w.

 

Date

0—7639

July 18,

LIBRARY

Michigan State
University

 

This is to certify that the

thesis entitled

Physicochemical studies of

magnesium complexes

presented by

Pierre-Henri C. HEUBEL

has been accepted towards fulfillment
of the requirements for

Ph. D. degree in _§hemi§:rx

flaw ,

Major professor

Professor Alexander I. POPOV

 

1978

PHYSICOCHEMICAL STUDIES OF
MAGNESIUM COMPLEXES

By

Pierre-Henri C. Heubel

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements

for the degree of

DOCTOR OF PHILOSOPHY

Department of Chemistry

1978

ABSTRACT

PHYSICOCHEMICAL STUDIES OF
MAGNESIUM COMPLEXES

By
Pierre-Henri C. Heubel

Magnesium-25 NMR measurements were made in a variety of solvents
on several magnesium salts as a function of salt concentration.
Magnesium chloride and bromide were prepared fOllowing the method
described by Ashby and Arnott [redox reaction between magnesium
turnings and a mercury (II) halide in tetrahydrofuran (THF)]. When
ion pairing takes place in solution, a chemical shift change (up
to 30 ppm in propylene carbonate [PC]) and a broadening of the NMR
signal (up to 260 Hz in PC) are observed.

The limiting chemical shift of magnesium-25 in different solvents
was determined by studying solutions of MgClz, MgBrZ, MgClzI4 and
MgBr214 in acetone, acetonitrile, dimethylformamide, dimethylsufoxide,
methanol, PC, THF and water.

Complexation of magnesium ions by the cryptand Cle in dry
methanol results in a small line broadening, while the complexation
by phosphonoacetate (PA) in water broadens the signal to such an
extent that the NMR signal disappears. The larger distortion of
the electric field by the ligand PA is thought to be responsible for

this behavioral difference.

Pierre-Henri C. Heubel

Infra-red studies of acetonitrile solutions of MgBrz showed
complexation of the magnesium ion by the solvent.

The acidity functions of phosphonoacetic acid (PAA) and its
complexation with magnesium ion were also investigated by other
physicochemical techniques. The behavior of the phosphorus-31 and
carbon-l3 NMR signals 1; pH, as well as the accompanying change in
Jc_p coupling constant show evidence for a relationship between the
acetic acid moiety of PAA and its second acid dissociation constant
(0K2).

The pKa's of PAA were studied potentiometrically by coulometric
titration at ionic strengths from 0.l to 0.02 and numerical values
were obtained by using the computer program MINIQUAD 76A. Tetra-
methylammonium bromide was used as a supporting electrolyte.
Thermodynamic pKa's of 2.0, S.ll i 0.04 and 8.69 i 0.05 were
obtained by linear least squares fitting and extrapolation to zero
ionic strength.

The thermodynamic quantities for the second and third protonation
of PAA were obtained by studying the temperature dependence of the
formation constants. These two protonations of PAA are entropy and
enthalpy stabilized (entropy stabilization strongly dominant) as

shown below:

003 = -6.9 i 0.3 kcal/mole A03 = -ll.7 i 0.4 kcal/mole
Aug = -0.2 i 0.3 kcal/mole AH3 = -l.3 i 0.4 kcal/mole
A53 = 22.6 i 0.9 cal/mole, °K 053 = 35 i l cal/mole, °K

In the case of K], the experimental results were not sufficiently

precise for thermodynamic calculations.

Pierre-Henri C. Heubel

The stoichiometry of the completely deprotonated PAA-cadmium
complexes was investigated by cyclic voltammetry. A l:l and a
2:l (ligandzmetal) complex were found. The respective formation
constants of 8 x 103 and 8 x 101 were determined at 0.4 ionic
strength.

Different complexes (namely mono- and completely deprotonated
ones) were observed by potentiometry in the case of alkali and
alkaline earth metal ions. Thermodynamic values of the magnesium-
PAA complexes were determined in the same manner as the thermodynamic
pKa's. Log Kf values of 3.0 i 0.3 and 5.58 i 0.09 were obtained
for the mono- and deprotonated complexes, respectively.

The thermodynamic quantities:

AGO = -7.2 i 0.7 kcal/mole

AHO 3.0 i 0.7 kcal/mole
0

AS 35 i 2 cal/mole, °K
were calculated from the temperature dependence of the deprotonated
complex formation constant. The change in charge-to-size ratio of
the different ions upon complexation is thought to be responsible
for the entropy stabilization observed.

The stability of the deprotonated complex decreases when the
size of the alkaline earth ion increases (log Kf from 4.50 i 0.02
for magnesium ion to 3.67 i 0.02 for barium ion at 0.05 ionic strength),
or when the charge of the ion decreases (log Kf of 1.43 i 0.02 for
sodium ion at 0.07 ionic strength).

The pKa's and magnesium complex f0rmation constants of several

PAA analogs: phosphonoformic acid, 2-phosphon0pr0pi0nic acid, and

Pierre-Henri C. Heubel

3-phosph0nopropionic acid were determined. The results show
evidence for a relationship between the amount of magnesium complexed

at pH 7 (physiological pH) and the biochemical activity of the

ligand.

AUX "QUATRE AUTRES" ET A NANCY

11°

ACKNOWLEDGMENTS

The author wishes to thank Professor Alexander I. Popov for his
counseling and for his guidance into the field of chemical research
and scientific writing. Thanks are also extended to Professor
Bernard Tremillon for introducing me to the analytical field and
for his friendship.

Financial aid from the Department of Chemistry, Michigan State
University, and the National Science Foundation was appreciated.

The help of Frank Bennis and Wayne Burkhardt in keeping the
NMR spectrometers in operating condition is acknowledged.

Dr. John G. Hoogerheide (also known as Long John, Mutt, the
siamese brother, etc.) is thanked for his friendship and constant
encouragement during this whole research project. The friendship of
Dr. Jenshang Shih is also acknowledged.

The author would like to extend his appreciation to Dr. Neil
Armstrong f0r being the second reader, to Dr. Andrew Timnick for his
willingness to discuss technical problems, and to Dr. Donald Hard for
his help in computer programming and usage.

The author wishes to thank Nancy A. Fedewa fbr her love, patience
and concern throughout this research project. Her everyday smile was
greatly appreciated.

Finally, the author is deeply grateful to his family. The love,
consideration and encouragement of the "four others" contributed to

a large extent to the successful completion of this research project.

TABLE OF CONTENTS

Chapter Page
LIST OF TABLES ....................... ix
LIST OF FIGURES ....................... xi
PART I: MAGNESIUM-25 NMR STUDIES .............. 1
l HISTORICAL REVIEW .................... 2
1.1. Introduction ................... 3
1.2. Magnesium-25 NMR ................. 3
1.3. Proton NMR .................... 3
1.4. Ligands ...................... 6
1.5. Conclusion .................... 9
2 MATERIALS AND METHODS .................. 10
2.1. Materials ..................... 11
2.1.1. Reagents ................. 11
2.1.2. Solvents ................. 13
2.1.3. Molecular Sieves ............. 15
2.2. Instruments .................... 15
2.2.1. Flame Emission .............. 15
2.2.2. Nuclear Magnetic Resonance ........ 15

2.2.2.1. DA-60 NMR Spectrometer . . . . 16
2.2.2.2. Bruker HFX-90 NMR

Spectrometer ......... 16
2.2.2.3. Bruker WH-180 Superconducting
NMR Spectrometer ....... 16
2.2.2.4. Operating Procedures ..... 17
2.2.3. Infra-red Spectra ............ 18

iv

PART II:

1

2.3.
2.4.

NMR Sample Preparation ..............

Data Handling and Miscellanei ...........

RESULTS AND DISCUSSION .................

3.1.
3.2.
3.3.

3.4.

3.5.
3.6.
3.7.
3.8.

1

.1.
1.
1.

\l 0" 01 b (A) N
O C O O 0

Introduction ...................

Problems Specific to Magnesium-25 NMR .......

Solubilities of Magnesium Salts in Nonaqueous

Solvents

Study of Magnesium Salts .............

3.4.1.
3.4.2.

Water ..................
Nonaqueous Solvents ...........
3.4.2.1. Chemical Shift ........
3.4.2.2. Line-Width ..........

Donor Number Correlation .............

Complexation ...................

I.R. Studies ...................

Conclusion ....................

PHOSPHONOACETIC ACID STUDIES ...........
HISTORICAL REVIEW ....................

Introduction ...................

Phosphonoacetic Acid ...............

Analytical Techniques ...............

Nuclear Magnetic Resonance Studies ........

Acidities and Actinide Complexes of PAA ......

Industrial Applications ..............

Biological Properties of PAA ...........

1.7.1.
1.7.2.

Herpesviruses ..............

Herpesvirus DNA polymerase ........

18
19
20
21
21

23
24
24
24
24
33
36
39
42
42
44
45
46
47
49
49
50
52
52
52
53

1.7.3. Herpesvirus Mode of Action ........

1.7.4. Inhibitory Power of PAA .........
1.7.5. Drugs Against Herpesviruses .......
1.7.6. Toxicological Effects of PAA .......
1.8. Phosphonoacetic Acid Derivatives .........
1.9. Conclusion ....................
MATERIALS AND METHODS ..................
2.1. Reagents .....................
2.2. Resins ......................
2.3. Spectroscopy ...................
2.3.1. Visible Spectroscopy ...........
2.3.2. Nuclear Magnetic Resonance
Spectroscopy ...............
2.4. Cyclic Voltammetry ................
2.5. Potentiometry ...................
2.6. Coulometer ....................
2.7. Sample Preparation ................
2.7.1. Cyclic Voltammetry ............

2.7.2. Sodium Ion Electrode Measurements

2.7.3. pH Measurements .............

2.8. Data Handling ...................

RESULTS AND DISCUSSION .................

3.1. Acidity Studies of PAA ..............

3.1.1. Spectroscopic Studies of PAA .......
3.1.1.1. Phosphorus-31 and Proton

NMR . . . . . .........

3.1.1.2. Carbon-13 NMR .........

vi

53
54
57
57
60
61
62
63
64
64
64

64
65
65
67
7O
7O
7O
70
72
73
74
74

74
81

3.1.2. Potentiometric Studies of the pK's

of PAA .................. 87
3.1.2.1. Coulometric Titrations . . . . 87
3.1.2.2. pK's of PAA .......... 90
3.1.2.3. Thermodynamics of
Protonation .......... 99
3.2. Complexation Studies of PAA ............ 102
3.2.1. Spectroscopic Studies .......... 102
3.2.1.1. Visible Spectra ........ 102
3.2.1.2. Sodium-23 NMR ......... 103
3.2.1.3. Phosphorus-31 NMR ....... 103
3.2.2. Cyclic Voltammetry ............ 105

3.2.3. Potentiometric Studies of
Complexation ............... 111

3.2.3.1. Sodium Ion Electrode Study . . 111
3.2.3.2. Glass Electrode Study ..... 111

3.2.3.3. Thermodynamics of
Magnesium Complexation . . . . 120

3.3. Phosphonoacetic Acid Analogs ........... 120

3.3.1. Spectroscopic Studies of PFA ....... 121

3.3.2. Determination of the pK's ........ 121

3.3.3. Complexation ..... ' .......... 122

3.4 Conclusion .................... 125
APPENDICES

A SODIUM-23 NMR ...................... 126

A.1. Introduction ................... 127

A.2. Results and Discussion .............. 127

B APPLICATIONS OF THE COMPUTER PROGRAM KINFIT4 ...... 135

vii

8.1. Calibration of a pH Glass Electrode ........

8.1.1. Program Function .............

B.1.2. Subroutine EQN .............

8.1.3. Data Sample ...............
8.2. Determination of the Formation Constants

of 1:1 and 2:1 Complexes From NMR Data ......

8.2.1. Program Function .............

B.2.2. Subroutines ...............
8.2.2.1. Subroutines EQN for

Data Simulation ........

8.2.2.2 Subroutine EQN ........

8.2.3. Data Sample ...............

C APPLICATION OF THE COMPUTER PROGRAM MINIQUAD 76A
TO THE DETERMINATION OF EQUILIBRIUM CONSTANTS FROM
POTENTIOMETRIC DATA ...................
C.l. Program MINIQUAD 76A ...............
C.2. Data Input Instructions for MINIQUAD 76A .....
C.3. Data Sample ....................

BIBLIOGRAPHY ........................

viii

136
136

137
139

139

Table

II

III

IV

VI

VII

VIII

IX

LIST OF TABLES

Page
Key Solvent Properties and Correction
for Magnetic Susceptibility ............ 14
Line-Width of the Magnesium-25 NMR
Signal in Different Solvents ........... 34
Limiting Chemical Shift of Magnesium-25
in Various Solvents ................ 38
Vibrational Frequencies of Free and
Complexed Acetonitrile .............. 43
Carbon-13 Chemical Shift of 0.7 M_
PAA y§_pH ..................... 84
Acidity Functions of Citric Acid at
Ionic Strength 0.l ................ 91
Acidity Functions of PAA at Various
Ionic Strengths at 25°C .............. 97
Thermodynamic Quantities for Acid
Association and for the Mg-PA
Complex ...................... 101
Half-wave Potential of the Cd(II)/Cd
Couple in the Presence of PA
y§_Thallium Amalgam ................ 110

ix

XI

XII

XIII

PAA Complexes With Metal

Ions .......................
Formation Constants for PAA-Magnesium

Complexes at Various Ionic

Strengths at 25°C .................
Complexation Formation Constants

of PAA Analogs at 25°C and 0.05

Ionic Strength ..................
Formation Constants of Sodium-(lZ-C-4)

Complexes and Their Chemical Shifts ........

112

115

123

134

Figure

LIST OF FIGURES

Page

Structures of Representative Cryptands and

Crowns (Cryptand: The Numbers Represent the

Number of Oxygens in Each Bridge, Crown: The

First Number is the Number of Atoms in the

Cycle, and the Second, the Number of Oxygens ..... 7
Magnesium-25 Chemical Shifts 0f MgBrz

in Acetonitrile ................... 26
Magnesium-25 Chemical Shifts of M9012 (A)

and MgC1214 (I) in PC ................ 27
Magnesium-25 Chemical Shifts in THF [MgClz (X),

MgC1214 (0)] andAcetone [MgC12 (A),

MgClZI4 (I)] .................... 28
Magnesium-25 Chemical Shifts of MgCl2 in

Methanol (A), DMSO (0) and DMF (I) and 0f

MgC1214 in DMF (x) .................. 29
Magnesium-25 Chemical Shift of MgCl2 !§_

Mole Ratio of Iodine in Acetonitrile ......... 31
Magnesium-25 Chemical Shift of MgCl2 y;

Mole Ratio of Iodine in PC .............. 32
Limiting Magnesium-25 Chemical Shift in Various

Solvents y§_Their Gutmann Donor Number ........ 37

xi

1O

11

12

13

14

15

16

17

18

19

Structures of Various Phosphonic Acids:
Phosphonoacetic Acid, Phosphonoformic Acid,

2-Phosph0n0pr0pi0nic Acid, and

3-Phosph0n0propionic Acid .............

DNA Synthesis and Mode of Action

of PAA .......................

Block Diagram of the Constant

Current Coulometer .................

Phosphorus-31 Chemical Shift of 0.3 M

PAA y§_pH .....................

Comparison of the Phosphorus-31 Chemical Shifts

and the Species Distribution of PAA y§_pH .....

Carbon-13 Chemical Shift (Referenced to Dioxane,

Which is 68.1 ppm Downfield of TMS) of the CH2

Carbon of 0.7 M_PAA y§_pH .............

Change in JC-P Coupling Constant of the CH

Carbon of PAA y§_pH ................

Plot of the pKZ of PAA gs Ionic Strength,

the Line is a Linear Least Squares Fit .......

Plot of the pK3 of PAA y§_Ionic Strength,

the Line is a Linear Least Squares Fit .......

Species Distribution Plot of PAA at 0.05

Ionic Strength ...................

Temperature Dependence of 1n Kf for Three PAA

Complexes, the Lines are Linear Least Squares Fits . .

xii

55

68

95

96

98

100

20

21

22

23

24

25

26

27

28

29

Phosphorus-31 Chemical Shift of 0.3 M_PAA and

of a 1:1 Mg:PAA Solution y§_pH ............ 104
Phosphorus-31 Chemical Shift of 0.3 M PAA Xi

Mole Ratio of Magnesium at pH Above 9 ........ 106
Plot of E1/2 as a Function of Log [PA] for .

the Cd2+-PA System (Reference: Thallium Amalgam) . . . 109
Coulometric Titration Curves of the PAA and

PAA-Mg2+ Systems ................... 113
Plot of Log Kf of the Monoprotonated PAA-Mg2+

Complex y§_Ionic Strength, the Line is a

Linear Least Squares Fit ............... 116
Plot of the Log Kf 0f the Completely

Deprotonated PAA-Mg2+ Complex y§_lonic Strength,

the Line is a Linear Least Squares Fit ........ 117
Species Distribution Plot of the PAA-Mg2+ System

at 0.05 Ionic Strength ................ 119
Species Distribution Plot of the PFA-M92+ System

at 0.05 Ionic Strength ................ 124
Sodium-23 Chemical Shift y§_M01e Ratio of 12-C-4.

NaBPh4: (O): 0.1 M in PC, (I): 0.034 M in THF:

(0): 0.04 M NaI in Methanol: (D): 0.04 M

NaClO4 in Water ................... 129
Sodium-23 Chemical Shift y§_Mole Ratio of 12-C-4.

NaBPh4: (I): 0.04 M in Pyridine, 0.1 M in DMSO

(A). Acetone (x). DMF (A), Acetonitrile (o) ..... 130

xiii

PART I

MAGNESIUM-25 NMR STUDIES

CHAPTER 1

HISTORICAL REVIEW

1.1. INTRODUCTION

The main purpose of this thesis was to study the complexes
formed between magnesium ion and phosphonoacetic acid (PAA) by
different physicochemical techniques. Techniques such as visible
UV spectroscopy are not very useful to study the environment of
group I or group II metal ions in their stable oxidation states.
In the case of most of the alkali metal ions, the observation of
the metal nuclide resonance and the study of both the line-width
and the chemical shift of the signal gave interesting information
about the chemical environment and the complexation of these
nuclei in their first oxidation state (1). For these reasons,
the magnesium-25 nuclear magnetic resonance (NMR) technique was

applied to the magnesium-phosphonoacetic acid system.
1.2. MAGNESIUM-25 NMR

However, the properties of the magnesium-25 nucleus are not
very favorable for NMR studies. The magnesium-25 nucleus has a
low nuclear magnetic sensitivity (the senstivity per nucleus at

constant field is only 2.68 x 10'2

of the sensitivity of the proton),
a low natural abundance (10.05%), a low resonance frequency and a
quadrupolar momentum (the nuclear spin is 5/2). In 1951, Alder and
Yu (2) located the resonance frequency of magnesium-25 in a 4.6 M
solution of magnesium chloride in water near a frequency of 3.9 MHz,
in a magnetic field of 11 k6. They also measured the magnetic
moment of magnesium-25 by NMR. However, because of the poor

properties of the magnesium-25 nucleus mentioned earlier, few

studies of magnesium-25 NMR have been performed.

4

Using continuous wave technique, Ellenberger and Villemin (3)
reported magnesium-25 NMR studies of aqueous solutions. They
found lineridths of 40 to 45 Hz for the different solutions used.
However, within the precision of the measurements, the resonance
frequency seemed to be independent of the salt concentration.

Subsequently, Dougangt 91. (4) and Dickson and Seymour (5)
determined the Knight shift of magnesium-25 in magnesium metal,
that is the paramagnetic shift of the nuclear resonance due to the
delocalized conduction electrons of a metal (named after W. D. Knight
who was the first one to observe it in copper in 1949). Bryant (6)
used magnesium-25 NMR to study aqueous solutions and reported a line-
width (full width at half height of the peak) of 3.8 i 0.5 Hz for
a 1.5 M_MgC12 solution, which is an order of magnitude smaller than
the one reported by Ellenberger and Villemin (3). Bryant's use of
magnesium-25 NMR was essentially to determine the chemical rate 0f
exchange (based on measurements of the line-width) of magnesium
between complexed and uncomplexed sites of ATP type phosphate
solutions.

In the absence of chemical shift change in aqueous magnesium
solutions, Magnusson and 80thner-By (7)used the change in line-
width gs pH of acidic ligand solutions (e.g., citric acid, ATP) to
approximate the formation constants of the corresponding complexes.
Although this method seems unsound because of the various factors
affecting the lineawidth of the signal and because of the precision
of such measurements, their values are within one order of magnitude
of other published values. The above authors reported a variation

of the line-width from 5 to 9 Hz when the concentration of magnesium

ions was increased from 1.0 to 4.0 M, They also reported that above
pH 7, a solution of magnesium ions and EDTA gave one broad peak,
thus showing that the exchange was fast compared to the NMR time
scale. The widest line-width reported for magnesium-25 NMR was

in the case of complex formation with ATP (1300 Hz).

Simeral and Maciel (8) are among the last ones so far to have
reported magnesium-25 NMR studies, and those were also done in
aqueous solutions. After having taken the viscosities of the
different solutions into account, the line-width data have been
discussed in terms of four models for quadrupolar nuclear spin
relaxation in the systems studied. The line-widths reported vary
from 3 to 11 Hz, but once again, no chemical shift change either

with the concentration or the counterion was observed.
1.3. PROTON NMR

Proton magnetic resonance has been used to study the solvation
number of magnesium ion in acetone-water and methanol-water mixtures
(9-15). The inner solvation shell consisted of six water molecules
as long as the mole ratio of water to magnesium was equal to or
higher than 6. Matwiyoff and Taube (10) also determined that in
methanol-acetone mixtures, methanol was a much better solvating
agent than acetone. In all those solutions the total solvation
number was 6.

The preparation of anhydrous magnesium salts presents a
difficult problem. With one exception [16, redox reaction between
magnesium and mercury (II) halide in tetrahydrofuran], all the

purifications or synthesis (17) of magnesium chloride or bromide

6
found in the literature involve the fusion of the halide and its
separation from magnesium oxides by filtration in a stream of dry
nitrogen in a quartz tube (the melting points of MgC12 and MgBr2
are 708° and 700°C, respectively).

The affinity of magnesium ions for water is such that direct
heating of the hydrated salt under active vacuum leads to the
formation of magnesium oxides. Thus, according to Halliwell and
Nyberg (18), the hydration enthalpy of magnesium is -459 kcal/mole,
respectively four, five, six and seven times the ones of lithium
(~124 kcal/mole), sodium (-97 kcal/mole), potassium (-77 kcal/mole)
and cesium (-63 kcal/mole). Since all of the above magnesium-25
studies were restricted to aqueous solutions, it was of interest

to us to study this resonance in nonaqueous solvents.
1.4. LIGANDS

Cyclic polyethers, or crown ethers, were developed by Pedersen
(19 and references therein). They have a remarkable ability to form
stable two-dimensional complexes with alkali metal cations. Struc-
tural formulae of some typical crown ethers are shown in Figure 1.
Formation of such complexes distorts the spherically symmetrical
electric field around the solvated metal ion which may result in
a broadening and/or a chemical shift change of the metal NMR signal.

There are (19-23) several reports of syntheses, spectroscopic
and crystallographic studies of the ligand lZ-crown-4 (12-C-4), also
known as the ethylene oxide cyclic tetramer (1,4,7,10 tetraoxa-
cyclododecane). The ligahd 12-C-4 is a liquid at room temperature

with a specific gravity of 1.104, a boiling point of 238°C at 760 mm

.Amcmmaxo mo .8952 m5 .283 mg..—
vce .298 05 5 mac; Co c3532 65 3 $452 amt... 0.: "586 6375 53 E. Swabs .8
$4.52 9.» 23¢..qu €2.52 2:. 653.28 mczocu E; mucougcu 2333333. “B 83323 .p 953...

moo 5%

cup—305-2 ticgu au—

3 0
.CL.

8

of mercury and a vapor pressure of 0.03 mm at room temperature (24).
The cavity size of this crown is 1.2 to 1.5 A, which is approximately
the size of the magnesium ion (1.3 A diameter from crystallographic
measurements). In a study of the complexation of crown ethers,
Chock (25) found that the enthalpy of hydration involved in the
desolvation step and the binding energy for a given ligand were the
most important factors in the mechanism of complexation and the
stability of the resulting complex. These observations, added to the
great affinity of magnesium ion for water, probably explain why the
crystal structure of MgClz, 6H20-12-C-4 complex (20) consists of
octahedral Mg(H20)§+ units very similar to those in MgClz, 6H20,
located on crystallographic axes, with 8 of the 12 water hydrogens
forming hydrogen bonds to six different chloride ions, which are
located in tetrahedral holes, while the remaining four hydrogens
of the Mg(H20)g+ unit form hydrogen bonds to ether oxygens from
four different cyclomer molecules.

Leong gt_gl, (24, 26) reported some toxicological studies
of 12-C-4. Rats exposed to the vapors of 12-C-4 for six hours daily
for six to eight days exhibited variable degrees of anorexia, loss
of body weight, asthenia, hindquarter incoordination, testicular
atrophy, auditory hypersensibility, tremors, convulsions, moribond
conditions and death, depending on the exposure concentration.

Cryptands are diaza-polyoxamacrocycles. These ligands with
bridging nitrogens were first synthesized by Lehn £3 21, (27, 28).
They can form very stable complexes with alkali and alkaline-earth
cations. The length and number of ether bridges may be changed to

vary the size of the inner cavity of the cryptand. The term cryptand

9

refers to the ligand and cryptate to the complex. The stability
of the cryptate complexes depends to a large extent on the size
relationship between the three-dimensional cavity of the cryptand
and the diameter of the complexed ion.

