary, but
was very difficult to maintain at a constant temperature.
Therefore, all adjustments (balance of A5 and the multiplier,
adjustment of RA' and the reading of the conductance with and
without compensation) were not made with AT = 0. This would
not be a problem in a realistic situation, since the tempera-
ture would be constant. Although compensation was excellent
for negative temperature deviations, it was not very good
above a positive deviation of 0.2500. Most likely the multi-
plier was not properly adjusted to obtain the best linearity
for a large negative input signal (especially since et is a
small negative signal).
It is important to observe that the conductance and the
temperature can both vary once proper adjustments are made.
This is very important when following a reaction or a titra-
tion conductometrically.
Ani‘a‘fi .
it
91
TABLE 14
Temperature Compensation in KCl Solutions
Measured G (pa-1)(PW : 100 usec)
thjoC) With Comp. Without Comp.
-0.90 4,851 4,680
—0.75 4,853 4,704
-0.63 4.853 4.726
-o.50 4.853 4.751
-0.25 4,854 4.802 T = 14.5°c
0.00 4,852 4,847
0.25 4,853 4,896
0.50 4,867 4,940
0.60 4,887 4,988
-1,25 5,492 5.195
-1.00 5,491 5,271
-0.75 5.49“ 5.323
-0.50 5,496 5,376
-0.25 5.493 5.438
-0.12 5,493 5,456 Temp. adj. and
0.00 5,492 5,484 $02.5288: remain
7.25 5.993 5.540 More KCl added.
0.50 5.503 5.590
0.75 5.525 5.646
0.90 5,542 5,680
VI. CONCLUSION
The bipolar pulse technique makes many experimental
parameters non-critical or at least not as critical as when
using a-c or d-c techniques. The method is unique in that
it combines all of the desirable features of high and low
frequency a-c techniques while making the actual measurement
using a d-c potential. Most of the limitations of the tradi-
tional techniques are eliminated or reduced to a tolerable
level. The most important advantage of this technique is
its relative independence of all cell capacitances.
Since this technique is independent of Cp, accurate
conductance measurements can be made in solutions of low
conductance. The cell design and lead length are no longer
critical. Since the Parker effect is eliminated, oil baths
are not needed. This techniQue should prove extremely useful
when working with non-aqueous solutions.
Since this technique is not strongly dependent on Cx’
many experimental difficulties are eliminated. Platinization
is not necessary. Therefore, many new chemical systems, in
which surface adsorption on platinized platinum is a problem,
may be studied conductometrically. It is also possible to
use small electrodes. This makes possible conductometric
determinations using small cells (such as would be used for
a conductivity detector in ion-exchange). This also has
92
93
application in the bio-medical field, such as in the study
of Na+/K+ exchange in nerves and blood. The lack of dependence
on Cx also permits accurate measurements on very high con-
ductance solutions (such as molten salts).
Independence of the cell capacitances has several
other advantages. The time-consuming bridge balance is not
necessary (bridge balance also requires a pure signal with
no harmonics, as the capacitive reactances are only balanced
at one frequency). This permits very rapid conductance
measurements. Rapid measurements with small electrodes are
very useful for fast kinetic studies (such as stopped-flow).
The bipolar pulse technique is also readily applicable to
other flow systems which require continuous monitoring, such
as ion—exchange or liquid-liquid chromatography. It may even
be possible to detect non-conducting liquids by injecting a
continuous flow of ions into the effluent stream. As the
non-conducting liquid comes off the column, it may change
the dielectric constant of the effluent and/or the solvation
spheres of the ions enough to detect a change in conductivity
(recall a sensitivity of up to 1 ppm).
The circuit of the bipolar pulse instrument permits
compensation for various experimental parameters. The tem-
perature compensation circuit of the new instrument is just
one example. Another example is the application of a current
ramp to the summing point of the output amplifier in order to
compensate for a linearly increasing (or decreasing) conduc-
tance. This was done using the prototype instrument to
follow the change in slope during a titration. This
94
technique may allow conductance to be used to follow titra-
tions in which the number of ions is unchanged, but in
which the ions undergo a change in salvation number (or in
which ions of similar, but not equal, equivalent conductance
are exchanged).
