EFFECT QF ION ~CLUSTERING AND
BIANION FORMATION 0N THE‘RATE
SF REACTION OF ANTHRACENIDE.
RADICAL ANION WITH ETHANOL IN THF

Thesis for the Degree of Ph. D..
MICHIGAN STATE UNIVERSITY
LE BINH LONG
1973

ig‘ i.) -‘ fl '. n '3', L

. ’5- ‘ V L ‘ Ii. 1. '31 AL '
n o. . o- I

Michigan {Edits -
Universrty

     

This is to certify that the

thesis entitled

EFFECT OF ION-CLUSTERING AND DIANION FORMATION
ON THE RATE OF REACTION OF ANTHRACENIDE
RADICAL ANION WITH ETHANOL IN THF

presented by

Le Dinh Long

has been accepted towards fulfillment
of the requirements for

Ph.D. Chemistry

degree in

 

 

9 7
n , .,
Aqt/L 1,” 1‘74 & / 1&{1 " (J
v , I/

.l'

( Major professor '

Date (Kali/(L, [5/ /9 7k

0-7 639

 

    
       

   

'5’ amime‘ av “‘7’
.I unsasns
”;5mxmwmvmu

LIBRARY DINOERS

._.—-'

 

I sralusrmmlcmq I
IILKs- M4949” I

 

M
'1’
’1

'zan

wee Of cc:
AraPid scar.

”a
IF
I.

J‘ antnrace

‘»

 

 

ABSTRACT
EFFECT OF ION-CLUSTERING AND DIANION FORMATION

ON THE RATE OF REACTION OF ANTHRACENIDE
RADICAL ANION WITH ETHANOL IN THF

BY

Le Dinh Long

The protonation of potassium anthracenide (K+AnT),
by ethanol in tetrahydrofuran (THF), was studied over a wide
range of concentrations of ethanol and free anthracene.
A rapid scanning stepped-flow apparatus was used in the study.
The anthracenide anion was produced by the reduction of
anthracene with potassium metal in THF.

The protonation rate was largely second-order in the
absorbance, thus revealing the existence of intermediate
steps prior to the protonation step. At low alcohol concen-
trations, the pseudo second-order constant depended upon the
ratio [ROH1/[An], where [An] denotes the concentration of free
anthracene. These results are consistent with the following
"dianion mechanism":

k
2(An ,K ) -;—* (An , 2K ) + An

k
= + + - + -
(An ,2K ) + ROH ——l+» K AnH + K RO
fast

+ - _
K AnH + ROH ————> AnH2 + K+RO

 

A: high con:
increased w:

0'01.

9 ‘ ‘ M V
; this

ZIPS. I}

In - ".v.
Inn ,A:

 

 

Le Dinh Long

At high concentrations of alcohol, the rate of protonation
increased with [ROH], but did not depend significantly on

[An]. This suggested the participation of a quadruple ion-
cluster intermediate as the protonated species. A "dianion

and ion-cluster mechanism" was then proposed:

2(ATK+)-—EQ+ ATK+ —k—+—>(A: + A
n , ( n , )2 k' n ,2K ) + n
k

= + + - + -
(An ,2K ) + ROH -—l9' K AnH + K RO

k

? + 2 + — + -
(An ,K )2 + ROH ———+- K AnH + K R0 + An

K+AnH— + ROH ———>faSt AnH2+ K+RO_

An attractive alternate mechanism was also proposed: the
"cation solvation mechanism", in which the intermediate Species
might be protonated by K+.ROH already present in the aromatic
system. This ”intra-complex" protonation scheme was further
supported by the apparent insensitivity of the protonation

rate at relatively high values of [ROH] toward the nature of
the alcohol used.

A preliminary study of the effect of dicyclohexyl-lB-
crown-6 (crown) on the rate of protonation was also undertaken.
Crown eliminated the second—order component, leaving only a
slow first-order contribution to the rate. The effect of crown
provides strong evidence that the second—order protonation

pathway requires contact ion-pair formation.

 

Thee

5i:i~n*t ‘w‘

.4 we”;

aian;cn I

contact 1

I) T
hm , K

I
s

$

1

 

 

Le Dinh Long

These results and those of other investigators are con—
sistent with a decrease in the protonation rate with decreas-
ing charge localization in the aromatic system in the order:

= + .
dianion (An ,2K ) > quadruple ion—cluster (AnT,K+)2

‘\

. . + .
contact ion-pair (AnT,K ) > solvent-separated ion pair

+ .
(AnTII K ) > free or solvent solvated ion (AnT).

EFFECT OF ION-CLUSTERING AND DIANION FORMATION
ON THE RATE OF REACTION OF ANTHRACENIDE

RADICAL ANION WITH ETHANOL IN THF

BY

Le Dinh Long

A THESIS

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

DOCTOR OF PHILOSOPHY

Department of Chemistry

1973

To:
My Parents
and

Ngoc-Xuan

ii

 

The ac
Professor
ad enccur
fins work.

He als:

: ‘ ~
.rienc, f3;

 

 

 

ACKNOWLEDGEMENTS

The author wishes to express his sincere appreciation to
Professor James L. Dye whose invaluable guidance, assistance
and encouragement greatly contributed to the completion of
this work.

He also wishes to thank Mr. Joseph M. Ceraso, a special
friend, for his untiring help during the tedious hours of
kinetics experiments as well as data analysis.

The author is indebted to Mssrs. Frederick J. Tehan,
Richard B. Coolen, Mei-Tak Lok, Nicholas Papadakis and Dr.
Marc G. DeBacker for their valuable suggestions.

Finally, financial support from the United States Agency

for International Development is gratefully acknowledged.

 

iii

'1' Eve—an

0". AoAU-J

2.2

 

I. INTRODUCTION.

II. HISTORICAL. .

TABLE OF CONTENTS

2. l--Early Studies . . . . . . . . . . . . . .
2.1.1--Production of Ar via Reduction with

Alkali Metals. . . . . . . . . . . .

2.1.2--Studies of Optical Spectra . . . . .
2.1.3--Electrochemical Studies. . . . . . .
2.1.4--E.S.R. Studies . . . . . . . . . . .
2.1.5--Ion-pairing Studies. . . . . . . . .

2.1.5.l--Evidence for a Dissociative
Equilibrium . . . . . . . .
2.1.5.2--Evidence for a Non—dissoci—
ative Equilibrium . . . . .
.3--Other Equilibria. . . . . .
.4--Conclusions Regarding Ion—
pairing . . . . . . . . .

2.1.5
2.1.5

2. 2--Kinetics Studies. . . . . . . . . . .
2.2.l--The Mechanism of Paul, Lipkin and

Weissman . . . . . . . . . . . . . .

2.2.2--Kinetics Studies in Ammonia. . . .
2. 2.3--Pulse- -radiolysis Studies . . . .
2.2. 4--Kinetics Study Using ESR and Polar-

ography. . . . . . . . . . . . . .

2.2. 5--Combined Polarography- -ESR- UV Absorp-

tion Spectroscopy Studies. . . . .

L 2. 6--Kinetics Studies Using Polarography

and StOpped- -flow . . . . . . . . .

2.2.7--St0pped-flow Studies . . . . . . . .

III. EXPERIMENTAL.

2.2.7.1--The Work of Minnich . . .

2.2.7.2--The Studies of Bank and
Bockrath. . . . . . . . . .

2.2.7.3--The Studies of Szwarc and
Co-workers. . . . . . . . .

3 o 1--Vacumn TeChniqueS o o o o o o o o o o o o o
3.2--Cleaning Techniques . . . . . . . . . . . .

iv

.b

&D\IO\U'IJ>

10

12
14

15
15

15
21
24
25
27
28
28
30
32
36

36
36

....4
P

h-

.T . 5 UR:

.

 

2
.

S 5 p3 u. u
“L
D.
Tu
hrs

.3.
\HA

 

TABLE OF CONTENTS--Continued Page

3.3--Purification Techniques. . . . . . . . . . . 37
3.3.l--Tetrahydrofuran . . . . . . . . . o . 37
3.3.2--A1cohols. . . . . . . . . . . . . . . 38
3.3.3--Alkali metals . . . . . . . . . . . . 39
3.3.4--Anthracene and Crown. . . . . . . . . 39
3.3.5--Helium. . . . . . . . . . . . . . . . 39

3.4--Preparation. . . . . . . . . . . . . . . . . 4O
3.4.l--Pre-rinsing Techniques. . . . . . . . 40
3.4.2--Anthracene and Crown Solutions. . . . 4O
3.4.3-—Alcohol Solutions . . . . . . . . . . 42
3.4.4--Anion Solutions . . . . . . . . . . . 42

3.5--The Stopped- -flow Experiment. . . . . . . . . 45
3.5.l-—General . . . . . . . . . . . . . 45
3.5.2--The Stopped- -flow Apparatus. . . . . 45
3.5.3--Calibration of the Stopped- -flow

System. . . . . . . . . . . . . . . . 49
3.5.4--Operations. . . . . . . . . . . . . . 51
3.5.5--Data Acquisition. . . . . . . . . . . 53
3.5.6--Data Analysis . . . . . . . . . . . . 58

3.5.6.l--General. . . . . . . . . 58

3.5.6.2--Program PUNDAT . . . . . . . 63

3.5.6.3--Program KINFIT . . . . . . . 64

IV. SURVEY AND TREATMENT OF THE DATA . . . . . . . . . . 65
4.l--Survey and Discussion of the Data. . . . . . 65
4.2--Data Treatment . . . . . . . . o . . . . . . 68
V. RESULTS AND CONCLUSIONS. . . . . . . . . . . . . . . 75

5. l-—Resu1ts. . . . . . . . . . . . . . . . . 75
5.1.l--Inapplicable Mechanisms . . . . . . . 75
5.1.2--Effect of Crown . c . . . . . . . . . 78
5.1. 3--Proposed Mechanisms . . . . . . . . 78

5.1. 3.l--Dianion and Ion— cluster

Mechanism. . . . . . . . . . 81

5.1. 3. 2--Cation Solvation Mechanism . 85

5. 2--Discussion . . . . . . . . . . . . . . . . 95
5. L l--Comparison of the Two Proposed

Mechanisms. . . . . . . . . . . . . 95

5. L 2--Discrepancies Between Some of
Minnich' 3 Results and Those of Bank

and.Bockrath. . . . . . . . . . . . . 98
5.3--Summary. . . . . . . . . . . . . . . . . . . 99
5.4--Conc1usion and Suggestions for Further Work. 100

REFERENCES 0 O O O O O O I O O O O O O O O O O 0 O O 6 O 10 3

tion
Minn

 

LIST OF TABLES

TABLE Page

I. List of Experiments. . . . . . . . . . . . . . . . 66
II. List of Minnich's Experiments Re—analyzed by the

Author (2K+An“' + ZROH -.'-Pfl1~>AnH2 + An + 2K+R0') . . . 69

III. Observed and Calculated Pseudo Second—order Rate
Constants for the Reaction of K+AnT with EtOH in
THF and Pseudo First-order Contribution at High
Ethanol Concentrations (from this study) . . . . . 86

IV. Calculated Constants from Equation 70. . . . . . . 89

V. Pseudo Second—orderTfigte Constants for the Reac—
tion 2K+An" + 2ROH —->- AnH + 2K+0R + An (from

Minnich's re-analyzed data)? . . . . . . . . . . . 96

vi

_.._ , . _ _ 1__. __' __

--r,~r~p':

..
. .Vvanu

 

d
E; :1 :
PSe

 

 

LIST OF FIGURES

FIGURE

1.

10.

ll.

12.

13.

14.

15.

Special set-up for pre-rinsing technique. . . . . .

Vessels used in the preparation of: (a) An, crown
or ROH solutions (b) K+AnT solutions. . . . . . . .

Schematic diagram of the stOpped-flow apparatus . .
Syringe used in the stOpped-flow system . . . . .

Four-jet mixing cell. . . . . . . . . . . . . . . .
Absorbance vs C/CO of the solutions of KMnO in H O

4 2

Decay of spectrum of K+Anr during reaction with
EtOH in THF 0 O O O O O O O O O O O I I O O O 0 I 0

Block diagram of the data—acquisition system. . . .

Decay of the 720 nm peak of K+Anf after mixing with
0.0052 M EtOH in THF (scanning technique) . . . . .

+ r . . I
Decay of the 720 nm peak of K An after mix1ng With
0.090 M EtOH in THF (fixed wavelength technique). .

Block diagram of the analysis system. . . . . . . .
Parallel first- and second-order kinetics applied
to th reaction K An" + EtOH in THF ([EtOH]O = 8.9

x 10' M) . . . . . . . . . . . . . . . . . . . . .

Log-log plot of kES (second-order) 13 [EtOH] for
the reaction of K An? with EtOH in THF. . . . . . .

Log-log plot of kps (second-order) gs (a) lEtOH]
and (b) [EtOH]/[An], for the reaction of K An? with
EtOH in THF 0 O C O C C C O O O I O O O O O O O O O

o g _ +
Pseudo first-order plot for the reaction of K An? +
(excess crown) with EtOH in THF . . . . . . . . . .

Vii

Page

41

43
46
48
50

52

54

55

59

60

62

72

76

77

79

ll I..|I|II’||I I II III I ll

 

LIST OF FIGURES--Continued

FIGURE ‘ * Page
16. C/Co vs time for the reaction: 2K+Anr + 2EtOH

THF AhH2 + An + 2K EtO . . . . . . . . . . . . . . 80

17. Log-log plot of f obs XE-fcalc . . . . . . . . . . . 88

18. C/CO vs time for the reaction of K ”An with EtOH in
THF0 at _three different values of [EtOH]O . . . . . 9O

19. Log-log plot of k 5 (second- order) vs [ROHJ/[An]

for the reaction 8f K+ An with i- -PrOH, H20, t- BuOH,
MeOH ' EtOH in THF 0 O O O O O O O O O O O O O O O O 9 3

viii

a

‘AY‘F‘
.v-u|

c}

IF C
LI.
A:

a:
H u

I. INTRODUCTION

An aromatic molecule Ar may accept an extra electron to
form a new species which we will denote by Ar?) The negative
sign in this symbol represents the negative charge of the
aromatic ion; the dot shows that the new species possesses an
odd number of electrons, and thus, a radical nature. Hence,
we will reserve for the new species ArT'the term aromatic
radical anion.

Aromatic radical anions are formed by three common
methods as follows:

a) reduction of the parent aromatic hydrocarbon gig

reaction with alkali metals;

b) electrolytic reduction of the aromatic molecule;

c) reduction by solvated electrons produced by pulse-

radiolysis.

Aromatic radical anions are paramagnetic, conducting and
strongly colored species. They are good reducing agents.
They react with air and with proton donors. (The reader is
referred to the Historical section for more details on the
properties of Ar? and for references.)

The protonation of aromatic radical anions by alcohols

has the following stoichiometry:

I

'3I.

, "fly
' r vu.uv—vv

where

""19
v.3-

.a
U)
'h

‘u \
cue :4
“f\

--u.. .

2Ar"’ + 2ROH -——> ArH2 + Ar + 2R0-

where ArH2 is the dihydroproduct of the parent aromatic mole-
cule Ar. Paul, Lipkin and Weissman assumed that this reaction
is first-order in ArT.’ Pulse-radiolysis studies by Dorfman
gt El' of a series of aromatic compounds in pure alcohols and
in mixtures of ethanol and other solvents did show first—order
behavior in Arr. However, when Minnich used the stOpped-flow
technique to study the protonation of sodium and potassium
anthracenide with various alcohols and water in tetrahydro-
furan (THF), he found that the reaction was second-order in
the Ar? concentration and of low order in the ROH concentra—
tion. This second-order behavior was consistent with the
existence of a quadruple ion-cluster (ArT',M+)2 as the pro-
tonated intermediate Species. However, Minnich did not vary
the concentrations of the various alcohols and the anthracene
over wide enough ranges to test completely the effect of An
and ROH concentration on the protonation rate. At about the
same time, Levin, Sutphen and Szwarc, also using the stOpped-
flow technique, observed a second-order decay of the absorbance
of sodium perylenide (Na+PeT) in its reaction with various
alcohols in THF. They also found the reaction to be of inverse
first-order in the perylene (Pe) concentration, consistent
with the protonation of the ion-paired perylene dianion

= + . . .
(Pe ,2Na ) as the intermediate Spec1es.

SI

H"

q~w uflu -H
up. qnfi I ‘
A: mg by
L; ~.. Lu

“I.
‘hb

F: at C
.I. . .

