arm ‘

_ - I " I I —' M‘ r“.— “—‘A :r‘u-ufflfinz—g-qfi-n‘rdh

— - V"- V‘ V'-

 

 

 

PART I
$TUDY OF SOME ISOMERS OF THE
1, ‘E-SF’EIICJCYCLCI'IEIIC445.NE-A4
-TETRAHYDRQINDAN-3-ONE SYSTEM
PART II
INVEgTEG-AHQH OF THE THERR’IEAL AND
FREE RADICAL REACTIONS OF PHENYL
SUBSTITU‘FED OXIRANES

Thesis For Hm Degree of DL. D.
MICHIGAN STATE UNIVERSITY
Ronald Weiss
1964

(“6.55:

Part I:

Part 11:

0-169

 
  

This is to certify that the

thesis entitled
Study of Some Isomers of the 1,1-Spirocyclo-
hexane-A -tetrahydroindane- --3 -one System
Investigation of the Thermal and Free Radical
Reactions of Phenyl Substituted Oxiranes

presented by

Ronald Weiss

has been accepted towards fulfillment
of the requirements for

Ph. D. degree in Chemistry

 

MIL/W

Major professor

Date October 8, 1964

 

“In-1.31am .1- ._

LIBRAP v ‘*
buCul 1’3”“ *

lfibbw ~vv

' I
Sufi“ 09—14‘ 0 ;.
“'5" ‘we. .:.. L

11:9".
fiil'iiifitfl‘fiwnn :(

u‘fl‘fllflm

 

ABSTRACT

PART I

STUDY OF SOME ISOMERS OF THE
l,l-SPIROCYCLOHEXANE-A4-TETRAHYDROINDANE-3-ONE SYSTEM

by Ronald Weiss

Marvel's cyclization products have been reinvesti-
gated and shown unequivocally to have the same carbon
skeletons, and previous structural assignments are incorrect

and must be reversed.

PART II

INVESTIGATION OF THE THERMAL AND FREE RADICAL
REACTIONS OF PHENYL SUBSTITUTED OXIRANES

by Ronald Weiss

'Four phenyl substituted oxiranes (i.e., styrene

oxide, trans-stilbene oxide, a-methyl-trans-stilbene oxide

 

and tetraphenylethylene oxide) were studied at high
temperatures (T > 1500). They were investigated as neat
samples and with the free radical initiators tertiary butyl
hydroperoxide and 2,4,4—trimethy1-2-formylcyclopentanone.
Some of the earlier work reported on these oxiranes

was reinvestigated and the results clarified.

Ronald Weiss

None of the oxiranes studied showed a true unimolecular
thermal reaction.

Substitution of phenyl groups on the oxirane ring
retards radical abstraction of benzylic hydrogens.

The product of surface reactions were explained in
terms of mild acid catalysis by Lewis acid sites on the

Pyrex glass surface.

PART I

STUDY OF SOME ISOMERS OF THE

L
l,l—SPIROCYCLOHEXANE-A4-TETRAHYDROINDANL3-ONE SYSTEM

PART II

INVESTIGATION OF THE THERMAL AND FREE RADICAL

REACTIONS OF PHENYD SUBSTITUTED OXIRANES

BY

Ronald Weiss

A THESIS

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

DOCTOR OF PHILOSOPHY
Department of Chemistry

1964

To Tante

ACKNOWLEDGMENTS

The author wishes to express his appreciation for
the guidance and assistance given him by Dr. William H.
Reusch throughout the course of this investigation.

Grateful acknowledgment is extended to the Research
Corporation of America for a special graduate research

assistantship from January. 1960, through January, 1961.

iii

TABLE OF CONTENTS
Page
PART I
STUDY OF SOME ISOMERS OF THE 1.1-

SPIR0CYCL0HEXANE-A4-TETRAHYDRDINDAN-
3—ONE SYSTEM 0 o o o o o o o o o o o o o o o o o o 1

INTRODUCTION AND HISTORICAL . . . . . . . . . . . . . 2

RESULTS AND DISCUSSION . . . . . . . . . . . . . . . . 6

EXPERIMENTAL . . . . . . . . . . . . . . . . . . . . . 9

SUMMARY . . . . . . . . . . . . . . . . . . . . . . . 11
PART II

INVESTIGATION OF THE THERMAL AND FREE RADICAL
REACTIONS OF PHENYL SUBSTITUTED OXIRANES . . . . . 12

INTRODUCTION AND HISTORICAL . . . . . . . . . . . . . l3
“SULTS O O O O O O O O O O O O O O O O O O O O O O O 18
I. Styrene Oxide 18

II. Trans-Stilbene Oxide 23
III. d-Methyl—Trans-Stilbene Oxide 26
IV. Tetraphenylethylene Oxide 28
DISCUSSION . . . . . . . . . . . . . . . . . . . . . . 3O
EXPERIMENTAL . . . . . . . . . . . . . . . . . . a . . 36
I. General Procedures 36

A. Apparatus 36

B. Furnace 37

C. Preparation of Pyrex Sample Vials 37

D. Degassing 37

E. Nitrogen 39

F. Handling of Samples 39

iv

Page

II. Synthesis of Oxiranes 40
III. 1,2-Diphenylallyl Alcohol 40

SUMMARY . . . . . . . . . . . . . . . . . . . . . . . 42

LITERATURE CITED . . . . . . . . . . . . . . . . . . . 43

Table

II.

III.

Iv.

Thermal and free radical reactions of
styrene oxide

Thermal and free radical reactions of
trans-stilbene oxide

The chemical shifts (T) for the
characteristic protons used for analysis
of a—methyl-trans—stilbene oxide and its
likely decomposition products

Thermal and free radical reactions of
a-methyl-trans-stilbene oxide

Thermal reactions of tetraphenylethylene

oxide

LIST OF TABLES

vi

Page

22

24

26

27

29

LIST OF DIAGRAMS
Diagram Page
I. Decomposition modes available to the oxirane
radical generated by a-hydrogen

abstraction . . . . . . . . . . . . . . . 16

II. The formation of intermediate ions on the
surface of Pyrex glass . . . . . . . . . . 31

vii

LIST OF FIGURES

Figure Page

I. % Transmittance at 5.78 u gs. moles/liter
of phenylacetaldehyde . . . . . . . . . . 20

II. % Transmittance at 11.43 u gs. moles/liter
of styrene oxide . . . . . . . . . . . . . 21

viii

PART I

STUDY OF SOME ISOMERS OF THE

l,l-SPIROCYCLOHEXANE-A4-TETRAHYDROINDAN-3-ONE SYSTEM

INTRODUCTION AND HISTORICAL

During the course of a general study of dieneyne

cyclization.l Marvel t l. attempted to prepare All-

dodecahydrophenanthrone-9 (I) from cyclohexanone and

acetylene through the following sequence of reactions.