In summary, the bipolar pulse technique can be used
under any conditions that traditional techniques can be used
(including the accurate determination of cell constants and
equivalent conductance). In addition, it can be used in
many circumstances that the traditional techniques cannot be
used. This technique is extremely simple to use, as it
requires no balancing, and the output is directly proportional
to the conductance. In addition, very rapid measurements may
be made. Thus, this technique should prove to be very useful
for the accurate determination of conductance under a wide
variety of experimental conditions.
REFER ENC ES
REFERENCES
l. Newberry. E. J. Chem. Soc. 113. 701 (1918).
2. Eastman, E. J. Am. Chem. Soc. 3;. 1648 (1920).
3. Kraus. C. and Fuoss. R. J. Am. Chem. Soc. 55, 21
(1933).
4. Gunning, H. and Gordon. A. J. Chem. Phys. ‘19. 126
(1942).
5. Elias, L. and Schiff. H. J. Phys. Chem. 59. 595
(1956).
6. Gavis, J. J. Chem. Ppys. 41. 3787 (1964).
7. Jaffe. G. and LeMay, C. J. Chem. Phys. 21, 920 (1952).
8. Douwes. C. and van der Waarder, M. J. Inst. Petrol.
9. Kohlrausch. F. Wied. Ann. 49, 249 (1893).
10. Kohlrausch. F. and Holborn. L. Das Leitvermdgeg Der
Elektrolyte Insbersondere Der Lbsungen. Leipzig:
B. G. Teubner, 1898.
ll. Washburn. E. and Bell. J. J. Am. Chem. Soc. 35. 177
(1913).
12. Washburn, E. J. Am. Chem. Soc. 38. 2431 (1916).
13. Hall. R. and Adams, L. J. Am. Chem. Soc. 41. 1515
(1919).
14. Gieringer. C. Rev. Sci. Instr. Z. 414 (1936).
15. Jones, G. and Josephs, R. J. Am. Chem. Soc. 50. 1049
(1948).
16. Taylor, W. and Acree, S. J. Am. Chem. Soc. 38. 2403
(1916).
17. Jones, C. and Bollinger, C. J. Am. Chem. Soc. 5;.
2407 (1929).
95
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30°
31.
32.
33-
34.
35.
36.
37.
38.
39.
96
Jones. C. and Bollinger, G. J. Am. Chem. Soc. 53.
411 (1931).
Parker. H. J. Am. Chem. Soc. 45. 1366. 2017 (1923).
Randall. M. and Scott, G. J. Am. Chem. Soc. 49. 636
(1927).
Smith, F. J. Am. Chem. Soc. 42. 2167 (1927).
Shedlovsky. T. J. Am. Chem. Soc. 53. 1793 (1930).
Roy. W. and Martell. A. Rev. Sci. Instr. 19, 628
(1948).
Cole. R. and Gross. P. Rev. Sci. Instr. 29. 252 (1949).
Hladek. L. Chem. Listy 69. 238 (1966).
Warburg. E. Wied. Ann. PhXSik. éZ. 493 (1899).
Winterhager, H. and Werner. L. Forschungsber. des
Witschafts. No. 341 (1956).
Feates. F.. Ives, D.. and Pryor, J. J. Electrochem.
SOC. 10:. 580 (1956).
Metcalf. W. J. SCi. InStr. fig. 742 (1965).
Robbins. G. and Braunstein, J. J. Electrochem. Soc.
116. 1218 (1969).
Bertram. R. and Cruse. K. Ber. Bunsenges. Physik. Chem.
92. 98 (1963).
Bertram, R. and Cruse, K. 2. Anal. Chem. 86, 213
(1962).
Cruse. K. and Bahr. E. Z. Anal. Chem. 192. 141 (1963).
Wershawé R. and Goldberg. M. Anal. Chem. 31. 1180
19 5).
Warshawsky. 1. Rev. Sci. Instr. g5. 711 (1955).
Griffiths. V. Talanta. g, 230 (1959).
Griffiths. V. ApalLChim. Acta. l8. 1?“ (1958).
Ferris. C. Rev. Sci. Instr. 33. 109 (1963).
Hanss M. and Guermonprez R. J. Chim. Phys. 6 . 663
(i966). ’ ‘1
40.
41.
42.
43.
an.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
97
Janz. G. and McIntyre, J. J. Electrochem. Soc. 1 8,
272 (1961). ”-
Walisch. W. and Barthel. J. Z. Physik. Chem. 4, 38
(1962).
Walisch. W. and Barthel, J. Z. Physik. Chem. 32. 235
(1963).