Rf...” ~a
vv..,.,‘

In order to investigate further the second—order be—
havior in the protonation reaction of aromatic radical anions
with alcohols in ethereal solvents, we decided to extend the
work of Minnich. However, instead of using various aromatic
radical anions and various alcohols, we chose only a single
system for study. The reaction of potassium anthracenide
with ethanol in THF was studied over a wide range of concen-
trations of EtOH and An. We found the same second-order
behavior as did Minnich. But we also found that neither the
quadruple ion-cluster nor the anthracene dianion mechanism
holds over a wide concentration range of EtOH and An.

At the end of this thesis, several mechanisms are pro-
posed which are consistent with our experimental data, backed
up by some of Minnich's data which have been re-analyzed by
us. We also used dicyclohexyl-lB-crown-G (crown), a good
complexing agent for alkali metal cations, to test the effect
of ion—pairing on the protonation rate. Finally, the reader
will find at the very end of this thesis some suggestions for

further work in this unpredictable but fascinating field.

n

.—

E

San

5.x»

II . HISTORICAL

The Historical section of this rapidly expanding subject
will be divided into two main parts: early studies and studies
of kinetics. The first part deals with the general properties
of the aromatic radical anions, with emphasis on ion-pairing
because, as will be shown later, ion-pairing drastically af-
fects the protonation rate. The second part presents all of
the studies which have dealt with protonation kinetics, from
the early studies of Paul, Lipkin and Weissman to the most
recent paper of Szwarc and co-workers which is about to be

published.

2.1--Early Studies

 

2.1-l-—Production of Ar? via Reduction with
Alkali Metals

 

'It has been known for a long time that aromatic hydro-
carbons can react with alkali metals. The first comprehen-
sive study on such a reaction can be attributed to Schlenk
and co-workers (1) who, in 1914, mixed sodium and anthracene
in diethyl ether solutions. They reported the formation of two
distinct compounds: a one-to-one and a two-to-one adduct
(respectively sodium anthracenide and disodium anthracene).

Later, Scott, Walker and Hansley (2) found that the reaction

“L,

.h‘h

0‘.

“L
Q

h V

\‘w\

u A I

v‘ 1‘

of aromatic hydrocarbons with sodium is greatly facilitated by
using dimethyl ether as the solvent in place of diethyl ether.
The sodium naphthalenide solutions are dark green in color.
These solutions are also electrically conducting. Two classes
of reaction of sodium naphthalenide were reported:

a) reversal of the reduction step by reaction with
mercury, oxygen or benzyl chloride, in which the
naphthalene is recovered unchanged;

b) an irreversible reaction with water, alcohols and
other proton donors, in which the naphthalene is
reduced to dihydronaphthalene or its derivatives
depending on the reagent used. This type of reaction
was represented by the following equation:

. . + + V
CIOHB Na2 C10H8 ZRH ——>-C10H8 CIOHIO + 2NaR (1)

in which R is the conjugate base.

2.1.2--Studies of Optical Spectra

Paul, Lipkin and Weissman (3) studied the optical spectra,
the stoichiometry and the electrical conductance of the
products of the reaction of metallic sodium with various
aromatic hydrocarbons in tetrahydrofuran (THF). All of these
reaction products were reported to have deep colors; for
example, green for sodium naphthalenide, brilliant blue for
sodium anthracenide, etc. Studies of the Optical spectra of
both the mono- and di-anions of the products formed by the

reaction of numerous aromatic hydrocarbons with Li, Na and K

.‘a

h ..
A:

I .

in 1,2-dimethoxyethane (DME) and in THF were also reported
(4,5). The polarizations of the electronic transitions were

studied by Hoijtink et_al. (6,7).

2.1.3--Electrochemical Studies

 

In addition to their production by reduction of the
aromatic molecules by alkali metals, aromatic radical anions
can be produced gig electrochemical reduction of the
parent aromatic compounds. The pioneering work of Laitinen
and Wawzonek (8) showed that polarographic reductions of

aromatic hydrocarbons (Ar) follow the scheme:

Ar + e- Z Arr (2a)
ArT + e- -——+- Ar= (2b)

followed by:
Ar= + 2H20 -——>- ArH2 + 20H- (2c)

since the experiments were conducted in dioxane-water mixtures.
Later, Maccoll (9) and Hoijtink and Van Schooten (10) (using
zero-order molecular orbital calculations) correlated the
half-wave potentials s%_ with the energies of the lowest
unoccupied orbitals, E. They found a linear relationship be-
tween 8% and E. Hoijtink and Van Schooten (10) also noted
another possible mechanism for the protonation of aromatic

hydrocarbons as follows:

c.-

+ ' -——->- "’

Ar e Ar (3a)
Arr + H20 ———9-ArH' + OH— (3b)
ArH‘ + e- ——> ArH' (3c)
ArH- + H20 ——>-ArH2 + OH- (3d)

Also, polarographic studies with varying dioxane-water
mixtures and with varying HI concentrations (11) showed one
two-electron wave in the case of high water concentrations
(or upon the addition of HI), and two one-electron waves for
low concentrations of proton-donors. These data indicated
that high concentrations of proton donors favor step (3b)
while step (2b) is favored at low concentrations. Finally,
the species (ArH‘) seems to protonate more readily than does
the parent radical anion Arr.

In later studies (12), the biphenylide anion (Biph?)
produced tut the reaction of sodium with biphenyl was used to
reduce other aromatic compounds. Potentiometric titrations

showed that dianions could also be produced, but in two steps:

Biphr + Ar ———>- Biph + Ar? (4a)

Biphr + Ar? ———>— Biph + Ar: (4b)
‘_._._

20104--EOSORI StUdieS

 

Aromatic radical anions (Arr) are paramagnetic since
they have an odd number of electrons. Thus one can use the

sensitive electron Spin resonance technique to study some of

 

their pro;
from aroma
Early ESR
due to the
:xclei in
splits the
spli ting
ing- the od
35R studie

ESR sperm»

V»;

 

Note

‘P
“01; d6”
n»:

which COr

U

s, tne

 

their pr0perties. On the other hand, the dianions formed

from aromatic hydrocarbons are generally diamagnetic (13).
Early ESR studies of Ar? (14,15) revealed hyperfine Structure
due to the interaction of the odd electron with the magnetic
nuclei in the investigated species. Each magnetic nucleus
Splits the ESR signal, and for a given type of nucleus, the
splitting constant is proportional to the probability of find-
ing the odd electron at that nucleus. In one of their early
ESR studies, DeBoer and Weissman (16) compared the observed
ESR spectra of numerous aromatic radical anions with those
calculated on the assumption of a linear relationship between
hyperfine coupling constant with a proton and the N-Spin dens-
ity on the adjacent carbon. Good agreement between observed
and calculated spectra was reported. For the anthracene mono-

anion the distribution of the spin densities is as follows:

-—-— 0.193
-———- 0.004
———— 0.096
‘———— 0.048

Note that while certain positions are favored, the elec-
tron density is distributed over the entire aromatic system,
which consequently makes the species very stable once formed.

Yet, the extra electron can jump back and forth from one

aromatic molecule to another, as shown by Ward and Weissman
(17). These authors observed a broadening of the ESR spectrum
of naphthalenide (Nap?) when they added extra naphthalene
(Nap) to the Nap? solutions. They concluded that the broaden—

ing resulted from the exchange reaction:

Nap + NapT' ———)- Nap? + Nap (5)

Second-order rate constants from 1 x 107 to l x 109 M-lsec-l

were reported, depending on the solvent and on the positive
ion used.
ESR studies also reveal another interesting phenomenon;

ion-pairing of the aromatic mono-anion species with cations.

2.1.5--Ion-pairing‘Studies

 

AS described above, the ESR technique can reveal the
hyperfine interactions of an aromatic radical anion. For
example, the ESR hyperfine structure of a free naphthalene
radical anion is due to the magnetic interaction of the odd
electron with two sets, a and B, of four equivalent protons.
These split the ESR signal into 25 lines. But the Spectrum
of the sodium naphthalenide ion-pair is more complex. The
odd electron has a finite probability density at the 23Na
nucleus. This nucleus has spin 3/2 so that each of the orig»
inal 25 lines is split further into a quadruplet. Atherton
and Weissman (18) reported this additional Splitting of the

spectrum of sodium naphthalenide in THF, and postulated the

a I “.0 FAR

 

..
nfiqfiev V
:vh&w1 5:
R“ ‘- V. .-
.:.'lte- C-

516 ’v’ariou

'G'V"
rt ~
~.:.\" we: 0

£5

 

\
f‘
th‘A..
“a
1
a"
‘8 H
ya
U
“Vduh
b
.'
.fzt
N

 

10

existence of an ion-pairing equilibrium:

+ .— + -.- -.
Na Nap ———9— Na + Nap (6;

They also went further and suggested that the existence of the
sodium Splitting would require the Na+ion to lie above the
center of either of the two benzene rings and at a distance
of about 2.5 A. If the NaIion Shifted from this equilibrium
position, the splitting would decrease. Later experiments
showed that this pioneering description of ion-pairing was
not adequate to describe all of the phenomena which can occur.
Other studies showed that aromatic radical anions can exhibit
two or more forms of ion-pairing with equilibria connecting
the various forms. Two general classes of equilibria were
reported:

- a dissociative equilibrium:

M+Ar"' Z M+ + Ar"
— A non-dissociative equilibrium:

+ 1- + 1" ‘
-——a> A. 8)
(M Ar )a E (M r )b I _

In the following sections, we will give evidence for both

classes of equilibria.

2.1.5.1--Evidence for a Dissociative Equilibrium

 

Conductance measurements (19) of the solutions of several
Arr produced by reduction with alkali metals in THF showed

. . +
that for pota351um anthracenide (K An?) the speCific conductance

.«l‘m. '
.I.

I- — ”'V"""Y"-‘IIV1'—r—"*

r
I

 

increased
and then d
behavior I:
crease in
solvent vi

be the res

Ii“. re An?
and Li ant
continues

stantial 3
he Optica

absorption

 

 

11

increased with temperature from -75°C to a maximum at 0°C
and then decreased slightly from 0°C to 25°C. While the
behavior below 0°C was interpreted as due to the normal in—
crease in conductivity of the ions with a decrease in the
solvent viscosity, the behavior above 0°C was postulated to

be the result of a shift to the left of the equilibrium:
+ r + T
g
M An M + An (9)

where An? represents the anthracenide mono-anion. For Na

and Li anthracenide solutions, the specific conductance
continues to increase from -75°C to 25°C, indicating no sub-
stantial shift of equilibrium (9) to the left. Studies of
the optical Spectra of these same solutions showed that the
absorption peaks occurred at higher energies for the smaller
alkali cations. It seems that low temperatures, large anions
and small cations enhance the dissociation of the ionmpairs
into free ions.

ESR studies of solutions of sodium anthracenide in
methyl-tetrahydrofuran (MTHF) at low temperatures (20) Showed
a superposition of two distinct ESR spectra. Changes in anion
concentrations led to changes in the relative intensities of
the two spectra. These results indicated the existence of a
dissociative equilibrium. When Dodson and Reddoch (21)
extended the ESR study of the naphthalenide solutions to in-
clude all of the alkali metals in THF and DME, they found the

same superposition of spectra at room temperature. In most

 

 

.
a
“SW-Fe?

‘VH
.. LI

3‘ .
u gl

§
'Izv.’

"‘11

my
5“e

5|;
D's”

't)

(I)

LICI

AC

..a

\‘L

an I

V“ ‘9
“e

‘7

n {‘1

12

cases, they found the same concentration dependence of the ESR
spectra. The spectrum of Li naphthalenide in DME showed
neither metal splitting nor concentration dependence. This
would be the case if the dissociative equilibrium (7) was
shifted far to the right. Slates and Szwarc (22), by studying
the conductance of the sodium salts of numerous aromatic
radical anions in THF, were able to determine the dissociation
constants of the respective ion pairs. These constants in-
creased with the size of the anion.
2.1.5.2--Evidence for a Non-dissociative
Equilibrium

Hogen—Esch and Smid (23) reported that the Optical spec-
trum of sodium fluorenyl (Na+FlT) in THF contained a number
of peaks whose relative intensities varied reversibly with
variations in the temperature. For example, when the tempera-
ture was changed from +25°C to -50°C, there was growth of a
peak at 356 nm simultaneously with decay of the peak at 373 nm.
Since neither dilution of the Na+FlT solution nor the addition
of a common ion affected this interconversion, the authors

eliminated the two following equilibria:

+ .- + -.-
(Na Fl )2 ———9' 2(Na Fl ) IIO)

.__4, (ll)

Na+FlT +— Na+ + F1..-
On the other hand, when they studied the spectra of lithium

.fihmorenyl (Li+FlT) in various solvents, they observed the

folloWin‘.

These res

non-dies:

 

 

13

following:
- only a single peak at 346 nm in dioxane at 25°C;
- only one peak at 373 nm in DME;
- both peaks in THF.
These results led the authors to suggest the existence of a

non-dissociative equilibrium:

.+ a .+
Fl?, L]. ————r Fl H L1 (12)

. + = , -
The symbol F17, L1 represents a "contact ion-pair", conSidered
to be a pair of ions held together by coulombic attraction.

+ I
'8' . . . . .
II L1 is a "solvent-separated ion—pair" in

The species Fl
which the cation and the anion are separated by at least one
solvent molecule. In order to further test for the existence
of such Species, the authors added dimethyl sulfoxide (DMSO),
a very strong solvating agent, to the Li+FlT solution in

dioxane. As expected, the 346 nm peak was converted to the

peak at 373 nm by the addition of DMSO:

.+ .+ .-
Fi", L1 Bis—9+ FIT H L1 (13.)

The results therefore indicated that the contact ion-pair was
converted to the solvent-separated pair. A graph of log
(A346/A373) gs. log (DMSO) gave a straight line with slope

1.15. Other studies (24,25) have also shown the existence of

non-dissociative equilibria.

fl

. .....'....u

1;_'- - v

"99%....

 

 

 

14

2.1.5.3 --Other Equilibria

 

Hogen-Esch and Smid (26) found that the combination of

both equilibria described above

Arr; M+ ———+— Ar? I] M+
REM /¢§/ (14)
+ -.-
M +Ar

fit their spectrosc0pic and conductance data. Yet another

set of equilibria was proposed by Hirota (27):

(Arr, M+)a : (Arr, M+>b —->—Ar"' || M+
N + ‘H, .— / (15)
M + Ar

+ . .
where (Ar?) M+)a and (Arr, M ) are both contact ion-pairs

b
but are solvated differently.
Another model of ion-pairing was prOposed by Chang,
Slates and Szwarc (28). These authors described ion-pairing
by a potential well whose shape is temperature and solvent
dependent. Presumably, the potential energy of the two ions
increases as they are pulled away from their contact position.
Initiallyg their mutualuattraction is not attenuated by the
solvent because the gap between them is too small to accommo—
date a solvent molecule. But when the gap is large enough to
allow a polar solvent molecule to squeeze in, the energy of
the system decreases. Further stretching of the pair again

increases the potential energy until a limiting value (corre-

Sponding to infinite separation of the ions) is reached.

Q

‘

We a'
81’. none C

c!
i
o

 

.‘I. .C .1 .r L r1. v‘
.th S .l S .l 3 C;
t nu VA d l l
O t .1 e .1 i t n...
b v .e L l a LI. nu a l
«a t Fl». a wig. r ‘5 t
We L . . V e L . a
”.7. 7.. 2. Wu . . an n .
.3 C k . 3 3 F. o. S
2. S .L ,1 5 q . C u
InuIIQIWjIIIJII. .
an...» .1 JIPIJ

 

15

2.1.5.4r-Conclusions Regarding Ionepairisg

We already have seen more than one model of ion-pairing.
Yet none of them is completely satisfactory. We can only
say at this time, by restricting the scope of ion-pairing to
solutions of aromatic radical anions in ethereal solvents,
that it is probably correct to admit the existence of contact
ion-pairs and solvent-separated ion—pairs along with the free,
solvated ions. These three forms of ion-pairs are probably
in equilibrium. Depending on the size of the ion, the con-
centration, the solvent and the temperature, one form is
usually predominant over the others. We will see later that
ion-pairing can drastically affect the kinetics of protonation

of aromatic radical anions in ethereal solvents.

2.2--Kinetics Studies

 

2.2.l--The Mechanism of Paul, Lipkin and Weissman

 

The stoichiometry of protonation of aromatic radical
anions (Ar?) in ethereal solvents has been shown (2,29,30,31)

to be:
2 Ar" + 2 ROH -——->—Ar + ArH2 + 2 R0- (16)

where Aer is the dihydro-adduct of the parent aromatic hydro-
carbon.