In general, cryptands form much stronger complexes with the
alkali metal ions than the crown ethers, provided that ligands with
approximately the same cavity size are compared. Recent reviews (29-33)
contain extensive compilations of complexation stability constants.

Cryptand 0211 is shown in Figure 1 (21] refers to the number
of oxygens in each hydrocarbon chain). Its cavity size is approximately
1.6 A. Very little work has been done on the complexation of Mg2+
ion by C211. Lehn and Sauvage (29) reported log Kf for a 1:1 complex
to be 2.5 i 0.3 and 4.0 i 0.8 in water and in 95/5 methanol/water

mixture respectively, while Morf and Simon (30) found a log Kf

value of less than 2 in water.
1.5. CONCLUSION

Very little work has been done concerning the NMR properties
of magnesium-25. In order to use this technique to study the
magnesium-PAA complexes, it is necessary to perform preliminary
studies of the influence of the environment of the magnesium ion

upon its NMR properties.

CHAPTER 2

MATERIALS AND METHODS

10

11
2.1. MATERIALS

2.1.1. REAGENTS - The following chemicals were all reagent
grade and used without further purification. The code is: Aldrich (A),
Alfa inorganics (AI), J. T. Baker (8), Drake brothers (0), Fisher
scientific company (F), Mallinckrodt (M), Matheson Colleman and Bell
(MCB) and G. F. Smith Company (S):

Barium chloride (M), barium oxide (MCB,F), cadmium nitrate
tetrahydrate (8), citric acid monohydrate (F), cobalt chloride (AI),
cobalt sulfate (8), copper sulfate (AI), dioxane (F), ethylene
diammine tetraacetic acid, disodium salt (MCB), glacial acetic acid
(M), hydrochloric acid (M), lithium bromide (MCB), magnesium acetate
tetrahydrate (8), magnesium bromide hexahydrate (F), magnesium chloride
hexahydrate (MCB), Magnesium perchlorate (S), magnesium sulfate (M),
manganese chloride (AI), nickel bromide (AI), potassium hydroxide
(MCB), potassium iodide (F), potassium permanganate (B),
3(trimethylsilyl)-1-propanesulfonic acid, sodium salt hydrate
(known as 055, the water soluble TMS: (CH3)3Si(CH2)3SO3Na, nHZO,
is an internal standard for use with 020) (A), sodium carbonate
(F), sodium hydroxide (0), sodium iodide (F), tetraphenylarsonium
chloride (S) and zinc chloride (M). Tetramethyl ammonium hydroxide
was from either Eastman organics or MCB and was obtained as a 10%
solution in water. Phosphonoformic acid, trisodium salt hexahydrate
and 3—phosphonopropionic acid were kindly furnished by J. Reno from
Dr. J. Boezi's group (Biochemistry Department) and used as received.
Cryptand 211 was from E. M. Lab, and 12-crown-4 from Aldrich;

3-phosphonopropionic acid was from Richmond Organics.

12

Several indicators have been used: bromocresol green and
eriochrome black T (F), phenol red (A), and phenolphthalein (F).

The following chemicals were dried in the vacuum oven, unless
otherwise Specified (Company, time, temperature): magnesium metal
(8, 6h., 60°C), sodium tetraphenylborate (8, 48h., 25°C), lithium
perchlorate (F, 96h., 190°C, regular oven), mercuric chloride,
bromide and nitrate (M, 6h., 60°C), sodium chloride (M, 24h., 150°C),
sodium perchlorate (S, 24h., 110°C) and mercuric thiocyanate
(Ventron Alfa Products, 6h., 60°C).

A flame and K. Fischer analysis of 12-crown-4 showed that the
mole ratios of Na+ ion and water to 12-crown-4 were 0.26% and 2%,
respectively. No traces of lithium or potassium were detected.

Iodine (Baker Analyzed reagents, ACS or USP grade) was purified
by double sublimation: 10 g of iodine are first sublimed with 5 g of
potassium iodide (in order to eliminate the iodine chloride and iodine
bromide) at 70-80°C. It takes about 10 hours to sublime 10 g of iodine.
Then the obtained iodine is sublimed with 4.3 g of granulated barium
oxide (in order to remove the water) under the same conditions.

Bromine (Dow Chemical) was dried over P205 for 24 hours and
then distilled under atm05pheric pressure.

Magnesium salts: anhydrous magnesium bromide was first prepared
by direct reaction of bromine over magnesium in ether, but contamina-
tion of the solution by bromine made the solution useless for NMR
purposes. Therefore, magnesium chloride and bromide were prepared
by using the method described by Ashby and Arnott (10).

A redox reaction between a mercury (11) halide and an excess

of magnesium metal in a tetrahydrofuran (THF) diethylether mixture

l3
(kept under reflux for 4 to 6 hours) leads to the formation of
the corresponding magnesium salt and a magnesium amalgam. It was
found here that THF alone gave better results for M9012 and MgBrz
than the THF-ethylether mixture. Mercury (11) bromide (86 g.,
0.24 mole) and metallic magnesium turnings (6 g., 0.25 mole) were
refluxing in 1.4 1. of THF for 6 hours (in the case of magnesium
chloride, the concentrations were doubled: 0.5 mole in 1.4 1.).
The solution then was transferred to a dry box and the amalgam
and excess of magnesium were removed by filtration through filter
paper (because of the presence of liquid mercury and of clogging
problems, a glass frit could not be used).

Tetrahydrofuran is then evaporated on a rotavap under vacuum
and the salt is dried under vacuum at 80°C overnight. Omission of
this step lead to supersolubility problems, due to the contamination
of the solution by the small amounts of THF remaining in the salt.
The salt was stored in a dry box where the nonaqueous solutions
were prepared.

Attempts to prepare Mg(N03)2, Mg(SCN)2 or Mg(BPh4)2 by the
same technique failed, as well as attempts to prepare Mg(8Ph4)2
by either using the method given in the literature for the lithium
salt (34), or the potassium salt (35). This latter method was used
to prepare Hg(8Ph4)2.

2.l.2. SOLVENTS - The solvents were obtained from the following
companies: Acetone (F), acetonitrile (MCB), dimethylformamide (DMF, F),
dimethylsulfbxide (DMSO, F), methanol (F), propylene carbonate (PC, A),
tetrahydrofuran (THF, M). (See code in Part II.2.1.l.). Useful

solvent parameters are listed in Table I.

14

 

 

 

ooo.o ooo.o o.mm em.mn seam:
mee.o+ e-.o- o.o~ mm.“ Razhv catacosuzeespah
omm.o+ owp.o- _.mp o.mm Auav eueeoeeeu eeeFAaoee
mmm.o+ m~3.o- n.m~ A.Nm Pegasus:
~m3.o+ Fem.o- w.m~ «8.83 Aomzov eewxoc_=mpxeoeswo
mpm.o+ mom.o- c.m~ p~.om Aazav eUPEessoc_»eoeewo
own.o+ omm.o- _.5_ m.~m ep_sopeoueo<
ao.p+ mem.o- o.AF A.o~ eeooae<
253 8 I. :53 8-5 89.52 8:8 28.28 85; Sm
co cowuomccou co cowuumccou m.c:aEp=o owcpuopmwc

 

 

.quFWaPpnmumsm uwumcmoz com cowuuoccou use mowpcmaocm uco>Pom xox .H mpnm»

15
All solvents except methanol were dried over calcium hydride

or activated 3A and 4A molecular sieves (3A has a cavity size that
just fits the water molecule) for 5 hours, followed by a distillation
at atmospheric pressure (except for DMSO for which a static low
pressure was applied). Usually 10 g of CaHZ or 100 ml of Sieves

were used for one liter of solvent. Methanol was dried by double
distillation over magnesium (36). The water contents of the solvents
used were measured with the help of a Karl Fischer automatic titrator

and were found to be always less than 50 ppm's.

2.l.3. MOLECULAR SIEVES - Molecular sieves 3A (Davison
Chemical Company, W. R. Grace and Co.) and 4A (MCB) were activated
by drying at a maximum of 400°C for 6 hours under nitrogen which

was dried by passing it through concentrated sulfuric acid.
2.2. INSTRUMENTS

2.2.1. FLAME EMISSION - Flame emission experiments were performed

 

using a Heath EU-703-70 flame unit, an EU 700 scanning monochromator
and EU-70l-30 photomultiplier module containing an R-44-UR Hammamatsu
photomultiplier tube. Output from the P.M. tube was recorded through

a Heath EU-703-3l photometric readout amplifier. The total consumption
burner used hydrogen at 1 psi and air at 15 psi. Samples were
aSpirated for at least 30 seconds with distilled water blanks between

samples. Duplicate runs were made for each standard and unknown.

2.2.2. NUCLEAR MAGNETIC RESONANCE - The NMR spectra were

 

collected on three different instruments, all operated in the

Fourier transform mode.

16
2.2.2.1. DA-6O NMR Spectrometer - A modified Varian

 

DA-60 NMR spectrometer was used for magnesium-25 and sodium-23 NMR
measurements. The instrument consists of the magnet of a Varian
DA-60 Spectrometer equipped with a wide band probe capable of multi-
nuclear operation, as described by Traficante gt_al, (137). The
field is at 14.09 kG and the frequency of the RF unit is 56.44 MHz.
A detailed description of the spectrometer has been given elsewhere
(133), The resonance frequency of the sample can be obtained by
changing the synthesizer frequency, instead of changing the magnetic
field.

Before each experiment, the probe and image filter were tuned to
the resonance frequency of the nucleus under consideration (3.672 MHz
f0r magnesium-25 and 15.87 MHz for sodium-23). The field is locked
by a home-built lock probe on the proton resonance of the upfield
sideband of H20). Pulse generation and data collection were performed
with the help of a Nicolet 1083 computer with 12K of memory. The tubes
were 25 and 15 mm 00 (magnesium-25 NMR) and 10 mm 00 (sodium-23 NMR).

2.2.2.2. Bruker HFX-90 NMR Spectrometer - The phosphorus-
31 and sodium-23 NMR spectra were obtained by using a Bruker HFX-90
NMR spectrometer equipped with a Nicolet 1083 computer with 12K of
memory, a Diablo disc memory unit and a Nicolet 293 I/O controller.
The spectra were obtained at a frequency of 36.44 MHz (phosphorus-31)
and 23.81 MHz (sodium-23). Spinning tubes of 10 mm 00 were used.

2.2.2.3. Bruker WH-180 SuperconductinngMR Spectrometer -
This instrument was used to collect magnesium-25 NMR spectra at

11.01 MHz, and proton spectra at 180 MHz. The field strength is

17

42.28 kG. An external 020 lock was used. Due to the superconducting
magnet, the field was very stable, and results were obtained without
using an internal lock for the nonaqueous samples.

Solutions were run in 20 mm 00 nonspinning tubes, the minimum
amount of solution for each tube is 8 m1. Preliminary studies
showed that due to the broadening of the peak by quadrupolar
relaxation, the absence of spinning did not affect the line width

to a measurable extent.

2.2.2.4. Operating Procedures - The chemical shifts
were measured against external references. A 4.0 M solution of
aqueous NaCl was used for sodium-23, and a 4.0 M solution of aqueous
MgClz (DA-60) or a 4.0 M 75/25 HZO/DZO MgClz solution (8-180) for
magnesium-25. AS shown by Maciel (8), there is no difference in
chemical shift for these last two solutions. The phosphorus-31
NMR chemical shifts are reported relative to 85% phosphoric acid.

According to the convention used in this thesis, a positive
change in chemical shift of sodium-23, magnesium-25 and phosphorus-31
NMR signals corresponds to a shift to higher field strength than
that of the reference signal.

All chemical shifts are corrected for the bulk diamagnetic
susceptibility difference between water and the pure nonaqueous
solvent. For each solution, the contribution of the salt to the
susceptibility of the solution is assumed to be negligible: a
reasonable assumption as shown by Templeman and Van Geet (37). The
formulae given by Live and Chan (38) for normal (equation 1) and

superconducting (equation 2) magnet were used:

18

acorr = 6obs -%m(x€ef _ Xsample) (1)
5corr = aobs +%fl(xcef _ xsample) (2)

In the case of magnesium-25 and phosphorus-31 NMR, when the
top of the peak was too flat to allow precise determination of the
chemical shift change, the precise frequency of the peak was obtained

by averaging the two frequencies at half height.

2.2.3. INFRA-RED SPECTRA - Infra-red spectra covering a range
of 4000-600 cm'1 were collected on two Perkin-Elmer grating IR
spectrophotometers: models 237 B and 457. The frequencies were
calibrated using polystyrene reference peaks. Solutions and nujol
mulls of solid samples were placed between sodium chloride mull

plates.
2.3. NMR SAMPLE PREPARATION

Nonaqueous solutions of magnesium salts were prepared in a
dry box. When the dissolution of the magnesium salt lead to the
formation of a light precipitate [presumably Mg(OH)2], the solution
was filtered in the dry box. Since magnesium has a large affinity
for water, the water content for each sample was measured after
each NMR measurement, and data points were kept only if the mole
ratio of water to magnesium was less than 20%. In these conditions,
the lowest concentration studied in nonaqueous solvents was 10"2 M,
which corresponds to a water content for the solution of 36 ppm.
In most cases, the lowest concentration was 3 x 10"2 M. The

magnesium concentration in the different solvents was measured

19
by diluting a small sample (usually 1 ml) of solution with distilled
water followed by the usual titration by EDTA in NH3/NH4Cl buffered
medium.

Since 12-crown-4 and cryptand 211 are liquid at room temperature,
the preparation of the solutions containing them was completed outside
the dry box. The amounts of the above ligands or sodium salts used
were determined by weight. Suitable diultions were performed using

2, 5, 10, 25 and 50 ml volumetric flasks.
2.4. DATA HANDLING AND MISCELLANEI

Extensive use was made of the CDC 6500 computer with several
Fortran IV programs (see Appendices B and C). Elemental analyses
were done by Chemalytics Inc. of Tempe, Arizona, and/or by Spang
Microanalytical Laboratory, Eagle Arbor, Michigan; the latter
giving more precise results. Melting points were determined on
the Fisher-Johns melting point apparatus. The instrument was
calibrated from 40°C to 300°C by means of melting point standards
(Hoover). Analyses for water were accomplished with a Photovolt

Aquatest II automatic Karl Fischer titration apparatus.

CHAPTER 3

RESULTS AND DISCUSSION

20

21
3.1. INTRODUCTION

Previous studies in this laboratory and elsewhere have Shown
that the NMR of alkali metal ions offers a sensitive Probe of the
environment of these ions in different solvents and solvent mixtures.
The purpose of this study is to determine whether magnesium-25 NMR
would be a suitable technique to study the influence of the cation
on the ionic equilibria and ionic species present in various
solvents. The exchange of ions between different environments is
usually rapid with respect to the NMR time scale, resulting in
only one resonance signal at an average frequency determined by
the magnetic shielding and nucleus population in each of the sites.
Alteration of parameters such as concentration, counterions and
solvent produces changes in the relative proportion and type of
environment which may be reflected by a change in chemical shift

and/or line-width of the observed resonance.

3.2. PROBLEMS SPECIFIC TO MAGNESIUM-25 NMR

In the first place, magnesium ion has a very large affinity
for water. Since most organic solvents cannot be obtained completely
anhydrous, it is obvious that in such cases meaningful chemical shifts
for magnesium can only be obtained if the concentration of water is
much smaller than the concentration of the salt. We found that in
the present study a mole ratio of water to magnesium of 20% was the
upper water content limit. For example, a solution of magnesium
chloride in acetonitrile, with a mole ratio of water to magnesium
of 50% gave rise to a change in chemical shift of 2 ppm, compared

to a solution with a mole ratio of 10%.

22

Secondly, prior to this study no one had performed any
magnesium-25 NMR studies of nonaqueous solutions of magnesium
salts, and, therefore, one of the problems was to determine whether
the change in chemical Shift would be large enough to be detected
easily with our instruments (after all, lithium ion which has
approximately the same size as magnesium ion has a chemical shift
range of only a few ppm).

Thirdly, several instruments have been used in this study,
with a large increase in sensitivity for each change of instrument:
thus, with the first insert and probe, the sensitivity of the DA-60
was such that the magnetic field did not need to and could not be
tuned: in these conditions a 0.05 M aqueous solution of magnesium
chloride required 8000 scans (50 minutes of time with 4K of memory
and 5000 Hz sweepwidth for the computer) for a signal to noise ratio
of 2.5 (the signal to noise ratio is defined as the ratio of the
signal over the largest noise, multiplied by 2.4). The first
improvement was to use a better probe, which allowed to study the
same solution in much better conditions (6000 scans gave a signal
to noise ratio of 8). The second and best improvement was to use
the Bruker 180 Spectrometer with which a 0.046 M aqueous solution
of MgClZ can be analyzed in 10 minutes.

Finally, another major problem specific to magnesium is the
following - either the ion is strongly solvated and does not form
contact ion pairs, resulting in no chemical shift change with salt con-
centration, and small line-width change (due to the change in viscosity
of the solution), or there is ion pairing, which, if it results in a

change in chemical shift, also significantly increases the line-width.

23
For this latter reason and because of the low sensitivity of
the instrument, only few solutions could be studied with the DA-60

(mainly aqueous, methanol and acetonitrile solutions).
3.3. SOLUBILITIES OF MAGNESIUM SALTS IN NONAQUEOUS SOLVENTS

The solubilities of MgCl2 and MgBr2 in nonaqueous solvents
are fairly small. The best cases studied are represented by
acetonitrile (2.0 M), acetone (1.2 M), methanol (1.0 M) and
dimethylformamide (DMF, 0.9 MD. For propylene carbonate (PC)
and tetrahydrofuran (THF), the solubility is approximately 0.3 M,
while it decreases to 0.1 M for dimethylsulfoxide (DMSO), and
to 10'3 M for pyridine and nitromethane. It was first thought
that the solubility in nitromethane was around 1.0 M, but it was
found later on that THF impurity in the salt (from its synthesis)
was responsible for this high value. At the time the salt was
only dried at room temperature under vacuum for 24 hours. Upon
dissolution in a given solvent, the THF molecules solvate the
cation thus increasing the "solubility" of the salt. On occasion,
after the preparation of a 1.0 M solution of MgClz, the salt
would precipitate out, leading to a final solubility of the salt
of less than 10"2 M,

The solubility of magnesium acetate (Mg(0Ac)2) was found much
smaller as a rule than the solubility of MgClz or MgBrZ, since
except in methanol (0.1 M) all the other solvents tested (aminoethanol,
formic acid, f0rmamide, DMF, ethylenediamine, DMSO, THF, nitromethane,
pyridine, acetone) dissolved at most 10"2 M_Mg(OAc)2, which is not

suitable for a magnesium-25 NMR study in the present conditions.

24
3.4. STUDY OF MAGNESIUM SALTS

3.4.1. MAI§R_- Five different magnesium salts have been
studied in water, Mg(OAc)2, MgClz, MgIZ, Mg(NO3)2, and Mg(ClO4)2.
No change in chemical shift was observed when the concentration
of each of these salts was changed. In the case of four of these
salts [MgClzz 4.6 to 0.1 M, MgIz: 1.8 to 0.1 M, Mg(NO3)2: 3.0
to 0.05 M and Mg(C104)2: 3.2 to 0.05 M) the change in lineewidth
was small: from 17 to 7 Hz, while for Mg(OAc)2 the line-width
changed from 43 to 8 Hz for a corresponding change in concentration
of 2.0 to 0.24 M, However, in the latter case, the concentrated
solution was much more viscous than the other salts solutions,
and this fact is probably responsible for the broadening of the
peak. These results are similar to the ones reported by Maciel
gt_al, (8). As expected, magnesium ion is strongly solvated by
water, and from these measurements, it appears that if ion pairs
are formed between the different ions in solution, these are at

least solvent shared ion pairs.
3.4.2. NONAQUEOUS SOLVENTS

3.4.2.1. Chemical Shift - It has been previously observed
(39, 40) that the contact ion pair equilibria strongly depend on the
donor ability of the solvent molecule as well as on the bulk dielectric
constant of the medium. Although PC and acetonitrile have high
dielectric constants (65 and 37.5, respectively), their donor
abilities are low on Gutmann's scale (41) (15.1 and 14.1, respectively).

Gutmann defined a donor ability scale for solvents as the negative

25
AH-value in kcal/mole for the interaction of the solvent with
SbCl5 in a highly diluted solution of dichloroethane. We see from
Figures 2 and 3 that the magnesium-25 chemical shifts of MgBrz
and MgClz solutions are concentration dependent and, therefore,
that there is contact ion pair formation. These results are in
agreement with the data obtained by sodium-23 and lithium-7 NMR
in PC and acetonitrile solutions where the chemical Shifts of NaI
and NaSCN or LiBr and LiI are strongly concentration dependent (42, 43).

Tetrahydrofuran and acetone are interesting solvents in that
they have low dielectric constants (7.58 and 20.7, respectively), but
reasonable donor abilities (20.0 and 17.0). As can be seen from
Figure 4, there is contact ion pair formation even between Mg2+
and the Cllé ions.

0n the other hand, solvents like methanol, DMSO or DMF that
have both high dielectric constants and high donor numbers and are
good solvating agents show no or very little ion pairing (see Figure
5). For the methanol and DMSO solutions, the chemical shift does
not change with concentration (1.1 and 3.5 ppm, respectively),
while for the DMF solutions, the chemical shift changes from -4
to 4 ppm for a concentration change of 0.9 to 0.03 M.

We mentioned above that we attribute the concentration
dependence of magnesium-25 chemical Shifts to the formation of
contact ion pairs, i.e., to cases where the anion directly replaces
a solvent molecule or molecules in the inner solvation shell of
the cation. It seems reasonable to assume that gross variations

in the chemical shift and/or the line-width of the magnesium-25

26

.e_peoweeueu< :0 News: ca mocpem _eupeeeu m~-e=mme=mez .N ee=m_a

 

2.58.84 5 «cm as.

0.. 0.0
_ 40m:

(wdd) ””9

 

26

rung ow

628.2362 5 N28: .3 3:5 .8255 3-5856: .N 2:2...

0..

0.0

 

 

o__.=_c2oo< c. N6 02

0

Own

 

27

rm

0.): .

no

.8 E i exam: 2; 3 ~53. .8 3:5 8:85 3.552%: .m 23:

ONO
H

0.0
0N:

 

 

.28

.2! :38: . 9
. mmmcmez e we: .0
N u< use :3 53%: .00 ~82: “.E 5 3:5 59:55 mm :5 0.0
a 0 0:30
3 5 E 0.0 0.0...
”to; 6.. s
_

 

29

be

as: o._

.8 “=3 5 3.6.sz

co 2; i “=3 nee i 85 .3 .828: 5 N89,. co Stem .6385 8.53886: .m 28.:

00

0

 

4

_

0N...

 

 

 

 

40.

30

NMR signal are related to a change of its direct chemical environment,
that is,it reflects the influence of its nearest neighbors. It is
well known that NMR gives information about the inner solvation
shell of the nucleus under consideration. Replacement of the solvent
molecule(s) by the anion may either increase or decrease the electron
density of the cation. Replacement of solvent molecules by halide
ions increases the electron density around the magnesium ion
resulting in a downfield chemical shift of the Mg-25 resonance.
Similar behavior has been previously observed with Li, Na, K, and
Cs salts studied by Li-7, Na-23, K-39 and Cs-l33 NMR (44).

Another evidence f0r the formation of contact ion-pair in
the above solutions is the result obtained by studying the ngz-I2
mixtures in different solvents. It is well known (45) that halogen
molecules interact with halide ions to form trihalide ions which

7

are quite stable in nonaqueous solvents (e.g., Kf of 106°6, 10 and

1010 for 13, Br§ and C13 ions respectively in acetonitrile). We
can expect, therefore, the IZCl‘ complexes to be stable, and, in
the concentration range used (0.7 to 0.05 M), we can assume that
at a 1:1 mole ratio of iodine:halide ion, all the halide ions will
be complexed. The size of the anion will become significantly
larger and thus will have a much smaller tendency to form ion-pairs
with magnesium ions.

The effect of iodine upon Mg2

+-X' ion-pairing is shown in
Figures 6 and 7 - there is a linear change in chemical shift
[similar to the one reported by Erlich and Popov (46) in the case
of sodium-23 NMR] y§_mole ratio of iodine to Mg2+ ion up to a

value of 2 where all the halide ions are complexed. The largest

31

62.5.5884 .... 9:8. ca 6.5.2. 6.6... m.» ~82. .6 at...“ .8265 3.5.528... ... 6.53.

0

 

_. oz_\_~__ 0.~ 0..
b . .

Rod ...
N.

 

362

o._.=_co.oo< c. «.5 as.

 

     

0N.

 

 

.iom

32

.um cw mcmuom mo ovumm upoz.mm N—umz mo ukwgm pmuwemso mmusswmmcmoz .m mgampm

 

    
  

_.~§_\_~._ oe o.~ o
_ _ 8..
o_-
mmnoé as:
m: I o
.8 c. «a 22
Lo.
lulllfllln‘ o
C

 

 

 

 

 

«En—3 scum

33

effect of ion-pairing was observed in the case of PC solutions,
where the change in chemical shift due to ion-pairing was more
than 30 ppm.

For solvents with a medium value of the dielectric constant
(PC, acetonitrile, DMF), the interaction of the magnesium ion with
the halide ion is decreased to such an extent that a magnesium-25
NMR concentration study of the complexed salt results in a constant
chemical shift. '

0n the other hand, for solvents with a low dielectric constant
(THF, acetone), the interaction between the two ions is decreased
but is still noticeable, as shown by Figure 4. The poor solvating
ability of acetone, and of THF had been shown by sodium-23 and

lithium-7 NMR (47, 48).