Schmidt. K. RevL Sci. Instr. 32. 671 (1966).
Robertson, R. Can. J. Chem. 33, 1536 (1955).
Robertson. R. and Hyne. J. Can. J. Chem. 35. 623 (1957).
Murr. B. and Shiner. V. J. Am. Chem. Soc. 84, 4672
(1962).
Sirs, J.w Trans. Faraday Soc. 4, 201 (1958).
Prince, R. Trans. Faraday Soc. 54. 838 (1958).
Strehlow, H. and Wendt, H. Inorg. Chem. g. 6 (1963).
Eurato. E. and Leimu, R. Acta. Chem. Scand. 29.
2029 (1966).
Dusek. K. and Lesek. J. Chem. Listy. 52. 1353 (1965).
Tregloam, P. and Laurence, G. J. Sci. Instr. fig. 869
(1962).
Mori. I. and Chikui. S. J. Chrom. 29, 627 (1965).
Duhne. C. and Sanchez, 0. Anal. Chem. 53. 1074 (1962).
Pecsoc. R. and Saunders, D. Anai. Chem. 39. 1756 (1968).
Pecsoc. R. and Saunders, D. Anal. Chem. 49, 44 (1968).
Knudson. G., Ramaley, L. and Holcombe, W. Chem. Instr.
1. 325 (1969).
Gerischer. R. Z. ElektrochemA 58. 9 (1954).
Cruse. K. Z. Chem. 5. 1 (1965).
Barthel. J. Angew. Chem. Int. Ed. 2, 260 (1968).
Fuoss. R. and Accascina, F. Electrolytic Conductance.
New York: Interscience Publishers, Inc.. 1959.
Daviesézc. Ion Association. London: Butterworths,
19 .
11' I! i l 1‘ ‘
.II III . Ill!
63.
64.
65.
66.
67.
68.
69.
70.
98
Monk. C. Electrolytic Dissociation. New York: Aca-
demic Press. 1961._
Nancollas. G. Interactions in Electrolyte Solutions.
New York: Elsevier Publishing Co.. 1966.
Thom€soné)H. and Rogers, M. Rev. Sci. Instr. 32, 1079
195 . .
Lund, H. and Bjerrum, J. Ber. B64 210 (1931).
Hartley. H. Technometrics 3, 269 (1961).
Daum. P. Ph. D. Thesis, Mich. State Univ.. East Lansing,
Michigan (1969).
Analog Devices. §_Selection Handbook and Catalog Guide
59 Operational AmElifiers. (1969).
Armitage, P. and French. C.) J. Chem. Soc. 1963, 743.
A PPENDI CES
APPENDIX A
PROGRAM FOR CALCULATING PSEUDO FIRST-ORDER RATE CONSTANTS
The general equation for the pseudo first-order rate
constant is:
C“ - Gt
0
where C is the concentration at time 0. t, and... and k is
the rate constant.
If the conductance is proportional to concentration,
Equation 1 becomes:
0‘ - Gt
m— : exp(-kt)
‘ O
or Gt + (G —Go)exp(-kt) - G. = 0 (2)
A computer program was written to minimize the equation:
n 2
E = ‘Z [Y. + B’exp(-kX.) - A] (3)
. 1 1
1:1
where Yi = the ith conductance reading
X.1 = the 1th time reading
A = G
O
B =Go-Go
Equation 3 is differentiated with respect to each curve-fit
variable. A. B. and k. The partial derivitive in each case
is set equal to zero and a value for AA. AB, or Ak is
99
100
calculated:
AA = n (4)
[Y1 + B'exp(-kXi) - A]
gflflfl
1
OE eXp(-2kXi)
AB - n 1‘1 f (5)
i§1 eXp(-kXi)‘[Yi + B-exp(-kXi)-—A]
2
igi Xi exp(-kXi)[Yi + 2B-exp(-kxi)-A]
Ak ‘
n (6)
iii Xi exp(-kXi)[Yi + B'exp(-kXi)-A]
Since A. B. and k are not independent variables. the
whole difference cannot be used as it can with the simple
Newton successive approximation. -It is assumed that the
least squares error. E, as calculated from Equation 3 is a
parabolic function of the fraction, v, defined by the
equations:
Al+1 = A5 + vAA (the (j+1)th iterative value of A) (7)
‘ (v is a fraction between 0 and 1)
B3+1 = Bj + vAB (8)
kj+1 = kj + vAk (9)
The optimum v occurs at the minimum of the parabola and is:
(4 E8 - 3 E0 - E1)
" - 2E
Op 4 (E1 + E0 %)
where E0 = E when v = 0
E% = E when v = %
E = E when v = 1
I IIIIII.‘ (1‘1:
101
The new values of A, B, and k are then calculated from
Equations 7 - 9 with v = vop; As soon as the values do not
change by more than the quitting tolerance, the iterative
procedure halts.