Paul, Lipkin and Weissman (3), basing their arguments
only on the stoichiometry of the reaction, proposed a mechan-

ism for the reaction of naphthalene radical anions (Nap?) with

C.

either C
an alccho

 

 

I. I L . r. L . a E A“ n . s .
C. S t t t .I T. 5. l\ C. C
G 4 m f O a K a II n
\rL In». 0 cl . Vs 2H \‘
L. y . . C. n . nu... .
e I I} l 2 x I S
r. L . a S .(t . a e.
A: I.— MIW .. e “e. 2 s S u .
.hu .. u . I
”I Tn .n a W . .I M «4 MN.» MW...

r... ‘AIK'AI- N.
. .nH,

1IH.I. «44h Ir. .-
> D

16

either CO2 or various proton donors. For the reaction with

an alcohol(ROH), the mechanism was:

Nap? + ROH -——+—NapH° + RO- (17a)
NapH' + NapT ———>-NapH- + Nap (17b)
NapH- + ROH ———>-NapH2 + R0- (170)

where NapH is the dihydro-adduct of the naphthalene molecule.

2
With this mechanism, one would expect to find a first-order
decay of the radical anion with the following assumptions:
- the electron transfer step (17b) is fast compared
to the proton transfer step;
- the second proton transfer step (17c) to NapH_ is
fast compared to the first one (l7a).

This pioneering mechanism has come to be known as the Paul,

Lipkin and Weissman mechanism.

2.2.2--Kinetics Studies in Ammonia

 

Krapcho and Bothner-By (32,33) conducted a study of the
metal-alcohol reduction of benzene and some substituted ben—
zenes in ammonia in the following way: they successively added
an aromatic hydrocarbon, an alkali metal and an alcohol to
liquid ammonia at -34°C. Since the reactions were slow (half-
times % 100 sec), they could extract aliquots periodically,
quench them and analyze them. From this product analysis, the

authors deduced the stoichiometry:

..

'r '9 awn,

.AJu‘ bbU-H
.

:' ‘ ;
.1 them

 

.L a

(I)

n
V

:r

J
u

>
»
r4 .

 

 

17

H H
/ \/
©+2M+2R0H-—> @ +2ROM (18)
/‘\
H H

Although benzene was present at the end of the reaction, the
investigators considered it to be unreacted starting material.

They found that a third-order rate law:

d[Ar]
dt

 

= k [Ar][M][ROH] (19)

fit their experimental data well. They prOposed the following

mechanism:
M + solvent ———+-M+(solv.)...e-(solv.) (20a)
+ " + T
M (solv.)...e (solv.) + Ar -——>»M (solv.)...Ar (solv.)
«e-—-

(20b)

+ 'r .
M (solv.)...Ar (solv.) + ROH ———)>ArH (solv.) + ROM(solv.)
(20c)

M+(solv.)...e-(solv.) + ArH'(solv.) ———9'M+(solv.)...
ArH_(solv.)
(20d)
M+(solv.)...ArH-(solv.) + ROH -——)--ArH2 + ROM(solv.)
(20e)

where M+(solv.), e-(solv.), etc. ... designate the solvated
species. In addition to the first-order dependence on the
metal concentration, the following observations were made:

- the reduction reaction proceeded more slowly with Na

than with Li;

 

 

 

 

' l
”CF" I! V M» .
"“'"~at“DL: .

 

KLT‘C‘.

 

 

18

- the reaction rate was not affected by the addition

of NaBr;

- The reaction rate increased upon addition of LiBr.

The authors explained the effect of LiBr by proposing the

equilibrium:
. + - + 1-
L1 (solv.)... Br (solv.) + Na (solv.)...Ar (solv.) ————e»
<+—
Na+(solv.)...Br_(solv.) + Li+(solv.)...ArT(solv.) (21)

Later, Jacobus and Eastham (34) reported a fourth-order

rate law for the same system and proposed the following

mechanism:

H + - . H '

C6 6 e ———+- C6 6 (22a)

' H -+ o H + -

C6 6 ROH ———>’ H C6 6 RO (22b)
. + -‘ : -

H C6H6 e ———>- H C6H6 (22c)

- slow -
H°C6H6 ———> C6H7 (22d)
- fast - -
C6H7 + ROH ——-> C6H8 + RO (22e)

In this mechanism, the alkoxide RO_ plays an important role.
In fact, inhibition of the reaction upon addition of R0- was
reported. Thus step (22b) was postulated as a reversible
step. Also, step (22d) was considered to be the rate deter-
mining step.

In addition, Kelly gt El. (35) have reported some

kinetics data for the reaction of sodium with ethanol in liquid

 

the react:

afunctlo:

‘

1

reil‘ t

0"

‘1

0

::.e rate

‘_

.I

'.

" J
'33:?

.. . .
m—w-na-‘v

I f).

——.‘

V
"Fn'ab flrv
ugy‘ ‘1“

1"“..n- I v ~

7“?" Fffirao.
9.1‘ a,

'- uy

 

.,
H I-
I” l, ‘

 

 

 

l9

ammonia at —33.4°C. These authors monitored the progress of
the reaction by measuring the amount of hydrogen evolved as
a function of time. Later, Jolly (36) showed that the data
of Kelly gt El. were at least qualitatively in agreement with

the rate law:

kl[EtOH][e-]
= rt :‘ _ (23)
(kZ/k3)[EtO ]+[e ]

d[EtOH]
dt

 

which corresponds to a mechanism involving a steady-state
concentration of ammonium ion:

k

1
+ -——q> + -
EtOH NH3 é——— NH + EtO (24a)
k2 4
NH + + ' —E3+- NH + l H (24b)
4 e 3 2 2

In View of the mechanism suggested above for the sodium-
ethanol reaction, Jolly (36) showed that the third-order rate
law (Equation 19) obtained by Krapcho and Bothner-By was also
consistent with the mechanism:

k

 

+ _
ROH + NH ——lé- NH + RO (25a)
3 T- 4
2
Ar + e- BEE‘ Ar? K = [Ari] (25b)
‘r-—-—- [Arlle‘]
+ '1- k3
NH4 + Ar -——%> NH3 + ArH' (25c)
ArH' + e- + NH + —£3§£>- ArH + NH (25d)

4 2 3

 

kinetics s

 

PA.‘,‘.“:/'\F'R
Dwauulx...)

.
”"VMVAd-a

"t
Vvuuh. _ .‘b

its rate 1
dEe

C

i'iEIe e-

twp :n N
nn. ‘V k .'

 

 

20

In fact, assuming that a steady-state concentration of the

ammonium ion was achieved, this mechanism yields the rate law:

k K[ROH][Ar][e-]

1k3
k2[RO ] + k3K[Ar][e 1

d[Ar] =
dt

(26)

 

When K is small and when the metal alkoxide is only slightly

soluble, this rate law reduces to that obtained by Krapcho

'v: awn-'— J
, .7!“

and Bothner-By (Equation 19).

hr. {2

J‘Lkr'

1
LL.

Very recently, Kahn and Dewald (37) have undertaken a

 

kinetics study of the reduction of benzene by sodium-ammonia
solutions at —45°C, using ethanol as the proton source. Their
conductance data showed that the reduction reactions followed

the rate law:

d[eam]

dt

k[EtOH][eam] (27)

where e;m is the solvated electron in ammonia. They proposed

the following mechanism:

1 + -
-———»
EtOH + NH3 NH4 + EtO (28a)

eam + NH4 —-)' NH4 (28b)

+
C6H6 NH4 ———9- C6H7 + NH3 (28c)

+ +
C6H7 NH4 —* C6H8 NH3 (28d)

BY applying the steady-state approximation to the ammonium

 

 

 

radical c

1315 rate

 

 

 

21

radical concentration, the rate law becomes:

_’ d[eam] = k3kl[EtOH][eam][NH3] (29)
at k2[EtO-]

 

When the reaction mixture is saturated with sodium ethoxide,

 

this rate law yields Equation 27 in which: T:
E‘ L
k k [NH ] j
k = 3 1 _3 (3o)
k2[EtO ]

 

2.2.3--Pulse-radiolysis Studies

Besides their production by electrochemical reduction
and by reduction with alkali metals, aromatic radical anions
can also be generated gig reduction of the parent aromatic
hydrocarbon by solvated electrons produced by ionizing radia-
tion. In the pulse-radiolysis technique, ionization is
produced by a burst of high-energy electrons from a linear
accelerator. In some early studies, aromatic hydrocarbons were
dissolved in pure solvents such as cyclohexane (38) and
ethanol (39), bombarded with high energy electrons and then
observed spectroscopically.' Absorbing species were formed which
then decayed completely within a few microseconds. The absorp-
tion spectra of these species often corresponded very closely
to those obtained when the same aromatic hydrocarbons were
reduced with alkali metals. Thus these transient Species could
be identified as the well known aromatic radical anions pro-

duced this time by the reaction:

 

 

{note how
Rate cons
30:109
39,40).

::lecules
In pure a

:” follow

C

.ae autho:

alcohols

The raze
filth: Ce."

'
.ange fr:

(5
K
d

«1x It;

 

 

 

 

 

22

Ar + e-(solv.) ———+-ArT (31)

(note however that radical cations have similar spectra).
Rate constants for this formation reaction were found to be
about lOQM-lsecml depending on the aromatic hydrocarbon used
(39,40). In some solvents, triplet states of the aromatic
molecules and aromatic radical cations were formed (39,41).

In pure alcohols, the aromatic radical anions were reported

to follow a fast first-order decay:

d [Ar-r]

dt = k' [Arr] (32)

The authors (39,40) suggested a proton transfer from the

alcohols as follows:

T

Ar + ROH ——>ArH’ + R0. (33)

The rate constants for this reaction were determined for
anthracene, biphenyl and terphenyl in four alcohols. They

range from 2 x 10‘2Mnlsec“1 for terphenyl in ethanol to

8.1 x 104M'lsec'l

for anthracene in methanol. For a given
anion, the more acidic alcohols gave the higher protonation
rates.

In a recent article (40), Dorfman presented an overall
picture of the mechanism of the formation of aromatic radical
ions by pulserradiolysis techniques.I In polar protic solvents,
such as the aliphatic alcohols, only the radical anion (Arr)

is formed. In certain chlorinated hydrocarbons, on the other

hand, only the radical cation (Ar'+) is produced. In non-polar

 

+' «
1'0th .

F
U
u

..4 Fae

:-
‘ul 5'“

In

“e actil

uai

 

“big e] en;
...v 4. V

'1; Then a se

(‘5'.

1
«or»: '

 

 

 

 

23

aprotic solvents such as cyclohexane, both the radical anion
and the radical cation are formed.
In the alphatic alcohols, the proposed mechanism for

the action of ionizing radiation was:

+ .—
ROH W4- ROH + e (34)

v‘
a
v.

The electrons produced in this step are rapidly solvated.

Then a sequence of reactions can occur as follows:

 

e-(solv.) + Ar ———>~ Ar? (31) ’
Ar" + ROH —-.> ArH' + m" (33)
+ + -
ROH + ROH ———-> ROH + R0 (35)
4——— 2
Ar" + ROH2+——->- ArH' + ROH (36)

. + e .
The solvent counter ion ROH rapidly forms the alkyloxonium

ion (ROH +) (Equation 35) which then reacts very rapidly with

2
Ar? (Equation 36).

Pulse radiolysis in mixed solvents (40,42) gave inter-
esting information about solvent effects on the proton-
transfer rates. In fact, irradiation of biphenyl solutions
in ethylene diamine and in triethylamine gave a lifetime of
biphenylide which was about 100 times greater than in ethanol.
When the investigators added ethanol to ethylenediamine as

solvent, they found that the lifetime of the biphenylide ion

remains virtually unchanged up to a concentration of about

 

65 mol %
'eactive

‘*e deer

vot

r:::;3n k
ZCTF'ors,

v...

‘ ' th‘ a
2'" “VOL

‘1'“.
‘

2.2.4-4,

 

Usir

.
..
‘ £"'-r~ L

V‘V

'32,-”

““41; z‘

u
I
“‘9 fit... .
. .
i¢«.‘t”‘ ‘
.

I” that

,.
“Ruler.

":3: the

 

3‘ 'L
1» the E
:‘l'e fl'c

 

 

 

24

65 mol % ethanol. An extensive hydrogen bonding of the
reactive alcohol proton to the amine was suggested to explain
the decrease in the rate constant. On the other hand,
weakening of the hydrogen bonded solvent structure of the
alcohol would be expected for non-polar hydrocarbon solvents.
Irradiation of a solution of biphenyl in ethanol-cyclohexane
mixtures showed that the first-Order decay constant actually

increased over the value in pure ethanol. This enhancement

 

of the protonation rate was interpreted as due to the dis-
ruption by the cyclohexane of an equilibrium system of
monomers, dimers and polymers of ethanol units held together

by hydrogen bonds.

2.2.4--Kinetics Study_Usinq_ESR and Polarography

Using a combination of ESR and polarographic techniques,
Umemoto (43) studied the protonation by water of anthracene
radical anions which were electrochemically generated in N,N'-
dimethylformamide (DMF). It has been mentioned earlier (9,
10) that d.c. polarograms showed two reversible one-electron
additions to the anthracene molecule in dry DMF as solvent.
Upon the addition of water, the first wave increased in height
at the expense of the second.’ Also, the reversible peak of
the first wave in the a.c. polarogram decreased remarkably
upon the addition of a small amount of water. Umemoto pro-

posed the following mechanism to explain these results:

(In-"Ad

 

1‘ n
mere ml

In
. I, .
:eca‘; 1.

"atom

3““:ce th

 

r
H
:15 ,
fi. .

 

25

T k' 0 _
An + H20 ———9- AnH + OH (37a)
AH°+A"-——>-AnH'+A 7
n n n (3 b)
- + O + -
AnH H2 -—+- AnH2 OH (37C)

where AnH2 is the dihydroproduct of anthracene. A first-order
decay in AnT was derived from this mechanism by assuming that
step (37a) is irreversible and that the steady-state concen-

tration for the species AnH' was achieved. This gives:

_ d [An']

dt = 2k'[An ][H20] (38)

Since the protonation reaction was relatively slow, the decay
of the radical anion in the presence of water could be studied
by observing changes in the ESR signal intensity with time.

A first-order curve was observed up to 6% H20. However, the
first-order rate constant did not increase linearly with water
concentration as predicted by Equation 38. This result was
interpreted as due to change in the solvent composition in

the presence of high water concentrations.

Umemoto also found that the benzophenone radical anion
followed a mixed first- and second-order decay.
2.2.5--Combined Polarography-ESR-UV Absorption

Spectrosc0py_8tudies
Recently, Hayano and Fujihira (44,45) studied the pro-

tonation of numerous aromatic radical anions in DMF-water

mixtures. They used three methods, polarography, ESR, and

3 CV absorp‘
Arr conce

l. currents

1 intensity

sicn cur:

 

i Of the as
‘ it be due
I Spin Exc}

? weenie
t: __:._3

tie 55st

 

Hayano a:

1,, .
‘taln' a

fast! th

 

.

4" “”5 ‘

 

26

UV absorption spectroscopy, to simultaneously measure the

Ar' concentration. A linear relationship between the diffusion
currents and the absorbance was found. But the ESR signal
intensity showed no linear relationship with either the diffu-
sion current or the absorbance. This non-linear dependence

of the ESR signal intensity on the Ar? concentration was found
to be due to the broadening caused by the rapid intermolecular
spin exchange, even when the concentration of the.parent
molecule, Ar, was constant. The change with time of the visi-
ble absorption band of Ar? followed a first-order rate law.

Hayano and Fujihira proposed the following mechanism:

Arf + H20 —3—+- ArH' + OH‘ (39a)
ArH’ + Ar" ——ar ArH- + Ar (39b)
_ kl _ .
ArH + H20 —->- ArH2 + OH (39c)

which is identical to the Umemoto mechanism (Equations 37a,b,c),
Again, assuming that step (39a) is slow, but step (39b) is

fast, the rate equation:

d[ArT]

dt = 2k [H20][Ar J (40)

was derived, which is identical to Equation 38.

Once again, the protonation rate by water was found to be
greatly accelerated by increasing the water concentrations in
DMF-water mixtures. The reason for this behavior was inter—

preted (45) as follows: the negative charge is more localized

 

 

be inte

the cha

P561150

27

on a certain atom in the activated complex than in the orig-
inal aromatic radical anions (ArT). Therefore, the activated
complex is more stabilized than is the.original Arr by
changing the solvent from aprotic to protic. The reaction can
be interpreted as a transfer process between the no-bond and
the charge-transfer structures. The activated complex is a
charge-transfer complex in which the no-bond and charge-
transfer structures contribute almost equally to the resonance.

