. ’UTTLEZATTONOFZZE-‘CRYPT ASHA' f . “

NTEORALKALI ‘

 

' " SOLUBTUZING AG

 

. ' Arms LA ALALAES' AAA mans ' .. ~
"f: Nissertation fer the Degree 5f Ph; D. . T A
MICHTGAN STATE UNEVERSTTY , g ...

' AEHAALAA ’ - n
1973 _ ..  _,

 

 

  
     

 
   
  

-
iv! - ut-

n

 

 

  
 
  

  

' P‘ '7
”47h”, :”' “..'.".’, "' ""'
H? .u l‘flfrtllrfll an ”1'27; {XI

355“”

M.“ an
LIBRA R 1'

Michigan St; 2‘ - f
‘1 University [
I

This is to certify that the

thesis entitled
Utilization of 2,2,2—crypt as a solubilizing
agent for alkali metals in amines and ethers

presented by
Mei-Talc Lok

has been accepted towards fulfillment
of the requirements for

57‘
Ph. D. degree in Chemistry

   
 

Major professor

 
  

, alumna: 31s.? »

‘5 llllAlHL

.1 ml mum ulc. 11 .
’LIIIARV ‘

   

 

 

ABS TRACT

UTILIZATION OF 2.2.2-CRYPT AS A SOLUBILIZING
AGENT FOR ALKALI METALS IN AMINES AND ETHERS

BY
MeidTak Lok

We have successfully synthesized the bicyclic complex-
ing agent, 4,7,13,16,21,24-hexaoxa-1,10-diaza4bicyclo(8,8,8)-
hexacosane(2,2,2-crypt) in our laboratory. The procedures
used are similar to those of Lehnllz, with some important
modifications which make the synthesis faster, the yield
higher, and the product purer. The use of flow synthesis to
replace the traditional high dilution method for the cycli-
zation steps was developed.3 It was found that both the
yield and the quality are comparable for these two methods.
However, the flow synthesis can save about half of the
total time needed for the complete synthesis.

A potentiometric study of the protonation of 2,2,2-
crypt in water gives two equilibrium constants, 1.53 i

10

0.08 x-10' M and 6.3 i 0.3 x 10"8 M for the dissociation

of the first and second protons respectively, according to:

k
CH+ 3—3 C+H+

+ k
CH2+ ——2-> CH+ + H+

<

The titration with acid of the chloride salt of Na+-cryptate

'3 :21. 3 2;. 6/

Mei-Tak Lok
gives an equilibrium constant of 1.3 i 0.1 x 104 for the

reaction
K3

 

+
Na + C > CNa+

<

 

The titration also indicates that the protonation of the
crypt requires that the sodium cation leave the crypt. The
rsults from all titrations were fitted by a nonlinear least-
squares program. The agreement between the experimental
data and the calculated values was excellent.

The cation complexing agent, 2,2,2-crypt, was also
used to enhance the solubility of alkali metals in amines
and ethers.4o5 The Optical spectra of a number of solutions
of alkali metals in amines and ethers were obtained. The

peak positions for solvated electrons (e- and alkali

solv)
anions (MT) were measured. The temperature dependence of
the peak maxima for M- anitmesolvated electron in most of
the solvents was determined. For each of the three Species,
the wave number of maximum absorbance decreased linearly
with an increase in temperature. Linear correlations were
found for the changes in the peak positions of e-solv’ Na-,
K- and Cs- with solvent at a given temperature. No satis-
factory correlation between the peak position of Na- and K-

and solvent properties has been found. The spectrum of

e solv has now been determined by dissolving metals as well

as by flash photolysis or pulse radiolysis techniques, for
the solvents ammonia, ethylenediamine, 1,2-pr0panediamine,
diethylamine, ethylamine, methylamine, diethyl ether, tetra-

hydrofuran, dimethoxyethane and diglyme. The agreement

MeidTak Lok

between the trends in peak positions in all cases is good.

An ESR study was carried out to test the importance of
the effect of the cation on the spin-pairing process in
dilute metal-ammonia solutions. Samples were surrounded
concentrically by a solution of 2,2-diphenyl-1-picryl-
hydrazyl radical (DPPH) in carbon disulfide. The DPPH solu-
tion was used to serve as a relative spin standard. The
relative intensities of the single ESR line of the solvated
electron in solutions of Na, K, Rb, and Cs in liquid ammonia
were measured as a function of temperature. If the nature
of the cation were important in the equilibria (for example
through species of stoichiometry M- and M2) one would ex-
pect substantial differences in the temperature dependence
of the esr signal intensities for Na, Rb and Cs relative to
that of K in the dilute concentration range (0.02-0.1M).
The results show that this ratio is not very sensitive to
metal, temperature or concentration and that it remains
constant to within at least 50 percent over the temperature
range -700 to -5°C. These results indicate that the tempera-
ture dependence of the spin-pairing equilibrium in metal-
ammonia solution is relatively insensitive to the nature of
the cation.

Two new kinds of solids were formed by combining metals
and 2,2,2-crypt. The first type of solid, which has a dark
blue color, is paramagnetic. It is formed by the combina-

tion of metals with dicyclohexyl-lB-crown—S and some metals

Mei-Tak Lok
(K, Cs, Ba) with crypt. The second type of solid is
formed with sodium and 2,2,2-crypt. This solid is gold in
color at low temperature. The color changes with changes
in temperature. From various composition studies, it was
shown that the solid has the empirical formula Na2C18H36N206.
Conductivity studies indicate that the solid is a semi-
conductor with a room temperature resistivity of about
1013 ohm-cm. The solid is indefinitely stable ig_y§gug_at
low temperatures. It reacts vigorously when it is exposed

to air or water.

REFERENCES

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

 

2. B. Dietrich, J. M. Lehn and J. P. Sauvage, Tetrahedron
22, 1629 (1973).

 

l

3. J. L. Dye, M. T. LOk, F. J. Tehan, J. Ceraso, and K.
Voorhers, J. Organic Chem., §§, 1775 (1973).

4. J. L. Dye, M. T. Lok, F. J. Tehan, R. B. Coolen, N.
Papadakis, D. M. Ceraso, and M. DeBacker, Ber. Bunsenges
Phys. Chem., 25, 681 (1971).

5. M. T. Lok, F. J. Tehan, and J. L. Dye, J. Phys. Chem.,
22' 2975 (1972).

UTILIZATION OF 2.2,2-CRYPT AS A SOLUBILIZING

AGENT FOR ALKALI METALS IN AMINES AND ETHERS

BY

Mei-Tak Lok

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Chemistry

1973

To

Jenny and My Parents

ii

ACKNOWLEDGMENT

The author wishes to express his deepest appreciation
to Dr. James L. Dye for his constant encouragement, guidance
and whole-hearted support which he so generously gave during
this research program.

Special thanks go to Dr. Frederick J. Tehan for so many
things we shared during the years we spent together and for
the valuable suggestions given on the collaborathe projects.
He would also like to thank Dr. Alexander I. Popov for being
second reader. Advice about this research from the research
committee members, Drs. Alexander I. Pepov, Ahsan Khan and
Gordon A. Melson is greatly appreciated. He would like to
extend his appreciation to J. M. Ceraso, R. B. Coolen, Dr.

L. D. Long, N. Papadakis and Miss E. Mei for their help and
suggestions in many ways.

Many thanks go to Drs.Richard H. Schwendeman and Barnett
Rosenberg for their help in providing the microwave spectrom-
eter and Faraday cage for conductivity studies.

Lasm,but not the least, many thanks go to my wife, Jenny,
and to my parents for their encouragement during these years.

Financial assistance from the Atomic Energy Commission
is acknowledged. A term of summer fellowship from The Dow
Chemical Co. is also acknowledged.

iii

TABLE OF CONTENTS

Page
I. INTRODUCTION . . . . . . . . . . . . . . . . 1
II. HISTORICAL PERSPECTIVE . . . . . . . . . . . 5
2.1. Metal-Ammonia Solutions . . . . . . . 5
2.1.1. General properties of dilute
metal-ammonia solutions . . . 6
2.1.2. Spin-pairing phenomenon in dilute
metal-ammonia solutions . . . 10
2.2. Alkali Metals in Amines and Ethers . . 14
2.2.1. General preperties . . . . . . 14
2.2.2. Major species involved in metal-
amine and metal-ether solutions 17
2.2.2.1. Solvated electrons . 17
2.2.2.2. Spin-paired species 18
2.2.2.3. Alkali anions . . . 18
2.2.2.4. Monomers . . . . . . 21

2.2.3. Problems involved in the study
of metal-solutions in amines
and ethers . . . . . . . . . . 21

2.3. Pulse Radiolysis and Flash Photolysis
Studies with Ammonia, Amines, and Ethers 22

2.3.1. Flash photolysis studies . . . 22
2.3.2. Pulse radiolysis studies . . . 24

2.4. Cation Complexing Agents—-The Macro-
bicyclic Diamine, Crypt . . . . . . . 25
2.5. Protonation of Amines . . . . . . . . 27
2.6. Solid State Studies . . . . . . . . . 29
III. EXPERIMENTAL . . . . . . . . . . . . . . . . 31
3.1. General Techniques . . . . . . . . . . 31

3.1.1. High vacuum techniques and

glassware cleaning . . . . . . 31
3.1.2. Metal purification . . . . . . 31

iv

TABLE OF CONTENTS (Continued)

IV.

V.

3.2.

4.1.
4.2.

4.3.

Solvent Purifications and Solution

Page

Preparation . . . . . . . . . . . . . . 34

SYNTHESIS OF 4.7,13,16,21,24-HEXA0XA—1,10-DI-
AZABICYCL0(8,8,8)HEXACOSANE ("2,2,2-CRYPT") . 37

Introduction . . . . . . . . . . . . . . 37
High Dilution Method . . . . . . . . . . 39

4.2.1.
4.2.2.

4.2.3.

4.2.5.

4.2.6.

Synthesis of starting materials. 39

Preparation of 5,12-dioxo-1,7,10,
16-tetraoxa-4,13-diazacycloocta-
decane (First cyclization) . . . 44

Preparation of 1,7,10,16-tetraoxa-
4,13-diazacyclooctadecane
(First reduction) . . . . . . . 45

Preparation of 2,9-dioxo-4,7,13,
16,21,24-hexaoxa-1,10-diazabi-
cyclo(8,8,8;hexacosane (Second
cyclization . . . . . . . . . . 47

Preparation of 4,7,13,16,21,24-
hexaoxa-l,10-diazabicyclog8,8,8)-
hexacosane ("2,2,2-crypt"

(Second reduction) . . . . . . . 48

Purifications of final product . 49

Flow Synthesis . . . . . . . . . . . . . 50

NMR AND TITRATION STUDIES ON 2.2,2-CRYPT AND
SODIIJM CRYNATE . . . . . . . . . O . . . . O 55

5.1.
5.2.
5.3.

Introduction . . . . . . . . . . . . . . 55

Experimental . . . . . . . . . . . . . . 56

Proton Exchange Studies . . . . . . . . 57

5.3.1. Titration studies . . . . . . . 57
5.3.1.1. Titration of 2,2,2—
crypt with HCl . . . . 57
5.3.1.2. Titration of Na+—
cryptate with HCl . . 61
5.3.2. NMR studies of protonation . . . 71

5.3.2.1. Protonation of 2,2,2-
crypt . . . . . . . . 71

5.3.2.2. Protonation of Na+-
cryptate . . . . . . . 77

Paramagentic Shift of Crypt Proton in

Metal-deuterated Ammonia Solutions . . . 82

V

TABLE OF CONTENTS (Continued)

VI. OPTICAL STUDIES OF ALKALI METALS IN AMINES AND
ETHERS . . . . . . . . . . . . . . O O O O O
6.1. Introduction . . . . . . . . . . . . . .
6.2. Experimental . . . . . . . . . . . . . .
6.3. Results and Discussion . . . . . . . . .

6.3.1. Spectrophotometric studies . . .
6.3.2. Pulse radiolysis studies . . . .
6.3.3. Other studies . . . . . . . . .
6.3.3.1. Solubility of potassium
in ethylamine . . . .
6.3.3.2. Na-KI-ethylenediamine
system . . . . . . . .
6.3.3.3. Cesium in methylamine
VII. STUDIES OF THE METAL DEPENDENCE OF THE SPIN-
PAIRING PHENOMENON IN DILUTE ALKALI METAL-
AMMONIA SOLUTIONS BY ESR TECHNIQUES . . . . .
7.1. Introduction . . . . . . . . . . . . . .
7.2. Experimental . . . . . . . . . . . . . .
7.3. Some Problems Involved in the
Measurement 0 o a o o o o o o o o o o o
7.4. DPPH Solution Studies and Some Prelimin-
ary Results for Metal-Ammonia Solutions
7.5. Saturation Problem . . . . . . . . . . .
7.6. Results and Discussion . . . . . . . . .
7.7. Preliminary Studies of Spin-Pairing
Process in Metal-Amine Solutions . . . .
7.7.1. Cesium-ethylenediamine solutions
7.7.2. Sodium-2,2,2-crypt-ethylamine
solutions . . . . . . . . . . .
VIII. SOLID STATE STUDIES . . . . . . . . . . . . .
8.1. Introduction . . . . . . . . . . . . . .
8.2. Solid Preparations . . . . . . . . . . .
8.2.1. Ba(NH3)6 and crown compounds . .
8.2.2. 2,2,2-Crypt compounds . . . . .

vi

Page

90

9O
93
95
95
114
123

123

125
126

128
128
129
133

134
136
140

158
158

159

162

162
163
163
165

TABLE OF CONTENTS (Continued)

IX.

Page

8.3. Na-2,2,2-Crypt Compound . . . . . . . . 168

8.3.1.

8.3.2.

8.3.3.
8.3.4.
8.3.5.
8.3.6.
8.3.7.

8.4. Summary

Physical appearance and

reactivity

. . . . . . . . . . . 168

Composition . . . . . . . . . . 169

8.3.2.1. Sodium analysis by flame
emission . . . . . . . 169
8.3.2.2. Existence of 2,2,2-crypt-
-NMR study . . . . . . 171
8.3.2.3. The ratio Of sodium and
2,2,2-crypt-—titration
study . . . . . . . . 172
8.3.2.4. Complete elementary
analysis . . . . . . . 173
Solubility . . . . . . . . . . . 173
Reducing power studies . . . . . 173
Stability . . . . . . . . . . . 180
Studies by ESR and NMR techniques 181
Conductivity studies . . . . . . 182

8.3.7.1. Two electrodes method. 182
8.3.7.2. Transformer current
methOd . . . . . . . . 184
8.3.7.3. Microwave conductivity
studies . . . . . . . 187
8.3.7.4. The use Of conducting

glass and aluminum
plates . . . . . . . . 189

. . . . . . . . . . . 196

SUGGESTIONS FOR FURTHER STUDIES . . . . . . . 198

APPENDIX . .

BIBLIOGRAPHY

. . . . . . . O . . O 201

. . O . . . . . . . . 203

vii

TABLE

1.

10.

LIST OF TABLES

Page

Computer generated data for 2,2,2—crypt pH
titration: concentrations Of 2,2,2-crypt and

its protonated species, ionic strength and
activity coefficients . . . . . . . . . . . . 63

Computer generated data for sodium ion-cryptate

pH titration: concentrations of protonated
2,2,2-crypt and sodium ion-cryptate, ionic
strength, and activity coefficients . . . . . 69

NMR peak positions for 2,2,2-crypt protons
in ND3 o o o o o o o o o o o o o o o o o o o 86

Solvents in which K- has been Observed . . 96

Peak positions and temperature coefficients for
Na , K‘, and eSOlv in various solvents . . . 102

Data from the pulse radiolysis study in ND3 . 118

Corrected area ratio Of the solvated electron

ESR Signal in 0.02M Na, K, Rb, and CS-ammonia
solutions to that Of the center DPPH radical

peak . . . . . . . . . . . . . . . . . . . . 151

Corrected area ratio Of the solvated electron

ESR signal in 0.04M Na, K, Rb., and Cs-ammonia
solutions to that of the center DPPH radical

peak . . . . . . . . . . . . . . . . . . . . 152

Corrected area ratio Of the solvated electron
ESR signal in 0.06M Na, K, Rb, and Cs-ammonia
solutions to that Of the center DPPH radical
peak . . . . . . . . . . . . . . . . . . . . 153

Corrected area ratio Of the solvated electron

ESR signal in 0.10M Na, K. Rb, and Cs-ammonia
solutions to that Of the center DPPH radical

peak . . . . . . . . . . . . . . . . . . . . 154

viii

LIST OF TABLES (Continued)

TABLE
11.

12.

13.
14.

15.

16.

Area ratio Of the solvated electron ESR sig-
nal in liquid ammonia for Na, Rb,
that Of K at the same concentrations and

temperatures . . . . . . . . . . . .

Temperature-dependence Of the ESR signal for
a solution Of Na-2,2,2-crypt in ethylamine

Titration data for the Na-cryptate solid

Complete elementary analysis Of Na-cryptate

solid .

Solubility test of Na-cryptate solid on some

common solvents . . . . . . . . . . .

Microwave studies of the conductivities Of

metals,

semiconductors and insulators

ix

and Cs to

Page

155

160

175

176

177

191

LIST OF FIGURES

FIGURE page
1. Dicyclohexyl-lB-crown-6 and 2,2,2-crypt . . . 3
2. Optical absorption bands of solutions of
metals in ethylenediamine . . . . . . . . . . 16
3. Tube for preparing metal samples . . . . . . 33

4. Setup for flow synthesis . . . . . . . . . . 52
5. Titration curve for 2,2,2-crypt solution . . 59
6. Computer fit Of 2,2,2-crypt titration curve . 62
7. Plot Of the concentration Of crypt and Of mono-
protonated and diprotonated crypt versus the
amount Of HCl added . . . . . . . . . . . . . 64
8. Titration curve for Na+—cryptate solution . . 65
9. Computer fit Of Na+-cryptate titration curve. 68
10. Plot Of concentration of monoprotonated and
diprotonated 2,2,2-crypt, and Na -cryptate

versus the amount Of HCl added . . . . . . . 70

11. pH-dependent NMR spectra for 2,2,2—crypt solu-
tions . . . . . . . O . . . . . . . . . . . . 72

12. Plot Of NMR shift versus pH . . . . . . . . . 74
13. Computer fit Of the NMR shift with pH . . . . 76

14. pH—dependent NMR spectra for Na+-cryptate in
ND3 o o o o o o o o o o o o o o o o o o o o o 78

15. Temperature-dependent NMR studies of 2,2,2-
crypt in ND3 o o o . o o o o o u o o o o o o 83

16. Temperature-dependent NMR studies Of K+-
cryptate in ND3 o o o o o o o o o o o o o o o 84

LIST OF FIGURES (Continued)
FIGURE Page

17. Temperature-dependent NMR studies of Na-
cryptate in ND3 o o o o o o o o o o o o o o o 85

18. Plot Of logarithm of volume susceptibility
versus (1/T) for sodium-cryptate solution in
ND: 0 o o o o o o o o o o o o o o o o o o o o 89

19. Temperature-dependent studies Of the position
of the band maximum for Na in various sol—
vents . . . . . . . . . . . . . . . O O O O O 98

20. Temperature-dependence Of the position Of the
band maximum for K" in various solvents . . . 99

21. Temperature-dependence of the position of the
band maximum for Cs‘ in methylamine . . . . . 100

22. Temperature-dependence Of the position Of the
band maximum for eSOlv in various solvents . 101
23. The relation between the Na—peak position at
25° and its temperature coefficient. The
slope Of the straight line has a magnitude of
298°K . . . . . . . . . . . . . . . . . . . . 104

24. The relation between the peak position Of I-
and Of Na' in various solvents at 25° . . . . 105

25. The relation between the peak position of Na-
and Of K' in various solvents at 25° . . . . 107

26. The spectra Of the solvated electron at 25° in
various solvents. . . . . . . . . . . . . . . 109

27. The Spectrum of the solvated electron in 1,2-
propanediamine at various temperatures . . . 110

28. The spectrum of the solvated electron in ethyl-
amine at various temperatures . . . . . . . . 111

29. The relation between the peak position Of K-

- o o O
and Of esolv in various solvents at 25 . . . 112

30. The relation between the peak position Of eSOlv
Obtained by dissolving metals and by pulse
radiolysis. The straight line has unit slope 115

31. Comparison of the band Shape Of eSOlv at 25°
in DEE and THF Obtained by dissolving metals
and by pulse radiolysis . . . . . . . . . . . 116

xi

LIST OF FIGURES (Continued)

FIGURE

32.

33.

34.

35.
36.

37.

38.

39.

40.

41.

42.

43.

44.

Page

Comparison Of the absorption spectra Of sol-
vated electron produced by pulse radiolysis and
by metal SOlUtiOnS in ND3 o o o o o o o o o o 117

The spectrum Of the solvated electron in 1,2-
propanediamine at various temperatures Ob-
tained by pulse radiolysis . . . . . . . . . 120

Pulse radiolysis studies Of pure THF, THF
with Na and THF with Na+ and 2,2,2-crypt . . 122

Dewar for ESR study .., . . . . . . . . . . . 132

Temperature-dependent ESR studies Of 0.02M
DPPH SOlution O . . . . . . . . . . . . . O O 135

Plot Of peak-to-peak width versus the square

root of the microwave power for the solvated
electron Signal in cesium solutions and for

the DPPH radical signal . . . . . . . . . . . 142

Plot Of peak-to-peak amplitude versus the square
root of the microwave power for the solvated
electron signal in cesium solutions and for the
DPPH radical Signal . . . . . . . . . . . . . 143

Plot Of peak-to-peak width Of the solvated
electron Signal in 0.1M cesium solution versus

the square Of the amplitude Of the DPPH

radical signal . . . . . . . . . . . . . . . 144

Plot Of logarithm of the area ratio Of the
solvated electron absorption in cesium solution

to that Of the DPPH radical central absorption
versus (1/T) . . . . . . . . . . . . . . . . 146

Changes in the solvated electron ESR band
shape with changing microwave power . . . . . 147

Plot Of the percent Lorentzian character for
the solvated electron signal in Na, K, Rb, and
Cs-ammonia solution versus microwave power . 148

Plot Of the logarithm Of the area ratio
against (l/T) for Na, K, and Rb solutions . . 149

Plot Of the area ratio of the solvated electron

signal of Na, Rb and Cs to that Of K versus
temperature . . . . . . . . . . . . . . . . . 150

xii

LIST OF FIGURES (Continued)

FIGURE

45.

46.

47.

48.

49.
50.
51.

52.

53.

54.

55.

Plot Of the spin concentration against con-
centration for Na and K—ammonia solutions

ESR Spectra Of the postassium-dicyclohexyl-

18-CIOWD-6 SOlid o o o o o o 0

Flame emission study Of the sodium content in

the sodium-cryptate solid . . .

Acid titration curve of the decomposed sodium-

cryptate solid . . . . . . . .

Apparatus for the reducing power study

Two electrode conductivity cell

Apparatus for the transformer current con-

ductivity study . . . . . . . .

Block diagram for the microwave
setup . . . . . . . . O . . . .

Plot Of logarithm Of resistance
for the thermistor . . . . . .

Plot Of Ohm's law study for the
cryptate solid . . . . . . . .

Plot of logarithm Of resistance
for the sodium-cryptate solid .

xiii

conductivity

versus (l/T)

sodium—

versus (l/T)

Page

157

166

170

174
179

183

185

188

190

194

195

I. INTRODUCTION

The study Of alkali metals in liquid ammonia solutions
was first reported by Weyl in 1863.1 Since that time the
topic has been Of great interest to many people and both
the physical and chemical properties Of such solutions have
been widely studied.2 However, since the Optical and ESR
spectra each Show only a single absorption band, the study
Of the complicated interactions hidden behind the apparent
simplicity Of these solutions has been severely hampered.
Unfortunately, many Of these interactions influence the
solution prOperties. In spite Of the shortage Of structural
information, a number Of models have been developed for
metal-ammonia solutions.3"9 The main Species in very dilute
solutions have been shown to be the ammoniated electron and
the alkali cation. As the concentration increases, the
interactions between the solvated electrons themselves and
between the solvated electrons and the cations become more
and more important. The nature Of these interactions is far
from clear.

In 1926, Gibson and Phipps1° discovered that the alkali
metals could also be dissolved in certain ethers to form
blue solutions. Thereafter the alkali metals were found to

form'blue solutions with a number Of aliphatic amines and

1

2

even hexamethyl phosphoric triamide (HMPA).11.12 The Opti-
ca111o13'22 and ESR?1o23:24 results for these solutions not
only show the solvated electron absorption but also indicate
the existence of other Species such as the monomer (M)13'15'
19:23:2‘ and the alkali anion (M-)}3'25'31 Although we are
able to Obtain more information about the nature Of the
species involved in these metal solutions, there are only a
limited number Of solvents which dissolve alkali metals in
high enough concentrations to study. Recently, the discovery
that the use of certain cation complexing agents (Figure 1)
can greatly enhance the metal solubility has solved many of
the prOblems.1°v31v32 In addition, by adjusting the ratio
of added complexing agent to dissolved metal, we can form
solutions enriched in either the solvated electron or the
alkali anion.

By extending these studies to various solvents, a new
type of solid has been discovered in which the metal combines
with the cation complexing agent.

The research described in this thesis is centered around
the preparation, properties and use of one Of the cation com-
plexing agents--4,7,13,16,21,24-hexaoxa-1,10-diaza-bicyclo-
(8,8,8)-hexacosane(2,2,2-crypt)(Figure 1). The area covered
by this research is very broad.1 Therefore, it is not feas-
ible to present the detailed background for all Of the topics
and only those which are closely related are included. The
thesis begins with a historical background. After a chapter

which describes the general experimental techniques, the

Figure 1. Dicyclohexyl-lB-crown-6 and 2,2,2-crypt.

4

detailed synthesis Of 2,2,2-crypt is presented. Studies of
2,2,2-crypt and its cation-cryptate, utilization of 2,2,2—
crypt in metal solutions, and the study of metal-crypt
solids are discussed separately in different chapters.
Finally the effect Of the cation on the spin-pairing pro-
cess in dilute metal-ammonia solutions is described. A
chapter with suggestions for future work is also included
at the end of the thesis.

Since much of the work was done in collaboration with
Frederick Tehan, his Ph.D. thesis should be consulted for

additional information.

II. HISTORICAL PERSPECTIVE

2.1. Metal Ammonia Solutions

Alkali and alkaline earth metals dissolve in liquid
ammonia to form blue-colored, paramagnetic, electrically
conducting solutions. Due to the high solubility of the
metal, solutions with a wide range of concentrations can
be prepared. Usually, experimental studies can be separated
into three different concentration regions -- very dilute
(lower than 10-3M), intermediate (IO-SM to 1M), and concen-
trated (above 1M) solutions. In the very dilute solution
region, most investigators agree that the solvated electron
and M+ are the most important species. However, in the
intermediate region, the interactions between solvated elec-
trons themselves and between solvated electrons and cations
become important. It is these interactions which influence
the physical and Chemical properties and complicate our
attempts to understand the nature Of the solutions. The
non-metal to metal transition occurs in the concentrated
region. From here on, we shall focus our attention only on

the very dilute and intermediate concentration regions.

6

2.1.1. General properties Of dilute metal-ammonia solutions

When alkali or alkaline earth metals dissolve in liquid
ammonia, a blue color appears immediately. The optical
spectrum Of this blue solution shows a strong absorption
band in the near infrared which tails into the visible
region and gives the blue color. The molar absorptivity
and band shape are independent Of concentration and the peak
position is only slightly concentration dependent.33‘35
The area under the absorption curve gives an oscillator
strength near unity. The recent independent pulse radiolysis
studies on pure ND3 by Fletcher36 and COdworkers and by us,37
leave no doubt that the species mainly responsible for the
absorption is the solvated electron. However, the contri-
bution from spin paired species might still be responsible
for the slight change in band position with concentration.

The ESR studies give a single very narrow absorption
(half width «0.05 G) with a g-value equal to 2.0012.38 The
absence Of any hyperfine splitting, and the narrowness Of
the line indicate that there is no strong long-lived mag-
netic interaction between the solvated electron and either
the cation or the solvent molecules. However, the marked
decrease in molar susceptibility with increasing concentra-
tion38 suggests that the interaction among solvated elec-
trons is strong and a stabilized diamagnetic state must be
formed. (We will deal with this subject in more detail

later in this chapter.)

7

A weak and short-lived cation-electron interaction does
exist as shown by Knight shift (paramagnetic shift) studies
by cation NMR techniques39‘42 and also by conductivity
data.“”"45 The Observed Knight shift Of the resonance
position Of the cation results from the interaction Of the
spin Of the solvated electron with the spin Of the metal
nucleus. The conductivity data also show the existence Of
interactions between cations and electrons. When the Con-
centration is increased, the equivalent conductance de-
creases tO a certain limit. This decrease in equivalent
conductance presumably results from the formation Of neutral
complexes or Species with lower mobility. Several species
such as M, M+-e-, M2, M- etc have been proposed.""9 Among
them, the ion-pair between the solvated electron and the
solvated cation is the simplest and is most compatible with
all Of the data.

As summarized by Dye,46 the models for electron-pairing
in metal-ammonia solutions can be classified into three
general categories: (1) metal-based species; (2) double
occupancy Of cavities; (3) electrostatic aggregates. On the
other hand, the theoretical treatments Of the solvated elec-
tron in liquid ammonia47 can be classified into four general
categories: (1) Landau's polaron model; (2) Ogg's cavity
model: (3) Alder's cluster model: (4) Jortner's semicontinum
model.