Table A-1 shows the three-parameter, least-squares,
curve-fit program which was written to obtain the results
listed in Table 7. It was written for a PDP 8/i computer,
but could easily be modified for other computers.
As written, there are three independent options for
operation, depending on the code that is typed:
odd code
type log plot using final values
code < 5
print values after each iteration
2 < code < 7 = pause after each set of data
Thus. the input data necessary is:
(1) the experiment name or number
(2) the number of seconds in each time
interval first line
(3) the infinite conductance
(4) the iteration stopping tolerance (%)
(5) time
succeeding lines
(6) conductance
(7) time = -l.0 end of data
Typical input and output is shown in Table A-2.
THREE-PARAMETER. LEAST-SQUARES. CURVE-FIT PROGRAM
102
TABLE A-l
C THIS PROGRAM CALC
C PROPERTY WHICH I
C
108
21
22
101
12
103
13
15
S PROPORTIONAL T0 CONCENTRATION.
A J-PARAMETER LEAST SQUARES CURVEFIT IS DONE.
DIMENSION X(25).Y(25).Q(3)
READ (1.108) 111
FDRHAT ('CODE : '12)
K.:1
IF (K1)21.21.22
K4:0
R3:(K1-3)t(R1-6)
K2:R1-2.(Kl/2)
K1:K1-5
READ (2.101)ALPH.TH.FL.TOL
FORMAT (A6.2X3F8.3)
TOL:(.01#TOL>#*2
1:0
I:I+I
READ (2.103)X(I).Y(I)
FORMAT (2F8.3)
X(I):X(1).TH
IF (X(I)) 13.12.12
Rzl-l
FN:N
VX:0
IF (FL)3.4.4
1:0
X2:(X(R)-X(1))/2.
1:1+1
IF (X2-X(I))6.6.5
12:1-1
Y2:Y(I)-(Y(I)-Y(12)).(X(I)-x2)/(X(1)-X(12))
FL:Y2+(Y2-Y(1)):(Y(N)-Y2)/(2.tY2-Y(1)-Y(N))
G=ALOG((FL-Y(1))/(FL-Y(N)))/(X(N)-X(1))
8:(FL-Y(1)).EXP(G*X(1))
81:0 7
52:0
53:0
54:0
85:0
C SUMHING ROUTINE
DO 50 1:1.N
T:X(1)
A:EXP(-Gtt)
R=Y(I)+BtA-FL
$1=SI+N
ULATES A FIRST ORDER RATE CONSTANT FOR ANY
50
32
200
53
103
TABLE A-1 (continued)
SZ:SZ+A¢R
$5:SS+A..2
Sa:84+TttZ.A.(R+A.B)
SS:SS+TtAtN
IF (K1)32.33.33
HRITE (1.200) FL.G.B.VX
FORMAT (3E12.6.F5.3)
DL=.5t51/Fl
DB:-.5¢52/53
ox:.5:ss/sa
C CONVERGENCE GUARANTEE
39
40
3A
35
56
26
59
106
60
101
61
TEST
DO 40 J:1.3
55:0
DO 59 13'."
SS:SS+(Y(I)+B.EXP(-GtX(I))-FL)..2
0(J):SS
B:B+D8
G:G+DN
FL:FL+DL
V:-.S*(4.t0(2)-3.*0(1)-O(3))/(Q(3)+O(1)-2.t0(2))-3.