2.2.6--Kinetics Studies Using Polarography
and Stopped-flow

 

 

Another way of investigating the same problem as described
in sections 2.2.4 and 2.2.5 was undertaken by Fujihira, Suzuki
and Hayano (46). These authors used electrochemical reduction
to generate the radical anions of anthracene and 1,2—benzan-
thracene. They used the stopped-flow technique to study the
protonation with H20 in DMF.

Direct measurements by the stopped-flow method gave a

pseudo first-order decay:

_ d[ArT]
dt

-2k[H20][Ar?] (40)

= -kapp [Ar ] (41)

which was in agreement with mechanism (39).
The results obtained from the polarographic method were
analyzed by using rigorous mathematical treatments for the two

possible mechanisms (3) and (39). The values of kapp obtained

 

 

28

by the application of mechanism (39) to the polarographic
data agreed well with those obtained by the stopped-flow

technique.

2.2.7--Stopped-flow Studies

 

The stopped-flow technique is suitable for the study of
fast protonation reactions of aromatic radical anions with
alcohols in ethereal solvents. In fact, this technique has
been adopted recently by many investigators including Minnich
and Dye (47,48), Bank and Bockrath (49,50) and Szwarc and co-

workers (51,52).

2.2.7.l--The Work of Minnich

 

Minnich studied the reactions of the radical anions of
anthracene (An) and terphenyl (Ter) with alcohols and with
water in two ethereal solvents, THF and DME (47,48). The
radical anions were produced by reduction of the parent mole-
cules with Na and K. The reaction rates were followed Spectro-
photometrically by using a scanning stOpped—flow apparatus.

In THF, the reactions of K+AnT with EtOH, MeOH, t-BuOH,

i-PrOH and H 0 were reported for the first time (47) to have a

2
second-order behavior in the absorbance. The order in ROH

was less than unity. Only limited ranges of the alcohol con-
centrations were studied. Also the concentration of anthracene
remained nearly the same in many experiments. Variation of the

alkoxide concentration (RO-) had no effect on the reaction

rates.

29

Similar results were obtained for the reactions in DME,
but the pseudo second-order rate constants were an order of
magnitude lower (kTHF W l x lO4M-lsec-l; kDME % l x 103M-lsec-1).

The reactions of Na+AnT with EtOH in THF were similar to
those of K+Anr in the same solvent. The reaction with H20 of
Na+AnT in THF was studied in two separate experiments. In one
case, pseudo second-order decay was observed, but in the other
case, a mixed pseudo first- and second-order was seen with a
large first-order contribution. Due to this inconsistency,
no quantitative results were obtained for this reaction.

Striking differences were reported for the reaction of
Na+AnT'with EtOH or with H20 in DME. The decay of Na+Anv
in DME was much slower than in THF and it followed a first-
order rate law in the absorbance. Also, the reaction of
K+TerT with EtOH in THF was found to be fast and first-order

in each reactant.

These results were consistent with the following prOposed

mechanism:
+ k'

+ + T +

An‘", M + An", M —k———» (An , M )2 (42a)

+ kw + ‘
(An', M ) + ROH ———-1——>- M AnH- + An + M+,Ro" (42b)
2 slow

M+AnH" + ROH ——:2—>- AnH + M+ R0- (42 )
fast 2 ' C

An" M+ + ROH -k—3——> M+ R0. + A H' 42d)

’ slow ' n (
An , M + AnH —————>- M AnH + An (42e)

'fast

30

If a steady-state concentration for the species (AnT,.M+)

 

 

 

2
was made, the general rate law could be derived:
I
d [Anrj M+] = 2k+ [An-c M+] 2
dt k: ’
l +
ki [ROH]
T +
+ k3[An , M ][ROH] (43)

In cases where contact ion-pair formation is favored (at
least at low values of [ROH]), the second-order process pre-
dominates (Equations 42a,b,c). When such ion-pairing is not

favored, the first-order process is seen (Equations 42d,e,c).

2.2.7.2--The Studies of Bank and Bockrath

 

For their first kinetics study with a stopped-flow device,
Bank and Bockrath (49) chose the reaction of sodium naph-
thalenide (Na+NapT) with H20 in THF. The naphthalene radical
anion was produced by the normal method of reduction of
naphthalene (Nap) with metallic sodium. The authors found that
the decay of Na+NapT followed a pseudo first-order rate law in

agreement with the Paul, Lipkin and Weissman mechanism (Equa-

tions l7a,b,c):

k
Nap? + H20 ——l+- NapH' + OH- (44a)
k2 _
NapH' + Nap? ———+- NapH + Nap (44b)
.. k3 ..
NapH + H20 ———+— NapH2 + OH (44c)

31

where NapH is the dihydroproduct of naphthalene. Again,

2
assuming a steady-state concentration of the species (NapH°),

the pseudo first-order rate law was derived:

_ Elia—9:1. = 2k [H o] [Nap-r] = 2k'[NapT] (45)
dt 1 2 l
where ki = k1[H201
The value of k1 at 20°C was found as:
kl = 1.01 x 104M‘1sec‘l,

Very recently, the same authors studied the kinetics of
the reaction of sodium anthracenide (Na+AnT) with H20 in
THF, DME and mixtures of these solvents (50). For this study,
instead of preparing the An? in the usual way, they used

naphthalenide (Nap?) (produced from reaction of naphthalene

with sodium metal) to reduce An to AnT :

1-

NapT + An ———+- An + Nap (46)

+—

They made the three following assumptions to validate this
method:
a) the electron transfer to An is kinetically faster
than the protonation of NapT;
b) the concentration of Na+NapT in equilibrium with
Na+AnT is kinetically insignificant;
c) the equilibrium concentration of the anthracene

dianion is also kinetically insignificant.

 

32

Once more, the reaction of Na+AnT with H20 in THF was re-
ported to be similar to the reaction of Na+NapT with H20 in
the same solvent. Thus, the Paul, Lipkin and Weissman mechan-
ism was apparently once more confirmed. However, Na+AnT was
reported to be considerably less reactive than Na+NapT
(k(Nap)/k(An) = 182). When sodium tetraphenylboron was added
to the stock solution of Na+AnT in DME, the observed rate
constant increased more than 20-fold. Further addition how-
ever did not have much effect on the reaction rate. The
authors presumed that the sodium ion solvates a water molecule
significantly better than the ether. The water molecule so
solvated is more "acidic" than either free or solvent—
solvated water. Note however that these experimental results
are in disagreement with the results of Szwarc when DME was

used as solvent and also with our results in THF (see later).

2.2.7.3--The Studies of Szwarc and Oo-workers

ifimakineticsstudy of Levin, Sutphen and Szwarc (51) in-
volved the reaction of sodium perylenide (Na+PeT) by alcohols
and by water in THF at 25°C. The mode of preparation of the
_perylene radical anion was the standard reduction of perylene
(Pe) with Na. The protonation reaction was shown to be second—
order in Na+PeT, first-order in ROH and inverse first-order
in Fe. The ion-pair of the perylenide dianion served as the

base according to the "dianion mechanism":

 

33

K

2Pe , Na --9- Pe , 2Na + Pe (47a)
= + - + - \
Pe , 2Na+ + ROH —E—>’ Na PeH + Na RO (47b)
_ + _
Na+PeH + ROH —£§§£>- PeH2 + Na RO (47c)

The rate expression was:

1- + R P N
_ d[Pe , Na ] = 2kK1[ OH][ e ' a 1 (48)

dt [Pe]

 

 

The relative values of k were found to increase with the
acidity of the alcohol used.

Very recently, Rainis, Tung and Szwarc (52) investigated
the protonation of sodium anthracenide (Na+AnT) with alcohols
and with water in DME, a stronger solvating agent than THF.
The following interesting results were reported:

First, the protonation by MeOH, EtOH, i—PrOH and H20 was
first-order in Na+AnT in agreement with the results of Minnich

and Dye (47,48), but mixed first- and second-order in ROH.

The authors proposed that the dimeric Species (ROH)2 formed by:

K

d
2ROH -——%- (ROH)2 (49)

participates in the protonation process. The reactions:

+

+ . -
A11", Na + ROH ——->AnH + RO , Na (50a)

or

1-

+ , — +
An , Na + (ROH)2 ———9-AnH + RO , Na + ROH (50b)

are rate determining.

 

 

34

Second, the protonation with t-BuOH was reported to

. + .
ve a second-order dependence in (Na An?) . A new mechanism

3 proposed to explain this result:

k k'
+ 'r + =
.An-r, Na ) -,-—*-f (An , Na ) -—£—>— (An , 2Na+; An) (51a)
kb 2 b

k.

V + - + — +
.An , 2Na ; An) + ROH —P-> AnH , Na + An + RO , Na (51b)

‘ . = + .
The speCies (An , 2Na ; An) is called the "solvent caged"

complex and is presumed to be the most probable protonated

base. This species can also generate the "diffused out of

cage" complex (An=, 2Na+) which is the ion-paired dianion of

anthracene:

+ + ' + +-
2(An", Na ) 3+ (AnT, Na )2 -I-<—->- (An , 2Na ; An)
n = +
——>-K An , 2Na + An (52)

Thus the pseudo second-order rate constant. is only slightly
affected by the variation of [An] although it increases with

[ROH]. In their interpretation of the results, Rainis, Tung

and Szwarc made use of a preprint of the work reported in this

thesis (53).

In summary, the protonation of aromatic radical anions
by alcohols has been investigated in diverse media, with
various techniques. On the one hand, the Paul, Lipkin and

Weissman mechanism, which leads to first-order behavior, seems

to have been verified in some systems by many investigators.

 

35

On the other hand, recent stOpped-flow studies reveal a
second-order behavior in the absorbance. This proves the
existence of intermediate Species which are protonated.
However, the detailed nature of these intermediate species is
uncertain and different mechanisms propose different inter-
mediates. One of the major contributions of the present work
is that it permits us to rule out certain mechanisms and to

suggest the nature of the intermediate Species.

'. in
'-

 

I I I . EXPERIMENTAL

3.l-—Vacuum Techniques

 

ESiru3e solutions of aromatic radical anions are very un-
stable in the presence of air, high vacuum techniques are
required to handle every transfer of liquid from one glass
container to another. Agreaseless vacuum system was built for
this purpose. The vacuum manifold was first pumped to about
3 x.llffi3 torr with a mechanical pump. A Veeco air-cooled

diffusion pump which uses Dow Corning 704 diffusion pump oil
then brought the system to pressures of 2 x 10-6 torr. The
vacuum line was separated from the two pumps by a liquid
nitrogen trap. The manifold was also connected to a reservoir
of helium gas through a purification system to be described

later.

3.2--Cleaning Techniques

 

hiorder to eliminate insofar as possible every trace of

hwmrfljeS‘which could react with anion solutions and eventual-

lylewdto erronous results, careful cleaning of all glassware
was necessary prior to solution preparation.

Thegeneral cleaning procedure for every piece of glass-
mnevmsas follows: The item was first soaked in HF cleaner

meecmmosition (in volume percent) was:

36

 

“If

37

6 O % distilled water

33% HNO3, Fisher reagent grade

5% HF, 48%, Baker reagent grade

2% acid soluble detergent.
he glassware was then rinsed many times with distilled water

-

Jefore being soaked in aqua regia for at least half a day.

Finally, about ten rinses with distilled water followed by five

rinses with conductance water terminated the cleaning pro-

cedure. The item was then flamed-out and pumped to pressures

of 5 x 10"6 torr to get rid of traces of water before being

used in the preparation of various solutions.

The stOpped-flow system which constitutes the main tool
for this study was also cleaned as described above, except

that HF cleaner was not used in this case since it might at-

tack the inner parts of the mixing cell. After cleaning,

the system was also pumped to at least 1 x 10-5 torr for

several days before a kinetics run.

3.3--Purification Techniques

3 . 3 . l--Tetrahydrofuran

Tetrahydrofuran (THF) , the only solvent used in this

study, was obtained from Burdick and Jackson ("Distilled in

Glass“ type). The purification procedure was slightly modified

from that of Minnich (48) . About 3 liters of THF were poured

into a large glass vessel containing approximately 10 grams

of purified grade CaH2 (from Fisher Scientific Company).

 

38

A Teflon-coated magnet stirred the THF-CaH2 mixture for

several days until bubbles were no longer formed. The THF

solvent was then vacuum-distilled into a spherical glass
vessel which contained about 6 grams of benzOphenone (from

Eastman-Kodak Company) and an excess of a l to 3 sodium-
potassium alloy. A purple color appeared almost immediately

because of the formation of benzophenone ketyl. Because of

its reactivity with water, this served both as a drying agent
This purple solution re-

The THF

 

and as an indicator of dryness.

tained its color for months at room temperature.

actually used in the kinetics experiments was vacuum dis-

tilled from the purple solution thus prepared.

3.3.2--Alcohols
Spectral grade absolute ethanol (from Commercial Solvent

Co.) was degassed on the vacuum line by freezing with liquid
nitrogen, pumping and thawing many times until the pressure

of the alcohol in the frozen state stabilized to about

2 x 10—5 torr. In the second series of experiments, ethanol

was first distilled into a glass vessel containing a thin
mirror of sodium and pumped out from time to time to avoid

possible excessive pressure build-up due to the formation of

hydrogen gas. The treatment of ethanol with sodium was aimed
at the elimination of possible traces of peroxide in the

Llccfiuol. 'Then the ethanol was distilled into another glass
’essel where the freeze—pump—thaw technique described earlier

as used.

 

39

3.3.3--Alkali metals

 

Potassium metal as commercially available is neither pure
nor convenient to use. Therefore, it was stored in short
lengths of thin glass tubing such that the desired amount of
metal could be easily provided.

For this purpose, lumps of potassium metal were cut into
small pieces and quickly put into a tall glass vessel. Many
long, thin glass tubes of 3 mm I.D. with their top ends
sealed were also introduced into the same vessel which was
then pumped to a pressure of about 3 x 10--5 torr. Then the
vessel was moderately heated until the metal melted. At this
point, helium gas was introduced into the vessel to force the
molten metal up into the glass tubes. Potassium thus prepared,

after solidifying, was convenient and ready for use.

3.3.4--Anthracene and Crown

 

Zone-refined anthracene from James Hinton Co., Valparaiso,
Florida, and dicyclohexyl-lB-crown-6 from E. I. du Pont de

Nemours Co., were used without further purification.

3.3.5-rHelium
Grade A helium from Liquid Carbonic Company was treated
as follows before being admitted into the reaction vessels.
It was successively passed over hot c0pper shavings, hot copper
oxide, through Ascarite and finally through a liquid nitrogen

trap.

 

4O

3.4--Preparation

 

3.4.l--Pre-rinsing Techniques

 

Before being used in the preparation of the solutions,
every glass vessel was pre-rinsed with a solution of potassium
anthracenide in THF especially prepared for this purpose.

The glass bottle A (Figure l) to be pre-rinsed was attached F?
to the closed glass vessel B which contained the rinsing -.

solution. The attachment was made by means of a T-joint which

 

could be connected to the vacuum line. After a pressure of 1;“
l x 10-5 torr or less had been reached, the whole system was 5
disconnected from the vacuum line. Then, the rinsing solution
was poured back and forth two or three times. In the final
step, the rinsing solution was poured back into vessel B,
and the THF solvent was distilled several times from B into
A to eliminate traces of solute remaining on the walls of
veSsel A.

This technique, while time-consuming, has proven to be
very helpful in the preparation of stable solutions for the

kinetics experiments.

3.4.2--Anthracene and Crown Solutions

 

Weighed amounts of anthracene (or crown) were put into a
glass tube which had a break-seal on one end and a 5 mm Fisher-
Porter joint on the other. The glass tube containing An (or
crown) was then attached to the vacuum line through a liquid

nitrogen trap and pumped to a pressure of 5 x 10-.6 torr before

41

9 mm Penton
couplings

LEXIS II Tl

coarse frit

"4,To vacuum line

 

 

 

 

 

 

 

 

 

 

rinsing solution
iJ am.)

A B

Figure 1. Special set-up for pre-rinsing technique.

 

42

being sealed off. The amount of An (or crown) used was de-
termined by noting the weights of the break-seal with and
without An (or crown). The break-seal prepared in this way
was then sealed to a side-arm of the reaction vessel

(Figure 2a). This vessel was then weighed and pumped to

a pressure of about 3 x 10-6 torr. The solvent (THF) was then
distilled into the reaction vessel and covered with helium
gas at a pressure of about 1 atmosphere. The whole apparatus
was weighed again to determine the exact amount of THF used.
Finally, a Teflon-enclosed magnet in the side—arm of the
vessel was used to break the break—seal and An (or crown) was

dissolved in the THF solvent.