The polaron mode17t48'51 is based on the possibility Of

binding an electron in a continuous dielectric medium.

8

Landau's potential well is used for V(r) as r approaches
zero, i.e. the continuous dielectric constant is used even
for distances smaller than molecular dimensions. The calcu-
lated results from this model fit reasonably well for ex-
tremely dilute solutions, but are inconsistent with the
experimental data Obtained from other ranges Of concentration.

The cavity model postulates that the electron which
dissociates from the metal ion is trapped in a self-formed
or pre-existing hole in the solution. This model was first
proposed by Ogg in 1946.52 An oversimplified Spherical po-
tential well with infinite energy boundaries was used. The
model was later modified by Lipscomb53 and Stairs54 who
also used the model Of an electron in a spherical box.
Lipscomb53 also suggested that interactions between the
electrons and the protons Of the oriented polarized solvent
molecules should be included. Kaplan and Kittel55 developed
this idea and presented their own model in which the un-
paired electron exists entirely in delocalized molecular
orbitals on all Of the protons which define the cavity.

The cluster model was proposed by Becker, Lindquist
and Alder (BLA model).6 This model presents a picture of a
valence electron moving in the expanded orbitals of the sol-
vated positive ions. It explains the formation Of electron-
pairs by two metal ion centers weakly bound by two electrons
Of Opposite spin similar to the ground state Of the hydrogen

molecule. Gold, Jolly and Pitzer8 modified the BLA model

9

by considering that the monomers and dimers are ion-pairs
or aggregates held together by Coulomb forces.

The semicontinuum model was developed extensively by
Jortner in a series Of papers.5°‘°° In the latest paper
by Copeland, Kestner and Jortner,66 interactions Of the
electron in thetzwity with the first layer Of solvent mole-
cules (first solvation shell) through short-range attractive
and repulsive forces were explicitly included. Starting
with the second solvent layer, the polar medium was con-
sidered to be continuous. The potential used for this model
is the averaged long-range polarization energy. The total
energy of the electron in the polar medium is given by the
sum Of an electrostatic microscopic short-range attraction,
a Landau type long-range interaction, a short-range repulsive
interaction and a medium rearrangement energy.

The semicontinuum model seems to be the most satis-
factory one available. It can be used to rationalize the
experimental results although its predictive value is limited
Very recently, Fueki, Feng and Kevan61 applied this model to
trapped electrons in solids and found that the calculated
results fit the experimental results very well. The use of
this model tO explain the existence Of dielectrons in polar
liquids was also tried. However, the results Of COpeland,
and Kestner62 do not agree with the results Of Feng, Fueki

and Kevan.63

10

2.1.2. Spin-pairing phenomenon in dilute metal-ammonia
solutions

As early as 1897, Cady Obtained conductivity data for
sodium-ammonia solutions.64 He found that the solutions were
excellent conductors. Later Kraus and others published a
series Of papers all dealing with the conductance Of metal-
ammonia solutions.“3:55'68 It was shown by Kraus that the
conduction process in solutions Of alkali metals in ammonia
was ionic with metal cations as positive carriers and elec-
trons as negative Carriers. He also found the mobility Of
the negative ion was much greater than that Of the positive
ion and that the conductivity Of concentrated solutions
reached half of the conductivity Of liquid mercury. However,
the conductivity Of dilute solutions varies with concentra-
tion in a most unusual way. At first the conductivity Of
the solutions decreases with increasing concentration. It
reaches a minimum value at about 0.04M. The data from mag-
netic susceptibility studies Of dilute metal-ammonia solu-
tions come as another surprise. The static magnetic sus-
ceptibility studies performed by Freed and Sugarman69 and
the ESR paramagnetic susceptibility results Of Hutchison and
Pastor38 both showed a drastic decrease in paramagnetic sus-
ceptibility per mole with an increase in concentration. At
0.6M, the molar paramagnetic susceptibility is only about
2.5 percent of that at 10-2M. It is, therefore, Obvious
that some kind Of electron-pairing process must exist in

dilute metal-ammonia solutions. There are a number Of models

11

which purport to explain these phenomena.7° They propose

Species such as e- M, M-, M2, and e:. Among all these

solv'
species, there are many equilibrium expressions possible.
The following equations are commonly used:

+ — K1

 

 

 

 

 

M +e < > M
.. K2 _.

M +e < > M

M +M < > M2.

 

Dye calculated K1 from the conductivity data for
sodium-ammonia solutions.46 He also used the magnetic sus-
ceptibility data for potassium-ammonia solutions tO Obtain
the equilibrium constant K2. By assuming K1 equal tO K3
(the iOn-cluster model) all three equilibrium constants
were obtained. Based upon the values Of K1, K2, and K3,
the cationic transference number of metal-ammonia solution
was predicted and compared with the measured value.46 The
result for sodium in ammonia showed a remarkable difference
between the calculated values and experimental data. Also
the comparison between the measured activity coefficient and
the calculated relative activity coefficient Of sodium in
ammonia indicates a large discrepancy between two values?6
Based upon these comparisons, Dye emphasized the dilemma of
metal-ammonia models.46 Since this dilemma was based upon
the assumption that the electrochemical data Obtained with
sodium solutions could be combined with the paramagnetic
susceptibility data Obtained for potassium solutions, the

dilemma wOuld not exist if the Spin pairing process in sodium

12
solutions were considerably different from that in the
potassium solutions. Unfortunately, most Of the ESR data
of Hutchison and Paster were for potassium solutions.38
The few data for sodium-ammonia solutions give essentially
the same values as for potassium solutions at —33°. How-
ever, the data are not extensive enough to draw any definite
conclusions. Recently DeMortier, at 31.71 made ESR studies
on different concentrations Of sodium solutions. They
found that in sodium- ammonia solutions with concentrations
less than 0.01M there is practically no spin—pairing. Ex-
trapolation Of their data tO higher concentrations indicates
much less spin-pairing in sodium solutions than in potassium
solutions. From these results they concluded that the few
sodium data of Hutchison and Pastor at -33° were in error.
If DeMortier's data were correct, the dilemma Of metal-
ammonia solutions would vanish. In order to resolve this
discrepancy, careful ESR studies with Spin standards were
carried out in this laboratory for solutions Of Na, K, Rb,
and Cs in liquid ammonia in the concentration range 0.01M
to 0.1M. Details are discussed in Chapter VII.

Just as for solvated electrons, there are a number Of
models for the dielectron in dielectric media. Ogg applied
his cavity model to the dielectron.52 Again he used an over-
simplified picture Of particles in a spherical box with
infinite walls. The electrostatic repulsion term was evalu-
ated by the adaptation Of the analogous treatment for the

ground state helium atom. His calculations indicated that

13
two electrons trapped in the same cavity were appreciably
more stable than either two electrons in separate cavities
or one trapped and one conducting electron. Hill carried
out the same calculation but found a totally different re-
sult.72 The results from Hill's calculation indicated that
electron pairs were energetically unstable with respect to
single electrons. Land and O'Reilly made self-consistent
field calculations for a pair Of electrons trapped in a di-
electric medium.73 In their calculation, several approxi-
mations were made. The local electric field at an ammonia
molecule on the periphery of the cavity was only roughly
estimated, the polarization energy of the first coordination
shell ammonia molecules was neglected and the correlation
energy Of the trapped electron pair was not considered. It
was found in their calculation that the solvated dielectron
was unstable relative to two solvated electrons. Later,in
1971, O'Reilly improved these approximations and came to the
Opposite conclusion.74 Fueki, gt_§l, used a continuous di-
electric model tO develop a theory for the trapped dielec-
tron.75!7° This work is essentially an extension Of
Jortner's self-consistent field treatment for the one-elec-
tron problem tO the dielectron problem. They assumed that
two electrons were located in the same polarization center
with zero radius. The theory leads to the conclusion that
the dielectron is energetically stable provided that DB >
[Bop/(DOp-lfl, where Ds and Dop are the static and Optical

dielectric constants respectively. Kestner and Copeland77

14

applied the semi-continuum model for the solvated electron
to the dielectron system. They considered that the first
solvation layer was formed by either 12 or 18 ammonia mole-
cules. The calculation suggested a stable state for the
dielectron in liquid ammonia. Recently they tried again by
making a similar calculation for the one-electron cavity
Species. However, the result predicts that the dielectron
cavity Species is unstable relative tO two one-electron
Species. Feng, Fueki and Kevan used a semicontinuum model
for two electrons in the same cavity.61 They Claim the di-
electron is stable with rSSpect to two single trapped elec—
trons. From their calculation, they also found that the
cavity radius for the dielectron is slightly smaller than
that for the single electron. Therefore only a slight shift
in the Optical absorption for the dielectron relative tO
that Of the single solvated electron band should be eXpected.

It is clear from the above discussion that the stability
of a two-electron cavity relative tO two one—electron
cavities has not been demonstrated theoretically. However,

the calculations do not rule out its existence.
2.2 Alkali Metals in Amines and Ethers
2.2.1. General prOperties

Alkali metals dissolve in some Of the amines and ethers
to form blue solutions. The Optical spectra of the solutions
Show two absorption bands, one in the near infrared region

and one in the visible region. The peak position and shape

15
in the near infrared region is solvent dependent but almost
totally metal independent. On the other hand, the visible
absorption is only slightly affected by the solvent but is
definitely metal dependent (Figure 2). The species mainly
responsible for the near infrared absorption is well identi-
fied as the solvated electron. Recently this conclusion has
been further confirmed by pulse radiolysis studies.13‘17o23r
33'37 The excellent agreement between metal solution and
pulse radiolysis studies with this assignment leaves little
doubt about its essential correctness. The species which
gives the visible absorption was not identified until the
last few years. In 1968, Hurley, Tuttle and Golden79
demonstrated that potassium solutions could exchange with
sodium from borosilicate glass. After this demonstration,
it quickly became clear that the visible band could be
associated with the metal anion. Detailed consideration
Of this assignment will be discussed fully later.

The ESR Spectrum shows not only a very narrow single
absorption line, but also a hyperfine pattern superimposed
on the narrow line.21v33:24 The hyperfine structure comes
from the relatively strong interaction between the spin Of
the electron and the spin of the nucleus. This is the first
direct proof Of the existence Of a monomeric Species. The
narrow line is attributed to solvated electrons in the
solution.

Conductance studies have been made only with methyl-

amine°°o81 and ethylenediamine82 as solvents. In a solution

.Imz .v nix .mulmu .N “>Homw .fi .ocHEmflomcoawsuo CH
mamuwe mo mcoflusHom mo mocha coflumHOmnm HmOHumo .N ouomflm

Ania—x TE 8 soafiacwerukr

om M: 2 E E 2 m o s
_ _ _ — — _ _ _d

 

16

 

 

 

Nd

°°. ‘9
CD CD
)(flflMl‘Z/‘i

17
which shows only an infrared absorption, the conductance
(corrected for viscosity effects) is comparable to that Of
metal-ammonia solutions. However, when the metal-dependent
visible band is present, the conductivity value is much
lower than that predicted on the basis Of metal-ammonia
solutions. The decrease in conductivity is caused by the
presence Of the poorly conducting metal-dependent species.

2.2.2. Major Species involved in metal-amine and metal-
ether solutions

2.2.2.1. Solvated electrons

The existence Of solvated electrons in metal-amine and
ether solutions is based upon the following strong evidence:

(1) The similar infrared absorption found in the metal
solutions as well as in the pulse radiolysis Of pure solvents
is the strongest proof.

(2) The sharp single ESR line with the same g-value
as that of metal-ammonia solutkns also supports the exist-
ence of solvated electron.

(3) The comparable conductance data for metal-amine
solutions which show only the infrared band and metal-
ammonia solutflxs suggests that the same kind Of Species is
responsible for the electrical conductivity in these solu-

tions.

18

2.2.2.2. Spin-paired species

The existence Of spin-paired species other than M- in
solutions of metals in amines and ethers is still an Open
question. There are no direct studies of either conductivity
or magnetic susceptibility in these solutions which bear on
this question. However, some indirect evidence presented
at the end of Chapter VII seems to indicate the existence

Of such a species.
2.2.2.3. Alkali anions

The existence Of alkali anions (M-) may be questioned
by scientists who are not accustomed tO negative oxidation
states for the alkali metals. However, M- does exist in
the gas pahse as shown by Russian scientists in the 60's.82
Besides these experimental results, theoretical calculations
on electron affinities show positive values for both Li
(0.82 e.v.) and Na (0.47 e.v.).82

The species responsible for the intense visible absorp-
tion band is assigned as an alkali anion. The evidence for
this assignment is as follows:

(1) The relatively low intensity of the monomer ESR
absorption rules out the possibility Of assigning the monomer
to the large visible absorption band seen in the optical
spectrum.

(2) Solutions which contain no IR band give little
or no absorption in the ESR. Therefore the species respons—

ible for the visible absorption must be diamagnetic.

19

(3) The visible absorption band is strongly metal
dependent.

(4) The charge transfer to solvent (CTTS) comparison
Of the visible absorption spectra of metal-amine and ether
solutions with that Of the negative ion (I-) supports the
assignment.25 There are two major criteria for the CTTS
transition. First is a linear dependence Of the shift of
the absorption maximum with temperature. Second, there
exists a linear relationship between the change in absorp-
tion maximum Of two anions, both measured in the same set
Of solvents. For example, iodide ion (1-) has a marked blue
shift upon lowering the temperature. The same shift is also
Observed for the visible band. A straight line is obtained
when the position Of the absorption maximum is plotted versus
temperature. The visible absorption also satisfies the
other criteria. A quantitative study of the shift of the
visible absorption with solvent composition has been made
for metal solutions and the I- band by Matalon, Golden
and Ottolenghi.25 The peak position of the visible band
was plotted against the transition energy Of the iodide ion
measured at the maximum of the low energy I- band. A
straight line was Obtained. Therefore, the transition re-
sponsible for the visible absorption seen in the metal
solution appears to be similar to the CTTS band Of 11“"86

(5) Flash photolysis studies also suggest the exist-
ence Of M-.2°'3° Solutions Of sodium in propylenediamine

(PDA) and ethylenediamine (EDA) were photolysed by a giant

20
pulsed ruby laser.3° After the pulse, Huppert and Bar-Eli
found that the absorption Of the band at 670 nm was 70 per-
cent bleached, and at the same time an absorption at ~.1ooo
nm grew in. A plot Of 1/AA, where AA is the change in
absorbance, versus time shows a straight line for bleaching
of the 1000 nm absorption and recovery Of the 670 nm absorp-
tion. This indicates that the recovery after bleaching is
a second order reaction. The same result was also Obtained
by Gaathon and Ottolenghi.2° The results from pulse radiol-
ysis studies not only gave the same general results but also
provide more information.87 A growing-in Of the visible
absorption band was Observed when a solution which contained
only a sodium salt was pulsed. In this case the metal de-
pendence Of the visible band was once again proven and the
second order (in solvated electron) nature Of the reaction:
2e.801v + Na+ -9 Na- was shown.

(6) Recently in our laboratory, we have found that by
addition Of one Of the cation complexing agents, crown or
crypt, to metal solutions, the visible band can be totally
eliminated and at the same time the solvated electron band
grows in.18 This indicates that the species responsible
for the visible absorption is an aggregate Of the cation
and the electron.

(7) Finally, the oscillator strength calculation done
by Dye and DeBacker provides the most convincing evidence
for the stoichiometry M-.88 The oscillator strength calcu-

lated for Na- in ethylenediamine is about 1.9 i 0.2.

21

This value indicates that at least two electrons per sodium
are involved in the transition, a condition which is satis-

fied by the species Na-.
2.2.2.4. Monomers

The hyperfine structure Observed in ESR studies provides
direct evidence for the existence Of the monomer.89‘91
However, the low intensity Of the ESR Signal tells us that
the monomer probably is not a major constituent of the solu-
tion.

Very recently, pulse radiolysis studies done by Bockrath
and Dorfman16 and by Fletcher and co-workers14 Showed that

a monomer is an intermediate in the formation Of M- from

e solv and M+ and that the monomer has a detectable

Optical absorption.
2.2.3. Problems involved in the study Of metal solutions
in amines and ethers

Metal- amine and metal-ether solutions seem to give
more information about the species involved in metal solu-
tions than is the case for metal-ammonia solutions. Un-
fortunately, the number Of solvents which dissolve enough
metal to form stable blue solutions is limited. SO far only
methylamine, ethylamine, ethylenediamine, propanediamine,
THF, dimethoxye thane and certain poly-ethers have been
found to dissolve alkali metals. Besides these amines and

ethers,hexamethyl phosphoric triamide is the only other

kind Of solvent which dissolves metal. The small number of

22

solvents available and the extremely low solubility severely
hampered progress. In 1970, Dye and co-workers,32 dis-
covered that the use Of the cation complexing agent dicyclo-
hexyl-lB-crown-G, enhanced the solubility of metals in
amines and yielded solubility even in diethyl ether.

Shortly thereafter, the use of another, even more effective
complexing agent, 2,2,2-crypt was initiated. Since that
time a wide range Of new solvents has been found that dis-
solve alkali metals with the aid Of the cation complexing
agents.

2.3. Pulse Radiolysis and Flash Photolysis Studies
with Ammonia, Amines, and Ethers

2.3.1. Flash photolysis studies

Flash photolysis studies on solutions Of potassium in
THF and in dimethoxyethane (DME) were done by Eloranta and
Linschitz in 1963.92 They found only one absorption band
for the metal in THF and in DME solutions at 900 nm and 720
nm reSpectively. Upon flashing, this visible band was com-
pletely bleached out and led to a new broad infrared band.
This transformed rapidly to an intermediate band around
900 nm for both solvents. The secondary product finally
decayed slowly back to the original substance. They ex-

plained the entire transformation by the processes:

le: hv > 2e- 3 = 1 = .

 

 

 

Now With more understanding of the nature of the species

23

involved, we probably would suggest another route:

M- —h—Y-—> M+ + 2e; L 2M+-e- ---—-> M_ + M+.

The following year, Ottolenghi, Bar-Eli and Linschitz
performed flash photolysis experiments on solutions of
potassium in ethylamine.93 They Observed three absorption
bands at 660, 900, and 1300 nm and assigned them to the
monomer, dimer, and solvated electron bands respectively.

(Now we know these bands are actually Na-, K-, and e-solv

hands.) When the solution was exposed to the light flash
the 900 nm band disappeared and the 1300 nm band grew in.
This result is the same as that observed by Eloranta and
Linschitz92 for metal-ether solutions. The experiments done
later by Gaathon and Ottolenghi in solutions Of sodium in
prOpanediamine gave the same result.2° However, they first
pointed out that the primary step Observed in photolysis of

the solution could be

 

and the second step could be

 

+ _
> M + e .

M

In the same year, Huppert and Bar-Eli used a pulsed ruby
laser to photolyze solutions Of sodium in propylenediamine.3°
They found similar results and came up with the same general
explanation. Kloosterboer, st 31,29 extended the ethereal

solution work and demonstrated for the first time the

24

presence Of an intermediate absorbing transient with the
stoichiometry M. Glarum and Marshall27 followed the ESR sig-

nal Of eSOlv after flashing M- in ethereal solutions.

2.3.2. Pulse radiolysis studies
Pulse radiolysis is another way to study the behavior
Of solvated electrons in liquid media. Studies Of the
solvated electron in water and in alcohols by this technique
have been extensive. The use of fast response detectors has
overcome the problem Of the very short lifetime Of the sol-
vated electron in these solvents.94 Pulse radiolysis ex-
periments had also been done in liquid ammonia,95:96 ethyl-
enediamine°7v98 and a few other amines99 in the 60's. Most
Of the spectra Obtained were only partially mapped, usually
on the high energy side. The first complete spectrum with
ammonia was reported for a "long-lived" component (milli-
second time scale) by Compton g£_gl,95 Dye, DeBacker and
Dorfman determined the spectrum in ethylenediamine in the
microsecond time range.26 A partial spectrum Of the sol-
vated electron in ammonia was also included. Recently
Fletcher and co-workers36 and we37 (independently) have
mapped the entire Spectrum Of the solvated electron in
liquid deutero-ammonia. Jou and Dorfman reported studies
Of the spectrum Of the solvated electron in a number Of
ethers.15 Nauta and Huis studied thesolvated electron in
HMPA by the pulse radiolysis method.1°° Fletcher and co-
workers Obtained the solvated electron spectrum in ethyl-

amine.1‘

25

Pulse radiolysis studies of the'solvated electron in
amines and ethers have not been limited to spectrum mapping.
Kinetics studies were also reported by many investigators.
Dye, DeBacker, Eyre, and Dorfman87 studied the kinetics Of
formation Of Na- in ethylenediamine by the reaction Of sol-
vated electrons with sodium salts. Bockrath and Dorfman
reported the Spectrum and kinetics of formation Of the
sodium-electron pair in THF.16 Fletcher, Sedden, Javcak and
Sopchyshyn published a pulse radiolysis study of solutions
of the alkali metals in methylamine and ethylamine in which
the intermediate formation Of the electron-cation ion-pair
(or monomer) was followed.13

2.4. Cation Complexing Agents -- The Macrobicyclic
Diamine, Crypt

Many interesting macrocyclic polyethers and polyether
amines have been synthesized.101 All Of these compounds
Show remarkable cation-binding ability. It was in our
laboratory that some of these complexing agents were first
used to enhance the solubility Of metals in amines and
ethers.1°o32 Among the macrocyclic compounds, the macro-
bicyclic diamine class (which we call the crypt class) has
a stronger complexing ability than the others. Therefore,
the crypts were used more extensively in this research.

The macrobicyclic diamines were first synthesized by
Lehn and COdworkers in 1969.1°2 The general structure Of

this class Of compounds is shown below:

Where m, n, and p may have the values of 0, l or 2.

{Hwy found that the compound formed cage complexes with
metal cations and they named this new class Of inclusion
compounds “cryptate".m3 We have shortened the name Of the
bicyclic diamine polyethers, which form cryptates, to
"crypt“. In order to distinguish the various compounds Of
this class, the prefixes m, n, and p, as used by Lehn, are
applied in;front Of the name to denote the number of
-CH2CH2-O- linkages in each Of the Chains.

In the past few years, Lehn and coaworkers have pub—
lished eleven papers,1°3"111 all dealing with the crypt com-
pounds. They measured the NMR shift Of the protons on the
metal ion—cryptates and reported the equilibrium constants
for complexation of the cations. From the equilibrium
constants, a remarkable cation selectivity of the various
crypts with regard to both the size and the charge of the
cation was found. Because Of the presence Of the crypt, the
solubility Of metals and their salts increases drastically
in many organic solvents. The X-ray structures for the Lit,
Nat, Ki, Rbi, Cst, Cat+ and Bai+ cryptates Of various salts
have been determined.11‘-"119 Lehn and coaworkers also syn-

thesized a number Of different-sized crypts.1°7a1°8 With

27
the availability of different-sized cavities, another way
is provided for the formation Of stable complexes with

metal cations Of different sizes.
2.5. Protonation of Amines

Various NMR studies on proton exchange in ammonia and
amine solutions have been made. The first piece Of informa-
tion came from the work Of Gutowsky and Saika.12° They
studied solutions Of ammonium hydroxide and ammonium chlor-
ide. In 1954, Gutowsky and Fujiwara found that in ammonium
hydroxide solutions there is a significant association be-
tween ammonium ion and water in dynamic equilibrium.121
Ogg,122 in a study Of the NMR spectra Of such solutions from
the gaseous to the liquid phase, Observed a marked shift Of
ammonia protons. He attributed this shift to hydrogen bond-
ing in the liquid phase. He also pointed out that the ex-
change rate for NH4+ and for H20 could be changed by a
Change in pH value. Later, in 1956, Meiboom and his co-
workers proposed a mechanism for the proton exchange in am-

monia and amines.123‘125

 

R3NH+ + B > RaN + BH+ (1)
k .

R3N + 13H+ ———’4-+ R3NH+ + B (2)
ks

RaN +13 --—-->R2NH +B'R (3)

For the ammonium ion, R = H, B is a basic molecule and B'
indicates that one Of its hydrogens has been replaced.

Reaction (1) can futher be specialized for the case Of

28

aqueous solutions with the following:

R3NH+ + H20 ——1—> R3N + H30+ (4)

. + _ k5 .

RsNH + OH -——-¢ R3N + H20 (5)
+ ' k3 + ' .

R3NH + NR3 -—-A R3N + HNR3 (o)

R3NH+ + O-H + NR3 —El—o RéN + H-O + +HNR3 (7)
A A
Meiboom and COdworkers further extended the work from
ammonia to amines.123 Among the amines, methylamine, di-
methylamine, trimethylamine and triethylamine were studied
in detail. The rate constants, k1 to k7, for these amines
were determined.

Bottini and Roberts studied the nitrogen inversion
rate for N-ethylallenimine.126 It is understood that for
most Of the amines, the rate Of nitrogen inversion is too
fast to be measured by NMR. However, due to the steric
restrictions Of the N-ethylallenimine molecule, the nitrogen
iversion rate is slowed down considerably. Bottini and
ROberts measured the mean lifetime Of the imine between
nitrogen inversions and found it to be 0.017 sec at 110°.
They also reported a number Of nitrogen inversion rates for
other imines. Saunders and Yamada described a new method
for the use Of NMR techniques to study the rate Of nitrogen
inversion for tertiary amines.127 They attached two benzyl
groups to the nitrogen and Observed the change in the -CH2-
hydrogens. A rate constant of 2 r 1 x 105 sec.1 was Ob-

tained for the nitrogen inversion in dibenzylmethylamine.

29

2.6. Solid State Studies

Liquid ammonia dissolves certain metals to form gold
metallic solids. These metals are lithium, calcium, stron-
tium, barium, europium, and ytterbium. Among them only
the lithium compound has been extensively studied.128 The
empirical formula for the lithium—ammonia compound is
Li(NH3)4. An x-ray diffraction study at 77°K showed no Li
or NH3 pattern, but rather a new set Of lines.129 Between
82° and 890K, the solid is in a cubic phase with unit cell
length of 9.55 A. At temperatures below 82°K, the hexagonal
phase with a=b=7.0 A and c=11.1 A is the principal
phase.129 The conductivity 2°:130-132 Of the solid is very
high and the Hall effect shows a carrier concentration equi—
valent to one lithium valence electron per lithium atom.133
Magneto-resistivity studies for fields up to 43 killogauss
at 4.20K were done by Thompson and co-workers.2a.13°"132
Extensive ESR studies were performed by Levy134 and by
SienkO and co-workers.136 Magnetic susceptibility was also
measured for the lithium compound in the temperature range
of 1.5-194°K.136

For the alkaline earth metal—ammonia compounds, the
empirical formula is belived to be M(NH3)6. Crystallographic
studies reveal that the crystal structure for these compounds
is body-centered cubic.137 The melting point of these com-
pounds, as suggested by resistivity measurement513° and by

phase diagrams, is close to room temperature.138

30
The lanthanide metal-ammonia compounds are also hex—
ammines.139 X-ray studies indicate a body-centered cubic
structure for both europium and the ytterbium compounds.14°
The unit cell lengths were found to be 9.55 A and 9.30 A
for Eu(NH3)6, Yb(NH3)6 respectively. Magnetic suscepti-
bility data for Eu(NH3)6, Yb(NH3)6, and Sr(NH3)6 were ob-

tained by Sienko and co-workers.140

III . EXPERIMENTAL

3.1. General Techniques
3.1.1. High vacuum techniques and glassware cleaning

High vacuum was achieved with an Oil diffusion pump
(Vactronic HVP-100) and a mechanical pump (Cenco Hyvac-7).
Dow Corning 704 diffusion pump fluid was used in these
pumps. The pressure was measured with a Veeco RG-75P ioni—
zation gauge and a RGLL 6 ionization gauge control. Teflon
needle-valve stopcocks (Fisher Porter CO. and Kontes) were
used exclusively both on the vacuum lines and on the sample
preparation tubes. All connections were made with 5 mm,

9 mm, or 15 mm Solv-Seal joints (Fisher Porter CO.).

Glassware was cleaned first with an HF cleaner”1
followed by a thorough rinsing with distilled water, then

with boiling aqua regia. Then it was allowed to remain in

 

the tube for at least 10 hours. After rinsing 6 times with
distilled water and six times with conductance water, the

tube was dried in an oven (110°) overnight.
3.1.2. Metal purification

Small pieces Of sodium and potassium were cut from the
center Of a large piece and washed with hexane. The freshly

31

32

cut metal was placed in the top portion Of a 5 cm diameter
piece of tubing (Figure 3). The tubing was then sealed and
pumped under high vacuum. The metal was melted and the
molten metal was run down along the tube through 2 to 3
constrictions. Small-sized pieces of tubing with one end
sealed off were placed in the lower portion of the large
tubing with the sealed ends pointing up. When the metal
reached the bottom Of the large tube, one atmosphere Of
helium pressure was introduced to force the molten metal up
into the small tubing. A fairly pure metal sample can be
Obtained in this way. Further purification was done by
high vacuum distillation through a sidearm on the sample
tube.

For cesium and rubidium, small samples were prepared by
using the method of Dewald and DeBacker.141:142 Further
purification Of rubidium and cesium was done as follows:

A small amount Of cesium was first sealed in a 4 mm
O.D. piece Of tubing. The glass tubing containing the
cesium sample was scratched and placed in the sidearm Of
tube Of a sample tube. A piece Of shrinkage tubing (POpe
Scientific CO.) was used to make a high vacuum connection
between the sidearm tubing and a glass rOd (or a piece Of
tubing with one end sealed). The sample tube was pumped to
high vacuum and the cesium sample was broken in the shrink-
age tube. The broken cesium sample was then dropped down
along the sidearm and passed through a constriction. The

sidearm was then sealed off at the constriction under high

33

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 3. Tube for preparing metal samples.