3.4.2.2. Line-Width - Another interesting feature of
the magnesium-25 NMR signal is the line-width. In the case of strong
solvating agent (i.e., with both high dielectric Constant and high
donor number), the line-width of the signal is small. It is
approximately l0 Hz in water, methanol and DMSO solutions. In DMF,
concentration change from 0.03 to 0.9 fl_increases the line-width
from 30 to 127 Hz.

As can be seen from Table II, the line-widths in other solvents
are large, which also implies contact ion-pair formation. Indeed,
although PC solutions are more viscous than aqueous solutions, this
is not the case for acetone solutions, and, therefore, line-broadening
of the signal must be due to other effects.

It is well known that the line-width of an NMR signal is

related to the relaxation time T; of the nucleus by the equation:

34

Table 11. Line Width of the Magnesium-25 NMR Signal in Different

Solvents

 

 

 

Magnesium Line Width
Solvent Concentration (m) at Half Height (Hz)
Methanol 0.768 12
0.0543 10
THF 0.325 235
0.1248 220
0.0325a 146
PC 0.369 460
0.111 390
0.111a 134
DMF 0.913 127
0.0319 30
0.0319a 15
DMSO 0.1227 37
0.01227 11
Acetone 1.221 3l0
0.1195 215
0.1196a 150
Acetonitrile 2.045 62
0.423 27
0.6303 24

 

 

aWith a mole ratio of iodine to

magnesium equal to or higher than 2.

35

 

 

 

_ 1
Av (3)
1/2 nTE
with
1 = 1 + YAHO
* (4)
T2 T2 2

AH
where g represents the contribution to the line-width due to

 

inhomogeneity of the magnetic field AHO, v is the magnetogyric ratio,
a constant for a given nucleus, and T2 is the spin-spin relaxation
time. Since the observed relaxation time is assumed to be in the
motionally narrowed limit so that T1 = T2, with T1 being the spin
lattice relaxation time, we can relate the line-width of the signal

to the five categories of spin lattice relaxation mechanisms:
dipole-dipole relaxation, chemical shift anisotropy, scalar coupling,
spin rotation and quadrupole relaxation.

We can then express the observed results as follows:

'I =

1 " = . . 1"
_TT 'fifg 1, exp Rl, d1p-d1p Rl

, chem shift an +

R + R + R 5
l, scal l, spin rot l, quad ( )

It has been shown (49) that in the case of quadrupolar nuclide
resonance, the line-width is generally determined by the quadrupolar
interaction, the contributions from the other relaxation mechanisms
being negligible. The interaction Hamiltonian for the quadrupole
relaxation has the form: Hq = 1.0.1 and represents the interaction
between the nuclear spin (I>l/2) and the electric field gradient at

the nucleus. As reorientation occurs, the components of the quadrupole
coupling tensor Q become random functions of time and provide a

relaxation mechanism. We then have:

36

 

R=R= 3 21+3, n2 92 2 (6)
1 2 40 X 1?}21-1)\1 +7) (—hg'g) TC

where n is the asymmetry parameter and (equ)/h is the quadrupole
coupling constant, I is the nuclear spin of the observed species and
Tc is the translational correlation time.

Thus, the line-width of the NMR signal depends on three terms:
n, Tc and the quadrupole coupling constant. When ion-pairs are
formed, the asymmetry parameter and the field gradient (and hence
the quadrupole coupling constant) will change, and we will observe
an increase in line-width. In the absence of ion-pairing, the line-
width will be affected by the solvent cage. In a strongly solvating
medium, the solvent molecules are held tightly around the metal ion,
and the change in electric field around the nucleus is minimal,
therefore, the line-width of the signal is small. 0n the other hand,
when the magnesium ion is surrounded by a less solvating agent, like
acetone or PC, the solvent molecules around the magnesium are held
less tightly, and having more latitude to move, cause an increase
in the inhomogeneity of the field, thus allowing the magnesium ion

to relax faster, which in turn is reflected by a wider line-width.
3.5. DONOR NUMBER CORRELATION

It has been shown previously (50, SI) that in the case of the
sodium-23 and potassium-39 resonances there is a roughly linear
relationship between the magnitude of the chemical shift in a given
solvent and the donicity of the latter expressed by the Gutmann donor
number. Similar plot is shown in Figure 8 for the magnesium-25

resonance, and the data are given in Table III.

37

.20 06

...mnsaz cocoa 5555 ..EE. 1m.» 3539.... maotm> ..., p.25 $3.55 3:533:32 9.3.2.3 .m 939...

 

 

  

Om ON 0.
_ 4 ”=9: _ O?!
l O
.4025...sz
OM20 O NZOPNUQ
39.".
l m m w
l O.
L V—

 

BEEP—MS

 

38

Table III. Limiting Chemical Shift of Magnesium-25 in Various

 

 

 

Solvents.

Solvent 5 (ppm)a Donor Number Dé:;::§:1c
Acetone 3.0 17.0 20.7
Acetonitrile 14.0 14.1 37.5
DMF 5.5 26.6 36.71
DMSO 3.5 29.8 46.68
Methanol 1.1 25.7 32.7

PC 13.5 15.1 65.0
THF -4.0 20.0 7.58
Water 0.0 33.0 78.9

 

 

a!§_4.0 M_aqueous
susceptibility.

MgClz as external standard.

Corrected for bulk

39
It is readily seen that for a majority of solvents there is a
definite correlation between the magnitude of the downfield chemical
shift and the donor number. Five of the eight solvents studied
seem to fall on a reasonable straight line, but the correlation is
not good in the case of THF, acetone and methanol. For the latter,
a similar behavior was observed in the case of cesium-133 and

potassium-39 resonances.
3.6. COMPLEXATION

Only preliminary studies have been performed in this area.

An aqueous solution of phosphonoacetate and 0.05 M_magnesium chloride at
pH above 8 (so that all the PAA is under the deprotonated form)

gave the following results: for a mole ratio of 0, 0.05, 0.1 and

0.2, the line-width changed from 7 to 16, 25 and 50 Hz, respectively,
and no change in the chemical shift was observed. For higher mole
ratios, the line-width was too large and the signal could not be
detected within a reasonable amount of time.

It is seen from the above results that we have a fast exchange,
compared to the NMR time scale, between the free and the complexed
magnesium ion, therefore, the line-width and the chemical shift
are the average of the free and complexed magnesium signals. If
we define Ta and Tb as the lifetime of the magnesium nucleus in

sites A and B, u and “b as their resonance frequencies and 'l' as:

T = T 1:
(Ta + Tb) (7)

we will have two resonance signals if

1:) 2

nIva-vb) (3)

otherwise, we will have the population average resonance.

40

Likewise, the line-width of the complexed magnesium ion
is large, due to quadrupolar broadening since the ligand, being
unsymmetrical, produces a nonuniform electrical field around
the nucleus.

Similar studies performed with EDTA at pH 9-10 showed a
somewhat different result, the line-width of the peak stays
constant when the mole ratio ligand to metal increases from zero
to close to one. At the same time, the signal becomes weaker and
weaker, and finally for a mole ratio of one or above, it disappears.
These results show that we have a slow exchange, compared to the
NMR time scale, of free and complexed magnesium. The line-width of
the complexed magnesium signal is again much too large to allow its
detection. These results are in contradiction with the ones reported
by Magnusson and Bothner-By (7).

Since the formation constant of the MgZ+-C21l complex in water
and in 95/5 methanol/water mixture was reported as 102'5 and 104'0,
respectively (29), and since C211 and Li+ ion form a complex with
an exchange between the free and complexed Li+ ion that is slow
compared to the lithium-7 NMR time scale, the complexation between
C211 and MgClZ has been studied in methanol. Kauffmann (52)
from carbon-13 MIR studies of 0.1 [1 solutions of C211 reported the
formation of strong complex between magnesium and C211 in 95/5
methanol/water mixture. She found that in the 25°C-85°C temperature
interval, the spectra of a 2:1 mixture of C211 and MgClz contained
two series of signals, one corresponding to the free ligand, while

the other one was characteristic of the cryptate.

41

From equation (8), it can be seen that if the rate of exchange
(r = 1/1) between the free and complexed magnesium is slow compared
to the C-13 NMR time scale, such is not necessarily the case for
the magnesium-25 NMR time scale, since magnesium-25 resonates at a
lower frequency and has a smaller chemical shift range than carbon-13.

Indeed, in the present case, the exchange between the two sites
is fast since only one signal is observed. The line width increases
from 10 to 22 Hz for the solutions of 5 x 10'2 11 MgC12 with a mole
ratio of ligand to magnesium of O and 1, respectively. The change
in chemical shift, however, was small - from 1.1 to 0.7 ppm and,
therefore, could be attributed to a random experimental error.

As an aside, the perchloric acid salt of C211 is insoluble in
methanol. This property could be used to purify C211 or to recover
it.

Another complexation study with lZ-crown-4.(12-C-4, which has
a cavity size comparable to the size of the magnesium ion) in DMF
has been performed. In this case, no change in either line-width
nor in chemical shift was observed. Either there is no complexation
between Mgz+ ion and 12-C-4, or both free and complexed ions have
the same chemical shift. The first hypothesis seems more reasonable
in view of the solvating properties of DMF (e = 36.71, donor number =
26.6).

On the other hand, at this early stage of development of
magnesium-25 NMR in nonaqueous solutions, only qualitative results
can be obtained as far as complexation is concerned, since there is

no quantitative results concerning the ion-pairing in these solutions.

42
3.7. I.R. STUDIES

Some I.R. studies of 2.0 M_solutions of MgBr2 in acetonitrile
showed a behavior of the solvent similar to the one observed by Janz
et al.(53)by Raman spectroscopy in the case of silver ion. The
splitting of the solvent bands corresponding to the C-C and CEN
symmetric stretches (a]) for complexed solvent molecules is shown
in Table IV. The splitting is due to the presence in solution

of both free and complexed acetonitrile.
3.8. CONCLUSION

We determined in this study that the chemical shift range
of the magnesium-25 NMR signal is larger than 30 ppm, and that the
NMR of magnesium-25 can be a useful tool to study solvent effect
and ion-pairing of magnesium salt in nonaqueous solutions. The
line broadening and chemical shift change have been explained
by quadrupolar relaxation in terms of ion-pairing and solvent-

magnesium interaction.

43

 

 

Aceveepom Ncmmz.m Nm~.~v

 

_-eU OF-mom~ .ommm .oemm _-EU mmm .epm .owm upmcumeepee< eexepeEOU
2-58 omNN .oeNN F-ea m_m epwcuweeeee< mesa
eeeetem ewcpeseam zmu sebacem uwceeeexm 8-8 eeweapom

 

 

.mpwepwcoumu< nmxmpasou uco amen we

mmwocmzamcm pmcomuwenw> .>H mpnmh

PART II

PHOSPHONOACETIC ACID STUDIES

44

CHAPTER 1

HISTORICAL REVIEW

45

46

1.1. INTRODUCTION

Just as the use of antibiotics revolutionized the treatment
of bacterial disease in the 1940's, new antiviral agents promise to
revolutionize the treatment of viral diseases in the 1970's and
1980's. For the first time, man will be able to combat viral
infections, halting the replication of viruses and preventing
their spread. The advent of this era may well be one of the most
important milestones in the continuing battle against infectious
diseases.

The genesis of the antiviral era has been quite different from
that of the antibiotic era. The latter was launched by the discovery
of penicillin, a relatively broad spectrum antibiotic whose value
was readily apparent. The antiviral era is being launched with a
number of narrow-spectrum agents whose value has been more difficult
to establish. Antibiotics were introduced at a time when only a
limited amount of testing was necessary to introduce a new drug to
the market and when the need to treat wounded soldiers during a
major war greatly accelerated that testing. Antiviral agents are
being introduced at a time when consumer safety is the paramount
concern, necessitating a great deal of expensive, time-consuming
clinical testing.

The development and testing of antibiotics was largely
subsidized by the federal government because of the national
emergency, and the first antibiotics were quite profitable for the
companies that put them on the market. The development of antiviral

agents, in contrast, has been financed by the drug companies with

47
only limited support from the government, and no company has
made a profit on one. In some ways, it is remarkable that the

antiviral agents have been developed as rapidly as they have been.

1.2. PHOSPHONOACETIC ACID

Herpes viruses are a large family of deoxiribonucleic acid
(DNA) viruses that are the causative factor of diseases, such
as cold sores of the mouth, genital lesions, eye infections,
varicella (chicken pox) and shingles. By some estimates as
many as 10% of all Americans over the age of 18 have recurrent
Herpes infections three or more times per year, and more than
70% of Americans are thought to have antibodies in their blood
that indicate a prior Herpes infection. Some types of persistent
Herpes infections are believed to be associated with initiation
of cancer.

Few antiviral drugs exist against Herpesviruses, and most
of them are nucleoside analogs, and, therefore, they are potential
teratogenic agents (i.e., they could cause birth defects if
injected during pregnancy). Phosphonoacetic acid (PAA), which
is not a nucleoside analog, and has been found active against
quite a few different Herpesviruses is, therefore, a very
promising compound.

Phosphonoacetic acid (Chemical Abstract #4408-78-0) is an
organophosphorus compound (see Figure 9), it has an inhibitory
effect on the replication of Herpesviruses, and the mode of

inhibition is thought to involve complexation with magnesium ion.

48

.umu< umcowaogqocosamogalm ucm .ku< upcownogqococqmogmim

.uwo< owELomococamosg .uwo< owuwomococamosa

IO IO

1 _
suumxuumxo .. 4.0:
o =

O
0.04 U_Zo.n_0mn_OZOIn_mOIn_lm

IO
IO _

mu-zo-e-o:
mu m” _ =
180

0.04 U_ZO_QomQOZOIQmOIQ1N

”mcmu< upcosamosm maowgm> mo mmgzuuzgum .m wczawu

10

...o, _
.Ahvnu nuiunvv+

o __

o

«6.1.x: 0.04 0.2m0u0201amOIa

IO

IO, N . _
So... IUunTOI

. 2.

A43": 904 Ochoa. OZOIQmOIa

49

This triacid, as well as several of its salts (calcium,
barium, copper, manganese, zinc and silver), were first synthesized
in 1924 by Nylen (54). Several syntheses have been proposed
since then (55). Phosphonoacetic acid is a triacid; it is a
white solid, with a molecular weight of 140.03, and has a sharp
melting point at l42-l43°C; it will also be referred to as
phosphonoacetate (PA), its completely deprotonated salt.

1.3. ANALYTICAL TECHNIQUES

Several infra-red (IR) studies of PAA or its derivatives
have been published in the literature (56-59). Its trimethylsilyl
derivative has been studied by gas chromatography-mass spectroscopy
and by gas chromatography (60); a single peak is obtained when
examined in polar (0V-17) and nonpolar (SE-30) gas chromatographic
phases. The above study is of importance for the biochemist, and
the technique has been used to monitor the amount of PAA in blood.
Phosphonic acid derivatives have also been separated by thin layer
chromatography (61 ) .

It has been found (62) that phosphonates can inhibit the
transfer of copper ions or c0pper complexed ions to a mercury
electrode because of their surface active properties. Investigations
have been conducted on the adsorption of PA on hematite, silica and

graphitized carbon (63, 64).
1.4. NUCLEAR MAGNETIC RESONANCE STUDIES

Nuclear magnetic resonance studies involving the measurement

of the carbon-l3 or phosphorus-31 chemical shift or of the coupling

50
constants between C and P or P and H have been conducted on PAA
or its ethylester. For carbon-l3 NMR studies, the acid was
dissolved in methanol (65) and the chemical shifts vs TMS are
35.8 and 168.6 ppm, for the CH2 and the COOH carbons, respectively,
while the coupling constant between C and P is 132.0 Hz. The
ester has been studied as a neat liquid (66).

Some phosphorus-31 NMR studies have been performed on PA
(67). The solution was basic (pH=l4) with tetramethylammonium
hydroxide (TMAH). The chemical shift reported !§_85% phosphoric
acid (contained in a capillary tube inside the sample tube) is
-l3.l ppm, no indication being given whether it is a downfield
or an upfield shift.

Riess §t_al, (67) found that a several fold change in
either phosphorus or alkalinity concentrations caused less than
0.2 ppm change in the chemical shift. No mention was made by
them of a triplet signal due to the coupling between the P and
the CH2 hydrogens. Complexes between lithium and two PAA esters
have been studied in tetrahydrofuran (THF) (68).

1.5. ACIDITIES AND ACTINIDE COMPLEXES OF PAA

Aside from these industrial uses, all the other studies
made so far were concerned with either the extraction of rare
earth elements, or biological applications. It is interesting
to note that most of the studies concerning the biological
applications of PAA have been published in the last five years,
and that they represent approximately 80% of all the studies

ever performed on PAA.

51

Organophosphorus compounds have been found to be good
extractants of rare earth elements (69), and more specifically,
Elesin _$“El- (70, 71) investigated the complexation of PAA with
americium, curium and promethium ions, as well as determined
the acidity dissociation constants (pK‘s) of PAA.

Several comments can be made about this work: the dissociation
constants of PAA were determined against a background of ammonium
perchlorate at an ionic strength I = 0.2, although ammonium
perchlorate is known to be a weak acid. Moreover, sodium hydroxide
was used as a titrant in this research, and since it has been
shown (72, 73) that phosphate ions form weak complexes with alkali
metal cations, the values of the pK's obtained (1.14, 4.33, and
7.31 at ionic strength 0.2, which, extrapolated to zero ionic
strength gives 1.37, 4.84, and 8.50) are probably too small.

The investigation of the complex formation of Am3+, Cm3+,
and Pm3+ was conducted by the method of ion exchange on the cation
exchange resin KU-Z, also against a background of ammonium
perchlorate, at 0.2 ionic strength. Only protonated complexes
were studied, and the log Kf values extrapolated to zero ionic
strength are 2.75, 5.15 and 3.3 for the M(H2L)2+, M(HL)+ and
M(HL)§ complexes, respectively (LH3 represents the freePAA).

The formation constants were found to be equal for the three
cations.

Mao £3 11. (74) reported pK's of PAA of 2.6, 5.0 and 8.2 for

the carboxylic and two phosphono groups, respectively. Their

correlation between pK's and acidity functions is most probably

52
incorrect (see p 74), besides, the conditions under which those

values have been determined are not specified.
1.6. INDUSTRIAL APPLICATIONS

The practical applications of PAA or its esters are numerous,
and several patents mention it for widely different purposes: it
has been used as a flame retardant in thermoplastics (75), as a
corrosion inhibitor for iron-containing water conducting systems
(76), to prevent the formation of spots in photographic materials
due to the presence of particles of heavy metals or their
derivatives (e.g., iron and rust) (77), as an insecticide (78),
as an herbicide (79), or a plant growth regulator (80).
Phosphonoacetic acid has also been proposed as a treatment against

warts (81), or Herpesviruses (82, 83).
1.7. BIOLOGICAL PROPERTIES OF PAA

The popularity of PAA in the biochemical field lies in its
ability to inhibit the replication of Herpesviruses. Three good
reviews have been, or are going to be published on that matter

(84-86).

1.7.1. HERPESVIRUSES - Herpesviruses (Herpesvirus in the U.S.,
Herpes virus in England) is the name of a genus of viruses that
cause diseases affecting both man and animal. Some of these diseases
are fatal, and it has been suggested recently that Herpesvirus may

be directly or indirectly associated with cancer in several animals.

53
Herpesviruses have also been indirectly associated with naso-
pharyngeal and cervical carcinomas in humans [carcinoma: solid
tumor derived from epithelial tissues,is the major form (85%)
of human cancer].

Other types of Herpesviruses are: Equine Herpesvirus,
Varicella-zoster virus (causes chicken pox, shingles), Marek‘s
disease Herpesvirus (the main cause of death among chicken in
commercial poultry, and the reason of the creation of the Poultry
Research Laboratory at M.S.U.), Epstein-Barr virus (responsible
for infectious mononucleosis), Cytomegalovirus (among the
diseases: pneumonitis, Cytomegalovirus mononucleosis), Herpes
Simplex virus [causing cutaneous lesions (Herpes Dermatitis),
mouth sores, cornea infection (Herpes Keratitis: responsible
for an estimated 18,000 cases of blindness each year in the
United States), venereal diseases]. The list is far from

being exhaustive.

1.7.2. HERPESVIRUS DNA POLYMERASE - It has been found for
several types of Herpesviruses that the virus induces its own
DNA polymerase (DNA polymerase is an enzyme involved in the
production of DNA). Three general classes of DNA polymerases are
found in cells of all vertebrates: DNA polymerase a, B, and y.
The greek letters designate their order of discovery: 0 and B
are the most important ones. Herpesvirus induced DNA polymerases

have been reported for several Herpesviruses.

1.7.3. HERPESVIRUS MODE OF ACTION - The mode of action of

 

Herpesvirus is as fOllows: the virus enclosed in a capsid protein

54
comes in contact with a cell, at this point the protein opens
and the viral DNA penetrates into the cell, it then travels
through the cell to the nucleus, enters it, and uses the nucleus
system to produce ribonucleic acid typical of the virus (RNAHV).
The RNAHV leaves the nucleus, and following the usual pattern,
forms a DNA polymerase typical of the virus. This DNA polymerase
goes back to the nucleus where it creates its own DNA (DNAHV);
from this point on, the loop is closed, the production of DNAHV
increases a lot, and within a few hours or a few days, depending on

the virus, the cell dies and many new viruses are then released.

1.7.4. INHIBITORY POWER OF PAA - Leinbach gt_gl, (87)

 

proposed two solutions to explain the inhibitory power of PAA on the
Herpesvirus of turkey induced DNA polymerase. The basic pathway used
by the enzyme to build up the Herpesvirus specific DNA is shown in
Figure 10.

In the first step, the DNAHv binds to the enzyme, in the second
step deoxynucleoside triphosphate (dNTP) binds to another site of the
enzyme, upon which the length of the viral DNAHV molecule is increased
and inorganic pyrophosphate (PPi) is formed. In a third step, PP,
is released and in the final step DNAHV is released. The loop is
closed, and the procedure can start again. Actually, in each cycle,
the DNAHV molecule is not exactly increased by one link, but rather
the two complementary strands of DNA are separated and each forms
the template for synthesis of a complementary daughter strand.

From the results of their kinetic studies, Leinbach gt_gl, (87)

proposed the mechanism presented in Figure 10. In the presence of

55

.42.. ..e 5:2 .3 so: 2; 205.5 55 .2 23:

11......m
...-22¢ a:

ll m
epz. . . .=:<zo

 

 

 

 

 

 

 

1. <20 L <e..:.\ A
...

A .l (s 1.

...-.(20m
A r

fithns .3 (51m52sm ‘20-: as
.l (20
fi =
V x (mu-.22. x V

 

 

 

 

 

...m. , .. . m

(26...

56
PA, an alternate pathway exists in addition to the basic polymeri-
zation pathway. Phosphonoacetate binds to the polymerase at the
PP: binding site (believedly via complexation with magnesium ion)

and is thus a competitive inhibitor of PPi in the exchange

DNA(n+1)
PA

complex or may undergo reaction with the nucleotide at the 3' end

reaction. Phosphonoacetate may simply dissociate from the E

of the DNA primer chain to yield the postulated nucleotide,

DNA

dNMP-PA, and E (dNMP: deoxynucleoside monophOSphate). Thus,

PA inhibition occurs because the EDNA(n+])

complex is diverted
by PA from the main polymerization pathway into an alternate
pathway.

The kinetics results are not consistent with the second
scheme proposed - PA being a simple dead-end inhibitor in which

it could only bind and dissociate from the EDNA<n+1)

complex, but
are consistent with the first scheme proposed.

To fully accept this scheme, however, the postulated
nucleotide, dNMP-PA, must be identified in the reaction mixtures.
Attempts to do so have not proved successful.

From the alternate pathway scheme it can be seen that PA and
PPi inhibit DNAHV synthesis in an analogOus manner. Studies
performed by Leinbach gt_al, (87) confirmed this fact; the
apparent inhibition constants for pyrophosphate, however, are two
to three orders of magnitude greater than those for phosphonoacetate,
that is, the amount of PPi required to produce 50% inhibition is

100 to 1000 times larger than the amount of PA required for the

same result.

57

1.7.5. DRUGS AGAINST HERPESVIRUSES - In the case of
Herpesviruses, several drugs having an inhibitory power have
been studied, although only one is currently used commercially
in the U.S., and two or three outside the U.S. (see references
88 and 89 for a good review of most of them). These drugs are,
however, usually specific for one or two types of viruses only,
and may also be incorporated into host DNA. The potential for
genetic damage, infertility or even carcinogenesis nay not be
negligible.

At least some of these compounds have been compared to PAA.
Ara-A (9-B-D-arabinofuranosy1 adenosine) is about as effective as
PAA against Herpes Simplex virus type 1 (90), while Ara-A was
unsuccessful (91, 92) in clinical trials in humans with Herpes
Genitalis. In rabbits, Ara-A and PA were equally effective in the
treatment of Shope fibroma (93), while idoxuridine and PA had the
same effect in treatment of Herpes Keratitis in rabbits (94, 95),
however, idoxuridine has been shown to be carcinogenic and
teratogenic. Tilorone hydrochloride (96) was found uneffective
against Herpes Simplex virus type 1 in mice as opposed to PAA.
Against Cytomegalovirus, Ara-A was also found less effective than

PAA.

1.7.6. TOXICOLOGICAL EFFECTS OF PAA - Thus, all the drugs
other than PAA that have been used against Herpesvirus are usually
specific of one virus only, and may be or have been proved to be
carcinogenic or teratogenic because of their mode of action. The

best solution would be a compound that inhibits DNA synthesis by

58
a direct interaction with viral DNA polymerase. This solution is
represented by PAA. However, PAA may have some toxic effects of
its own, and its clinical drug usefulness is uncertain.