0 OH OH
II + BngC 5 CMgBr -—-—-! .—C 5 C
I
HCOZH C E C
II + +_____.

\\0

Two isomeric ketones, a solid I (melting point 940)
and a liquid II.were obtained. The structural elucidation
of II was carried out by means of the Clemmensen reduction
to give a compound, C14H22, whose properties agreed closely
with those of a dodecahydrophenanthrene which had been
prepared by the reduction of phenanthrene with phosphorous
and iodine. Other reductions using Raney Nickel gave
products with properties agreeing with those reported
for perhydrophenanthrenes.

When the resulting dodecahydrophenanthrene was
partially dehydrogenated by heating with selenium, the product

was found to be identical with a known octahydrophenanthrene.2

2

The dodecahydrophenathrene was also dehydrogenated
over platinum on a charcoal catalyst at 3000 giving
phenanthrene as the only product.3

Reinvestigation of Marvel's cyclization product II
by R. P. Linstead gt _l.4 showed this to be a low melting
(melting point 390) ketone, isomeric with II. Attempts to
prove the skeletal configurations employed Clemmensen
reduction and vapor phase dehydrogenation. Compounds ob-
tained by these means were compared with those having the
phenanthrene skeleton.

The position of the keto group was indicated by
hydrogenation over palladised charcoal followed by dehydro—
genation in the vapor phase.

The Linstead group found the lower melting isomer
to be more stable than the higher melting isomer. Subse-
quent research by I. N. Narasov and I. I. Zaretskaya5
showed that longer heating of the dieneyne produced a higher
percentage of the lower melting isomer; also, when a crude
mixture of the two iSomers was heated, only the lower melting
isomer was found. Furthermore, heating the higher melting
isomer with acetic acid and hydrochloric acid for eight
hours also produced the lower melting isomer. Correlating
these findings with data already available on hydronaphthalenes
led to the assignment of structure III to the more stable
isomer (m.p. 39°) and structure IV to the higher melting

isomer (m.p. 94°).

o
o\\ \\
III IV

A further study of the isomeric ketones III and IV
by M T. Bogert at _al.6 introduced the pDSSibility that
these compounds are not hydrogenated phenanthrones but
isomeric spiranones (this possibility was earlier rejected
by Marvel7). Support for the required rearrangement of
these spiranes was found in the reaction of cyclohexanespiro-
cyclopentane with selenium at 2800-320O to give naphthalene.

The spirane skeleton proposed by Bogert was later
accepted by Linstead gt al.8, for this best supported their
experimental results.

0
II

 

Structure VII was eliminated as a possibility because
it could not give the same ketone on hydrogenation as
either V or VI. In converting to the spiran system, III is
V, and IV becomes VI.

Schwartzman9 dehydrogenated the 390 isomer over a
palladium charcoal catalyst and produced a ketone having a

melting point similar to that of a C14H220 ketone obtained

by Bogert and co-workershw'an.alternate route. Derivatives

of the dehydrogenated ketone were also similar to those

reported by Bogert, although no mixed melting points were

reported.

RESULTS AND DISCUSSION

Thus far, all of the experimental evidence used to
show that the isomeric ketones were of the same skeletal
configuration, has been obtained under conditions where
Vrearrangement could have occurred (e.g., acidic media and
lmetallic catalysts at high temperatures). Interpretation
of such results has consequently been ambiguous. It was
therefore necessary to find a reaction where rearrangement
would not occur. The reduction of a,B—unsaturated ketones
to the saturated ketone by lithium in liquid ammonia is

known to be mild.10

A generalization of the outcome of

the reaction by Stork gt_‘gl. states that "In reduction of

an octalone system with lithium in ammonia the product will

be the more stable of the two isomers (gi§_or trans) having

the newly introduced hydrogen axial to the ketone ring."11
The action of lithium and liquid ammonia upon the

two isomers in question should therefore proceed along the

following lines,

0 -
4’ le'

 

   

terminating in a common product.
Reduction of the two isomeric ketones with lithium
and liquid ammonia did in fact give products whose infra-
red spectra and vapor phase chromatography retention times
were identical. Melting points and mixed melting points
of the oximes of the reduced products were also identical.
This proves identical skeletal configuration.
12,13

Assignments based upon ultraviolet absorption data

gave the 390 isomer structure VI and the 940 isomer structure

 

V. However'when the correction factor for a cyclopentenone

I.R. (Found) U.V. (Found)
lower melting isomer 5.93 u 241 mu (6 12 x 103)

9 x 103)

higher melting isomer 5.85 u 247 mu (e
system14 was applied to the calculations’the results were so

close that a clear cut assignment could not be made.

V VI
base 215 mu base 215 mu
a-substituent 10 a-substituent 10
ZS-substituent 24 B-substituent 12
(calculated uncorrected)249 mu exo-double bond 5
correction -11 242 mu
238 mu

The key to the problem was nuclear magnetic resonance
spectrometry. There was no olefinic proton absorption
for the lower melting isomer while a broad doublet (T = 3.45)
was found in the vinyl hydrogen region for the higher
melting isomer. This data confirmed the presence of a

tetrasubstituted double bond in the lower melting isomer.

Nuclear magnetic resonance spectrum of:

 

 

 

 

 

Structure V is therefore the lower melting isomer

and VI is the higher melting isomeric ketone.