34

vacuum. Further distillation in the sidearm under high
vacuum serves the purpose Of purification. For rubidium,

a small sample Of metal was first cooled in liquid nitrogen
and then cut Open in the air. The opened rubidium sample
was quickly dropped into the sidearm tubing. The sidearm
was sealed Off immediately. Further purification was car-

ried out by high vacuum distillation.
3.2. Solvent Purification and Solution Preparation

All solvents were first dried with either barium oxide
(BaO) or calcium hydride (CaHz) for at least 24 hours and
then degassed a few times. After initial drying, ethylene-
diamine (EDA), ethylamine (EA), 1,2-propanediamine (1,2-PDA)
were distilled onto Na-K (1:3) alloy which yielded blue
solutions. TO avoid decomposition, after remaining in
contact with the metal for two or three days, these solvents
were distilled from the Na-K alloy into storage bottles.
The solvents tetrahydrofuran (THF), diethylether (DEE),
diethylamine (DEA), diisopropylether (DIPE), di-gfpropyl-
ether (DNPE), di-prropylamine (DNPA), and triethylamine
(TEA) were distilled onto a mixture Of benzophenone and an
excess of Na-K (1:3) alloy. The resdlting blue to purple
color of the ketyl radical served as an indicator of dryness.
If the color faded, the solvent was redistilled onto a fresh
drying mixture.

Ammonia, deuterated ammonia and methylamine were dis—

tilled intO a heavy walled glass bottle through a "T" joint

35

under high vacuum. The ammonia or methylamine tank was
cooled tO about -30° to avoid high pressure problems. The
Na-K (1:3) alloy was distilled through a sidearm into the
heavy walled bottle prior to the distillation of ammonia or
methylamine. After the metal solution had been brought to
~'0° and allowed to stand for a half day, the blue solution
was frozen and stored at liquid nitrogen temperature until
used.

PrOpylene carbonate (PC) was vacuum distilled through
a Teflon spinning-band column («I100 theoretical plates).
‘The middle portion was collected and pumped to a pressure of

10“5

torr while frozen. The solvent was then poured onto a
potassium mirror under vacuum.

All solutions were prepared just before the experimental
work was done. Clean sample tubes with the desired amount
Of metal and cation complexing agent, crown or crypt, were
pumped under high vacuum overnight. Then the metal was dis-
tilled at manifold pressures of 10.5 torr or less. Purified
solvent was distilled onto the freshly prepared metal and
warmed up tO room temperature to further purify the solvent.
The cation complexing agent was then mixed with the metal
solution in the sidearm. The solution was then kept cold
in an isoprOpanOl-dry ice bath. Just prior to measurement,
the blue solution was poured into the optical (or ESR) cell
and the cell was put into the cooled compartment Of the

instrument. The tube was kept cold as long as possible to

minimize problems of decomposition.

36

In the case of propylene carbonate, the purified sol-
vent was poured through an evacuated "T" joint onto the
metal. If solvent had been kept for months, it had to be
degassed.

Dicyclohexyl-lS—crown-6, purchased from E. I. Dupont
de Nemours and Co., Inc., was further purified by passing
it through an alumina column. In the initial work, 2,2,2-
crypt was introduced as Obtained without further purifica-
tion.143 Later work was done with 2,2,2-crypt which had

been synthesized and purified in this laboratory.

IV. SYNTHESIS OF 4.7,13,16,21,24-HEXAOXA-1,10-

DIAZABICYCLO (8 , 8 , 8 )HEXACOSANE (2 , 2 , 2 -CRYPT)

4.1. Introduction

Macrobicyclicdiamine compounds were first synthesized
by J. M. Lehn and his co-workers in 1969.”3 The compounds
have the general structural formula shown below. Since the
initial work, a variety of different bicyclic and tricyclic

compounds have been synthesized.1°3o1°‘

can4 -O-(C2n4 —o )m-C2H4

/

N

 

 

C3H4-O-(C3H4-O)n-C3H4

C3H4-O-(C3H4-O)p-C2H4

Where m, n, and p may have the values 0, 1, or 2.

The principal reason for the great interest in these
compounds is their ability to form complexes with metal
cations,"°3’111 and to form a new class of inclusion com-
pounds called cryptates. The remarkable cation selectivity
Of the macrobicyclic diamines was studied by Lehn and
Sauvage.1°1"'-‘111 It was found that the complexation constants

depend on the size Of the cation as well as the size Of the

macro4bicyclic diamine compound. The value Of log Ks' (Ks

stability constant) for 2,2,2-crypt in water changes from

37

38

4 . + + 2
< 2.0 for Li , Cs , and Mg + to ~.9,5o for Ba2+.106 Proton

magnetic resonance studies Of process

11+ <—

[L,M]2; + m[H20]

Show an upfield shift Of both triplets (—CH2N and -OCH2) and
Of the singlet (-O-CH2CH2-O--).1°3 The amount of the shift
is dependent on the metal cation inside the cage. A sodium
NMR study for the same process in ethylenediamine has also
been performed.144 The temperature dependence Of the ex-
change time, the first order rate constant and an activation
energy were reported for the sodium cryptate dissociation in
the above solvent. Detailed X-ray structural studies“?-119
Of Li, Na, K, Rb, and Cs cryptates reveal three different
kinds Of crystal lattices, tetragonal, hexagonal and mono-
clinic. Potassium rubidium and cesium cryptate crystals all
have the monoclinic form.

It was first noted in this laboratory32 that a similar
complexing compound, dicyclohexyl-lB-crown-G, increases the
solubility Of alkali metals in amines and ethers. It was
then found that the macrobicyclic diamine.2,2,2-crypt.readily
formed blue solutions Of potassium in diethylether.31 This
has led to a new method for the study of metal solutions.
The desire for large amounts Of the macrobicyclic diamine
and the absence Of a commercial source at the time led us to
synthesize this compound. The complex organic synthesis was
performed in two different ways. First, with certain modi-

fications, we followed the usual high dilution method of

39

Lehn.1°3 Second, a flow mixing technique was developed.145
Both methods give about the same yield and quality Of pro-
duct.

Detailed procedures for both methods as well as the
NMR spectra, melting points and mass Spectra for intermediate
compounds and the final product are presented in the follow-

ing sections.
4.2. High Dilution Method102

The high dilution method described here is mainly
based on the method Of J. M. Lehn.“3 However, some im-
portant modifications were made, either tO make the synthesis
easier or to improve the yield Of an intermediate. All the
NMR Spectra, mass spectra, melting points and IR Spectra
were taken in this laboratory. A schematic description Of
the procedure is shown at the beginning Of each step. The
description Of the synthesis is given in more detail than
may be commonly done because Of the many frustrations en-

countered when we tried to follow "abbreviated" directions.
4.2.1. Synthesis Of starting materials

A. Preparation Of (1,8-DiaminO-3,6-dioxaoctane) (hereafter

called "diamine").

40

 

CO~\\ _ +
ClCHZCH20CH3CH20CH2CH2Cl + N K
COI”
1,2+Bis(2-chloroethoxy) Potassium phthalimide
ethane
O O
II II
C C
DME > \NCHZCHZOCHZCHZOCHZCHZN/
C”/’ \\‘C
O O

Triethylene glycol diphthalimide

CO \
NHzNHz NH
—--—-—> HzNCHz CH2 OCHQ CH2 OCHZ CH2 NHz + I
EtOH ”,NH
O

C

1,82Diamino-3,6-dioxaoctane Phthalhydrazide

COC1
- + + _
39.1...» C1 N H3CH2CH20CH3CH20CH2CH2N H3Cl +
COC1

OONa

NaOH
extract> NH? CH2 CH2 OCHz CH2 OCH; CH3 NH2 + a

COONa

1,8-DiaminO-3,6-dioxaoctane (1)

One mole Of potassium phthalimidel“6 (185 grams), 1/2
mole of 1,2-(bis—(2-chloroethoxy))ethane147 (93.5 grams) and
1 liter Of DMF are mixed in a 3-liter flask equipped with a
mechanical stirrer.1“8 The DMF mixture is heated overnight
with an Oil bath at 95-100°. The solution is then cooled to
room temperature and poured into 2 liters Of ice water while

it is stirred. A white precipitate forms. The whole

41

solution is filtered and the white precipitate is recrystal-
lized from glacial acetic acid. The recrystallized product
is then first washed with 5%’Na2C03 solution and then with
distilled water until no strong smell Of acetic acid re-
mains. (The yield at this stage should be ~.2o4 grams.)
The white solid so obtained is suSpended in'~ 2050 ml Of
95% ethanol. The ethanol solution is heated tO boiling
while it is mechanically stirred, under reflux. Just after
it is boiled, 102 ml of 85% hydrazine hydrate is added and
refluxing is continued for 2 hours. (After the addition Of
hydrazine hydrate, a clear yellow SOlUtion is first formed
and then in about 10 minutes the whole solution solidifies
at once. At this point the stirrer is stOpped.) Two
hundred and twenty-five ml of 10N HCl is slowly added and
the solution is refluxed for another 1/2 hour. (Note: The
HCl solution should be added in small portions with care.)
Most Of the solvent (~A80%) should be distilled Off.

Any precipitate Should be filtered and solid NaOH is
added to the filtrate tO make the solution strongly basic.
This basic solution is extracted with diethyl ether in a
continuous liquid-liquid extractor for 1-2 days. The di-
ethyl ether is then stripped Off in a rotatory evaporator,
and the condensed solution, which contains a large amount
Of diamine, then is vacuum distilled twice. The final yield
is ~v50 grams. (The reaction scale can be proportionally

increased if desired.)

42

B. Preparation Of Triglycolyl Chloride (hereafter called

diacid chloride).

 

‘ NH4v03
Hocnzcnzocnzcnzocnzcnzon + HNO3 > HOCOCHZOCHZCHZOCHZOCOH
triethylene glycol J triglycolic acid (2)
SOClz

——-——> C1COCH20CH2CH20CH20CC1

triglycolyl chloride (3)

Into a 5-liter flask which contains 3150 grams Of 60%
nitric acid is added 5 grams Of triethylene glycol and 3
grams Of ammonium metavanadate (NH4V03). The solution is
heated to 68-73° and stirred with a mechanical stirrer. As
soon as the brown fumes form, 745 grams Of triethylene glycol
is dropped into the flask by means Of a dropping funnel over
a period of “.4 hours. The temperature should be maintained
at 68—73°. After the addition has been completed, the solu-
tion is stirred for another hour. Then 80% of the nitric
acid is distilled away under vacuum. A green syrup is Ob-
tained after distilling away most Of the nitric acid.

Further evaporation is carried out in an evaporating dish.
The color Of the solution changes from green to brown to

dark brown to purple and finally to sky-blue at about 140°.
After it is cooled it becomes a hard skyéblue colored solid.
An ether Soxhlet extraction is done on the solid. A white
solid, triglycolic acid, 2, is recovered from the ether solu-
tion. (Because the diacid chloride is unstable, large

amounts of triglycolic acid can be prepared for long storage.)

43

Sixty grams Of triglycolic acid, 2” and 180 ml Of SOClz
(redistilled from commercially available SOClz) are dissolved
in 200 ml Of diethyl ether in a 1-liter flask. The solution
is refluxed for 4 hours. Then the diethyl ether is evapor—
ated Off in a rotatory evaporator. The yellow residue is
washed twice with diethyl ether and then recrystallized in
an etherepetroleum ether mixture at ~a-50°.

Recrystallization is used instead Of vacuum distilla-
tion, because triglycolyl chloride, Q} decomposes at high
temperature. By vacuum distillation, a colorless liquid
is collected at ~.900 and 2 x 10.1 torr. The NMR spectrum
of this liquid shows two singlets at 3.88 ppm and 5.56 ppm.
The spectrum is similar to, but not the same as the NMR
spectrum of the liquid prior to distillation. Decomposition
Of the triglycolyl chloride must occur before reaching the
distillation temperature of 90°. The recrystallized pro-
duct is pumped tO dryness at 15°. The white solid product
can be stored in the freezer. (The product can also be
stored in the freezer before stripping Off the diethyl ether.)
Properties Of 3;

4M.P. - 20425° NMR ("product Obtained by
.distillation):
NMR: 3.76 ppm (singlet)
. 3.88 ppm (singlet)
4.48 ppm (singlet) .
- 5.56 ppm (singlet)

44

4.2.2. Preparation of 5,12-dioxo-1,7,10,16-tetraoxa-4,13-
diazacyclooctadecane (1st cyclization)

HzNCH3 CH2 OCHg CH2 OCHz Cir-ms2 + C1COCH20CH2CH20CH20CC1

(,1) (g)

Benzene > N I,’CH2CH20CH20H20CH2CH2~‘~
High dilution

 

H NH
\C-CHzoCH2 CH20CH2 -C/
u n

O O

5,12-Dioxo-1,7,10,16-tetraoxa-4,13-diazacyclooctadecane (4)

A solution Of 29.6 grams of diamine (I) in 1000 ml
dry benzene and a solution Of 22 grams Of the diacid chlor-
ide (3) in 1000 ml dry benzene are dropped into 1 liter Of dry
benzene in a 5-liter flask. The solution is stirred vigor-
ously with a mechanical stirrer and the system is purged
with a dry nitrogen stream. The entire addition is completed
in 12 hours.149 After the addition has been completed, the
benzene solution is filtered and evaporated to dryness. A
white crystalline solid, 4, is Obtained with a yield Of 75-
80%. This white solid is further purified by passing it
through an alumina column with benzene as an eluant. In
preparation for the cyclization, the solvent -- benzene, is
first dried by refluxing over calcium hydride, CaHz, for a
day and then poured on top of freshly cut sodium metal and
refluxed for another day. Finally the benzene is distilled
in a dry N2 atmosphere. Twice the amount Of the diamine, 1,
is used for cyclization instead of using triethylamine as

an HCl scavenger. It was found that double the amount Of

45

diamine resulted in a purer product. The excess diamine
used is recovered at the end of cyclization in the following
way:

The solid from the cyclization (amine-HCl-salt) is dis-
solved in 200 ml of concentrated NaOH solution. This strong-
ly basic solution is poured into a continuous liquid extrac-
tor and extracted with diethyl ether for 2 days. Finally
the ether is evaporated Off and the diamine, l, is recovered.
The recovered yield is 85-90%.

Properties Of g;
M.P. a 110°
NMR (CDC13): -CO-CH2-O 4.00 ppm (singlet)
-CH2-N and -CH2-O 3.6 ppm (multiplet)
Mass spectrum: 290 parent peak.

4.2.3. Preparation Of 1,7,10,16-tetraoxa-4,13-diazacyclo-
octadecane (lst reduction)

,,CH3CH,OCH,CH,OCH2CH2\

l/NH + AlLiH4
“~C-CH,0CH2CH20CH2-C
I. u

o o
2.
[CH3CH20CH3CH20CH2CH3\
HN\ I’NH
CH3CH20CH3CH30CH3CH3

THF
-—-->

1,7,10,16-tetraoxa-4,13-
diazacyclooctadecane (5)

Twenty-nine grams Of purified 1st cyclization product,
4, is added, in small portions with care, into a mixture of
600 ml Of dry THF,15° and 7.6 grams Of LiAlH4. ¢JA1H4 is

added into the THF in small portions.) During the addition,

46

the reaction mixture is purged with a slow stream Of nitro-
gen. When the addition has been completed, a drying tube

is placed at the top Of the reflux condenser and the solu-
tion is refluxed for 24 hours. The solution is then cooled
to 5° in an ice bath and 7.6 ml Of H20 is added with caution,
followed by 7.6 ml Of 15% NaOH solution and finally 25 ml of
H30. The solution is stirred for an additional hour at room
temperature. The white precipitate is filtered and washed

3 times with THF and 3 times with diethyl ether. The com-
bined filtrate is evaporated tO dryness. A white solid, é/
is formed and purified by passing it through an alumina

column with benzene as an eluant. The yield is ~o70%.

Properties of g:
M.P. = 115°
NMR: -CH2-O 3.60 ppm (singlet & triplet)

-CH2-N 2.80 ppm (triplet)

-NéH 2.25 ppm (singlet) (with a wet sample,
this singlet may shift to as
high as 3.00 ppm)

IR: No absorption band Of >C=O can be detected.
Mass spectrum: N0 262 parent peak was found.
Elemental analysis:“’1

Calculated values: C, 54.96: H, 9.92; N, 10.69%

Analysis results: C, 54.88; H, 9.88: N, 10.48%

47

4.2.4. Preparation Of 2,9-dioxo-4,7,13,16,21,24-hexaoxa-
1,10-diazabicyclo(8,8,8)hexacosane (2nd cyclization)

,CHgCH30CH3CH20CH2CH
‘:NH + c1COCH20CH2CH20CH20CC1
CH3CH20CH3CH20CH2CH2

(2) ($2,)

CHacnzocnzcnzocnzcn2
N -' CH2C1120CH2CH20CH2CI'12 ""
C-CHZOCH2CH20CH3-C’ll’
II II
o o

Benzene

High dilution N

2,9-dioxo-4,7,13,16,21,24-hexaoxa-1,10-
diazabicyclo(8,8,8)hexacosane (Q)
A solution Of 26.2 grams Of the first reduction product,
5, in 500 ml Of dry benzene and a solution Of 11.0 grams Of
diacid chloride, 3, in 500 ml Of dry benzene are added into
1 liter Of dry benzene in a 5-liter flask. The solution is
stirred vigorously with a mechanical stirrer and is flushed
with a dry nitrogen stream. The entire dropping time is 8
hours. At the end Of the addition, the solution is filtered
and the filtrate is evaporated to dryness. A white (slightly
yellowish) product, 6, is Obtained, with a yield Of 80-85%.
The product is purified by passing through-an alumina column
with benzene as an eluant.
Properties of Q;
M.P. = 114°
NMR (CDCls): several peaks between 3.3 and 3.6 ppm

Mass spectrum: a parent peak at 404.

48

4.2.5. Pre aration Of 4,7,13,16,21,24-hexaoxa-1,10-diazabi-
cyc O(8,8,)hexacosane("2,2,2—crypt") (2nd reduction)

//,CH2CHZOCHZCHZOCHZCH2\\\

N -CH2CH20CH2CH20CH2CH2 -N + BH3 in THF

C-CHZOCHZCH20CH2-C

O O

(6)

~

.. + +_
> H3BN(CH2CH20CH2CH20CH2CH2)3-N B H3

Diborane Of "2,2,2-crypt"

-+ +-
.JEE£___> HClN (CHZCHZOCH2CH20CH2CH2),NClH

“2,2,2-crypt" HCl salt

 

CH2CH20CH2CH20CH2CH2
ion exchange
Dowex 1-X8 > N CH2CH20CH2CH20CH2CH2 -N
CHZCHZOCHZCH20CH2CH2

4,7,13,16,21,24-hexaoxa-l,10-diazabi-
cyclic(8,8,8)hexacosane (“2,2,2-crypt")

(Z)

Ten grams Of the second cyclization product, 6, is dis-
solved in 200 ml Of THF. One hundred fifty ml Of 1M solu-
tion Of borane in THF152 is added to the flask slowly at 0°.
After the addition has been completed, the solution is
stirred for a half hour at this temperature and then refluxed
for an additional hour. (A white precipitate forms during
the process.) The solution is cooled to room temperature
and excess reagent is decomposed by adding 50 ml Of H20.

(The solution becomes clear after the addition of H20.)

49

Solvents are evaporated. Approximately 10 grams Of the
product (diborane adduct) are formed. The diborane adduct
is changed to "2,2,2-crypt"-HCl salt by adding 200 ml Of 6N
hydrochloric acid to the compound and refluxing for an hour.-
The solution is then evaporated to dryness. The white crys—
talline solid is dissolved in 100 ml Of conductance water
and the solution is passed through an anion exchange column
(Dowex 1-x8, 20-80 mesh). The column is washed continuously
with conductance water until the solution coming out Of the
column is neutral. The water is evaporated and further dry-
ing is done by using the azeotrOpic mixture evaporation
technique. In this case, dry benzene is added to the wet
solid and evaporated. The white solid "2,2,2-crypt" so Ob-
tained is pumped for a day before any further purification.
The yield is ~v90%
Properties Of 2,2,2—crypt:

M.P. = 68°

NMR: -CH2-N: 2.65 ppm (triplet)

-O-CH2CH2-O-: 3.78 ppm (singlet)
-CH2-O-; 3.65 ppm (triplet)
Final elemental analysis:151 .
Calculated values: C, 57.45; H, 9.57; N, 7.45%

Analysis results: C, 57.31; H, 9.52; N, 7.65%.

4.2.6. Purification of final product

Purification Of the final product, Z, was carried out

in three steps. The first step was gfhexane extraction.

50

This was followed by vacuum sublimation and finalbrby the
zone melting method.

The crude material was dissolved in a small amount of
conductance water. Then the aqueous solution was extracted
with pure gfhexane. The extraction was continued with
fresh gfhexane until no further product could be extracted.
A white crystalline solid can be collected from the combined
portions by stripping-Off the pfhexane. This extraction
separated the final product from the uncyclized second re—
duction product, 5, The white solid was further dried by
pumping under high vacuum for at least a day, then transfer—
ring into a sublimation tube and heating tO «11000 with an
Oil bath. The sublimed product was collected on a cold
finger. Further purification was carried out with a zone
refiner. The solid was packed into a vacuum-sealed quartz
tube. The temperature Of the zone refiner was set atA~ 10°
higher than the melting point Of the compound. Over one
hundred passes were completed before the final pure com-
pound was collected from the middle portion Of the tube.

The compOund was stored in the dark under vacuum.
4.3. Flow Synthesis

In the synthesis of the polyoxa-macrobicyclic diamine,
1, the procedure for synthesizing both Of the intermediates,
5,12-dioxo-1,7,10,16-tetraoxa-4,13-diazacyclooctadecane, 4,
and 2,9-dioxo-3,7,13,16,21,24-hexaoxa-1,10-diazabicyclo(8,8,8)—

hexacosane, 6, requires a slow addition with vigorous

51
stirring (over a period of about eight hours) of dilute
(~r0.1M) solutions of the two agents, the diamine, l, and
the diacid chloride, 3, into large amounts of benzene in a
reaction flask under an nitrogen atmosphere. This not only
requires a long period Of addition during the cyclization
but also gives a large amount Of solution which must be
evaporated after the cyclization has been completed. On a
few occasions, a faster dropping rate was tried to test the
effect Of the rate on the yield and the quality Of the
product. We found that neither the yield nor the quality
were greatly reduced by speeding up the addition process,
provided the stirring was sufficiently vigorous. This, and
the fact that good yields could be Obtained even when an
additional equivalent Of the diamine was present, suggested
that a ring closure reaction probably is much faster than
the various chain reactions which might compete. Therefore,
efficient stoichiometric mixing seems tO be the most impor—
tant factor. This supposition was tested by using a flow
cell to carry out these steps.

The entire flow synthesis set up can be divided into
three major portions -- two storage vessels for the reagents,
a mixing chamber, and a receiving bottle (Figure 4). The
storage bottles were made of heavydwalleleyrex tubing in
order tO withstand the high pressure. The mixing chamber
is Of the type used in our laboratory for stopped-flow
kinetics studies153 (Figure 4). It has four tangential

1.0 mm inlets “drilled" into a 2.0 mm I.D. Pyrex capillary.

 

 

 

52

 

 

 

 

 

jall‘l

 

 

 

 

 

Figure 4. Setup for flow synthesis.

53

An airbrasive unit (S. S. White CO.) which uses helium to
drive Alundum through a nozzle at supersonic velocities was
used to make the inlet holes. (Presumably, conventional
mixing chambers made Of Teflon Or metal could be used.) The
receiving bottle is an ordinary l-liter round-bottom flask.
The reactant solutions were driven through the mixing
chamber at flow velocities high enough tO insure turbulent
mixing by applying about three atmospheres of nitrogen pres-
sure Over the stock solutions. The heavy-walled vessels
which contained the stock solutions were connected to the
mixing chamber with 5 mm solv~seal joints (Fischer-Porter
CO.) through Teflon needle-valve stopcocks (Kontes). The
entire apparatus was surrounded by a metal safety shield.

In a typical flow reaction, 200 ml Of 0.06M solution
Of the diamine, l, and 200 ml Of a 0.03M solution Of diacid
chloride, 3, in dried benzene were allowed to flow through
the flow cell under 3 atmospheres pressure. Blockage Of
the flow tube by the precipitation of the diamine dihydrogen
chloride salt can occur at higher reagent concentrations.

A high polymer coating on the wall Of the mixing cell, or
a larger inner diameter capillary tube might help to solve
this blockage problem.

In the high dilution method, the byproduct Of the
cyclization, hydrogen chloride, is removed with the diamine.
The same procedure is used in the flow technique. However,
the required excess Of amine may either be present in the

amine stock solution or it may be in the receiver flask.

54

The diamine can be recovered from the diamine dihydrogen
chloride salt with very little net loss Of material.

The cyclization product, the dilactam, g, was collected
by evaporating the solvent in a rotatory evaporator and
purified by elution through an alumina column (80-100 mesh)
with benzene. The yield after purification is ~v70%.
Properties Of 2 prepared by flow synthesis:

M.P. = 108-9°

NMR: 3.90 ppm (singlet)

3.50 ppm (multiplet)

Similar results were Obtained in the synthesis Of the
second cyclization product, 3, The yield and quality for
both cyclizations were found to be comparable with those
Of the high dilution method.

Consideration Of the kinetics tO be expected in cycli-
zation reactions suggests the requirements which must be met
if the flow method is to be a useful replacement for high-
dilution techniques. 1) Mixing must be rapid and complete
SO that the prOper stoichiometry is maintained. 2) The
initial step in the reaction must be fast enough to be sUb-
stantially cOmplete during the time Of flow. 3) The cycli-
zation step must be fast enough to compete with intermolecu—
lar reactions. It is likely that the type of reaction con—

sidered here meets these criteria.

V. NMR AND TITRATION STUDIES ON 2.2,2-CRYPT AND

SODIUM-CRYPTATE

5.1. Introduction

NMR techniques have been very powerful tools not only
for structure analysis but also for the study of (1) pro-
tonation Of amines, (2) chemical exchange, (3) nitrogen
inversion, (4) electron transfer reactions, (5) hindered
rotation about a chemical bond, (6) hydration Of carbonyl
groups and (7) paramagnetic shifts and magnetic suscepti-
bilities. The procedures behind these studies are well
developed. With help from the computer, analysis Of the
data can be readily accomplished.

The experiments reported here were done in order to
study: (1) the protonation of a diamine, (2) the chemical
exchange reaction between a cation and its complexing agent,
and (3) the paramagnetic shift in metal-deuterated ammonia
solutions. We have chosen 2,2,2-crypt as the basis for these
studies, because it has several unique characteristics.
First, it is a new compound with great potential for further
study. Second, it has two protonation sites and therefore
provides an Opportunity to study both inter and intra-molec-
ular exchange. Third, it is a bicyclic diamine which could

have a slow nitrogen inversion rate.

55

56

Because Of time limitations, some Of the studies de-
scribed here are still in the very preliminary stages.
However, the author believes that with this study as an
introduction, further detailed studies can be performed
without much difficulty. Suggested possible work is men-

tioned in the last chapter of this thesis.
5.2. Experimental

All of the pH titrations were carried out with either
a Sargent model LS pH meter or a Corning model 7 pH meter.

A Corning combination pH electrode was used for all the
studies. A Varian A56/60 NMR was used for the NMR work.

The complexing agent, 2,2,2-crypt, was zone purified as
described elsewhere in this thesis. Commercially available
sodium chloride was introduced without further purification.
Deuterium chloride (35% DCl in D20; min 99% D) and deuterium
oxide (99.7% D30) were purchased from Sigma Chemical CO.
Deuterated ammonia (prepared in this laboratory byDr. Gale
Smith) was purified in the same manner as ammonia.

The amounts Of 2,2,2-crypt and sodium chloride needed
were weighed on an analytical balance. For the sodium
chloride cryptate solutions, a slight excess Of sodium
chloride was introduced over that required to form a one to
one ratio Of sodium ion to 2,2,2-crypt. Besides using
deuterated solvents, conductance water was used as a solvent
for the titration studies. Hydrochloric acid was standard-

ized with a sodium hydroxide solution which in turn was

57

standardized with a known amount Of potassium hydrogen
phthalate. The sodium 2,2,2-crypt-ammonia solutions were
prepared by dissolving the gold colored sodium-2,2,2—crypt
solid into purified liquid ammonia.

Tetramethylsilane (TMS) was used as an NMR standard.
In the case Of metal-ammonia solutions, a dual sample tube
was constructed from a normal NMR tube (5 mm O.D.) and a
thin walled quartz tube (3 mm O.D.). Teflon Spacers were
used to properly center the inner tube. The metal-ammonia
solution was sealed in the quartz tube and TMS in deuterated
cthrOform was placed in the outside tube to act as a
standard.