Meyer gt_gl, (95) reported that intravenous PAA in a
concentration of 300 mg/kg produced severe tetanic muscular spasms
in rabbits, often resulting in death; the same dose given to mice
orally or intraperitonally was well tolerated (97, 98).

Roboz §t_gl, (99) developed an analytical technique for the
determination of PAA in blood. The method is based on the monitoring
of selected ions of PAA and phosphonopropionic acid (internal standard),
formed in the chemical ionization (methane) source of a combined gas
chromatography-mass spectrometer system. The drug is quantified in
blood after removing proteins and lipids as the trimethylsilyllated
derivative. The detection limit is 20 ng/ml when 0.2 ml of serum is
analyzed. The technique is applicable to serum from mice, rabbit,
monkey and man.

In the case of mice, oral administration of 500 mg/kg drug
results in slow increase of blood levels (to 50 pg/ml in one hour)
followed by a drop to 2 ug/ml by two hours. A dose of 500 mg/kg of
PAA given subcutaneously kills monkeys in two days and results in
blood levels in the 0.5-3.0 mg/ml range. A total dose of 100 mg/kg
given subcutaneously in 25 mg/kg doses (every two hours, four times)
is well tolerated and yields blood levels in the ug/ml range. Owl
monkeys with kidney disease develop higher blood levels than squirrel
monkeys, but both yield levels down to 2 ug/ml in 24 hours. A total
dose of 1000 mg/kg a day is well tolerated when infused continuously

resulting in a slow increasing blood level (from 155 to 400

59
pg/ml) which decreases to about 55 ug/ml 24 hours after terminating
infusion.
For therapeutic trials, a minimum of 50 ug/ml continuous blood
level is recommended. This is achievable with a dose of approximately
230 mg/kg a day.

14C, determined the amount of PA

Bopp gt 31. (100), using PA-
retained by the different parts of the body.

Orally administered PA was poorly absorbed in rat (14%, 100
mg/kg), rabbit (2%, 20 mg/kg) and monkey (8%, 20 mg/kg) as determined
from urinary excretion. Absorption averaged 60% in dogs given
1 mg/kg, but was quite variable. The total recoveries in seven days
after i.v. administration of the compound (20 mg/kg) accounted for
104% in adult dog, but only 88% in rat and monkey, 66% in rabbit
and 62% in puppy, suggesting that a substantial amount of 14C
remained in the body.

Whole body autoradiography studies in rats demonstrated that
the 14C was retained in bone. Seven days after i.v. administration
of the drug (20 mg/kg) the levels of radioactivity expressed as PAA
in dry femur were 55, 62, 24, 4, and 57 ug/g in rat, rabbit, monkey,
adult dog and puppy, respectively. During a 25-50 day period after
i.v. administration (20 mg/kg), the elimination of 14C was at least
biphasic, with half-lives of 5.8 days in rat, 8.7 days in rabbit,
and 13.9 days in both monkey and dog. The levels of PAA in dry femur
on day 50 were 9.6 09/9 in rat, 24.6 09/9 in rabbit, 0.8 09/9 in
monkey and 3.0 09/9 in dog. Long-term retention studies in rabbits

demonstrated the persistence of 14C in bone for over 200 days.

60
1.8. PHOSPHONOACETIC ACID DERIVATIVES

For the preceeding reasons and to try to obtain a better
antiviral agent as well as to determine the mode of action of PAA,
derivatives of PAA have been tested.

Herrin gt 21, (101) and Lee gt_al, (102) have studied the effect
of PAA derivatives on some Herpesviruses. Esters of the acetic function
of PAA eventually proved to be inhibitors1nyerpes Simplex virus-induced
DNA polymerase, but always to a smaller extent than PAA.

Replacement of either the carboxylate or the phosphonate moiety
resulted in a loss of effectiveness of the compound. An amino,
methylamino or phosphono group cannot substitute for the carbonyl
group, neither can a carboxyl or a sulfo group for the phosphono
group.

The only two other phosphonates active against Herpesvirus
besides PA are 2-phosphonopropionic acid (see Figure 9) which is
about 50 times less active, and phosphonoformic acid (PFA, see
Figure 9) which has approximately the same efficiency as PAA.

Since PFA decomposes in acidic medium (103, 104), it seems
to have a better future than PAA. Phosphonoformic acid was first
synthesized in 1924 by Nylen (105), who also reported the following
salts: Zn, Mn, Cu, Pb and Ag.

Esters of PFA have been used in oil (106) because they provide
extreme pressure lubricity in lubricating oils and are useful as
hydraulic fluids.

According to Warren and Williams (104), the pKa's of PFA are
0.49, 7.27 and 3.41 for the phosphonic and carboxylic moieties,
respectively. Nylen (103) reported 0.39 and 3.18 in 4 M_NaCl.

61

The crystal structure of trisodium phosphonoformate (PF)
hexahydrate has been determined (107), it consists of sodium
ions surrounded octahedrally by six oxygen atoms, some from
water' molecules, some from the PF ions.

So far, very few studies have been performed on the antiviral
action of PFA. Reno gt al. (108) found that it was at least as
effective as PAA against Herpes virus of turkey and Herpes Simplex
virus, but PFA was not found to be an effective inhibitor of a PAA
resistant mutant of the Herpesvirus of turkey, nor of its induced
DNA polymerase.

These facts suggest that PFA and PAA have basically the same
biochemical properties and mode of action against Herpesvirus
incuded DNA polymerase, aside from the acidic decarboxylation reaction

of PFA which makes it a very promising clinical drug.

1.9. CONCLUSION

It is seen from the above that the physicochemical properties of
PAA are not well known and even its acidity constants have not
been accurately determined. 0n the other hand, it is an interesting
molecule from the point of view of its acid-base equilibria and
its complexing abilities. In addition, it appears to have very
important physiological and even therapeutic prOperties. It was
of interest to us, therefore, to investigate in some detail the

acid-base and the complexing behavior of PAA.

CHAPTER 2

MATERIALS AND METHODS

62

63
2.1. REAGENTS

Phosphonoacetic acid (PAA, from Richmond Organics) was
dissolved in glacial acetic acid (22 g of PAA in 100 m1 of
glacial acetic acid) at a temperature close to its boiling point,
and cooled down to room temperature. Crystals obtained after this
recrystallization were rinsed with ethyl ether. The product was
then dried at 60-70°C in the vacuum oven for at least 24 hours.

The melting point is l42.5—143.5°C (stem correction included, heat
rate: 2°C/minute), in agreement with the literature value (54):
l42-143°C. Elemental analysis (Spang) gave the following figures:
calculated/experimental: C: 17.15/17.16; H: 3.60/3.55; P:
22.12/22.l3. A mass spectrum using the field desorption technique
gave the following peaks: H0(0) P+(CH2COOH); +P(OH)3 (CHZCOOH)2
and [(0) P(0H)2 CHZCOOHJZ HT.

Tetramethylammonium bromide (TMAB, from Eastman organic
chemicals, technical grade) was purified in the usual manner (109)
by recrystallization from methanol (100 g in 1 to 1.5 1 of boiling
methanol, then cooled down with a dry ice acetone mixture), followed
by two precipitations from methanol solution by the addition of
ethyl ether (dissolution of the precipitate in 3 liters of methanol,
filtration, addition of 4.5 lb of ether, filtration, drying of the
precipitrate at room temperature under vacuum for 24 hours, and
repetition). Tetramethylammonium bromide was then dried under vacuum
at room temperature for 48 hours. The compound decomposes slowly
with time and/or heat and had to be repurified every two to three
months. The melting point is higher than 300°C [literature (110, 111):

decomposition above 240°C].

64
The other chemicals used have already been described in

Part 1.2.1.
2.2 RESINS

The anion exchange resin was Dowex 1x2, 50-100; the cation
one was Dowex, 100-200; 50-X-8. The anion transfer membrane used
for the coulometric titrations was from Ionics, type 103-PZL-183,
and was treated with KBr before use to convert it to the bromide
form. The membrane should be always kept wet, otherwise, it becomes

porous and useless.
2.3. SPECTROSCOPY

2.3.1. VISIBLE SPECTROSCOPY - Absorbances in the UV visible
range were measured with a Cary 17 double beam spectrometer, and

the base line has been adjusted through the entire range.

2.3.2. NUCLEAR MAGNETIC RESONANCE SPECTROSCOPY - The
instruments used to observe the sodium-23, magnesium-25, phosphorous-
31 and proton NMR spectra have been described previously (Part 1.2.2.2.).
For carbon-l3, a Varian CFT-ZO Fourier transform NMR spectrometer
was used. The instrument is equipped with a computer controlled
pulse generator and data collection. The field strength is 18.682
kG and carbon-13 resonates at 20 MHz. When an external reference
is used, it is enclosed in a capillary tube coaxially centered in
the 8 mm 00 NMR tube containing the sample solution; otherwise, an
internal reference (dioxane) is used. The chemical shifts are
reported v§_TMS. For carbon-13 and proton NMR, a positive chemical

shift is downfield of the TMS.

 

‘/_“rt

65
2.4. CYCLIC VOLTAMMETRY

Cyclic voltammograms have been measured with a PAR model
174 polarographic analyzer and recorded with a Houston omnigraphic
2000 recorder. The solutions were contained in a cell consisting
of two compartments separated by a frit. The solution on both
sides was at the same ionic strength (0.4 M_with TMAB). The
solutions were degazed for 15 minutes with a stream of nitrogen.

The electrodes were as follows: reference electrode:

thallium amalgam (Ecalomel - Eref = 840 mV); counter electrode:
platinum wire; working electrode: mercury dropping electrode.
The sweeping rates were 50, 20 and 10 mV/s. Triplicate runs

were made for each solution and for each sweep rate. Usually

the average of the maximum and minimum (frontward and backward
peaks) were the same, and were independent of the sweep rate.

The half wave potential of the metal ion for a given solution was

obtained by averaging all the maxima and minima of the different

spectra.
2.5. POTENTIOMETRY

The potentiometric measurements were done with a line powered
voltmeter (Analogic), #AN 2546 (i 2 V at i 0.1 mV). A high impedance
buffer was built by the electronics shop to connect the electrode to
the voltmeter. The following electrodes were used: Sargent Welch
300 72-15 (combination pH electrode, with a thallium amalgam
reference, which is more stable at high temperature than the usual

calomel reference), and Corning NAS 11-18 (sodium ion electrode)

66

used with a calomel reference electrode. The Sargent Welch
electrode was replaced after 3 months, because its response
became unstable. .

Measurements made outside and inside of a Faraday cage
showed that 60 cycle noise was up to 80 mV peak-to-peak outside
and less than 1 mV inside the cage. Static noise was detected
by random potential fluctuations, up to 40 mV outside and down
to less than 0.1 mV inside the cage. Static noise is due to
the operator and to charges building up on the instruments. Cyclic
and static noise were picked up by the cables bringing the current
from the coulometer to the electrodes, resulting in potential
fluctuations of up to 0.8 mV. These fluctuations completely
disappeared when shielded cables were used.

The titrations were performed in a thermostated cell, 3.5 cm
1.0. and 9.0 cm in length. The temperature was maintained constant
by a water bath. The cell is closed with a teflon top, with an
inlet for a stream of nitrogen. Before reaching the cell, the
nitrogen goes through two bubblers containing an aqueous solution
of Co(II)SO4 and a slightly basic solution of BaCl2 to eliminate
oxidizing impurities and C02. Before each titration, the solutions
were temperature equilibrated for at least 30 minutes. The
solutions were stirred with an air driven magnetic stirrer. After
the addition or generation of the titrant, the solution was stirred
for 30 seconds, and the voltage was recorded after 1.5 minutes;
therefore, it took two minutes to get one data point.

In the case of the pH measurements, extensive use was made

of a coulometer described in the next section. The anode compartment

67

of the coulometer consists of a 12 mm 00 glass tube, 15 cm in
length, with a piece of anion transfer membrane at the end. The
membrane is held tight with a teflon cap (16 mm 00, 20 mm long,
with an 8 mm 1.0. hole at the end), and an 0 ring. Care must be
taken not to have an air bubble against the membrane when putting
the compartment in solution, otherwise, the electric contact is
not possible. When working at temperatures higher than 35°C,
a slightly larger glass compartment was used, since teflon has
a slightly larger expansion coefficient than glass.

The two electrodes connected to the coulometer consisted of
a platinum rectangle (17 x 10 mm) and a platinum grid (30 mm in
length, 32 mm in diameter). After each run, each electrode was
cleaned by immersion in 6 M_HNO3 and generation of current for
60 seconds using each electrode in turn as an anode and as a

cathode.
2.6. COULOMETER

The base in the potentiometric experiments was generated via
the use of a constant-current coulometer. A block diagram is shown
in Figure 11. The scheme for the different parts of the coulometer
was kindly furnished by Dr. C. G. Enke's group, Marty Rabb (electronic
designer) and the electronic shop. The instrument itself was assembled
with Dr. J. Hoogerheide.
The time base consists of a Crystek 1 MHz crystal oscillator
and a Mostek MK 5009 counter time base circuit. The gate is formed
of NAND gates (two SN 7400, Texas Instruments). The pulse generator

contains a 555 timer chip (T.I.). Five of seven 7490 decade counters

68

.emumsopaou acmggau acoumccu on» $6 Eagmo_o xuopm

 

 

 

 

 

 

 

 

 

 

 

 

>385 1 1 1
m2: boarL 78853:
5.238 - 250m
088 E38 1 .sz53

mmqm m2: _ H mofimmzmo

.35ch _1rm»<o: mmbaa :54;—

 

 

fibEquW
723m

 

 

 

 

 

not". thm 1
mpqmmjg

 

.Fp mesa?“

69
were multiplexed to provide a five-digit display. The current
source consists of a 741 CP operational amplifier, with a
temperature compensated Zener diode to provide an accurate and
stable voltage reference. An Intersil 00 191 solid state
switch was used to decrease the residual current between
current generations. The power board converts 115 V, 60 Hz
into i 15 & 5 V D.C.

The general mode of operation of the instrument is the
following:

Operational mode: activation of start (by depressing the
start switch) causes generation of a pulse (+5 V) by the pulse
generator. This pulse turns on the current until the phase
returns to a logical zero (0 V). The gate is simultaneously opened,
allowing the time base oscillations to be counted, thus giving a
measure of time which the current was on. In the operational
mode, the time base frequency is 104 Hz.

Calibrate mode: manual opening and closing of gate allows
102 Hz time base to be counted. Readout is in seconds. The
calibrate mode is used to count clock pulses up to 20,000 against
a reference time source. The total of 20,000 seconds allows
considerable error in the start and stop actions (1 ppt corresponds
to 20 seconds).

The characteristics of the coulometer are as f01lows:
current rise and fall time on the order of 1 us, residual current
of approximately 2 nA, current stable to 11 ppt and time base
accurate to better than 0.5 ppt. Further information can be

obtained in the operating manual (133).

70
2.7. SAMPLE PREPARATION

2.7.1. CYCLIC VOLTAMMETRY - In order to prevent the

 

precipitation of Cd(OH)2 for this experiment, a solution of
Cd(N03)2, TMAB and PAA was titrated by TMAH up to a pH of 11
(all the ligand was then in the phosphonoacetate form). The

concentration of cadmium ion was 5.046 x 10'4.

The ionic strength
of the solution was high (0.4), so that the solution with a ligand
to metal mole ratio of 100 could be obtained without changing the
ionic strength. Eleven solutions were run at mole ratios from 0
to 100 by increment of 10. The solutions were deoxygenated with

a flow of nitrogen for 15 minutes prior to run.

2.7.2. SODIUM ION ELECTRODE MEASUREMENTS - Sodium ion
measurements have been performed by using a Corning NAS 11-18
electrode. The calibration was performed by titrating 10 ml of
1.0 [1 basic TMAB solution with 30 ml of a basic solution of 0.02
M NaCl-l.0 M TMAB. Tetramethylammonium hydroxide was used to
obtain the basic pH. The above procedure gave a potential-sodium
ion concentration calibration curve. Then 20 m1 of the 0.015 M
basic solution of NaCl were titrated with 20 m1 of basic PA at
an ionic strength of 1.0. In these conditions, the highest mole
ratio of PA to sodium ion attained was 10. These titrations were

performed in the Faraday cage.

2.7.3. pH MEASUREMENTS - Since laboratory distilled water

 

is contaminated with a small amount of morpholine (morpholine

distills with water), tap water 2 x 10'2.M_in KMnO4 and 2 x 10'2 M

71
in KOH was distilled. This distilled water was boiled for at
least 15 minutes to expell the carbon dioxide, and then cooled
down and kept under nitrogen. The distilled water obtained had
a conductance corresponding to 1 ppm of sodium ion equivalent,
and had a neutral pH.

A purified stock solution of 6 M HCl was prepared by the
distillation of an approximately 6 M_HCl solution under atmospheric
pressure. A 7 x 10'2 M_stock solution was prepared from the 6.M
solution, and titrated in the usual way against NaZCO3, with
bromocresol green as indicator. A further dilution brought the
concentration down to 7 x 10'4 M, This last solution was used to
calibrate the pH electrode. The amount of PAA required for each
titration was weighed on a CAHN model 4100 microbalance.

The stock solution of citric acid had to be standardized
against Na2C03 with phenolphthalein as indicator, since a
titration of a known amount of citric acid monohydrate gave an
equivalent weight slightly smaller than the theoretical one. The
magnesium bromide stock solution was standardized with EDTA in the
usual manner.

When PAA alone or citric acid were titrated, 30 m1 of

7 x 10-4

M HCl at the working ionic strength were used for the
calibration of the electrode and 14 points were obtained at
30-second intervals with a current setting of 5 mA. The titration
was stopped before the end point (at approximately 390 seconds)

in order to remain in the acidic medium. This way a plot of

potential y§_proton concentration and not activity was obtained.

72

Then, without removing the electrode of the solution, 5 ml of
4.9 x 10'3 PAA was added, and the solution was titrated coulo-
metrically. Data were collected at 25-second intervals. The
equivalence points in this last titration were determined by the
first derivative method. It has been found that taking the electrode
out of solution to rinse it and putting it back in another solution
can give rise to a change of up to 0.4 mV in potential reading.

It was often found that the experimental final end point
corresponding to the titration of HCl and the weak acid was off
compared to the calculated value. However, the experimental
elapsed time between two equivalence points of the weak polyacid
was always equal to the calculated value (determined from the
number of equivalents measured with the microanalytical balance).

When the complexation of PAA with a metal ion was studied,
the amount of metal ion necessary for the titration was present
in the calibrating solution. The concentration of metal ion was
7 x 10"4 [1 except for sodium ion which was 2.8 x 10'2 11.

When work was done at different temperatures, the volumes
of solution used were corrected by using the values given in the
CRC Handbook (112). The temperature of the solution was measured

and kept constant to i 0.1°C.
2.8. DATA HANDLING

Extensive use was made of the CDC 6500 computer, with
several programs: MINIQUAD 76A for the PAA titrations, and KINFIT4
for the calibration of the electrode and the sodium-23 NMR

measurements with 12-crown-4 (see Appendices B and C).

CHAPTER 3

RESULTS AND DISCUSSION

73

74

3.1. ACIDITY STUDIES OF PAA

3.1.1. SPECTROSCOPIC STUDIES OF PAA - It has been reported

 

in the literature (74) that in the deprotonation of PAA, the
first proton results from the dissociation of the acetic acid
group, while the last two protons come off the phosphoric acid
moiety. The pK's of phosphoric and acetic acid are 2.15, 7.20,
12.40 and 4.75, respectively, while the ones of PAA are
approximately 1.5, 5 and 8 (see p 97). Therefore, intuitively
one would correlate the second pK of PAA to the acetic acid

function.

3.1.1.1. Phosphorus-31 and Proton NMR - The proton
NMR spectrum of the free acid in 020 consists of a doublet at
2.964 and 3.082 ppm, with a coupling constant of 21.2 Hz. The
signal is split due to a JP-H coupling. No signal was observed
for the acidic protons, due to the fast exchange between the
deuterium of the solvent and the protons.

A phosphorus-31 NMR spectrum of 0.3 M_PAA consists of three
peaks because of the coupling between P and the two protons on
the neighboring carbon. The coupling constant has a value of
21.0 t 0.5 Hz, in good agreement with the value obtained from
proton NMR measurements. The chemical shift of the middle peak,
referenced to 85% H3PO4 is -16.0 ppm at pH 1.0. Since it has
been found that PA forms a weak complex with the sodium ion, in
the study of the change in phosphorus-31 chemical shift 1; pH, the

latter was adjusted with tetramethylammonium hydroxide.

75

The result of this study is shown in Figure 12. An upfield
shift (to -l3.2 ppm at pH 3.2) is followed by a downfield one
(to -16.5 ppm at pH 6.5) and a final upfield shift (to -13.3 ppm
for pH 10.5). At pH above 10.5, the phosphorus-31 chemical shift
does not change. These results are in good agreement with the
results of Riess gt a1. (67), who reported a chemical shift of
-l3.3 ppm for a PAA solution at pH 14.

A comparison of the change in chemical shift to the distribution
of PAA species at different pH values (see Figure 13) shows that the
chemical shift change seems to be related to the removal of the
different protons. It seems quite reasonable to relate an upfield
shift to the removal of a proton closely related to the phosphorus.

Crutchfield et_al, (113), however, found a downfield shift of
0.5, 3.0 and 2.5 ppm, respectively, for each deprotonation step of
H3P04. Jones and Katritzky (114) observed the same results and
stated that it was not clear whether shielding of the phosphorus
in the species H3P04, Hzpog, Hpofi‘ and P02" would increase with
the ionization (because of the increasing charge) or decrease
(because of distribution of the same number of electrons over a
larger volume as a consequence of expansion).

It has been found that phOSphorus-Bl chemical shift is related
to p_orԤ_electrons in 90% of the cases and to g_electrons in the
remaining 10%.

In a series of papers and books (115-118), Letcher and van Watzer
formulated a quantum mechanical derivation of the phosphorus-31
chemical shifts using s, p_and g_orbitals and allowing full latitude

in the n bonding. This has been accomplished by using an

76

.Ie.m» <<e.m m.o co Seeem .eepeeeu _m-m=ceeemeee

 

 

 

 

 

 

Ia .m> <43 .8 «.22 em

.mp 323322

77

in w.» <<n_ mo 5533me 8.88% 2: EB Sigm 23.55 5133.385. 9: .3 eaten—=8 .2 953...

 

 

 

.m> 44¢ m0“.
w>m30
ZO_._.Dm_m._.m_n_
mwawam

3:0. 0 wk 0 n v

_ _ _

 

ON
CV
00

0m
00.

o\o

 

 

WE “.0 «=22 mm

m.

3 Egg
0.
m_

 

 

78

artificially chosen coordinate notation which allows the chemical
shift expression to be stated in a particularly simple functional
form, using a minimum number of hybridization parameters. This
derivation has been applied to several forms of molecules,
including molecules of the MPZZT type.

According to this derivation, the chemical shift may be treated
as the summation of o-bond and n-bond contributions. The distinction
made between a and n bonding is that 0 molecular orbitals involve
only §_and p_orbitals while g orbitals alone are involved in the n
bonding. Moedritzer (119) further developed the quantum derivation
of Letcher and van Watzer, and applied it to the case of the
chemical shift change of a series of oxyacids of phosphorus .22 pH.

Thus, for molecules in which the phosphorus has a coordination
number of 4, the phosphorus-31 chemical shift is comprised of a 0-
bond contribution determined solely by the prorbital occupation
and a n-bond contribution corresponding to g_orbitals only. The
chemical shift can be expressed as:

6 = 60 + 60 + 6Tr (9)

where 6 is the chemical shift referenced to 85% H3P04, 60 is the
absolute chemical shift of the reference standard and 60 and 61r are
the o and 0 contributions, respectively.

The theory showed that 61r is negative and is proportional to
the increase in total occupation of the d1r orbitals of the phosphorous,
being independent of the distribution of this total character among
the various bond. For P atoms having 4 neighbors, 61r = -l47nfl,

where n1T is the total number of electrons in the dTr orbitals of

79
the phosphorous atom being viewed by phosphorus—31 NMR. On
the other hand, the prorbital occupation was treated in terms
of several parameters: the angular geometry, consisting of the
molecular symmetry and bond angles, and the polarity of the
bonds.

Moedritzer then proposed that the change in phosphorus-31
chemical shift could be expressed as the sum of small variations
in: l) the effective electronegativity of the neighboring oxygens
due to association of hydrogen, 2) the bond angles and 3) the total
occupation of the g_orbitals. Therefore, according to Moedritzer,
equation (9) becomes:

A6 = CAXO - 147An1r - Aae (10)

where 4X0 is the change in the effective electronegativity of
the oxygens caused by their association with the hydrogen(s), n1T
is the change in the total number of electrons in the d1r orbitals
of the phosphorous induced by variations in the character of the
P-O bond brought about by hydrogen association, A0 is the increase
in 0P0 bond angle caused by association of the hydrogen, and A and
C are numbers which may vary from one oxyacid to another.
In the case of PAA we can write equation (10) slightly

differently:

‘ A6 = CAxo + BAxR ~147Ann -DAn: - A00 (11)

where B and D are constants, AXR reflects the change in electro-

negativity of the organic group (acetic acid group presently),

R
n

of n electrons, and the other terms have been defined previously.

An corresponds to the change in P-O n bond from rearrangement

80

Upon protonation of the phosphoric acid moieties, we have
an increase in electronegativity of the corresponding oxygen
atom, corresponding to a positive contribution to the chemical
shift. On the other hand, we also have feedback of n electrons
from the P-O bond onto the phosphorus resulting in an increase
in n electrons and a negative contribution to the chemical shift.
The change in bond angle has not been considered here and is
probably not completely negligible although of minor contribution.