3.4.3-—Alcohol Solutions

 

Pure alcohol was distilled from its storage vessel into
a glass tube with a break-seal on one end. It was then frozen
with liquid nitrogen and pumped on before the glass tubing
was sealed off. Weights of the break-seal tubing With and
without alcohol in it gave the exact amount of alcohol used.
The break-seal was then connected to the side—arm of a reac-
tion vessel (Figure 2a) and the process of distilling THF
solvent and admitting helium gas into the vessel was exactly

the same as described earlier in section 3.4.2.

3.4.4.--Anion Solutions

 

Solutions of potassium anthracenide were prepared in

special glass vessels as shown in Figure 2b. A break-seal

 

43

fl __

. . u .
1.. 1
1 I. Hrllgfiflqi. ill-"B

mom no n3ouo .s4 Amo

 

 

 

 

3o

 

.mcoHHSHOm.rc<+M Ago mcowunaom
"no coaumummoum on» cw poms mammmw>

 

 

 

 

23

.N musmwm

 

 

 

44

containing a known amount of anthracene was sealed onto a
Side-arm on one side of the glass vessel. A second side-arm
on the other side of the vessel was constricted at four or
five places. A small length of tubing which contained potas-
Sium metal (as described in section 3.3.3) was introduced
into the side-arm and it was sealed off at the end. The
vessel was then pumped to a pressure of 5 x 10”6 torr or less
before the potassium metal was gently heated. The glass con-
strictions were then sealed off one by one after the metal
had been distilled through each of them. In this way, the
metal underwent four or five independent distillation steps,
leaving oxide behind, before it was distilled into the Special
vessel in the form of a Shiny mirror of pure potassium metal.
After introduction of the metal, THF was distilled into the
non-metal side of the vessel, and helium gas was admitted.
After the break-seal had been broken by a Teflon enclosed
magnet, the anthracene was dissolved in the THF. About half
of the anthracene solution was poured into the metal part of
the vessel through a coarse frit. A blue color appeared as
soon as the anthracene solution contacted the potassium metal.
This color grew deeper with time. When it was judged that it
would give an absorbance between one and two absorbance units
in the flow system, the blue solution was poured back into
the non-metal part where it was mixed with the remaining
anthracene solution. This method of preparation of anthra-

cenide solutions prevented the formation of the anthracene

-.....)

pr

h}'nfl.‘i-‘.4.vvrn_

 

‘It;;

45

dianion since an excess of anthracene was always present in
the blue solution. Differences in the weights of the vessel
with and without THF gave the amount of solvent present.

The vessel was then attached to the stopped-flow system, and

the long-awaited kinetics experiment could begin.

3.5--The Stopped-flow Experiment

 

3.5.l--General

Conventional methods of mixing and analysis are too slow
to be used in the study of fast reactions whose half-times
range from a few milliseconds to a few seconds. StOpped-flow
techniques possess two main advantages over conventional
techniques. These are fast mixing and fast analysis, thus

allowing one to study very fast reactions.

3.5.2--The Stopped-flow Apparatus

 

The stopped-flow system can be roughly divided into three
parts (Figure 3) which we will refer to as the left, the right
and the central parts. The left and right parts are the same.
Each one is composed of a burette and a mixing vessel which
were used to prepare solutions of various reactants. These
were introduced into the system gig 9 mm Fisher-Porter Solv-
Seal joints located above each burette. The burettes were
standard 50 ml Pyrex burettes calibrated to 0.1 ml. Volume
readings were estimated to 0.01 ml. Mixing was performed by

stirring the solution with a Teflon-encased magnet which had

 

46

+ 7' .
K An solution

1 An (or crown) ROH sol tion
'0
s Intl n T THF solvent

+-

Vac.

' ran-r-

 

 

Trigger unit A \ bi

Stopping‘block l

I

 

 

 

 

 

Burette-——+

 

 

 

Mixing chamber and
Vac

E: _,' II,””optica1 cell

 

 

 

 

 

 

 

 

)1
)

 

 

Teflon valve-3&6 H t

()
+ Mixing vessel
C D
L_____. V

‘1
’ 71‘
e L: l: +

Waste Waste

/ \Syr inge
Pushing block

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 3. Schematic diagram of the stopped-flow
apparatus.

47

been sealed inside the mixing vessel. To empty the bottom
tubes, the solution was often warmed to boiling with a heat
gun or with just the heat from one's hand.

The central part of the stopped-flow apparatus is more
complex and thus deserves more attention in its description.
The main components of this part are: a pair of pushing
syringes with Teflon plungers, apprOpriate valves, a mixing
cell and a stopping syringe. All of these components were
tightly held together by a set of aluminum plates with
threaded 3/8 inch brass rods, and the whole system was bolted
to a rigid table made of two inch angle iron of 1/4 inch
thickness.

Since THF attacks Viton—A "O-rings" and dissolves most
stopcock greases, a Special design was needed for a greaseless
and air-tight syringe for the stopped—flow system. The final
model of such a syringe (Figure 4) was evolved after many
tests by Drs. V. A. Nicely, M. G. DeBacker and E. R. Minnich.
These syringes were made of precision bore tubing (Trubore
8700-60, Ace Glass Inc.) with a side-arm sealed at the mid-
point of the syringe and a 5 mm Fisher-Porter Solv-Seal joint
at one end. The Teflon plungers have two sets of double Teflon
ridges so that gases which are trapped between the pairs of
ridges can be pumped out constantly through the side-arms.

Needle-valve type stopcocks used in the stopped-flow
system were of two kinds: Fisher-Porter Lab-Crest 4 mm Quick

Opening Valves and Delmar-Urry valves. The first ones had

 

“ff

48

 

<——Teflon ridge

 

Teflon plunger— ——>-

—> Vacuum

 

 

 

 

...—L
L

 

Figure 4. Syringe used in the stopped-flow system.

 

 

49

Teflon tape wrapped around the threaded parts. The second
ones had two of their four "O-rings" replaced by Teflon.
To accomplish this, the entire Teflon valve stem was con-
structed in the Chemistry Machine Shop.
The design and construction of the four-jet quartz mix-
ing cell (Figures 5a,b) used in this stopped-flow system is F?
described by Hansen (54). The cell has two flat faces and is tar

approximately 1.0 mm I.D.

 

3.5.3--Calibration of the Stopped-flow System H
The stopped-flow apparatus was calibrated in the follow—
ing way:
A freshly prepared stock solution of KMnO4 in H20 at a
concentration CO was successively diluted to make three more
solutions at concentrations 0.75 Co’ 0.50 CO and 0.25 CO.
The absorbances of the 526 nm peak of these various KMnO4
solutions were then measured on a Cary 14 spectrophotometer
with (1.00 :_0.01) mm SCC cells. On the other hand, a digital
voltmeter (Heath Universal Digital Instrument, model EU-805)
was used to read the corresponding voltage values of the
neutral density filters (from Oriel Co.) through the stopped-
flow system. For this purpose, the fixed wavelength technique
was used: the center of scan of the scanning monochromator
was set at 526 nm, and the nutation set at zero. Then the
same KMnO solutions were successively pushed through the mix-

4
ing cell of the stopped-flow apparatus. Their respective

50

 

 

Optical
path ———+

 

 

 

    

 

 

 

 

 

 

 

 

(b)

Figure 5. Four-jet mixing cell: (a) cross—section
(b) entire flow cell.

51

absorbances were recorded by the digital voltmeter as voltage
values. A calibration curve of the voltages obtained for the
neutral density filters was then used to convert voltage

values to absorbance values of the various KMnO4 solutions
recorded through the stopped-flow device. The plots of
Absorbance from the Cary 14 and the Absorbance from the digital

voltmeter XE KMnO concentrations give two straight lines

4
(Figure 6). From the values of their slopes, an effective

path length of the mixing cell of (0.90 i 0.02) mm was obtained.

 

From the fixed wavelength data, the effective stopping
time of the stopped-flow system was estimated to be at most

5 msec.

3.5.4--Operations

 

After the entire stopped-flow apparatus had been pumped
to a pressure of l x 10-5 torr, the stopcocks connecting the
flow system to the vacuum line were shut off. Solutions of
various reactants were let into the mixing vessels through the
pair of burettes. The stopcocks A and B were Opened; the
pushing block which held the two pushing plungers was pulled
downward with a lever. This allowed the solutions to flow
from the mixing vessels into the two syringes on the bottom.
Then the stopcocks A and B were closed and the stOpcocks C
and D were Opened. The pushing block was pushed upward very
rapidly by a manual pushing lever. The solutions were thus

forced to mix and were permitted to react in the mixing cell,

52

N v

.o m an case no mcoausaom may no oo\o mm mocmnnomna

O

o.H m.o m.o

U\U

v.0

 

 

r _ _ _ i _

N ¢

0 m nuflz coHuDHOm oczx mo swam

Ana nufls EcumMm 30am may Eouw ADV

 

_ _

 

wage cofluSHOm

v

OGZM

nufl3 Emummm 30Hm1owmmoum on» Scum Aoo
«a mnmu on» scum Axe

N.o

o.H

m.H

 

 

.w mHDmHm

eoueqlosqv

53

and the resulting product solution flowed into the stOpping
syringe. When the stOpping plunger hit the stopping plate,
the flow of liquid was suddenly stOpped, marking the end of a
typical push and the beginning of the acquisition of useful

rate data.

3.5.5--Data Acquisition

 

The decay of the Spectrum of potassium anthracenide
(Figure 7) was followed spectrOphotometrically. A Schematic
of the system for data acquisition is shown in Figure 8.

Light from an apprOpriate light source was scanned by
a scanning monochromator over a desired range of wavelength
before being Split into two matched beams by a beam-splitter.
The output light beams then passed through a reference cell
or the mixing cell before striking the photomultipliers. A
dual-log-circuit which employed a differential amplifier and
Philbrick SPl-A Operational amplifier received the output
currents from the two multiplier phototubes and converted trans-
mittance to absorbance. An FM tape recorder (Ampex SP-300)
stored in four independent channels absorbance signals from the
log circuit, trigger signals from the monochromator, verbal
explanations from a microphone and a blank signal for later
common-mode noise rejection. An oscillOSCOpe connected to
the FM tape recorder allowed signals to be monitored during
the entire eXperiment.

Since in this study, the wavelength range of interest ex-

tended from 400 to 900 nm, the only light source used was a

54

.mme :a noun spas cowuomon mcwuso.kgm+x mo Eduuommm mo mmoco

£00.

 

.25 Eozmsm><3
o3 o8 can. con
~ — —

_

 

 

 

 

  

)/I\

one
1)

l'.‘

000

q

 

(aims Douaqlv) aouvaaosav

.h musmwm

 

55

.Empmwm cowuflmasqomnmumo can no EMHoMHO xoon

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.w musmwh

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_ i ll Hoouooom I)
(II). k ,
ILA meme 2m mcosmouoflz
mmoomOHHHomo uHSOHHUImOA
mnsuouoam mnsuouoamg mmnsuoummm
a #1 Vlh
HHOU HHOU
wocmnmmmm . mcflxflz
FIJI. . V
.
"younaamm sown _
. AWMV ” cmmonuflz
oaoo
fulAIll. I V ..V\
- .t
um3oHHom_ HoumEonnoogoz
moonumo cmomlowmmm 1A. lllll ; condom ”2.3.3

 

 

 

 

 

 

 

 

 

56

Bausch and Lomb tungsten-iodine lamp. The monochromator, a
Perkin-Elmer Model 108 Rapid-Scan Monochromator has a scanning
capability of from 3 to 150 scans per second over any wave-
length region within the range of the quartz prism, provided
that the appropriate light source and detector are used.

Its detailed description has been given by Feldman (55). It
is no longer commercially available. Often a scanning rate

of 60 scans per second was high enough to permit one to ob—
serve a complete absorption Spectrum of potassium anthracenide
during reaction. For very fast reactions, the scanning
technique was replaced by a fixed wavelength technique. In
this case, the motor of the scanning monochromator was stOpped
and a desired single wavelength corresponding to a particular
peak position of the anion spectrum could be chosen.

In the fixed wavelength case, a trigger signal was pro-
duced every time the plunger hit the stopping plate. When the
scanning technique was used, a special device described by
DeBacker (56) produced the trigger signals in the following
way. A slotted wheel attached to the drive mechanism of the
monochromator let the light from a small neon light bulb hit
a CdS photocell once per rotation of the mirror. A trigger
signal of about 1 Volt in l millisecond was thus produced at
the beginning of each scan. This trigger signal was passed
into the FM tape recorder through a cathodafollower. Prior to
the second series of experiments, the neon light bulb and the

CdS photocell were repositioned by R. B. Coolen and

57

N. Papadakis behind a reflecting gear tooth of the metallic
gear which drives the mirror.

The stacked-mirror beam-splitter was of the type used in
the Bausch and Lomb Spectronic 505 SpectrOphotometer. Spherical
mirrors of 98 mm focal length (from Karl Lambrecht Co.) focused
the two light beams onto the mixing cell and the reference
cell respectively.

Two types of RCA photomultipliers were used, depending
on the wavelength region of interest: an RCA 6199 for the
region between 400 and 600 nm and an RCA 7102 for the region
between 600 and 900 nm. It was necessary to set up a cooling
system for the RCA 7102 phototubes to reduce thermal noise.
Nitrogen gas passed through liquid nitrogen gig copper tubing
was used for this purpose.

f Prior to each run, the two currents from the photo-
multipliers had to be matched by adjusting the input light
received by each phototube. Also the linearity of the ab—
sorbance readings from the log-circuit was checked by using
neutral density-filters (from Oriel Co.). The log-circuit was
built with the following units purchased from Philbrick
Research Inc.: a power supply PR-30C, a high impedance input
amplifier P25A, a dual logarithmic transconductor Pll-P and
an operational output-amplifier P65AU.

The tape recorder was an Ampex SP-300 PM Direct Recorder/
Reproducer. Finally, a Tektronix 564 Storage Oscilloscope

with a Type 2A63 Differential Amplifier Unit and a 3B4 Time

58

Base Unit served as a monitor.

3.5.6--Data Analysis

 

3.5.6.l--General

After the kinetics data had been prOperly stored on
magnetic tapes as voltage fluctuations, the next problem was
their analysis.

With the aid of a Computer of Average Transients (CAT),
the kinetics data could be read from the tape recorder onto
graph paper in analog form for display purpose, and onto
computer cards in digital form for the purpose of rigorous
analysis. Typical displayed data in analog form are shown in
Figures 9 and 10. These graphs correspond to scanning and
fixed wavelength techniques respectively. When the scanning
technique was used, a series of spectra was recorded during
an entire push. In this way, each Spectrum after the flow
stOps is a decayed form of the preceding one. The basic
idea, then, is to sample in digital form identical portions of
the Spectrum from successive spectra and to use a computer
program to study them. A Varian C-1024 Computer of Average
Transients (CAT) was used to produce the digital record.

This small computer has 1024 channels, each of which can be
independently triggered from an external trigger source. The
CAT can convert absorbance data stored on magnetic tape (as
voltage versus time) into digital words which are stored in

successive channels. A diagram of the analysis system is

.Amavwcsomu maflccmomv has
cw scum z mmoo.o nufi3 mcwxwfi Honmm_kc¢+m mo xmmm as one on» no mmoma .m musmflm

Afisuuommm Hum come n.mao mafia

59

 

 

(sqrun Kxexqrqxe) eoquJosqv

J
“:1“.—

 

 

60

omm

oom

.Amnqflcnomu numcmam>m3 poxflwo
was an moan z omo.o nuns maaxae Hanna +a<+s no read as ems can no smoma

AOOmEo mafia
oma

00H

om

.OH ousmflm

 

)1

_

 

 

_

—

 

 

(areas Kierqrqle) eoueqlosqv

‘-

61

shown in Figure 11. In order to better understand this some-
what complex network, let us give an example.

Suppose that we wish to study the decay of the 720 nm
peak of the potassium anthracenide Spectrum as a function of
time. We then want to store a small portion of the 720 nm
peak of each spectrum in 8 channels. In this way, 64 succes-
sive decay peaks will occupy the first 512 channels of the CAT.
In order to implement this idea, it is necessary to use a
pulse-delay unit. Although a trigger signal is produced at
the beginning of each spectrum, the pulse-delay device can
delay this trigger Signal for any desired time. This delayed
trigger signal can be used to produce 8 square-wave pulses
from a waveform generator. These pulses command the CAT to
successively Open 8 channels for data storage. The frequency
of the waveform generator determines the time-interval between
consecutive pulses. The next 8 channels of the CAT will be
triggered by the next delayed trigger, and so forth until all
512 channels are filled. Thus absorbance values of the first
64 spectra sampled at 8 different wavelengths were stored in
the CAT, ready to be punched onto computer cards for analysis.

For more Specific details refer again to the diagram in
Figure 11. The absorbance signal and the signal from the blank
channel of the tape recorder were sent through a variable gain
differential amplifier to reduce common mode recorder noise.