The temperature was determined from the splitting Of
methanol peaks and confirmed with a thermocouple. Vari-
able temperatures were achieved by the use of a Varian V-4540

temperature control unit.
5.3. Proton Exchange Studies

5.3.1. Titration studies

5.3.1.1. Titration Of 2,2,2-crypt with HCl

The molecule, 2,2,2-crypt, has two nitrogen sites avail-
able for protonation. Two equations to represent the equi-

librium can be written as follows:

+ K1 +

CH < > C+H

 

 

K
++ 2
CH2 —""""'>':

58

The equilibrium constants, K1 and K2, were determined by

Lehn and coeworkers to be 2.5 x 10.10 and 5.2 x 10.8 re-
spectively (molarity basis).154 The difference between

these two constants is just large enough to distinguish two
breaks in the titration curve as shown in Figure 5. In

order to compare the experimental results with the calculated

values, calculations were performed as follows:

From the equilibrium conditions, we have

 

 

e (C) H11 1
K1 '(CH+) ( )
and
- (cH*L (H*) AA
K .. . - (2)
2 (CH2++) 72

Where (C) is the concentration Of 2,2,2-crypt, (H+) is the
concentration Of hydrogen ions, (CH+) is the concentration
Of the monoprotonated 2,2,2-crypt, (CH2++) is the concentra—
tion of the diprotonated 2,2,2-crypt, 71 is the activity
coefficient Of the monovalent ion (assuming y+ = y_), and

72 is the activity coefficient Of the divalent ion. From
the material balance conditions we have

1000 ml

(c) + (cn+> + <cn.*+> = '—'§;—-' (3)

where n1 represents the total number Of moles Of 2,2,2-

crypt added, and V is the total volume in milliliters.

t

From the charge balance conditions, we have

(H+) + (CH+) + 2(CH2++) = (C1') + (OH-) (4)

59

 

 

.cOHuoHOm DQMHOIN.N.N How o>soo coflumuufle .m muomflm
wooed Hum as
o.m m.H o.H m. o
- E la .
[00
/ I1
O
aw!
OOIO O
I I0,
O/OIOIO,
010
I
010
OOJ'O
'? I
,. i
ll

 

 

Hd

60

and we know
_ 1000 n
(c1) = -—————-—Vt 2 (5)
in which (Cl-) is the concentration Of chloride ion, (OH-)
is the concentration Of hydroxide ion and n2 is the.
number Of moles of HCl added.
Assuming that the extended Debye-Huckel theory is valid,

we have

4.509sz J‘S
logmhi) = 1 + 0.33 3J3 (6)

 

0
where S is the ionic strength, a is the distance param-
eter in Angstroms. (5A was used for these calculations).

The ionic strength is calculated from

_ 1 2
S - 2 f Zi Ci (7)

where Zi is the number Of charges on the ith ion, C1
is the concentration, and the summation extends over all
the kinds Of ions in the solution. In the present case, S

is given by
S = 0-5 X [(H+) + (CH+) + 4(CH2++) + (Cf) + (OH-)] (7')

Rearranging equations (1) and (2) and substiuting into

equation (3), we have

 

 

( ) 1000 n, 1 (3+) (H+)27: 8)
c = -——————- + + I, ‘.
_ Vt [ K1 lea Y2 ‘l (

Once we have Obtained the concentration Of 2,2,2-crypt, we

61

can calculate the concentration Of the other species in a
similar way.

With the help of the computer and the non-linear least
squares program, KINFIT155'155 (with the modified subroutine
EQN listed in Appendix 1), the titration curve was fitted
as shown in Figure 6. The two equilibrium constants, K1
and K2, were calculated to be 1.53 i 0.08 x 10-10 and
6.3 i 0.3 x 10"8 reSpectively (molarity basis). With these
two equilibrium constants, the concentration Of each species
involved at various pH values was also calculated by the
KINFIT program. Figure 7 shows the plot Of the concentra-
tions Of 2,2,2-crypt (C), monoprotonated 2,2,2-crypt (CH+),
and diprotonated 2,2,2-crypt (CH2++) versus the amount Of

HCl added.
5.3.1.2. Titration of Na+-cryptate with HCl

A similar study was done by titrating a sodium chloride
cryptate solution. A solution containing ~'0.1M 2,2,2-crypt
and a slight excess Of sodium chloride was titrated with
hydrochloric acid. The result is shown in Figure 8. In
this case, the titration curve showed only one break. Be-
cause a slight excess Of sodium chloride was in the solution,
the 2,2,2-crypt molecules were mainly in the form Of the
complexed sodium cryptate. X-ray studies have shown that,
at least in solids, the two nitrogens Of the crypt are in
the "in-in" form, i.e., both nitrogen electron lone pairs

are pointing towards the sodium ion. In order to protonate

pH

62

 

 

 

o
" "8 ..
-xO
x0
x0
x o
Ox
— Ox
x0
x0
Ox
ox
ml HCl added
Figure 6. Computer fit Of 2,2,2-crypt titration

curve .

 

63

 

 

 

Table 1. Computer generated data for 2,2,2-crypt pH titra-
tion: concentrations Of 2,2,2-crypt and its
protonated species, ionic strength and activity
coefficients.

pH (c) x 102 (CH+)x102 (CH:+)X102 s x 10 VI

11.30 8.81 0.306 0.29x10‘4 0.31x10‘1 0.94

10.80 8.08 0.922 0.31x10‘3 0.92x10‘1 0.90

10.41 6.83 1.99 0.18x10‘2 0.20 0.87

10.16 5.66 2.99 0.53x10‘2 0.30 0.85

10.02 4.91 3.63 0.92x10‘2 0.37 0.84

9.96 4.59 3.91 0.12x10"1 0.39 0.84
9.73 3.36 4.95 0.26x10'1 0.50 0.83
9.57 2.60 5.59 0.44x10‘1 0.57 0.82
9.40 1.92 6.14 0.73x10'1 0.64 0.81
9.02 0.88 6.88 0.20 0.75 0.80
8.75 0.48 7.01 0.39 0.82 0.70
8.51 0.27 6.87 0.68 0.89 0.79
8.37 ’0.19 6.66 0.93 0.94 0.79
8.12 0.95x10‘1 6.04 1.54 1.07 0.78
7.82 0.38x10'1 4.89 2.60 1.27 0.77
7.62 0.19x10’1 3.96 3.43 1.42 0.76
7.45 0.10x10‘1 3.16 4.14 1.56 0.75
7.30 0.57x10'2 2.50 4.72 1.67 0.75
7.08 0.23x10'2 1.70 5.43 1.80 0.75
6.98 0.15x10'2 1.41 5.69 1.85 0.74
6.64 0.35x10'3 0.70 6.31 1.96 0.74
6.43 0.14x10'3 0.45 6.53 2.00 0.74
6.22 0.53x10“ _0.28 6.68 2.03 0.74
5.20 0.49x10'° 0.27x10‘1 6.90 2.07 0.74
3.00 0.20x10‘1° 0.17x10'3 6.91 2.08 0.74
2.80 0.78x10‘11 0.11x10‘3 6.90 2.09 0.74

 

Concentration ( M )

64

 

 

 

 

.08 -°
I- A
o A .6 + XXX ++
(C) A (CH) Xx (CH2)
X
.06 - A A
O A X
X
L- o a A
o x
X
.Oh r A A
A
0 x
A
h- A
x
o x A
.02 — A K
O
A
X
A
x A
0 A
{ xxoo A
O x oh: 4; l
O l 2

ml HCl added

Figure 7. Plot of the concentration Of crypt and
Of monoprotonated and diprotonated
crypt versus the amount Of HCl added.

65

.COfiuSHOm wumummuol+mz How w>uso :ofiumuufia .m musmflm

 

 

cocoa Homdu
m m a m o
d . _ a _ _ _ — . q _ a _ _ _
'OIII'OIO
O
1
OIOIO
If

 

 

OH

Hd

66

the nitrogen, it seems necessary to have the trapped sodium
come out of the cavity or, to have the proton squeeze into
the cavity which is not likely. In addition to the two
equilibria mentioned above for the protonation of 2,2,2-

crypt, a third equilibrium is involved.

+ K3
Na + C jf-# CNa

The equilibrium constant, K3, as obtained by Lehn and co-
workers, was found to be 7.9 x 103.106

Because of the additional equilibrium, the calculations
also had to be modified. From the above equilibrium, we

have

= (CNa+) 9
K3 (Na‘F) (c) ( )

where (Na+) is the concentration of Na+ and (CNa+) is
the concentration of Na+-cryptate. The material balance

conditions become

1000 ml

(CNa+) + (CH+) + (CH2++) + (c) = v

. (10)
t

In addition, we have the sodium balance represented by,

1000 n
(Nan + (wan = 7.2.

(11)
. t n

where n3 is the total number of moles of NaCl added. The
charge balance becomes
+4")

(mi) + (cum + (11*) + (an?) +(cn2 = (of) + (on') (12)

and the expression for (Cl-) is

67

 

1000 n 1000 q
Vt Vt

The ionic strength is calculated as follows:

S=O.5[(H+)+(Na+)+(CNa+)+(CH+)+4(CH2++)+(Cl-)+(OH-)] (14)

From all of these equations, we have

 

 

1000 ql (H+) (H+)271 1000 naKs
(C) = -——————— [1 + +--——————— + ](15)
Vt K1 K1K272 Vt (I+K3TC77

Again the concentration of the other species can be derived
in a similar manner.

In order to check this postulate, a computer fitting
was performed in the same manner as was done for the 2,2,2-
crypt case. The result is shwon in Figure 9. The equi-
librium constant K3 was found to be 1.3 i 0.1 x 104. A
plot of the calculated concentration of CNa+, CH+, CH2++
versus the amount of HCl added is shown in Figure 10. The
excellent agreement between the experimental data and the
calculated values indicates that the presumed protonation
scheme is correct. Furthermore the initial amount of 2,2,2-
crypt used can, if desired, be determined by treating it as
an adjustable parameter. The answer obtained from the
computer fitting (1.68 i 0.01 x 10-4) is in excellent

agreement with the experimental value (1.68 x 10-4).

pH

68

 

 

X0

X0
X0
X0
X0

X0

 

Figure 9.

ml HCl added

Computer fit of Na+-cryptate titration
curve.

 

69

 

 

 

Table 2. Computer generated data for sodium ion-cryptate
pH titration: concentrations of protonated
2,2,2-crypt and sodium ion-cryptate, ionic
strength, and activity coefficients

pH (CNa+)x102 (CH+)x102 (CH:+)x102 s x 102 yl

9.41 3.13 0.991410"1 0.11x10‘2 3.72 0.34
9.27 3.03 0.13 0.20x10‘2 3.73 0.34
3.31 2.34 0.27 0.12x10‘1 3.79 0.34
8.60 2.63 0.36 0.27x10’1 3.35 0.34
3.39 2.45 0.47 0.56x10‘1 3.94 0.34
3.19 2.13 0.57 0.11 4.07 0.34
3.11 2.06 0.61 0.14 4.13 0.34
7.90 1.67 0.69 0.26 4.37 0.33
7.30 1.49 0.71 0.33 4.50 0.33
7.70 1.25 0.71 0.44 4.63 0.33
7.63 1.21 0.71 0.46 4.71 0.33
7.57 0.97 0.70 0.53 4.92 0.33
7.50 0.32 0.67 0.66 5.05 0.33
7.42 0.67 0.64 0.76 5.21 0.32
7.32 0.50 0.53 0.33 5.39 0.32
7.20 _ 0.34 0.50 1.01 5.59 0.32
7.07 0.21 0.42 1.14 5.77 0.32
6.97 0.14 0.35 1.22 5.39 0.32
6.36 0.39::10'1 0.29 1.29 5.93 0.32
6.66 0.33x10‘1 0.19 1.33 6.11 0.31
6.24 0.57x10'2 0.79x10‘1 1.43 6.24 0.31
3.60 0.30x10'7 0.19x10'3 1.53 6.31 0.31
3.03 0.27x10‘“ 0.55x10" 1.51 6.30 0.31

 

Concentration ( M )

70

 

 

 

 

0.03“
o
o
o
o
0.02, 0
o (CNa+)
o x )‘ X
x
o o X X
X
(L01 p o X
X
o X
A DADA“ Q
A
A X g A
A
+ X ++ A
A(CH) x (CH2) 0 A
A X A
o A
A X“ o A
o 14x 1 J J l, L o a 4
O 2 h 6

ml HCl added

Figure 10. Plot of concentration of monoprotonated
and diprotonated 2,2,2-crypt, and Na -
cryptate versus the amount of HCl added.

71

5.3.2. NMR Studies of protonation
5.3.2.1. Protonation of 2,2,2—crypt

The NMR Spectra of 2,2,2-crypt in water were obtained
at different pH values. The most pronounced shift occurs
for the -NCH2- hydrogen triplet. At a pH equal to 11.8
(at this pH no hydrochloric acid has been added) the trip—
let is at 3.00 ppm. As the pH is decreased, the triplet
shifts downfield. The peaks merge to a broad singlet at a
pH value of 8.3 which then further shifts down to 3.76 ppm.
The NMR Spectra are shown in Figure 11. The NMR peak posi-
tion versus the pH is plotted in Figure 12.

The total shift (from 3.00 ppm to 3.76 ppm) results
from the protonation of the two nitrogens. Unless we could
determine the exact location of the NMR peak for both the
monoprotonated and the diprotonated species, it would not
be possible to draw further conclusions. A computer least-
squares fitting of the NMR shift was performed.155r155
It basically followed the same equations as those used for
the titration of the 2,2,2-crypt. However, the NMR shift

is calculated from the equation:

= + v +
vobserved vCPC CH+PCH+ VCH2++PCH2++

where v VCH+ and v represent the chemical shifts

c' CH2++
of the unprotonated, monoprotonated and diprotonated

2,2,2-crypt. P PCH+ and P represent the fractional

CI
pOpulations of these three species as computed from the

72

11.15;,

jfi J

.mcofluDHOm um>HoIN.N.N How

 

 

_
«.0 m H.0H M

manommm mzz ucmccmmwnlmm

m.3 m 9:

 

 

.HH madman

 

73

“~‘,*. '~.‘.LL' '

 

Apmscflucouv .Hfi muzmflm

3 w. . K J) 1.x.

 

7h

 

 

.mm m5mnm> umanm mzz mo uoam .NH wasmflm
ma
ma m o
u d a —
/
O
/
01/ I.
o
z
0 l
l
{0'
1

 

 

o.n

m.n

i

m.n

m.n

(mdd) uotqrsod need awn

75

acidity constants. At very high pH values, it can be
assumed that nearly 100 percent of the material is in the
and v

unprotonated form. By using as two

vCH+ CH2 ++
adjustable parameters, a computer fit was attempted. The
result is shown in Figure 13. The chemical shiftsof CH+
and CH3++ are found to be 3.25 i 0.02 ppm and 3.76 i 0.02
ppm respectively. The first protonation induces a relative-
ly small shift compared to the second protonation. The
merged broad singlet is located at ~43.36 ppm. Since 3.36
ppm is between 3.25 ppm and 3.76 ppm, the merging of the
triplet is probably due to the exchange of the second pro-

ton. The exchange rate at which the collapse just occurs

is given by

1 _ 2wJ

?‘T
where 1 is the mean life time of a given nucleus before
exchange and J is the spin-spin coupling constant ex-
pressed in cycles per second. The result for 1/1 in
this case is 22.2 sec”1 (1 = 0.045 sec). A sextet is ex-
pected at very low pH values due to the further splitting
of the triplet by the nitrogen protons. However, this
splitting is not clearly observed in the Spectra. A close
study at low pH values is needed to clarify this point.

The proton exchange mechanism, as mentioned in Chapter

II, was proposed by Meiboom and co-workers.133'124 The

deprotonation pathways are shown below:

76

.mm suAB uwwnm mzz wsu mo paw amusmeou
dopvm Hum HE

.mH wusmflm

 

 

OX

OX

H

0::

O):

 

OX

 

uoritsod xaad HWN

77

 

k
R3NH+ + H20 ‘ > R3N + H30+ (4)
* + - k -
R3NH + OH ——-L-> R3N + H20 (5)
+ k6 ' +
R3NH + NR3 —--—-> R3N + HNR3 (6)
' + k7 '
R3NH + O-H + NR3 ——--> RSN + H-O + HNR3 (7)
I I
H H ‘

Equation (6),in this case, can be further divided into intra
and inter-molecular exchanges.

The mechanism prOposed by equations (4), (6), and (7)
are all independent of pH while mechanism (5) is pH de-
pendent. A careful study of the water resonance line for
a concentrated solution of 2,2,2-crypt could tell how im-
portant reaction (7) is. If no line broadening is observed,
the exchange presumably does not go through the step indi-
cated by reaction (7). The sharp triplet observed at both
low and high pH values seems to indicate that reaction (4)
is prdbably not very important. Reaction (5) should proceed
much faster at higher pH values. In order to determine
whether reaction (5) makes a significant contribution or not,
a detailed study of the change of r with pH is needed.

The evaluation of k3 and k4 would require quantitative

treatment of the spectral data.
5.3.2.2. Protonation of Na+-cryptate

The NMR spectra of sodium-cryptate in D20 at different
pH values are shown in Figure 14. The sodium-cryptate solu-

tion (1.01 moles of sodium chloride to 1.00 mole of

78

 

.maz Ca mumummuol+mz How manommm mzz ucmocmmmnlzm .¢H musmwm

m.m

 

i4 6:.

 

 

 

 

_

_ macaw

o.oa up Agapo-m.m.m ,

“W A

80

mwgw

 

n: .a.-_'_'..._'_..;_. _ ' ‘

~.m

 

m.w

Anmscaucouv

 

 

 

 

“wv-

.va musmflm

>.>

 

81

2,2,2-crypt) has a pH value of approximately 10. The NMR
spectrum indicates that the major species in solution is
sodium-cryptate while only a small amount of free 2,2,2-
crypt is present. As the pH value changes from 10 to ~8.2,
the ratio of sodium-cryptate to free crypt stays the same.
In this region, the peak positions for the sodium-cryptate
protons remain unchanged, while the free crypt signals
shift downfield due to the protonation. At a pH of ~48.2,
the ratio of sodium-cryptate to crypt (protonated) begins
to decrease. However, the positions of the sodium-cryptate
protons do not shift at all. The signal from the crypt
moves further downfield. Finally at a pH value oflv 6.7,
the peaks from the sodium-cryptate protons are no longer
detectable and the NMR pattern becomes exactly the same as
that of the protonated 2,2,2-crypt at the same pH value.
These observations are explained in the following way.
As the pH decreases, the free crypt molecules start to pick
up protons and the signal shifts downfield. However, those
crypt molecules which hold sodium ions do not protonate
until the pH value reaches ~o8.2. At this pH value, the
sodium cryptate starts to dissociate into free crypt mole-
cules and sodium ions. As soon as the crypt becomes free in
the solution, the protonation starts. Actually, two protons
attach to the crypt molecule at about the same time, since
at pH values lower than 8 the second protonation is also
favored. Finally'when the pH reaches ~16.7, all the sodium-

cryptate complexes have dissociated into protonated crypt

82

molecules and sodium ions. The NMR spectra become identical
to those for 2,2,2-crypt alone at the same pH value.

This study indicates that the protonation reaction does
not take place unless the caged sodium ion leaves the cavity,
probably because the sodium-cryptate molecule is in an
"in-in" form112 and the sodium ion is held at the center of
the cavity. Since both nitroqens are in the "in" form, no
protonation can take place on the outside of the molecule.
The cavity contains the sodium cation and probably cannot
accommodate any other cations. Therefore, the sodium ion
must leave the cavity before protonation can take place.
Another possibility is that after the sodium ion leaves the
cavity, the nitrogens on the crypt molecule invert to the
"out-out" positions to make themselves available to a proton
on the outside of the cavity. Further studies are needed

to clarify this question.

5.4. Paramagnetic Shift of Crypt Protons in
Metal-deuterated Ammonia Solutions

Solutions of 2,2,2-crypt, KI-2,2,2-crypt and Na-2,2,2-
crypt in deuterated ammonia at various temperatures were
studied by NMR techniques. The results are summarized in
Table 3. The NMR spectra are shown in Figures 15-17. In
the case of the free 2,2,2-crypt in ND3, no change in the
spectrum was observed with changing temperature. However,
the whole pattern is shifted upfield by as much as 27 Hz

compared with the Spectrum of the free 2,2,2-crypt in water

83

 

.mnz Ca ummHUIN.N.N
mo poflcsum mzz unoccmmmplmnsumummame .mH ousmfim

_
in: g 5: rim} 4?

‘ om.mH| _ . om.NI _ om.:H

 

 

 

 

84

.mnz CH wumummnon+m mo mmficsum mzz ucmccmmwcuwusumummfime

a a a
A. filfl §§j A 5 {1;

_v

 

 

 

 

--———
_ -.-.- .—.-..-—"’
_nua—a—

oo.mm- N om.mH- om.>u om.:H

 

 

m,

85

.moz ca wumummuolmz mo mmflcsum mzz ucmccmmmclmusumummeme .ba ousmflm

14.4.1.4...444:.m... E”... # a .4... 4 4.. 444...... .1744 4. 4.... .r. :fiifzn. .4. )i
4 .. .
4. .4

 

 

om.m om.HH- . om.>mu oo.mmu

 

 

 

 

 

 

86

 

 

 

Table 3. NMR peak positions for 2,2,2-crypt protons in ND3
T(°C) .2,2,2,-Crypt . HI + 2,2,2-Crypt .
- Trip- Trip- Sing- Trip- Trip- Sing—
let let let let let let
-35 2.54 3.51 3.58 2.58 3.52 3.57
-15 2.53 3.50 3.57 2.57 3.57 3.59
- 7.5 2.53 3.50 3.58 2.58 3.48 3.52
- 2.5 2.55 3.50 3.58 2.60 3.52 3.56
+14.5 2.56 3.52 3.58 2.59 3.50 3.53
Na + 2,2,2—Crypt
Triplet Triplet and Singlet
8.5 2.04 2.88
-11.5 2.10 2.98
-37.5 2.18 3.12
—58.0 2.25 3.18

 

 

 

87

and the N-CHz- triplet appears to be much closer to that of
the metal cation-cryptate triplet. In the presence of
potassium ions the peak positions are only slightly shifted
from the positions of 2,2,2-crypt in ND3. Let us assume
that in the metal-cryptate case, all of the cryptate mole-
cules are in the "in-in" form. Then, one explanation for
the NMR spectra would be that the free crypt molecules have
almost 100 percent "in-in" form in ND3 but have mixed forms
in CDC13 and H20. If this is the case, the "in-in" to
"out-out" conversion must be relatively fast compared to
the NMR time scale which results in the appearance of only
one triplet for the solution.

In the case of sodium metal and 2,2,2-crypt in ND3,
the spectra are shifted further upfield. This shift may be
identified as a Knight shift (paramagnetic shift). The
shift at 8.50 is about 32 cycles per second. As the tempera-
ture is lowered, the magnitude of the shift decreases. The
volume susceptibility of the solution may be calculated from
the theoretical expression:157

6V _ 271' _
'0'; ' ‘3' OR. Xv')

where 0v is the paramagnetic shift, Va is the applied
field, Xv is the volume susceptibility of the paramagnetic
solution, and Xv' is the volume susceptibility of the
reference solution. (In this case it is I<+9cryptate solu-

tion.) If we assume:

XV. = O

88

then the volume susceptibility of the paramagnetic solutions

6 at 8.50.

is found to be 0.253 x 10-
In this preliminary study, even though the exact amounts
of sodium and 2,2,2-crypt added were not known, some impor-
tant qualitative results were obtained. First, since only
a small paramagnetic shift was observed, the unpaired elec-
tron spin density near the crypt is small. Second, if we
compare the shift of the -O-CH2CH2-O- and -O-CH2- hydrogens
with the -N-CH2- hydrogens, we find that the central portion
of the crypt seems to have a higher unpaired electron density.
Third, the volume susceptibility increases as the temperature
is increased. This probably indicates that the following

equilibrium shifts to the right when the temperature is in-

creased

e -—4
zsolv <—- 26

Solv

a plot of logarithm of the volume susceptibility versus
(l/T) is shown in Figure 18. A straight line was obtained
with a slope of -2.0 x 102.

Temperature studies of the free 2,2,2-crypt in ND3
showed no sign of broadening for either triplet. This indi-
cates that the nitrogen inversion rate is either too fast or
too slow to be detected by NMR techniques at these tempera-
tures. However, the temperature dependent studies of potas-
sium ion-cryptate in ND3 did show a broadening of the triplet.

At about 15°, the triplet nearly merges to a singlet. There—

fore, the exchange rate of the potassium ion in and out of

. -1
the crypt in ND3 at this temperature is about 22.2 sec .

89

 

Ibgfgr
6x
:4
I

 

l .L 1

 

 

3.5 14.0 4.5
l/T (°K'l x 103 )

Figure 18. Plot of logarithm of volume susceptibility
versus (l/T) for sodium-cryptate solution

VI. OPTICAL STUDIES OF ALKALI METALS IN

AMINES AND ETHERS*

6.1. Introduction

Alkali metals have been found to dissolve in some
amines and ethers to form blue solutions.12 At least three
species have been identified; solvated electrons, monomers,
and alkali anions. However, the number of solvents in
which stable alkali metal solutions could be studied was
severely limited by solubility and decompositionproblems.
So far only methylamine (MA), ethylamine (EA), ethylene-
diamine (EDA), propylenediamine (PDA), tetrahydrofuran (THF),

dimethoxyethane, hexamethyl phosphoric triamide, diglyme

and certain polyethers have been found to dissolve alkali
metals unassisted. Even in some of these solvents, true
solubility in the absence of impurities has been questioned.
The resulting blue solutions have one or more broad optical

absorption bands. The absorption bands of species with

stoichiometries M- and e are influenced by both

solv
temperature and the solvent. A blue shift with decreasing
temperature is characteristic of these bands. The position
of the M- peak is also strongly influenced by the metal

used, whereas the e- band is independent of the

solv

.4

*-
Some of the data included in this chapter are taken from

In J. Tehan's Ph.D. Thesis, Michigan State University (1973).
90

91

metal. In some solutions, ESR studies also prove the exist-
ence of relatively low concentrations of a species with the
stoichiometry n.21723724793194 It is believed that the
equilibrium scheme shown in equations 1-3 involving these
three species adequately describes the experimental results

for all solvents.

2 M (s) 2:3'M+ +'M- (1)

LL: M + eSolv
IL'""'""""M+ + e'

solv (3)

(2)

For example, metal-ammonia solutions lie far to the right

in this equilibrium scheme, metal-methylamine, and ethylene-
diamine solutions show all three reducing species, and metal-
ethylamine and diglyme solutions have mainly M- present.
Solvents such as diethyl ether are unable to dissolve the
metals at all without assistance. However, by using cyclic
polyethers of the crown and crypt classes to complex the
alkali cations, the equilibrium scheme can be shifted by the

introduction of the additional equilibrium.13.31.32

 

M++C<

 

> MC+ (4)
The cation complexing agents not only made it possible

to dissolve alkali metals in a wide variety of solvents but

also to obtain relatively large concentrations of M— and/or

e solv by selecting appropriate amounts of complexing agent

and metal.

92

Besides the continuing investigation of alkali metal
solutions in ammonia and in various amines and ethers,
pulse radiolysis studies of the solvated electron in many
organic solvents have been made.14'17'23I36'3"!106 Despite
the general agreement between these two methods (as shown
for example by Dye, DeBacker and Dorfman26 and by the work
to be described later in this chapter), there still exist a
number of unanswered questions. For example, what is the
effect of the cation and its aggregates upon the shape of
the optical absorption bands? In order to answer some of
these questions, pulse radiolysis studies of solvents both
with and without cations and cation-complexing agents are
needed. We report some preliminary experiments by pulse—
radiolysis methods which begin to clear up some of the
questions.

The studies performed here can be divided into two
parts, spectrOphotometric studies and pulse radiolysis
studies. In the SpectrOphotometric study we have (1) ex-
tended the number of solvents in which the alkali metals
can be made to dissolve; (2) measured the peak position and
its temperature dependence for the species Na-, K-, and,

in most cases, e- and (3) compared the result for

solv

e solv with those obtained by pulse radiolysis and flash

photolysis. In the pulse radiolysis study we have (1)

mapped the Spectrum of e solv in ND3; (2) studied the

effect of the complexing agent in Na+-THF solutions: (3)

93
carried out studies of the temperature dependence of the
band shape in 1,2-propanediamine.

Besides these two major tOpics, a study of the solubil-
ity characteristics of potassium in ethylamine and in ethyl-
amine-ammonia mixtures, a brief study of the Na-KI-EDA
system and of the cesium-methylamine system was performed
in order to clear up some questions which have existed in

the field for a long time.
6.2. Experimental

All the optical spectra were taken with either a Beck-
man DK-2 Spectrophotometer or with a scanning system which
was designed for stopped~flow measurements.158 The UV
Spectra of the iodides were determined with a Cary 14
spectrophotometer. The cell compartments of the DK-2 and
the Cary 14 were modified to permit variable-temperature
measurements over the range of -1000 to +700. The tempera-
ture was controled by means of a Varian V-4540 Variable-
temperature control unit and measured with a thermocouple.

The optical cell with a pafl11ength of 2.0 mm was made
of fused Infrasil quartz (Kontes). When a smaller path
length was needed, a quartz plate of thickness 1.5 mm was
introduced as an insert. In some cases, optical path
lengths of 10 mm or 0.1 mm were used for very dilute solu-
tions or concentrated solutions respectively.

The pulse radiolysis experiments were performed in the

laboratory of Professor Leon M. Dorfman at the Ohio State

94

University. A Varian V-7715A electron linear accelerator,
delivering 3-4 MeV electrons at a pulse current of about
350 mA was used as the source of the electron pulse. The
pulse duration was 100 ns to 1500 ns. The infrared de-
tector used was a solid state diode employing a diffused
junction of indium antimonide. The detector was cooled with
liquid nitrogen. The Specified spectral range is from 1000
to 5500 nm. For the visible range, an H.T.V. 196 detector
with an S-1 response was used. A Bausch and Lomb grating
monochromator, type 33-86-25, f/3.5 was used with several
different gratings. Appropriate Corning filters were used
to eliminate second-order components. A 500 watt xenon
light source was used.