VX=(V+3.)/2o
G:G+V#Dx
FL:FL+DLtV¥tK4
B:B+V.D8
FOR CONVERGENCE
IF ((DK/G)..2-TOL)34.15.15
IF ((DL/FL)t#2-TOL)35.15.15
IF ((DB/B)..2-IOL)36.15.15
55:0
DO 26 I:1.N
SS:SS+(Y(I)+B¢EXP(-G*X(I))-FL)#:2
SC:IO0.t(SS/(Fl-l.))tt.5/(Y(N)-Y(1))
URIIE (1.100) ALPR.G.FL.SC
FORMAT (A6.' N: ' E12.6.' INF G = ' F10.6.' SCAT = 'FT.3.° 2')
IF (N2)61.51.59
WRITE (1.106)
FORMAT (' TIME LN(GI-G)')
DO 60 I=I.N
FN:ALOG(FL-Y(I))
WRITE (1.107)X(I).FN
FORMAT (2(1XF10.A))
IF (N3)l.|.62
PAUSE
GO TO 1
END
A1.1
K:
1
T
INPUT AND OUTPUT OF LEAST-SQUARES PROGRAM
INFIN
.707209E-01 INFIN G
IME
12.0000
18.0000
24.0000
30.0000
36.0000
42.0000
48.0000
54.0000
60.0000
66.0000
72.0000
78.0000
84.0000
104
TABLE A-2
Al.1 2.0
6. 4150.
9. 5297.
12. 6041.
15. 6529.
18. 6847.
21. 7056.
24. 7194.
27. 7282.
30. 7339.
33. 7377.
36. 7403.
39. 7420.
42. 7430.
“lo 0.
I LN(GI-G)
-.8150
-
.—
-1.2419
”1.6657
-2.0905
‘2.5135
-209383
“3.3683
‘3.7878
-4.1996
-4.6146
-5 . OLI'8L“
’50n87l
”5.8785
7457.
7463.2 SCAT r
0.01
.052%
1.111 ‘I I111 11‘. 1|] Ill!‘
11"}! '11} I'll
APPENDIX B
CIRCUIT DESCRIPTION OF
IMPROVED BIPOLAR PULSE CONDUCTANCE INSTRUMENT
The circuit diagram of the analog portion is
shown in Figure 8-1. A description of each critical com-
ponent is given below.
The -5V reference supply is an integrated circuit
voltage regulator (LM 304, made by National Semiconductor).
Its rated accuracy is 0.01%.
The resistors from -5V to ground and to A2 have
tolerances of 0.005%. as do the resistors from A2 and A3
to the summing point (bottom electrode of cell). The
feedback resistance of A2 is a 5 decade. 100 kCZresis-
tance sub-assembly with a rated accuracy of 0.005%. All
of these are metal film resistors and have temperature
coefficients of l ppm/Co. These resistors have very low
capacitances, having rise-times of less than 1 nsec.
They are the only resistors involved with the actual
pulses.
A3 and A4 are very fast (to 0.01% in 1 usec)
differential amplifiers (Analog 1488). Their common-mode
rejection ratio (CMRR) is not very high. but since the
non-inverting input is grounded. the CMRR is unimportant.
105
106
8
0
too
10
I
DO
'\..
I
.OIAF
Figure B-1
CIRCUIT DIAGRAM OF THE IMPROVED CONDUCTANCE INSTRUMENT
(A11 resistances in kohm)
107
They have low input currents (necessary for sample and hold)
and up to 20 mA of output current. A4 must sink all of the
current coming to the summing point during 5.
A1 (1498) has similar specifications to A3 and A4. but
it also has a high CMRR (15.000). This is necessary since
it functions as a voltage follower.
A2. A5. and A6 (119K) are low drift (5 uV/Co) ampli-
fiers. They have relatively high input currents (35 nA),
but they are not used in any portion of the circuit that
would be affected by this current. A7 (1428) is used simply
as an inverter and is not at all critical.
The multiplier (Hybrid Systems 1070) has a rated
accuracy of 1% over four quadrants and has an output of
exey/IOV.
The 100 k9. pot (10 turn) to A6 has a linearity of 0.1%.
There is a 10 kfll% resistor in series with it. and the dial
reads from 10 kIZto 110 k9. This potentiometer is used to
adjust for temperature during temperature compensation. RT
is a thermistor (YSI, 1%) with a resistance of 30 Id? at 25°C.
Without adding any more resistance to the 100 kfl pot. this
thermistor permits temperature balance from -4 to +520C.
The feedback resistance of A5 is a 2 RR. 10 turn pot
(0.1% linearity). It is used to adjust the temperature
coefficient (see Equation 10 on page 81). The feedback re—
sistance of A6 and the output resistance (5 k9) of the multi-
plier have a 1% tolerance. A11 previously undefined resis-
tors are 0.1%. The 10.2V zener diodes across A6 prevent it
from limiting when the temperature is unbalanced.