EXPE RI MENTAL

l,1-Spirocyclohexanehexahydroindan-3—oxime

 

4(9)—tetrahydroindan—

290 mg. of 1,1-spirocyclohexane—A
3-one16 in 50 ml. of a dried 50:50 mixture of ethyl ether
and dioxane was added to a flask containing 20 mg. of
lithium metal in 25 ml. of liquid ammonia. The cooling
bath used to liquefy the ammonia was removed, and the
solution was stirred with the aid of a mechanical stirrer.
After 5 minutes the reaction was quenched with dry tertiary—
butyl alcohol. The solution was allowed to stir until all
the ammonia had evaporated. The residue was extracted with
ether several times; the ether extracts were washed with
water and dried over anhydrous magnesium sulfate.

The ether was removed and the resultant residue taken
up in benzene. The benzene solution was then run through a
chromatography column containing Merck acid washed alumina
in benzene. The eluted benzene solution was collected and
stripped of solvent. An oxime was prepared from this material
giving 1,l-spirocyclohexanehexahydroindan-B—oxime. The
oxime, m.p. 138—1400, was recrystallized from water-ethanol.

Anal. Calcd. for C14H23NO: C, 75.96; H, 10.47; N, 6.33.
Found: C, 76.09; H, 10.49; N, 6.31.
The same procedure was used for converting 1,1-

4(5)

spirocyclohexane-A tetrahydroindan-3-one to

9

 

lO

1,1—spirocyclohexanehexahydroindan-3-oxime. A mixed melting
point of the oximes of the reduced product of the two isomeric
ketones showed no depression.

Infrared spectra were obtained on a Perkin-Elmer,
Model 21, recording spectrophotometer, using sodium chloride
optics. The ultraviolet spectra were determined in 1—cm.
quartz cells using a Cary, Model 11, spectrophotometer.

Proton magnetic resonance spectra were determined
in carbon tetrachloride solution using a Varian, A-60,
high resolution spectrometer. All spectra were obtained at

60 MC using tetramethylsilane as an internal standard.

 

SUMMARY

Marvel's cyclization products have been reinvesti-
gated and shown unequivocally to have the same carbon skeletons.
The use of proton nuclear magnetic resonance spectroscopy
was the key factor in showing that previous structural

assignments are incorrect and must be reversed.

ll

 

PART II

INVESTIGATION OF THE THERMAL AND FREE RADICAL

REACTIONS OF PHENYL SUBSTITUTED OXIRANES

INTRODUCTION AND HISTORICAL
The oxirane system is well-known for its ability to

react and undergo rearrangement.
rather mild conditions suffice

As a general rule,
118 and 1119)

for acid or base reactions (e.g.,

 

 

 

 

0
n 0 H
BFB-EtZO
g e (i)
5 minutes
Room temp
ép
Et NHLi
::p 2 4% (ii)
10 minutes

and the abstraction of oxirane hydrogen by tertiaryfibutyl
hydroperoxide is reported20 to be rapid at 00 (iii).

0 +oc1
_____1 (iii)
0°, hY cr

Reports of oxirane rearrangement at elevated temperatures
(iv,21 v22 and vi 3) are consequently of more than passing
interest.

13

14

CW)

\ 200° \ (vi)
__—_—q ‘\0
o//

The question arises as to the role the higher temper—
ature plays in the thermal rearrangement of the oxirane system.
Since the intrinsic strain associated with the oxirane
system makes it vulnerable to ring opening under a variety
of conditions, and since early work on this system left
ambiguities concerning experimental procedure, one may ask
if the high temperature rearrangements are indeed caused by
thermal rupture of the three-membered ring. It was toward
this end that this investigation of the decomposition of the
oxirane system was undertaken.

In much of the early work concerning decomposition
of oxirane systems, experimental procedures were followed
which gave equivocal results. The following points are of
particular importance: (1) Yields were often poor and only
the major product was isolated. (2) The carbonyl compounds
produced in the decomposition of the oxirane systems were
characterized as hydrazones, oximes or semicarbazone.

Preparation of those derivatives usually involves an acid

15

medium which could cause rearrangement of any unreacted
material. (3) The presence of oxygen may give rise to
radical species which could then promote further reaction.24
(4) Reactions may occur at the surface of the reaction vessel.
Rondestvedt25 has demonstrated that acidic sites on silica
gel catalyze the decomposition of 1,3-dioxanes. A similar
catalysis could presumably occur at acidic sites on the
surface of the Pyrex reaction ampule. No previous attempt
has been made to differentiate between this possibility and
true thermal reactions; consequently one must proceed with
caution when interpreting data obtained by early investi—
gators.

Previous studies of oxirane systems at high tempera-
tures suggest that at least three types of reactions can be
observed: (1) a free radical chain reaction initiated by
hydrogen abstraction, (2) a unimolecular decomposition
involving rupture of an oxirane bond and (3) a heterogeneous
reaction catalyzed by acidic sites on the surface of pyrex
glass.

Several decomposition modes (Diagram I) are formally
available to the oxirane radical generated by o—hydrogen
abstraction. Examples of some of these possibilities have

been reported,26’27

while others are unknown and improbable.
Hydrogen abstraction at the B-position of cyclohexene oxide
has also been observed.28

A true thermal unimolecular decomposition involving

rupture of an oxirane bond is a rare phenomena, a likely

example being Pulegone oxide.

 

l6

Diagram I

R o R' R o R' R o
\ / \ / \ / / \' || ,
/C—C\ A.l /c—-c —' c—C—R
R" H R" / R"
o R/ R' o
\ /*' l u o
R—C—C R—C —C« R
o \o . \H n '
R" R" c- c ——R
1’
l R"
0
ll .
R—c —c-R' R' 8 R'
1'1" R-c-C-H R’C'H
I I
l R" R"
o H

II I
R—C — c—R'

R"

Catalysis by acidic sites on the surface of pyrex

glass may be suSpected when the products parallel acid

catalyzed products, e.g., equation 1.29

. CH
BF3 ::O
—-+

O
w270 , UNHCH

0
II.
CH

::O 2:0
-—————o

 

3

 

17

Extreme care must therefore be taken in assuring
cleanliness of the apparatus used.

In our work four epoxides (I-IV) were investigated
at high temperatures (T > 150°), and the decomposition

products identified.

\/\/ \/\/ \/\/ \C/ \C/

/C—\ /C_\ /C_—'C /C C\

H H H U CH3 [5 I1 {5
I II III IV

Each epoxide was investigated as a neat sample and
with the free radical initiators tertiary butyl hydroperoxide
and 2,4,4-trimethy1-2-formylcyclopentanone. Surface reactions
were identified by observing the effect of added Pryex

glass.