Reaction cells (the same type used by DeBacker141)
with 3.0 mm optical path length were used for all of the
studies. The exact amount of metal or other desired com-
pound was predweighed in a breakable bulb and placed in the
sidearm of the reaction cell. The solvent to be studied was
distilled into the reaction cell under high vacuum. Then
the desired compound was introduced into the solvent by
breaking the appropriate bulb. The cell was placed in a
variable temperature compartment and the temperature was
controlled by means of a Varian V-4540 variable-temperature

control unit and measured with a thermocouple.

95

6.3. Results and Discussion
6.3.1. Spectrophotometric studies

With the aid of dicyclohexyl-lB-crown—6 and 2,2,2-crypt,

the optical Spectra of Na , K-, and e-solv have been ob-
tained in a variety of solvents. The number of solvents
for the study of these species has now been extended to
include primary mono- and di-amines, secondary amines, and
straight, branched chain, and cyclic ethers. The solvents
in which the optical band of K- has been observed are
listed in Table 4. It was not possible to dissolve potas-
sium in triethylamine (D = 2.42) even in the presence of
2,2,2-crypt. Because of the solubility enhancement and
solvent range extension it has now become possible to better
understand the equilibria existing among different Species
in solution. The overall equilibrium scheme given previously
must Simply be modified as mentioned earlier to include the
effect on reactions (1) and (3) of the complexation reaction
(4).

When a metal such as Na or K dissolves in liquid ammonia,
the equilibria lie far to the right and only e-solv can be

detected optically or by the ESR techniques. On the other

hand, with EDA, MA, EA, etc., mixtures of M-, M and

e solv result, and the solubility is drastically reduced.

In all cases, M appears to be a minor Species and only

for a few solvents is e- optically detectable. In

solv

solvents such as DEE, DEA, etc., the metals are normally

.mmmwsucmumm CH cm>flm ma mnsumumm50u Eoou ummc ucmumcoo Ufluuomawfln
*

.x.

 

nonuw ammoumnmrfla

m NV
6 mV
0.4V uwbum ammonomflnfla
m mV 0cflem 4mn004o
m 4V

40:00 Hanumflo

Um>ummno Omam >Homm
'*
Av.h V cmusmouomcmuume
Am.h V mcmnummxozuwefln
. wcHEm ammoumlmnfln mamaman

mcHEmflcmcmmoumIN.a

Am.® V mcflEmaxnum
Av.m V *wcflsmamnuwz
Am.N~V *mcfiEMHomcmahzum
4*Ao.omV mofiEMSHu *Ufluozmmonm Hanummemm

 

oumummno no czouo monasomm

pmumwmmmcb mm>aommwo

 

 

.@m>uwmno coma um: I& L0H£3 CH mucm>aom .v magma

97

insoluble and a cation complexing agent is required to
dissolve the alkali metal. In all cases tested, the addi-
tion of crown ether or 2,2,2-crypt increased the solubility

of the metal and tended to give M- and/or e- The

solv'
spectrum of the latter Species was always observed (fre-
quently along with the spectrum of M-) when 2,2,2-crypt
was added (in excess of the amount of metal dissolved) to
those solvents listed in the second column of Table 4, and
when either crown ether or 2,2,2-crypt was added to solu-
tions of potassium in the solvents listed in the first
column. These results are adequately described by equi-
libria 1-4.

At 25°, the positions of the absorption maxima of Na-,

K-, and e- in these solvents vary over the range of

solv
12,700 to 15,200 cm-l, 10,700 to 12,200 cm"1 and 4,330 to
7,810 cm-l, reSpectively. The peak positions for Cs-

in methylamine and ethylamine were also determined to be

8910 and 9520 cm-1, reSpectively. In all the solvents tested,
the peak positions proved to be strongly temperature de-
pendent. A linear blue shift with decreasing temperature

was Observed in all cases as shown in Figures 19, 20, 21,

and 22 for Na-, K-, Cs-, and e- The temperature coef-

solv'
ficient (determined by least-squares analysis) together with
the corresponding positions of the maxima at 25° are re—
ported in Table 5. Although the variation of the tempera-

ture coefficient with solvent is only a few times larger

than the error in measuring it, the results are in agreement

98

 

 

  

   
 

 

 

4
L 4 o
15-—
1717 \0
\DIGLYME
IA —- f
z
2.: «\gk
0- 'o N
3 ‘ .orHr
23‘ - '
g I2 I” \3 g
g 3 "” \‘
9 6 ” \ \.
u e 5 ‘
z I; \ A
a p ‘ ‘
15_— ‘\‘ \\A , I2PDA
\A 5“
a . \
r- a B
\
a 8 B\A
a 8
L4 - ‘\a \\\
\ ‘\DME
_ "DEE
1 1 1 1 1\ J 1
—100 -—60 -—20 20
t ( ° c )
Figure 19. Temperature-dependent studies

of the position of the band
maximum for Na in various sol-
vents. (Note the displacement
of the vertical scale.)

99

 

 

 

 

 

 

1212 -
114 >
§.V|thl.06—- X
:2 o
a ':
2— ~—
2 5 65 £5
a -.
=- 6‘
I E
x |h -
1.22
114
106 *7 J I l l I l __,
-40 1—60 ~20 20
t(°c)
Figure 20. Temperature-dependence of the position of

the band maximum for K— in various 501-
vents. (Note the displacement of the
vertical scale.)

100

.mcHEmamnumE Ca ImU How
Eseflxme pawn 0:» mo coHuHmom on» no mocmwcmmmpimusumummeme

AooV whopwhmasme

ON 0 ON... Opal

.HN madman

 

 

 

Tm

 

 

( g_OI XI_w0 ) xoqmnu GABM

101

 

 

 

 

A
A
1,2PDA
9 F—
A '— A
. Z
9.70 A
.- o-
?n' x 8 "" A
0!-
5'2 A
x ta
(V F
3 s
> E X
'3 l}. A
d” 7 7‘ \\\\
" EA
‘\b\\\\\
‘b\o x
\
oO DEE
6 —— °\
0 K\\
_ \\\\\\\\\ Ox\\\
5 J 1 .4 .L 1 1
400 —60 —20 20
o .
t( c )
Figure 22. Temperature-dependence of the position of

the band maximum for e in various

solvents. SOIV

102

 

.Amnmav muflmum>flco mumum :mmflnoflz .mflmmza .o.:m m.cmnm9 .h .m scum cmxmu mama
*

.omnIo

.Ammaflv omen xmm .Emnu .mwnm .n ~Emsmo .m .m cam mxooum .2 .no

.COflUMH>w© cumccmumn

.cwuoc mmfl3uo£uo mmwacs can um cowufimomm

 

 

 

 

 

 

III III III NN.a III III mumummuo madam Hmmoumkmrfln
III III III mo.H III bN.H wumummuo Hmnuw Hamoum0mfiIfin
III Hm.m III vfi.dz mm.o «m.mfl mm.” mumumauo mcflamamnumfia
mm.o mm.mH mH.m mo.fl wH.mH so.H mm.H NH.>H om.fi mumummuo umgum Hsgumflo
III III III bo.H III Hm.H c3ouu Hmnum Hmzuwfln*
III $5.0 um.o mm.b Na.~ vm.o pm.w~ mm.” s3ouo mcmsummxonumEHQ*
III mm.v III HH.H mm.o n®.ma mm.H mumumauo nonsmouomnmnums
III III mv.o ¢b.m NH.H III mm.H cBOMU CMHSMOHomnmuumB*
II. can.» mm.o mm.MH mH.fi um.H mm.mH ov.H azouo mamamfla*
nm.m no.0m om.m hv.o mm.oH oH.H NN.H mm.NH m¢.a GBOHU mcfiEMflomcmmonmIN.n
mm.o mm.mm vm.m vfi.o no.» mH.H hm.H om.oH av.a :3ono mCHEmawzum
om.H mm.mH ow.> mm.o «v.5 wa.H m>.o mm.mH mn.fi III mcfismfiomcmHmsum*
III . III . III . III moHEMMHu canonmmonm
ovv v so H fin H Iamnumemxmm*
m Ion m Ion m oax
no xwe vme 90 xwe vxms no wa wwwe usmm<
poI 9 DUI p pol p mcflxwa ucm>Hom
IsdqmeIll Ix Imz Imsoo
>Hom m.mu:m>aom

macaum> Ca Im 6cm . x .Imz How musmwoflmmwoo musumnmmeou ocm mcowuwmom xmmm

.m magma

103

with the prediction of charge-transfer-to-solvent (ctts)
theory34135a158 that the position of the maximum be a
linear function of the temperature coefficient. According
to Smith and Symons' theory84 the SIOpe of a plot of
Vmax XE-dvmax/dT should be T. The results for Na- are
shown in Figure 23. The solid line has a slope equal to
the absolute temperature.

Based upon these results, the shift of the Na- and
K- absorption bands with both solvent and temperature is
presumed to be characteristic of a ctts band as first sug-
gested by Matalon, Golden, and Ottolenghi.25 As a further
test of the applicability of ctts theory, a linear correla-
tion of the absorption maxima of two anions measured in the
same set of solvents is expected.86 Matalon, Golden, and
Ottolenghi25 found such a linear dependence of the peak
position of the iodide band upon the peak position of the
Na- band at the same temperature in ethylamine-ammonia
mixtures. We attempted to extend this test from mixtures of
two solvents to a variety of solvents. To do this, the
spectrum of K1 was also studied by us in these solvents.
To enhance the solubility of KI, crown was added to the sol-
vents, EA, EDA, DME, THF, and DEE, while cryptate was used with
DIPE. No complexing agent was used for the KI-HMPA system.
However, as shown in Figure 24, the correlation failed.

The lack of correlation might be associated with the

iodide absorption Spectrum for the following reasons.

1. The iodide ctts transition is strongly dependent

104

 

 

 

 

 

EDA
15*-
I2PDA
2;“
O
z '2
I71 x
2T: 1.4 *—
x u
:5 V.
a. o
'uliE
Z
13'“
1 I 1
-10 -14 -18 #22
damn/Ur
Figure 23. The relation between the Na peak posi-

tion at 25° and its temperature coef—
ficient. The slope of the straight line
has a magnitude of 2980K.

105

.omN um mucm>aom m50aum> ca

 

 

Imz mo cam IH mo coauwmom xmmm on» cmeuwn sowumeH one .vN musmwm
A IH wISxTzuorEh
0v 3. On
# _ _ _ N—
0
win.
0 o l «A
mmo (m2:
4.
w
0 0
m1» x
o M
320 o J 3 w
(m I.
X
m.
a 7
N
. o.
I nF,I

 

 

106

unpon the cation in solvents of low and intermediate di-
electric constant, presumably because of ion-pairing
effects.”""'161

2. Determination of the position of the first ctts
transition may be complicated by overlap with other
transitions.

3. The onset of solvent and/or impurity absorptions
can interfere with the determination of the I-absorption
band. These difficulties may also account for the poor
correlation of vmax with deax/dT for I-observed by
several authors with solvents of low dielectric constant.

Although the Spectrum of Na- fails to correlate
with that of I_, the spectra of Na" and of K- are well-
correlated with each other. The peak position of Na- XE
that of K- at 25° is shown in Figure 25. The linear cor-
relation is excellent for all solvents in which both the
Na- and K_ bands were observed. It is clear that what-
ever property (or properties) affects the absorption maximum
of Na-, it has a similar effect on the absorption maximum
of K-. Interestingly, the ratio of the variation of the
peak position of Na- to that of K- with solvent is about
the same as the ratio of their temperature coefficients.
Correlation of the peak position with various solvent proper-
ties such as the static (DS) and optical (Dop) dielectric
constants, dipole moment, polarizability, Kosower's Z
valueslazol63 and (l/DOp — l/DS)2 have been tried without

success .

Na'PsAK POSITION

107

 

 

 

 

EDAgX/
L5—-
12PDAo
14 EA”
° F ODIGLYME
ODME
°THF
HMPA L
055°
13r-
ONT
0
1.2 ' I l
1.0 1.1 1.2 1.3
_1 __
( Cl“ ><IO 3 )
K" PEAK POSITION
Figure 25. The relation between the peak position

of Na and of K' in various solvents
at 25°.

108

Compared with the M- absorption, the e- band

solv
is often much broader and has a larger solvent dependence

and a greater variability of the temperature dependence.
Figure 26 shows the solvated electron absorption spectrum

at room temperature in a number of different solvents. AS

is readily seen from this figure, there are marked variations
from one solvent to another in both the band shape and the
peak position. For example, while ethylamine and diethyl
ether give rise to nearly the same position of the absorp-
tion maximum, there is a very pronounced difference in the

width of the band. The spectrum of e- in ethylamine

solv
is much narrower than it is in diethyl ether. 0n the basis
of similar solvent properties one might expect similar band
widths in EDA and 1,2-PDA. However, EDA yields a much

broader band for e than does 1,2—PDA.

-solv
The solvated electron band width and position are not
only affected by the solvent, but are also strongly influ-
enced by temperature. As shown in Figure 27, the solvated
electron band shape in 1,2-PDA shifts to higher energies
and becomes broader as the temperature is decreased. This
same phenomenon was also observed in the case of ethylamine
(Figure 28). In spite of the greater sensitivity of the
Solvated electron spectrum to solvent and temperature, the
position of the absorption maximum correlates well with
those of Na- and K- as shown in Figure 29. Therefore,

the predominant effect of solvent on the Spectra appears to

be the same for all three cases. The band width of M- is

109

A.mm>uso may
mo meow mo ucmEmomHmmHU HmUHuHm> mzu muozv .mucm>H0m
msoflum> Ca com um couuomam owum>aom on» mo mnuommm one

Adm—x129 7..

OF 0 m u o n

.om musmflm

 

l

o.—

 

 

 

 

110

.mmusumummfimu mSOHHm> um mafiEMAo.

 

ImcmmoumIN.H Ca conuomam omum>aom mnu mo Esuuowmm one .bm wusoflm
A123 $5232 m><>>
ooo__ oooo ooon ooon
fl _ a . . l _ _ \\%® 0
.® \\\
/ \ \x I S
o o
x \ J v.
\ o
o l o.
/o / ®\V\M\m
Ill/lI/IN/llllllltd .1 m.
0..ka o
occulk
RT 4

 

.32 o

 

 

111

. .mwusumummewu
msoflum> pm mCHEmHmLDm Ga couuomam owum>aom on» mo Eduuowmm one .wm wusmflm

Annoa x Huao V hopes: m>m3

 

 

m m w m m
A i _ _ _ o
u\
\\

is/
\Q
\\°
mmv/v

 

 

>H0m
.onm um mucw>aom msoaum> Ga I0

mo cam x mo coaufimom xmmm may smm3umn coaumHmH one .mm musmflm

20....»05 8<un Iv.
AVIIOP X FIEUV XOEP‘.

. fi 2 o.—

112

 

 

(an

(nan—

 

_

l
v

l
o

 

 

)XDW—A-

OlX W)
(I:— I-
NOILISOd )Ivad ”"9

113

also dependent upon temperature, but it decreases with a

decrease in temperature in contrast to the band width of

e solv. For example, the width at half-height of the K’

band in ethylamine decreases from 3320 cm”1 at 39° to
2480 cm”1 at -24°. Similarly a decrease from 4150 to 3190
cm"1 occurs between -9° and -71° for the Na- band in di-
ethyl ether. A similar but smaller effect has been noted
for the absorption of I- in water.85

It is important to determine whether the change in

shape of the spectrum of e with temperature is an

solv

intrinsic prOperty of the solvent or whether the influence
of solute concentration is important. The formation of

. . + - . .
ion-pairs between M and e or of a spin—paired

solv
Species”.68 could result in temperature-dependent equi-
libria among species with slightly different spectra.

Several authors19I33v29'154v135 have reported the presence

of Optical bands which cannot be attributed to either M-

or e solv' Such bands could arise from the interaction of

e solv with the cation28.29,164,165 and might be respons-

ible for the variations in band shape which we Observe.

Although the Spectrum of e could be studied as

-solv
a function of concentration to determine whether the band
shape changes, intense solvent absorptions limit the path
length which can be conveniently used. Therefore, perhaps
the best method to separate concentration effects from sol-

vent effects is to compare these results with those deter-

mined by pulse-radiolysis. In the latter case, the

114

concentration of ions can be kept very small so that ion-
pairing effects can be minimized.

As shown in Figure 30, the position of the absorp-
tion maximum found in this work correlates well with the
results of pulse radiolysis17 and flash photolysis.28v29

However, the spectrum of e- obtained by dissolving

solv
metals is, in some cases, broader than that observed by
pulse radiolysis. This is shown in Figure 31 in which the
spectra obtained in this work for the solvents THF and DEE
are compared with the corresponding spectra measured by

pulse radiolysis.17 The broader band observed in the

present study may also account for the fact that the peak
positions obtained from metal solutions tend to be at
slightly higher energies than those found by pulse radiolysis
(Figure 30). Clearly, additional information about the
concentration dependence of the absorption band of e-solv

is needed to separate the effect of solvent from that of

solute interactions.
6.3.2. Pulse radiolysis studies*

Careful spectrum mapping of the solvated electron
absorption Spectrum in pure deutero-ammonia has been done.
The measured Spectrum covered the entire range of the ab-
sorption from «1800 nm to ~v2500 nm as shown in Figure 32.

The complete data are given in Table 6. The Spectrum

*
This work was done in collaboration with F. J. Tehan.

115

 

 

 

 

 

7.5 L-
Y:
Ui/~\
5‘?
g o
2'} 6.5 r-
“‘l
2'
1“ IJ
U’~_/
g s
E
“‘ "‘ 5.5 —
4.5 THF 1 l J
4.5 £15 6:5 215
—— ——1 -—3 '
vm.x(cm on )
AflEflllL S<>LllTl<>F6
Figure 30. The relation between the peak position

Of esolv

and by pulse radiolysis. .The straight

line has unit lepe. (Comparison in

the case of diglyme is with flash photol-
ysis rather than pulse radiolysis.)

obtained by dissolving metals

116

.mma mo mammHoaomn wmasm IIIIIII .mumumhuo £ua3 was Ca
mTllll..me mo mammaoaomu madam I.I.I.I .mumummuo nua3 mma ca
M mam>H0aomu mmasm an cam mamume msa>aomma© wn cocamuno

Mme ocm mmn Ga onm um >Homw mo mmmcm flown ms» mo cowanmmeou .fim musmfim

AmIOPX—IEJ tuhEacuZu;

 

 

e m k e . m w
A _ e e _ e _ J A 2 0.0
0.0 I II N6
NO I. II v.0
V
we I I 8 W
m
0
8 I I no x
m0 .1. l O.—

 

 

 

117

.mumo mflmhaoflomu Omasm m.cmamuoa cam waommma .O>Q QQQQQ
“mumo mammaoflomn mmasm H50 0000 “GOHuSHOm Hmuwe
“moz CH mcoflusaom Hmuma an ocm mammaoflomu mmasm mg omosoOHm

 

 

couuomam omum>aom mo muuommm coflumnomnm may no conflummaoo .Nm musmflm
A mnoa x HIEO V gonad: m>m3
0H m m e
. _ . (a . q _ _ o
oAV Q Ox
0 0 II
o \
O 0 IL

 

CO

Ilumfio
IL

Ill
'1, 00H

 

 

V/V

X301

118

Table 6. Data from the pulse radiolysis study in ND3.

 

 

 

Spectrum NO. A (nm) A
21 800 0.0471
15 850 0.0552
25 900 0.0728
19 950 0.0879
23 1000 0.0989
17 1050 0.1648, 0.1528
22 1100 0.1828
20 1150 0.2222
24 1200 0.2887
14 1250 0.3199
18 1300 0.4084, 0.3917
16 1350 0.3406
5 1400 0.3585, 0.3665
8 1450 0.3678, 0.3812
1 1500 0.431 , 0.415
11 1550 0.3165
4 1600 0.278
7 1650 0.2179
2 1700 0.2205
12 1750 0.1608
10 1800 0.1266
6 1850 0.0901
9 1900 0.0910, 0.0885
3 1900 0.0632

13 2000 0.0211

 

119

Obtained for metal-ammonia solutions compares very well
with that obtained by pulse radiolysis on both the high
energy and the low energy sides. The data also coincide
with the earlier data on the high energy side -- reported
by Dye, DeBacker and Dorfman.26 Recent unpublished data of
Fletcher and co-workers36 Obtained by the pulse radiolysis
of ND3 are also in complete agreement with our results.
This total agreement once again furnishes confirmatory evi-
dence that the solvated electron formed by metal solutions
in ammonia is the same as that produced by pulse radiolysis
and that the effect of the cation on the spectrum is very
small.

The temperature dependence of the solvated electron
band shape in 1,2-prOpanediamine was also studied by pulse
radiolysis. The solvated electron Spectra in 1,2-propane-
diamine at room temperature and at —25° are Shown in Figure
33. The half-widths at half-height for the room tempera—
ture and -25° Spectra were found to be ~'5800 cm.1 and
~ 6400 cm-1 respectively. This result is consistent with
the results found for metal solutions.31 Therefore, in this

case at least, the change in shape of the Spectrum of

e solv with temperature is likely to be an intrinsic

property of the solvent. However, it must be noted that

the intensities obtained in the pulse radiolysis study were

low and the Signal-to-noise ratio was not very good.
Finally, the study of the cation effect in THF was

carried out in the following way. First pure THF was

120

.mosum mfim>HOHomH wmasm 4 pom O “mosum :Oflusaom

 

 

HmumE umflm>HOHomu mmasm an UOGHMDQO mmusumummemu msownm>
um OCHEmHUmcmmOHmIN.H Ga couuomam omum>aom on» mo Esuuommm 038 .mm wusmflm
AOIOH x auao V gonads m>w3
ma 0H m m e
A e . o

 

 

O.H

 

 

XmV/V

121

studied and the complete spectrum of the solvated electron
was Obtained (Figure 34). The result is in good agreement
with the results Of Jou and Dorfman15 except on the low
energy side. The difference might be due to the fact that
the light source reached its cut-off region and the solvent
absorbed strongly to give a relatively poor spectrum. After
the solvated electron Spectrum in the pure THF had been ob—
tained, sodium tetraphenyl borate, (Na+¢4B-) was added.

The spectrum changed drastically. The solvated electron
band was completely absent and a new band with a maximum at
11,300 cm-1 was formed (Figure 34). The same result had
been obtained earlier by Bockrath and Dorfman.16 They at-
tributed this new absorption band to the sodium cation-elec-

tron pair (Na+- e- ). Further addition of the 2,2,2-

solv
crypt almost completely restored the spectrum of the sol-
vated electron in pure THF (Figure 34). The 2,2,2-crypt
presumably prevents formation of the sodium cation-electron

pair by Shifting the equilibrium to NaC+ and e- The

SOlv'
solvated electron spectrum in the sodium-cryptate solution
was broader than the solvated electron Spectrum in pure THF.
The nature Of this broadening cannot be understood without
additional information. These results are similar to the

-recent results of Fletcher, 32 31.13 who studied the effect

of "crown" in amine solvents.

122

.mumo o + +mz + mus use O “mumo
+mz + has H50 x .mumo mma ousm Moo 0 .mumo m .smamuon paw sumnxoom nu
.mumo m .OMEquQ ocm son O .aosum OOHusHom Hmuma d .ummHOIN. N. N cam
m2 zuHB has won +mz nuH3 mma .hmB musm mo mmHosum mHmmaoHomu OmHsm .vm musmHm
+ A OI oax HI So V Mugabe m>wz

 

2H ma NH HH OH m m N m m
u u - a # d u \ a d d
X
HU\ l NAV
o/o /“

IX'UIX QmQICOCQ 01.0...”

 

 

 

123

6.3.3. Other studies
6.3.3.1. Solubility of potassium in ethylamine

The solubility characteristics of potassium in ethyl-
amine have been a subject of controversy for a long time.
In some laboratories, it has been claimed that metallic
potassium dissolves immediately in pure ethylamine to form
a dark blue solution; However, this result is not charac-
teristic of the results in our laboratory. The only solu-
tion of potassium-ethylamine ever formed in our laboratory
is a very light blue solution which indicates that only a
very small amount of potassium is dissolved. We were able
to dissolve large amounts of potassium in ethylamine with
the help of cation complexing agents. The only other way
which we have been able to produce a dark blue solution with
potassium in ethylamine is by letting the solvent remain in
contact with the metal for a long time. During the long
waiting period, small amounts of NH3 may be formed as found by
Nicely.166 In order to clarify the nature of this problem,
the experiment was carefully repeated with pure potassium
and pure ethylamine.‘ First an experiment was done to rule
out the possibility of forming an oxide layer on the metal
surface and therefore, preventing the direct contact between
the metal and ethylamine. In order to do this, ethylamine
was distilled into the sample tube and kept frozen, then
the potassium metal was distilled through a sidearm onto

the frozen ethylamine. The sample tube was sealed off

124

immediately. Ethylamine was melted and mixed with the metal
film. The solution turned only a very light blue. Then

the solution was frozen again and more metal was heated to
form a fresh new metal surface. Remixing the liquid and

the fresh metal surface failed to turn the solution dark
blue. There was some effect on the solution but not enough
to form a dark blue color. (With a 2 cm O.D. tube, the
light can easily be seen through the solution.)

Second, an ethylamine-methylamine mixture (mole ratio
= 1.74) was studied. The potassium solution in the ethyl-
amine-methylamine mixture also gave a light blue solution
with an absorbance of ~i0.2 in a 0.1 cm cell. Finally,
ammonia (less than 1%) was added to the solution. A much
darker blue solution was obtained upon mixing. This time
the absorbance went up to ~'0.4. However, upon cooling the
sample, a remarkable and uniform color change throughout
the whole solution was observed. The color changed revers-
ibly from dark blue at room temperature to very light blue
at dry-ice iSOpIOpanol bath temperatures. A Tyndall effect
was also observed at low temperature. If the solution was
filtered at the low temperature, the darkening upon warming
was no longer observable. When the solution was allowed to
remain in the dry-ice isopropanol bath for a half-day or
longer, shiny metal particles could be seen by eye. An ESR
study also supported the assumption that metal precipitates
out at low temperatures. At room temperature, the ESR

spectrum gave a four-line pattern indicating the existence

125

of the potassium monomer. As the solution was cooled, the
cavity "Q" changed dramatically and at about 0° a broad line
with much lower intensity was observed. This broad line

was probably the signal from suspended particles of potas-
sium metal. This study reinforced our previous finding

that the solubility of potassium in ethylamine (even with

small amounts of added methylamine and ammonia) is very low.
6.3.3.2. Na—KI-ethylenediamine system

In order to study the reaction of the sodium anion
with added potassium cation in ethylenediamine originally
Observed by DeBacker,”1 purified ethylenediamine was dis-
tilled onto a vacuum-distilled sodium film. A dark blue
solution was formed. The solution was found by both its
Optical absorption spectrum and its ESR spectrum to be
very stable. The addition of recrystallized solid KI did
not result in any observable decomposition of the sodium
ethylenediamine, the Optical Spectrum shows only a single
absorption band at ~'660 nm which has been assigned to the
Na- species. The addition of KI did not greatly alter
this band. However, an observable near-infrared absorption
was found and at the same place as that of the solvated
electron band. NO optical absorption peak around 850 nm
attributable to the K‘ band, was found. An ESR study gave
Similar results. Before the addition of KI, there was only

one very narrow and weak absorption with a peak to peak

width of ~ 0.3 G. With KI in the solution again only a

126

single narrow line was Observed but with about 3 times the
intensity. Since the ESR spectrum had no observable multi-
ple lines, we conclude that neither the Na monomer nor the
K monomer was present at a detectable concentration (of
course, a monomer with rapid exchange could be present).
Therefore, the effect of K+ on a sodium ethylenediamine
solution is apparently a Shift to the right of the equi—

librium

+ + 2K+°e-

11+

Na- + 2K+ :———> Na

+ 2e solv'

However, the sodium anion is much more stable in solu-
tion than the potassium anion, SO that another possible

equilibrium

 

lies far to the left. The present results are in complete
agreement with those of DeBackerJ"1 However, he did not

map the complete spectrum.
6.3.3.3. Cesium in methylamine

Cesium apparently dissolves in methylamine to form
mainly Cs- with the solvated electron and the cesium
monomer as minor constituents. The solubility is high at
room temperatures, but decreases at lower temperatures. At
dry-ice iSOprOpanol temperatures, cesium started to precipi-

tate out as fine particles. Even for the saturated solution

127

at room temperature no bronze phase separation was observed.
When the methylamine was distilled out, cesium metal could
be recovered.

At very high concentration, the solution gave an ESR
pattern with a small and relatively broad signal superimposed
on a small 8-line monomer pattern. The broadening of the
signal is probably due to Spin-spin exchange. The Optical
Spectrum showed a very high intensity of the Cs- band
(~v9400 cm.1 at 0°) along with a small Shoulder for e-Solv'
These results implied that at high concentration, Cs- is
the most important Species. Upon diluting the solution or
lowering the temperature, the ESR spectrum gave a single
narrow line with higher intensity. The Optical spectrum

Showed a large decrease of the CS- band and a Slight in-

crease in the e

solv spectrum. This phenomenon can be

explained by Shifts of both of the following equilibria to

the right:

> C8 + 2 e

 

Cs

 

SOlv

- + -
> S e
< C + solv

 

Cs+-e

 

as the solution is diluted or the temperature lowered. In
addition, the overall solubility decreases markedly as the

temperature is lowered.