A Tektronic Storage Oscilloscope displayed the output waveform

for monitoring purposes and this analog voltage was also

 

62

 

 

 

 

 

FM

Absorbance data

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Speaker Tape Recorder Blank channel
External
Trigger V
Differential
Amplifier
Input (-)
Oscilloscope (+) Input
Gate out
PUDZan Trigger Input
‘ r
Circuit
1
Fj'fi Wavetek Address
Advance ”

 

 

 

Figure 11.

 

 

 

j

Binary Output
To Keypunch

 

 

Block diagram of the analysis system.

Analog Output
To X-Y Recorder’

L.

63

connected to the input of the CAT. The pulse-delay unit re-
ceived gated trigger pulses from the oscilloscope and sent
the delayed Signals into a Wavetek Model 116 Signal Generator.
This waveform generator produced square-waves which were sent
into the channel advance input of the CAT and also into the
second channel of the differential input amplifier of the
oscilloSCOpe. In this way, we could follow on the oscilloscope
screen not only the entire spectrum but also which portions
of the absorption spectrum were being stored in the CAT in
digital form. After acquisition of the data, they could be
punched onto computer cards by an IBM type 526 Keypunch which
was coupled to the CAT by a Varian C-lOOl coupler.
Simultaneously, analog plots of the data could be obtained.
Two computer programs, PUNDAT and KINFIT were then used with a
CDC-6500 computer to analyze those kinetics data.

When the fixed wavelength technique was used to collect
the data, the pulse-delay circuit and the Wavetek were not

needed in the analysis system.

3.5.6.2--Program PUNDAT

 

The data obtained from the CAT had to be transformed into
a form acceptable to program KINFIT, the main tool for data
analysis. Program PUNDAT was written for this purpose by
Dr. V. A. Nicely. It was modified later by Dr. E. R. Minnich.
A detailed description of this program was given in Reference

48.

64

3.5.6.3-—Program KINFIT

 

This program was written by Nicely and Dye and is

described in detail elsewhere (48,57,58). Although program

KINFIT cannot decide whether a given rate law is suitable for

a particular set of data, it gives the following information:

the best estimates of the parameters obtained from

a least-squares treatment of the data;

the sum of the squares of the residuals;

the multiple correlation coefficients and the variance-
covariance matrix;

the comparison between calculated and experimental
values;

a print-plot of calculated and observed values.

Having all the above information, an experienced program

user can decide whether to accept or to reject the trial rate

law.

IV. SURVEY AND TREATMENT OF THE DATA

As mentioned in the Introduction, this study is a continu-
ation of Minnich's work. However, instead of continuing to
examine various aromatic hydrocarbons and numerous alcohols in
tetrahydrofuran (THF) and dimethoxyethane (DME), we chose only
one system. This study deals with the protonation of potassium
anthracenide (K+Anr) with ethanol in THF over a wide concen-
tration range of ethanol and of anthracene. We also used
dicyclohexyl-lB-crown-6 (crown) to test the effect of ion-

pairing on the protonation rate.

4.1--Survey and Discussion of the Data

 

There were two main series of experiments as shown in
Table I. In the first series, the initial concentration of
the K+Anv stock solution in THF was rather low (n.2 x 10”4 M
after mixing). The reason was that, through inexperience, we
did not spread the mirror of potassium metal over a large area
on the glass walls of the reaction vessel (Figure 2b).
Therefore, the anthracene solution did not react with enough
potassium metal to produce a sufficiently concentrated K+AnT

solution. In certain runs of this series, we could not analyze

the collected data for one of the following reasons:

65

6
6

 

 

 

 

 

 

 

 

 

Nica x m.~ mica x m.m ii cma wim
Nica x c.a mica x N.v ii cma him
mica x m.m mica x a.m ii oma mim
mica x m.a mica x v.m ii cma aim
mica x m.a mica x m.a ii cmaicm Oim
mica x c.a Nica x c.m ii a cmxaw mim
mica x n.a mica x m.v ii cc dim
mmaoc ccsoumxomn £098 009 ca x v.m ca x c.m ii a cmxam mia
mmowc vi ai

HOUROiumHam occcmm 30am vica x v.c aica x c.w wica x c.v cmaicc wia
30am 008 h_vica x v.c mica x m.m vica x c.v cc mia
wica x m.c aica x m.a ii cma Mia
ummm OOH vica x m.c aica x c.c ii cma aia

msvacnomu
ocaccmom Mom ummm ooa aioa x m.m aioa x e.e ii oma Oia
wica x m.m mica x h.m ii cc mia
vica x m.c mica x 5.x ii cc fiia

o O o
mucmfifiou 2 ammo z am0#m_ z ac3ouoo omm\mcmom cam
mucmEaHOmxm mo umaa
H Manda
H.

67

 

wmumo Hmcuo

iumuam 30am an Um3oaaom
mmomw HmUHOIUcoomm ummm
mmumc

Hm@H0iumuam 0650mm 30am

 

 

 

 

cm

cmxam
om

omxam
cmxam
umxam
umxam
om

oma

 

aim
mim
OIN

ZIN

68

- the reactions were too fast for the scanning tech-
nique (high ethanol concentration, no crown);
- the reactions were too slow (low EtOH concentration,
withicrown);
- there was too much background noise for the fixed
wavelength technique because of low absorbance.
The majority of the data come from the second series of experi-
ments. This time, we wanted to have a high enough concentra-
tion of K+AnT to work with easily. In fact, we were so careful
in spreading out the active surface of K metal on the glass
walls that we almost overshot the desired initial K+An7 con—

3 M after mixing).

centration (’b2.4 x 10-
Besides these two main series of experiments, we re-
analyzed most of Minnich's data. In fact, Minnich did a great

deal of screening work, and obtained many data on a variety
of systems. However, in most cases he analyzed only a few
pushes at each concentration in order to test various mechan-
isms. The reader will see later that we needed nearly all

of Minnich's data for the THF solvent system in order to
distinguish between several prOposed mechanisms. This need

became apparent only after completion of our study. A summary

of Minnich's data as re-analyzed is shown in Table II.

4.2--Data Treatment

 

In conducting this study, we had three main purposes in

mind:

69

TABLE II

List of Minnich's Experiments Re-analyzed by the Author

(2K+An"' + ZROH

THF

-———+- AnH + An + 2K+Ro’)

2

 

 

Range of ROH

Number of ROH

 

Run ROH concentrations (M) concentrations
KR7,5-6 EtOH (4.32-2.20) x 10'3 2
KR9,ll-l4 EtOH (5.0-0.48) x 10"3 4
KR7,l-4 H20 (3.6-0.38) x 10'1 4
KR8,4-5 H20 (2.2-O.8) x 10'1 2
KR9,9-10 H20 (4.7-0.41) x 10’2 2
KRlO,l-3 i-PrOH (1.14-0.11) x 10’1 3
KR10,4-6 t-BuOH (2.0-0.54) x 10‘2 3
KRlO,7-9 MeOH (3.7-0.5) x 10'3 3

 

 

 

 

70

a) to reconfirm the existence of a second-order
protonation of potassium anthracenide in THF;

b) to test the ion-cluster mechanism (Equations 42, 43)
and the dianion mechanism (Equations 47, 48) over a
wide range of concentrations of ethanol and anthra-
cene. The ethanol concentration was varied from a
very high value (“10.5 M) down to a value less than
the initial concentration of K+AnT. .In addition the
concentration of free anthracene was also varied from
an upper limit of “10.04 M to a concentration which
was less than that of K+Anv.

c) in case these prOposed mechanisms failed to fit our
experimental data over the entire concentration range,
search for new mechanisms.

Simultaneously, a number of attempts were made to fit the data
at different concentrations directly to a particular mechan-

ism. However, only three or four pushes could be examined at
a time and this procedure was not successful. Therefore, the
data (without crown) were fit to a pseudo parallel first- and
second-order rate law where possible. The integrated form of

the following differential rate law:

2 (53)

 

+ 1-
d [K An ] + 1- ' + 1"]
_ = +
dt kps[K An ] kpS[K An
was fit to the data by using a non-linear least-squares pro-
gram (57). Usually when [EtOH] and [An] were appreciably

higher than [K+An'], the data were well-described by this

71

rate law. A representative fit of this rate law to our data
is given in Figure 12. The first-order contribution to the
decay of the absorbance only became appreciable above [EtOH]
% 0.1 M. In those cases for which [EtOH] and [An] were com-
parable to [K+AnT], pseudo order kinetics were no longer
valid. For these cases, we fit the data by using the equa-

tion:

_ d[K; An
k :TAn] [K+ An (54)
k1 [ROH]

derived from the dianion mechanism:

 

 

k
+ . =
2(AnT,K ) ——i+- (An ,2K+) + An (55a)
Ik
= + - k1 + — + -
(An ,2K ) + ROH ————+- K AnH + K RO (55b)
K+AnH + ROH EEO-E)- AnH + K+RO- (55C)

2

with the assumption that a steady-state concentration of
(An=,2K+) was achieved. For these data, the KINFIT program
was used in its differential form. For comparison with the
data at higher concentrations, the "pseudo second-order rate
constant" was obtained fromzi

2k+
kps = k [An] (56)

1 + El‘IEROH]

 

 

In this case, [An] and {ROH} must be adjusted to their values

72

 

 

———-Calculated line

0 Experimental points

 

 

E
V
O
H
x
F“ ‘0
b: s F’ \\
fit
+
E.

4 _-

‘0
3 b \
°\.
2 _.
\\\o\\\
‘0
1 -— \\\“~<>
\o
\—
<3‘~—~cy—
l J L l cJ
0 50 100 150 200 250 300
Time (sec)
Figure 12. Parallel first- and second-order kinetics applied

to the reaction K+AnT + EtOH in THF ([EtOH]O =
8.9 x 10-2

M).

iii}

 

ELI—""m

73

at the point at which kpS is to be calculated. The values of
[An] and [ROH] were calculated from the mass balance equa-
tions:

([K+An*] - [K+An*]t) (57)

NIH

[Ath = [An]O + 0

_ _ + r _ + T
[ROH]t — [ROH]O ([K An 10 [K An ]t) (58)

It was necessary to use this procedure because at low concen-
trations of ROH, the second-order rate constants are sensitive
to variations in the ratio [ROH]/[An]. By this procedure, we
found that the dianion mechanism fits our data well at low
[EtOH] ( <0.0l M). In addition, by using this procedure it
was possible to calculate the values of the rate constants
not only at the initial point of a decay curve, but also at
any point along this curve. In one extreme case, for which
[EtOH] was less than [K+AnT], neither the pseudo parallel
first- and second-order rate law nor the dianion mechanism
could be used to fit the data. For this case, we used the

empirical rate law:

_ d[K+AnT]
dt

 

= k[ROH][K+AnT]2 (59)

with the mass balance equation (58) to fit the data over only
the first half-life of the reaction. For this case, of

course,

kpS = k[ROH] (60)

74

By these procedures, it was possible to obtain values
of kps (2nd) for all cases. This proved to be a valuable aid

in testing new mechanisms.

V. RESULTS AND CONCLUSIONS

5.l--Results

5.1.l--Inapplicable Mechanisms

 

As mentioned in the last chapter, one of our goals was
to test the ion-cluster mechanism (Equations 42, 43) and the
dianion mechanism (Equations 54, 55). In fact, both mechan-
isms predict that the pseudo second-order protonation rate
constant would reach an upper limit set by k; and k+ respec-
tively, and become independent of [EtOH] at high ethanol
concentrations. This clui.not happen in our case: the rates
continued to increase as the ethanol concentration was in-
creased as shown in Figure 13. Figure 13 also shows that the
rates become less sensitive to the anthracene concentration at
high [EtOH]. Clearly neither the dianion mechanism nor the
ion-cluster mechanism alone can fit the data at high [EtOH].
On the other hand, for low [EtOH] (€50.01M), the dianion
mechanism fits the data well, while the ion-cluster mechanism
failed to predict the observed effect of [An] on the protona-
tion rate. Figures l4a,b show the superiority of the former
mechanism over the latter: while the plot of kpS (2nd) vs.
[ROH] diSplays scattered data points which correlate with
[An], the same values of kps (2nd) when replotted vs. [ROH]/

[An] show a more ordered trend of variation.

75

76

.3: 33.2: :3; iv :5: $33.." 539: :c
30333880 :4 $3 amamca 30a 3c "gamma moum flag ...:4 m
mo coauommu on“ How amouma m> ammoHOiocoommv x mo HOaQ moaimma .ma madmam

2 2.0.:

-. .0.
4nd _ #7.. . ~__a

 

 

 

 

77

to ‘0
0 a5: no? 7: . a5: 532 A:
N an<c 30a Aoc name ca moum cua3.iqm+x mm :oapommu may now
.aaac\amoumc Inc can amOpmH “ac m> Iumcuoicaoommc M NO uoam moaimoq .va mummam

Abc Amv

“53:2: 2 «0.120.:
; a .o. .oo. o. a

 

 

 

la__a_14a_Jal..aaa a .0. a. a aaafi . __fi.

1

‘
I114

('-,..'-w) ”'0' ‘ optical).
H
(I',.'l'w) :-°' ' Options).

D!
53'
L

o i.
r... -

o. n. ..

LLIJ

 

 

 

 

 

78

5.1.2--Effect of Crown

 

The effect of solvent on the rate of protonation has been
reported in the literature (48,49,50,52). When the formation
of contact ion-pairs is not favored, the protonation rate is
markedly lower. In order to test this effect of ion-pairing,
we added dicyclohexyl-lB-crown-6 (crown) (59), to a solution
of K+Anr in THF prior to protonation. The crown compound has
been proven to be a good complexing agent for alkali metal
cations (60,61). ESR results indicate that crown breaks up
the contact ion-pairs of K+Anf. The effect on the protonation
rate is dramatic. The addition of crown essentially wipes
out the second-order protonation, provided that [crown]
> [K+AnT]. Only a slow first-order component remains, as
shown in Figure 15.

Although the studies with crown are only preliminary in
nature, it is interesting to note that the first-order proto-
nation rate of the solvent (and/or crown)-Separated ion-
pair, An? ll K+ (in THF with excess crown) seems to be about
100 times smaller than the first-order protonation rate of the
contact ion-pair, AnT, K+ (in THF without crown). In the
intermediate case for which [crown] <:[K+An7], the decay curve
is broken into two portions: a fast initial decay followed by

a much slower first-order decay as shown in Figure 16.

S.l.3--Proposed Mechanisms

 

We have seen that neither the dianion mechanism nor the

ion-cluster mechanism fit our data over the entire range of

[K+An"] x 104 M

79

 

0.6 ’—

0.4 ——

0.3 _-

0.2 _-

 

0.1

l

|

l

0 Experimental points

-—-Calculated line

 

Figure 15.

0.2

0.6 0.8 1.0

Pseudo first-order plot for the reaction
(excess crown) with EtOH in THF.

Time (sec)

 

80

"noum+MN + cc + mac

w.—

0a

.G3OHU on auc

N

Era.

.mommo aam How 2 mv.c n
«a... 5+ v: v .aczouUH Am:
T mOUMN + FG<+ MN

“90$ 25...
a c.

camoumc

.a ...EN +5 A HESHE AS

.coauomou may you mEau m> OU\U

at

.ma musaam

 

 

 

—

 

 

 

81

concentrations of EtOH and An. In the search for a new

mechanism, we must take into account the three main effects

already mentioned in part 5.1.1:

- the protonation rate continued to increase with in-

creasing [EtOH];

- the rate at high [EtOH] did not depend significantly

on the anthracene concentration;

- the dianion mechanism fit the data well at low

[EtOH].

A suitable mechanism would seem to require the formation of

. . = + .
the anthracene dianion (An ,2K ) to account for the-behaVior

at low [EtOH]. It also might include the formation of the

, +
ion-cluster (AnT,K )

leads to a mechanism which does not depend on [An].

S.l.3.l.-—Dianion and Ion-cluster Mechanism

2 since the protonation of this species

For the reason given above, we tested the following

+
,2K + An

mechanism:
K k"
7' + Q 1' + +
2(An ,K ) (An ,K )2 -——*' An
= + k1 + - + -
An ,2K +ROH——:->KAnH +KRO
1- + k2 + "' + -
(An ,K )2 + ROH ———+vK AnH + K R0 + An
K+AnH- + ROH 33111» AnH2 + K+RO-

which leads to the general rate law:

(61a)

(61b)

(61c)

(61d)

82

 

 

 

 

 

+ 1- 2k"K
_ d[K An ] _ :KQ + T 2
dt _ k: [An] + 2k2KQ[ROH] [K An ]
1 + 1?; [ROH]
._ _. (62)I

provided a steady-state concentration of the dianion
(An=,2Kf) is reached.