 

RESULTS

I. Sytrene Oxide

30studied reactions of

Ti ffeneau and co—workers
styrene oxide (I) with a variety of acid catalysts. They
found that distillation in the presence of a catalyst
(e.g., A1203) produced phenylacetaldehyde.

In beginning our work, styrene oxide was distilled
(b.p. >2000) from a Pyrex vessel in the presence of air;

a thin layer chromatograph (T.L.C.) of the distillate
demonstrated that no reaction had occurred.

Considering the possible radical and thermal reactions
of styrene oxide (e.g., diagram I, see page 16), one arrives
at phenylacetaldehyde, acetophenone and toluene as the most
likely products from thermal decomposition. With the
exception of toluene these were distinguished from each
other and from styrene oxide by the T.L.C. technique employed
here. Toluene is easily recognized by vapor phase chroma—
graphic analysis.

The infrared spectra and T.L.C. of the high temper—
ature reaction products from I showed that acetophenone was
not present among the products: acetophenone was found to
be stable under the reaction conditions. Consequently a
reliable quantitative method for following the decomposition

of I need only encompass the starting material and

18

 

l9

phenylacetaldehyde. For this purpose several experimental

procedures were eliminated.

(a) Gas chromatographic procedures while valuable for
detecting toluene could not be used as.a general purpose
tool, since the epoxide decomposed on the column (toluene
is not a product of this decomposition).

(b) Ultraviolet spectrophotometry was likewise undesirable

since peak overlapping occurred.

 

Fortunately, infrared spectra showed a characteristic
absorption for phenylacetaldehyde at 5.78 u and styrene
oxide at 11.43 H: thus permitting determination of the
percentage of I which had gone to recognizable product
(i.e., phenylacetaldehyde) and the percentage to other
product (i.e.. tars).

The plot of percent transmittance, at the respective
absorbancies, gs, molar concentration of the pure materials
in question gave smooth curves (see Figures I and II).

An infrared spectrum of a weighed amount of decom-
position sample was taken. The peak heights of reactant
and product were measured, and their corresponding concen-
trations determined.

Table I shows the results of the thermal and free
radical reactions of styrene oxide as followed by infrared

spectrophotometry.

56‘
52
49‘
‘4}
40'

36‘

28!

% Transmittance at 5.78 u

24‘

2.0V

 

20

 

 

Figure I.

T“ 7

'7

.1! .03 .04 ,3; .Jg 3'07 .3? '.J'9'.p“"'.II ' 712. J:- .74

moles/liter of phenylacetaldehyde

% Transmittance at 5.78 u gs. moles/liter of
phenylacetaldehyde(Perkin Elmer Model 21).

21

S3

2555"?

% Transmittance at 11.43 u
5'

 

 

 

 

.3: .3: .03 .‘m .3: .36 .57 o‘os foo Jo J: 4'2. .33 .\'4

moles/liter of styrene oxide

Figure II. % Transmittance at 11.43 u gs. moles/liter of
styrene oxide (Perkin Elmer Model 21).

22

Table I. Thermal and free radical reactions of styrene oxide.

 

 

Furnace

 

 

Time Sample . . a o
Temggrature (hrs.) Condition Additives A UCHZCHO
170° 24 neat none 6.1
200° 6 neat air *
2000 6 10% decalin none b
200° 24 neat none 7.7
200° 24 neat 100% increase of 21.7

surface area
230° 4 neat none 14.4
230° 5 neat none 17.7
2300 78 neat none *
170° 16 neat c 9.1
170° 18 neat c 19.1
200° 6 neat c 6.8
200° 6 neat 5x increase of c 16.3
2000 12 neat c 15.9
2000 16 neat c 22.5
200° 18 neat c *
230° 1 neat c 10.8
230° 2 neat c 28.1
2300 4 neat c *
170° 16 neat d 13.2
170° 18 neat d 23.4
200° 6. neat d 9.5
2000 12 neat d 17.5
200° 16 neat d 34.7
2000 18 neat d *
230° 1 neat d 31.4
2300 2 neat d 35.2
2300 4 neat d *
aUnless otherwise stated142%.
b

A qualitative analysis was made, starting material
and phenylacetaldehyde were the only substances present.

cTertiarybutyl hydroperoxide.

d2,4,4-trimethy1-2-formylcyclopentanone.

*
The percentage of product varied and the concen-
trations of tar were considerable.

IS

23

II. Trans-Stilbene Oxide

House et. 21,31 have studied the acid catalyzed

rearrangement of trans-stilbene oxide (II) with various Lewis
acids. When treated with BFB-etherate, II formed diphenyl-
acetaldehyde. Small amounts of deoxybenzoin were found
accompanying the aldehyde when magnesium bromide and zinc

bromide were the Lewis acid catalysts. Tiffeneau et- al.32

reported that the distillation of Egaggestilbene oxide
gave benzaldehyde and trans-stilbene: under acid conditions
deoxybenzoin and diphenylacetaldehyde were observed.

II was distilled (b.p. > 250°) in this investigation
and only starting material was recovered.

The use of infrared spectroscopy for following the
decomposition of II had to be abandoned, because the percent
yield of recognizable product (i.e., deoxybenzoin and
diphenylacetaldehyde) was low compared to the amount of tarry
uncharacterizable material, thus producing a large error.

Nuclear magnetic resonance was a useful tool, the
methine hydrogens of II appearing at 8.171, the methylene
hydrogens of deoxybenzoin at 7.9T and the methine hydrogen
of diphenylacetaldehyde at 7.651. Therefore one may
determine which substances were present and their relative
amounts.

Table II shows the results of the thermal and free

radical reactions of trans—stilbene oxide, as followed by

nuclear magnetic resonance spectroscopy.