According to the theory developed by Moedritzer, in the case
of phosphonic acids, we must have a larger contribution to the
chemical shift from the feedback mechanism than from the change

in polarity (while the opposite is true f0r the H3PO case). No

4
quantitative evaluation of each term has been done so far, and

the above argumentation explains why the above theory of phosphorus—
31 NMR chemical shift has been referred to as a "tug of war" (118),
but the theory fits our experimental results.

Studies performed by Riess gt_gl, (67) showed that although

R

the terms involving changes in n1T

or 0 are not negligible, they

are much smaller than the contribution of the electronegativity

of the carboxyl group. Furthermore, no n bonding is involved in

the P-CH2 bond, and the effect of the rearrangement of the n
electrons around the P-O n bonds due to change in electron distribu-
tion of the CH2 group is most probably negligible. We can also
assume that the change in A6 (O-P-O angle) upon protonation of the

acetic acid moiety which is at the other end of the molecule is

negligible.

81

By bonding a proton to the carboxylic group, the electronega-
tivity of the related oxygen increases, thus increasing the electro-
negativity of the two carbons [corresponding to a decrease in
prelectrons, as shown by 13C NMR (see below)] which corresponds to
a positive contribution to the chemical shift as shown in equation
(11). This formalism also applies to the replacement of the
acetic acid moiety by the formic acid moiety as shown in Part II.3.3.1.

No conclusion can be drawn from the change of the coupling
constant JP-H y§_pH because of the precision of the measurements when
JP-H is determined from 31F MR spectra. Whatever is the change,
it will be small anyway since JP-H changes between 19.5 and 21.5 Hz
with a precision of':0.5 Hz. The value is as expected (118) for the
P-H coupling through an intermediate nucleus, i.e., between 15 and
25 Hz. On the other hand, the line-width of the central peak changes
from 3 Hz at pH 1.0 to 8 Hz at pH 6.0 and back down to 3 Hz at pH 9.0.

A study of the chemical shift of phosphorus-31 in the free acid
y§_concentration shows an upfield shift from -16.4 ppm at 1.0 M_to

2

-13.3 ppm at 10' M. If we again take a pK1 of 1.5, a simple

calculation shows that a 1.0 M solution of PAA contains 84% of
triprotonated acid, while a 10'2 M solution contains 80% of
diprotonated acid. According to Figure 13, the chemical shift should
then be below -16 ppm, and at -13.3 ppm for the l M and 10'2 M
solutions, respectively, in good agreement with the experimental

results.

3.1.1.2. Carbon-13 NMR - 0f more interest for the

 

pK-acidity function correlation is the study of the change in

82
carbon-13 NMR chemical shift and JP-C coupling constant v_s_ pH.
Generally speaking, an upfield shift for any NMR measurement can
be related to an increase in §_(diamagnetic) electron density or
a decrease in p_(paramagnetic) electron density. As a rule,
shifts in carbon-13 NMR are mostly related to §_electron density.
However, the PAA case is slightly different, and resembles the
studies performed on the deprotonation of carboxylic acids.

Upon deprotonation, the carbon-13 NMR chemical shift of the
carboxylic carbon moves downfield by approximately 5 ppm, due
to a decrease of the electronegativity of the oxygen, which
corresponds to an increase in p_electron density around the
oxygen atom and, therefore, around the carbon. The decoupled
carbon-l3 NMR spectrum of 0.4 M aqueous PAA shows two doublets,
corresponding to the COOH and the 9H2 carbons (at 172.8 and
173.1 ppm, and 34.4 and 40.7 ppm, respectively at pH 1.0). The
doublets correspond to the P-C coupling, with a larger coupling
constant for the carbon closer to the phosphorus. Figure 14
shows the change in chemical shift !§_pH for one of the peaks
of the CH2 carbon signal, and the values for the four peaks are
listed in Table V.

As expected, we have a downfield chemical shift. When the pH
is increased, however, the change starts with the removal of the
phosphoric acid proton, and the overall change corresponding to
the removal of the three protons is 7.5 ppm for the COOH carbon
signals, and 6.2 and 5.5 ppm for the CH2 carbon peaks. In this

case, again, the pH was adjusted with tetramethylammonium hydroxide.

83

.3... .3» s: m we .8 .523 9...... 2:
ee Amze ee upe_eezeo see ..me 3, £3.33 .eeexe_o e3 emceeceeeav Sewem .eu_eeeu m_-eeeccu .3. eceu.e

 

In
0_ m. m .h m n 3» mm m“ _ 0
.VM1
1. NMI
Emu
- w
1 ON!
4®N1

 

 

 

Io
_
Ia .m> 681901410: ...o «:22 on.
o

Tab1e V.

84

Carbon-13 Chemica1 Shift of 0.7 fl PAA !§_pH

 

 

 

pH -£H2
0.76 173.0 172.7 40.4 34.2
1.48 173.7 173.4 40.9 34.8
2.44 174.4 174.0 41.3 35.4
3.51 174.7 174.5 41.6 35.7
4.46 175.4 175.1 42.1 36.2
5.47 176.8 176.5 43.1 37.2
6.45 177.6 177.3 43.7 37.7
6.93 177.7 177.4 43.8 37.9
8.11 178.3 178.1 44.3 38.4
8.95 179.4 179.1 45.2 39.4

 

 

85
On the other hand, if we consider the JC-P coupling constant for the
ng carbon (see Figure 15), we have a decrease in JC-P from 125.5 to
118.3 Hz for a change in pH from 0.8 to 4.0 followed by a slight
increase from 118.3 to 119.5 Hz (pH changes from 4.0 to 6.5)and a
final decrease from 119.5 to 112.2 Hz when the pH is brought to 10.6.

The JC-H coupling constant has been related by quantum mechanical
calculation to the §_e1ectron density around the carbon and the
hydrogen (120), an increase in the electronic density corresponding
to an increase in JC-H‘ From this assumption, several authors (121,
122) related the JP-C value to the §_electron density around the
carbon. 0n the other hand, Tebby (123) reported that the effects of
e1ectronegative substituents on coupling are different for phosphorous
and carbon. Either they do not increase the §7character of the
remaining bonds, or another factor such as n-bonding is introduced.

Because of the numerous variables and unknowns concerning the
chemical shift and specially the JC-P coupling constant (very little
work has been published on that matter and the theory is still missing
at the present time, presumably because of the lack of data), it seems
unrealistic to try to explain the shape of the JP-C !§_pH curve
quantitatively. Qualitatively, it seems reasonable to relate the
two decreases in the JP-C coupling constant to the deprotonation of
the phosphoric acid moieties.

In conclusion, we have correlated the second pK of PAA to the
acetic acid moiety in the molecule. The change in direction of the
phosphorus-31 NMR signal shift has been explained in terms of changes
in the g_electron density around the phosphorus and of changes in

electronegativity of the different substituents.

.=a_m».<<a co concau sz «no

mo «caumcou mcwpaaou m-ua :_ mmcmgu

 

 

( ...

l

mm mm _ AV

 

 

...
Ia .m> :00???

o
_
no: ...o “To:
0

mm. _
~V__
mw__
mw__

/. afiumw.

mme_
.umw_

.mwmw_

ANIV
n-0n

.m. ae=m_a

87

3.1.2. POTENTIOMETRIC STUDIES OF THEApK'S 0F PAA - Once
the pK's of PAA assigned to the different acidity functions, it
was of interest to determine precisely the pK's of PAA at low
ionic strength and to get the thermodynamic values by extra-
polation to zero ionic strength.

It was decided to determine these values from measurements
performed with a pH electrode, however, because of the reported
complexation (72, 73) between phosphate ions of various types and
the sodium ions, it was decided to use a much bulkier cation for the
base and the supporting electrolyte, i.e., the tetramethylammonium
ion. Some titrations have been performed with tetramethyl-
ammonium hydroxide as a base, but since the tetramethylammonium
hydroxide commercially available is an approximately 1 fl_solution,
problems occurred due to the presence of carbonate ions in the
base which interfered with the determination of the pK's of PAA
(002 cannot be removed in a solution with appH above 7.5 by
degassing). Moreover, the commercial product from Eastman contained
a precipitate soluble in acid.

Attempts to synthesize the base in the laboratory by ion
exchange resins were not successful. The use of a cation exchange
resin resulted in incomplete exchange of the sodium ion, while
the use of an anion exchange resin resulted in only partial
exchange of the bromide ion. For these reasons, it was decided

to generate the base electrically with a home-built coulometer.

3.1.2.1. Coulometric Titrations - The advantages of

that method are the elimination of the preparation and handling

88
of a multiplicity of standard solutions, the increase in the
precision in the amount of titrant used, and in the purity of the
titrant since the solvent is used to generate it.

The reactions involved in the electrode processes are:

2Ho++2e' Z H+2H0 (12)

3 2 2 .
_ + _ cathodic

2 H20 + 2 e + H2 + 2 0H reactions (13)

4 0H" Z 02 + 2 H20 + 4 e' . (14)

+ + _ anodic

6 H20 + 02 + 4 H30 + 4 e reactions (15)
+ - +

02 + 2 H30 + 2 e + H202 + 2 H20 (16)

- -) .-

02+2H20+2e + H202+20H (17)
+ - ->

H202 + 2H30 + 2 e + 4 H20 (18)

H202 + 2 e" Z 2 0H" (19)

As can be seen from the last four equations, oxygen must be removed
from the solutions to prevent formation of H202 and unwanted
cathodic reactions. In order to keep the solutions free from
oxygen, they were kept under nitrogen atmosphere as explained in
the experimental part. In nonbasic conditions, the anodic reaction
will generate protons.

Since the mobility of proton through the membrane separating
the anodic and cathodic compartments is much larger than the
mobility of the tetramethylammonium ion, this would result in
pollution of the solution by an acidic impurity. To remedy that
problem as well as to decrease the ohmic resistance of the system,
concentrated (0.2 to 0.5 fl) solutions of the supporting electrolyte
tetramethylammonium bromide were used in the anodic compartment.

The anodic reactions then become:

89
Br' .2 l/2 Br2 + e" (20)

In these conditions, the increase in negative charges due to the
formation of base in the cathodic compartment is counterbalanced
by a flow of tetramethylammonium ions through the membrane (the
membrane is impermeable to the solvent).

As mentioned previously (see Part II.2.2.7.3.), a titration
of 7 x 10'4 fl_HCl at the working ionic strength is first performed.
The EMF of the cell:

Glass electrode//Acid solution, 7 x 10'4 [1 at/Thallium amalgam

constant ionic strength reference electrode
°. -031 4' _R_T.
15. E - E + F 1n [H ] + Ej + Easym + F 1n yH+ (21)

where E° is the standard potential referred to the thallium amalgam

reference, R, T and F have their usual meaning, E is the junction

3

potential, E is the asymmetry potential and yH+ and [H+] are the

asym
activity coefficient and the concentration of the hydrogen ion. At
a constant ionic strength, the Ej and Easym potentials as well as
the activity coefficient of the hydrogen ion are assumed to be

constant. Therefore, the EMF of the cell can be expressed as:

E = E°' + El x 2.303 log [11*] (22)

Using the computer program PHCALIB-KINFIT4 (see Appendix B),

the experimental points are fitted to the equation:

E = intercept + slope x log ([HT]0 - t : l + residual) (23)

F o

where [H+]o is the initial concentration of HCl (7 x 10"4 M), i0 is
the intensity of the current used to generate the base in mA, t

is the time during which the base is generated in seconds, and

v0 is the volume of acid in ml. Residual is a concentration

90
of acid or base impurity that could be present in the solution.
Initially, the three terms intercept, slope and residual were
refined, but it was found that the last two terms are coupled.
Since the temperature of the solution is constant to 0.l°C, and
pH glass electrodes are known to have nernstian response, the
slope term was calculated from the solution temperature and kept
constant, and only the two other terms were refined.

Once the calibration of the electrode is performed, and in
order to keep the junction and asymmetry potentials as constant
as possible, the acid to be titrated is added without removing
the electrode from the solution.

Using this procedure, citric acid was titrated at an ionic
strength of 0.1 by tetramethylammonium bromide (TMAB). The pK's
obtained are shown in Table VI and compared to literature values.
As can be seen the agreement is quite good when compared to the
tetramethylammonium chloride supporting electrolyte case (the
largest difference between the two sets of values is 0.11 pK

units and the smallest 0.01).

3.1.2.2. (pK's of PAA - The pK's of PAA were then
determined at different ionic strength. The data reduction was
performed by using the computer program MINIQUAD 76A (see Appendix C
for an explanation of the data handling) and the three following

expressions of the K's:

K1 = [Lug] [H+]/[LH3] (24)

K2 = [LH=] [H+]/[LH'] (25)

91

Table VI. Acidity Functions of Citric Acid at Ionic Strength 0.l.

 

 

 

Temperature Medium pK] pK2 pK3 Reference
20 KCl 2.96 4.39 5.67 a
20 KC1 3.08 4.39 5.49 b
25 KN03 2.79 4.30 5.65 c
25 KN03 2.92 4.39 5.72 d
25 KN03 3.04 4.47 5.80 d
20 NaC104 2.87 4.35 5.68 e
20 NaC104 2.96 4.38 5.68 f
30 NaN03 2.94 4.44 5.82 g
25 (Me)4NC1 2.88 4.36 5.84 h
25 (Me)4NBr 2.89 4.32 5.78 this work

 

 

aA. Okac and Z. Kolarik, Collect. Czech. Chem. Commun., 24,1 (1959)

dT. N. Briggs and J. E. Stuehr, Anal.

K. S. Rajan and A. E. Martell,

K. K. Tripathy and R. K. Patnaik, J.

Indian Chem. Soc., 43, 772 (1966)

Inorg. Chem., 4, 462 (1965)

eE. Campi, G. Ostacoli, M. Meirone, and G. Saini, J.
Chem. , 26, 553 (1964)

f

C. F. Timberlake, J. Chem. Soc., 5078 (1964)

Chem., 47, 1916 (1975)

Inorg. Nuclear

9R. C. Warner and I. Weber, J. Am. Chem. Soc., 75, 5086 (1953)

hS. S. Tate, A. K. Grzybowski, and S. P. Datta, J. Chem. Soc.,

3905 (1965)

92
= [L3'1 [H*1/[LH=1 (26)

The relationship between the concentration and thermodynamic

formation constant is for K3:
3-
t_ KCYL YH
K3 27
K311;FF_—— ( )
by using the Guntelberg modification of the Debye-Huckel equation:
log yn = - A zfi «T7(1 + 7T) (28)

with
_ 1.823 x 106 (29)

- (cT)3/2

 

we can rewrite equation (28) as

c t 2 2 _ 2
pK3 pK3 A (ZL + 2H ZLH) 777(1 + «7)

pkg - pkg 6A «T7(1 + «T1 (30)

Similarly we have:

pK; = pK: 4A 7T7(1 + «T1 (31)

pK: A /T/(1 + If) (32)

C
PK]

Half-way between the second and third equivalence points we have:

[TMAE] = 2.5 [PAA] (33)

with [TMAE]: concentration of tetramethylammonium ion corresponding
to the formation of the equivalent amount of base, on the other
hand:
[LH‘] = [L3'] = l/2[PAA] (34)
therefore, we have:
= [S.E.] + 1/2 ([LHE] + 22 [LH=] + 32 [L3'])
or I z [S.E.] + 4.5 [PAA] (35)

93
where [S.E.] is the concentration of supporting electrolyte.
Similarly, at half-way between the first and second equivalence
points:

I = [S.E.] + 2 [PAA] (36)

With a concentration of PAA: 7 x 10’4 M and a concentration of
supporting electrolyte of 0.02, the change in ionic strength during
the titration of the third acidity function corresponds to an increase
of the ionic strength of 15%. This fact explains why the lowest ionic
strength used was 0.02. On the other hand, at an ionic strength of
0.1, a l ppt acid or base impurity in the supporting electrolyte
represents an error of 1.5% when referred to PAA. This factor
itself limited the acid molarity to 7 x 10'4 minimum.

Several comments can be made about the use of equation (31).
First, it has been shown that the validity domain of the Guntelberg
equation extends to an ionic strength of approximately 0.1, and,
therefore, the pK's of PAA were determined at ionic strength lower
than that value. Second, when the Guntelberg equation is used,
it is equivalent to the Debye-Huckel equation with a diameter of
the solvated ion of 3.05 A (at 25°C). Bates (124) proposed to
use a value of 9 R for the proton, but we have different ions
which sizes are not directly available, and, therefore, the use
of the Guntelberg equation seems to be a reasonable approach to
the problem.

The computer program MINIQUAD 76A reduces the data by calculating
the overall formation constants of the equilibria involved, and

first starts by refining the pK value, then the pK2K3 and finally

3
the pK1K2K3, therefbre, the error in the value of this last term

94
is a cumulation of the error on each term. For this reason and
because the first acidity of PAA is fairly strong (pK1 2 1.5),
the extrapolation of the concentration values of pK1 to zero
ionic strength was not possible. The thermodynamic values of
pK2 and pK3 are 5.11 t 0.04 and 8.69 i 0.05, respectively. The
slope of the ionic strength plot is -0.942 and -2.98, respectively,
while the theoretical values are -2.04 and -3.05 at 25°C. The
thermodynamic plots and the corresponding values are shown in
Figures 16 and 17, and Table VII.

A non-negligible factor in these potentiometric measurements
is the fact that the glass electrode aged. At the end of a two-
month utilization period, pK values at different ionic strengths
started to fall consistently below the straight lines of Figures 17
and 18 by 0.05 pK units. At the same time, the readings were
becoming less stable (precision of i 0.2 or 0.3 mV instead of
0.1 mV). Upon replacement of the combined glass electrode by a
new one, the pK values fell back on the line, and no more unstability
problems were encountered. The EMF's recorded in two different
buffers (acidic and basic) were 165.1/181.4 and -174.6/—155.6 mV
for the new/old electrode. The calibration of the electrode is
done in the acidic medium and the difference between the two
electrode potentials increases from 16.3 to 19.0 mV from acidic
to basic pH, or a change of 2.7 mV or 0.046 pK units, which
corresponds to the pK difference experimentally observed between

the two sets of data.

A species distribution plot at ionic strength 0.05 and 25°C

is shown in Figure 18.

95

3:... «33cm 33.. .8ch a mm «5.. m5 .5933 FEB in.» 53 mo NV... 93 mo aoE .2 3:3...

ONO

 

 

0.0 9.0 00.0 0
_ _

memo: , 3o.»
1 on
main. mxa

 

 

Kn

. N 8
H3 «5 5.5:? <41 3 v.

96

.....E 8:53 33.. ..8:5 a mm 2.5 23 .535ng 3:3 m.» <5 mo mva 2: ma SE .3 «.53...

NO

151.
Q.

..0

 

 

 

mmml mono?

«mod . mxa

mN

Qm

..Q

Md

v.0

mum

m6

 

 

LED
Q.

.9 8.6:? 44a .6 my...

xa

97

Table VII. Acidity Functions of PAA at Various Ionic Strengths
at 25°C.

 

 

Ionic Strength

 

x 102 9K1 9K2 9K3

9.95 2.26 i 0.04 4.90 i 0.02 8.04 i 0.02
6.954 2.22 i 0.04 4.93 i 0.02 8.07 i 0.02
5.144 1.97 i 0.04 4.91 i 0.02 8.11 i 0.02
5.104 2.18 i 0.04 4.94 i 0.02 8.14 i 0.02
3.224 2.01 i 0.04 4.96 i 0.02 8.23 i 0.02
2.492 1.91 i 0.04 4.95 i 0.02 8.22 i 0.02
1.963 2.13 i 0.04 5.02 i 0.02 8.32 i 0.02
1.876 1.93 i 0.05 4.99 i 0.02 8.32 i 0.02
0.00 2 2.0 5.11 i 0.04 8.69 i 0.05

 

 

98

 

.gpmeaeum 68:68 mo.o 88 <<a to papa =o_p=awcummn mawuaam .mp weaama

L

 

 

 

o\o

 

 

 

00.
no.0... m. 6

24.93.8346 “.0 29596.55 88%

99
3.1.2.3. Thermodynamics of Protonation - The
thermodynamic equilibrium constant K is related to the appr0priate

thermodynamic functions 00°, AH° and 05° by the relation:

-RT 1nK = AGO = AHO - TASO (37)

or:
K = '95? + 45° 8
1" RT 1?" (3)

Thus, the slope and intercept of a plot of 1nK v§_l/T will provide us
with the values of AH° and 05°, provided this plot is a straight line.

The values of the pK's were determined at an ionic strength of
0.05 at different temperatures, and the thermodynamic values at a
given temperature were determined by using the Guntelberg equation.
When the temperature changes, several parameters must be changed:
the volume of solution used, measured at 22°C must be corrected for
the expansion (contraction) of the solution with temperature, and,
therefore, the ionic strength of the solution is also affected. It
can be seen from equation (30) that e and T are involved in the
calculation of the activity coefficient, and since the change in
temperature also involves a change in dielectric constant, corrections
must be made.

The values of A at different temperatures were modified in the
following way: the value f0r A at temperature T was extrapolated
from the values given by Harned and Owen (125) and multiplied by the
ratio of the experimental value of A obtained at 25°C (see Part II.
3.1.2.2.) over the theoretical value at that temperature. For the

same reason as previously, AH° and 05° could only be determined
for the second and third acidity functions. These determinations
are shown in Figure 19 and the corresponding values are listed

in Table VIII together with the corresponding values for acetic

100

.mumm
mmcmacm ummmb gmmcwb mgm mmcwb mgu .mmxmpasou <<a wmgch 508 8x cp 8o mucmncmamo wgaumgwnsm» .mp mgammm

 

 

 

 

._.\0. on m.» 3 mm hm mu
m
_ _ fl _ _
lo...
N
v. c.
_ c + _ _
l\ s + . J) a
JON.
\
39>: 8x c. .u:
...M
low.
“A“
Low.
n
v.— c.
{\j
IrocN

101

momp ..86 AAA ..68 xoom .FA=-;Aeouz .PN-P a .Aepmweaeu PAUAA»_8=< to 3668888: .AmAAazA

 

 

 

com.“ NA.m- m.m~ ‘ mo.o- ocean: 46 mg

Am~.e ~4.8- 8.FN no.9- 86<o= A6 x

mo.o A 88.m 4.o A A.FP- A A mm 4.8 A m.P- <<a to my

Ao.o A mp.m m.o A 8.8- 6.0 A 8.NN m.o A ~.o- (<8 A6 mg

mo.o A mm.m ~.o A ~.A- N A mm N.o A o.m <a-mz co 83

Ac n “V Ngimpos .quxv «Fuxo .pumpos .quv ~pumpos .puuxq xmpnsou
comm AA 3 86_ gomm A8 .8A .ma wox<

 

 

.xmpasou <aimz mg“ 60% van cowumwuomm< umu< com mmwumucmao umsmcacoEngh .HHH> open»

102
acid and H3P04 (126). The protonation reactions are entropy
stabilized. This is expectable since we are replacing two
species of opposite charges by one species with a smaller overall

charge.
3.2. COMPLEXATION STUDIES OF PAA

3.2.1. SPECTROSCOPIC STUDIES

 

3.2.1.1. Visible Spectra - Aqueous solutions of
manganese chloride are yellow-green. Upon the addition of PAA,
the solution turns pink. A possible explanation for this color
change is given by Cotton and Wilkinson (127). The majority of
manganese complexes are high spin, in octahedral field this
configuration gives spin forbidden as well as parity forbidden
transitions, thus accounting for the extremely pale color of
such compounds. In tetrahedral environments, the transitions
are still spin forbidden, but no longer parity forbidden, the
energy of the transitions is about 10 times stronger, and the
compounds have a noticeable pale yellow-green color.

These results show that there is complexation between the
metal ion and PAA.

Interaction of PA with copper ion was indicated by changes
in the absorbances of copper ion solutions upon addition of PA.
The molar absorptivity at Amax of 780 1 5 nm is 10. When PAA
is present at a mole ratio PA/Cu of 1.8, in basic medium (pH

higher than 9), Ama is at 765 z 3 nm with a molar absorptivity

x
of 30.

103
3.2.1.2. Sodium-23 NMR - When enough PA was added

 

to a 2 x 10'2 M_solution of NaCl at pH above 9, to make the

PA/Na+ ion mole ratio equal to 8, the sodium-23 resonance

shifted slightly downfield from 0.36 to -0.05 ppm. At the

same time, the line-width increased from 11.2 to 21.2 Hz. While
those changes are small, they indicate unambiguously the formation

of a weak sodium complex.

3.2.1.3. Phosphorus-31 NMR - Results similar to the
sodium-23 NMR ones were observed. The phosphorus-31 resonance
shifts slightly downfield from -l3.3 to -l4.9 ppm for a sodiumzPA
mole ratio change from 0 to 10.

The change in chemical shift of a 0.3 M_PAA - 0.3 M_MgClz
solution v§_pH is shown in Figure 20 and compared to the results
previously obtained for PAA alone. The plot presents several
interesting points. At pH above 9, when the potentiometric results
indicate quantitative complexation of magnesium ions, the chemical
shift is -l4.7 ppm compared to ~13.3 ppm for the free PA. We
have seen that protonation of the acetic acid moiety results in
an upfield shift of the phosphorus-31 NMR signal, while protonation
of the phosphoric acid moieties causes a downfield shift. At a
pH of 9, PAA is completely deprotonated and, therefore, becomes
a tridentate ligand. Formation of the Mg2+ complex affects all
three sites and the result is a downfield phosphorus-31 shift.

Secondly, at pH between 3 and 6, the NMR signal of a PAA
solution moves downfield when MgBr2 is added, indicating evidence

2-

of complexation with the LH ion (i.e., formation of a mono-

protonated complex).

104

.18 w.» 8.5.38 558: E A. .6 2; s: a ma .8 :23 829.... 5.38885 .8 2.5:

In
gawkmncnm.
dd-

<<n_ mum-... 1 N.