At high values of [ROH], the second term of this rate
law becomes predominant, and the decay becomes pseudo second-
order in [K+Anf] and first-order in [ROH] after correction for
the first-order protonation. At low [ROH], the second term
becomes less important and the rate expression reverts to that
of the dianion mechanism.

An important point should be made here: the "solvent
caged" complex (An=,2Na+;An) suggested by Szwarc and co-

workers (52) cannot be kinetically distinguished from the ion-

 

. +
cluster spec1es (AnT,K )2
If the formation of the cluster involves only electro-

static forces, the value of the equilibrium constant for the

reaction:
KQ
-———+- f +
2Anf,M+ (——— (An ,M )2 (63)

can be estimated from the Fuoss expression (62,63):

 

 

Y ' -
KQ = (g )y e 1 1 1/2
2000 DkT y7/2 2X2—- (64)

. , . . . + i
where: u = dipole moment of the ion-pair AnT,M con51dered

as an ellipsoid;

 

83

 

a = major axis of the ellipsoid;
Aa = minor axis
and:
U2
Y = ‘T’j (65)
(la) DkT

Using the observed value of the ion-pair association constant
of Na+AnT (64), we can calculate the value of a from the Fuoss

equation for ion-pair formation (63):

 

3
40a N
_ Av b
Ka ’ 3000 e (66)
with
e2
b = aDkT (67)

0
The value of 3 obtained this way is 3.: 5.7 A. From this

value, we get KQ z 48 M-1

(65) for the equilibrium constant for triple-ion formation

. Also the Fuoss-Krauss expression

and the assumption that the equilibrium constant for the

reaction:

+ + a +
AnT,K ,AnT + K 7r-*' (An ,K )2 (68)

is the same as that for ion-pair formation, Equation 66,

gives K = 67 M-l. Therefore, we can estimate a reasonable

Q
value of 50 M‘1 for KQ.

Returning to the rate expression (62), we see that it
contains 3 adjustable parameters, kiKQ, kz/kl and szQ with
Strong coupling between them. We used a lot of computer time

ttying to couple various data sets and fit them directly to

84

Equation 62 in order to extract reasonably stable values of
these three parameters. Finally, we realized that the sep-
aration of these strongly coupled parameters depends upon the
use of many different values of [An] and [ROH]. This cannot
be satisfied by coupling just a few data sets. Thus we,
decided to use an indirect approach as follows. We write

Equation 62 as:

 

+ -.-
- dIKdin 1 = f(An,ROH)[K+AnT]2 (69)
where:
2k1K '
f(An,ROH) = ""Eg‘TXET" + 2k2KQIROH] (70)
1 ”LE—1’ [ROH]

depends upon [An], [ROH] and the 3 parameters. From the _
material balance equations (Equations 57, 58), we can get

[An] and [ROH] at any point during the reaction. At high
values of [ROH] and [An], f(An,ROH) is just the pseudo second-
order rate constant. At low values of [ROH], f(An,ROH) is ob-
.tained from Equation 56 of the dianion mechanism as previously
described. It can be calculated at the initial concentration
as well as at various extends of reaction. For the one case
mentioned previously at [ROH] <:[K+An'], an empirical fit of
the data (Equations 59,58) over the first half-life of the
decay curve gives the corresponding values of f(An,ROH).
Finally, Equation 70 was fitted by least-squares to all the

data over a wide range of values of [ROH] and [An]. The data

85

are given in Table III. Figure 17 allows us to judge the
"goodness of fit" of the expression to the data over almost

3 orders of magnitude of the function f(An,ROH). Most of the
points agree with the calculated values to within one or two
standard deviations of repeated pushes. The final values
obtained for the 3 parameters in Equation 70 are given in
Table IVa.

The validity of this indirect method was checked in the
following way: we fit all of the decay curves to Equation 70,
with the 3 parameters replaced by their corresponding values
from Table IVa, allowing only the initial concentration to be
adjusted. Figure 18 shows some representative observed and
calculated curves which were obtained in this way for varying
initial conditions. The particular curves which are dis—
played were chosen on the basis that their deviation from the
corresponding calculated curves represents about the average

deviation of all of the data.

 

5.1.3.2--Cation Solvation Mechanism
Another hint about a possible mechanism came from the
research of Fujihira and co-workers (46) who observed a pro-

nounced effect of [H O] on the protonation rate of An? in

2
DMF-water mixture. This effect was interpreted by them as
due to the stabilization of charge localization in the transi-

tion state by hydrogen bonding (45,66). Such an effect, if

applied to our system, might increase the rate of formation of

86

 

 

 

 

 

 

 

Ill iii 0N.0 £00.H 050.0 m 0HN.0 00.0 0m.m
iii iii 000.0 0000.0 000.0 0 000.0 00.0 00.0
ii) iii 00.0 000.0 000.0 0 000.0 00.0 00.0
iii iii 000.0 0000.0 000.0 m 00.0 00.0 00.0
iii iii 000.0 0000.0 000.0 0 000.0 00.0 00.0
ii) iii 000.0 0000.0 000.0 m 00.0 00.0 00.0
ii) iii 000.0 0000.0 000.0 0 00.0 00.0 00.0
iii iii 000.0 0000.0 000.0 0 00a 0 00.0 00.0
iii iii 00.0 n0a.a 000.0 0 000.0 00.0 0a.a
ii- iii 000.0 030.0 000.0 0 00.0 00.0 00.a
III III m0.0 Q001m m0m.0 N PMN.0 VN.0 05.0
III III mH0.0 QHNH.0 NEH.0 N ©0.H 05.0 00.0
iii iii 000.0 0000.0 0000.0 0 00.0 0a.a 00.0
ill III III £00000.0 HNH0.0 H Nv.v Mb.0 NH.0
ii. iii iii na0000.0 00000.0 a 00.0 00.0 00.0
a mno oamo mosmsm o o 0
mo .m 00 .M ac0.cwmc .M mo umnesz z oax acac Emoax 00:40 zmcax amoumc
a 0000 mu: umcou mumm avioaxaiowmaizv musmumcou

HI.

 

umpMOIumuam ocsmmm

 

GUMM HQUHOIUCOUQW OUDOmm

 

 

 

ampsum manu Eouwv

 

mQOaumuucmosoo aocmnum gmam um soausnanucou HmpuOiumHam Opzmmm paw
mme ca mOum nua3 .:¢+M mo COHuommm mnu How musmumcou mumm HwUuOicsoomm OUDmmm cmumasoamu 0cm Uw>HmeO

 

HHH mam<fi

87

.xmomc HmUHOIUsoomm can iumuam amaamnmm on 0006 m0 paw Eoum

.mump mo paw amoauamaw Scum ucmuwcoo mumu amauaaam

U

.vm coaumswm mo paw Scum mpcvuwcoo mumu amauasao

.vm coaumswm mo paw Eouw coauommu mo musmuxm msoaum> Hmumm muCMUmcoo mumm

 

 

a.mm

a.mm
00.0
mm.m
Nv.a
mn.m
mm.o

 

m.m
vv.a
0m.a
ON.o
ma.o
00.0
mo.o
aa.o
©N.0
ma.o
aa.o

.mho.o

 

0.00
00.0
00.0
00.0
00.0
00.0
00.a

000.0

00.a
o0a.a
0000.0
0000.0

'U'U'U'U'U'U

n

.> magma EOHM anc wuocuoom mmmm

 

 

m.ma

N.Na
mm.v
mN.m
NN.m
mo.m
mm.a
vo.a
mm.a
no.a
ma0.o
amv.o

 

MKOOKOMQ‘VUIVMKOM

 

mmm.o
Nv.v
N®0.0
mma.o
ooa.o
mm.v
mv.m
mmm.o
Noo.o
mwa.o
mmm.o
mv.m

 

hm.o
om.o
ON.O
hm.o

vn.a.

mm.a
hm.a
va.o
ma.o
®®.a
mo.N
av.m

 

m.mvv

m.w¢v

v.ama
mm.mm
mm.©m
wh.mm
®©.bm
mh.m
on.m
mm.v
mm.v

mm.m

88

 

IOO

IO

fol”): IO'3 (Md soc...)

 

.Ol

 

 

11! i 111] i 11111 11111

 

Figure 17.

.I I. I0 I00
feel: 100-3 (In-'13:")

Log-log plot of fobs gs fcalc: (O) fobs at
various extents of reaction; (0) fobs from

initial rates; ([3) fobs from parallel first-

and second-order fit of entire decay curve.

TABLE IV

89

Calculated Constants from Equation (70)

 

(a)
with EtOH

(b)

with all proton donors.

 

2 Q

 

(5.33i0.32)x 10

0.3l6i0.063

(l.25:0.07)x.10

3

5

M—lsec

M—

2

SEC

-1

-l

 

(6.09:0.7l)x 10

0.35:0.14

(l.33i0.08)x 10

3

5

M-lsec

M'Zsec

-l

-l

 

90

.m>H magma ca nm>am mum soap
imasoamo mac» MOM @005 ch coaumavm mo mwamumcoo was .mm>uno pmuma
isoamo pcmmmnmmu cllic mwcaa caaom . amoum_ mo mmsam> uswmwmmav

 

 

mmunu um has ca moum nua3.b:<+m mo coauommu 0:» How mfiau m> U\o .ma mhsmam
2.03 05:.
On. nu. ON. 3. . Op. 80. O
u q q u u . O
11.11, 0
O

 

 

..i . a,
/ .
(7.

,

‘0

 

fi’D
0
HIV

 

91

the anthracene dianion, or speed up the protonation rate of
the ion-cluster intermediate species, because of the proximity
of ROH to the aromatic anion when the ROH is complexed to K+.
Cram and co-workers (67) proposed a "cation solvation" of ROH
for proton or deuteron exchange on carbanions. According to
this mechanism, the potassium cation may experience the E

substitution of ROH for one molecule of THF in the primary

 

solvation layer.

K
+
K + ROH ——§+v K+.ROH (71a)
.4...—
or
1- + KS 1- +
An ,K + ROH -—9> An ,K..ROH (71b)

This cation solvation could increase the charge localization
on An? in the vicinity of the cation. If this increases the

formation rate of the anthracene dianion, then one obtains the

 

expression: - T
_2 2k+ + 2Kska[ROH]
f(An,ROH) = (1 + KS[ROH]) k_ [An] (72)
1 + F— [ROH]

 

 

L 1 ~
It is necessary to assume that the same anthracene dianion
species is formed. In this expression ka is the accelerated
dianion formation rate constant. On the other hand, if the
cation-solvation helps to increase the protonation rate of

the ion-cluster, then:

92

 

_ "2 2k1KQ
f(An,ROH) — (1 + KS[ROH]) far— + ZKQKSké[ROH]
k: [An]
1 +__ _———
' kl [ROH] L

 

 

(73)

Equations 70, 72 and 73 turned out to be nearly the same as
far as the least-squares fitting routine is concerned, since
they all contain 3 parameters, and KS[ROH] is apparently I

small compared to unity. This was tested by fitting the data :

 

with Equation 72. Moreover, when [ROH] is high enough to make
the first-order term in [ROH] predominant, the value of the
term which contains [Anl/[ROH] becomes small compared to unity.
In order to try to distinguish between the various
mechanisms, we used the re-analyzed data of Minnich for the
protonation of K+Anley EtOH, MeOH, i-PrOH, t-BuOH and H20 in
THF. When we used both Minnich's data and ours to test the
dependence of the pseudo second-order protonation rates upon
the nature of the alcohol used, we found, not without surprise,
that the nature of the proton donors did not influence the
protonation rate significantly. This is shown in Figure 19.
We would like to emphasize here again that Minnich did not
cover a wide range of concentrations of ROH and An, and,
except for ethanol, the data in Figure 19 were obtained at
values of [ROH] above that for which the dianion mechanism is
valid. In order to include the data for ethanol on this
graph, the rate constants are plotted XE [ROH]/[An] rather

than [ROH] alone. However, in order to do this, the ethanol

93

N ..mma ....n moum .00
£00.. .4. £500.... :3 .0 ..L4. 3.00010 .0000 50...: 0:50. 00
mcoauommu map How ~:40\amom0 m> amm©H0i©soommv x no uoam mOaimoa .ma musmam

 

 

...<<c.o.:
oo. o. .. .. .0. _oo.
...4. 7..aa . .20 . _..fi . .....
\..
O 1.. x.
.d
1 u
n
.. ...
M X
1.. o.
e
.. w
....
l S
I O
I 3-
12(1
J
I.

 

 

 

94

data for which both [EtOH] and [An] were high could not be
included in Figure 19. This figure shows us that all the data
points fall on nearly the same curve, independent of the
nature of the alcohols used. This apparent insensitivity of
the protonation rate to the nature of ROH indicates that the
breakdown of the dianion mechanism at high [ROH] may be due :‘

to a modified cation solvation mechanism. If we add to

 

reactions:
1" + k+ = + '
2(An ,K ) -——+-An ,2K + An (55a) .
.4__
k- g»
= + k1 + - + -
An ,2K + ROH ——> K AnH + K OR (55b)
0 + Ks . +

the reactions which effectively yield:

k
+ - + + - + -
AnT,K + An ,K .ROH J?» K AnH + K R0 + An (74)

we can derive another general rate law:

5 1
2k+

 

f(An,ROH) = (1 + Ksinom)’2 + 2k KS[ROH]

k [An] d

+ _: __
kl[ROH]

 

 

L1

_J

(75)

The reaction described by Equation 74 might have the anthra—
cene dianion as an intermediate species which would then be

protonated by the cation-solvated alcohol K+cROH. Step (55b)

95

should be very fast (nearly diffusion controlled) in order
to explain the insensitivity to the nature of ROH. Equation
75 has the same form as Equation 73. And since KisOH]<< l
as expected, Equation 75 is similar to Equation 70 from the
dianion and ion-cluster mechanism. When we used Equation 70
to fit the combined data for all proton donors in THF:

- data for EtOH only (Table III)

— Minnich's data re-analyzed (Table V)
we obtained the new values of the 3 parameters of Equation 70
shown in Table IVb. They are nearly the same as the corre-
Sponding constants in Table IVa. The goodness of fit is

indicated by the standard deviation of f /fob . With the

calc s

data from Table III and the values of the 3 parameters from

Table IVa, we obtained a standard deviation of fC of

alc/fobs
0.27. When all of the data from Tables III and V are com-

bined, and the values of the constants from Table IVb are

used, we get a standard deviation of f /fOb of 0.36.

calc s

This shows once more that the nature of the proton donor did

not strongly affect the rate of protonation.

5.2--Discussion

 

5.2.l--Comparison of the T 9 Proposed Mechanisms

 

Now that we have proposed alternate mechanisms for the
reaction of K+AnT with alcohols in THF, it is in order to ask
which mechanism might be preferable. The answer to this

question might lie in the apparent insensitivity of the

96

039

.cmam 0p £05m Eoum >uaaanaosvoummu man no mudmmma m 00>am mane .

I i . if»... 1.1.5.5”. It

.Uaa0> an no: 005 moauwcax “memo ovummm cowonu sm>w paw HwGMOicsoomm Opsmmm @000

v

.0000 mo paw amoanamEm Eonw ucmumsoo mumu amauacao

....

.m mo Houomm m an
umaamam maaMHmcmm ma mm>uso amomc a05©a>apca msauuam an cwCaauno .m mo coauma>mc Unmcsmum Uwumsaumw

...

. . H and Hi:
. Hr ' I H i
.01. l .0; .0. a0 0

 

 

 

 

 

 

 

m\a n
.0005 003 Ed omh um mca x 0.m mo unmaoamwmoo coauUCauxm cm .ma.c¢0xi u uU\a smavccom: vwmmmm
0).... 0 000 . a 0 00-0.! 0. 0 1...,00 . 0
«0.0 000.0 0 a0.a v.0 mmo.c mow:
00.0 00.0 0 0mm... Iii 0.0 00.0
cm.c a0.a m ch.a m.m om.o .
ma.o 0m.a m m0.a 0.0 00.0 mosmiu
mm.o mv.v 0 mm.a m.m ov.aa
mm.c am.m 0 mm.a 0.0 om.m
aa.o cc.m m mm.a «.0 om.a moumia
00.0 m~.m m 0m.m m.v om.mm
am.c 0m.m v 00.0 v.~ ca.aa
mm.o 0m.N m 0m.m m.m co.m
0m.o 00.0 m 0a.a o.m a0.v
aa.o om.m v 00.0 0.m m0.m m
ma.o ma.a m mo.a m.m av.c o m
0a.o mm.a v 00.0 m.m 000.0
mm.o 00.0 0 00.0 v.m aav.o
Na.o mm.c 0 00.0 m.m omm.o
aa.o mm.m v mm.o m.m Nam.o moum
. mmnmsm o o o
omm Ev oax o omm z oax 0 mm 2 z x cm 2 x cm x z x mom
0.... .. 0- ..- 7.0- .... 0 95 mo. . _ 00. .... +0 00. . 0

 

 

 

 

 

 

 

 

m Amumv vmuhaMGMiwu m.noaccaz Eoumc

8.. + imo+o.0 + 05. a 0000 + 156.0 0030000 0n... .80 0000000000 0000 000001000000 000000

> quflfi

97

protonation rate to the nature of the alcohol used eSpecially
at relatively high alcohol concentrations. If the "dianion and
ion-cluster mechanism" is correct then we can reason as
follows: Equation 70 derived from Equations 61 and 62, shows
(that at high [ROH], the second term is predominant. Since

this term arises from protonation of the ion-cluster, at high
values of [ROH], protonation of the ion-cluster rather than

protonation of the dianion becomes the predominant pathway.