24

 

 

 

Table II. Thermal and free radical reactions of trans-
stilbene oxide.
o
H
s
u c
S .2 a .o
o u -H m
04 --I O 'O
E A '0 N I>1
Q)" 0 C: m G Hug
510 m o m m >1m
o u U o D at:
m" S > a mr4
o ~r m -H x r:m
m H u o m+J
: U Q -H m H o
*4 .5 s :3 '° W:
a E m d 32 3%
230 c
200° 24 neat none 0 0
230° 12 neat none 2 3 0
2300 6 neat benzoic acid 3 < 10
2300 12 neat air 11 0
250° 12* neat none < 3 < 3
200° 24 neat d 0 0
230° 12 neat d E 3 < 3
2500 12* neat d 3 < 3
200° 24 neat e 0 0
230° 12 neat e 2 3 E 3
250° 12* neat e 10 10

 

aUnless otherwise stated..2%.

bWhen diphenylacetaldehyde was heated at 2000 for
six hours a gummy unrecognizable material was obtained.

CWhen p-toluenesulfonic: acid was used as a catalyst,
a very viscous material was obtained after degassing:
which showed no epoxide band in the infrared.

dTertiaryLbutyl hydroperoxide.

e2,4,4-trimethy1-2-formylcyclopentanone.

25

III. o—Methyl-Trans-Stilbene Oxide

Tiffeneau33 reported that distillation of a-methyl-
‘tgagg-stilbene oxide (III) in air gave a,o-dipheny1propion—
aldehyde, and that reaction of III with hydrochloric acid
produced methyldeoxybenzoin.

When III was distilled in air (b.p. > 300°) in our
laboratory and the product examined by T.L.C., three spots
(1,2-dipheny1ally1 alcohol, methyldeoxybenzoin and
o,o-diphenylpropionaldehyde) appeared in addition to the
starting material. This is a good illustration of the
tendency in early investigations of identifying and reporting
only the major product.

When III is treated with BF3-etherate for one minute
at room temperature, T.L.C. established that quantitative
conversion to a,o-dipheny1acetone and o,oedipheny1propronalde-
hyde had occurred.

Considering reaction by o-hydrogen abstraction from
III (see diagram I, page 16) three possible products are
methyldeoxybenzoin, a,o-diphenylpropronaldehyde and
o,o—dipheny1acetone.

In View of Gritter's findings regarding B-hydrogen
abstraction28 the possibility of encountering 1,2-dipheny1a11y1
alcohol should also be considered.

o—Methylatgagg-stilbene oxide and the four likely
products (i.e., methyldeoxybenzoin, o,ordipheny1acetone,
o,a-diphenylpropionaldehyde and 1,2-dipheny1a11y1 alcohol)

are easily resolvable on thin layer chromatography, making

 

26

this a useful tool for the qualitative identification of the
products from the thermal and free radical reactions of
III.

Quantitative analysis of the reaction mixture was
by nuclear magnetic resonance spectroscopy. The chemical
shifts (T) for the characteristic protons used for analysis
are given in Table III.

Table III. The chemical shifts (T) for the characteristic
protons used for analysis of a-methyl-trang-

stilbene oxide and its likely decomposition
products.

 

 

 

 

Compound Chemical Shift values (T)
I?! 0 H
\ /
(a) c —c< 8.60
/
CH3 0
H P.
(b) ¢2—‘ c — c—CH3 7.90
3 EL
I
(c-) 9! —- c -- c -—-U 8.50(d)
H
(d) CH3—c-UZ— 3H 8.05
(e) UCH(-0H)c¢=CH2 . 4.62 and 4.72

 

Table IV shows the findings of the thermal and free

radical reactions of o-methyl-trans—stilbene oxide.

27

. m
.omEHom mm3 mmmE >Hnmu m .mun em you coaxoumoono_£HXD o mamaunmfio

 

 

 

 

   

ooom 0m emumme was Honouam asaamasememflcuw.a ewes mcoemucmsoflosu
A m -Hseuomnmuassumeflunue.m.mn
.mun em you ooom um magnum an on

ocsom ouo3 osoumomahcmsmflold.o Dom :Housonzxomoahnumzo .ummc can UHDB macauummn Haflm
O
om .ea we .em as .om s .OH 0 .o .mue «m
ooom
n
«a .0m em .om Hm .w mm .He .mu: m
comm
Q
mm .om mm .ma om .em mm .SN .mue em
ooom
. 0:02
a .ma m .NH ma mo .mo .mu: m
comm
mmmam omnmn3om
as .mm Hm .me m .mH as o .o .mun em
ooom
. 0:02
om .ma em .mm Hamsm o .0 me .mm .mue em
ooom

4 mm mac n mmo
_
u I o I o In «no N u .m/I m mmouolonmu no I 0 Ian 133032
m . : . _. _ ._ DEHB
so 0 m u .e o m o m o u

o o Aoov ousumnoofima

30303 no 33.» x

 

 

mucofluaocoo

 

.opflxo wchHHumnmamuulamsumEIo mo mcofluumou HMUAUMH mmum one Hmfiumsu one .>H magma

28

IV. Tetraphenylethylene Oxide

Beaninacalone was the only product obtained when
tetraphenylethylene oxide (IV) was treated with acid.34
Since the aromatic hydrogen atoms in tetrapheny1ethy1ene
oxide are relatively unreactive to radical attack, this
substrate affords an opportunity to study thermal rearrange-
ment in the absence of interfering; radical reactions.
Benzpinacalone was the only product found in the
thermal decomposition of IV, using infrared spectrophotometry,
nuclear magnetic resonance spectrometry and thin layer
chromatography for analysis.
In one case beaninacalone was actually isolated
and identified by classical methods.
The decomposition of IV was followed by nuclear magne-
tic resonance spectrometry. A multiplet at 2.821 repre-
sents resonance by the twenty aromatic protons of tetraphenyl-
ethylene oxide. The nuclear magnetic resonance spectrum
of beaninacalone shows a characteristic resonance at
2.67m. This peak is due to the fifteen aromatic protons
B to the carbonyl. The five o—aromatic protons appear at
lower field.
Table V shows the results of the thermal reactions
of tetrapheny1ethy1ene oxide as followed by nuclear magnetic

resonance spectroscopy.

 

29

Table \h The thermal reactions of tetraphenylethylene oxide.

 

 

% benzpinacalone produced
in neat samples of tetra-
phenylethylene oxide.