 

 

.0.

 

 

In .m) <<n_ .....O «:22 En

105

Thirdly, at pH below 3, the chemical shift of the PAA
and PAA-magnesium ion solutions are different, showing evidence
of the formation of diprotonated complex. Lowering of the pH
of a free PAA solution upon addition of magnesium reinforces
this hypothesis.

At pH above 8, a mole ratio plot at the concentrations
required by our phosphorus-31 NMR spectrometer (0.3 M) is
limited to an Mg:PA mole ratio of approximately 1 because of
the precipitation of Mg(0H)2. As can be seen in Figure 21,
the chemical shift goes downfield when the mole ratio increases,
but the change in chemical shift is too small (1 to 1.5 ppm)
to allow us to determine the stoichiometry of the complexes.

At the same time, the J coupling constant changes

P-H
from 20.6 to 18.6 Hz.

3.2.2. CYCLIC VOLTAMMETRY - The stoichiometry of the
complexes between divalent cations and PA ions was determined
by cyclic voltammetry. Since the magnesium ion is not reducible
in water, the Ni(II)/Ni and Zn(II)/Zn systems were first tried,
because of the similarity in the size of the ions (crystallographic
radii of 0.65 and 0.74 K, respectively, and of 0.65 R for the
magnesium ion). However, the Ni(II)/Ni couple does not show a
reversible behavior at the dropping mercury electrode (DME). The
free Zn(II)/Zn couple is reversible, but shows an irreversible

behavior when ligand is added to the solution.

106

.8 2,22 :8 AA 5886: cc 8.58. 26: ....» 53 a m... t :5 3385 5.38885

46.862 .oEE .502
a... N._ 0.. m6 m6 «.0 N6 0
d - _ _ u _

 

 

 

 

-
on _
\
_\
A
\
\\1\. .Jnv4v_
\\
\4 Eng
\L¢\\. I
«1 141 \41 w:
.. 0.9
16 22m

2 <6 60 «.22 En

.pm mgampm

107
Finally, the Cd(II)/Cd couple was tested (the radius of

the Cd2+ ion is 0.97 A) and the rate of exchange of the electron
at the DME for the free or complexed metal ion is fast enough
to have a reversible diffusion controlled behavior in the different
experimental conditions.

Lingane and De Ford and Hume equations (128, 129) are well
known and their derivations are out of the scope of this study.
To be able to use these equations several requirements have to
be met. The electrode reaction must be reversible, this is
ascertained by the difference in mV between the forward and
backward peak extrema. It was found to be 35 mV at sweep rates
of 20 mV or less [for the Cd(II)/Cd couple the theoretical value
is 30 mV]. The ligand to metal concentration ratio must be high.
This way the change in free ligand concentration upon complexation
is small compared to its analytical concentration, and free and
analytical concentrations can be equated. The diffusion current
constants of the free and complexed metal ions must be approximately
equal. Finally, the experiment was run at a constant ionic strength
to alleviate the activity coefficient problem.

We can then express the Lingane equation for the following
equilibrium:

M+jL Z MLJ. (39)

with
[ML.]

BMLJ g [MJELJJ (4°)

as:

_ _ 2.303111 j
AE1/2 ' (El/2)free - (El/2)complex ' nF log BMLj CL (4])

108

CL: analytical concentration of ligand in solution, j: stoichio-

' ff"t,E .5 =
metr1c C09 1c1en ( 1/2)free ( 1/2)comp1ex

of the free and complexed metal ions, respectively. This equation

1/2 wave potentials

applies to only one type of complexes in solution.
De Ford and Hume (129), using the same assumptions derived

an equation for stepwise reactions:

E A 2.303RT j
A 1/2 ' n'F— “’9 BMLJ. CL
_ 2

This equation predicts that the half-wave potential of the cadmium
couple should shift towards more negative values when PA is added
to the solution, which is actually observed.

In the case of step equilibria, a plot of AE1/2 v§_log CL
will give [equation (42)] a curve with a variable slope. At low
ligand concentration, the major term is BMLCL and successively
becomes BMLZCE, BML3CE’ etc... when CL increases, corresponding
to a change in slope from 30 mV to multiples of this value. As
can be seen from Figure 22 and Table IX, the slope of such a plot
changes from slightly above 1 to slightly below 2 when the ligand
concentration increases, showing evidence of a CdL and a Csz
complex in solution.

De Ford and Hume extended their derivation to allow for
calculation of the overall formation constants. A series of
Fronaeous equations (130) can be defined as:

F0 (L) = 1 + B][L] + 82[L]2 = exp (AE1/2 nF/RT) (43)

F1 (L) = Fo (L) ' ‘ = e] + 82 [L] (44)
[L]

109

NE

.Asmmpms< Eswppmgh "mucmcm$mmv emumam <g-+~vu mg» so» m<¢g me; we cowuucam c mu m 4o pop; .NN mcammu

 

 

 

    

 

 

 

_<n__8_
c...- m..- QT ow.- m.~.. tw-
- - d 4 u - — — -
1, ..AuAv_
.0:
40m.
85.8 863
mmmm u 302.. < - on. _w
E . u
> 8mm 6 2A -8.
a
3.9.8.. .. «no.8

_<n__8_ .m> Nanozvoo ...o «Sm

110

Table IX. Half-wave Potential of the Cd(II)/Cd Couple in the
Presence of PA v§_Tha11ium Amalgam.

 

 

 

[50315/11-3 log [PA] E1/2 (mV)

0.00 194.2

5.022 -2.299 145.3
10.06 -1.997 133.3
15.09 -1.821 124.8
20.13 -1.595 118.5
25.16 -1.599 113.3
30.20 -1.520 109.0
35.23 -1.453 105.3
40.27 -1.395 100.5
45.30 -1.344 98.0
50.34 -1.298 94.3

 

 

lll

-FL-B
F2 (L) - 1([21 1 = 82 (45)

assuming we only have a lzl and a Zzl complex.

 

A plot of F0 (L) v§_[L] will be a steeply rising curve, however,
as [L] approaches zero, the graph will have a limiting slope of B],
and an intercept on the Fo (L) axis of l. A preliminary value of B1
is obtained. A plot of F] (L) y§_[L], on the other hand, will have
a limiting slope as [L] tends to zero, of 82 and an intercept on the
F1 (L) axis of B]. A confirmative estimation of B] is possible and,
in addition, a preliminary value of 82 is obtained. The values of
B1 and 82 found are given in Table X.

The precision in the determination of B] and hence of 62 is

limited by the previously mentioned assumption that we have a high

ligand to metal mole ratio.
3.2.3. POTENTIOMETRIC STUDIES OF COMPLEXATION

3.2.3.l. Sodium Ion Electrode Study - Due to the low
stability of the sodium-PA complex and the high charge of the PA
ions, the experiment was performed at l.0 ionic strength. It has
been shown (lBl) that at high ionic strength, the activity coefficient
is much less sensitive to the variations of the ionic strength. Thus,
in our experimental conditions, the change in ionic strength was
small enough during the course of the titration to be neglected in a
first approximation. The value of the formation constant is 3.3

(Table X).

3.2.3.2. Glass Electrode Study - Figure 23 represents
a coulometric titration of PAA and PAA with MgBrz. In the latter

112

o.P

ll
o—c

¢.o

ll
H
U

mo.o
wuo.o

II
H

 

 

 

P.o a m._ “Nazvcx mop
F.o a m.m mm.o Aszvcx mo_
one unz
mo.o a om.N oo.o a mm.~ mo.o a ~m.~ mo.o a om.~ mo.o a m~.o Azozvcx no.
35 a $5 85 ... 3...” No.0 .. 85 No.0 a o: 85 a m: :5: 9:
om mam (mag no: maz

 

 

.mcoH papa: cur: mmxmpaeou <<a .x a_n~»

113

.msmumam +Nmzl<<m vcm <<a wgp we mm>L=u :omumguw» owguosopzou- .mm mgammu

 

 

mFZwJ<>50w
m N _ O
_ _ _
\...,.i.. . _ .IgnvAvmvl
>Em_uw< .13!)
.6 I l I
.. .. .e. loo—I
>Ehhum< ., - .
... _ . Io
6.49%on . e . 3,5
. . . . m
. . . <3
. . . mum... _ loo.
JOON

 

 

<<a mo ZO_._.<m._..._. Baku—20.500

114

case, the following equilibria are taking place in solution:
Mg2+ + LH; 2 MgLH + H+ (46)
Mg2+ + LH2' 2 MgL' + H+ (47)
Experimental results showed that the latter complex was more
stable than the former one. Phosphorus-31 NMR results showed
evidence of a diprotonated complex, however, in our experimental

4 n PAA and 7 x 10‘4 u MgBrz) the amount of

conditions (7 x 10'
magnesium ions complexed and the corresponding release of protons
are small, and the resulting change in pH is not detectable.

The formation constants of the magnesium-PAA complexes
were determined at different ionic strengths. The data are
listed in Table XI and plotted in Figures 24 and 25.

The Guntelberg modification of the Debye-Huckel equation

allows us to express the thermodynamic values of the complexes

of equations (46) and (47) as follows:

K§(MgL) = M" LL x 3191-. = K§(MgL) x Y_MQL_ (48)
g 7"ng YMng
log K$(MgL) = log K§(MgL) - l—ffififll) (49)

and similarly:

C _ t 8A .5

The theoretical slopes for the deprotonated and protonated
complexes at 25°C are -6.ll and -4.07, respectively. The experi-
mental values are -5.99 and -2.27. As in the case of the protonation
of PAA, the completely deprotonated form follows the theory quite
well, while the value for the monoprotonated form is off by a factor

of 2. In these conditions the thermodynamic values of the formation

115

Table XI. Formation Constants for PAA-Magnesium Complexes at Various
Ionic Strengths at 25°C.

 

 

Ionic Strength

 

x 10+? Iog Kf(M9LH) log Kf(M9L)
9.94 2.4 i 0.1 4.22 i 0.05
7.04 ' 2.6 i 0.1 4.38 i 0.05
5.112 2.6 i 0.1 4.50 i 0.05
4.185 2.4 i 0.1 4.53 i 0.05
3.176 2.6 i 0.1 4.66 i 0.05
2.425 2.7 i 0.1 4.80 i 0.05
2.047 2.8 i 0.1 4.88 t 0.05
0.00 3.0 1 0.3 5.58 i 0.09

 

 

116

.uwu mucosa ammo; gmmcpg

w m. we: 93 .fiacmbm FEE w)» xmasou +~mzu<§ “03839.35: 93 mo V. 9... mo uoE .3. 953...

“\r

Nflv _Av MWN

 

 

 

 

.x 02
/

 

 

 

 

 

bNNIu 32m on
mmmuiooz. v. 8.

 

 

 

H:
C.

.2, 8.681582} 8.

117

.pwm mmcmzam ammo; gmwcwg a mm mew;

mfi .cumcmbm FEE 1......» 6358 91-53 umumcouogamo 33238 23 mo of 3.. m5 to “SE

...N

NO 9.0 ..O 00.0

 

Z

.x 8.

man I. 32m
own . Y 8.

 

 

 

 

HH\\((+_ .m> 5.8:? 39):; 00.

.mm ac=m_a

118
constants of the mono- and deprotonated complexes are 1.0 x 103
and 4.0 x 105, respectively.

By analogy with the cadmium-PAA system, similar complexes
were assumed in the magnesium-PAA case. However, our potentiometric
results disagree with this hypothesis and show that the magnesium-
PAA complexes are different from the cadmium-PAA ones. These
results show evidence of completely deprotonated and monoprotonated
magnesium-PAA complexes.

These complexes have one or no charge. In these conditions
the change in ionic strength during the titration is negligible,
and the value of the ionic strength can be equated to the concen-
tration of supporting electrolyte. A species distribution plot
at an ionic strength of 0.05 is shown in Figure 26. It can be
seen that for a 1:1 mole ratio of PAzmagnesium, approximately 50%
of the magnesium (or PAA) is complexed at pH = 7, which is the
physiological pH.

The formation constants of the PAA-calcium ion complexes
were determined at 25°C at an ionic strength of 0.05, and compared
to the values of the magnesium ion complexes. A similar study
was performed with strontium and barium ions. The stability of
the complexes decreases with increasing radius of the metal (see
Table X).

The complexation of sodium ion has been studied by the same
method (Table X). At 0.078 ionic strength, the sodium-PA complex
has a formation constant of 27. Owing to the low ionic strength,
this value is compatible with the formation constant of 3.3 found

by sodium ion electrode measurement at 1.0 ionic strength.

119

 

.gumcaepm u_coH mo.o be Emumsm uz-<<a any co papa eovuanwcumpa mm_uaam .om mesmwa

+N

0.

 

o\o

 

 

 

 

N . 8.0%. 6 8.
emu: s. 4.0...» <3 2 To...» do 20.59590 88%

120

These results definitely show that sodium salts cannot be

used as supporting electrolytes.

3.2.3.3. Thermodynamics of Magnesium Complexation —
Measurements of the formation constants of the magnesium complexes
at different temperatures lead to the determination of the AG°,
AH° and AS° of the reaction. The precision in the determination
of the monoprotonated complex formation constant was low and
prevented the calculation of the thermodynamic quantities for
that reaction. These quantities were determined for the deprotonated
complex only. A Van't Hoff plot of these formation constants is
shown in Figure 19. The thermodynamic quantities are listed in
Table VII. As can be seen, the complexation reaction is entropy
stabilized and enthalpy destabilized. For EDTA, the changes in
entropy and in enthalpy at 0.1 ionic strength have been reported
to be 50.5 to 52 cal/mole, °K and -3.14 to 2.9 kcal/mole (132).
The entropy stabilization effect of complexation is readily
explained by the fact that we go from two ions with high charge

to size ratios to a larger entity with a single negative charge.

3.3. PHOSPHONOACETIC ACID ANALOGS

In order to determine why PAA is an inhibitor of the repli-
cation of Herpesviruses, biochemists have conducted studies with
analogs or derivatives of PAA. As mentioned earlier, it has been
found that phosphonofbrmic acid (PFA) is the only compound found

so far which shows a biological activity similar to that of PAA.

121
On the other hand, 2-phosphonopropionic acid (2-PPA) has an
activity 50 times smaller, while 3-phosphonopropionic acid (3-PPA)
does not show any activity at all. The complexes between
magnesium and these compounds have been studied in an attempt

to correlate activity and complexing power of these ligands.

3.3.1. SPECTROSCOPIC STUDIES OF PFA - A phosphorus-31
NMR spectrum of the trisodium salt of phosphonoformic acid (PFA)
consists of one peak with a chemical shift of -0.34 ppm vs_
85% H3PO4. If we consider the differences between the PAA

and PFA molecules, we can rewrite equation (10) as:
A6 = BAXR - AABOPO - 147m1T (51)

Again, the change in electronegativity of the R group between
the two molecules is the major factor influencing the chemical
shift [see Riess gt_gl, (67)]. The formic acid carbon has a
larger electronegativity than the 9H2 carbon of the acetic acid
group, and in going from PAA to PFA, a poSitive chemical shift

is expected, which is indeed experimentally observed.

3.3.2. DETERMINATION OF THE pK'S - The determination has
been done at 0.05 ionic strength to allow for eventual use of
the Guntelberg equation (28) to determine the thermodynamic
values.' Since PFA undergoes a decarboxylation reaction in
acidic medium, its trisodium salt was used instead. Complexation
between sodium and PF ions is suspected, however, owing to the

low concentration of sodium ion, calculations showed that the

122
influence of the sodium ion on the values of the pK's and
the formation constants of the complexes is negligible.
The values obtained for the pK's of PFA, 2-PPA, 3-PPA

are compared to the PAA values in Table XII.

3.3.3. COMPLEXATION - Once the pK's of the above-

 

mentioned acids had been determined, it was possible to study

the complexation of magnesium ion by these different ligands.

The formation constants determined by potentiometric measurements
are listed in Table XII.

Experimental results show that less stable complexes are
formed when the length of the PAA molecule is modified. Loss
of stability is also observed when a hydrogen of the CH2 group
of PAA is substituted by a methyl group.

These results are only partly satisfactory since biochemical
studies have shown that PFA has about the same biochemical activity
as PAA. However, if we compare the species distribution plots of
the two acids when magnesium ions are present in the solution (see
Figures 26 and 27), we see that at pH 7.0 approximately 60% of
the magnesium ions are involved in a complex with PFA, while less
than 50% are with PAA. Similar calculations show that at a pH
of 7.0, 5% of the magnesium ions are complexed by 3-PPA,
and 30% by 2-PPA. These results confirm the hypothesis that
the mode of action of PAA involves complexation with magnesium

ions.

123

 

 

 

mo.o . mm.~ mo.o n mm.¢ mo.o a mm.m mo.o a om.¢ Asmzvcg mop
F.o a N.F m.o H e.~ m.o a N._ F.o a m.~ Azsazvcx mop
No.o n mu.e No.0 n em.m No.o H mm.“ No.9 n ep.m mga
No.o H me.e No.o . m_.m No.9 a mm.m No.o a em.¢ Nxa
eo.o a om.~ _.o a m._ _.o a N.P eo.o n mp.~ .xa
<aa-m <aa-~ (ma <<a

 

 

cumcacpm uwcofi mo.o ego oomm um mmo_mc< <<a ecu <<a co mpemumcou cowuastoa =o.uaxm_asou .HHx m_nme

124

..pmcmLHW awco. mo.o pa emamxm +~uz-<.a 8.» co pop. cowasnw.pm.o mmmumam .NN mc=m..

 

 

hum“
.«r

 

 

 

Aunv_

 

madam. .6
N59). .2 4.03.4“... .2 .70...» ...o 20.59590 myomam

125
3.4. CONCLUSION

A relationship between the pK's of PAA and its acidity
functions has been established by phosphorus-31 and carbon-l3 NMR.
The chemical shift changes have been explained in terms of changes
in electronegativity of the substituents, n-bonding feedback, and
§_and p electron shielding. Spectrosc0pic studies (UV, NMR)
showed evidence of complexation of sodium and magnesium ions by
PAA. The stoichiometry of the PA-Cd2+ ion complexes has been studied
by cyclic voltammetry. The acid dissociation constants of PAA
and the stability of the complexes of PAA with metal ions have
been determined by potentiometric measurements. The pK's and log
Kf of magnesium ion complexes with several phosphonic acids have
been determined, and related to the biochemical activity of these

ligands.

APPENDICES

 

APPENDIX A

SODIUM-23 NMR

126

127
SODIUM-23 NMR

A.1. INTRODUCTION

Complexation of sodium ions by the crown ether 12-crown-4
(lZ-C-4) was studied in different solvents. In order to minimize
the number of equilibria taking place in solution, sodium salts
which do not exhibit chemical shift dependence on concentration
were chosen. In this case, it is reasonable to assume that the ion
pairing for these salts in the solvent considered is negligible.
All the studies reported in this appendix have been done at a
constant concentration of salt. Since there is no ion pairing,
and the crown ether bears no charge, the activity correction is

small at best and Kf's obtained can be taken as thermodynamic values.
A.2. RESULTS AND DISCUSSION

The stability of the crown complex is largely determined by
the size relationship between the inner diameter of the crown and the
diameter of the desolvated ion, and by the solvating ability of the
solvent. If the rate of exchange of the sodium ion between the two
sites (free and complexed ion) is greater than Avn/Z, where Av is
the difference between the characteristic resonances (in Herz) of
each site, only one population resonance is observed. This is the
case with 12-C-4 and sodium ion. The ligand 12-C-4 has a cavity
size of 1.2 to 1.5 A while sodium ion has a diameter of 1.9 A
(from crystallographic measurements),therefore, in the cases where

the solvent does not have too strong of a solvating ability, the

 

128
formation of 2:1 sandwich complexes is expected (2 crowns for 1
sodium ion). This is the case for acetone, acetonitrile, propylene
carbonate (PC), pyridine, and THF. As can be seen from the graphs
(Figures 28 and 29), there is a break in the chemical shift change
for the various solvents at a mole ratio of 2. In several cases
(acetone, acetonitrile, tetrahydrofuran (THF)],the complex precipi-
tated out of solution for a mole ratio above 2. In the case of
methanol or water, a precipitate appeared as soon as the ligand was
added to a solution of sodium tetraphenyl borate (water, methanol)
or sodium perchlorate (methanol). In the former case, an elemental
analysis showed that we had a 2:1 complex: NaBPh4(12-C-4)2 calculated/
experimental: C: 69.16/69.30, H: 7.55/7.42. The melting point
of the tetraphenyl borate complex was found to be 248.5-249.5°C.

In the case of dimethylsulfoxide (DMSO), methanol or aqueous
solutions, the solvent molecules compete quite successfully with
the ligand, consequently, only a weak complex is formed, and a
large excess of ligand is necessary to produce a variation of the
chemical shift from the position characteristic of the solvated sodium
ion in the given solvent. It is quite interesting to note that for
all solvents except water, methanol and DMSO, the chemical shift
above a mole ratio of 2 is the same, showing that once complexed,
the sodium ion is not affected by the solvent: the two crowns
around the sodium ion shield it effectively from the solvent.
Figures 28 and 29 show that in the case of the three other solvents
(DMSO, water, methanol), the chemical shift of the sodium ion tends

to the same limiting value of 5.0 i 0.5 ppm when the ligand

"A9 3:: E.

 

 

 

 

 

 

I I ...32. .... .582. m 8... .6. 282:: fi :2 m 3...
... 485 A! .8 .... z ..o 29 "same: 4-8-2 8 0.52. £3. a... it; .3285 8.5.68 .8 28.;
.+oz_\_¢-o-m__ a m N _ o .
a i q 4 n
.. o
1 )DIIIIMIIIID 0 $1 in). (kw
In) I .
w
a 0 L.\...
T D b D l‘E'JQ't \J c
O
0
09
O 3
o O
o u.
o m
o M.

’I

 

co—

130

.6. 8:52.38... .3 .51 . so 288... ..s 8.3 ....

 

a .... 6.3.... .... a 3... .... .....amaz 4-8-2 t. 2...... 2o: .2. 3:... .885 8.52.8 .a 952..
D I I
... E2.» 0 N: c m N _ o
. n . d 4 . . -
k .0
o
--1-1nli 1 1 1 11 u m
4 A.“
u 09
a
u o
1
m:
d
w.

... 0..

 

131
concentration increases. In the case of dimethylformamide (DMF),
no chemical shift change is observed, and, therefore, no conclusions
can be drawn as far as the strength of the complex is concerned,
since it is obvious from the preceeding remarks that the solvated
and complexed sodium ions have the same chemical shift.

We assume that in all solutions containing the crown ether
and the sodium ion, the latter is found in three environments:
free solvated sodium ion, 1:1 12-C-4zsodium ion complex and 2:1
12-C-4zsodium ion complex. The exchange between the three sites
is fast compared to the sodium-23 NMR time scale, therefore, only

the mass average chemical shift is observed:

5obs = 6MXM * 6MLXML * 6MLZXMLZ (52)
where dobs 15 the observed chemical shift, XM’ XML’ XMLZ are the
respective mole fractions of the free, 1:1 and 2:1 complexed metal
ions, while 6 , 6 and 6 are the respective chemical shifts of

M ML MLZ

the three species. Assuming the following equilibria:

M + L ML (53)
ML2 (54)

++ ++

M + 2L

where L is the ligand, the formation constants of the complexes in

concentration units are:

_ - ML
K1 ' 31 ‘ M L (55)
[ML2]
K - _________ 56
2 [MLJIL] 1 1
82 = ["L2] (57)

[NHL]2

 

132
where [L], [M], [ML] and [MLZ] represent the equilibrium concentra-
tions of free ligand, free metal, 1:1 and 2:1 complexes, respectively.

By using the ligand and metal mass balance equations:

c; = [M] + [ML] + [MLzl (58)
CI = [L] + [ML] + [ML2] (59)

where c; and CE represent the total concentrations of metal and

ligand, respectively, we can express the concentration of free

metal as:
mP+me+qml+r=o mm
with:

(cfik§ - 8cfikzk1 + 4cEK2k1 - K§c[)

 

 

 

(61)
(4K2K1 - K?)
T T2 T2 T T
_ (1 + K-ICL + KIKZCL + 4K1K2CM - 4CLCMK1K2)
q ' 2 (62)
-CT
r = M 2 (63)

This cubic equation has three real roots for a mole ratio of ligand
to metal less than 2, and one real root after. Once the concentra-
tion of free metal is known, equation (52) can be written as a
function of the two formation constants of the 1:1 and 2:1 complexes,
and the chemical shifts previously defined. Since 5M can be easily
determined from measurements of solutions of sodium salt without

the ligand, equation (52) contains 4 unknowns. For a fairly strong

2:1 complex, GMLZ can be determined experimentally by the addition

133

of such excess of L that essentially all of the metal ion is
in the 2:1 complexed form. For weak complexes, however, either
thelimitingchemical shift value is unobtainable, or such large
excess of L is needed that the solution loses even a semblance
of ideality.