VII. STUDIES OF THE METAL DEPENDENCE OF THE SPIN-PAIRING
PHENOMENON IN DILUTE ALKALI METAL AMMONIA

SOLUTIONS BY ESR TECHNIQUES*

7.1. Introduction

The ESR studies of Hutchison and Pastor38 on the para—
magnetic susceptibility of metal-ammonia solutions are the
most reliable results available. Although the results were
Obtained mostly for potassium solutions, a few solutions of
sodium at several concentrations were studied. The results
from both the potassium and sodium spectra ShOW'a decrease
in molar magnetic susceptibility with increasing concentra-
tion. At all concentrations which were examined, (0.04M
to 0.7M for potassium) the molar magnetic susceptibility
increases with increasing temperature. However, the molar
magnetic susceptibilities for sodium solutions are about the
same as those of potassium solutions at 240°K, but are
Slightly higher than those of potassium solutions at 298°K.
A recent study by DeMortier, 35 31,71 indicated that spin-
pairing in sodium-ammonia solutions was much less extensive
than for potassium solutions. If their results are correct,

AAV

*The work was done in collaboration with F. J. Tehan.
128

129

then the dilemma raised by Dye‘6 would vanish, because the
electrochemical data for sodium solutions and the magnetic
susceptibility data for potassium solutions could no longer
be used together to calculate the equilibrium constants.
Also, if DeMortier's statements are correct, we would ex-
pect the cation to play an important role in the process of
Spin-pairing. This implies that the Spin-paired species
might well be a "genuine" alkali anion, M-, or a dimer, M2,
rather than the dielectron, e: , ion pair, M+-e: or loose
aggregate, e-°M+'e-. If this is the case, it is puzzling
that Species such as M_ or M2 would Show the same Opti-
cal absorption spectrum as the solvated electron and that
the Spectrum would be independent of the metal. Strong
evidence is available from the solutions of metals in amines
and ethers to Show that the distinct Optical band of M- is
definitely metal-dependent.31 In order to solve this prob-
lem, the following ESR experiments were designed in order to

determine the effect of the cation on the Spin-pairing

process.
7.2. Experimental

In order to have a known concentration of a metal-
ammonia solution, it is necessary to know separately the
amount of metal and the amount of ammonia used. The puri-
fied metal was used to fill a tube Of 1.0 mm I.D. A
desirable length of the metal was measured and sealed in

the tube. When the experiment was ready, the tube

130

containing a known amount of metal was cut open (in the case
of Rb and C8, the metal was first cooled in liquid nitrogen
to avoid decomposition problems) and placed in the sidearm
of the sample tube.

‘The metal was further purified by vacuum distillation
through the sidearm of the sample tube. Purified ammonia
was first vacuum-distilled into a small thickawalled stor-
age bottle and then vacuum-transferred to the sample tube.
The metal and ammonia were mixed thoroughly inside the
sample tube immersed in a dry-ice iSOprOpanol bath. The
well-mixed solution was then quickly poured into the 1 mm
I.D. quartz ESR tube. The ESR tube was rinsed with this
solution a few times to clean its surface and to assure a
uniform concentration. A small portion Of the solution was
then frozen and sealed in the tube. The solution was kept
frozen until we started to make the measurements. Precau-
tions were taken to avoid changes in the concentration of
the solution but it is admitted that the absolute concentra-
tions might be uncertain to i 5 percent.

2,2~Diphenyl-1-picrylhydrazyl free radical (DPPH,
Aldrich Co., 86% pure) was used as a standard for the meas-
urement of relative Spin concentrations. A weighed amount
of DPPH was dissolved in purified carbon disulfide to give
an approximate spin concentration of 0.1 molar. The DPPH
solution was first sealed Off in a 1 mm I.D. tubing and
then placed inside of a 3 mm I.D. ESR tubing containing a

metal-ammonia solution. This method was soon found to be

131

unsatisfactory. The DPPH standard was frequently broken
when the metal-ammonia solution was frozen. Placing this
DPPH sample tube on the outside of the ESR tube is also not
desirable, because we cannot reproduce the exact geometric
location of different samples relative to the DPPH standard.
This might affect the intensity of the DPPH standard rela-
tive to that of the sample. Finally, a dual sample holder
was constructed as shown in Figure 35. The sample tube was
placed in the center Of a Dewar concentric with the DPPH
standard. The DPPH standard and the metal-ammonia solution
were both placed in a temperature controlled Dewar. The
temperature was controlled by a Varian V-4540 temperature
control unit, and was measured with a thermocouple. The
temperature dependent studies were done without changing
the position Of the sample. In this way, relative changes
with temperature Should be accurately measurable.

The ESR instrument used was a Varian V-4500 X-band
spectrometer. The Signal was recorded as its first deriva-
tive. The intensity of the signal was calculated from the
area under the signal, i.e. the square of the peak-to-peak
width times the peak-to-peak height. When we first made
the area comparison with the DPPH standard, we assumed that
the line Shapes of both the DPPH and the signal were the

same (we will discuss the validity Of this assumption later).

132

 

 

 

 

 

 

 

 

__ [1 \\\0K\y\\7€:\\\\

 

 

 

<7 _ ////////Z? SZ////

 

 

 

L \\i\\\?§\\\;

 

Dewar for ESR study.

Figure 35.

133

7.3. Some Problems Involved in the Measurement

During the experiment, we ran into some difficulties.
Most of them were overcome but some were not, mostly be-
cause of limitations of the instrument. Admittedly, some
problems do affect the results we obtained. However, they
are not serious enough to alter the conclusions which we
draw at the end of the chapter.

The first problem was that of determining the relative
spin concentration. This was overcome by using a concentric
dual sample tube. The intensity Of the signal from the
sample was compared with the intensity of the DPPH standard.
Thus, a relative spin intensity was obtained instead of an
absolute spin intensity. The second problem is that of tun-
ing the instrument. AS we know, metal-ammonia solutions
have very high conductivities. The high conductivity tends
to Spoil the microwave cavity “Q". This was Solved by
using a very small diameter ESR tube which contained much
less solution than does a standard tube. The problem of
variations in sample position within the cavity was over-
come by taping the tube to the instrument and then completing
the entire temperature study before removing the sample.
Although we paid Special attention to the concentration
problem, it still existed to a certain extent. Variations
in diameter of the ESR tubing also affected the data, but
corrections were made for this effect.

Finally, the largest problem of all was that posed by

instrumental limitations. The best instrument available in

134

the Department is an X-band ESR Spectrometer. The extremely
narrow line (~a0.06 G) makes a low power radio frequency ESR
Spectrometer ' much more desirable. Even at X-band, a low-
power instrument would give better results. The use of the
available X-band instrument rather than a low power Spec-
trometer causes saturation of these narrow lines. The ef-
fects of saturation were eliminated as far as possible.
(The saturation problem will be fully discussed later.)
The instrument also caused a small change in the solvated
electron band Shape for sodium, potassium and rubidium solu-
tions when the microwave power was low (probably because of
modulation Sidebands). Unfortunately, we were not able to
correct for this change in the band shape. However, this
effect was identical for all three metals and should cancel
out.
7.4. DPPH Solution Studies and Some Preliminary
Results for Metal-Ammonia Solutions
Dilute solutions of DPPH in carbon disulfide give a
five line pattern. The center peak of the pattern was used
as a relative standard for Spin concentrations of the metal-
ammonia solution. A temperature dependence study of the
0.02M DPPH solution was performed. It was found that there
was no significant change in the DPPH Signal intensity with
temperature as Shown in Figure 36. A saturated solution of
DPPH in carbon disulfide gave a similar result. This means

that even a drastic change in the concentration of DPPH does

135

.SOHDSHOm mama Emo.o mo mmHosum mmm unwocmmmo mnsumummEmB .om wusmHm

S

AooV OHSHonQBOB

om o ONI OJI omI owI OOHI

 

 

_ H IH H fi HI _ om

 

 

 

 

 

 

0:

mo- 3 Kirsuequl
a

136

not affect the conclusion that the peak intensity does not
change with temperature.

In a preliminary study, solutions of sodium and potas-
sium in liquid ammonia were studied by using a Varian E-4
ESR spectrometer. A study of the temperature dependence of
Solutions of the same concentration with no spin standard
showed that the Signal intensity of both potassium and
sodium solutions changed significantly with changing tempera-
ture. Furthermore, the area ratio of the Signal from the
sodium solution to that of the potassium solution stayed
relatively constant over the concentration range studied
(0.02M and 0.04M). Because of these promising preliminary
results, a detailed study of four different metals with a

DPPH spin standard was performed.
7.5. Saturation Problem167

Because of the extreme narrowness of the solvated elec-
tron Signal (~'0.06 G), line broadening, due to saturation,
is unavoidable in an X-band ESR spectrometer which Operates
at normal power levels. However, the problem can be solved
by studying the change in line width and amplitude with
microwave power. The mathematics involved is shown below.

If we assume that the Bloch equations are valid, then
the magnetic resonance absorption Y and its first deriva-

tive Y' for Lorentzian line shapes are found to be

137

 

 

yo
Y = 1 TH-Hffi (17)
+ T...—
EA“;
2
4 2 _ 0'
Y' . (30 (H H0)Ym (18)

 

1 H-H
$13pr 1 +341 ° )2]2
_ {Aflpp

where y; and y$ are the maximum amplitude below satura-
tion, AH; is the half width at half height for the absorp-
tion curv: and AH is the width between points having
maximum SlOpe; H and Ho are the magnetic field and the
resonance field respectively. If one writes these equations
as a function of Spin lattice and Spin-Spin relaxation times

T1 and T2 and includes the effect of the microwave ampli-

tude H1 then,

H1 Y;

 

I
1 + (H-Ho)2Y2T22 + Z’H1272T1T2

 

and
16 (H-HO)YT2H1Yg
y' 2 . H—V 1V1V_V ll
33;2 [1 + (H-Ho)2v2Tz2 + EHIZIZTITzl (20)

where y is the gyromagnetic ratio given by

y e gg/h = 0.87934 x 107 ngs/G

Let us define the saturation factor S as

138

 

 

1
1 + Z 1112szsz ~
then,
SHIYmO
Y = (22)
1 + S(H-Ho)2y2T22
and

16(H-Ho)yT282H1yg
Y. = 3 *2 ' (23)
3/3 [1 + s(H-Ho)2y2T22]

 

If the signal is unsaturated, S = 1. The AH1 and AH

2 PP
are given in terms of their unsaturated values, AH? and
. E
Aflgp respectively, by
0 -1
AH1 = [AHil S é (24)
E E
and
0 -1/2
AH = AH S . 25
pp [ pp] ( )

The amplitudes ym and y$ can be obtained by letting

H H0 and H - H0 = i 1/2 Apr reSpectively. Then we

have the forms

(Ym/Hi)

[Y3] S (26)

and

(yg/HI) Iyg'I s39 . (27)

From the definition Of S and the above equations, we can

substitute and obtain the following:

139

F 122
1 + 27 H1 T1T2

[AHl/H1_I:OAH_]2
2

lim 2
pp/H ~40 mpp]

UJIH
II
A

(28)

[$i2,0 (ym/HI>/<ym/HI>I

 

I;:3,0 (yg/H1>/(yg/HI>I%9 .

The microwave magnetic field H1 at the sample may be
computed from the microwave power, PW, incident on the

resonant cavity and has the form
H1 = KP (29)

where H1 is in gauSS and Pw is in watts. In a typical
case, K is of the order of unity. Therefore, if we plot

1/ he

y' versus P 2 and AH (Gauss) versus P , we find y'
n1-—————- w pp -————- w m
gives a linear relationship with P39 (or H1) below satura-
tion. y; become proportional to P.1 (or Hi) when the
resonance line is strongly saturated. The peak-to-peak
linedwidth, AH , approaches a minimum as P approaches

PP
. . 1/2
zero and is proportional to P

as the line becomes
strongly saturated. Hence if a sample is only Slightly
saturated, i.e. y; is still in linear realtion with Pgé,

the area under an unsaturated signal can be obtained by

multiplying the square of the unsaturated line width by the

“MAL u

140

measured amplitude. The unsaturated line width can be ob-
tained easily from a plot of Apr versus P39. However,
a more accurate result can be Obtained with the help of the
following equation:
AH

PP )2 z

( .
lim
H1->0 Apr

1.22
1 + 4V H1 T1T2 o (30)

We know, in the microwave power range studied, that the
DPPH sample was not saturated at all. Therefore, the
amplitude of the DPPH signal is proportional to the inci-

dent microwave power. Thus,

'(DPPH)

H1 = a Ym (31)

where a is a proportionality constant. Substituting

y$(DPPH) for H1, we obtain
2 _ lim 2 lim 2 1 2T1T2 I(DPPH)
(wisp) —(H1_,o pp) + (H140 pp) (3v )(aym
(32)
The plot of (AH )2 versus (y'(DPPH))2 should give a

PP m

straight line with an intercept equal to the unsaturated
lim

line width, i.e. (H-90 Apr)”-

7.6. Results and Discussion

We measured the ESR signal amplitude (yé) and peak to
peak width (Apr) of solutions Of Na, K, Rb, and CS in
liquid ammonia at four different concentrations (0.02M,

0.04M, 0.06M, and 0.10M) as a function of temperature from

141

-60° to 0°. We found that the e signal in Cs solu-

SOlv
tions had a much larger linewidth (AHpp -0.12 G) than the

e signal for Na, K, and Rb solutions (Apr~v0.06 G).

solv
Because of the linewidth difference, the solvated electron
Signal in Cs solutions was found to be much less saturated.
Plots Of the solvated electron signal amplitude (y$) and

the peak-to—peak width for Cs solutions versus the square
root of the microwave power are shown in Figure 37 and
Figure 38 respectively. Furthermore, the line Shape of

the signal of the solvated electron in Cs solutions does

not change with changing microwave power. The data Obtained
from Cs solutions are quantitatively more reliable than the
data Obtained from the other solutions. On the same plots
(Figure 37 and Figure 38), the signal for the DPPH solu-
tion is also shown. It demonstrates that in the microwave
power range studied, no saturation of the DPPH signal
occurred. Therefore, no correction was made in calculating
the intensity of the DPPH absorption. In order to obtain
the unsaturated peak-to-peak width, a plot of AH;p versus
(yé‘DPPH)2 was made for all of the solutions and a straight
line was Obtained in all cases. Figure 39 Shows an example
of such a plot. The intercept appears to be the square of

the unsaturated peak-to-peak width of the signal. The

intensity, (AA), of the signal was Obtained by the equation

2 I
AA 3 ° ' 33
c Apr ym ( )

where C is a constant equal to BIT/3v2 for a

142

 

 

 

 

 

4 _.
DPPH
. 3 ~P___x_x_x_x__x_x—X———-—-x——"X"—X
2 -
Cs
In 0
o///
.3" /
o
o
0 O
a 2L
1p 0/,o
4 1 /’
. /°’do
.l—
0 l .2 .3 .4 .5 .6 .7 .8
Figure 37. Plot of peak-to-peak width versus the

square root of the microwave power for
the solvated electron signal in cesium
solutions and for the DPPH radical signal.

143

 

a
20*- DPPH

 

 

 

J I l 1

.02 ./04 .06
2 .
P! ( W I/2)

Figure 38. Plot of peak-to-peak amplitude versus
the square root of the microwave power
for the solvated electron signal in
cesium solutions and for the DPPH radical
signal.

144

 

 

 

 

 

 

l2r-
Cs O.IM
IO--~
N 8 —
9.
x
a _-
2
o
eI2;6-
I:
4
" 0
4*— o
/
0°
/
7."
2 V
l l Mini .1
O I 2 3
(YnfioppHI’ (cm’x I63)
Figure 39. Plot Of peak-to-peak width of the sol-

vated electron signal in 0.1M cesium
solution versus the square of the
amplitude of the DPPH radical signal.

145

Lorentzian line. The plot of the logarithm of the area
ratio for the solvated electron signal of Cs solutions to
that of the DPPH radical is plotted against (1/T) is Shown
in Figure 40. The Slopes of the lines for different con-
centrations are relatively close to each other. This
indicates that the relative change in spin concentration
with temperature is essentially concentration independent.
In the cases of Na, K, and Rb, the Situation becomes
more complicated because of the change in band shape with
microwave power (Figures 41 and 42). The cause for this
shape change is not certain but it is probably caused by
the presence of modulation sidebands. All three solutions
changed in the same manner. Despite this shape change, the
data for Na, K, and Rb solutions were treated in the same
way as those for CS solutions and are shown in Tables 7 to
10. A similar plot of the logarithm of the area ratio of
the solvated electron signal for Na, K, and Rb to the Sig-
nal of the DPPH radical versus (1/T) is Shown in Figure 43.
The straight lines with nearly the same slopes indicate
that the relative signal intensities change Similarly with
temperature for the three different metals. If we take the
area ratio of sodium, rubidium, and cesium to that of potas-
sium at the same concentration and plot these values versus
temperature (Figure 44), we find the points remain rela-
tively constant throughout the temperature range. (The
complete ratios for each metal at the different concentra—

tions are listed in Table 11.)

Log(Area ratio)

I
m

146

 

 

 

 

 

.Al.__ 1
OJI 0J6
l/T ( °K'l x 103 )
Figure 40. Plot of logarithm of the area ratio of the

solvated electron absorption in cesium
solution to that of the DPPH radical central
absorption versus (1/T).

.Hm3om m>m3OHOHE
mchdmso suH3 mmmnm pawn Mmm SOHDUOHO cmum>HOm mnu SH mmmdmno .Hv musmHm

e. N 0 «II VI

 

147

H A _ A _ _ _ _ . _

 

 

 

 

 

148

 

 

 

 

¢
100—
¢ ¢
¢ ¢ Cs
50v
OL—
" fl:
69 100—
g B
a
g ,3Rb
50k ONO
a
3K 2
I I 1
16 20 24 28
PHD)

Figure 42. Plot Of the percent Lorentzian charac-
ter for the solvated electron signal in
Na, K, Rb, and Cs-ammonia solution
versus microwave power.

Log ( area ration )

149

 

 

 

 

 

- \O X
A 0.10M
.0 I- X \0
‘A~\\\\\~ () Na
0
X
14 P A~\\EB~\\\“‘
as
.8 _
0.06M
?" X
2 — g‘\\\\\\\\\‘
P
\: 0.04M
X
.9 I- ‘77777‘0
A A x
” O
o
A~\\\\\\\\\
.5‘¥; O
]_ P 0.02M
z x
_ x
0 O X
S - A
0
L l I .l
3.5 1+.O 14.5 5.0
l/T ( °K'1x 103 )
Figure 43. Plot of the logarithm of the area ratio

against (l/T) for Na, K, and Rb solutions.

150

 

.mnsumummamu msmuw> M mo was» 0» mo cam am

 

 

 

 

 

.m2 m0 HmcmHm COHHUOHO omum>Hom 0:» mo OHumu mmHm on» no HOHm .v¢ w
A x. V .H.
owm 0mm 04m 0mm oom
H _ A -
m
0 pm
O m2
0
_U
0
0 0 mo

 

 

HsmHm

m.o

o.H

H/w

m.H

151

 

 

 

Table 7. Corrected area ratio of the solvated electron ESR
Signal in 0.02M Na, K, Rb, and Cs-ammonia solutions
to that of the center DPPH radical peak.

T(°C) AH C Area Ratio 0 Corrected
pp Area Ratio

Na (0.02M)
-59 0.0840 0.0024 0.00606 0.00022 0.00292
-48 0.0692 0.0021 0.00885 0.00031 0.00427
-27 0.0486 0.0026 0.00888 0.00043 0.00428
- 8 0.0405 0.0072 0.01680 0.0054 0.00810
K (0.02M)
—59 0.0754 0.0054 0.00820 0.00093 0.00374
-48 0.0584 0.0038 0.0124 0.0013 0.00566
-27 0.0340 0.0036 0.0146 0.0014 0.00667
- 8 0.0157 0.0067 0.0152 0.00046 0.00694
Rb
-59 0.0894 0.0012 0.00410 0.00052 0.00285
-48 0.0756 0.0013 0.00512 0.00075 0.00356
~27 0.0564 0.0052 0.00629 0.00068 0.00437
- 8 0.0492 0.0030 0.00896 0.0014 0.00622
CS
-59 0.109 0.0047 0.00522 0.00064 0.00464
-48 0.101 0.0021 0.00766 0.00048 0.00682
-27 0.0776 0.0073 0.00940 0.00084 0.00836
- 8 0.0719 0.0059 0.00878 0.0018 0.00781

 

152

 

 

 

Table 8. Corrected area ratio of the solvated electron ESR
Si al in 0.04M Na, K, Rb, and Cs-ammonia solu-
tions to that of the center DPPH radical peak.

T(°C) AH 0 Area Ratio 0 Corrected
pp Area Ratio

Na (0.04M)
-59 0.0983 0.0050 0.00746 0.0013 0.00717
-48 0.0922 0.0049 0.00451 0.0023 0.00434
-27 0.0933 0.0062 0.0126 0.0049 0.0121
- 8 0.0670 0.0058 0.00669 0.00044 0.00643
K
-59 0.0990 0.0052 0.00649 0.00075 0.00919
-27 0.0723 0.011 0.0101 0.0017 0.0143
- 8 0.0470 0.0070 0.0127 0.0041 0.0180
Rb
-59 0.0916 0.0048 0.00586 0.0012 0.00542
-48 0.0600 0.014 0.0104 0.0021 0.00961
-27 0.0632 0.0048 0.0111 0.0023 0.0103
- 8 0.0678 0.0085 0.0166 0.0054 0.0153
CS
—59 0.137 0.0051 0.00516 0.00031 0.0105
-48 0.127 0.0029 0.00642 0.00072 0.0131
-27 0.122 0.0047 0.00878 0.00079 0.0179
- 8 0.130 0.0048 0.0428 0.0063 0.0873

 

153

 

 

 

Table 9. Corrected area ratio Of the solvated electron ESR
Signal in 0.06M Na, K, Rb, and Cs-ammonia solu-
tions to that of the center DPPH radical peak.

0 . Corrected

T( C) Apr 0 Area Ratio 0 Area Ratio

Na (0.06M)
-59 0.0831 0.0025 0.00541 0.00027 0.00564
-48 0.0706 0.0032 0.00689 0.00081 0.00718
-27 0.0488 0.0039 0.00837 0.0010 0.00872
- 8 0.0395 0.0037 0.00917 0.00067 0.00955
K
-59 0.0737 0.0069 0.00743 0.00056 0.00839
-48 0.0666 0.0058 0.00917 0.00067 0.0103
-27 0.0436 0.0088 0.110 0.0050 0.0124
Rb
-59 0.0927 0.0030 0.00653 0.000361 0.00653
-48 0.0710 0.010 0.00926 0.00102 0.00926
-27 0.0456 0.0042 0.0128 0.00152 0.0128
- 8 0.0221 0.0075 0.0179 0.0053 0.0179
Cs
-59 0.139 0.0012 0.0102 0.00068 0.0115
-48 0.129 0.0027 0.0160 0.0012 0.0181
-27 0.109 0.0027 0.0258 0.0034 0.0291
- 8 0.120 0.0028 0.0274 0.00362 0.0309

 

154

 

 

 

Table 10. Corrected area ratio of he solvated electron ESR
Signal in 0.10M Na, K, Rb, and CS-ammonia solu-
tions to that of the center DPPH radical peak.

0 . Corrected

T( C) Apr 0 Area Ratio 0 Area Ratio

Na (0.1M)
-59 0.104 0.0036 0.0157 0.0040 0.0130
-48 0.0935 0.0047 0.0136 0.0015 0.0112
-27 0.0924 0.010 0.0194 0.0067 0.0160
- 8 0.0724 0.0094 0.0320 0.012 0.0264
K
-59 0.116 0.0044 0.0113 0.0027 0.00933
-48 0.0959 0.0075 0.00887 0.0010 0.00732
-27 0.0590 0.016 0.0345 0.0036 0.0285
- 8 0.0598 0.0067 0.0189 0.0039 0.0156
Rb
-59 0.125 0.0059 0.0142 0.0032 0.00685
-48 0.100 0.0055 0.0209 0.0044 0.0101
-40 0.0840 0.0063 0.0247 0.0036 0.0119
-28 0.107 0.011 0.0276 0.011 0.0133
- 8 0.0952 0.011 0.401 0.0092 0.0193
CS
-59 0.158 0.0036 0.0130 0.0010 0.0111
-48 0.154 0.0023 0.0232 0.0037 0.0199
-28 0.228 0.0063 0.0752 0.0081 0.0645
- 8 0.149 0.0042 0.0386 0.0037 0.0331

 

155

Table 11. Area ratio of the solvated electron ESR signal
in liquid ammonia for Na, Rb, and Cs to that of
K at the same concentrations and temperatures.

 

 

Area Ratio

 

 

o o g
Concentration T( C) Na/K Rb/K CS7K
0.02M -59 0.7807 0.7620 1 .241
-48 0.7544 0.6290 1.205
-27 0.6417 0.6552 1.253
- 8 1.167 0.8960 1.125
0.04M -59 0.7801 0.590 1.14
-43 -_- _-_ _-_
-27 0.846 0.720_ 1.25
- 8 0.357 0.850 4.85
0.06M -59 --- --- ---
-48 0.672 0.778 1.370
-27 0.697 0.899 1.757
- 8 0.703 1.03 2.35
-48 1.530 1.626 2.719
-27 0.514 0.467 2.263
- 8 1.692 1.237 2.122

 

156

Comparing our potassium data with the data of Hutchi-
son and Pastor, we find that the temperature dependence of
the ratio of the integrated intensity for potassium solu-
tions to that of DPPH agrees well with their data as does
the concentration dependence.’ The comparison requires
that we normalize our relative data to their absolute spin
data at one temperature and concentration. Based on the
assumption that the potassium data of Hutchison and Pastor38
are valid, we can conclude that the effect of the cation on
the Spin-pairing process is not large. This result is in
agreement with the conclusions of Hutchison and Pastor for
potassium and sodium but is at variance with the extrapola-
tion of DeMortier, 33 31.71

Finally, if we place the sodium and potassium data
from Hutchison and Pastor and the sodium data from DeMortier,
leg 31., together with our results for sodium and potassium
solutmms on one graph, Figure 45, we find that the data
need not strongly disagree with each other if an appropri-
ate extrapolation is made. It is possible that a plateau
exists in the concentration region of 0.01M to 0.1M. Cer-
tainly there are not enough data to draw any firm conclu-
sions about the validity of the plateau. It would be
profitable to make a careful study with a radio frequency
ESR spectrometer or a low-power X-band Spectrometer over

this concentration region to further study this point.

157

.0500 02 H5O . “mumo mz m.uoummm can GOmHsousm x “sumo x
m.uoummm cam somHnousm 4 HcOHuwHommnuxm m.u0HuHOZOQ.III!I
“mumo ESHoOm m.H0HuHOzOQ O «msoHusHOm chOEEMIx cam 02
How coHumHucoocoo umchmm :oHumuucoocoo :Hmm was no uOHm

:30

79 «.2 mm:

.mw musmHm

 

H _ H

 

(W)"N

 

 

 

158

7.7. Preliminary Studies of the Spin-pairing
Process in MetaleAmine Solutions

7.7.1. Cesium—ethylenediamine solutions

A cesium-ethylenediamine solution was prepared in the
usual manner. The solution was in the concentration range
which contains solvated electrons but no cesium anions or
cesium monomers. The Optical Spectrum of the solution
showed only a near-infrared absorption at 1280 nm. The ESR
spectrum gave a single absorption line with a peak-to-peak
width of -1.75 Gauss. The optical spectrum was obtained
again after the experiment was completed and no absorbance
change was observed. This implies that the concentration
of the solution did not change during the priod of the
experiment. A calculation based on the absorbance of the
Optical spectrum with an extinction coefficient of 1.8 x 104
(as determined by M. DeBacker and J. L. Dye83) revealed the
concentration to be 3.4 x 10'"4 moles liter"1 for this
cesium-ethylenediamine solution. The ESR signal of the
cesium-ethylenediamine solution was compared with the sig-
nal intensity of the strong pitch sample (Varian). The
spin concentration of the strong pitch is approximately
1.2 x 10.4 moles liter-1. The area ratio for the sample
to that of strong pitch was found to be about one-to-one.
Therefore, the spin concentration as Obtained from the ESR
study was ~10.4 moles liter-1. The difference between

the calculated concentration from the optical study and the

159

spin concentration study indicates that a Spin-pairing
process is probably also involved in cesium—ethylenediamine
solutions. Perhaps more important is that the spin con-
centration is comparable to the total concentration so

that we can conclude that Spin-pairing is not complete at

these concentrations.
7.7.2. Sodium-2,2,2-crypt-ethylamine solutions

Sodium and 2,2,2-crypt were both dissolved in ethyl-
amine to form a dark blue solution. At dry ice-isopropanol
bath temperature, a gold-colored solid was precipitated out
of the solution. (A detailed study of the gold-colored
solid is discussed in the following chapter.) The solution
was passed through a frit at this temperature to filter out
all of the solid precipitate. A clear blue solution was
Obtained after the filtration. An ESR study of this solu-
tion gave a strong single absorption with a peak-to-peak
width of ~o0.28 G at 0°. A temperature dependent study on
the solution was performed. The data are listed in Table
12. AS the temperature decreased, the linewidth of the Sig-
nal increased (0.36 G at -57°) but the intensity of the sig-
nal decreased. The intensity at 20° is about 6 times
greater than the intensity at -57°. The change in signal
intensity with changing temperature indicates that a spin-
pairing process is involved in sodium-ethylamine solutions
even in the presence of the cation complexing agent,

2,2,2—crypt.

I160

Table 12. Temperature-dependence of the ESR signal for a
solution of Na-2,2,2-crypt in ethylamine.