 

 

 

ngggigiure Regigéon Initiators
(0C) (hrs.) None a b
2000 C, d
5 74.8
12 , 84.5
24 > 90 > 90 > 90

 

aTertiarybutyl hydroperoxide

b2,2,4-trimethy1—2-formylcyclopentanone

CIncreasing the surface area by the addition of

powdered Pyrex glass gave complete conversion to benzpinacalone
after one hr. at 200°.

When methylaniline was added with the powdered glass.
complete conversion to benzpinacalone was noted in two hrs.

DISCUSSION

Recently Russell gt..al.35 have shown that electron-

withdrawing substituents deactivate and electron-supplying
substituents activate substituted styrenes in addition
reactions involving peroXy radicals. In contrast, the Iowa
State group also found that the autoxidation of benzyl
phenyl ethers and benzyl alkyl ethers which involves peroxy
radicals is insensitive to polar effects of substituents
in either the phenyl or benzyl group. It was suggested
that the oxygenzatom acts as a powerful electron supplying
group and overshadows any effects of polar substituents.

The oxirane system is an ether and its oxygen atom
should also be capable of acting as an electron supplying
group. In fact, the findings of Walling20 pertaining to the
chlorination of propylene oxide (see page 13, iii) show
that oxirane hydrogen abstraction is favored over methyl
hydrogen abstraction.

That the activation due to the oxirane-oxygen
may not be as great as that observed in unstrained ethers
may be seen in the reluctance of radical initiators to
abstract a hydrogen atom from trans—stilbene oxide (Table II)
and the absence of acetophenone in the radical reactions of
styrene oxide (Table I). Apparently, the electron with-

drawing phenyl group overcomes the activating effect of the

30

31

oxygen in these instances. Since the polar effects studied

by Russell t al.35 were due to substituents in the para

position of the benzyl and phenyl groups, his data may not
apply here where a phenyl group is being substituted for a
hydrogen atom in the different oxiranes.

Gritter gt .al.28 have reported that cyclohexene
oxide undergoes B—hydrogen abstraction in a free radical

chain reaction. This mode of reaction becomes possible

in the case of a-methyl-trans-stilbene oxide (III); and

 

in fact, thermal decomposition of III in the presence
of radical initiators gave large amounts of 1,2-dipheny1—
allyl alcohol.

In addition to the radical chain reactions described
above, surface reactions seem to be occurring in some of the
more highly substituted oxiranes(i.e., III and IV). If we
accept Rondestvedt's25 hypothesis suggesting the presence of
Lewis acid sites on the surface of silica-—said sites
being responsible for the surface reactions--the formation
of intermediates B and C (Diagram II) follows logically.

These intermediates may aid in rationalizing the course of

the surface reaction.

.,..sp... unu
1 I
R1 9 E R1 0 P
\C/ \C/ \E—IC/

 

/ \ ,/ \

R2 H A R2 H _13
MSN
0
R I5
l | +
\ c___c/
.// \\
Diagram II R2 H -9

32

An important fact is that III and IV show significant
surface reaction (at 200°) while I and II do not. I and II
give rise to intermediate ions of lower stability than those
derived from III and IV; therefore this aspect of the hypo-
thesis agrees with the observed reactivities.

IV yields only one possible carbonium ion intermediate,
which upon formation may collapse and return to reactant.
or undergo phenyl migration to beaninacalone. Since the
reaction goes rapidly and in high yield, one must conclude
that the intermediate ion persists long enough to twist into

a higher energy conformation (i.e., IV C) favorable to a

 

 

1,2-pheny1 shift. -S-~
MSW MS 0
g T. I
0 I3 A 0 I?! .
\C/-'\C/ _. \€§_é/ _-_- :5 v a
0/ \Il ra/ \91 9’ P
JI‘
IVB MS“
I I?!
WISM
(9 l0 +""
l
{53— c—C—fi ’5 F
I5
IVC

It is assumed here that optimum geometry (i.e., an
orientation of the migrating phenyl group as in equation vii)

is a critical factor in the occurrence of rearrangement, and

33

that conformation IVC is subject to severe steric interaction

of a non-migrating phenyl group with the surface.

IVB

The surface reaction of III is slower than that
observed for IV. Consideration of the possible intermediates
from III (i.e., IIIB and IIIC), permits one to eliminate

1118 as an important contributor (note the low reactivity

of II).
a 0 H \ I1— -0 H
,/” GB
\ c/_:.\c< / \ t— c/ IIIB
CH3/ 0 CH3/ \I5
\x MSW
I
g' 69 H
\c I/ .
CH3/ \0 IIIC

In order for IIIC to be an effective intermediate in
rearrangement, it must persist long enough to convert to the
high energy conformations (IIID and IIIE) necessary for

l I Z'Shifts 1

 

Since IIIE should be more stable than IIID (fewer
and less severe non-bonded interactions) one would expect
_to find ova-diphenylpropionaldehyde among the major products
rfrom the reaction if rearrangements are occurring. One way
to account for the fact that this aldehyde is not a product
in the high temperature surfaCe reaction of III is to assume
that the'mean life spaniof intérmediate ion IIIC is so
short that cohversion to the high energy confirmations IIID
and IIIE is improbable. If ion IIIC is not undergoing
rearrangement one of the few alternate reaction paths is
the elimination of a proton to produce an olefin. Indeed
the neat reactions of III produce 1,2;dipheny1ally1 alcohol
in relatively good yield, by a surface reaction (equation Viii.
path a). The other major product. methyldeoxybenzoinfl
may also be accounted for”by{ap§oton loss (equation Viii.

path b).

w S 51—“. W s W W SW
.' I I
0 0 0
I + I b F I g
CHZ—c—c—IJ a cH2_.c_c——IJ CH3... -__c_
I I I a It) I I
I! H H - 9!, H I3
3.3 {.5
B B
(viii)

We make here the reasonable assumption that the
transition state geometry of E1 elimination is less
demanding than that for migration.

Another explanation for the products found in the
decomposition of III would parallel current views of the

36 That is to say. the

aldehyde-ketone rearrangement.
aldehyde is the primary product of the thermal reaction as

is predicted from consideration of ions IIID and IIIE.

but the aldehyde is not stable under the conditions of
thermalysis and is converted to methyldeoxybenzoin.
iUnfortunately, we have not been able to obtain this aldehyde
in sufficient quantitY'to test this fact.) However, experience
with other aldehydes suggests that a rapid clean thermal
rearrangement of this aldehyde to methyldeoxybenzoin is
unlikely.