The procedure we used to solve equation (52) was to substitute
the experimental parameters 60b5, Cg, CT

L
6ML’ and 6ML2 until the calculated chemical shifts would correspond

and 6M and vary K1, K2,

to the experimental values within the error limits on concentration
(1%) and shift (0.05 ppm). The data were analyzed on a CDC—6500
computer using the Fortran IV program KINFIT4 (135). A description
of the subroutine EQN used is given in appendix 8. Actually, before
the data could be processed, the correct root of the cubic equation
(60) to be used had to be determined. He had the choice between
three roots, all mathematically valid, but for a given mole ratio,
only one has a physical significance.
In order to determine which root to use, the best solution
was to simulate data by using KINFIT4 (see appendix 8) with the
expected values of the chemical shifts of the complexes, and their
formation constants. From these results. it readily appeared
in which region of ligand to metal mole ratio each of the roots
could be reasonably used. Once the domain of validity of each
root was determined, the experimental data were processed, and
as a check, the refined values of the chemical shifts and
formation constants were used to generate data for each root.
Following this procedure, the chemical shifts and formation

constants were determined for the fbllowing systems: NaBPh4-(12-C-4)

 

134

Table XIII. Formation Constants of Sodium-(lZ-C-4) Complexes
and Their Chemical Shifts

 

 

 

1:1 Complex 2:1 Complex
Solvent Kf 6(Epm) Kf 6(ppm)
Pyridine 200 :10 3.9:O.l 150140 5.4:O.l
Methanol 60:10 4.7:O.l 20:25 4.9:O.1

 

 

in pyridine, and NaI-(lZ-C-4) in methanol [NaBPh4-(12-C-4)2 and

NaClO4-(12-C-4)2 complexes are not soluble enough in methanol for

 

NMR purposes with our instrument]. The formation constants and

“-E:

chemical shifts are listed in Table XIII. As can be seen, it is
doubtful that we have a 2:1 complex in methanol. In the case
of PC, the formation constant K2 is larger than 104, and for the
other solvents, precipitation of the complex at a mole ratio near
2 prevented the determination of the formation constants. However,
it can be seen from Figures 28 and 29 that 6ML2 is 5.0 i 0.5 ppm
for all the solvents studied, except water and DMSO. In these
last two cases, the complexes are too weak, due to the good solvating
ability of water and DMSO, and the formation constants could not
be determined.

A comparison of the IR spectra.of 12-C-4 alone and of the

NaBPh4(12-C-4)2 complex between 700 and 4000 cm"

did not show any

new band which could be attributed to the complex. This result

correlates well with the literature result for lithium ion and

12-C-4 complex (135). The complex bands are in the far IR region.
Thus, it has been shown that sodium ion and 12-C-4 form

2:1 complexes in most of the solvents studied and that the crowns

shield the nucleus from the solvent.

APPENDIX B

APPLICATIONS OF THE
COMPUTER PROGRAM KINFIT4

135

 

136
APPLICATIONS OF THE COMPUTER PROGRAM KINFIT4

B.1. CALIBRATION OF A pH GLASS ELECTRODE

B.1.1. PROGRAM FUNCTION - The program KINFIT4 is a general
curve-fitting routine (134). The subroutine EQN of this program
was modified to allow for calibration of the electrode. The

electrode calibration equation is:

E = intercept + slope x log ([H+]o - t X 1.o + residual) (23)

F x vo

where intercept and residual are the unknowns [labelled as U(1)
and U(Z) in the computer program], E (experimental potential) and
t (time) are the variables [XX(l) and XX(2), respectively] and i0
(coulometric current in mA), vo (initial volume of acid), [H+]0
(initial proton concentration) and slope are the constants
[CONST(l) to CONST(4)]. The subroutine EQN used in the present
case is listed in B.1.2., and a sample of data input in 8.1.3.
The data input is as follows: the first card contains the
number of data points, the maximum number of iterations to be
performed, the number of constants and the convergence tolerance.
After the second card (title card) comes a card containing the
values of the constants (if any), and then the card containing
initial estimates of the unknowns. The cards following are data
cards. In the present case, the input is observed potential as
a function of the time during which base was generated. Each

datum is followed by its estimated variance.

 

PLEASE NOTE:

Dissertation contains small
and indistinct print.
Filmed as received.

UNIVERSITY MICROFILMS.
43

I

137

SUBROUTINE EQN

B.1.2.

 

‘fetéJ-Z'
0 J
I‘WaTv.

C

F

l
7:; roe »' ~
...:1:. J5h¥
.F " 1 ‘3
2 1.IS ’
s = i.

J
a
X
‘
1

~c=V,L;J:e,ng
:
L
7
F
AHF-F S

*vp
.T.C N
I 1
2 v
3 r
I 1

210;;
>1
:3 o
).l_ .
.Cl. 6
(" .
t). 7
:tva
7 &::3
LT;3.'I

$:

)1
:y s:.;:

659'. N'-

Tgl
L

{IN I|.;
Y

'L‘av’

:t'll

3'\

i3 ”

0.

v0 JH

:osfiu

9LT-VYI - v
{FNST'
«909:?

3110:
I
I

9’
$5693.12 UL)

gua:tur

3:
OULQ'L "
2L

ll.)

vii-Q9
' OOOHOJJ)OYO

D

(

1

I

V

73 3!
:P". 3‘.
‘ Z
T‘.
or N
St I
I1 I
1’ u

L‘: 3 ANAI‘QL

F:
:1 so
<u.r~a?gzg

1
2
_ 4
5

v
{V
.C
/

I L

5 U

E 1

1

P V
(

305oto1o‘vao‘1‘012911.12|

F 1
ui°:7LCE.J?n=t.:u'
:o'

I
D
i
l
l
(
F
J
0
O
I

4.0
,
V
U
3
2
I
3
J
3 I
L
6
3
i

£vl51o. DJ :20 ..b’.... 101- ... ... ... ... ... on. ... .3
T ...XNP/v...oc it... ..IU).,..3S. .J .... U? Unit-(l l:T».N.§.-nea:u .J J, Jo J U U u) U
UN... OnuNV.) 9.11.310. N : 3 : ..NN‘VVN...NNQIY. Tel-l u‘ :08.lal,. ... VNNNNV‘.N NVNN‘V‘VN‘
DOT LCGONOOTGEOI? K Eran v .31.!!!QISSSS..’.1’ECC3.N1353313 c521- RIRIQIRIRI.

S

P-

I

C
0'ohio’1’ GC 73

.UT.YCOIFVhNLOOCOTTtCtCECFtOLCCC!IUUUUCLFT2CLEOSCtOFEOEO¢C$Ck0£fi

“CHHVCCD 90 n9. GCN NIJPCDECIFC (VHSOhPPPFCPCF‘ .‘ECFCPCFCFI.
I. .

2!. 1:3 I. b 5 .U 9 .u 1 4‘

1 762 “u z I... 1

cease

DATA SAMPLE

8.1.3.

 

.. ...........
c g c . . . .
...... 5...... . I:
...... O

n I 7. a? t 1 t

I C O O O O O

O h.’ C p. c v I.

099075).

I. z........11
A
c.
r»

.c b C - o .o a

rb . o o o o n c

O .as_-I.IO:.-:Ib-

. o o o o o o o

‘5 0...:OOQOlono
.-
v

1.15300;

as Cruccose9.°

. o o o o o o o

.- Geezer-r- .

39.3173.

.24 1,2633
49
(if.
)- o
I:

5 23233.22

‘
It. a o c o o o a
o ..hilblofo’lb-
v.3 0 O O O O O O
I. 0......‘QAQQIQ

V
a
o
..h ”I. .6.’5
'94 00000-0
- 6232’“
f... .4363...‘
0 221.111
N
0‘
col
77
n nutuuuuu
o
I. .EE
.1 3.0000000
Is 610511.110.
fi“.
1!;
1 o
...-4

CTR)?
1

0.9.3090
... 52980.5
512 119‘33
1.8 o
If
oh
5.-

l

1 AND 2:

DETERMINATION OF THE FORMATION CONSTANTS OF 1

COMPLEXES FROM NMR DATA

8.2.

PROGRAM FUNCTION - The data of the plot chemica1

 

8.2.1.

d were fitted to equation (52)

igan

shift vs mole ratio of l

(52)

6obs 3 6MM + 6MLXML + 6MLZXMLZ

138

which was rewritten as:

c
a =EM_(6-5)+1L_(a -5 )+5 (64)
obs Ctot M ML2 tot ML ML2 ML2
M CM

The constants are: CfiOt, total concentration of metal ion

[CONST(l) or CMT in subroutine EQN], 6", chemical shift of the
free metal [CONST(Z) or DELO]. The variables are CEOt, total
concentration of ligand [XX(l) or CLT] and the experimental
chemical shift [XX(2)]. The unknowns are: KML’ the 1:1 complex
formation constant [K1 or U(1)], and its chemical shift 5ML [DELl
or U(2)], and KMLZ’ the 2:1 complex formation constant [K2 or
U(3)] and its chemical shift GMLZ [DEL2 or U(4)].

The concentration of free metal in solution is obtained by
solving the cubic equation (60). Before a mole ratio of 2 (ligand
to metal), there are three real roots, and only one after. The
root used before a mole ratio of 2 was determined by choosing the
value of the angle 0 (e = 0/3, 0/3 + 2n/3, 0/3 + 4n/4). In order
to find the domain of validity of each root, the program KINFTI4
was used to simulate data with approximate values of the unknowns
for each root (see B.2.2.l. for a listing of EQN in these conditions).
Once this step accomplished, the experimental data were refined with
the proper roots. The subroutine EQN used is listed in 8.2.2.2. and
a sample of data input in 8.2.3.

The order of the cards in the data deck has been given in
B.1.l. The only change is in the data input: we now have the ligand

concentration in solution as a function of the chemical shift, with

their respective estimated variances.

139

SUBROUTINES

 

Subroutine EQN for Data Simulation

B.2.2.].

 

B.2.2.

. e .
hNn
ten-.0 r.
‘5'": E
0:. I T
I o .u C I
.q. o ... .7 .7.
F1. '. TT .
T 6 (I! up D 5
.. ..0 0.1.9 .1 .... 1|
.7 . .. o.V. s o ‘2
U o o 0‘ o t O o 0
CDT IV I ‘ H 1..
In 6 c '0 o o “1
oUO (I... C C 0 K
Mo!“ ...-5‘ «A.
\ o 9 1.3. N1. . ‘2
4...... 31... 0.1 I a .K
0.. .- C 1!! VOT cl .0
N the T3 on. ‘1.
ODJ Is 0 I." K I‘
FC 9 [I I N‘ o o I.
3.0T ... '.u 900 0 I5
v I 7.0... ..ouu". 2‘
:K\ .I .C H . It... 0 I
h 110 P2? L GLO a I
9' O 0" I ‘00; Coll
Tn? F Lo..- 0 TELHL" K1
9 l: 31.. N E o 0 To
09" IVFT 5 HF :22 o T
.lvO 6... v! r... . . TL
0.1L :1 TI! ... 721.21 L:
..t a [$01 9 L ON o o 9.0 ...
L H? I 1. v v Tie; .2 L
h .0. VIII T .5. N N I ’K ..
. .IC ICIOS .. 001°th 2‘ c
I)! .11... . roan. My N 0 o 6
u no C!!! NT FTFT oT I
.....T 335T 1 OdATO 3 1.. 7.
L7: .5115 Z .825? .r r K: I ...
...... b o 0" 1 T ITOTC (b I 3 I.
1.V (110 o LLNPN N o 1 7 I I n.
T 0 O I.o. 2 u u. ‘oiuoiol‘vv .5 T‘ ,. .x. Q 0
IS? TC.) 9 .. ...-av: TFTFS< I0 I . .. ..
9.99 MIST- C “To :2 9 c- I I O L
0... a I 0.1T. . .a .2‘ INN‘" 01. OT 1. I. 1 3. 0A ‘-
.oY XILri .. ..., (£720....136K IL .I u .. r. r. I C .t
a .1 o 2313... o 3 N4 v... .. 2 o 23 .1. o ... 3 3 . I ...
...J T). 0N1 a OCLH . .01 (.I I7.) 1‘ H ... c
on. .J F‘I 9' I a c ‘....“N‘51 O 2 I). CI 14‘ 9! 1 0 I2 3 l‘ .. 1.
J at. OU..Iv 6 T THOCOCZIt 1 D II '1 C. ( I0 T T...- T I I T
5 o J v’. .. 0 o ...). 25.911 «x 4‘. .0 0 I51. 2.! .1 T2 I ...... I I O
...In a I .1... 7 0 ATT Ti. 81!... It: 733.. 1 30 ... v .0 D o .0 a I U
... o . . CC 0‘ o B TN 3531...). A. on . 2. QXI ...! Ir. 01 3.... 1..; 3r. ( 2 ..u
L .L . grizc... l ...-... ”INK. I 0 IT 00... ... .09.. 9. O... I T... I o L
T5 0 Milan‘s}. o I TM IECECO 7.- THZ Oi hI 0 Ce. . I T). I F... I 2 IT. I
l . .9 .I 1 i. .. 1 ...-:20 n. ....q. PLK 0!. Th TI 5‘. .... o 1.. 0| 1.. o P... ..
.quD IIbCSP... o . I nEIF3734 1.. C003! .Ibi II1I1I Q1 ( 0 $2 ... o o T .C .
..TA 9 TTTrTO-L b 0 I? T S I21/n L7 2 If 09.. so: 0... 02 of. 3 I L ...-X o
N T .JMII.) 5 v ... 1‘ IPTSlIlI2b o 9K(I.:I.‘..1 02.310. 01... 5‘ .1T 0‘ I .... LI ...
..U..v.::... ck .1! I. N (T 11¢.“ Z BKouY ......oIID o .In..3(..2 5 I. Z ... Io c. 0.7.... gm
NC TTRONIOITKoA o o ITS I I‘ENCI‘I‘I T5 T1“. SZI‘X 1.. C‘ITnu/IE and: II! 2’. .002" I a.» o
..(TSSFPD..:.4 963... ...1 ... ... Eu. ...151 2 Bl 3:133U1JC... 5(150 0 T5... . (59.830.5qu TT riasT TT .130, :1 .1... ... ... ...> ... ... ... ... ...
o. a.YN//1.... 25:7...(o: .55 ..J J J. UTNOIIIT'IC l. ..I (( Unb.IoIOOJTHII‘eJfl.B°!.T-I§: aJJoToIéaf. azuaTaltq. W. J 1.9 U U J .J J
{...-ONES oI(:.(.. N::2:NKKNNN...NNIO....UTU :Usqa... TN0.0T..45:.0 (0. NT 1..-C...) se.:-E§1-Nlt.|8£‘ . N OI .432...VKEVNN:.VN.\N\N\%NN..
abT LCOONJOTG 2.0.1.1 (...-5.121.111.94! Ir... 1: II! a In 5:21 . OH. O TIEIvTCIIO 01....11TNI1058122TNQ. 7 0.. I152HI0R£I¢ I. .- .a.‘. 8‘. ..Q. .3 9....
3d 01 o1‘..:..13... ’TT1'33 Jv J..Jo {JV t ta4J1Jfi. 2 JIXI?‘T(.T(1.-3,.3.ITHV 3.. A :TTI TTTnIa‘uoITTY TTTT. T T 1..-1......IJTJV Jo.o.JTJo.J.oJTJTJ
...NIIVQIIITTGQ. VVJIITITNTNCY‘TTLthTIY L L L 9.. "(To-\(TI "150:9... S‘s-e... It.‘s.|s...l I ... V II ..LILSTNTNTNITVTNTN? NT NT.
. L? b' LOTF vHNE‘OLOTYLCIblCFLmC LIE..E?|.L.IFL.1EEFP I .. ...: ..llr‘ OFCIINFLBIVIT‘Fépabtf SAX-.753 L ROLF ..EECI.C...UICCI.U.LCLn.Ir 2."
...(u. .v (LIL flux-.29... ....CnhwjbrrnsCIPCCCFVDKOCXCOKDKCINFCSQPIBP. “FEDTSquTh39‘NCo-Tog N C (15109 ECPLEK. .....r. RC; ....» (EFL...
017.1. I. o
1 7.25 2 .2751 111 122 2.. 1.L.w.u9|u1’.
3 222 b.- b l- .0 0 o o 0 6 .... o 2 1 e. .
”- o 1 2 ‘. O. 2 ‘ c
a 1 2 3

CE

140

 

Subroutine EQN

8.2.2.2.

 

 

 

1111111111
5.0088000
.... 000000.000
NNN 0
.00 0
....It ...
2.. T T
v- 0 o a .I
1... o R. ... 3.
PVT 1 on. 0 22577.01:
T .9 ()I I l o 7.0.1,??501
. .-.... UCD .... I. 1.! 18356100512
1’. ..1T. . I. (... 0000000000
J o v 0‘ o 9 w .0 “120.05.00.50.
9.5. ’V' m 2. o
V. 6 I. van. 0 ‘1
IU.U It" 6 c . ‘
(.5 " ‘.b‘ II.
\ o 0 (v... N u L (.2
u fiT ~25 a. an. 0 0 V 3.362200 A :0.
3.... r...“ 101T... ......o...
. 1 T ... f. 11 0.3%.
or} To... LA. I. /K 000000000
. x. o ...... II WC.- . a 1. ssgo-O‘Sz’
.... L o - Colo 0 al.-
V .l. .5): 1..."..NIZ...
\L. ... .l .1 ”.....pug.’
-..“ “...? I. fvhfl‘) ..
'7 O .(V’ 05%. 0., 0
. n17 fLo .. n. YELHLFV... .. ‘.227I.21..1..
9.5 1a.... "ERGO... n.- ...-......
C... 30...! J HF 1C1COT .5 tit—...:
Ni O bru OI r o n T. s .-
Tr. 1...]! o. “(131% O ‘14.}? 3.1.9
«I v Iii! J L H . . ... ... 23.301.73.505
. . .7 ) v . v v T511... . .... l 10%.-536:9...
u. . . 0.l!) . .0 N ... h )1 ... 2S..7‘.°I.9.3‘.
..IC 1....5. .. OTOMIWT.M.2.I 9 00.00.0000
I u h . .... ... ..v. a O 0 5.5.0:. 71.11....
v v. 0TH... N.T FTFTOT D -—- ... ..........1............
DD. 1.11),: .l D‘s-To .J 1 . 0. h.o.b.b.t.$.§b“f.
1': 1-..... 2 159.7 f r‘: .1 . L coco-0000000
5T .. h 0 OK 1 Cl T..l1rT-L‘O I. 3 ... ... .’o
1.”! ‘110 I LLNPN N 0.. .. I no ...; P D.
T v 9 l. . ... 1 ...........IT!.TT< .. ... 0 n V
45' T...) o 1 T.... TIT! I. I . .. 1 a.
..JD '“31’ 0 do.“ £5.09:U' , , . L a.
c .. o X 0" .J .u ...‘F.‘N1N H OT 1.. on 1 ... a. ... H.
... IV 31:54! . ...!L.L°t.n\<..l|. 7 O T. L a. V I»! I. S V
3 c. 0 01.2.7. 0 NH .... .. a... ... o ... I. I. . .15 ... .v ....C‘... .9.‘...§7
.. .J .-.. INV . 0.. ..4 . L‘ . .1 I1 ... u I 1 b 0557...... .7133?
. . .J 3.1!. .V s .. A .NIJAOTZ D? a... T .. .J 1.).. .J a. 10 .J A N 21.39 1160.....‘a.
....J .vU ....v 6 LTHOCOC1H. FIG. I. ’2 T T; T I ..T 7 .. ...... none...
1 . 0 J 9.: 3 o 9 .1... .. .....1‘. .o o -2... s T o I ...... ... .l O T D.C.... .- 5b.. ('5...
...H . ... ...) v ...T Tn.-~((.. 7.. .T .30 ...1 00 J . .... n )0 0 v ...
I u I o .. ...l. a O ..I. 81.5151... 0 is .6. 1.? I: not 1;... ‘9. 1x... rt 2‘. (r A U 0
s . . ...: (L). I... . . ...... h F" Vito! . 00 .59.. 20 ‘2 I To. I 0 . D l 1.
v .I 0 Pl...‘ 1. v o P 15.10;»... 6“. 0|. T 0 TI ‘2 0.! 9. ..TI .1 .u
. ..I .‘v I I. . c. . .120 n. .9. .K on..." 13 . .. .... f. 0 .... o D... 1 .
L oVC l/LCSD... o I. I...(.v-.va.oo3t.o oIvTv.I .U C. ...). ‘0 an 7 .o . .... ...... (...... r...
3 L 0 T1 (....TC... .- Ia ...lT 7.1/0 «1.1.9 0P1 511 o... 0.1. I ... IL .0 o a ... o. . . . . . . . . . .
.0 .0 . .1 ll. . ... I ...! .... 01:81.. 19(‘39I 107.510 . 0 T o .I o 1'- 0 V I .l. 0.. ... u 2. FI.E..0.C..€I.:.
....I...9......V. a... at... v. ...v ..VC.N .1. ... lt..’.../h..9 1.09.1... -0 z 0. I! o s 0.7...." N ‘1 ... 00.00000.
N. . 3.7.. CN’Clo.(1 0 .vv ... .- .II...N3’II:IT1 .0 3.2.1.6. .(ITCI..¢. 1... I33. 2,. .v 2". D. on. o o . 0 Ca 5511...: . .
. .4.-. ...... D a). .. .r. c... .... 5.. ... ... ...1..V ... cl .1... .1)! J..(.ISO o . 159.1103TT TT .....J).T '0! ...I.‘ o . v. .110... ... ... ... ...“ ... ... ... ... ... 95:. L.
7 SIN/[:1 1......:CJ.....8.. U U Uqu.)T’(l:.. T I. C. /.ool(c.Ctfi.1 .....1.)... n.U.T .15... ..UQ...To/OT.. U U U UT U U U U U 3 ... o
-.N ..«JNH...)T...(1 N: ......\NNNP.n.1-J..1J1.J ..J..U:Obv.noc..1T..3....... ...-... ....I..£..NIQ.II...S.....NOIL1L.U.....V‘V‘VNNV.&VNKNNNNNNVN o c...
....T -...OCN. ...: ... 6...... (3:23.19 1.. 1X. ... (.... .. .. I... b. H. 0 (31C... 3T1TNI1§.2TN3§H..‘P 11.101.- 1011110....1‘10131‘10. .
.11 ... .1N........D ... ....r 711.. 3 Jv 1T JT .. :JJlJ? ... .... 1.. ...(1. .3321... ....TZTTTTTI.‘ . TTTT TTII. .TTI..(I..I9.’.JTJY Jo JT..JTJT Jo-JTITJ 2 ...
...!!VVNNHIIDTia VIJIIV NTQTNTYLaLILTCTL L LCI.((T(.\.......L.. S 5!... .....NS‘ .. 32.1.. .. ..LILTTTVTNTNY ‘TTNY ‘TITNTNT... 1.5.3333336 1......
UOTGV 0.. ...Y. N5. .UCOUTT ..OLLECLH!..I2! .7? L11....... .. .. .. :E'Cfit.hu.0|..v.l$‘.rlr.0.t.r .PFUQRKFBLI1TIEOIOBEOF1015tthEL o . o . . . . . . . . . . .
«Pt-1.! (tn. 9?. 1.9 TGCNNIJFSCFECT((UPCiTLVOYDQZLfiIE CFFTCGTQTN‘. INT.TN’iNE’EXgI'ECPESFECVCFCFCS.... B P (.L111§11l L...
I151. ..Z'. N C
1 7 I. ... 1 ..U. 10. H H... ..u n P. uh “... .... B 5 .1... 0. I. nu ..(a ..L....J... .. ... a 1......
a 1 ’ I. 1‘2
0. ... o. 2 ..., o. 2 ... O «up...un!.w...«..nt’.u.......u7 .
1 ... ... . o ......c...’.\.....
SE Q 00000000000
61‘.- 00‘ 111.1.

APPENDIX C

APPLICATION OF THE
COMPUTER PROGRAM MINIQUAD 76A TO THE
DETERMINATION OF EQUILIBRIUM CONSTANTS
FROM POTENTIOMETRIC DATA

I41

I42
APPLICATION OF THE COMPUTER PROGRAM MINIQUAD 76A TO THE DETERMINATION
OF EQUILIBRIUM CONSTANTS FROM POTENTIOMETRIC DATA

C.I. PROGRAM MINIQUAD 76A

This program (136) determines the formation constants from
potentiometric data. Up to 20 equations involving 5 reactants
can be considered at the same time, and up to 3 electrodes can be
used to determine the concentration of free species.

The equation constants are calcuiated for the reaction

+
aA + bB + CC +’ AaBch (65)

with the formation constant

 

A B C
f = [aa bbC] c (66)
[A] [B] [C]
In such a case the reaction
LHZ’ + Mg2+ Z MgLH (67)

with the fbrmation constant

M LH
Kf(MgLH) = "9 LH (68)

becomes
L3” + H+ +Mg2+ Z MgLH (69)
with
. M LH
or
, K M LH
Kf (MgLH) = f( g ) (71)

K3IPAA5

 

 

l43

With the computer program MINIQUAD 76A, the formation
constants can be either held constant or refined in the calculations.
The coulometric (volumetric) titration data consist of the measured
electrode potential as a function of the time during which the
titrant was generated (the volume of titrant added). In these
acid base titrations, one of the mass balance equations describes
the concentration of species involving dissociable hydrogen ions.
The hydroxide ion is considered as a "negative proton", and the
proton concentration in a basic titrant is equal to the negative

value of the hydroxide ion concentration.
C.2. DATA INPUT INSTRUCTIONS FOR MINIQUAD 76A
C.2.l. 1 card /20A4/ : descriptive title

C.2.2. 1 card /815/ : LARS, NK, N, MAXIT, IPRIN, NMBEO.
NCO, ICOM.

LARS is an indicator for the data points to be considered in
the refinement: with LARS=l all the data points are used, with
LARS=2 alternate points, with LARS=3 every third point, 239. (last
points on all titration curves are always used).

NK is the total number of formation constants.

N is the number of formation constants to be refined.

MAXIT is the maximum number of iteration cycles to be
performed: with MAXIT=0 and according to the values of JPRIN and
JP (see below) the residuals on mass balance equations and/or the
species distribution are evaluated for the given formation constants

and conditions.

 

T44

IPRIN=O is normal; IPRIN=l monitors the progress of the
refinement at each cycle; IPRIN=2 produces an additional listing of
the experimental data at each titration point.

NMBEO is the total number of reactants (mass balance equations)
in the system under consideration.