On the other hand, from the estimated value KQ = 50 M.1 and
the computed value k2KQ = 1.25 x lOSM-zsec-l (Table IVa),
we can estimate k 2 2.5 x 103M-lsec-l. This small value of

2
the protonation rate constant for the ion-cluster would pre-

dict a strong influence of the acidity of the alcohols on the
protonation rate. Minnich's data combined with ours (Figure
19) showed no such effect. However, in contrast, the cation-
solvation mechanism provides a different picture. The proton
donor is already present in the ion-cluster as K+.ROH. Hence,
the so-called intra-complex protonation might be much faster
since the interaction of ROH with K+ and with the negative
charge on the aromatic system could very well make the alcohol
more acidic. If the intra-complex protonation rate was suf—
ficiently fast, the system need not be affected strongly by the
acidity of the proton donors. Therefore, the cation solvation
mechanism in combination with the dianion mechanism seems to

be preferable when compared with the dianion and ion-cluster
mechanism, at least at relatively high concentrations of proton

donors.

98

5.2.2--Discrepancies Between Some of Minnich's
Results and Those ofIBank and Bockrath—

 

 

It may be profitable to take a closer look at some
apparent discrepancies which have occurred between the results
obtained in this laboratory and some which have been recently
published.

As described in the Historical section, Minnich (48)
and Bank and Bockrath (50) independently studied the protona-
tion of Na+AnT with H20 in THF. While Bank and Bockrath
reported only a pseudo first-order behavior of the protonation
rate (first-order in the absorbance) with a rate constant
2k = 1328 M-lsec-l, Minnich found in one attempt a mixed
pseudo first- and second-order behavior with a relatively
large first-order contribution, and in another attempt, only
a pseudo second-order decay. This inconsistency forced Minnich
to eliminate this reaction from quantitative consideration.

On the other hand, although the experimental details are
sketchy, Bank and Bockrath reported in their experimental part
(50) that considerable (up to 90% or more) decomposition of

the sodium naphthalenide solution occurred in the syringe prior
to data collection. This fact suggests that perhaps the

pseudo first-order decay observed by Bank and Bockrath and in
one of his runs by Minnich was caused by the reaction of

Na+AnT with impurities. Alternatively, impurities could cata-
lyze the protonation reaction. If this were the case, then

the reported value of the rate constant (2k = 1328 M‘lsec—l)

99

would be larger than the true one. The reported rate constant
predicts a maximum half-life of 32 msec for [H20] = 0.016 M.

At about the same water concentration, a typical push from
Minnich's data for the run which showed little first-order decay,
recorded successive half-lives of 180, 450 and 700 msec starting
from [Na+AnT] = 1.2 x 10-4 M. However, for a typical push

from the run which showed a large first-order contribution,

we found successive half-lives of 26, 34 and 26 msec for

[H20] = 0.047 M. At this water concentration, Bank and Bock-
rath's results predict a half life of 11 msec. Thus, it seems
to us that the pseudo first-order protonation rate reported

by Bank and Bockrath is too fast to be reliable. It is, of
course, possible that impurities in our system inhibit the
first-order reaction. The factors which accelerate or inhibit

the first-order protonation deserve further study.

5.3--Summary

In summary, our results show that the disappearance of
K+Anr in its reaction with EtOH in THF is second-order in the
absorbance. At low concentrations of ROH, the major protona-
tion pathway is gig the dianion intermediate species. At
higher concentrations of ROH, the rate becomes less sensitive
to the anthracene concentration while remaining largely
second-order in K+Anr. This suggests the participation of an
ion-cluster (AnT,K+)2, or "solvent caged complex" (52)

= + . . . . .
(An ,2K ,An) in the protonation process. This intermediate

100

might be directly protonated or it may yield a dianion species,
An=,2K+ and an anthracene molecule by rapid electron-transfer
prior to the protonation step. It is likely that one of
these intermediate species can be rapidly protonated by a
cation-solvated ROH molecule (K+.ROH), which is present with-
in the complex.

We also found that direct protonation of the contact ion- E
pairs AnT,K+ by ROH is slower than the second-order processes

at the concentrations which we used.

 

'Fz _- '.

The crown effect strongly indicates that the pseudo
second-order pathway requires formation of contact ion-pairs.
Also the pseudo first-order.component obtained with the pres-
ence of crown is much smaller than the corresponding one ob-
tained without crown.

5.4--Conclusion and Suggestions
for Further Work

 

 

In the concluding part of this work, we would like to
present a general picture of the protonation reaction of
aromatic radical anions with alcohols, by combining our results
with the results found in the recent literature. This general
view is that the protonation rate depends not only upon the
acidity of the proton donor, but also on the degree of charge
localization in the aromatic system. We expect the charge
localization to decrease in the following order:

1) dianion (An=,2M+)

. +
2) ion-cluster (An ,M )2

101

3) contact ion-pair (AnT,M+)
4) solvent-separated ion-pair (An? || M+)
5) free or solvent solvated ion (Anr).
Let us denote the protonation rate constants (with EtOH) of

these species by k1' k2, ... k5, reSpectively. The results

4 l “1

obtained in DMF as solvent (45) give k k 2 x 10- M- sec .

5
+ -.- . . . .
Our result for K An With crown in THF allows an estimation

of k4.¥ 2M-lsec-l. The data of Rainis, Tung and Szwarc (52)

for DME as solvent give k % 6M-lsec-l (provided that we

4
+ 0- . .
assume that Na An eXists in DME largely as the solvent-

separated ion-pair). From our data, k3 R 200 M-lsec-1 at

high values of [EtOH] but we are not sure that the direct
protonation reactions are even bimolecular. The protonating
agents may well exist under aggregated forms (ROH)n (46,52).
An estimation of the value of k2 is somewhat more compli—
cated, since it depends upon the nature of the protonated
species. The ion-cluster (AnT,K+)2 may transform to (An=,
2K+,An), the "solvent caged complex" (52) prior to protonation.

In this case, the protonation rate constant k is expected to

2
be near the diffusion controlled limit. On the other hand,

if the protonated species is the ion-cluster itself, then
3 l -1

k2 & 2.5 x 10 M- sec (section 5.2.1).
k
Finally, Table IVa gives ii): 0.3, and k+ (or k+KQ) =

5 x 103M-lsec_l. Potentiometric data (68) yield a value of

about 1 x 10.5 for the overall disprOportionation equilibrium

constant for the reaction:

102

k

2(An ,K ) -}—(-—> An ,2K + An (55a)
- ii "" 8 -]- -l 4’
in THF. Thus k_ (or k_) = 5 x 10 M sec . Hence k =

1
1.5 x 109 M-lsec—l.

From all the values of the protonation.rate constants

l 2 3 4
agrees well with the degree of charge localization in the

obtained above, we can see that the order k In>k > k > k > k5 E

aromatic system:

= + 1- — 1"
(An ,2M ) > (An ,M+)2 > (Anni?) > (An ll M+) > (An?)

 

For further work in this area, we would like to propose
a careful and extensive study of the system Na+AnT with H20
in THF in order to clarify the discrepancies between the
results of Minnich on one hand and those of Bank and Bockrath
on the other hand. Also, in order to obtain a better under-
standing of the "solvent effect", a study of one system over
the entire range of solvent composition would be very inter-
esting. The system DMF-ethanol would be appropriate since
only the first-order process was observed (46). Pulse-
radiolysis techniques, the stopped-flow method and conventional

studies would be required to Cover the high, intermediate and

low ethanol concentration cases.

ll.

12.

13.

14.

REFERENCES

W. Schlenk, J. Appenrodt, A. Michael, and A. Thal,
Chem. Ber., 61, 479 (1914).

 

N. D. Scott, J. F. Walker, and V. L. Hansley, J. Amer.
Chem. Soc., 66, 2442 (1936). i

 

D. E. Paul, D. Lipkin, and S. I. Weissman, J. Amer. i
Chem. Soc., 16, 116 (1956).

 

.A. I". I'_.- 0 ."'1‘_A

P. Balk, G. J. Hoijtink, and J. W. H. Schreurs,
Rec. Trav. Chim., 16, 813 (1957). i

 

 

E. DeBoer, and S. I. Weissman, Rec. Trav. Chim., 16,
824 (1957).

 

G. J. Hoijtink and P. J. Zandstra, Mol. Phys., 2, 371
(1960).

 

K. H. J. Buschow, J. Dieleman, and G. J. Hoijtink,
Mol. Phys., 1, 1 (1963-64).

 

H. A. Laitinen, and S. Wawzonek, J. Amer. Chem. Soc.,
64, 1765 (1942).

 

A. Maccoll, Nature, 163, 178 (1949).

 

G. J. Hoijtink, and J. Van Schooten, Rec. Trav. Chim.,
2;, 1089 (1952).

 

G. J. Hoijtink, J. Van Schooten, E. DeBoer, and W. Ij.
Aalbersberg, Rec. Trav. Chim., 26, 355 (1954).

 

G. J. Hoijtink, E. DeBoer, P. H. Van Der Meij, and W. P.
Weijland, Rec. Trav. Chim., 16, 487 (1956).

 

T. L. Chu and S. C. Yu, J. Amer. Chem. Soc., 16, 3367
(1954).

T. R. Tuttle, Jr., R. L. Ward, and S. I. Weissman,
J. Chem. Phys., 36, 189 (1956).

 

103

104

15. E. DeBoer, J. Chem. Phys., 26, 190 (1956).

 

16. E. DeBoer, and S. I. Weissman, J. Amer. Chem. Soc., 80,
4549 (1958).

 

17. R. L. Ward, and S. I. Weissman, J. Amer. Chem. Soc.,
12, 2086 (1957).

 

 

18. N. M. Atherton, and S. I. Weissman, J. Amer. Chem. Soc.,
.83, 1330 (1961).

19. K. H. J. Buschow, J. Dieleman, and G. J. Hoijtink,
J. Chem. Phys., 42, 1993 (1965).

 

 

20. A. Crowley, N. Hirota, and R. Kreilick, J. Chem. Phys.,
4_J 4815 (1967).

21. C. L. Dodson and A. H. Reddoch, J. Chem. Phys., 46,
3226 (1968).

 

22. R. V. Slates, and M. Szwarc, J. Phys. Chem., 69, 4124
(1965). "

 

23. T. E. Hogen—Esch, and J. Smid, J. Amer. Chem. Soc., 66,
307 (1966).

 

24. P. Biloen, T. Fransen, A. Tulp, and G. J. Hoijtink,
J. Phys. Chem., 16, 1581 (1969).

 

25. N. Hirota, R. Carraway, and W. Schook, J. Amer. Chem.
Soc., 96, 3611 (1968).

 

26. T. E. Hogen-Esch, and J. Smid, J. Amer. Chem. Soc., 66,
318 (1966).

 

27. N. Hirota, J. Amer. Chem. Soc., 26, 3603 (1968).

 

. Chang, R. V. Slates, and M. Szwarc, J. Phys. Chem.,
0, 3180 (1966).

w

 

28. P
7

29. W. Schlenk, and E. Bergman, Ann., 463, 134 (1928).

30. W. E. Bachman, J. Org. Chem., 1, 347 (1936).

 

31. J. F. Walker and N. D. Scott, J. Amer. Chem. Soc., 66,
951 (1938).

 

32. A. P. Krapcho, and A. A. Bothner-By, J. Amer. Chem. Soc.,
81, 3658 (1959).

 

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

105

A. P. Krapcho, and A. A. Bothner-By, J. Amer. Chem. Soc.,
66, 751 (1960).

 

O. J. Jacobus, and J. F. Eastham, J. Amer. Chem. Soc.,
61, 5799 (1965).

 

E. J. Kelly, H. V. Secor, C. W. Keenan and J. F. Eastham,
J. Amer. Chem. Soc., 66,-3611 (1962).

 

W. L. Jolly, "Solvated Electron", Advances in Chemistry
Series No. 50, American Chemical Society, Washington,
D. C. (1965).

A. M. Kahn,.and R. R. Dewald, Proceedings of Colloque
Weyl III, Hanita, Israel, June 1972 (in press).

J. P. Keene, E. J. Land, and A. J. Swallow, J. Amer.
Chem. Soc., 61, 5284 (1965).

 

S. Arai, and L. M. Dorfman, J. Chem. Phys., 61, 2190
(1964).

 

L. M. Dorfman, ACC. Chem. ReS., 6, 224 (1970).

 

S. Arai, and L. M. Dorfman, J. Phys. Chem., 66,-2239
(1965).

 

J. R. Brandon, and L. M. Dorfman, J. Chem. Phys., 66,
3849 (1970).

 

K. Umemoto, Bull. Chem. Soc. Jap.,.g6, 1058 (1967).

 

S. Hayano and M. Fujihira, Bull. Chem. Soc. Jap., 66,
1496 (1971).

 

S. Hayano, and M. Fujihira, Bull. Chem. Soc. Jgp., 66,
2046 (1971).

 

M. Fujihira, H. Suzuki, and S. Hayano, J. Electro. Anal.
Chem., 66, 393 (1971).

 

E. R. Minnich, and J. L. Dye, Abstracts, lGlst National
Meeting of the American Chemical Society, Los Angeles,
California, March 1971.

E. R. Minnich, Ph. D. thesis, Michigan State University,
(1970).

S. Bank, and B. Bockrath, J. Amer. Chem. Soc., 66, 430
(1971).

 

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

62.

63.

64.

65.

66.

67.

106

S. Bank, and B. Bockrath, J. Amer. Chem. Soc., 94, 6076
(1972). '—

 

G. Levin, C. Sutphen, and M. Szwarc, J. Amer. Chem. Soc.,
66, 2652 (1972).

 

A. Rainis, R. Tung, and M. Szwarc, private communication.
To be published in J. Amer. Chem. Soc.

E. R. Minnich, L. D. Long, J. Ceraso, and J. L. Dye, to
be published in J. Amer. Chem. Soc.

E. M. Hansen, Ph. D. thesis, Michigan State University
(1970).

L. H. Feldman, Ph. D. thesis, Michigan State University
(1966).

M. G. DeBacker, Ph. D. thesis, Michigan State University
(1970).

J. L. Dye, and V. A. Nicely, J. Chem. Ed., 12"443'
(1971).

 

V. A. Nicely, Ph. D. thesis, Michigan State University
(1969).

C. J. Pedersen, J. Amer. Chem._Soc., 66, 7017 (1967).

 

J. L. Dye, M. T. Lok, F. J. Tehan, R. B. Coolen, N.
Papadakis, J. M. Ceraso, and M. DeBacker, Ber. Bunsen
Gesell. Phys. Chem., 16, 659 (1971).

 

 

K. H. Wong, G. Konizer, and J. Smid, J. Amer. Chem. Soc.,
2_2_, 666 (1970) .

 

R. M. Fuoss, and C. A. Kraus, J. Amer. Chem. Soc., 61,
1 (1935).

 

R. M. Fuoss, J. Amer. Chem. Soc., 66, 5059 (1958).

 

M. Szwarc, "Carbanions, Living Polymers and Electron-
Transfer Processes", Interscience, New York, 1968.

R. M. Fuoss and C. A. Kraus, J. Amer. Chem. Soc.,-661'
2387 (1933).

 

S. Hayano, and M. Fujihira, Bull. Chem. Soc. Jap., 66,
2051 (1971).

 

S. M. Wong, H. P. Fisher, and D. J. Cram, J. Amer. Chem.
Soc., 66, 2235 (1971).

 

107

68. J. Jagur-Grodzinski, M. Feld, S. L. Yang and M. Szwarc,
J. Phys. Chem., 66, 628 (1965).

 

 

‘Ilfim
-*.

 

WHEN“

7405

   

"TIT{I‘Mflflfl‘tflnflfiflfljflfli