No evidence of a true unimolecular thermal reaction

was found in any of the oxiranes investigated.

EXPERIMENTAL

I. General Procedures

The experimental procedure employed to prepare samples
was similar to that used by C. K. Johnson in his study of

some a, E-epoxy ketones.37

A. Apparatus

The infrared spectra were obtained on a Perkin-
Elmer Model 21 or a Beckman I.R. 7 recording spectophotometer
using sodium chloride cells. The band positions were re-
corded in microns. The ultraviolet spectra were determined
in 1-cm. quartz cells using a Cary Model 11 spectrophotometer.
Proton nuclear magnetic resonance spectra of samples in
carbon tetrachloride solution were determined using a Varian
A-60 high resolution spectrometer. All nmr spectra were
obtained at 60 Mo using tetramethylsilane as an internal
standard.

'Vapor phase chromatography analyses were made using
a l/4-inch, 6-foot column of 20% silicon on chromasorb.
Thin layer chromatographic analyses were carried out on
glass plates with silica gel G, with a 30:70 pentane-benzene

mixture as the eluting solvent.

36

37

B. Furnace

The sample vials were heated in an aluminum block
furnace. A temperature constant to within i 20 was main-

tained.

C. Preparation of Pryex Sample Vials

For all sealed tube reactions, the tubes were pre-
pared in the following way. Pyrex tubing was cut and the
sections placed in a large beaker. The glass was covered
with concentrated nitric acid and heated on a steam bath
for 24 hours. The acid was removed and the tubes rinsed
thoroughly with distilled water. Concentrated ammonium
hydroxide was added and the tubes allowed to stand for 30
minutes. The ammonium hydroxide was removed, the tubes
thoroughly rinsed with distilled water, and finally baked
in an oven at 2000 for 24 hours. The tubes were converted
to vials which were heated for several hours at 2000 and
then stored in a desiccator over Drierite.

Powdered pyrex tubing, used to increase surface
area in the vials, was treated in the same manner as the

tUbing used to make the vials.

D. Degassing

Degassing is a procedure whereby all oxygen is
removed from a system and the vial is sealed under a nitro-
gen atmosphere. The degassing procedure used was:

1. The system containing the vial is evacuated by

means of a vacuum pump.

38

'2. A three-way_stopcock is adjusted to cut the vacuum
off and to allow dry nitrogen to flow into the vial.

3. The contents of the vial are frozen, the nitrogen
is cut off and the system is once again opened to
the vacuum pump."

4. The contents of the vial are allowed to warm tO‘
room temperature.* (Steps one through four are
repeated five times.

5. Under a nitrogen atmosphere, the tube is sealed
as close to the material in the vial as is feasible,
to insure a liquid phase reaction at-high temperature.
*At this point. it is best if the material in the

vial is a liquid so all the oxygen can be eliminated.’ If
'the Sample is a solid. the temperature required~for liquifi-
cation under these conditions may cause (a) decompOsition-of
free radical initiators and/or (b) partial reaction of the

sample.

Compound I.
Styrene oxide is a liquid and degassing proceeded

smoothly.

Compound II.

trans—Stilbene oxide is a solid melting at 68-690.

a. trans-Stilbene oxide was dissolved in decalin
and degassed in solutions This method was discarded because
of the difficulties in the degassingiprocedure.

b. ‘tgggs-Stilbene oxide was melted with the aid of

a hot water bath and the complete degassing sequence was

39

performed. The system was then opened to the atmosphere,
the initiators added, and the degassing sequence repeated

without heating.

Compound III.
o-Methyl-trans—stilbene oxide is a solid melting
at 38.5-400. At 400 no decomposition occurred and the

initiators were not eliminated.

Compound IV.

Tetraphenylethylene oxide is a solid melting at
209—2100. The system was degassed without heating since

melting would require high temperatures.

E. Nitrogen

Oil pumped nitrogen was purified by passing the gas

through Fieser’s solution38 before use.

F. Handling of Samples

In all cases, except where otherwise noted, 0.1 grams
of sample was used with one drop of initiator and the sample
was degassed. After the prescribed time, the sample was
removed from the furnace and quenched in dry ice. The
vials were not opened until the spectra were to be run.

In each case two identical samples were run in order to

check reproducibility.

40

II. Synthesis of Oxiranes

Compounds 1,39 11,40 111,41 and IV’42 have been

previously reported in the literature, and their synthesis

was carried out according to these procedures.
III. 1,2-Dipheny1ally1 Alcohol

Brown and Brown43 report the preparation of 1,2—
diphenylallyl alcohol by lithium aluminum hydride reduction
of 2-phenylcinnamic acid.

A nuclear magnetic resonance spectrum of the allyl
alcohol (T = 3.0 (m), 3.5 (t), 5.75 (d) and 7.68 (S)
which is concentration dependent) indicates it to be 2,3-
diphenylallyl alcohol rather than 1,2-dipheny1ally1 alcohol.

Lithium aluminum hydride reduction of methylene-
deoxy benzoin yielded a compound which has a nuclear
magnetic resonance spectrum [T = 2.93 (m), 5.67 (d) and
6.95 (S) which is concentration dependent], ultraviolet

EtOH

absorption spectrum (xmax t: 240 mu), thin layer chromato—

graphy retention time and an infrared spectrum which are
different from the compound reported by Brown and Brown.43
The physical evidence indicates this compound
(i.e., the reduced product of methylenedeoxybenzoin)
to be 1,2-diphenylally1 alcohol.
Ingold's method of allyl alcohol isomerization44
was attempted by Brown and Brown43 on their apparent
2,3-dipheny1ally1 alcohol and in this work on 1,2-

diphenylallyl alcohol, no evidence of isomerization was

noted.