NCO is the maximum number of unknown concentrations of free
reactants; if NCO=O the whole job is abandoned before refinement.

ICOM=O is normal; with ICOM=l data points are eliminated before
the refinement if the corresponding block of the normal equation
matrix is found to be not positive-definite.

C.2.3. 1 card /3Fl0.6, 8X,IZ/ : TEMP, ADDTEMP, ALPHA,

NOTAPE

TEMP is the reaction temperature in °C.

ADDTEMP is the titrant temperature in °C.

ALPHA is the coefficient of cubical expansion for the solvent
used, °C’].

NOTAPE=O is normal; NOTAPE=l reads values for EZERO and SLOPE
(see below) from device TAPE3. This allows calibration curve data
to be calculated and used in the same computer run.

C.2.4. NK cards /FlO.6,715/ : BETA(I), JPOT(I), JQRO(J,I)
(NMBEO values), KEY(I)

The formation constants are expressed in exponential notation
B] = BETA(I) - ioJP0T(I) -
JQRO(J,I) (J=l, NMBEO) are the NMBEO stoichiometric coefficients

of the 1th species with formation constant Bi' The order of

coefficients is arbitrary, except that those referring to reactants

 

l45
of which the free concentration is determined potentiometrically
must come last. Such a choice implies that a progressive integer
number (from 1 to NMBEO) is assigned to each reactant.
KEY(I) is the refinement key of the 1th formation constant:
with KEY=O the formation constant is not refined and with KEY=l

the formation constantis refined.

C.2.5. The following set of cards for each titration curve:
1 card /1215/ : NMBE, JNMB(I) (NMBE values),
NC, JP(I)

NMBE is the number of reactants (mass balance equations)
involved in the titration curve.

JNMB holds the integer numbers previously assigned to the
NMBE reactants involved.

NC is the number of unknown free concentrations at each
point of the titration curve; the number of concentrations experi-
mentally determined (j;g;, the number of electrodes) is
NEMF=NMBE-NC.

JP contains integer numbers corresponding to selected reactants:
in the subroutine STATS the formation percentages relative to these

reactants will be calculated, depending on the value of JPRIN.
1 card /5AlO/ : REACT(I) (NMBE values)

REACT contains the names of the reactants, listed in the same

order as JNMB.

146
1 card /4IS/ : JEL(I) (NEMF values), JCOUL

JEL holds the number of electrons transferred at each
electrode. If the decimal cologarithm of concentration (QLQL-
pH) is to be read in, put JEL(I)=O.

JCOUL=0 is normal; JCOUL=l if the total volume of the
solution does not change during the titration (gEQL, coulometric
experiments).

l (or 2) card(s) /8FlO.6/ : TOTC(I) (NMBE values),
EZERO(I) (NEMF values),
ADDC(I) (NMBE values),
VINIT

TOTC contains the initial number of millimoles of reactants
in solution; the order of reactants is the same as in JNMB.

EZERO(I) holds the standard potential of the 1th electrode
(mV); the value is ignored if JEL(I)=O.

ADDC contains the molar concentrations of titrant solutions
(there is one for each reactant); the order of the reactants is
the same as in JNMB.

VINIT is the initial volume of the solutions (cm3), and should

correspond to the volume expected at the temperature of the TITRANT.
1 card /8FlO.6/ : SLOPE (NEMF values).

SLOPE contains the slopes of the calibration curves for the
species measured, in units of mV per decade of concentration, the

value is ignored if JEL(I)=O.

 

l47
cards /15,8F8.3/ one for each point of the

titration curve: LUIGI, TITRE(I)
(NMBE values), EMF(I) (NEMF values)

LUIGI=O is normal, LUIGI=l indicates the end of a titration
cruve, LUIGI<O indicates the end of all titration curves, LUIGI=2
indicates that, for coulometric titration, current (mA) and fractional
current efficiency are read instead of a data point.

TITRE contains the volumes of titrant solutions (cm3) added
in volumetric titrations or time of current passage (sec) in
coulometric experiments.

EMF contains the potentials (mV) measured on each electrode
with non-zero JEL value (otherwise the decimal cologarithms of

concentration).
C.2.6. 1 card /IS/ : JPRIN

JPRIN controls the amount and type of output produced by STATS:

 

Statistical
JPRIN Analysis Tables Graphs
0 no no no
1 yes no no
2 yes yes no
3 yes no yes
4 yes yes yes

If JPRIN>l, the amount and type of tables and/or graphs is determined

by the values contained in JP for each titration curve.

C.2.7. 1 card /15/ : NSET

NSET=l for another set of formation constants - items C.2.l. -
C.2.4., C.2.6. and C.2.7. only -; NSET=O for another complete set

of data, NSET=-l for the termination of the run.

 

 

 

I48

DATA SAMPLE

C.3.

{til
. b

.
=4

.... (Lot-.....coz

{a
TA

- ,1. 0......Is..b ...
'- \4

rd
.-
I... ooo;!t1..).la$.~
1
...»
K
n.
,I L L $In . . .9
»\.L O
..r.

a. S. I-
n. .-
..1 I... .1 ‘4... 3......
0 ..II I.
.
v

.0
n O c

n t I 0 O
' ,- Or f
1 v Ls 1 J
.o . . . 1 a
‘ a O

..

Co 0. 35.06

6.

1
‘
‘
0
Q .

'IQ
‘-
u
r.

.c,
c
.9

O

...

55.‘t

 

fl.Sluggaala,93l.’°’55$550~951.5553‘.!I.9959956‘9I.°
...U.J b.UaJ.Lf,-. ..U-LPI. .1? .0325. .306. . 4°}! co; urtdsolqsflucat
oooocoonocoou...-cocoounocucocoa-......ooou
o‘c‘921“‘oz‘io*.‘:;§7\*i-i‘31.
VVVVVVV'VJ"'VUV'V'VV'VV"'"l'ivvvvvvvv‘vv

E04615... .t‘.... E. ..’...1.a... 3:53,! .5 ...; :3....7.5v....17n . 1339!... 3355....

OOOOOOOOOOOOOOOCOIDIOIOOOOO00.00.000.000...
.1 filth; . Clio-t. ‘5‘... Lt.r~‘....b..tt 7695...?» 4.75.. e: 31“...
. .. 2 {Fitzbl’ ......B ctiuz. 51.5.5351... 27....h5A . 112.335.33.79...
~L2~s§¢1..101..¢5Qo..110c511¢a.2110 . o o o o o o o o o o

. ..7.as.3.a7:1 .2. o? aSJaiacr P... v7? ’5. .51’5’552‘ 315613.“.‘3
. .ru)? 05.. . s. ....5 ~. .... cat's-1192.: . u, :2... II): ~. . L. 28:6. - Stole-Er.
IO0..........OI......‘II.......‘OCOOOCCO...
uOCC—LOOJDCD... ruck-s... C50...CSC:.O...C¢..USO¢. 650.25.553.3— C
.3. 3.102113... 05.51.... ’3)..), .4323, . o3..77-J;37 1,.."J)..,’ .d):”.d
137.1531... 5.7.9; .9 b n. b O71 77 3.31. 00 Q: .c..1....)~2223333 b
1113111102....111:

00.00.000.000000000000000... 0
5L ..Ltlfikslkkt 5....ch5 C‘s SI. LLL,5L.L.~cr-.0 3.3

Q ......IOOCOIIOIOOOIOOIQII......‘OI........I
7 ...-... ......I p. a. .. oo o. .- —‘: o. .... o. ..J... ..vnc-b.o.OQ-..Jqoo ..5.‘-.Jo varufJQOoo
QC
0
I.

. sf. . :5 5"» ......» ...; etc LL.L.§L LLU‘ hLLI- ..va tt‘tt‘t‘OIL;

 

 

 

BIBLIOGRAPHY

 

IO.

11.

12.

13.

14.

15.
16.

149
BIBLIOGRAPHY

(a) A. I. Popov, Pure and Applied Chem. 31, 3 (1975).

(b) w. J. Dewitte, R. C. Schoening and A. I. Popov, Inorg.
Nucl. Chem. Lett., 12, 251 (1976).

(c) J. 5. Shih and A. I. Popov, Inorg. Nucl. Chem. Letters,
1_3_, 105, (1977).

F. Alder and F. C. Yu, Phys. Rev., B2, 105, (1951).

M. Ellenberger and M. Villemin, C. R. Acad. Sci. Paris, Serie
B, 266, 1430 (1968).

P. D. Dougan, S. N. Sharma, and D. Llewelyn Williams, Can. J.
Phys., 51, 1047 (1969).

E. M. Dickson and E. F. u. Seymour, J. Phys. Chem., g, 666.
(1970).

FL G.Bryant, J. Magn. Reson., 6, 159 (1972).

J. A. Magnusson and A. A. Bothner—By in "Magnetic Resonance

in Biological Research," A. Franconi, Ed. Gordon and Breach,
London, 1969, p. 135.

L. Simeral and G. E. Maciel, J. Phys. Chem. §Q, 552, (1976).

J. H. Swinehart and H. Taube, J. Chem. Phys. 31, 1579, (1962).

N. A. Matwiyoff and H. Taube, J. Am. Chem. Soc., 99, 2796,
(1968).

R. D. Green and N. Sheppard, J. Chem. Soc. Faraday Trans. II
§§, 821, (1972).

F. Toma, M. Villemin, and J. M. Thiery, J. Phys. Chem., 11,
1294, (1973).

A. D. Covington and A. K. Covington, J. Chem. Soc. Faraday
Trans. 1, 11, B31, (1975).

R. N. Butler, J. Davies, M. C. Symons, Trans. Faraday Soc.,
66, 2426, (1970).

B. N. Chernyshov, V. V. Kon'shin, Koord. Khim. g, 365, (1976).

E. C. Ashby and R. C. Arnott, J. Organometal. Chem., 13,
l, (1968).

 

 

17.

18.

19.

20.

21.

22.

23.

24.

25.
26.
27.

28.

29.

150

(a) Nouveau traite de chimie minerale, Pascal, Masson et Cie
Editeurs, 1958, tome IV, p. 170, and references therein.

(b) Handbuch der anorganische chemie, Gmelin, Berlin, Verlag
Chemie GMBH, 1939, 27, teil B, p. 140 and references therein.

(c) Ikeuchi and Krohn, Acta Chem. Scand. 23, 2232, (1969).
(d) D. Bryce-Smith, Inorg. Synthesis, Vol. 6, p. 9.

(e) H. J. Abendroth, A. Peters, w. Fischer, Rev. Chim. Min.,
u, 556, (1974).

(f) L. V. Gorbunov, V. N. Deciatnik, S. R. Raspopin, K. I.
Trifonov, Zh. Neorg. Khim., 12, 3093, (1974).

(g) u. D. Treadwell and Th. Zurrer, Helv. Chim. Acta, 15,
1271, (1932).

H. F. Halliwell and S. C. Nyberg, Trans. Faraday Soc., 59,
1126, (1963).

F. L. Cook, T. C. Caruso, M. P. Byrne, C. N. Bowers, D. H.
Speck, and C. L. Liotta, Tetrahedron Lett., 46, 4029, (1974).

M. A. Neuman, E. C. Steiner, F. P. van Remoortere, and F. P.
Boer, Inorg. Chem., 14, 734, (1975).

R.D. McLachlan and V. B. Carter, Spectrochim. Acta, Part A,

39, 1171, (1974).

A. Pullman, C. Giessner-Prettre, and Y. V. Kruglyak, Chem.
Phys. Lett., 35, 156, (1975).

F. A. L. Anet, J. Krane, J. Dale, K. Daasvatn, and P. 0.
Kristiansen, Acta Chem. Scand., 21, 3395, (1973).

B. K. Leong, T. 0. Ts'o and M. Chenoweth, Toxicol. Appl.
Pharmacol., 22, 342, (1974).

P. B. Chock, Proc. Nat. Acad. Sci. US, 62, 1939, (1972).
B. K. Leong, Chem. Eng. News, 4, 5, (1975).

B. Dietrich, J. M. Lehn and J. P. Sauvage, and J. Blanzat,
Tetrahedron Lett., 2885 and 2889, (1969).

B.Dietrich, J. M. Lehn and J. P. Sauvage, and J. Blanzat,
Tetrahedron, gg_1629, (1973).

J. M. Lehn and J. P. Sauvage, J. Amer. Chem. Soc., 22,
6700, (1975).

30.
31.
32.

33.

35.
36.
37.

38.
39.

40.
41.

42.

43.

44.
45.

46.
47.

48.

151

w. E. Morf and w. Simon, Helv. Chim. Acta, 54, 2683 (1973).
J. M. Lehn, Structure and Bonding, l5, 2 (1973).

J. J. Christensen, 0. J. Eatough, and R. M. Izatt, Chem.
Rev., 24, 351 (1974).

J. 0. Lamb, R. M. Izatt, J. J. Christensen, and D. J. Eatough,
"Thermodynamics and Kinetics of Cation - Macrocycle Interaction,"
Chapter in Chemistry of Macrocyclic Compounds, G. A. Melson,

ed., Plenum, N.Y. in press.

0. N. Bhattacharyya, C. L. Lee, J. Smid, and M. Szwarc, J.
Phys. Chem., 55, 608 (1965).

J. S. Shih, Ph.D. Thesis 1978, Michigan State University.
J. G. Hoogerheide, Ph.D. Thesis 1978, Michigan State University.

G. J. Templeman and A. L. Van Geet, J. Am. Chem. Soc., 54,
5578 (1972).

D. H. Live, 5. I. Chan, Analytical Chemistry, 42, 791 (1970).

M. S. Greenber , R. L. Bodner and A. I. Popov, J. Phys. Chem.,
.22, 2449 (1973).

U. Mayer and V. Gutmann, Structure and Bonding, 12, 113 (1972).

(a) V. Gutmann and E. Hychera, Inorg. Nucl. Chem. Lett., 2,
257 (1966).

(b) V. Gutmann, "Coordination Chemistry in Nonaqueous Solvents"
Springer-Verlag, Vienna (1968).

M. S. Greenberg, D. M. Nied and A. I. Popov, Spectrochim. Acta,
29.6.. 1927 (1973).

Y. Cahen, Ph.D. Thesis 1975, Michigan State University, East
Lansing, Michigan.

A. I. Popov, Pure and Applied Chem., 42, 3 (1975).

A. I. Popov, "Polyhalogen Complex Ions" in Halogen Chemistry, vol. 1
V. Gutmann, ed., Academic Press, London and N.Y. (1967), p. 253.

R. H. Erlich and A. I. Popov, J. Amer. Chem. Soc., 93, 5620 (1971).

M. Greenberg, Ph.D. Thesis 1975, Michigan State University, East
Lansing, Michigan

Y. M. Cahen, P. R. Handy, E. T. Roach, and A. I. Popov, J. Phys.
Chem., 22, 80, (1975).

 

49.

50.

51.

52.
53.

54.
55.

56.

57.

58.
59.
60.
61.

62.

63.
64.
65.

66.
67.

68.

69.

152

T. C. Farrar and E. 0. Becker, Pulse and Fourier Transform
NMR, Academic Press, N.Y. and London, (1971), p. 65.

R. H.)Er1ich and A. I. Popov, J. Amer. Chem. Soc., 93,5620
1971 .

J. S. Shih and A. I. Popov, Inorg. Nucl. Chem. Letters, 19,
105 (1977).

E. Kauffmann, personal communication.

K. Balasubrahmanyam and G. J. Janz, J. Am. Chem. Soc., 92,
4189 (1970).

P. Nylen, Ber., 979, 1032 (1924).

M. I. Kabachnik and T. A. Mastryukova, Izvest. Akad. Nauk.,
S.S.S.R., 163 (1953).

A. A. Shvets, 0. A. Osipov, and A. M. Shakirova, J. Gen.
Chem., USSR, 92, 2588 (1967).

E. M. Popov, M. I. Kabachnik, and L. S. Mayants, Russian
Chem. Rev., 99, 362 (1961).

L. J. Bellamy and L. Beecher, J. Chem. Soc., 475 (1952).
L. J. Bellamy and L. Beecher, J. Chem. Soc., 729 (1953).

. J. Harvey and M. G. Horning, J. Chromatogr., 29, 65 (1973).

D
R. J. Maile, Jr. G. J. Fischesser and M. M. Anderson,
J. Chromatogr. , 132,366 (1977).

H

1.1.

. Sohr, J. Electroanal. Chem. Interfacial E1ectrochem.,
. 188 (1966).

D. Balzer and H. Lange, Colloid Polym. Sci., 259, 643 (1975).
D. Balzer and H. Lange, Colloid Polym. Sci., 255, 140 (1977).

w. Althoff, M. Fild and H. P. Rieck, Z. Naturforsch. B. Anorg.

Chem. Org. Chem., 919, 153 (1976).
G. A. Gray, J. Am. Chem. Soc., 99, 2132 (1971).

J. G. Riess, J. R. Van Nazer, and J. H. Letcher, J. Phys. Chem.

11, 1925 (1957).

T. Bottin-Strzalko and J. Seyden—Penne, J. Chem. Soc.,Chem.
Commun., 22, 905 (1976).

L. L. Burger, J. Phys. Chem., 62, 590, (1958).

 

 

 

 

70.

71.

72.
73.

74.
75.
76.
77.
78.

79.

81.

82.

83.
84.

85.

86.

87.

88.
89.
90.

91.

A. A. Elesin, A. A. Zaitsev, S. S. Kazakova, and G. N. Yakovlev,

153

Radiokhimiya, 14, 541 (1972).

A. A. Elesin, A. A. Zaitsev, S. S. Kazakova, and G. N. Yakovlev,

Nuc1. Sci. Abstr., 25, 18030 (1973).

S. M. Lambert and J. I. Hatters, J. Am.Chem. Soc., 29, 4262 (1957).

.1.
Soc., 7_9, 3651, (1957).

I. Hatters, S. M. Lambert and E. D. Loughran, J. Am. Chem.

J. Mao, E. Robishaw, Biochemistry, 14, 5475 (1975).

J. F. Cannelongo, U.S. Patent #3,370,029.

Th. Auel, Ger. Offen. #2,505,435.

Anonymous, Neth. Appl. #6,508,180.

G. T. Bottger and A. P. Yerington, U.S. Dept. Agr., Bur.
Entomol. Plant Quarantine E-863, (1953).

E.

E.

Nl—Ca—i

84

F.

manor-TI

Rogers, A. A. Patchett, Ger. Offen. #2,318,254.
Rogers, A. A. Patchett, Ger. Offen. #2,318,254.

. Clark,U.S. Patent #4,016,264.
. Herrin, J. S. Fairgrieve, U.S. Patent #4,052,439.
. Schleicher, w. R. Roderick, Ger. Offen. #2,208,787.

Overby, R. G. Duff, J. C. Mao, Ann. N.Y. Acad. Sci.,
310 (1977).

J. A. Boezi, in press.

J. Hay, S. M. Brown, A. T. Jamieson, F. J. Rixon, H. Moss,

0. A. Dargan and J. H. Subak-Sharpe, J. Antimicrob. Chemother.,

9, Suppl. A, 63 (1977).

S. S. Leinbach, J. M. Reno, L. F. Lee, A. F. Isbell, and J. A.

Boezi, Biochemistry, 15, 426 (1976).

B. H. Fox, J. Antimicrob. Chemother. 9, Suppl. A, 23, (1977).

1. H. Maugh II, Science. L92. 128.

R. J. Klein and A. E. Friedman-Kien, Antimicrob. Agents Chemother.,

_7_, 289 (1975).

H. G. Adams, E. A. Benson, E. R. Alexander, L. A. Vontver,

M. A. Remington, K. K. Holmes, J. Infect. Dis. 199 (Suppl.):

A151-A159 (1976).

 

 

 

 

92.

93.

94.

95.

96.

97.

98.

99.

100.
101.

102.

103.
104.
105.
106.
107.

108.

109.
110.
111.

154

E. L. Goodman, J. P. Luby, M. T. Johnson, Antimicrob. Agents
Chemother. 9, 693 (1975).

A. E. Friedman-Kien, A. A. Fondak, and R. J. Klein, J. Invest.
Dermatol., 55, 99 (1976).

D. D. Gerstein, C. R. Dawson, and J. 0. 0h, Antimicrob. Agents
Chemother., 2, 285 (1975).

R. F. Meyer, E. D. Varnell, and H. E. Kaufman, Antimicrob.
Agents Chemother., 9, 308 (1976).

J. F. Fitzwilliam and J. F. Griffith, J. Infect. Dis., 199,
Suppl. A221, (1976).

N. L. Shipkowitz, R. R. Bower, R. N. Appell, C. W. Nordeen,
L. R. Overby, W. R. Roderick, J. B. Schleicher, and A. M.
von Esch, Appl. Microbiology, 25, 264 (1973).

R. J. Klein and A. E. Friedman-Kien, Antimicrob. Agents
Chemother., 2, 289 (1975).

J. Roboz, R. Suzuki and G. Bekesi, and R. Hunt, Biomed. Mass.
Spectrom., 4, 291 (1977).

B. A. Bopp, Ch. 8. Estep, D. J. Anderson, Fed. Proc., 95, 939 (1977).

. R. Herrin, J. S. Fairgrieve, R. R. Bower, N. L. Shipkowitz,
. Mao, J. Med. Chem., 29, 660 (1977).

. F. Lee, K. Nazerian, S. S. Leinbach, J. M. Reno, J.A. Boezi,
. Natl. Cancer Inst., 55, 823 (1976).

. Nylen, Z. Anorg. Allg. Chem., 295, 33 (1937).
. Warren and M. R. Williams, J. Chem. Soc., (B), 618 (1971).

. Denham, U.S. Patent #2,629,731.

. R. Naqvi and P. J. Wheatley, and E. Foresti-Serantoni,
. Chem. Soc. A, 12, 2751 (1971).

. M. Reno, L. F. Lee, J. A. Boezi, Antimicrob. Agents Chemother.,
3. 188 (1978).

A. Unni, L. E1135, and H. Schiff, J. Phys. Chem., 52, 1216 (1963).

T
J
L
P
S
P. Nylen, Ber., 529, 1036 (1924).
H
R
J
J
1

D. Duv111ier, Ann. Chim. Phys. 29, 327 (1881).
B. Schmidt, Ann., 252, 265 (1892).

 

112.

113.

114.
115.
116.
117.
118.

119.

120.

121.

122.

123.

124.

125.

126.

127.

128.
129.
130.
131.

155

Handbook of Chemistry and Physics, Chemical Rubber Company,
ed., (1967-8) p. F-4.

M. M. Crutchfield, C. F. Callis, R. R. Irani and G. C. Roth,
Inorg. Chem., 1, 813 (1963).‘

R. Jones and A. Katritzky, J. Inorg. Nucl. Chem., 15, 193 (1960).
J. H. Letcher and J. R. Van Wazer, J. Chem. Phys., 44, 815 (1966).
J. H. Letcher and J. R. Van Wazer, J. Chem. Phys., 45, 2916 (1966).
J. H. Letcher and J. R. Van Wazer, J. Chem. Phys., 45, 2926 (1966).
J. H. Letcher and J. R. Van Wazer, Chapters 2 and 3 in Phosphorus
31 Nuclear Magnetic Resonance, M. Grayson, E. J. Griffith, ed.,
Interscience Publishers, N.Y. and London, (1967).

K. Moedritzer, Inorg. Chem., 5, 936 (1967).

Carbon-13 NMR for Organic Chemists, Levy Nelson, ed., Interscience
(1972), p. 23.

J. C. Tebby, Organophosphorus Chemistry, vol. 1, Specialist
Periodical Reports, The Chemical Society, ed., London (1970),
p. 294.

G. Mavel and M. J. Green, Chem. Comm., 742 (1968).

J. C. Tebby, Organophosphorus Chemistry, vol. 2, Specialist
Periodical Reports, The Chemical Society, ed., London (1970)
p. 252.

R. G. Bates, Determination of pH Theory and Practice, J. Wiley
and Sons, Inc., N.Y., London, Sydney, (1964), p. 53.

H. S. Harned and B. 8. Owen, The Physical Chemistry of Electrolytic
Solutions, Reinhold Publishing Corp., N.Y., 3rd edition, (1958),
p. 165.

L. Meites, Handbook of Analytical Chemistry, McGraw-Hill Book Co.,
N.Y., San Francisco, Toronto, London, Sydney, Tst edition, (1963),
p. 1-21.

C. Cotton and J. Wilkinson, Advanced Inorganic Chemistry,
Interscience, 3rd edition, p. 848.

J. J. Lingane, Chem. Rev. 29, 1 (1941).

D. D. DeFord and D. N. Hume, J. Am. Chem. Soc., 29, 5321 (1951).

S. Fronaeous, Acta Chem. Scand.,4, 72 (1950).

H. S. Harned and B. B. Owen, the Physical Chemistry of Electrolytic

Solutions, Reinhold Publishing Corp., N.Y., 3rd edition, (1958),
p. 738.

156

132. L. Meites, Handbook of Analytical Chemistry, McGraw-Hill Book Co.,
N.Y., San Francisco, Toronto, London, Sydney, lst edition, (1963),
p. 1-46.

133. P.-H.)Heube1 and J. Hoogerheide, "Constant Current Coulometer,"
1977 .

134. J. L. Dye and V. A. Nicely, J. Chem. Educ., 49, 443 (1968).

135. F. A. L. Anet, J. Krane, J. Dale, K. Daasvatn, and P. 0.
Kristiansen, Acta Chem. Scand., 22, 3395, (1973).

136. A. Sabatini, A. Vacca, and P. Gans, Talanta, 21, 53 (1974).

137. D. D. Traficante, J. A. Simms and M. Mulvay, J. Magn. Reson.,
E, 484 (1974).

138. D. A. Wright and M. T. Rogers, Rev. Sci. Instr., 44, 1189 (1973).