108
Switch 1 is a non-critical three-position switch (it
may have high contact resistance since no current flows
through it). Switches 2 and 3 are high-quality switches
with silver contacts since they must have resistances which
are negligible with respect to l kfl.(S3) or 2 kg (82).
The transistor switches (a~1) are driven by the output
of a Fairchild 9624 interface gate (which converts a + 5V
input to -15V, and a CV input to +5V as used in this instru-
ment). All six of the N-channel junction FETs are T1873
transistors." A gate signal of -3V turns them ”off". The
P-channel. enhancement mode MOS-FETs (a, c, f, g, i. and k)
do not turn "on" until the gate signal is -6V. Since all of
the gates in each set of transistors are driven by the same
signal, a "break-before-make", single. double, or triple
pole, double-throw switch is obtained when driven by "A",
"B", or ”C“. respectively.- During condition A (see page 75),
"A" is -15V, during condition 8. "B“ is at +5V, and during
condition C. ”C" is at -15v,. The choice of logic level to
use is explained as follows.
Transistor “a" is further ”on" if connected to +5V
rather than -5V. and A1 has to be connected to the minus
voltage during the second pulse. Therefore. “A" must be -15V.
The N-channel FETs have a much lower "on" resistance
than do the P-channel FETS. It is desirable. therefore, to
have transistors d and e "on" during condition 8. Thus,
"B" = +5V.
Amplifier A4 must sink considerable current (up to
10 mA) during 5. Therefore. the P-channel transistors are
109
”on" during 0 (they have an 8 mA rating).
Notice that the “on“ resistances of transistors a. b,
c. e. i. j. and k are not critical since no current flows to
the inputs of the operational amplifiers. As long as the
amplifiers are functioning, the electrodes themselves are the
controlled points. since they are connected to the inverting
inputs of Al and eitherA3 (during C) or A4 (during 0) by
transistors e. i. and j, respectively.
The 0.01 nF holding capacitor is connected to ground
during C and to virtual ground during C. Thus. eO remains
(during C) at the voltage that is proportional to the con-
ductance.
et is proportional to the total conductance (with zero
intercept). This value is multiplied by the temperature~
deviation in order to feed a temperature-correcting signal
to A3 that is_proportional to the product CAT.
It should be mentioned that separate signal and power
grounds (connected at a common point) were used. Also.
separate power supplies were used for the amplifiers involved
with the pulses and those working at d-c. The analog portion
is covered with a metal shield and is completely separate
from the +5V digital logic supply.
The digital portion of this instrument consists of:
(1) a Heath EU 800 27 crystal oscillator card
(2) a Heath EU 800 28 7adecade scaler card
(3) a card consisting of:
(a) two SN7400N quad. 2-input NAND gates
(b) one SN7410N triple, 3-input NAND gate
(c) a SN7476N Dual J-K Flip-flop
(d) the interface gates (switch drivers)
110
A picture of the front panel is shown below.
Figure B-2
The Bipolar Pulse Instrument Front Panel
The instrument panel has been arranged for the convenient
measurement of conductance. The pulse width should be adjus-
ted for the conductivity of the solution with a rough guide
being:
PW = 10 usec for G > 1 m971
PW = 100 usec for 1 ma"1 > G > 0.1 m9'1
PW = 1000 usec for _ G < 0.1 m9'1
Normally the pulse repetition frequency is at least ten times
the PW.
The maximum cell current is 5 mA. The "c" switch
(lower right) controls the pulse amplitude to be consistent
with this current.
111
c = 100 (v = 50 mV) for 10 m RX < 1000
c = 10 (v = 500 mV) for 100 11 < Rx < 1 1:11
C = l (v = 5V) for Rx > 1 k9
The lower the pulse voltage. the higher will be the
relative noise-level. so that unless Cx is small (so that ZF
might be important), the highest voltage that provides under
5 mA should be used. One should begin with c = d = 100 and
decrease c until "e“ is between one and ten volts.
Then ”a" (the feedback resistor of A2) and "b" (the
offset multiplier) are adjusted until e = 0. The output
amplifier (A3) gain may then be increased (d decreased), and
"a" readjusted. The meter may also be made more sensitive
in order to achieve very sensitive balance.
.er
I1“
"
H
H
H
II,
Mm
II
I
3 03047 0631
1
3 129
11111111111111