41

Five grams of methylenedeoxybenzoin45 in 25 m1. of
dry ether was added slowly to a cooled stirred solution
of 0.95 gms. of lithium aluminum hydride in ether. The
mixture was allowed to warm to room temperature, then with
continued stirring, refluxed for three hours. A saturated
solution of soidum sulfate was added to the cooled flask,
the mixture was then filtered. The filtrate was taken up
in ether and dried over anhydrous magnesium sulfate. The
residue was extracted with methanol. The methanol extracts
were evaporated and the remaining material taken up in
ether to remove any inorganic material. The combined
ether extracts were stripped of solvent and an oily compound
was obtained. This substance had a retention time on thin

layer chromatography, using silica gel G and 30:70 of HCCl

3:
benzene as the eluting solvent, different from the substance
reported by Brown and Brown,43 Afiggfl e: 240 mu compared to
Afing = 257 mu reported by Brown and Brown.43 Nuclear

magnetic resonance showed a doublet (T = 4.62 and 4.72)
which had an area ratio of 3:1 compared to the hydroxyl

hydrogen (T = 6.95 whicn is concentration dependent).

SUMMARY

None of the oxiranes (i.e., I—IV) studied showed
a true unimolecular thermal reaction.

Some of the earlier work reported on these oxiranes
were reinvestigated and their results were shown to be
ambiguous.

Substitution of phenyl groups on the oxirane
ring retards radical abstraction of benzylic hydrogens.

The products of surface reactions were explained
in terms of mild acid catalysis by Lewis acid sites on the

Pyrex glass-surface.

42

10.

ll.

12.
13.
14.

15.

16.

LITERATURE CITED

P. S. Pinkey. G. A. Nesty. R. H. Wiley and C. S.
Marvel, J. Am. Chem. Soc.. 58, 972 (1963).

P. S. Pinkey, G. A. Nesty, D. E. Pearson and C. S.
Marvel. J. Am. Chem. Soc., 52, 2666 (1937).

C. S. Marvel, R. Mozingo and E. C. Kirpatrick.
J. Am. Chem. Soc., 61, 2003 (1939).

R. P. LinStead and A. L. Walpole. J. Chem. Soc.,
842 (1939).

I. N. Nazarov and I. I. Zaretskaya, Zhur. Obshchei
Khim., 22, 1558 (1959). (C.A...§&. 8751 (1960).)

M Levitz, D. Perlman and M. T. Bogert. J. Org. Chem..
.6, 105 (1941).

C. S. Marvel and W. L. Walton. J. Org. Chem., 1, 88 (1942).

R. P. Linstead and W. E. Doering, J. Am. Chem. Soc.,
.63. 1991 (1942) .

L. H. Schwartzman. J. Org. Chem., 15, 517 (1950).

C. Dyerassi, Ed., "gteroid Reactions," J. E. Starr,
Metal ammonia reductions of steroidal enones,
saturated ketones and ketols. Holden-Day, Inc.,
San Francisco. 1963, p. 299.

G. S. Stork and S. D. Darling, J. Am. Chem. Soc.,
(see ibid: 86, 1761 (1963) for complete article).

R. B. Woodward. J. Am. Chem. Soc., 64, 76 (1942).
A. E. Gillam and L. K. Evans. J. Chem. Soc., 815 (1941).
A. E. Gillam and T. F. West. J. Chem. Soc., 486 (1942).

L. F. Fieser and M. Fieser, "Steroids," Rheinhold,
New York. 1959, pp. 15-21.

Prepared by the method of Marvel gt. a1. reference 1
and purified by the method of Linstead e3, 31..
reference 2.

43

44
17. R. E. Parker and N. S. Isaccs. Chem. Rev., 59, 737
(1959).

18. H. 0. House and R. L. Wassen, J. Am. Chem. Soc.,
18,4394 (1956).

19. A. C. Cope and B. D. Tiffany, J. Am. Chem. Soc.. 73.
4158 (1951).

20. C. Walling and P. S. Fredricks, J. Am. Chem. Soc.,
84, 3326 (1962).

21. J. Reese, Ber., 15, 384 (1942).

22. K. Alder, F. Fleck and H. Lessenick, Ber., 20, 1709
(1957).

23. C. K. Johnson, Thesis, Michigan State University.
1963. p. 56.

24. Ibid., p. 28.

25. C. S. Rondestvedt, Jr., J. Am. Chem. Soc., 84, 3319
(1962).

26. C. K. Johnson, Thesis, Michigan State University.
1963.

27. T. J. Wallace and R. J. Gritter. J. Org. Chem.. 21.
3067 (1962).

28. E. C. Sabatine and R. J. Gritter. J. Org. Chem..
22. 3437 (1963).

29. C. K. Johnson, Thesis, Michigan State University.
1963. p. 38.

30. M. M. Fourneau and Tiffeneau. Compt. Rend.. l 0,
1595 (1905).

31. H. 0. House, J. Am. Chem. Soc.. 77. 3071 (1955).

32. M. Tiffeneau and J. Lévy, Bull. Soc. Chim., 32, 763
(1926).

33. M. Tiffeneau and J Lévy, Bull. Soc. Chim., 49, 1806
(1931).

34. H. J. Gebhart. Jr. and K. H. Adams, J. Am. Chem. Soc.,
19. 3925 (1954).

35. G. A. Russell and R. C. Williamson. Jr., J. Am. Chem.
Soc., gg, 2357 (1964).

36.

37.

38.

39.

40.

41.

42.

43.
44.

45.

45

J. W. Huffman and L. E. Browder. J. Org. Chem., 21,

3208 (1962).

Taken from the Ph.D. thesis of C. K. Johnson.

Michigan State University, 1963.

F. Fieser. "Experiments in Organic Chemistry."
Third edition, D. C. Heath and Company. Boston.
1957. p. 299.

Purchased from Matheson. Coleman and Bell.

N.

Rabjohn, Ed., "Organic Synthesis." Vbl. IV,
D. J. Reif and H. 0. House, John Wiley and Sons.
New York, 1963, p. 860.

Ogata and T. Tabushi. J. Am. Chem. Soc.,‘gg, 3440
(1961).

C. Cope, P. A. Trumbell and E. R. Trumbell.
J. Am. Chem. Soc., 89, 2 44 (1958).

R. Brown and P. E. Brown,.Tet. Let., 4, 191 (1963).

Burton and C. K. Ingold. J. Chem. Soc.. 904 (1928).

.Mehr, E. L. Becker and P. E. Spoerri, J. Am. Chem.

Soc.. 11, 984 (1955).