 

 

I AH y
0
T( C) Apr (G) Ymm (cm)_ (G pcfim
-56.5 0.355 10.77 1.357
-38.1 0.292 36.44 3.107
-18.4 0.268 74.00 5.315

+ 1.1 0.275 102.50 7.752

 

161

These studies are only very preliminary. Detailed
studies with a known concentration of the Spin standard
are definitely needed to give quantitative results about
the spin-pairing process in metal solutions in amines

and ethers.

VIII. SOLID STATE STUDIES

8.1. Introduction

As mentioned in the second chapter, solid compounds of
the metal-ammonia type have been found for lithium, the
alkaline earth metals (calcium, strontium and barium),128
and some lanthanide metals (europium and ytterbium). How—
ever, only the lithium compound haS been studied in detail.
When the cation complexing agent -- crown ether was first
used in this laboratory to form metal amine and ether solu-
tions,32 the prOSpect of forming solid compounds between a
metal and its complexing agent was considered. A number of
studies of potassium and barium complexes with crown ether
have been made. Blue solids were formed in all the cases.
Recently, 2,2,2-crypt has been synthesized in our laboratory.
The high cation complexing ability of this compound gave uS
the hope Of developing a completely new field of research
not only in solution chemistry but also in solid-state
chemistry. An extensive study of the interaction of alkali
metals and barium with 2,2,2-crypt has been made with
Special attention paid to the sodium system. A solid com—
pound forms between sodium and 2,2,2-crypt which is bright
gold in color and which has a remarkable stability toward

162

163

thermal decomposition. The solid is indefinitely stable at
-20° and appears to be stable for long periods at room
temperatures. This provided us storage and handling conven-
ience which was lacking in the metal-crown ether solid
complexes. In this chapter, studies of the composition,
solubility, stability, and conductivity Of the solid com-
pound of sodium and 2,2,2-crypt are reported along with some

preliminary studies of solids which contain other metals.
8.2. Solid Preparation
8.2.1. Ba(NH3)6 and crown compounds

Barium metal has a very high melting point and reacts
with quartz glass at high temperature. This makes it very
difficult to purify by high vacuum distillation. Instead, ‘
relatively pure barium was obtained by applying the technique
described below. Inside an evacuable dry box under an
atomosphere of helium pressure, commercially available
barium (Fairmount Chem. Co. 98% purity) was thoroughly
washed with pure hexane. Then the oxidized surface layer
was cut Off with a knife and the center portion was cut into
small pieces. The freshly cut barium was placed into a pair
of sealed-off 5 mm Fischer-Porter Solv-Seal joints and
weighed outside the dry box. The known amount of barium was

then transferred into a sample tube inside the dry box and

was ready for use. The barium prepared in this way still

had a Shiny surface without much oxide or nitride visible.

164

The sample tube containing a known amount of barium
was connected to the high vacuum line and pumped down to
< 10"5 torr. Purified ammonia was then distilled into the
tube. The ammonia was mixed well with barium metal. The
resulting solution formed two layers, one dark blue at top
and the other bnnnze in color at the bottom. After all the
barium had dissolved, the ammonia was distilled out at
I~ -78°, and a bronze-colored solid was obtained. As the
temperature was raised Slowly to room temperature, the color
remained unchanged. Further increase of temperature to
about 60° caused some of the bronze color to turn gray.
However, upon cooling the bronze color reappeared. If the
sample was pumped at room temperature, the change in color
from bronze to gray could not be reversed upon cooling.

This indicates that the reaction

 

> Ba(NH§)6

<—_

Ba ‘1' 6NH3

is reversible as long as some of the ammonia still remained
in the tube.

The addition Of 2,2,2-crypt to the barium ammonia solu-
tion did not affect the separation of the bronze layer from
the dark blue layer. In the presence of ammonia the solid
was still bronze in color. A dark blue solid, presumably
consisting of barium and 2,2,2-crypt,was obtained, when the
ammonia was completely pumped out. The dark blue solid was

very stable even at ~a60°.

165

The potassium-crown solid was prepared in the same way
as described above for the barium case. However, ethylamine
was used as the solvent in this case. Only a dark blue
colored solid was obtained. An ESR study of the blue solid
Showed a single absorption line at —20° (or higher) with a
peak to peak width of 8.3 G. The Shape of the line was
Lorentzian and the intensity of the Signal increased with
increasing temperature. As the temperature decreased to
-80° (or lower) the signal became a four line pattern as
shown in Figure 46. These four lines are not equally Spaced
and, therefore, cannot be a pattern of hyperfine Splitting.
Also the separation between points of maximum lepe (line
width) changed in value when the temperature was changed.
This fact eliminates an anisotropic 9 value as the source
of the splitting. Hence the most reasonable eXplanation is
the presence of two different paramagnetic centers. Two
possible species which might account for the two absorptions
are finely-divided potassium metal particles and trapped
electrons. Of course, two different environments of trapped

electrons could give the same result.
8.2.2. 2,2,2-Crypt compounds

Sodium metal-2,2,2-crypt solids were made by similar
procedures to those used in studies Of Ba(NH3)3 and metal-
crown compounds. In this case, ethylamine was used instead
of ammonia as a solvent. First, all of the 2,2,2-crypt was

dissolved in ethylamine. Then the solution was poured onto

166

0°C
o"./ ',_,a—.'_.\ \“"‘-. __
"‘ -‘.——'~—~
310°C

-100°C

~120°C

Figure 46. ESR Spectra of the potassium-dicyclohexyl—IS—
crown-6 solid.

167

an excess of metal which had been heated to produce a film.
Although a dark blue colored solution formed immediately,
the metal did not dissolve immediately to form a stoichio-
metric solution Of M+-crypt-M-. The dark blue solution
usually was passed through a frit into a sidearm and cooled
to ~v-78°. (The purpose of the frit is to filter off any
suspended metal particles.) After the solution had remained
in the dry-ice isopropanol bath for five to ten minutes, a
precipitate was formed and the solution became relatively
light blue in color. If we passed the light blue solution
back onto the sodium metal film again and warmed the solu—
tion to room temperature, more metal would dissolve in the
solution and the solution became dark blue once again. By
passing the solution through the frit and repeating the whole
process 6 to 7 times the stoichiometric amount of metal
could be dissolved. The solid obtained by precipitation or
by evaporation of solvent was either dark blue in color (in
the cases of K, Cs, and Ba), or bright gold in color (in
the cases Of Na and Rb). In the case of rubidium, the gold-
colored solid was very unstable. It first turned blue and
then gray. Only the gold-colored sodium-2,2,2-crypt com-
pound wasstudied in detail. Studies of the ESR Spectra were
performed on potassium and barium-2,2,2-crypt compounds.

In order to obtain good data for these studies, pure
samples which showed only the gold color were needed. A
washing procedure for the solid product was develOped for

this purpose. Diethyl ether and hexane were used as washing

168

solvents. Each washing was done at room temperature and at
least six washings were made. After careful washing, a
Shiny gold-colored material without a trace of blue color
can be Obtained.

In some instances, single crystals rather than powder
samples were desired for study. Two different methods of
growing Single crystals were tried, (a) slowly lowering the
temperature of a saturated solution, and (b) Slowly evapor-
ating the solvent. The first procedure was found to be the
most effective. Single crystals with diameters in the milli-

meter range can be obtained by this method.
8.3. Na-2,2,2-Crypt Compound
8.3.1. Physical appearance and reactivity

As mentioned above, the compound is gold in color and
the color also changes with temperature. AS the temperature
is increased from liquid nitrogen temperatures to room tem-
perature, the color of the compound changes from bright gold
to bronze-gold. The compound is sensitive to air and re-
acts with oxygen to form a white-colored solid @resumably
Nazo and crypt). The compound reacts immediately with
Imater to form a gas and a basic solution. The compound is
stable at room temperature under vacuum and is soft enough
to permit extrusion though with much more difficulty than

is the case with sodium metal.

169

8.3.2. Composition
8.3.2.1. Sodium analysis by flame emission

Because of the composition of the solution from which
the gold-colored solid was precipitated, its composition
was thoujt to be two sodium atoms for each 2,2,2-crypt.
However, proof of the composition had to be obtained. The
easiest way to determine the sodium content of a small
sample is flame emission. The method not only provided the
necessary proof for the presence of sodium in the sample
but also can be used to determine the amount of sodium. It
was assumed (and later proved) that sodium and 2,2,2-crypt
were the only constituents in the compounds.

The solid for the study was prepared and washed in an
NMR tube as mentioned earlier. The sample was weighed and
reacted with ethanol-water mixture (10% water by volume).
The approximate concentration of sodium in the solution
was calculated by the presumed formula Na2C18H38N206. The
flame emission spectrophotometer (Heath) was calibrated
with standard solutions. A good linear relationship be-
tween the detector currents and the concentrations of the
standard solutions in the region of interest was Obtained
(Figure 47). The results are shown below:

The weight of sample (gold-colored solid) 8 41.69 mg

The expected weight of sodium in the sample
as calculated by the formula Na2C18H36N206- 4.55 mg

The weight of sodium as determined by the
flame emission technique = 3.73 mg.

 

1.0

Detector reading

170

 

 

 

 

 

.- A
b 'u
H /A
I
h- /o
A
/“
J l 1 l l L
10 50 50

Na concentration (mm)

Figure 47. Flame emission study of the sodium con-
tent in the sodium-cryptate solid.

 

171

With the assumptions made earlier, the ratio of sodium to
2,2,2-crypt obtained by this method was found to be 1.6 to 1.
One of the Obvious problems was the error introduced by
weighing ~I40 mg Of sample in a 2 gram NMR tube. The loss
of small glass fragments when we cut the NMR tube might be
another reason for low results. Finally, this sample might
have had an excess of 2,2,2-crypt if the maximum amount of
sodium had not been dissolved. Because of the other methods
of analysis available, this analysis technique was not

repeated.
8.3.2.2. Existence of 2,2,2-crypt -- NMR study

The existence of 2,2,2-crypt in the solid compound was
proved with the help of an NMR spectrometer. Because the
compound to be studied was a solid, the study was carried
out in an indirect way.

First, a carefully washed gold solid sample was dis-
solved in deuterated ammonia. The NMR spectrum Of this
dark blue solution, clearly gave the pattern of the 2,2,2-
crypt -- an overlapped singlet and triplet downfield and a
triplet upfield. The integrated Signal gave a ratio of 2
to 1 for the combined triplet and singlet to the other
triplet. (These peaks were paramagnetically shifted up-
field from the Na+-cryptate solution without metal present
as mentioned in Chapter V). Second, the compound was
thermally decomposed in 32222_and then Opened to the air.

Deuterated chloroform was used as a solvent and TMS was

172

used as an internal standard. The NMR Spectrum so Obtained
was exactly the same, both in pattern and location, as that
obtained with a solution of a sodium salt in the presence
of 2,2,2-crypt. These results leave no doubt that 2,2,2-
crypt is a major constituent of the solid gold-colored
compound.
8.3.2.3. The ratio of sodium and 2,2,2-crypt -- titration
study

In order to have some information about the ratio of
sodium to 2,2,2-crypt, a titration of the aqueous solution
formed by reaction of the compound with water was carried
out. Aqueous hydrochloric acid was used as the titrant.
The pure compound was prepared and washed in an NMR tube
and sealed off under vacuum. The sample tube was weighed
accurately and then opened in a dry bag under a dry nitrogen
atmosphere. The compound was then allowed to react with
conductance water. The empty NMR sample tube was weighed
again SO that the difference in the two weights gave us the
‘weight of the compound. If there were only one sodium atom
for each molecule of crypt, then the aqueous solution would
require three equivalents of acid per mole of compound.
(There are two tertiary amine sites per crypt.) However,
if there were two sodium atoms per molecule, then four equi-
valents of acid would be required. The amount of solid
used in this study was 44.89 mg. The result from the titra-

tion with an assumption of four equivalents per mole was

173

44.72 mg. The titration curve is shown in Figure 48, and

the data are given in Table 13.
8.3.2.4. Complete elementary analysis

Two carefully washed samples were delivered to Spang
micro-analytical laboratory Of Ann Arbor for C, H, N, and
Na analysis. The results shown in Table 14 are in excellent
agreement with the calculated data based upon the presence

of two sodium atoms per 2,2,2-crypt molecule.

8.3.3. Solubility

It is important to know the solubility of the compound
in a variety of solvents in order to find a suitable solvent
for single crystal growing, crystal washing and the forma-
tion of different solids. A solubility test on some common
solvents has been made and the results are listed in Table
15. A comparison Of the solubility of the gold solid and
of sodium metal in the presence of 2,2,2-crypt in the same
solvent is also given in the Table. No surprising results
‘were found since the two results Show the same trends. It
should be noted that the other alkali metals generally

Show higher solubilities in the presence of 2,2,2-crypt.

8.3.4. Reducing power study

The reducing power of the compound was determined by
Iiecomposing the solid with degassed conductance water and

‘then collecting the hydrogen evolved by the reaction.

174

 

 

.UHHOm wumummHOIEsHoOm Ummomaoomo mnu mo m>uso aoHumuuHu UHod
cocoa Hum as
O.H ®.O W.O .2.0 N.O

.wv musmHm

 

I'I
OI.

/

O

O/

 

 

 

Hd

175

Table 13. Titration data for the Na-cryptate solid.

 

 

 

pH HClmAdded pH HClmAdded
11.05 0 6.83 0.60
10.99 0.09 6.35 0.65
11.01 0.20 3.58 0.71
10.55 0.30 2.67 0.80
10.00 0.34 2.45 0.88
8.00 0.40 2.31 0.97
7.27 0.50 2.25 1.09

 

Table 14. Complete elementary analysis of Na-cryptate

176

 

 

 

 

solid.
Sample NO . %c %H %N %Na
1 51.04 8.56 6.62 11.17
2 51.21 8.68 6.77 10.92
Average 51.13 8.62 6.70 11.05
Calculated Value
(Based 0“ formUIa 51.18 8.53 6.64 10.90

Na2c18H36N206)

 

177

Table 15. Solubility test of Na-cryptate solid on some
common solvents.

 

 

 

Solvent Conditions Solubility
NH3 T < 0°C Very soluble
Methylamine T > 0°C Very soluble
T -78°C for a day Some gold-colored
precipitation
Ethylamine T > 0°C Very soluble
T < 0°C Gold-colored
precipitate
THF T > 0°C Very Soluble
T < 0°C Gold-colored
precipitate
Diethyl ether Room temperature Slightly soluble
(Forms a very light
blue solution)
Diethylamine Room temperature Slightly soluble
(Forms a very light
blue solution)
EfPrOpyl ether Room temperature Insoluble
Di-gfpropylamine Room temperature Insoluble
Hexane Room temperature Insoluble

 

178

Na2C18H36N206 + H20 —> H2) + 211a+ + (C18H36N206) + 20H"

The apparatus consisted of three major parts -- a
decomposition vessel, a liquid nitrogen trap and a measur-
ing burette with a movable mercury leveling bulb as shown
in Figure 49. The volume of gas in the measuring burette
of total volume 1 ml could be estimated to within $0.002 ml.

The gold-colored solid, prepared and washed as de-
scribed above, was sealed off in either an NMR tube or in a
breakable bulb.168 In the case of breakable bulb, the bulb
was weighed first and then connected to a vacuum line with
heat-shrink tubing. The solid was poured into the bulb and
sealed off under vacuum. The bulb with the sample and the
stem were then weighed again. The weight of the sample
was determined by difference and the necessary buoyancy
correction was applied. With this method an accurate weight
of pure sample (to i 0.1 mg) could be Obtained. The bulb
containing a known amount of sample was then placed in the
sidearm of the decomposition vessel. The whole vessel was
evacuated and the conductance water in the bottle was de-
gassed with many freeze-pump—thaw cycles until no detectable
gas remained.

The entire system was pumped down to a pressure of
< 10-5 torr and the amount of gas remaining in the system
‘was measured and found to be negligible (0.05 ml at 43.5
mm Hg).

The bulb was then broken and the contents mixed with

conductance water. The hydrogen gas formed in this way was

. ‘ It.
i ' ' _; i.
a!“ -‘g It -'

179

.hosum

Hmzom mcHosomH 030 How msumummm<

 

 

 

 

 

 

 

.ov musmHm

 

 

 

 

 

180

pumped into the burette with the aid of a mercury leveling
bulb (Figure 49). The volume and the pressure were then
measured. The pumping process was continued until there
was no change in the product of the measured volume and
pressure. The reducing power was calculated from the amount
of hydrogen collected by assuming ideal gas behavior. In
the first experiments, alcohol and samples contained in NMR
(tubes were used. However, this procedure yielded only 0.5
to 0.8 mole of H2 per mole of compound. This may have been
caused by impurities in the alcohol and by errors in weigh-
ing. With the bulb technique and degassed conductance
water, 0.95 and 0.94 i 0.03 mole of H2 per mole of compound
were formed in two separate experiments. The theoretical
value of 1.0 is expected on the basis of two available elec-
trons per molecule. The posSible use of this compound as a
reducing agent in organic synthesis should be investigated.
The five different approaches described above leave no
doubt that the empirical formula of the compound is

Na2H36N206: i.e., two sodium atoms per 2,2,2-crypt.
8.3.5. Stability

Stability of the compound towards temperature and air
was studied in a number of ways. It is found that the pure
compound is very sensitive to air. When the compound was
placed directly in contact with the air, a white solid
formed in a matter of seconds. However, if the sample were

Opened inside of an evacuable dry box, which was filled with

181

dry helium gas, or even inside of a dry bag flushed with
dry nitrogen, it was stable with a shiny surface for at
least a few minutes. Also if the sample were Opened under
a non-oxidizing dry solvent such as diethyl ether, the
sample remained unchanged in appearance even for an hour.
The solid is very reactive towards water, with gas evolu-
tion observed during the reaction. A hydrogen test re-
vealed the presence of this gas. The solid is relatively
stable at room temperature under vacuum. It decomposed at
about 83°. As the temperature is raised from that of
liquid nitrogen to the decomposition temperature, a gradual
color change from a bright gold color to dark brown color
was Observed. Above the decomposition temperature, a white
cOlor mixed with gray Spots was seen. After the solid had
been decomposed at~ 83°, it would redissolve back into
ethylamine to form a dark blue solution. Therefore, it is
believed that the solid decomposed at least partially into
its components, 2,2,2-crypt and solid sodium at ~I83°

according to

 

N32C18H38N206 >' 2Na ‘1‘ C18H36N206'

If the temperature of the solid was further increased, first
a blue liquid formed, then it turned to dark brown, then

black.
8.3.6. Studies by ESR and NMR techniques

The ESR Spectrum of a very pure gold-colored solid

showed only a very weak paramagnetism (almost undetectable).

£111.;
. R "3} I Q

 

182

A peak to peak width of «’13 G was Obtained from the first
derivative curve.

NMR studies of the 2,2,2-crypt protons in a pure solid
gave no observable signal. The signal was probably too

broad and completely buried under the base line.
8.3.7. Conductivity studies

The conductivity studies were important towards the
understanding of the nature of the solid. From the physical
appearance of the solid we expected the solid to be either
a metallic conductor or a semiconductor. If it were metal-
lic it would be of extreme interest in view Of the largely
organic nature Of the compound. At the beginning of this
study a sizable single crystal was not available. There-
fore various powder methods were tried in order to determine
the magnitude of the conductivity and its temperature de-

pendence.
8.3.7.1. Two electrode method

Due to the relatively soft nature of the solid, it was
hoped that the solid could be compacted in a Specially de-
signed cell, Figure 50. With two platinum electrodes
contacting the ends, a DC voltage was applied. The re-
sistance of the compacted powder was measured with a con—
ductivity bridge.

The solid was prepared and washed in an NMR tube and

then covered with purified diethyl ether. The solution was

183

 

 

 

 

 

Figure 50. Two electrode conductivity cell.

184

then frozen in liquid nitrOgen and the tube was broken and
opened to the air. The NMR tube was connected to the con-
ductivity cell with the help of shrink-tubing (Figure 50).
After the connection had been made, the conductivity cell
was pumped to high vacuum while the diethyl ether was
frozen. When the pressure reached <10-5 torr, the diethyl
ether was melted and the gold—colored solid was poured onto
one of the electrodes. The solid was still covered with
diethyl ether which was again forzen. The cell was broken
at the tOp in order to introduce the second electrode.
Before the second electrode was inserted, a glass rod was
used to pack the solid as tightly as possible. Then the
second electrode was introduced. Measurement of conductivity
was attempted with the aid Of a standard conductivity bridge
(Beckman Instruments, Inc.).

The result showed that for a centimeter of packed solid
the resistance was >106 ohms. Due to the soft nature of the
solid, we expected that the solid was packed reasonably well.
However, gaps between the particles could be the cause of

the high resistivity. Therefore further studies were needed.
8.3.7.2. Transformer current method

The apparatus used for this study is shown in Figure 51.
It consisted of two Ferrite cores, a sample tube in the
form of a ring, an AC voltage generator and an oscilloscope.
The system was designed to measure the resistance of a com-

pound with a resistance close to the resistance of sodium

185

 

 

 

 

 

 

 

 

 

 

I
m

l

 

 

 

' \J

 

Figure 51. Apparatus for the transformer current
conductivity study.

186

metal. (These studies were done before we were certain that
the compound was not metallic.) A sodium sample was made to
test the system. Purified sodium metal was dissolved in
purified ammonia. The sodium-ammonia solution filled the
ring section of the sample tube. Ammonia was then slowly
distilled out. A layer of finely divided sodium metal
formed in the ring. In order to give continuity of the
sodium film, the ring section was sealed off from the sample
tube and left in a 500° oven overnight. Sodium metal was
then found to be evenly distributed throughout the whole
ring. The resistance of the sodium film ring was calculated

according to the equation169

0
va = constant .
fV

 

Where R is the resistance of the film, V° is the voltage
output, Vi is the voltage input, f is the frequency of
the AC voltage applied. With a ring made of OOpper wire as
a standard, the ratio of the resistance of copper to that of
sodium was found to be 0.15. The calculated value using
the data obtained from a handbook was 0.40, if we make the
unjustified assumption that the lengths and cross sections
are equal. These results Show that this method should at
least be valid for order-of-magnitude estimates of the
resistance.

A film of the solid was obtained by dissolving both

sodium and 2,2,2-crypt in methylamine and then quickly

187

evaporating the methylamine. (Since methylamine has a much
higher vapor pressure than ethylamine, quick evaporation is
much easier.) A bright gold-colored film covered the whole
glass ring. However, probably due to either the high re-
sistivity of the compound or to the absence of current
paths no output voltage could be detected. Neither changes
in the input voltage nor in the frequency gave positive

results.
8.3.7.3. Microwave conductivity study.*

The complete set up for this study is shown in Figure 52.
It consisted of a power supply, an X-band microwave generator,
a frequency meter, an attenuator, a microwave cavity and a
crystal detector. The detected current signal was converted
to a DC voltage output by means of a load resistor. The
voltage was measured with an oscilloscope. A powdered
sample was sealed into a 3 mm O.D. low-loss quartz tube and
placed inside of a temperature-controlled Dewar which was
also made of low-loss quartz. The temperaturewas adjusted
by using a Varian 4540 temperature controller and was
measured with a digital thermocouple (Doric, model DS—350
type T). .

All the metals and nonmetals used in the experiments
were commercially available. No further purification was
done. A thermistor with a room temperature resistance of

A __A .4

*The author is grateful to Dr. Schwendeman of this depart-
ment for profiding the instrument and valuable advice.

IUBB

.msumm muH>Huosocoo m>mBOHoHE mnu How EmHmMHo xOOHm .Nm wusmHm

oncomOHHHomo

®

 

HOpmHmmH @509

).

 

 

 

 

fil. .moms..w0Hm
sz>mo Mom mHmQSm
HOHOOHOU _ mpH>mo .Omo Hmsom
Hopmhho . - HOHmssmpH< ,

 

 

 

coco

189

.~ 1000 ohms was used as an example for the temperature

dependent study. The results are listed in Table 16 and
the temperature dependent study of the thermistor is shown
in Figure 53. From the results obtained, an indication of
general trends can be given. Metals with a high conductivity
gave small signals («.5 mV) and large frequency shifts.
Salts or organic compounds with low conductivity gave a high
output voltage (~o15 mV) and semiconductors gave inter-
mediate values ofI~ 12 mV. This was because the dielectric
constant of a sample has an effect on the cavity Q. There-
fore, the measurement of microwave losses can be represented
as the AC conductivity of the sample. Quantitative results
could be Obtained if pellets with diameters only several
mils smaller than that of the cavity I.D. were used instead
of powdered samples.17°

The measured voltage for the gold-colored solid is
about the same as the voltages of the semiconductor, but is
much higher than the voltages expected for metals. This
result indicates the solid is probably not a metal conductor.
rather a semiconductor. However, since the temperature
study failed to give any detectable change in the output
voltage, we were unable to further clarify the nature of the

compound by this technique.

8.3.7.4. The use of conducting glass and aluminum plates

The DC conductivity was measured by placing powdered

samples between two conducting glass plates. The apparatus

190

 

hng

 

 

1 J
2.5

 

 

3.0
l/T( °K’Jx 103)

Figure 53. Plot of logarithm of resistance versus (l/T)
for the termistor.

191

Table 16. Microwave studies of the conductivities of
metals, semiconductors and insulators.

 

 

Output Voltage

 

Sample (mv)
Empty tube 16.2, 15.0 (Quartz)

NaCl 15.3
Triphenylmethane 15.0
Tellurium 12.7
Thermistor 12.8
Na-cryptate 14.0

Zn 4.0, 5.0

Cu 4.4

A1 6.5

Ti 5.0

 

192

consisted of a variable DC power supply, a OOpper shielding
and thermostating container, and a Keithly Model 417K Chro-
matograph Electrometer with a range of 10.5 to 10"14 amperes.
The temperature was adjusted by cooling or heating a copper
.rod outside of the container. The copper rod made direct
contact with the sample holder. The actual temperature was
Obtained from a thermocouple taped to the sample cell. The
input voltage from the power supply was accurately measured
with a Heath Universal Digital Instrument, Model EU-805.

The sample was prepared, washed and vacuum sealed in an
NMR tube. The conducting glass plates were cleaned with HF
cleaning solution and aqua regia as described in Chapter III,
then washed with conductance water and dried. Teflon tape
(one mil) was used as an insulating material placed between
the glass plates. It usually took three layers of Teflon
sheet to give complete insulation. A hole was cut in the
middle of the Teflon sheets so that the sample could contact
both plates. The sample was Opened in a pre-evacuated box
under helium atmosphere and placed in the hole in the Teflon
sheet. The plates were tightly clamped to each other to
insure direct contact with the sample as well as to provide
an air-tight seal between the plates. In the final experi-
ments, aluminum plates were used to avoid the problem of the
glass cracking. Glass had been used initially in an unsuc-
cessful attempt to measure photo-conductivity. The resist-
ance of two directly contacting glass plates was ~130 ohms

and < 10 ohms for the aluminum plates. These values were

193

found to be negligible compared to the resistance of the
sample. The resistance between two plates with only the

+
14 Ohms.

Teflon Spacers present was found to be > 10

In order to be sure that the current measured was not
caused by migration of ions and/or surface moisture, an
Ohm's law study was done with both positive and negative
polarities. A plot of input voltage versus current is shown
in Figure 54. Reversing the voltage exactly reversed the
current with no time lag. At a given temperature and
voltage, the current was independent of time. A temperature-
dependent study was done on a sample which was clamped be-
tween aluminum plates with Teflon sheets for insulation as
before.

The reversible variation of the resistance with tempera-
ture, together with rough estimates of the contact area and
thickness yield specific resistances which range from 5 x
1013 Ohm cm at 0° to 7 x 109 ohm cm at 60°. This pronounCed
variation in resistance suggests that the material behaves
as a semiconductor.

A plot of log R against l/T is shown in Figure 55.

In the temperature region studied, the graph Shows a definite
curvature. If the curvature were a result Of experimental
error, a straight line still could be drawn. The Slope of

the straight line is ~v6.5. From the slope an activation

energy ofI~ 1.3 ev was calculated by the equation

In R = Ea/kBT + ln R0

This indicates that the energy gap is ~72.6 ev.

194

.oHHOm mumuQ>HOIEsHoOm OLD How wosum BmH m.E£O mo uon .vm muson

A 30H x sodas V Homhsso

 

 

d .
o.H o o.HI
H H J
O
G I 0s.
O\
OO
I.o
O
O

A om

0:

 

 

SBSiIOA

195

 

 

 

 

98°
0 O
CO 0
ll - o o
O
8
/8 8
9 8/
O
10 I-
8/ 8/
o
0 O
(a/O
CD
9: OO
00
.3 9 " 0y
0
8 F"
7 l l l l
2.8 3.0 3.2 M 5.6
l/T ( °K‘l x 103 )
Figure 55.

Plot of logarithm of resistance versus
(1/T) for the sodium-cryptate solid.

 

 

196

However, the curvature could be caused by the varia-
tion of the band gap with temperature. The compound,
which forms with sodium and the large organic molecule,
2,2,2-crypt, is different from a "normal" semiconductor.
The change of distance between the sodium ions and the
organic molecules with changing temperature might also
affect the magnitude of the band gap.

Finally, another way to analyze the results is to
separate the data into two temperature regions. The high
temperature portion (> r.t.) could represent the intrinsic
conductivity. The points in the region fall on a straight
line with a lepe of ~I7.8. From this lepe, an energy
band gap of ~t3.0 e.v. is Obtained. However, the low tem-
perature points fall below the straight line. The devia-
tion from the straight line suggests that the activation
energy changes with the temperature. In this low tempera-
ture region, the number of electrons in the valence band is
very small. Therefore the electrical properties could be
controlled by impurities. The difference between the intrin-
sic conductivity and the impurity conductivity would then

be the cause Of the curvature of the line.
8.4. Summary

As a result Of the studies described in this chapter,
we prepared two new kinds of solid compounds of alkali
metals with 2,2,2-crypt or crown ether. The first type of

solid is formed by the combination of various alkali metals

197

with crown. This also can be achieved by combining K, Cs,

or Ba with 2,2,2-crypt. These solids, which have dark

blue colors, are relatively unstable. The ESR study re-
vealed its paramagnetic nature. It is possible that these
blue-colored solids resemble the alkali halide F-centersi71v172
The exact nature and properties of these compounds need to

be studied further.

The second type of new solid is formed by combination
of sodium with 2,2,2-crypt. The color of this compound is
gold. When the temperature changes from liquid nitrogen
temperature to room temperature, the color changes from
bright gold to brownish-gold. The compound is very stable
at low temperatures and decomposes at ~o83°. The compound
reacts vigorously with oxygen and water. From the ESR study
we presume that the solid is diamagnetic. Conductivity
studies of powdered samples give a resistivity of «'7 x
10+12 Ohm-cm at toom temperature. The resistivity decreases
with decreasing temperature and the solid behaves as a
semiconductor. A study of the reducing power gives approx-
imately one available electron for each sodium atom in the
compound. Finally various composition studies indicate the
structural formula of the compound is Na2C18H36N206. The
mode of formation of the compound and its prOperties suggest
that it is the first known solid compound which contains

the sodium anion, Na-. The counter-ion is the sodium cation

trapped in the crypt.

IX. SUGGESTIONS FOR FUTURE STUDIES

Because of the limitations of time, many interesting
projects could not be attempted or completed. It is the
author's intention to present them here as suggestions for
future work.

(1) The possibility of using 1,8-diamino-3,6-dioxa-
octane as a solvent for metal.

As we know, the diamine, 1,8-diamino-3,6-dioxaoctane,
is one of the starting materials for the first cyclization.
It is a dense liquid with two ether linkages and two pri-
mary amine groups. Therefore, it may be a very good sol-
vent for the alkali metals. If it dissolves metals as
expected, a new group of solvents will be added to the
present solvent list.

(2) Studies involving 1,7,10,16—tetraoxa-4,13—diaza-
cyclooctadecane (the first reduction product).

The first reduction product is similar to crypt in
many ways. For example, it forms complexes with metal
cations,”2 it also has two protonation sites, etc. On the
other hand, it is a secondary amine with only two chains.
The structure probably more closely resembles the crown
ether compounds than the crypt compounds. Therefore, the

possibility Of forming sandwich or "club sandwich" types
198

199

of complex with metals or metal cations in the solid or in
liquid solutions is high. With a different crystal struc-
ture from that of the cryptate compound, the prOperties of
this compound might be totally different.

(3) Concentration dependent conductivity and magnetic
susceptibility studies on metal-crypt-ammonia, amine and
ether solutions.

A metal solution containing excess crypt should con-
tain only the solvated electron, and complexed metal cations.
Conductivity data for such solutions can be used to test
the importance of ion-pairing in these solutions, since
ion-pair formation should be less favorable when the crypt
is present. The conductivity data along with the magnetic
susceptibility data could also be used to understand the
effect of the cation on the spin-pairing process.

(4) Use Of NMR shifts to determine the paramagnetic
susceptibility of a metal solution.

The preliminary results Showed that the crypt proton
positions shift to higher fields for a metal-crypt solution.
The shift is caused by the paramagnetic centers (solvated
electrons) in the solution. The amount of shift should be
directly proportional to the concentration of the para-
magnetic center in the solution. Therefore, a detailed
study of this paramagnetic shift can give paramagnetic sus-
ceptibility data for the solution.

(5) Further studies on the effect of the cation on the

spin-pairing process in dilute metal-ammonia solutions.

200

Better absolute results could be obtained with either
a radio frequency ESR or a low-power ESR instrument. Of
particular importance would be the concentration dependence
of the spin-pairing process.

(6) Solid state studies

Both the gold-colored solid and the blue-colored
solid could be studied in a variety of ways. First the
conductivity of a single crystal Of the solid Should be
measured with the four probe method. A temperature de-
pendent study of the conductivity can give the value of
band gap for the solid. Static and ESR studies of the
magnetic susceptibility of both gold-colored and blue-
colored solids should be made. Finally, a photo-conduc-
tivity study on these solids can give more information
about the nature of the compound. Hall effect measurements
might also give information about the mobility and the

type of the charge carriers.

APPENDIX

“is

 

00

APPENDIX

KINFIT

Subroutine EQN

2 CONTINUE
GA1=1 . 0
IF NOUNK-EQ-i UR=CONST(9)
IF NOUNK-EQ-l NS=CONST 10)
IF NOUNK-EQ-l UT=U(1;
IF NOUNK-EO.2 UR=U 1
IF NOUNK-EQ-Zg US=U(2)
IF NOUNK-EQ-z UT=1.0
NB=O
VT=CONST(4)
GA1 IS THE ACTIVITY COEFFICIENT FOR A MONOVALENT ION
VT 18 THE TOTAL VOLUME AT A PARTICULAR INSTANCE
200 CONTINUE
IF(NR-NE-0) VT=CONST(4)+A
ATEMP=A
IF(VT-LE-0.0) VT=0.0001
NB=NB+1
AGA1=GA1
HYD=(10.**(-XX(1)))/GA1
HYD IS THE HYDROGEN ION CONCENTRATION
NC=O
CA=O
ABC=1000.*CONST(1)/VT
BAC=1000.*CONST 2)/VT
GAA1=GA1**2
201 CRY=BAC/(1.+HYD)/UR+HYD**2/(UR*US*GAA1)+ABC*UT/
(1.+UT*CA))
CRY Is THE CONCENTRATION OF THE FREE CRYPT
IF(CRY.LE.0.0) CRY=0.0001
RATIO=CA/CRY
DIF=ABS(1.-RATIO
NC=NC+1
CA=CRY ‘
IFENC.GE.100) GO TO 202
IF DIF.GE.0.0001) GO TO 201
200 CONTINUE
SOD=ABC/(1.+UT*CRY)
SODCRY=SOD*UT*CRY
CH=CRY*HYD/UR
CHH=CH*HYD/(US*GAA1)
OH=CONST(8)/(HYD*GAA1)
CL=SOD+SODCRY+HYD+CH+CHH*2.-OH

201

 

00000000

203

300

35

202

ON3=CL*VT/1000.-CONST(1)

A=ON3*1000./CONST(3)

SOD IS THE CONCENTRATION OF SODIUM ION

SODCRY IS THE CONCENTRATION OF SODIUM-CRYPTATE
CH IS THE CONCENTRATION OF MONOPROTONATED CRYPT
CHH IS THE CONCENTRATION OF DIPROTONATED CRYPT
CL IS THE CHIDRIDE ION CONCENTRATION

OH IS THE HYDROXIDE ION CONGENTRATION

0N3 IS THE MOLES OF HCL ADDED

A IS THE CALCULATED NUMBER OF ML. ADDED
S=0.5*(SOD+SODCRY(CH+4.*CHH+HYD+CL+OH)
IF(S.LF.0.0) s=0.0001

AB=SORT(S)

5 IS THE IONIC STRENGTH
GAA=-(CONST(6)*AB/(1.+CONST(5)*CONST(7)*AB))
GA1=10.**GAA

IF(GA1.LE.0.0) GA1=0.0001

ARAT=A/ATEMP

 

IF NB.GE.100) GO TO 203
IF ABS GA1-AGA1).GE.0.0001) GO TO 200
IF ABS 1.—ARAT).GE.0.0001) GO TO 200

CONTINUE

IF(IMETH.NE.-1) GO TO 35

WRITE(JTAPE,300) xx(1),CRY,CH,CHH,SODCRY,S,GA1
FORMAT(4X.8E14.5)

xx(2)=A

RETURN

CONTINUE

RESID= A-xx(2)

RETURN

 

BIBLIOGRAPHY

10.

11.

12.

13.

14.

BIBLIOGRAPHY

W. Weyl, Ann. Phys., 197, 601 (1863).

 

For references to metal ammonia solutions see (a) M.
H. Cohen and J. C. Thompson, Adv. in Physics, 17, 857
(1968); (b) U. Schindewolf, Angew Chem., Int. Edit.,
I, 190 (1968); (c) Metal-Ammonia Solutions, W. L. ,_,
Jolly, ed. Dowden, Hutchinson and Ross: Strondsburg, ‘
Pa. (1972).

 

 

 

 

 

R. A. Ogg, Jr., J. Chem. Phys., Lg, 295, 1141 (1946).

 

W. Bingel, Ann. Physik., 12, 57 (1953).

 

M. F. Deigen, Trudy: Inst. Fiz. Acad. Nank, U.S.S.R.,
é/ 11971954: Zh. Eksp, Teor. Fiz., 2E, 300 (1954)?

 

 

E. Becker, R. H. Linquist, and B. J. Alder, J. Chem.
Phys., 25, 5, 971 (1956).

E. Arnold and A. Patterson, Jr., J. Chem. Phys., $1,
3089, 3098 (1964).

 

M. Gold, W. L. Jolly and K. S. Pitzer, J. Amer. Chem.
Soc., g2, 2264 (1962).

 

S. Golden, C. Guttman, and T. R. Tuttle, Jr., ibid.,
g1, 135 (1965); J. Chem. Phys., 22, 3791 (1966).

 

G. E. Gibson and T. E. Phipps, J. Amer. Chem. Soc.,
fig, 312 (1926).

 

K. Bar-Eli and G. Gabor, J. Phys. Chem., 11, 323 (1973).

 

For other references to solutions of metal-amines
and ethers see M. T. Lok, F. J. Tehan, and J. L. Dye,
J. Phys. Chem., 26, 2975 (1972).

 

J. W. Fletcher, W. A. Seddon, J. Jevcak, and F. C.
Sopchyshyn, Chem. Phys. Lett., lg, 592 (1973).

 

J. W. Fletcher, W. A. Seddon, and F. C. Sopchyshyn,
Can. J. Chem..§l, 2975 (1973).

 

203

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

204

F. Y. Jou and L. M. Dorfman, J. Chem. Phys., 58, 4715
(1973). '””

 

R. Bockrath and L. M. Dorfman, J. Phys. Chem., 11,
1002 (1973).

 

L. M. Dorfman, F. Y. Jou, and R. Wageman, Ber.
Bunsenges Phys. Chem., 15, 681 (1971).

 

J. L. Dye, M. T. Lok, F. J. Tehan, R. B. Coolen, N.
Papadakis, J. M. Ceraso, and M. DeBacker, ibid.,
IQ. 659 (1971).

G. Gabor and K. Bar-Eli, J, Phys. Chem., 15, 286
(1971).

 

A. Gaathon and M. Ottolenghi, Israel J. Chem., 8, 165
(1970).

 

R. Catterall, J. Slater, and M. C. R. Symons, J. Chem.
Phys., Q2, 1003 (19701 ‘“"““"

For other references to optical properties of metal
solutions see I. Hurley, T. R. Tuttle, Jr., and S.
Golden, Pure Appl. Chem., 449 (1970).

 

R. Catterall, I. Hurley, and M. C. R. Symons, J. Chem.
Soc., Dalton Trans., 139 (1972).

 

For other references to ESR studies of metal solutions
see V. A Nicely and J. L. Dye, J. Chem. Phys., 53,
119 (1970).

 

S. Matalon, S. Golden, and M. Ottolenghi, J. Phys.
Chem., 23” 3098 (1969).

J. L. Dye, M. G. DeBacker, and L. M. Dorfman, J. Chem.
Phys., 52, 6251 (1970)

S. H. Glarum and J. H. Marshall, ibid., 52, 5555
(1970).

L. J. Giling, J. G. Kloosterboer, R. P. H. Rettschnick,

and J. D. W. Van Voorst, Chem. Phys. Lett., 8, 457
(1971).

 

J. G. Kloosterboer, L. J. Giling, R. P. H. Rettschnick,

and J. D. W. Van Voorst, ibid., 8, 462 (1971).

D. Hu pert and K. Bar-Eli, J. Phys. Chem., 74, 3285
(1970§.

 

31.

32.

33.
34.

35.

36.

37.

38.

39.

40.

41.
42.
43.

44.

45.

46.

47.

48.

49.

50.

205

M. T. Lok, F. J. Tehan, and J. L. Dye, ibid., 16,
2975 (1972).

J. L. Dye, M. G. DeBacker, and V. A. Nicely, J. Amer.
Chem. Soc., 92, 5226 (1970).

 

R. c. Douthit and J. L. Dye, ibid., 32, 4472 (1960).

M. Gold, and W. L. Jolly, Inorg, Chem., 1, 818 (1962).

 

D. F. Burow and J. J. Lagowski, Adv. in Chem. Series,
No. 59, 125 (1965).

 

J. W. Fletcher, Private Communication, to be published.

M. T. Lok and F. J. Tehan, This Laboratory, unpublished
work.

C. A. Hutchinson and R. C. Pastor, J. Chem. Phys., 21,
1959 (1953).

 

H. M. McConnell and C. H. Holm, ibid., 26, 1517 (1957).

J. Acrivos and K. S. Pitzer, J. Phys. Chem., 66, 1693
(1962).

 

 

D. E. O'Reilly, J. Chem. Phys., 4;, 3729 (1964).
C. Lambert, ibid., 2g, 2389 (1968).

C. A. Kraus, J. Amer. Chem. Soc., 43, 749 (1921).

 

R. R. Dewald and J. H. RobertS, J. Phy_. Chem., $2,
4224 (1968).

 

J. L. Dye, R. F. Sankuer, and G. E. Smith, J. Amer.
Chem. Soc., 82, 4797 (1960).

 

J. L. Dye, Metal-Ammonia Solution, Coloque Weyl II,
P1. J. J. La owski and M. J. Sienko, eds. Butterworths:
London (1970?.

For general review on this subject see B. C. Webster
and G. Howat, Radiation Res. Rev., 4, 259 (1972).

 

N. F. Mott and R. W. Gurney, Electronic Processes in
Ionic Crystaha Oxford University Press: London (1948).

 

 

s. Pekar, J. Phys., (U.S.S.R;), 10, 341,347 (1946).

 

 

A. S. Da dow, J. Exptl. Theoret. Phy_, U.S.S.R., 18,
913 (1948).

206

51. M. F. Deigen and Y. A. Tsvirko, Ukr. Fiz. Zh., 1, 3,
245 (1956). ~

 

52. R. A. 099, Phys. Rev., 33, 669 (1946); J. Amer. Chem.
Soc., 33, 155 (1946).

 

 

 

53. W. N. Lipscomb, J. Chem. Phys., 33, 52 (1953).
54. R. A. Stairs, ibid., 21, 1431 (1957).
55. J. Kaplan and C. Kittel, ibid., 33, 1429 (1953).

56. J. Jortner, N. R. Kestner, S. A. Rice, and M. H.
Cohen, J. Chem. Phys., 23, 2614 (1965).

 

57. N. R. Kestner, J. Jortner, S. A. Rice and M. H.
Cohen, Phys. Rev., 140, 456 (1965)

 

58. B. E. Springett, M. H. Cohen, and J. Jortner, J. Chem.
Phys., 43, 2720 (1968).

59. J. Jortner and N. R. Kestner, MetaleAmmonia Solutions,
Colloque Weyl II, p. 49, J. J. Lagowski and M. J.
Sienko, eds., Butterworths: London (1970).

 

60. D. A. Copeland,N. R. Kestner, and J. Jortner, ibid.,
33, 1189 (1970). "-

61. K. Fueki, D. F. Feng, and L. Kevan, J. Amer. Chem.
Soc., 33, 1398 (1973).

 

62. D. A. Copeland, and N. R. Kestner, J. Chem. Phys., 33,
3500 (1973).

 

63. D. F. Feng, K. Feuki, and L. Kevan, ibid., 33, 3281
(1973).

64. Cady, J. Phys. Chem., ,1, 707 (1897).

65. C. A. Kraus, J. Amer. Chem. Soc., 29, 1557 (1907); 30,

653, 1157, 13W 190 ; w, 6 (19’1’4); 4’3, 749 (1921’).

66. C. A. Kraus and W. W. Lucasse, J. Amer. Chem. Soc., 23,
2529 (1921): 4,4,, 1941 (1922).

 

67. R. M. Fristrom, Ph.D. Thesis, Stanford University (1949).

68. E. Huster, Ann. Phys., 33, 477 (1938).

 

 

69. S. Freed and N. Sugarman, J. Chem. Phys., 13, 354
(1943).

70. For other references to the models of metal-ammonia
solutions see 46. (

71.

72.
73.
74.
75.

76.

77.

78.

79.

80.

81.

82.

83.

84.

85.

86.
87.

88.
89.

207

A. DeMortier, M. DeBacker, and G. Lepoutre, J. Chim.
Phys., 380 (1972).

T. L. Hill, J. Chem. Phys., 13, 394 (1948).

 

R. H. Land and D. E. O'Reilly, £Q£Q., $3, 4496 (1967).
D. E. O'Reilly, 123g” 33, 474 (1971).

K. Fueki, $233,, 33, 5381 (1969).

K. Fueki, and S. Noda, Metal-Ammonia Solutions,

Colloque Weyl II, J. J. LagOWSki and M. J. Sienko, eds.
p 19. Butterworths, London (1970).

 

N. R. Kestner and D. A. Copeland, Metal~Ammonia Solu-
tions, Colloque Weyl II, J. J. Lagowski and M. J.
Sienko, eds. p 27. Butterworths, London (1970).

 

J. L. Dye, Metal-Ammonia Solutions, Colloque Weyl III,
to be published.

I. Hurley, T. R. Tuttle, Jr., and S. Golden, J. Chem.
Phys., 23, 2818 (1968).

D. S. Bern, E. C. Evers, and P. W. Frank, Jr., J. Amer.
Chem. Soc., 79, 5118 (1957).

 

R. R. Dewald and K. W. Browall, J. Phys. Chem., 11,
129 (1970).

 

R. R. Dewald and J. L. Dye, J. Phys. Chem., 33, 128

For other references to experimental evidences and
theoretical calculations of M in gas phase see S.
Golden, C. Guttman, and T. R. Tuttle, Jr., J. Chem.
Phys., 22, 3791 (1966).

M. Smith and M. C. R. Symons, Trans. Faraday Soc.,
33, 338, 346 (1958).

G. Stein and A. Treinen, Trans. Faraday Soc., 33, 1086,
1091 (1959)

I. Burak and A. Treinen, ibid., 33, 1490 (1963).

J. L. Dye, M. G. DeBacker, J. A. E re, and L. M. Dorf-
man, J. Phys. Chem., 23, 839 (1972 .

 

M. G. DeBacker and J. L. Dye, ibid., 13, 3092 (1971).

K. D. Vos and J. L. Dye, J. Chem. Phy§,, 33, 2033 (1963).

 

Infir—

 

‘3‘

90.

91.

92.

93.

94.

95.

96.

97.

98.

99.

100.

101.

102.

103.

104.

105.

106.

208

L. R. Dalton, J. D. Rynbrandt, E. M. Hansen and J. L.
Dye, J. Chem. Phys., 23, 3969 (1966).

 

L. R.)Dalton, M. S. Thesis, Michigan State University
1966 .

J. Eloranta and H. Linschitz, J. Chem. Phys., 38, 2214
(1963). ""

 

M. Ottolenghi, K. Bar-Eli and H. Linschitz, ibid., 43,
206 (1965). "‘““' ”W

W. A. Seddon, E. B. Selkirk and J. W. Fletcher, Int.
J. Radiat. Phys. Chem., 3, 323 (1973).

D. M. J. Compton, J. F..Bryant, R. A. Cesena, and B. L.
Gehman in Pulse Radiolysis. Academic, New York, 1965,
p. 43.

B. Ward, Advan. Chem. Ser., 31, 601 (1968).

 

M. Anbar and E. J. Hart, J. Phys. Chem., 39, 1244 (1965).

 

L. R. Dalton, J. L. D e, E. M. Fielden, and E. J. Hart,
ibid., 23, 3358 (1966 .

 

S. Arai and M. C. Sauer, J. Chem. Phys., 22, 2297 (1966).

H. Nauta and C. V. Huis, J. Chem. Soc., Faraday_Trans. I,
4, 647 (1972).

 

For general review see (a) J. J. Christensen, J. 0.
Hill, and R. M. Izatt, Science, 174, 459 (1971).
(b) A. Leveque and R. Rosset, Analysis, 3, 154 (1973).

 

The author is grateful to Mr. F. J. Tehan for his help
in preparing the starting materials and working on the
first and second cyclization, and to J. Ceraso and K.
Voorhees for their help in synthesizing some of the
starting materials.

B. Dietrich, J. M. Lehn and J. P. Sauvage, Tetrahedron
Lett., 2885 (1969); Tetrahedron, 33, 1629 (1973);
Tetrahedron Lett., 2889 (1969)i_

 

 

 

J. M. Lehn, J. P. Sauvage, B. Dietrich, J. Amer. Chem.
33, 2916 (1970). -

 

B. Dietrich, J. M. Lehn and J. P. Sauvage, Chem. Comm.,
487 (1971).

 

J. M. Lehn and J. P. Sauvage, Chem. Comm., 440 (1971).

 

209

107. J. Cheney and J. M. Lehn, J. C. S. Chem. Comm., 487
(1972).

108. J. Cheney, J. M. Lehn, J. P. Sauvage, and M. E.
Stubbs, J. C. S. Chem. Comm., 1100 (1972).

 

109. J. M. Lehn, and F. Montovon, Tetrahedron Lett.,
4557 (1972).

 

110. B. Dietrich, J. M. Lehn and J. P. Sauvage, Tetrahedron,
32, 1647 (1973).

 

111. B. Difitrich and J. M. Lehn, Tetrahedron Lett., 1225
(1973 .

 

112. B. Metz, D. Moras, and R. Weiss, Chem. Comm., 444
(1971).

 

113. D. Moras and R. Weiss, Acta C;yst., 333, 440 (1973).

 

114. D. Moras, B. Metz and R. Weiss, ibid., 333, 383
(1973).

115. D. Moras, B. Metz, and R. Weiss, 3233,, 333, 388 (1973).
116. B. Metz, D. Moras, and R. Weiss, 3333,, 333, 1377 (1973).
117. B. Metz, D. Moras, and R. Weiss, 3313,, 332, 1382 (1973).
118. B. Metz, D. Moras, and R. Weiss, 1333,, 333, 1388 (1973).

119. R. Wiest and R. Weiss, J. C. S. Chem. Comm., 678 (1973).

 

120. H. S. Gutowsky and A. Saika, J. Chem.Phys., 33, 1688
(1953).

 

121. H. S. Gutowsky and S. Fujiwara, ibid., 33, 1782 (1954).
122. R. A. Ogg. ibid., 33, 560 (1954).

123. E. Grunwald, A. Loewenstein and S. Meiboom, ibid., 25,
382 (1956); 31, 630 (1957). ‘““

124. A. Loewenstein and S. Meiboom, ibid., 31, 1067 (1957).

125. S. Meiboom, A. Loewenstein and S. Alexander, ibid.,
32, 969 (1958).

126. A. T. Bottini and J. D. Roberts, J. Amer. Chem. Soc.,
33, 5126 (1956); 33, 5203 (1958).

127. M. Saunders and F. Yamada, ibid., 33, 1882 (1963)-

128.

129.

130.

131 O

132.

133.
134.
135.

136.

137.

138.

139.

140.

141.

142.

143.

210

N. Mammano, Metal-Ammonia Solutions, Colloque Weyl II,
J. J. Logowski and M. J. Sienko, eds., p. 367.
Butterworths, London (1970).

N. Mammano and M. J. Sienko, J. Amer.Chem. Soc., 90,
6322 (1968) . ”w

 

W. J. McDonald, D. E. Bowen and J. C. Thompson, Bull.
Am. Phys. Soc., 23, 735 (1964).

 

J. A. Morgan, R. L. Schroeder, and J. C. Thompson,
J. Chem. Phys., 23, 4494 (1965).

W. J. McDonald and J. C. Thompson, ibid., 11, 4607
(1967).

 

H. Jaffe, z. Physik, 22, 741 (1935).

R. A. Levy, Phys. Rev., 102, 31 (1956).

 

W. S.)Glaunsinger, Ph.D. Thesis, Cornell University
(1973 .

W. S. Glaunsinger, S. Zolotov, and M. J. Sienko,
J. Chem. Phys., 23, 4756 (1972).

 

F. W. Cagle and H. J. Holland, presented at the 145th
National Meeting of the American Chemical Society,
New York (Sept. 1963).

M. J. Sienko, Metal-Ammonia Solutions, Colloque Weyl I.
G. Lepoutre and M. J. Sienko, eds. p. 23. Benjamin,
New York (1964).

 

D. S. Thompson, M. J. Stone and J. S. Waugh, J. Phys.
Chem., 13, 934 (1966).

H. Oesterreicher, N. Mammano, and M. J. Sienko, J. of
Solid State Chem.,le 10 (1969).

M. G.)DeBacker, Ph.D. Thesis, Michigan State University
(1970 .

R. R.)Dewald, Ph.D. Thesis, Michigan State University
1966 .

The author is grateful to Dr. C. H. Park of E. I.
duPont de Nemours Co. and to Professor J. M. Lehn of
the Institut de Chemie, Strabourg, for providing
samples of crypt for part of the studies performed in
this thesis.

144.

145.

146.

147.

148.

149.

150.

151.

152.

153.

154.
155.

156.

211

J. M. Ceraso and J. L. Dye, J. Amer. Chem. Soc., 95,
4432 (1973). '””

 

J. L. Dye, M. T..Lok, F. J. Tehan, J. Ceraso, and K
Voorhees, J. Org, Chem., 33, 1775 (1973).

 

Potassium phthalimide was brought from Matheson
Coleman and Bell Manufacturing Chemists and used
without purification.

1,2(bis-(2-chloroethoxy))ethane was bought from J. T.
Baker Chemical Co.

The method used for preparing diamine is different
from the way Dr. J. M. Lehn explained in his prepara-
tion. With commercially available 1,2(bis-(2-chloro-
ethoxy))ethane, the synthesis appeared to be much
simpler and faster. The yield and quality are found
equivalent for both syntheses.

We found out that a long period of addition might not
be necessary. We have added both the diamine and di-
acid chloride very quickly into a large flask con-
taining 1 liter of benzene with vigorous stirring.
Product is obtained with both good yield and quality.
However, we have not compared the quality of the pro-
duct from the methods of slow and fast addition very
carefully.

THF with no preservative is purchased from Burdick and
Jackson Laboratories, Inc. Freshly cut sodium is
added to the THF as soon as the bottle is opened.
Immediately prior to use, THF is refluxed on top of
sodium for two hours, and poured into the reaction
vessel.

Final elemental analysis was done by Spang Microanaly-
tical Laboratory, Ann Arbor, Michigan.

1M solution of borane in THF is obtained from Alfa
Chemical Co. The anion exchange resin must be in the
hydroxide form.

J. L. Dye and L. H. Feldman, Rev. Sci. Instr., 31,
154 (1966).

 

J. P. Sauvage, Thesis, Strasbourg (1971).

 

V. A. Nicely and J. L. Dye, J. Chem. Education, 23”
443 (1971).

F. J. Tehan and J. L. Dye. This laboratory, unpub-
lished data.

157.

158.

159.

160.

161.

162.

163.

164.

165.

166.

167.

168.

169 O

170.

171.

172.

212

J. A. POple, W. G. Schneider, and H. J. Bernstein,
High-resolution Nuplear_M§gnetic Resonangg, McGraw
Hill, New York (1959;.

M. Blandaner and M. Fox, Chem. Rev., 22, 59 (1970).

 

M. Blandaner, T. Gough, and M. C. R. Symons, Trans.
Faraday Soc., 33” 286 (1966).

 

T. Griffiths and R. Wijayake, Molec. Spectrosc.,
1562 (1970).

 

M. Fox and E. Hayon, Chem. Phys. Lett., 34, 442 (1972).

 

E. M. Kosower, An Introduction to Physical Organic
Chemistry, Wiley, New York(1968).

 

 

E. M. Kosower, J. Amer. Chem. Soc., 32, 3253 (1958).

 

M. Fisher, G. Ramme, S. Claesson, and M. Szwarc,

Chem. Ph 5. Lett., 9, 306, 309 (1971); Proc. Roy. Soc.,
A527, 281 (1972). ~

G. Ramme, M. Fisher, S. Claesson, and M. Szwarc,
Proc. Roy. Soc., A327, 476 (1972).

 

 

V. Nicely, Ph.D. Thesis, Michigan State University
(1969).

C. P. Poole, Jr., Electron Spin Resonance, Interscience,
New York (1967).

 

F. J. Tehan, Ph.D. Thesis, Michigan State University
(1973).

S. Naiditck, Metal-Ammonia Solutions, Colloque Weyl I,
P. 113, J. J. Lagow§kigafidfiM. J. Sienko, eds. Butter-
worths, London (1973).

 

M. A. Abkowitz and A. I. Lakatos, J. Chem. Phys.,
31, 5033 (1972).

J. H. Schulman and W. D. Compton, Color Centers in
Solids, MacMillan Co., New York (19627.

 

Jordan J. Markham, F-Centers in Alkali Halides,
Academic Press, New York (1966).

 

      
  

SIT

will$611611!lefilflfljflfilsmmlflifllj‘mfifs